diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,38443 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 14.951603498542275, + "eval_steps": 500, + "global_step": 1600, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2606.0, + "completions/mean_length": 587.7176513671875, + "completions/mean_terminated_length": 536.0668334960938, + "completions/min_length": 22.0, + "completions/min_terminated_length": 22.0, + "epoch": 0.009329446064139942, + "grad_norm": 0.1598890870809555, + "learning_rate": 1e-06, + "loss": 0.003, + "num_tokens": 559851.0, + "reward": 0.5290178656578064, + "reward_std": 0.27369576692581177, + "rewards/verify_math_reward/mean": 0.5290178656578064, + "rewards/verify_math_reward/std": 0.49943605065345764, + "step": 1 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2699.0, + "completions/mean_length": 614.9989013671875, + "completions/mean_terminated_length": 543.6343994140625, + "completions/min_length": 15.0, + "completions/min_terminated_length": 15.0, + "epoch": 0.018658892128279883, + "grad_norm": 0.15343371033668518, + "learning_rate": 1e-06, + "loss": 0.0051, + "num_tokens": 1124418.0, + "reward": 0.4732142984867096, + "reward_std": 0.24653121829032898, + "rewards/verify_math_reward/mean": 0.4732142984867096, + "rewards/verify_math_reward/std": 0.4995608925819397, + "step": 2 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2975.0, + "completions/mean_length": 568.5892944335938, + "completions/mean_terminated_length": 536.8108520507812, + "completions/min_length": 63.0, + "completions/min_terminated_length": 63.0, + "epoch": 0.027988338192419825, + "grad_norm": 0.14869461953639984, + "learning_rate": 1e-06, + "loss": -0.0066, + "num_tokens": 1699650.0, + "reward": 0.515625, + "reward_std": 0.22965605556964874, + "rewards/verify_math_reward/mean": 0.515625, + "rewards/verify_math_reward/std": 0.5000349283218384, + "step": 3 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 549.2511596679688, + "completions/mean_terminated_length": 525.3404541015625, + "completions/min_length": 75.0, + "completions/min_terminated_length": 75.0, + "epoch": 0.037317784256559766, + "grad_norm": 0.16647283732891083, + "learning_rate": 1e-06, + "loss": -0.0016, + "num_tokens": 2246635.0, + "reward": 0.5390625, + "reward_std": 0.24555623531341553, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 4 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3643.0, + "completions/mean_length": 586.9765625, + "completions/mean_terminated_length": 547.371337890625, + "completions/min_length": 14.0, + "completions/min_terminated_length": 14.0, + "epoch": 0.04664723032069971, + "grad_norm": 0.14486758410930634, + "learning_rate": 1e-06, + "loss": 0.005, + "num_tokens": 2815174.0, + "reward": 0.5145089626312256, + "reward_std": 0.2665913701057434, + "rewards/verify_math_reward/mean": 0.5145089030265808, + "rewards/verify_math_reward/std": 0.5000685453414917, + "step": 5 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3903.0, + "completions/mean_length": 618.943115234375, + "completions/mean_terminated_length": 551.6962280273438, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 0.05597667638483965, + "grad_norm": 0.13758042454719543, + "learning_rate": 1e-06, + "loss": 0.0065, + "num_tokens": 3390539.0, + "reward": 0.5323660969734192, + "reward_std": 0.24333493411540985, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 6 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2265.0, + "completions/mean_length": 564.9498291015625, + "completions/mean_terminated_length": 521.06103515625, + "completions/min_length": 5.0, + "completions/min_terminated_length": 5.0, + "epoch": 0.0653061224489796, + "grad_norm": 0.1523338407278061, + "learning_rate": 1e-06, + "loss": -0.0124, + "num_tokens": 3950134.0, + "reward": 0.5223214626312256, + "reward_std": 0.24505950510501862, + "rewards/verify_math_reward/mean": 0.5223214030265808, + "rewards/verify_math_reward/std": 0.49978047609329224, + "step": 7 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3532.0, + "completions/mean_length": 633.7332763671875, + "completions/mean_terminated_length": 590.699462890625, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "epoch": 0.07463556851311953, + "grad_norm": 0.12796993553638458, + "learning_rate": 1e-06, + "loss": 0.0063, + "num_tokens": 4547295.0, + "reward": 0.512276828289032, + "reward_std": 0.24619702994823456, + "rewards/verify_math_reward/mean": 0.5122767686843872, + "rewards/verify_math_reward/std": 0.500128448009491, + "step": 8 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3821.0, + "completions/mean_length": 659.078125, + "completions/mean_terminated_length": 568.5292358398438, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 0.08396501457725948, + "grad_norm": 0.13264265656471252, + "learning_rate": 1e-06, + "loss": -0.0055, + "num_tokens": 5132597.0, + "reward": 0.535714328289032, + "reward_std": 0.2254144549369812, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 9 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3970.0, + "completions/mean_length": 606.78125, + "completions/mean_terminated_length": 563.4124145507812, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 0.09329446064139942, + "grad_norm": 0.128464475274086, + "learning_rate": 1e-06, + "loss": 0.0016, + "num_tokens": 5726457.0, + "reward": 0.5, + "reward_std": 0.22939391434192657, + "rewards/verify_math_reward/mean": 0.5, + "rewards/verify_math_reward/std": 0.5002792477607727, + "step": 10 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3976.0, + "completions/mean_length": 646.8995971679688, + "completions/mean_terminated_length": 584.1885986328125, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 0.10262390670553936, + "grad_norm": 0.10745050758123398, + "learning_rate": 1e-06, + "loss": 0.0094, + "num_tokens": 6324887.0, + "reward": 0.4821428656578064, + "reward_std": 0.1643511950969696, + "rewards/verify_math_reward/mean": 0.4821428656578064, + "rewards/verify_math_reward/std": 0.4999600946903229, + "step": 11 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1931.0, + "completions/mean_length": 592.1886596679688, + "completions/mean_terminated_length": 524.42431640625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "epoch": 0.1119533527696793, + "grad_norm": 0.11800146847963333, + "learning_rate": 1e-06, + "loss": 0.0057, + "num_tokens": 6888336.0, + "reward": 0.5368303656578064, + "reward_std": 0.17595963180065155, + "rewards/verify_math_reward/mean": 0.5368303656578064, + "rewards/verify_math_reward/std": 0.49892017245292664, + "step": 12 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3596.0, + "completions/mean_length": 626.4185791015625, + "completions/mean_terminated_length": 579.3201904296875, + "completions/min_length": 17.0, + "completions/min_terminated_length": 17.0, + "epoch": 0.12128279883381925, + "grad_norm": 0.1384453922510147, + "learning_rate": 1e-06, + "loss": -0.0084, + "num_tokens": 7488447.0, + "reward": 0.5424107313156128, + "reward_std": 0.25145965814590454, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763562679291, + "step": 13 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2343.0, + "completions/mean_length": 583.0848388671875, + "completions/mean_terminated_length": 519.213623046875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 0.1306122448979592, + "grad_norm": 0.13699589669704437, + "learning_rate": 1e-06, + "loss": 0.0028, + "num_tokens": 8041467.0, + "reward": 0.5569196939468384, + "reward_std": 0.2198163866996765, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 14 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4009.0, + "completions/mean_length": 631.1886596679688, + "completions/mean_terminated_length": 580.1777954101562, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 0.13994169096209913, + "grad_norm": 0.11382047086954117, + "learning_rate": 1e-06, + "loss": 0.006, + "num_tokens": 8647228.0, + "reward": 0.5424107313156128, + "reward_std": 0.19178421795368195, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763562679291, + "step": 15 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 608.078125, + "completions/mean_terminated_length": 560.7307739257812, + "completions/min_length": 17.0, + "completions/min_terminated_length": 17.0, + "epoch": 0.14927113702623906, + "grad_norm": 0.11619266122579575, + "learning_rate": 1e-06, + "loss": 0.0076, + "num_tokens": 9234378.0, + "reward": 0.5446428656578064, + "reward_std": 0.1831049621105194, + "rewards/verify_math_reward/mean": 0.5446428656578064, + "rewards/verify_math_reward/std": 0.49828118085861206, + "step": 16 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2097.0, + "completions/mean_length": 586.5346069335938, + "completions/mean_terminated_length": 554.9177856445312, + "completions/min_length": 44.0, + "completions/min_terminated_length": 44.0, + "epoch": 0.158600583090379, + "grad_norm": 0.11830838024616241, + "learning_rate": 1e-06, + "loss": 0.0244, + "num_tokens": 9812769.0, + "reward": 0.5680803656578064, + "reward_std": 0.21124452352523804, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 17 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3639.0, + "completions/mean_length": 641.7879638671875, + "completions/mean_terminated_length": 562.9246215820312, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 0.16793002915451896, + "grad_norm": 0.12457925081253052, + "learning_rate": 1e-06, + "loss": 0.0153, + "num_tokens": 10387123.0, + "reward": 0.543526828289032, + "reward_std": 0.21065111458301544, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 18 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4090.0, + "completions/mean_length": 593.310302734375, + "completions/mean_terminated_length": 557.7699584960938, + "completions/min_length": 2.0, + "completions/min_terminated_length": 2.0, + "epoch": 0.1772594752186589, + "grad_norm": 0.14370805025100708, + "learning_rate": 1e-06, + "loss": 0.0007, + "num_tokens": 10976321.0, + "reward": 0.5892857313156128, + "reward_std": 0.274520605802536, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 19 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3708.0, + "completions/mean_length": 574.005615234375, + "completions/mean_terminated_length": 530.2293701171875, + "completions/min_length": 11.0, + "completions/min_terminated_length": 11.0, + "epoch": 0.18658892128279883, + "grad_norm": 0.12871408462524414, + "learning_rate": 1e-06, + "loss": 0.0069, + "num_tokens": 11536742.0, + "reward": 0.6205357313156128, + "reward_std": 0.2054648995399475, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 20 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4067.0, + "completions/mean_length": 589.8058471679688, + "completions/mean_terminated_length": 521.9954223632812, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 0.19591836734693877, + "grad_norm": 0.154408261179924, + "learning_rate": 1e-06, + "loss": 0.0099, + "num_tokens": 12077768.0, + "reward": 0.6026785969734192, + "reward_std": 0.17671848833560944, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 21 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 634.7444458007812, + "completions/mean_terminated_length": 591.7232055664062, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 0.20524781341107873, + "grad_norm": 0.10341142117977142, + "learning_rate": 1e-06, + "loss": 0.0198, + "num_tokens": 12696875.0, + "reward": 0.5301339626312256, + "reward_std": 0.17517825961112976, + "rewards/verify_math_reward/mean": 0.5301339030265808, + "rewards/verify_math_reward/std": 0.49936988949775696, + "step": 22 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3575.0, + "completions/mean_length": 624.786865234375, + "completions/mean_terminated_length": 561.673828125, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 0.21457725947521866, + "grad_norm": 0.12730124592781067, + "learning_rate": 1e-06, + "loss": -0.0083, + "num_tokens": 13275484.0, + "reward": 0.621651828289032, + "reward_std": 0.20564965903759003, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 23 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2504.0, + "completions/mean_length": 621.6517944335938, + "completions/mean_terminated_length": 570.5005493164062, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 0.2239067055393586, + "grad_norm": 0.11972032487392426, + "learning_rate": 1e-06, + "loss": 0.02, + "num_tokens": 13863532.0, + "reward": 0.5089285969734192, + "reward_std": 0.21004310250282288, + "rewards/verify_math_reward/mean": 0.5089285969734192, + "rewards/verify_math_reward/std": 0.5001994967460632, + "step": 24 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4090.0, + "completions/mean_length": 671.953125, + "completions/mean_terminated_length": 593.7785034179688, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 0.23323615160349853, + "grad_norm": 0.1088431179523468, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 14468914.0, + "reward": 0.5390625, + "reward_std": 0.18656732141971588, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 25 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 632.453125, + "completions/mean_terminated_length": 557.4161987304688, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 0.2425655976676385, + "grad_norm": 0.12452881783246994, + "learning_rate": 1e-06, + "loss": 0.0028, + "num_tokens": 15045240.0, + "reward": 0.5055803656578064, + "reward_std": 0.193512424826622, + "rewards/verify_math_reward/mean": 0.5055803656578064, + "rewards/verify_math_reward/std": 0.5002480745315552, + "step": 26 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2228.0, + "completions/mean_length": 597.177490234375, + "completions/mean_terminated_length": 557.6873779296875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "epoch": 0.2518950437317784, + "grad_norm": 0.12795555591583252, + "learning_rate": 1e-06, + "loss": 0.0071, + "num_tokens": 15627471.0, + "reward": 0.574776828289032, + "reward_std": 0.2006830871105194, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 27 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2195.0, + "completions/mean_length": 565.5256958007812, + "completions/mean_terminated_length": 541.7247314453125, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 0.2612244897959184, + "grad_norm": 0.1300457864999771, + "learning_rate": 1e-06, + "loss": 0.0192, + "num_tokens": 16199598.0, + "reward": 0.543526828289032, + "reward_std": 0.21666006743907928, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 28 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2955.0, + "completions/mean_length": 611.7076416015625, + "completions/mean_terminated_length": 564.4095458984375, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "epoch": 0.2705539358600583, + "grad_norm": 0.12154428660869598, + "learning_rate": 1e-06, + "loss": -0.0021, + "num_tokens": 16787520.0, + "reward": 0.5055803656578064, + "reward_std": 0.1938129961490631, + "rewards/verify_math_reward/mean": 0.5055803656578064, + "rewards/verify_math_reward/std": 0.5002480745315552, + "step": 29 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3164.0, + "completions/mean_length": 557.9397583007812, + "completions/mean_terminated_length": 522.0405883789062, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "epoch": 0.27988338192419826, + "grad_norm": 0.1334248185157776, + "learning_rate": 1e-06, + "loss": 0.0164, + "num_tokens": 17345898.0, + "reward": 0.5546875, + "reward_std": 0.20459476113319397, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 30 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3644.0, + "completions/mean_length": 652.2611694335938, + "completions/mean_terminated_length": 605.5136108398438, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "epoch": 0.2892128279883382, + "grad_norm": 0.10987009853124619, + "learning_rate": 1e-06, + "loss": 0.0149, + "num_tokens": 17966052.0, + "reward": 0.5390625, + "reward_std": 0.19181953370571136, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 31 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2406.0, + "completions/mean_length": 593.5580444335938, + "completions/mean_terminated_length": 558.020263671875, + "completions/min_length": 5.0, + "completions/min_terminated_length": 5.0, + "epoch": 0.29854227405247813, + "grad_norm": 0.14382220804691315, + "learning_rate": 1e-06, + "loss": 0.0002, + "num_tokens": 18551856.0, + "reward": 0.5569196939468384, + "reward_std": 0.2547265291213989, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 32 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.004464285714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3736.0, + "completions/mean_length": 534.732177734375, + "completions/mean_terminated_length": 518.7623291015625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 0.30787172011661806, + "grad_norm": 0.12861202657222748, + "learning_rate": 1e-06, + "loss": 0.0108, + "num_tokens": 19101128.0, + "reward": 0.5245535969734192, + "reward_std": 0.2000095248222351, + "rewards/verify_math_reward/mean": 0.5245535969734192, + "rewards/verify_math_reward/std": 0.4996756613254547, + "step": 33 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3980.0, + "completions/mean_length": 619.2623291015625, + "completions/mean_terminated_length": 560.0670166015625, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 0.317201166180758, + "grad_norm": 0.13320207595825195, + "learning_rate": 1e-06, + "loss": 0.0014, + "num_tokens": 19677939.0, + "reward": 0.535714328289032, + "reward_std": 0.20632296800613403, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 34 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3976.0, + "completions/mean_length": 621.6049194335938, + "completions/mean_terminated_length": 574.4412231445312, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 0.32653061224489793, + "grad_norm": 0.12315374612808228, + "learning_rate": 1e-06, + "loss": 0.008, + "num_tokens": 20269401.0, + "reward": 0.5837053656578064, + "reward_std": 0.20899002254009247, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 35 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2104.0, + "completions/mean_length": 623.90625, + "completions/mean_terminated_length": 564.7900390625, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 0.3358600583090379, + "grad_norm": 0.13047567009925842, + "learning_rate": 1e-06, + "loss": -0.0014, + "num_tokens": 20859805.0, + "reward": 0.5658482313156128, + "reward_std": 0.2144019603729248, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 36 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3519.0, + "completions/mean_length": 609.5, + "completions/mean_terminated_length": 558.1698608398438, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 0.34518950437317786, + "grad_norm": 0.12048663944005966, + "learning_rate": 1e-06, + "loss": 0.0183, + "num_tokens": 21429397.0, + "reward": 0.559151828289032, + "reward_std": 0.192121222615242, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 37 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 603.1495971679688, + "completions/mean_terminated_length": 559.735595703125, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 0.3545189504373178, + "grad_norm": 0.1148107573390007, + "learning_rate": 1e-06, + "loss": 0.008, + "num_tokens": 22014099.0, + "reward": 0.5580357313156128, + "reward_std": 0.17678125202655792, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 38 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2729.0, + "completions/mean_length": 590.3125, + "completions/mean_terminated_length": 546.739013671875, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.3638483965014577, + "grad_norm": 0.13146887719631195, + "learning_rate": 1e-06, + "loss": 0.0206, + "num_tokens": 22593827.0, + "reward": 0.5524553656578064, + "reward_std": 0.20670093595981598, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 39 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2003.0, + "completions/mean_length": 568.5558471679688, + "completions/mean_terminated_length": 532.7643432617188, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 0.37317784256559766, + "grad_norm": 0.12316915392875671, + "learning_rate": 1e-06, + "loss": 0.0002, + "num_tokens": 23168741.0, + "reward": 0.5167410969734192, + "reward_std": 0.1902451068162918, + "rewards/verify_math_reward/mean": 0.5167410969734192, + "rewards/verify_math_reward/std": 0.4999987483024597, + "step": 40 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2972.0, + "completions/mean_length": 671.5558471679688, + "completions/mean_terminated_length": 593.3721313476562, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 0.3825072886297376, + "grad_norm": 0.11921396851539612, + "learning_rate": 1e-06, + "loss": 0.0009, + "num_tokens": 23790319.0, + "reward": 0.4977678656578064, + "reward_std": 0.22278764843940735, + "rewards/verify_math_reward/mean": 0.4977678656578064, + "rewards/verify_math_reward/std": 0.5002742409706116, + "step": 41 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 627.9442138671875, + "completions/mean_terminated_length": 552.8095703125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 0.39183673469387753, + "grad_norm": 0.11759841442108154, + "learning_rate": 1e-06, + "loss": 0.0072, + "num_tokens": 24366333.0, + "reward": 0.5223214626312256, + "reward_std": 0.1943764090538025, + "rewards/verify_math_reward/mean": 0.5223214030265808, + "rewards/verify_math_reward/std": 0.49978047609329224, + "step": 42 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3704.0, + "completions/mean_length": 588.7232666015625, + "completions/mean_terminated_length": 565.0786743164062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 0.40116618075801747, + "grad_norm": 0.12721504271030426, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 24955389.0, + "reward": 0.5546875, + "reward_std": 0.22973981499671936, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 43 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.004464285714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3066.0, + "completions/mean_length": 522.84375, + "completions/mean_terminated_length": 506.8206481933594, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.41049562682215746, + "grad_norm": 0.12136313319206238, + "learning_rate": 1e-06, + "loss": 0.022, + "num_tokens": 25493665.0, + "reward": 0.6049107313156128, + "reward_std": 0.1739308387041092, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 44 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2946.0, + "completions/mean_length": 621.4989013671875, + "completions/mean_terminated_length": 562.3416748046875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.4198250728862974, + "grad_norm": 0.12345373630523682, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 26082480.0, + "reward": 0.5680803656578064, + "reward_std": 0.22804690897464752, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 45 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2603.0, + "completions/mean_length": 667.021240234375, + "completions/mean_terminated_length": 596.7232666015625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 0.4291545189504373, + "grad_norm": 0.12356200069189072, + "learning_rate": 1e-06, + "loss": -0.0039, + "num_tokens": 26694851.0, + "reward": 0.5558035969734192, + "reward_std": 0.22623679041862488, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715372920036316, + "step": 46 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3786.0, + "completions/mean_length": 635.5546875, + "completions/mean_terminated_length": 592.5435180664062, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 0.43848396501457726, + "grad_norm": 0.12838619947433472, + "learning_rate": 1e-06, + "loss": 0.0261, + "num_tokens": 27304148.0, + "reward": 0.559151828289032, + "reward_std": 0.23304061591625214, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 47 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 613.497802734375, + "completions/mean_terminated_length": 542.1025390625, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 0.4478134110787172, + "grad_norm": 0.1245860904455185, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 27874762.0, + "reward": 0.5323660969734192, + "reward_std": 0.19607001543045044, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 48 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3030.0, + "completions/mean_length": 619.3839721679688, + "completions/mean_terminated_length": 568.1992797851562, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 0.45714285714285713, + "grad_norm": 0.12745042145252228, + "learning_rate": 1e-06, + "loss": 0.0226, + "num_tokens": 28459498.0, + "reward": 0.4966517984867096, + "reward_std": 0.21571393311023712, + "rewards/verify_math_reward/mean": 0.4966517984867096, + "rewards/verify_math_reward/std": 0.5002680420875549, + "step": 49 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3924.0, + "completions/mean_length": 663.5725708007812, + "completions/mean_terminated_length": 548.7623901367188, + "completions/min_length": 13.0, + "completions/min_terminated_length": 13.0, + "epoch": 0.46647230320699706, + "grad_norm": 0.13521279394626617, + "learning_rate": 1e-06, + "loss": 0.0049, + "num_tokens": 29037811.0, + "reward": 0.5457589626312256, + "reward_std": 0.2123749852180481, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981797933578491, + "step": 50 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3186.0, + "completions/mean_length": 645.3069458007812, + "completions/mean_terminated_length": 582.5670166015625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.47580174927113705, + "grad_norm": 0.12615244090557098, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 29631678.0, + "reward": 0.5837053656578064, + "reward_std": 0.21906569600105286, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 51 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2363.0, + "completions/mean_length": 580.3660888671875, + "completions/mean_terminated_length": 532.642578125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 0.485131195335277, + "grad_norm": 0.14009985327720642, + "learning_rate": 1e-06, + "loss": 0.0185, + "num_tokens": 30182222.0, + "reward": 0.6104910969734192, + "reward_std": 0.22289825975894928, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 52 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3577.0, + "completions/mean_length": 593.5256958007812, + "completions/mean_terminated_length": 537.9308471679688, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 0.4944606413994169, + "grad_norm": 0.12409912794828415, + "learning_rate": 1e-06, + "loss": 0.0229, + "num_tokens": 30742429.0, + "reward": 0.5613839626312256, + "reward_std": 0.20845939218997955, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 53 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3689.0, + "completions/mean_length": 531.890625, + "completions/mean_terminated_length": 487.5909729003906, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 0.5037900874635568, + "grad_norm": 0.12954339385032654, + "learning_rate": 1e-06, + "loss": 0.0077, + "num_tokens": 31263787.0, + "reward": 0.5613839626312256, + "reward_std": 0.18028958141803741, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 54 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0033482142857143016, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3588.0, + "completions/mean_length": 520.5625, + "completions/mean_terminated_length": 508.55096435546875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 0.5131195335276968, + "grad_norm": 0.1351110339164734, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 31812315.0, + "reward": 0.598214328289032, + "reward_std": 0.19579969346523285, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 55 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3240.0, + "completions/mean_length": 611.9096069335938, + "completions/mean_terminated_length": 556.6065673828125, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 0.5224489795918368, + "grad_norm": 0.12937261164188385, + "learning_rate": 1e-06, + "loss": 0.0257, + "num_tokens": 32394650.0, + "reward": 0.5055803656578064, + "reward_std": 0.22338427603244781, + "rewards/verify_math_reward/mean": 0.5055803656578064, + "rewards/verify_math_reward/std": 0.5002480745315552, + "step": 56 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3190.0, + "completions/mean_length": 560.9699096679688, + "completions/mean_terminated_length": 525.1014404296875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 0.5317784256559767, + "grad_norm": 0.12970662117004395, + "learning_rate": 1e-06, + "loss": 0.0044, + "num_tokens": 32947167.0, + "reward": 0.5725446939468384, + "reward_std": 0.1899854838848114, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 57 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 574.3660888671875, + "completions/mean_terminated_length": 538.633544921875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 0.5411078717201167, + "grad_norm": 0.12180390954017639, + "learning_rate": 1e-06, + "loss": -0.0028, + "num_tokens": 33514799.0, + "reward": 0.5580357313156128, + "reward_std": 0.21053025126457214, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689781665802, + "step": 58 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3807.0, + "completions/mean_length": 579.5089721679688, + "completions/mean_terminated_length": 511.4994201660156, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 0.5504373177842565, + "grad_norm": 0.13004224002361298, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 34058551.0, + "reward": 0.5680803656578064, + "reward_std": 0.18989993631839752, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 59 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2534.0, + "completions/mean_length": 567.296875, + "completions/mean_terminated_length": 511.2857360839844, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.5597667638483965, + "grad_norm": 0.12029591202735901, + "learning_rate": 1e-06, + "loss": 0.0025, + "num_tokens": 34594513.0, + "reward": 0.6026785969734192, + "reward_std": 0.15293002128601074, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 60 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3548.0, + "completions/mean_length": 631.0892944335938, + "completions/mean_terminated_length": 576.0906982421875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 0.5690962099125364, + "grad_norm": 0.1275714635848999, + "learning_rate": 1e-06, + "loss": 0.002, + "num_tokens": 35186649.0, + "reward": 0.5491071939468384, + "reward_std": 0.20741769671440125, + "rewards/verify_math_reward/mean": 0.5491071343421936, + "rewards/verify_math_reward/std": 0.49786055088043213, + "step": 61 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3602.0, + "completions/mean_length": 648.4866333007812, + "completions/mean_terminated_length": 601.6878051757812, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 0.5784256559766764, + "grad_norm": 0.12914977967739105, + "learning_rate": 1e-06, + "loss": 0.0104, + "num_tokens": 35802509.0, + "reward": 0.5267857313156128, + "reward_std": 0.25610974431037903, + "rewards/verify_math_reward/mean": 0.5267857313156128, + "rewards/verify_math_reward/std": 0.4995608627796173, + "step": 62 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3981.0, + "completions/mean_length": 567.466552734375, + "completions/mean_terminated_length": 535.6779174804688, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 0.5877551020408164, + "grad_norm": 0.13891862332820892, + "learning_rate": 1e-06, + "loss": 0.0165, + "num_tokens": 36356847.0, + "reward": 0.5881696939468384, + "reward_std": 0.22300700843334198, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 63 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2314.0, + "completions/mean_length": 584.021240234375, + "completions/mean_terminated_length": 552.3817749023438, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.5970845481049563, + "grad_norm": 0.13675697147846222, + "learning_rate": 1e-06, + "loss": -0.0053, + "num_tokens": 36937130.0, + "reward": 0.6037946939468384, + "reward_std": 0.200128972530365, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 64 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2379.0, + "completions/mean_length": 597.1864013671875, + "completions/mean_terminated_length": 557.6964111328125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 0.6064139941690962, + "grad_norm": 0.1162147969007492, + "learning_rate": 1e-06, + "loss": 0.0147, + "num_tokens": 37528929.0, + "reward": 0.5457589626312256, + "reward_std": 0.17048241198062897, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 65 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1626.0, + "completions/mean_length": 568.7388916015625, + "completions/mean_terminated_length": 508.683349609375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 0.6157434402332361, + "grad_norm": 0.12827178835868835, + "learning_rate": 1e-06, + "loss": 0.0108, + "num_tokens": 38075615.0, + "reward": 0.590401828289032, + "reward_std": 0.18054921925067902, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 66 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3799.0, + "completions/mean_length": 645.9855346679688, + "completions/mean_terminated_length": 603.1039428710938, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 0.6250728862973761, + "grad_norm": 0.1203441321849823, + "learning_rate": 1e-06, + "loss": 0.01, + "num_tokens": 38705034.0, + "reward": 0.5167410969734192, + "reward_std": 0.2141769379377365, + "rewards/verify_math_reward/mean": 0.5167410969734192, + "rewards/verify_math_reward/std": 0.4999987483024597, + "step": 67 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2966.0, + "completions/mean_length": 614.1015625, + "completions/mean_terminated_length": 550.7943115234375, + "completions/min_length": 70.0, + "completions/min_terminated_length": 70.0, + "epoch": 0.634402332361516, + "grad_norm": 0.12513791024684906, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 39273357.0, + "reward": 0.5502232313156128, + "reward_std": 0.18606878817081451, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 68 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3836.0, + "completions/mean_length": 659.1596069335938, + "completions/mean_terminated_length": 592.6905517578125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 0.643731778425656, + "grad_norm": 0.12893493473529816, + "learning_rate": 1e-06, + "loss": -0.003, + "num_tokens": 39884700.0, + "reward": 0.5167410969734192, + "reward_std": 0.23634636402130127, + "rewards/verify_math_reward/mean": 0.5167410969734192, + "rewards/verify_math_reward/std": 0.4999987483024597, + "step": 69 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3453.0, + "completions/mean_length": 644.1975708007812, + "completions/mean_terminated_length": 585.4268188476562, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 0.6530612244897959, + "grad_norm": 0.1353166550397873, + "learning_rate": 1e-06, + "loss": 0.0156, + "num_tokens": 40490501.0, + "reward": 0.5390625, + "reward_std": 0.2425432950258255, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 70 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2030.0, + "completions/mean_length": 615.7824096679688, + "completions/mean_terminated_length": 552.5056762695312, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 0.6623906705539359, + "grad_norm": 0.13289125263690948, + "learning_rate": 1e-06, + "loss": 0.0062, + "num_tokens": 41076290.0, + "reward": 0.5625, + "reward_std": 0.20632296800613403, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 71 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3932.0, + "completions/mean_length": 621.4989013671875, + "completions/mean_terminated_length": 566.3480834960938, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 0.6717201166180758, + "grad_norm": 0.12672410905361176, + "learning_rate": 1e-06, + "loss": 0.0119, + "num_tokens": 41673249.0, + "reward": 0.5111607313156128, + "reward_std": 0.2086862176656723, + "rewards/verify_math_reward/mean": 0.5111607313156128, + "rewards/verify_math_reward/std": 0.5001546144485474, + "step": 72 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3936.0, + "completions/mean_length": 611.7801513671875, + "completions/mean_terminated_length": 572.4548950195312, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 0.6810495626822157, + "grad_norm": 0.12580303847789764, + "learning_rate": 1e-06, + "loss": 0.0004, + "num_tokens": 42277220.0, + "reward": 0.559151828289032, + "reward_std": 0.1888575404882431, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 73 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4029.0, + "completions/mean_length": 694.0457763671875, + "completions/mean_terminated_length": 596.4006958007812, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.6903790087463557, + "grad_norm": 0.1160757765173912, + "learning_rate": 1e-06, + "loss": -0.0074, + "num_tokens": 42886389.0, + "reward": 0.5636160969734192, + "reward_std": 0.19313152134418488, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 74 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4004.0, + "completions/mean_length": 672.2410888671875, + "completions/mean_terminated_length": 594.0730590820312, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 0.6997084548104956, + "grad_norm": 0.12040586024522781, + "learning_rate": 1e-06, + "loss": 0.0222, + "num_tokens": 43491861.0, + "reward": 0.5089285969734192, + "reward_std": 0.19963862001895905, + "rewards/verify_math_reward/mean": 0.5089285969734192, + "rewards/verify_math_reward/std": 0.5001994967460632, + "step": 75 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3651.0, + "completions/mean_length": 668.1004638671875, + "completions/mean_terminated_length": 589.837890625, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 0.7090379008746356, + "grad_norm": 0.1331687569618225, + "learning_rate": 1e-06, + "loss": -0.0057, + "num_tokens": 44101799.0, + "reward": 0.5178571939468384, + "reward_std": 0.24333061277866364, + "rewards/verify_math_reward/mean": 0.5178571343421936, + "rewards/verify_math_reward/std": 0.4999600946903229, + "step": 76 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3003.0, + "completions/mean_length": 624.3326416015625, + "completions/mean_terminated_length": 553.1594848632812, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 0.7183673469387755, + "grad_norm": 0.14083652198314667, + "learning_rate": 1e-06, + "loss": 0.0047, + "num_tokens": 44679097.0, + "reward": 0.53125, + "reward_std": 0.21688465774059296, + "rewards/verify_math_reward/mean": 0.53125, + "rewards/verify_math_reward/std": 0.4993011951446533, + "step": 77 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3831.0, + "completions/mean_length": 636.3761596679688, + "completions/mean_terminated_length": 561.4241333007812, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 0.7276967930029155, + "grad_norm": 0.1320095807313919, + "learning_rate": 1e-06, + "loss": 0.0128, + "num_tokens": 45257378.0, + "reward": 0.5758928656578064, + "reward_std": 0.20192770659923553, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 78 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1760.0, + "completions/mean_length": 576.2980346679688, + "completions/mean_terminated_length": 512.3034057617188, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 0.7370262390670554, + "grad_norm": 0.14839765429496765, + "learning_rate": 1e-06, + "loss": 0.0043, + "num_tokens": 45797293.0, + "reward": 0.6104910969734192, + "reward_std": 0.22715014219284058, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 79 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3744.0, + "completions/mean_length": 572.716552734375, + "completions/mean_terminated_length": 524.88916015625, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "epoch": 0.7463556851311953, + "grad_norm": 0.1434706449508667, + "learning_rate": 1e-06, + "loss": -0.001, + "num_tokens": 46351343.0, + "reward": 0.5915178656578064, + "reward_std": 0.23867936432361603, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 80 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2544.0, + "completions/mean_length": 648.0513916015625, + "completions/mean_terminated_length": 593.322021484375, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.7556851311953353, + "grad_norm": 0.13106873631477356, + "learning_rate": 1e-06, + "loss": -0.0037, + "num_tokens": 46970469.0, + "reward": 0.5613839626312256, + "reward_std": 0.2206423580646515, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 81 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3441.0, + "completions/mean_length": 636.4453125, + "completions/mean_terminated_length": 557.4600219726562, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 0.7650145772594752, + "grad_norm": 0.12176746875047684, + "learning_rate": 1e-06, + "loss": -0.0007, + "num_tokens": 47552972.0, + "reward": 0.5580357313156128, + "reward_std": 0.20218874514102936, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 82 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3018.0, + "completions/mean_length": 591.935302734375, + "completions/mean_terminated_length": 556.3810424804688, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 0.7743440233236152, + "grad_norm": 0.1223415732383728, + "learning_rate": 1e-06, + "loss": 0.0084, + "num_tokens": 48131698.0, + "reward": 0.6305803656578064, + "reward_std": 0.19640770554542542, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 83 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3018.0, + "completions/mean_length": 641.7645263671875, + "completions/mean_terminated_length": 578.960205078125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 0.7836734693877551, + "grad_norm": 0.13270699977874756, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 48736799.0, + "reward": 0.4955357313156128, + "reward_std": 0.21409954130649567, + "rewards/verify_math_reward/mean": 0.4955357015132904, + "rewards/verify_math_reward/std": 0.500259280204773, + "step": 84 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3593.0, + "completions/mean_length": 611.154052734375, + "completions/mean_terminated_length": 547.7931518554688, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 0.793002915451895, + "grad_norm": 0.12596824765205383, + "learning_rate": 1e-06, + "loss": -0.0009, + "num_tokens": 49308585.0, + "reward": 0.566964328289032, + "reward_std": 0.19549660384655, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 85 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3890.0, + "completions/mean_length": 590.7076416015625, + "completions/mean_terminated_length": 571.0370483398438, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 0.8023323615160349, + "grad_norm": 0.12296757102012634, + "learning_rate": 1e-06, + "loss": 0.0167, + "num_tokens": 49904243.0, + "reward": 0.6049107313156128, + "reward_std": 0.19591733813285828, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 86 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2137.0, + "completions/mean_length": 552.4308471679688, + "completions/mean_terminated_length": 516.4757080078125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 0.8116618075801749, + "grad_norm": 0.1391582190990448, + "learning_rate": 1e-06, + "loss": 0.0171, + "num_tokens": 50443021.0, + "reward": 0.6183035969734192, + "reward_std": 0.20125369727611542, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 87 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3319.0, + "completions/mean_length": 578.0792846679688, + "completions/mean_terminated_length": 530.32470703125, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "epoch": 0.8209912536443149, + "grad_norm": 0.13660335540771484, + "learning_rate": 1e-06, + "loss": 0.0194, + "num_tokens": 50991900.0, + "reward": 0.6116071939468384, + "reward_std": 0.18948623538017273, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.4876568913459778, + "step": 88 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3661.0, + "completions/mean_length": 643.3761596679688, + "completions/mean_terminated_length": 564.549072265625, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 0.8303206997084548, + "grad_norm": 0.12995995581150055, + "learning_rate": 1e-06, + "loss": 0.0135, + "num_tokens": 51588293.0, + "reward": 0.527901828289032, + "reward_std": 0.20245903730392456, + "rewards/verify_math_reward/mean": 0.5279017686843872, + "rewards/verify_math_reward/std": 0.49949970841407776, + "step": 89 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2473.0, + "completions/mean_length": 607.224365234375, + "completions/mean_terminated_length": 547.8240966796875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 0.8396501457725948, + "grad_norm": 0.11275593191385269, + "learning_rate": 1e-06, + "loss": 0.0199, + "num_tokens": 52158478.0, + "reward": 0.5636160969734192, + "reward_std": 0.1716410517692566, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 90 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3523.0, + "completions/mean_length": 613.1741333007812, + "completions/mean_terminated_length": 541.772216796875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 0.8489795918367347, + "grad_norm": 0.1297285109758377, + "learning_rate": 1e-06, + "loss": 0.0178, + "num_tokens": 52722114.0, + "reward": 0.5837053656578064, + "reward_std": 0.1964818835258484, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 91 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3583.0, + "completions/mean_length": 621.9710083007812, + "completions/mean_terminated_length": 566.8276977539062, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 0.8583090379008746, + "grad_norm": 0.10461857169866562, + "learning_rate": 1e-06, + "loss": 0.0009, + "num_tokens": 53309136.0, + "reward": 0.6015625, + "reward_std": 0.1629975289106369, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 92 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3135.0, + "completions/mean_length": 595.669677734375, + "completions/mean_terminated_length": 532.0272827148438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 0.8676384839650145, + "grad_norm": 0.18813636898994446, + "learning_rate": 1e-06, + "loss": 0.0134, + "num_tokens": 53874856.0, + "reward": 0.582589328289032, + "reward_std": 0.24276649951934814, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.493407279253006, + "step": 93 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2350.0, + "completions/mean_length": 607.4319458007812, + "completions/mean_terminated_length": 552.0578002929688, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 0.8769679300291545, + "grad_norm": 0.13008354604244232, + "learning_rate": 1e-06, + "loss": 0.0154, + "num_tokens": 54438523.0, + "reward": 0.582589328289032, + "reward_std": 0.20685499906539917, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 94 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 647.8917846679688, + "completions/mean_terminated_length": 589.1838989257812, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 0.8862973760932945, + "grad_norm": 0.11892709881067276, + "learning_rate": 1e-06, + "loss": 0.0069, + "num_tokens": 55045090.0, + "reward": 0.566964328289032, + "reward_std": 0.19858624041080475, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 95 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1937.0, + "completions/mean_length": 639.0614013671875, + "completions/mean_terminated_length": 564.1676025390625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 0.8956268221574344, + "grad_norm": 0.12884822487831116, + "learning_rate": 1e-06, + "loss": 0.0276, + "num_tokens": 55623321.0, + "reward": 0.5558035969734192, + "reward_std": 0.2165483683347702, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715369939804077, + "step": 96 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2844.0, + "completions/mean_length": 632.7991333007812, + "completions/mean_terminated_length": 573.8342895507812, + "completions/min_length": 54.0, + "completions/min_terminated_length": 54.0, + "epoch": 0.9049562682215744, + "grad_norm": 0.12199296057224274, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 56220165.0, + "reward": 0.5479910969734192, + "reward_std": 0.17250937223434448, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796947836875916, + "step": 97 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3313.0, + "completions/mean_length": 700.0145263671875, + "completions/mean_terminated_length": 626.4412841796875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 0.9142857142857143, + "grad_norm": 0.11525772511959076, + "learning_rate": 1e-06, + "loss": 0.0017, + "num_tokens": 56855482.0, + "reward": 0.5167410969734192, + "reward_std": 0.17250937223434448, + "rewards/verify_math_reward/mean": 0.5167410969734192, + "rewards/verify_math_reward/std": 0.4999987483024597, + "step": 98 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4059.0, + "completions/mean_length": 670.122802734375, + "completions/mean_terminated_length": 603.86572265625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 0.9236151603498542, + "grad_norm": 0.11693616211414337, + "learning_rate": 1e-06, + "loss": 0.0015, + "num_tokens": 57485496.0, + "reward": 0.494419664144516, + "reward_std": 0.19674287736415863, + "rewards/verify_math_reward/mean": 0.4944196343421936, + "rewards/verify_math_reward/std": 0.5002480745315552, + "step": 99 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3366.0, + "completions/mean_length": 662.5848388671875, + "completions/mean_terminated_length": 584.1963500976562, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 0.9329446064139941, + "grad_norm": 0.13173900544643402, + "learning_rate": 1e-06, + "loss": 0.0216, + "num_tokens": 58087516.0, + "reward": 0.5658482313156128, + "reward_std": 0.21294134855270386, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 100 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3616.0, + "completions/mean_length": 643.5558471679688, + "completions/mean_terminated_length": 576.7849731445312, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 0.9422740524781341, + "grad_norm": 0.11459070444107056, + "learning_rate": 1e-06, + "loss": -0.0025, + "num_tokens": 58680030.0, + "reward": 0.574776828289032, + "reward_std": 0.17998285591602325, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 101 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4017.0, + "completions/mean_length": 613.2232666015625, + "completions/mean_terminated_length": 569.9344482421875, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 0.9516034985422741, + "grad_norm": 0.12105487287044525, + "learning_rate": 1e-06, + "loss": 0.0095, + "num_tokens": 59284334.0, + "reward": 0.5401785969734192, + "reward_std": 0.19366146624088287, + "rewards/verify_math_reward/mean": 0.5401785969734192, + "rewards/verify_math_reward/std": 0.49866142868995667, + "step": 102 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3876.0, + "completions/mean_length": 714.2689819335938, + "completions/mean_terminated_length": 633.107421875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.960932944606414, + "grad_norm": 0.10672217607498169, + "learning_rate": 1e-06, + "loss": 0.0054, + "num_tokens": 59925767.0, + "reward": 0.5446428656578064, + "reward_std": 0.17634011805057526, + "rewards/verify_math_reward/mean": 0.5446428656578064, + "rewards/verify_math_reward/std": 0.49828118085861206, + "step": 103 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4094.0, + "completions/mean_length": 611.6495971679688, + "completions/mean_terminated_length": 568.3412475585938, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "epoch": 0.970262390670554, + "grad_norm": 0.1321568340063095, + "learning_rate": 1e-06, + "loss": -0.0015, + "num_tokens": 60521693.0, + "reward": 0.5837053656578064, + "reward_std": 0.2094464898109436, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 563.388427734375, + "completions/mean_terminated_length": 515.4343872070312, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 0.9795918367346939, + "grad_norm": 0.1289939135313034, + "learning_rate": 1e-06, + "loss": 0.0012, + "num_tokens": 61065233.0, + "reward": 0.5390625, + "reward_std": 0.1993015855550766, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4006.0, + "completions/mean_length": 601.0592041015625, + "completions/mean_terminated_length": 553.6165161132812, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 0.9889212827988338, + "grad_norm": 0.11271937191486359, + "learning_rate": 1e-06, + "loss": 0.0007, + "num_tokens": 61636822.0, + "reward": 0.5625, + "reward_std": 0.1666862815618515, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 106 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005681818181818232, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1951.0, + "completions/mean_length": 522.0056762695312, + "completions/mean_terminated_length": 501.5828552246094, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 0.9982507288629737, + "grad_norm": 0.13385462760925293, + "learning_rate": 1e-06, + "loss": 0.0289, + "num_tokens": 62209340.0, + "reward": 0.5446428656578064, + "reward_std": 0.20534475147724152, + "rewards/verify_math_reward/mean": 0.5446428656578064, + "rewards/verify_math_reward/std": 0.4982811510562897, + "step": 107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2726.0, + "completions/mean_length": 584.3638916015625, + "completions/mean_terminated_length": 552.7274780273438, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 1.00932944606414, + "grad_norm": 0.13441289961338043, + "learning_rate": 1e-06, + "loss": 0.0199, + "num_tokens": 62790586.0, + "reward": 0.5770089626312256, + "reward_std": 0.22912222146987915, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3852.0, + "completions/mean_length": 605.4475708007812, + "completions/mean_terminated_length": 550.0419311523438, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 1.01865889212828, + "grad_norm": 0.13123084604740143, + "learning_rate": 1e-06, + "loss": 0.0068, + "num_tokens": 63372171.0, + "reward": 0.5647321939468384, + "reward_std": 0.19857734441757202, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 109 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 673.0814819335938, + "completions/mean_terminated_length": 578.8726806640625, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 1.0279883381924197, + "grad_norm": 0.12347409874200821, + "learning_rate": 1e-06, + "loss": 0.0005, + "num_tokens": 63969948.0, + "reward": 0.5334821939468384, + "reward_std": 0.18693894147872925, + "rewards/verify_math_reward/mean": 0.5334821343421936, + "rewards/verify_math_reward/std": 0.49915632605552673, + "step": 110 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3905.0, + "completions/mean_length": 617.4921875, + "completions/mean_terminated_length": 558.2667846679688, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 1.0373177842565597, + "grad_norm": 0.12921538949012756, + "learning_rate": 1e-06, + "loss": -0.0015, + "num_tokens": 64566157.0, + "reward": 0.5234375, + "reward_std": 0.2080143243074417, + "rewards/verify_math_reward/mean": 0.5234375, + "rewards/verify_math_reward/std": 0.49972933530807495, + "step": 111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3642.0, + "completions/mean_length": 580.630615234375, + "completions/mean_terminated_length": 536.9367065429688, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 1.0466472303206997, + "grad_norm": 0.1322759985923767, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 65135922.0, + "reward": 0.590401828289032, + "reward_std": 0.23255206644535065, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3906.0, + "completions/mean_length": 608.5435791015625, + "completions/mean_terminated_length": 565.1965942382812, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 1.0559766763848397, + "grad_norm": 0.1161135882139206, + "learning_rate": 1e-06, + "loss": 0.0066, + "num_tokens": 65726753.0, + "reward": 0.5613839626312256, + "reward_std": 0.17893047630786896, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3164.0, + "completions/mean_length": 616.8828125, + "completions/mean_terminated_length": 549.5961303710938, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 1.0653061224489795, + "grad_norm": 0.12141867727041245, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 66321296.0, + "reward": 0.486607164144516, + "reward_std": 0.17284637689590454, + "rewards/verify_math_reward/mean": 0.4866071343421936, + "rewards/verify_math_reward/std": 0.5000997185707092, + "step": 114 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3190.0, + "completions/mean_length": 596.0792846679688, + "completions/mean_terminated_length": 548.5690307617188, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "epoch": 1.0746355685131195, + "grad_norm": 0.13341371715068817, + "learning_rate": 1e-06, + "loss": 0.012, + "num_tokens": 66894143.0, + "reward": 0.5546875, + "reward_std": 0.19697293639183044, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 115 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3760.0, + "completions/mean_length": 618.6875, + "completions/mean_terminated_length": 559.482421875, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 1.0839650145772595, + "grad_norm": 0.12516039609909058, + "learning_rate": 1e-06, + "loss": -0.0082, + "num_tokens": 67474887.0, + "reward": 0.6283482313156128, + "reward_std": 0.2076791673898697, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 116 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3073.0, + "completions/mean_length": 588.2299194335938, + "completions/mean_terminated_length": 556.62841796875, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 1.0932944606413995, + "grad_norm": 0.12874138355255127, + "learning_rate": 1e-06, + "loss": 0.0058, + "num_tokens": 68054293.0, + "reward": 0.5524553656578064, + "reward_std": 0.23160958290100098, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 117 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2461.0, + "completions/mean_length": 625.7578125, + "completions/mean_terminated_length": 554.6138916015625, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 1.1026239067055394, + "grad_norm": 0.1242959201335907, + "learning_rate": 1e-06, + "loss": -0.0089, + "num_tokens": 68615596.0, + "reward": 0.640625, + "reward_std": 0.21383853256702423, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2827.0, + "completions/mean_length": 592.6707763671875, + "completions/mean_terminated_length": 520.8485107421875, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "epoch": 1.1119533527696792, + "grad_norm": 0.140296071767807, + "learning_rate": 1e-06, + "loss": -0.0005, + "num_tokens": 69146525.0, + "reward": 0.6439732313156128, + "reward_std": 0.18521295487880707, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 119 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3486.0, + "completions/mean_length": 638.2355346679688, + "completions/mean_terminated_length": 583.350341796875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.1212827988338192, + "grad_norm": 0.12419833987951279, + "learning_rate": 1e-06, + "loss": 0.0078, + "num_tokens": 69757680.0, + "reward": 0.5859375, + "reward_std": 0.1885533183813095, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 120 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3251.0, + "completions/mean_length": 636.1830444335938, + "completions/mean_terminated_length": 577.27587890625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 1.1306122448979592, + "grad_norm": 0.12782782316207886, + "learning_rate": 1e-06, + "loss": 0.0063, + "num_tokens": 70363460.0, + "reward": 0.5368303656578064, + "reward_std": 0.21733155846595764, + "rewards/verify_math_reward/mean": 0.5368303656578064, + "rewards/verify_math_reward/std": 0.49892017245292664, + "step": 121 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 614.8527221679688, + "completions/mean_terminated_length": 567.5972900390625, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 1.1399416909620992, + "grad_norm": 0.1233368068933487, + "learning_rate": 1e-06, + "loss": 0.0084, + "num_tokens": 70943784.0, + "reward": 0.6383928656578064, + "reward_std": 0.18257687985897064, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 122 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3999.0, + "completions/mean_length": 671.841552734375, + "completions/mean_terminated_length": 593.6643676757812, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.149271137026239, + "grad_norm": 0.12879662215709686, + "learning_rate": 1e-06, + "loss": 0.0019, + "num_tokens": 71553922.0, + "reward": 0.5926339626312256, + "reward_std": 0.22225745022296906, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161848425865173, + "step": 123 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3305.0, + "completions/mean_length": 614.4710083007812, + "completions/mean_terminated_length": 559.2086181640625, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 1.158600583090379, + "grad_norm": 0.13115562498569489, + "learning_rate": 1e-06, + "loss": -0.0093, + "num_tokens": 72136656.0, + "reward": 0.6104910969734192, + "reward_std": 0.18160048127174377, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 124 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3528.0, + "completions/mean_length": 676.7332763671875, + "completions/mean_terminated_length": 590.6647338867188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.167930029154519, + "grad_norm": 0.11384246498346329, + "learning_rate": 1e-06, + "loss": -0.0005, + "num_tokens": 72743241.0, + "reward": 0.5078125, + "reward_std": 0.15293073654174805, + "rewards/verify_math_reward/mean": 0.5078125, + "rewards/verify_math_reward/std": 0.5002182126045227, + "step": 125 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2237.0, + "completions/mean_length": 578.9877319335938, + "completions/mean_terminated_length": 547.3029174804688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.177259475218659, + "grad_norm": 0.1207195520401001, + "learning_rate": 1e-06, + "loss": 0.0134, + "num_tokens": 73317726.0, + "reward": 0.6127232313156128, + "reward_std": 0.1758943349123001, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 126 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3481.0, + "completions/mean_length": 604.044677734375, + "completions/mean_terminated_length": 564.632080078125, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 1.186588921282799, + "grad_norm": 0.13039085268974304, + "learning_rate": 1e-06, + "loss": 0.0203, + "num_tokens": 73900590.0, + "reward": 0.6183035969734192, + "reward_std": 0.18160118162631989, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 127 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 659.796875, + "completions/mean_terminated_length": 569.2669067382812, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 1.1959183673469387, + "grad_norm": 0.12605169415473938, + "learning_rate": 1e-06, + "loss": 0.0072, + "num_tokens": 74489480.0, + "reward": 0.5345982313156128, + "reward_std": 0.21861176192760468, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3424.0, + "completions/mean_length": 636.1652221679688, + "completions/mean_terminated_length": 581.2471923828125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.2052478134110787, + "grad_norm": 0.1338040828704834, + "learning_rate": 1e-06, + "loss": 0.0035, + "num_tokens": 75095228.0, + "reward": 0.5647321939468384, + "reward_std": 0.23991748690605164, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 129 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3753.0, + "completions/mean_length": 720.646240234375, + "completions/mean_terminated_length": 627.7465209960938, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 1.2145772594752187, + "grad_norm": 0.1312447488307953, + "learning_rate": 1e-06, + "loss": -0.0006, + "num_tokens": 75735767.0, + "reward": 0.5089285969734192, + "reward_std": 0.22199741005897522, + "rewards/verify_math_reward/mean": 0.5089285969734192, + "rewards/verify_math_reward/std": 0.5001994967460632, + "step": 130 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4022.0, + "completions/mean_length": 557.8348388671875, + "completions/mean_terminated_length": 525.95947265625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 1.2239067055393587, + "grad_norm": 0.13229545950889587, + "learning_rate": 1e-06, + "loss": -0.0062, + "num_tokens": 76287971.0, + "reward": 0.6339285969734192, + "reward_std": 0.17476709187030792, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199838399887085, + "step": 131 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2593.0, + "completions/mean_length": 575.1596069335938, + "completions/mean_terminated_length": 535.4210205078125, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 1.2332361516034984, + "grad_norm": 0.11969714611768723, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 76847850.0, + "reward": 0.6049107313156128, + "reward_std": 0.147451713681221, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 132 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2995.0, + "completions/mean_length": 574.390625, + "completions/mean_terminated_length": 526.5859985351562, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 1.2425655976676384, + "grad_norm": 0.1340891569852829, + "learning_rate": 1e-06, + "loss": 0.0031, + "num_tokens": 77404224.0, + "reward": 0.5658482313156128, + "reward_std": 0.18806618452072144, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 133 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1978.0, + "completions/mean_length": 607.1517944335938, + "completions/mean_terminated_length": 559.7918701171875, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 1.2518950437317784, + "grad_norm": 0.134235218167305, + "learning_rate": 1e-06, + "loss": 0.0026, + "num_tokens": 77980344.0, + "reward": 0.6160714626312256, + "reward_std": 0.21838605403900146, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 134 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3948.0, + "completions/mean_length": 614.239990234375, + "completions/mean_terminated_length": 530.677734375, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 1.2612244897959184, + "grad_norm": 0.1362096667289734, + "learning_rate": 1e-06, + "loss": -0.0123, + "num_tokens": 78528703.0, + "reward": 0.6573660969734192, + "reward_std": 0.211551234126091, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 135 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 632.708740234375, + "completions/mean_terminated_length": 545.531982421875, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 1.2705539358600584, + "grad_norm": 0.13689996302127838, + "learning_rate": 1e-06, + "loss": -0.0026, + "num_tokens": 79097098.0, + "reward": 0.5870535969734192, + "reward_std": 0.18836969137191772, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 136 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3911.0, + "completions/mean_length": 652.6640625, + "completions/mean_terminated_length": 582.07177734375, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 1.2798833819241984, + "grad_norm": 0.12383565306663513, + "learning_rate": 1e-06, + "loss": 0.0027, + "num_tokens": 79694381.0, + "reward": 0.590401828289032, + "reward_std": 0.19141653180122375, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 137 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4084.0, + "completions/mean_length": 647.2288208007812, + "completions/mean_terminated_length": 576.5250854492188, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 1.2892128279883381, + "grad_norm": 0.1292634755373001, + "learning_rate": 1e-06, + "loss": -0.0111, + "num_tokens": 80292882.0, + "reward": 0.5680803656578064, + "reward_std": 0.18825024366378784, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 138 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3978.0, + "completions/mean_length": 649.7589721679688, + "completions/mean_terminated_length": 599.021484375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.2985422740524781, + "grad_norm": 0.139671191573143, + "learning_rate": 1e-06, + "loss": 0.0021, + "num_tokens": 80917602.0, + "reward": 0.5189732313156128, + "reward_std": 0.23443815112113953, + "rewards/verify_math_reward/mean": 0.5189732313156128, + "rewards/verify_math_reward/std": 0.49991893768310547, + "step": 139 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 709.6964721679688, + "completions/mean_terminated_length": 596.4290161132812, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 1.3078717201166181, + "grad_norm": 0.12194110453128815, + "learning_rate": 1e-06, + "loss": 0.005, + "num_tokens": 81527762.0, + "reward": 0.5078125, + "reward_std": 0.19294606149196625, + "rewards/verify_math_reward/mean": 0.5078125, + "rewards/verify_math_reward/std": 0.5002182126045227, + "step": 140 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2433.0, + "completions/mean_length": 681.5413208007812, + "completions/mean_terminated_length": 599.5942993164062, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 1.3172011661807579, + "grad_norm": 0.1287652552127838, + "learning_rate": 1e-06, + "loss": 0.0193, + "num_tokens": 82146335.0, + "reward": 0.5234375, + "reward_std": 0.2050088495016098, + "rewards/verify_math_reward/mean": 0.5234375, + "rewards/verify_math_reward/std": 0.49972933530807495, + "step": 141 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3231.0, + "completions/mean_length": 669.6295166015625, + "completions/mean_terminated_length": 599.385009765625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.3265306122448979, + "grad_norm": 0.11956693232059479, + "learning_rate": 1e-06, + "loss": -0.0007, + "num_tokens": 82766003.0, + "reward": 0.5535714626312256, + "reward_std": 0.18809868395328522, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973995089530945, + "step": 142 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2632.0, + "completions/mean_length": 654.6830444335938, + "completions/mean_terminated_length": 580.127685546875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.3358600583090379, + "grad_norm": 0.12889018654823303, + "learning_rate": 1e-06, + "loss": 0.02, + "num_tokens": 83366343.0, + "reward": 0.5502232313156128, + "reward_std": 0.20729796588420868, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 143 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2575.0, + "completions/mean_length": 679.0971069335938, + "completions/mean_terminated_length": 616.9715576171875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 1.3451895043731779, + "grad_norm": 0.1147773340344429, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 84005494.0, + "reward": 0.5334821939468384, + "reward_std": 0.19974736869335175, + "rewards/verify_math_reward/mean": 0.5334821343421936, + "rewards/verify_math_reward/std": 0.49915632605552673, + "step": 144 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3538.0, + "completions/mean_length": 634.5881958007812, + "completions/mean_terminated_length": 563.6253051757812, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 1.3545189504373178, + "grad_norm": 0.13499262928962708, + "learning_rate": 1e-06, + "loss": 0.0012, + "num_tokens": 84587013.0, + "reward": 0.6015625, + "reward_std": 0.22105422616004944, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 145 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2492.0, + "completions/mean_length": 691.1373291015625, + "completions/mean_terminated_length": 585.3475341796875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 1.3638483965014578, + "grad_norm": 0.1364857703447342, + "learning_rate": 1e-06, + "loss": 0.0099, + "num_tokens": 85199064.0, + "reward": 0.5792410969734192, + "reward_std": 0.21346086263656616, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 146 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4075.0, + "completions/mean_length": 734.8158569335938, + "completions/mean_terminated_length": 602.2401123046875, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 1.3731778425655976, + "grad_norm": 0.12410472333431244, + "learning_rate": 1e-06, + "loss": 0.0057, + "num_tokens": 85805803.0, + "reward": 0.5345982313156128, + "reward_std": 0.18399836122989655, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3167.0, + "completions/mean_length": 681.4944458007812, + "completions/mean_terminated_length": 599.5462646484375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.3825072886297376, + "grad_norm": 0.14087340235710144, + "learning_rate": 1e-06, + "loss": 0.0176, + "num_tokens": 86413854.0, + "reward": 0.5803571939468384, + "reward_std": 0.2527284324169159, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 148 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3475.0, + "completions/mean_length": 619.2857666015625, + "completions/mean_terminated_length": 548.0091552734375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.3918367346938776, + "grad_norm": 0.12467251718044281, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 86979630.0, + "reward": 0.5848214626312256, + "reward_std": 0.17596174776554108, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 149 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1961.0, + "completions/mean_length": 648.8370971679688, + "completions/mean_terminated_length": 566.1051635742188, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.4011661807580174, + "grad_norm": 0.12112050503492355, + "learning_rate": 1e-06, + "loss": -0.0094, + "num_tokens": 87571228.0, + "reward": 0.6272321939468384, + "reward_std": 0.174800306558609, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111400604248, + "step": 150 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3286.0, + "completions/mean_length": 694.3270263671875, + "completions/mean_terminated_length": 584.5956420898438, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 1.4104956268221573, + "grad_norm": 0.12694108486175537, + "learning_rate": 1e-06, + "loss": 0.0083, + "num_tokens": 88175857.0, + "reward": 0.5457589626312256, + "reward_std": 0.19628755748271942, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981797933578491, + "step": 151 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2874.0, + "completions/mean_length": 627.2846069335938, + "completions/mean_terminated_length": 560.1990966796875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 1.4198250728862973, + "grad_norm": 0.1249031275510788, + "learning_rate": 1e-06, + "loss": 0.0057, + "num_tokens": 88752752.0, + "reward": 0.640625, + "reward_std": 0.1961027979850769, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3080.0, + "completions/mean_length": 653.1964721679688, + "completions/mean_terminated_length": 562.4925537109375, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 1.4291545189504373, + "grad_norm": 0.1260456144809723, + "learning_rate": 1e-06, + "loss": 0.0049, + "num_tokens": 89333032.0, + "reward": 0.5658482313156128, + "reward_std": 0.1965906322002411, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2784.0, + "completions/mean_length": 605.0078125, + "completions/mean_terminated_length": 561.616943359375, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 1.4384839650145773, + "grad_norm": 0.12526555359363556, + "learning_rate": 1e-06, + "loss": 0.0224, + "num_tokens": 89912591.0, + "reward": 0.6261160969734192, + "reward_std": 0.18829117715358734, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 154 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2644.0, + "completions/mean_length": 608.7846069335938, + "completions/mean_terminated_length": 529.1677856445312, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.4478134110787173, + "grad_norm": 0.13587446510791779, + "learning_rate": 1e-06, + "loss": 0.0061, + "num_tokens": 90460470.0, + "reward": 0.6395089626312256, + "reward_std": 0.19340364634990692, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3435.0, + "completions/mean_length": 611.8013916015625, + "completions/mean_terminated_length": 548.4522705078125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 1.457142857142857, + "grad_norm": 0.13532030582427979, + "learning_rate": 1e-06, + "loss": -0.0048, + "num_tokens": 91032412.0, + "reward": 0.6015625, + "reward_std": 0.19094981253147125, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 156 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 686.7221069335938, + "completions/mean_terminated_length": 608.8847045898438, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 1.466472303206997, + "grad_norm": 0.11518780887126923, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 91657299.0, + "reward": 0.5546875, + "reward_std": 0.18325015902519226, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 674.5245971679688, + "completions/mean_terminated_length": 592.4091186523438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 1.475801749271137, + "grad_norm": 0.11879177391529083, + "learning_rate": 1e-06, + "loss": 0.0106, + "num_tokens": 92266513.0, + "reward": 0.5636160969734192, + "reward_std": 0.17585085332393646, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 158 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3952.0, + "completions/mean_length": 666.7433471679688, + "completions/mean_terminated_length": 612.3106689453125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.485131195335277, + "grad_norm": 0.11994239687919617, + "learning_rate": 1e-06, + "loss": 0.0091, + "num_tokens": 92885115.0, + "reward": 0.6037946939468384, + "reward_std": 0.18362995982170105, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 159 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3906.0, + "completions/mean_length": 652.8850708007812, + "completions/mean_terminated_length": 594.26220703125, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 1.4944606413994168, + "grad_norm": 0.12814350426197052, + "learning_rate": 1e-06, + "loss": 0.0214, + "num_tokens": 93498764.0, + "reward": 0.613839328289032, + "reward_std": 0.19280339777469635, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 160 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3710.0, + "completions/mean_length": 651.8939819335938, + "completions/mean_terminated_length": 601.18798828125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 1.5037900874635568, + "grad_norm": 0.141206294298172, + "learning_rate": 1e-06, + "loss": 0.0054, + "num_tokens": 94112005.0, + "reward": 0.6104910969734192, + "reward_std": 0.23138949275016785, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 161 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3377.0, + "completions/mean_length": 676.8314819335938, + "completions/mean_terminated_length": 582.7258911132812, + "completions/min_length": 84.0, + "completions/min_terminated_length": 84.0, + "epoch": 1.5131195335276968, + "grad_norm": 0.1266036331653595, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 94715582.0, + "reward": 0.5535714626312256, + "reward_std": 0.19230598211288452, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973994791507721, + "step": 162 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4049.0, + "completions/mean_length": 638.4342041015625, + "completions/mean_terminated_length": 575.5692749023438, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 1.5224489795918368, + "grad_norm": 0.13367679715156555, + "learning_rate": 1e-06, + "loss": 0.0144, + "num_tokens": 95323699.0, + "reward": 0.5301339626312256, + "reward_std": 0.20426097512245178, + "rewards/verify_math_reward/mean": 0.5301339030265808, + "rewards/verify_math_reward/std": 0.49936988949775696, + "step": 163 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3274.0, + "completions/mean_length": 620.59375, + "completions/mean_terminated_length": 561.421142578125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 1.5317784256559768, + "grad_norm": 0.11846431344747543, + "learning_rate": 1e-06, + "loss": 0.0026, + "num_tokens": 95911799.0, + "reward": 0.5502232313156128, + "reward_std": 0.15007779002189636, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 164 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3400.0, + "completions/mean_length": 709.7266235351562, + "completions/mean_terminated_length": 616.5263671875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 1.5411078717201168, + "grad_norm": 0.12937229871749878, + "learning_rate": 1e-06, + "loss": 0.0096, + "num_tokens": 96541298.0, + "reward": 0.5446428656578064, + "reward_std": 0.19268827140331268, + "rewards/verify_math_reward/mean": 0.5446428656578064, + "rewards/verify_math_reward/std": 0.49828118085861206, + "step": 165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3652.0, + "completions/mean_length": 567.5814819335938, + "completions/mean_terminated_length": 519.6843872070312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 1.5504373177842565, + "grad_norm": 0.127097487449646, + "learning_rate": 1e-06, + "loss": 0.0111, + "num_tokens": 97094259.0, + "reward": 0.566964328289032, + "reward_std": 0.16232171654701233, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 166 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2755.0, + "completions/mean_length": 597.2131958007812, + "completions/mean_terminated_length": 569.6636962890625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.5597667638483965, + "grad_norm": 0.14329767227172852, + "learning_rate": 1e-06, + "loss": 0.0099, + "num_tokens": 97683594.0, + "reward": 0.6049107313156128, + "reward_std": 0.23837655782699585, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 167 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3254.0, + "completions/mean_length": 666.5725708007812, + "completions/mean_terminated_length": 600.246826171875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.5690962099125363, + "grad_norm": 0.12098225206136703, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 98295555.0, + "reward": 0.578125, + "reward_std": 0.1763821691274643, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3295.0, + "completions/mean_length": 623.2355346679688, + "completions/mean_terminated_length": 572.1075439453125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 1.5784256559766763, + "grad_norm": 0.1256924271583557, + "learning_rate": 1e-06, + "loss": 0.0037, + "num_tokens": 98888526.0, + "reward": 0.5881696939468384, + "reward_std": 0.18419267237186432, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 169 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3755.0, + "completions/mean_length": 666.9754638671875, + "completions/mean_terminated_length": 580.6613159179688, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 1.5877551020408163, + "grad_norm": 0.1224668025970459, + "learning_rate": 1e-06, + "loss": -0.0136, + "num_tokens": 99485192.0, + "reward": 0.598214328289032, + "reward_std": 0.1913723647594452, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 170 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3148.0, + "completions/mean_length": 654.7924194335938, + "completions/mean_terminated_length": 580.2394409179688, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 1.5970845481049563, + "grad_norm": 0.1412746161222458, + "learning_rate": 1e-06, + "loss": -0.0099, + "num_tokens": 100085150.0, + "reward": 0.551339328289032, + "reward_std": 0.21196381747722626, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 171 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2219.0, + "completions/mean_length": 652.4699096679688, + "completions/mean_terminated_length": 597.8106689453125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 1.6064139941690962, + "grad_norm": 0.12173449993133545, + "learning_rate": 1e-06, + "loss": 0.0204, + "num_tokens": 100704171.0, + "reward": 0.6227678656578064, + "reward_std": 0.23187625408172607, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 172 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4080.0, + "completions/mean_length": 661.8638916015625, + "completions/mean_terminated_length": 551.0852661132812, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 1.6157434402332362, + "grad_norm": 0.13657061755657196, + "learning_rate": 1e-06, + "loss": 0.0134, + "num_tokens": 101262961.0, + "reward": 0.6082589626312256, + "reward_std": 0.21053095161914825, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 173 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2149.0, + "completions/mean_length": 678.3292846679688, + "completions/mean_terminated_length": 547.6419067382812, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 1.6250728862973762, + "grad_norm": 0.13285422325134277, + "learning_rate": 1e-06, + "loss": -0.012, + "num_tokens": 101825888.0, + "reward": 0.5758928656578064, + "reward_std": 0.19268713891506195, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 174 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3792.0, + "completions/mean_length": 671.7957763671875, + "completions/mean_terminated_length": 589.6148681640625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.634402332361516, + "grad_norm": 0.12007246166467667, + "learning_rate": 1e-06, + "loss": 0.01, + "num_tokens": 102450169.0, + "reward": 0.5558035969734192, + "reward_std": 0.18156908452510834, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715369939804077, + "step": 175 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3228.0, + "completions/mean_length": 651.911865234375, + "completions/mean_terminated_length": 601.2061157226562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 1.643731778425656, + "grad_norm": 0.13123011589050293, + "learning_rate": 1e-06, + "loss": 0.012, + "num_tokens": 103071338.0, + "reward": 0.5926339626312256, + "reward_std": 0.21053346991539001, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 176 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2255.0, + "completions/mean_length": 643.7935791015625, + "completions/mean_terminated_length": 573.0194091796875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 1.6530612244897958, + "grad_norm": 0.13956408202648163, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 103665065.0, + "reward": 0.574776828289032, + "reward_std": 0.1870456039905548, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 177 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3612.0, + "completions/mean_length": 622.390625, + "completions/mean_terminated_length": 547.1356811523438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 1.6623906705539357, + "grad_norm": 0.12908336520195007, + "learning_rate": 1e-06, + "loss": 0.0018, + "num_tokens": 104228359.0, + "reward": 0.6194196939468384, + "reward_std": 0.19411608576774597, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 178 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2148.0, + "completions/mean_length": 608.5413208007812, + "completions/mean_terminated_length": 549.1634521484375, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 1.6717201166180757, + "grad_norm": 0.12952981889247894, + "learning_rate": 1e-06, + "loss": 0.001, + "num_tokens": 104794356.0, + "reward": 0.5892857313156128, + "reward_std": 0.1871204674243927, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 179 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3240.0, + "completions/mean_length": 598.8292846679688, + "completions/mean_terminated_length": 523.0638427734375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 1.6810495626822157, + "grad_norm": 0.1407703161239624, + "learning_rate": 1e-06, + "loss": 0.0121, + "num_tokens": 105339987.0, + "reward": 0.5993303656578064, + "reward_std": 0.19899991154670715, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 180 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3941.0, + "completions/mean_length": 640.2846069335938, + "completions/mean_terminated_length": 553.298583984375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 1.6903790087463557, + "grad_norm": 0.13343185186386108, + "learning_rate": 1e-06, + "loss": 0.0123, + "num_tokens": 105906418.0, + "reward": 0.6283482313156128, + "reward_std": 0.19738228619098663, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 181 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3778.0, + "completions/mean_length": 741.2813110351562, + "completions/mean_terminated_length": 637.0494995117188, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 1.6997084548104957, + "grad_norm": 0.12370902299880981, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 106564518.0, + "reward": 0.5033482313156128, + "reward_std": 0.19587455689907074, + "rewards/verify_math_reward/mean": 0.5033482313156128, + "rewards/verify_math_reward/std": 0.5002680420875549, + "step": 182 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2418.0, + "completions/mean_length": 544.9207763671875, + "completions/mean_terminated_length": 500.7830505371094, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 1.7090379008746357, + "grad_norm": 0.15225689113140106, + "learning_rate": 1e-06, + "loss": 0.0126, + "num_tokens": 107103487.0, + "reward": 0.6417410969734192, + "reward_std": 0.20474812388420105, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 183 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2579.0, + "completions/mean_length": 657.739990234375, + "completions/mean_terminated_length": 587.251708984375, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 1.7183673469387755, + "grad_norm": 0.12613576650619507, + "learning_rate": 1e-06, + "loss": 0.0062, + "num_tokens": 107711942.0, + "reward": 0.566964328289032, + "reward_std": 0.19760164618492126, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 184 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2920.0, + "completions/mean_length": 699.6652221679688, + "completions/mean_terminated_length": 577.9468383789062, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.7276967930029155, + "grad_norm": 0.12997154891490936, + "learning_rate": 1e-06, + "loss": 0.0011, + "num_tokens": 108297650.0, + "reward": 0.5457589626312256, + "reward_std": 0.17585083842277527, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 185 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3711.0, + "completions/mean_length": 660.046875, + "completions/mean_terminated_length": 565.4793090820312, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 1.7370262390670554, + "grad_norm": 0.13106532394886017, + "learning_rate": 1e-06, + "loss": 0.0048, + "num_tokens": 108882596.0, + "reward": 0.515625, + "reward_std": 0.2166917473077774, + "rewards/verify_math_reward/mean": 0.515625, + "rewards/verify_math_reward/std": 0.5000349283218384, + "step": 186 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3540.0, + "completions/mean_length": 565.5814819335938, + "completions/mean_terminated_length": 517.6572875976562, + "completions/min_length": 54.0, + "completions/min_terminated_length": 54.0, + "epoch": 1.7463556851311952, + "grad_norm": 0.13621224462985992, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 109427949.0, + "reward": 0.6640625, + "reward_std": 0.1840072125196457, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 187 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3957.0, + "completions/mean_length": 712.3370971679688, + "completions/mean_terminated_length": 611.216064453125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 1.7556851311953352, + "grad_norm": 0.11739154905080795, + "learning_rate": 1e-06, + "loss": -0.0054, + "num_tokens": 110050723.0, + "reward": 0.59375, + "reward_std": 0.17780686914920807, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 188 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 668.0803833007812, + "completions/mean_terminated_length": 585.810302734375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.7650145772594752, + "grad_norm": 0.13567480444908142, + "learning_rate": 1e-06, + "loss": 0.0007, + "num_tokens": 110659995.0, + "reward": 0.5803571939468384, + "reward_std": 0.21418648958206177, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 189 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3633.0, + "completions/mean_length": 729.0792846679688, + "completions/mean_terminated_length": 620.4688720703125, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 1.7743440233236152, + "grad_norm": 0.14293672144412994, + "learning_rate": 1e-06, + "loss": 0.0172, + "num_tokens": 111283346.0, + "reward": 0.5859375, + "reward_std": 0.2155335396528244, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 190 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3924.0, + "completions/mean_length": 643.1596069335938, + "completions/mean_terminated_length": 531.7776489257812, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 1.7836734693877552, + "grad_norm": 0.12812790274620056, + "learning_rate": 1e-06, + "loss": 0.01, + "num_tokens": 111835281.0, + "reward": 0.5970982313156128, + "reward_std": 0.15744362771511078, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075523018836975, + "step": 191 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3160.0, + "completions/mean_length": 669.7678833007812, + "completions/mean_terminated_length": 559.2442626953125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.7930029154518952, + "grad_norm": 0.14222432672977448, + "learning_rate": 1e-06, + "loss": -0.0152, + "num_tokens": 112404353.0, + "reward": 0.5848214626312256, + "reward_std": 0.21583479642868042, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 192 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3300.0, + "completions/mean_length": 706.1027221679688, + "completions/mean_terminated_length": 612.802734375, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 1.802332361516035, + "grad_norm": 0.12013652920722961, + "learning_rate": 1e-06, + "loss": -0.0086, + "num_tokens": 113031213.0, + "reward": 0.5680803656578064, + "reward_std": 0.16476880013942719, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 193 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2510.0, + "completions/mean_length": 643.9330444335938, + "completions/mean_terminated_length": 532.5760498046875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 1.811661807580175, + "grad_norm": 0.14051838219165802, + "learning_rate": 1e-06, + "loss": -0.014, + "num_tokens": 113578945.0, + "reward": 0.5915178656578064, + "reward_std": 0.17081941664218903, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 194 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2394.0, + "completions/mean_length": 614.1763916015625, + "completions/mean_terminated_length": 550.8704223632812, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 1.820991253644315, + "grad_norm": 0.12363959103822708, + "learning_rate": 1e-06, + "loss": 0.0016, + "num_tokens": 114148095.0, + "reward": 0.590401828289032, + "reward_std": 0.18411710858345032, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3743.0, + "completions/mean_length": 654.1183471679688, + "completions/mean_terminated_length": 571.5131225585938, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 1.8303206997084547, + "grad_norm": 0.1257123202085495, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 114731041.0, + "reward": 0.637276828289032, + "reward_std": 0.17686115205287933, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4025.0, + "completions/mean_length": 733.7913208007812, + "completions/mean_terminated_length": 629.3268432617188, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 1.8396501457725947, + "grad_norm": 0.13334035873413086, + "learning_rate": 1e-06, + "loss": 0.0097, + "num_tokens": 115378734.0, + "reward": 0.4464285969734192, + "reward_std": 0.225978285074234, + "rewards/verify_math_reward/mean": 0.4464285671710968, + "rewards/verify_math_reward/std": 0.4973995089530945, + "step": 197 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3640.0, + "completions/mean_length": 695.8035888671875, + "completions/mean_terminated_length": 565.7844848632812, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 1.8489795918367347, + "grad_norm": 0.13430985808372498, + "learning_rate": 1e-06, + "loss": -0.0022, + "num_tokens": 115962542.0, + "reward": 0.5011160969734192, + "reward_std": 0.18840248882770538, + "rewards/verify_math_reward/mean": 0.5011160969734192, + "rewards/verify_math_reward/std": 0.5002779960632324, + "step": 198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4055.0, + "completions/mean_length": 678.193115234375, + "completions/mean_terminated_length": 551.607666015625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 1.8583090379008746, + "grad_norm": 0.1244550496339798, + "learning_rate": 1e-06, + "loss": 0.0052, + "num_tokens": 116519699.0, + "reward": 0.5647321939468384, + "reward_std": 0.1658935248851776, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 199 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3505.0, + "completions/mean_length": 670.671875, + "completions/mean_terminated_length": 564.2462768554688, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 1.8676384839650146, + "grad_norm": 0.11111914366483688, + "learning_rate": 1e-06, + "loss": 0.0095, + "num_tokens": 117101421.0, + "reward": 0.6428571939468384, + "reward_std": 0.1320803463459015, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 200 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3310.0, + "completions/mean_length": 639.5513916015625, + "completions/mean_terminated_length": 544.419677734375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 1.8769679300291546, + "grad_norm": 0.14538905024528503, + "learning_rate": 1e-06, + "loss": -0.0136, + "num_tokens": 117656771.0, + "reward": 0.574776828289032, + "reward_std": 0.2198163866996765, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 201 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4096.0, + "completions/mean_length": 720.6607666015625, + "completions/mean_terminated_length": 607.7600708007812, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 1.8862973760932946, + "grad_norm": 0.12490031123161316, + "learning_rate": 1e-06, + "loss": 0.008, + "num_tokens": 118263779.0, + "reward": 0.5881696939468384, + "reward_std": 0.19091658294200897, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 202 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3561.0, + "completions/mean_length": 665.1574096679688, + "completions/mean_terminated_length": 578.7974853515625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 1.8956268221574344, + "grad_norm": 0.14387813210487366, + "learning_rate": 1e-06, + "loss": -0.0046, + "num_tokens": 118861440.0, + "reward": 0.5502232313156128, + "reward_std": 0.22428692877292633, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3222.0, + "completions/mean_length": 674.015625, + "completions/mean_terminated_length": 587.878662109375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 1.9049562682215744, + "grad_norm": 0.13967692852020264, + "learning_rate": 1e-06, + "loss": -0.0065, + "num_tokens": 119472198.0, + "reward": 0.582589328289032, + "reward_std": 0.21132442355155945, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 204 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2532.0, + "completions/mean_length": 670.1942138671875, + "completions/mean_terminated_length": 583.9610595703125, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 1.9142857142857141, + "grad_norm": 0.12676329910755157, + "learning_rate": 1e-06, + "loss": -0.0181, + "num_tokens": 120068396.0, + "reward": 0.5424107313156128, + "reward_std": 0.19636492431163788, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763562679291, + "step": 205 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3702.0, + "completions/mean_length": 768.0022583007812, + "completions/mean_terminated_length": 632.7177734375, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 1.9236151603498541, + "grad_norm": 0.12641918659210205, + "learning_rate": 1e-06, + "loss": -0.0208, + "num_tokens": 120692790.0, + "reward": 0.5613839626312256, + "reward_std": 0.18370550870895386, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 206 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4066.0, + "completions/mean_length": 699.052490234375, + "completions/mean_terminated_length": 573.2395629882812, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 1.9329446064139941, + "grad_norm": 0.13479377329349518, + "learning_rate": 1e-06, + "loss": -0.0054, + "num_tokens": 121281517.0, + "reward": 0.609375, + "reward_std": 0.18986602127552032, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 207 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3974.0, + "completions/mean_length": 702.3873291015625, + "completions/mean_terminated_length": 616.9644775390625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.9422740524781341, + "grad_norm": 0.12791851162910461, + "learning_rate": 1e-06, + "loss": -0.0002, + "num_tokens": 121911736.0, + "reward": 0.5546875, + "reward_std": 0.22090084850788116, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2966.0, + "completions/mean_length": 572.6261596679688, + "completions/mean_terminated_length": 512.6367797851562, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.951603498542274, + "grad_norm": 0.14209264516830444, + "learning_rate": 1e-06, + "loss": 0.0075, + "num_tokens": 122448593.0, + "reward": 0.6696428656578064, + "reward_std": 0.17728513479232788, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2354.0, + "completions/mean_length": 627.505615234375, + "completions/mean_terminated_length": 548.316162109375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 1.960932944606414, + "grad_norm": 0.15689514577388763, + "learning_rate": 1e-06, + "loss": 0.0121, + "num_tokens": 123011542.0, + "reward": 0.6417410969734192, + "reward_std": 0.22984997928142548, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 210 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2622.0, + "completions/mean_length": 625.239990234375, + "completions/mean_terminated_length": 578.1256103515625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.970262390670554, + "grad_norm": 0.1276608407497406, + "learning_rate": 1e-06, + "loss": 0.015, + "num_tokens": 123612525.0, + "reward": 0.621651828289032, + "reward_std": 0.1978694498538971, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3985.0, + "completions/mean_length": 634.7767944335938, + "completions/mean_terminated_length": 551.7073974609375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 1.9795918367346939, + "grad_norm": 0.12583675980567932, + "learning_rate": 1e-06, + "loss": 0.0092, + "num_tokens": 124179805.0, + "reward": 0.6205357313156128, + "reward_std": 0.17829149961471558, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 212 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3283.0, + "completions/mean_length": 607.0256958007812, + "completions/mean_terminated_length": 531.4378662109375, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 1.9889212827988338, + "grad_norm": 0.11668175458908081, + "learning_rate": 1e-06, + "loss": 0.0015, + "num_tokens": 124738628.0, + "reward": 0.6227678656578064, + "reward_std": 0.1591670960187912, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022727272727272707, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3538.0, + "completions/mean_length": 648.7698974609375, + "completions/mean_terminated_length": 568.6017456054688, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 1.9982507288629736, + "grad_norm": 0.13209541141986847, + "learning_rate": 1e-06, + "loss": 0.0008, + "num_tokens": 125315000.0, + "reward": 0.637276828289032, + "reward_std": 0.15916749835014343, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 214 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3347.0, + "completions/mean_length": 642.4342041015625, + "completions/mean_terminated_length": 563.5855712890625, + "completions/min_length": 84.0, + "completions/min_terminated_length": 84.0, + "epoch": 2.00932944606414, + "grad_norm": 0.13432128727436066, + "learning_rate": 1e-06, + "loss": -0.0043, + "num_tokens": 125897189.0, + "reward": 0.5959821939468384, + "reward_std": 0.20455309748649597, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.4909749925136566, + "step": 215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3815.0, + "completions/mean_length": 745.1439819335938, + "completions/mean_terminated_length": 580.3477783203125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 2.01865889212828, + "grad_norm": 0.1319049894809723, + "learning_rate": 1e-06, + "loss": 0.0042, + "num_tokens": 126482486.0, + "reward": 0.5792410969734192, + "reward_std": 0.18362995982170105, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 216 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3784.0, + "completions/mean_length": 691.177490234375, + "completions/mean_terminated_length": 569.1549072265625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 2.02798833819242, + "grad_norm": 0.1434415578842163, + "learning_rate": 1e-06, + "loss": -0.0032, + "num_tokens": 127068501.0, + "reward": 0.5848214626312256, + "reward_std": 0.18968310952186584, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 217 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 738.1116333007812, + "completions/mean_terminated_length": 649.6449584960938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.03731778425656, + "grad_norm": 0.12249483913183212, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 127726953.0, + "reward": 0.5412946939468384, + "reward_std": 0.19047404825687408, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 218 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2979.0, + "completions/mean_length": 612.349365234375, + "completions/mean_terminated_length": 528.74169921875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 2.0466472303206995, + "grad_norm": 0.1585049331188202, + "learning_rate": 1e-06, + "loss": -0.0165, + "num_tokens": 128276946.0, + "reward": 0.660714328289032, + "reward_std": 0.19697223603725433, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3931.0, + "completions/mean_length": 690.4531860351562, + "completions/mean_terminated_length": 632.469970703125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 2.0559766763848395, + "grad_norm": 0.11900050193071365, + "learning_rate": 1e-06, + "loss": -0.0073, + "num_tokens": 128928688.0, + "reward": 0.527901828289032, + "reward_std": 0.2089158594608307, + "rewards/verify_math_reward/mean": 0.5279017686843872, + "rewards/verify_math_reward/std": 0.49949970841407776, + "step": 220 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2485.0, + "completions/mean_length": 631.4888916015625, + "completions/mean_terminated_length": 548.340576171875, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 2.0653061224489795, + "grad_norm": 0.12185671925544739, + "learning_rate": 1e-06, + "loss": 0.004, + "num_tokens": 129496398.0, + "reward": 0.6205357313156128, + "reward_std": 0.15931907296180725, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 221 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3327.0, + "completions/mean_length": 696.169677734375, + "completions/mean_terminated_length": 598.5855102539062, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 2.0746355685131195, + "grad_norm": 0.13736842572689056, + "learning_rate": 1e-06, + "loss": 0.0084, + "num_tokens": 130109414.0, + "reward": 0.582589328289032, + "reward_std": 0.1922753006219864, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.493407279253006, + "step": 222 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3891.0, + "completions/mean_length": 603.3817138671875, + "completions/mean_terminated_length": 551.9614868164062, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 2.0839650145772595, + "grad_norm": 0.13463492691516876, + "learning_rate": 1e-06, + "loss": 0.0113, + "num_tokens": 130697916.0, + "reward": 0.6595982313156128, + "reward_std": 0.1837783008813858, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3457.0, + "completions/mean_length": 706.3772583007812, + "completions/mean_terminated_length": 588.9537963867188, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 2.0932944606413995, + "grad_norm": 0.1270330399274826, + "learning_rate": 1e-06, + "loss": -0.0104, + "num_tokens": 131293654.0, + "reward": 0.6573660969734192, + "reward_std": 0.16532759368419647, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 224 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3808.0, + "completions/mean_length": 744.7210083007812, + "completions/mean_terminated_length": 628.6258544921875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 2.1026239067055394, + "grad_norm": 0.12777270376682281, + "learning_rate": 1e-06, + "loss": -0.0034, + "num_tokens": 131936452.0, + "reward": 0.5345982313156128, + "reward_std": 0.19794613122940063, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 225 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3455.0, + "completions/mean_length": 712.6752319335938, + "completions/mean_terminated_length": 615.5648803710938, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 2.1119533527696794, + "grad_norm": 0.12400238960981369, + "learning_rate": 1e-06, + "loss": 0.0095, + "num_tokens": 132564457.0, + "reward": 0.6160714626312256, + "reward_std": 0.2014826238155365, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 226 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2392.0, + "completions/mean_length": 635.0111694335938, + "completions/mean_terminated_length": 560.0296020507812, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 2.1212827988338194, + "grad_norm": 0.13383056223392487, + "learning_rate": 1e-06, + "loss": 0.0012, + "num_tokens": 133138355.0, + "reward": 0.6462053656578064, + "reward_std": 0.1817513257265091, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 227 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2308.0, + "completions/mean_length": 729.9531860351562, + "completions/mean_terminated_length": 580.8741455078125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 2.130612244897959, + "grad_norm": 0.13521888852119446, + "learning_rate": 1e-06, + "loss": 0.0125, + "num_tokens": 133725153.0, + "reward": 0.6261160969734192, + "reward_std": 0.18532174825668335, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 228 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3899.0, + "completions/mean_length": 751.7991333007812, + "completions/mean_terminated_length": 639.9400024414062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.139941690962099, + "grad_norm": 0.1175265908241272, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 134374917.0, + "reward": 0.5703125, + "reward_std": 0.16758599877357483, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 229 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3413.0, + "completions/mean_length": 643.3671875, + "completions/mean_terminated_length": 568.5667114257812, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 2.149271137026239, + "grad_norm": 0.12349607795476913, + "learning_rate": 1e-06, + "loss": 0.0048, + "num_tokens": 134959142.0, + "reward": 0.6395089626312256, + "reward_std": 0.18867096304893494, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 230 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3900.0, + "completions/mean_length": 674.7578125, + "completions/mean_terminated_length": 568.4591674804688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 2.158600583090379, + "grad_norm": 0.12909038364887238, + "learning_rate": 1e-06, + "loss": -0.0006, + "num_tokens": 135542869.0, + "reward": 0.6026785969734192, + "reward_std": 0.15849308669567108, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 231 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3953.0, + "completions/mean_length": 806.5513916015625, + "completions/mean_terminated_length": 636.6737060546875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 2.167930029154519, + "grad_norm": 0.126162588596344, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 136175691.0, + "reward": 0.527901828289032, + "reward_std": 0.17333535850048065, + "rewards/verify_math_reward/mean": 0.5279017686843872, + "rewards/verify_math_reward/std": 0.49949970841407776, + "step": 232 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3747.0, + "completions/mean_length": 686.4486694335938, + "completions/mean_terminated_length": 628.3972778320312, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 2.177259475218659, + "grad_norm": 0.13025717437267303, + "learning_rate": 1e-06, + "loss": 0.0162, + "num_tokens": 136820573.0, + "reward": 0.5569196939468384, + "reward_std": 0.1826963573694229, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 233 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3287.0, + "completions/mean_length": 633.8560791015625, + "completions/mean_terminated_length": 562.878173828125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 2.186588921282799, + "grad_norm": 0.13308566808700562, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 137422044.0, + "reward": 0.5546875, + "reward_std": 0.1856248527765274, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3675.0, + "completions/mean_length": 647.9006958007812, + "completions/mean_terminated_length": 536.671630859375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 2.195918367346939, + "grad_norm": 0.13044235110282898, + "learning_rate": 1e-06, + "loss": -0.0081, + "num_tokens": 137975171.0, + "reward": 0.6283482313156128, + "reward_std": 0.18187008798122406, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3509.0, + "completions/mean_length": 646.7444458007812, + "completions/mean_terminated_length": 559.9210205078125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 2.205247813411079, + "grad_norm": 0.13473135232925415, + "learning_rate": 1e-06, + "loss": 0.0113, + "num_tokens": 138553974.0, + "reward": 0.625, + "reward_std": 0.170893594622612, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 236 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4016.0, + "completions/mean_length": 678.3817138671875, + "completions/mean_terminated_length": 584.3187866210938, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 2.2145772594752184, + "grad_norm": 0.14180408418178558, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 139146908.0, + "reward": 0.5546875, + "reward_std": 0.20688636600971222, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 237 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3590.0, + "completions/mean_length": 685.0391235351562, + "completions/mean_terminated_length": 570.9468994140625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 2.2239067055393584, + "grad_norm": 0.12916496396064758, + "learning_rate": 1e-06, + "loss": -0.0024, + "num_tokens": 139723655.0, + "reward": 0.5859375, + "reward_std": 0.17723983526229858, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 238 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 687.3460083007812, + "completions/mean_terminated_length": 544.6581420898438, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 2.2332361516034984, + "grad_norm": 0.13915696740150452, + "learning_rate": 1e-06, + "loss": -0.0056, + "num_tokens": 140283253.0, + "reward": 0.6350446939468384, + "reward_std": 0.1831093281507492, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 239 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3836.0, + "completions/mean_length": 685.9866333007812, + "completions/mean_terminated_length": 567.8568115234375, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 2.2425655976676384, + "grad_norm": 0.1293559968471527, + "learning_rate": 1e-06, + "loss": 0.0003, + "num_tokens": 140850161.0, + "reward": 0.6662946939468384, + "reward_std": 0.18013623356819153, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 240 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3344.0, + "completions/mean_length": 684.4576416015625, + "completions/mean_terminated_length": 586.5372924804688, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 2.2518950437317784, + "grad_norm": 0.13639135658740997, + "learning_rate": 1e-06, + "loss": 0.0034, + "num_tokens": 141449515.0, + "reward": 0.613839328289032, + "reward_std": 0.20136500895023346, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 241 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2907.0, + "completions/mean_length": 647.484375, + "completions/mean_terminated_length": 576.785888671875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 2.2612244897959184, + "grad_norm": 0.14444217085838318, + "learning_rate": 1e-06, + "loss": -0.0011, + "num_tokens": 142035901.0, + "reward": 0.640625, + "reward_std": 0.19114483892917633, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 242 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2298.0, + "completions/mean_length": 644.4207763671875, + "completions/mean_terminated_length": 549.4231567382812, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 2.2705539358600584, + "grad_norm": 0.1451975256204605, + "learning_rate": 1e-06, + "loss": 0.005, + "num_tokens": 142600654.0, + "reward": 0.598214328289032, + "reward_std": 0.19369465112686157, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 243 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2157.0, + "completions/mean_length": 631.1741333007812, + "completions/mean_terminated_length": 523.5213012695312, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 2.2798833819241984, + "grad_norm": 0.13140873610973358, + "learning_rate": 1e-06, + "loss": 0.0099, + "num_tokens": 143132658.0, + "reward": 0.6361607313156128, + "reward_std": 0.12726575136184692, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 244 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3885.0, + "completions/mean_length": 684.7846069335938, + "completions/mean_terminated_length": 598.9187622070312, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.2892128279883384, + "grad_norm": 0.1164131760597229, + "learning_rate": 1e-06, + "loss": 0.0106, + "num_tokens": 143739369.0, + "reward": 0.6004464626312256, + "reward_std": 0.15364499390125275, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 245 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2779.0, + "completions/mean_length": 695.786865234375, + "completions/mean_terminated_length": 598.1917114257812, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.298542274052478, + "grad_norm": 0.1390455663204193, + "learning_rate": 1e-06, + "loss": -0.0035, + "num_tokens": 144353386.0, + "reward": 0.5546875, + "reward_std": 0.22875411808490753, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 246 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2285.0, + "completions/mean_length": 648.5792846679688, + "completions/mean_terminated_length": 585.8988647460938, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "epoch": 2.307871720116618, + "grad_norm": 0.11788541823625565, + "learning_rate": 1e-06, + "loss": 0.0007, + "num_tokens": 144964665.0, + "reward": 0.590401828289032, + "reward_std": 0.14928476512432098, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 247 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4076.0, + "completions/mean_length": 652.0502319335938, + "completions/mean_terminated_length": 557.2625732421875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 2.317201166180758, + "grad_norm": 0.1455429196357727, + "learning_rate": 1e-06, + "loss": 0.0082, + "num_tokens": 145541598.0, + "reward": 0.5915178656578064, + "reward_std": 0.2091120034456253, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 248 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3990.0, + "completions/mean_length": 765.8772583007812, + "completions/mean_terminated_length": 614.3314208984375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 2.326530612244898, + "grad_norm": 0.12138891220092773, + "learning_rate": 1e-06, + "loss": -0.0091, + "num_tokens": 146158736.0, + "reward": 0.5837053656578064, + "reward_std": 0.1761525422334671, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 249 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2784.0, + "completions/mean_length": 676.435302734375, + "completions/mean_terminated_length": 574.2413940429688, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 2.335860058309038, + "grad_norm": 0.13957031071186066, + "learning_rate": 1e-06, + "loss": -0.0031, + "num_tokens": 146741390.0, + "reward": 0.6741071939468384, + "reward_std": 0.18224169313907623, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 250 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4029.0, + "completions/mean_length": 684.7422485351562, + "completions/mean_terminated_length": 574.7015991210938, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 2.345189504373178, + "grad_norm": 0.13807563483715057, + "learning_rate": 1e-06, + "loss": 0.0082, + "num_tokens": 147332071.0, + "reward": 0.5814732313156128, + "reward_std": 0.1879139393568039, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 251 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 663.5535888671875, + "completions/mean_terminated_length": 552.8294677734375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.354518950437318, + "grad_norm": 0.1375046819448471, + "learning_rate": 1e-06, + "loss": -0.0161, + "num_tokens": 147897255.0, + "reward": 0.598214328289032, + "reward_std": 0.19614629447460175, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 252 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3809.0, + "completions/mean_length": 698.0201416015625, + "completions/mean_terminated_length": 580.30712890625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 2.363848396501458, + "grad_norm": 0.12242113053798676, + "learning_rate": 1e-06, + "loss": -0.0052, + "num_tokens": 148484937.0, + "reward": 0.6417410969734192, + "reward_std": 0.17878076434135437, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 253 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3486.0, + "completions/mean_length": 735.7656860351562, + "completions/mean_terminated_length": 639.3179931640625, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 2.373177842565598, + "grad_norm": 0.1236719936132431, + "learning_rate": 1e-06, + "loss": -0.01, + "num_tokens": 149124239.0, + "reward": 0.5502232313156128, + "reward_std": 0.19235128164291382, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 254 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3532.0, + "completions/mean_length": 633.0256958007812, + "completions/mean_terminated_length": 562.03076171875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 2.3825072886297374, + "grad_norm": 0.13827461004257202, + "learning_rate": 1e-06, + "loss": 0.0011, + "num_tokens": 149702734.0, + "reward": 0.6462053656578064, + "reward_std": 0.1798752099275589, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 255 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3421.0, + "completions/mean_length": 687.4063110351562, + "completions/mean_terminated_length": 625.4318237304688, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.3918367346938774, + "grad_norm": 0.12812216579914093, + "learning_rate": 1e-06, + "loss": 0.0055, + "num_tokens": 150346810.0, + "reward": 0.5881696939468384, + "reward_std": 0.1626291275024414, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924396276473999, + "step": 256 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3126.0, + "completions/mean_length": 762.3214721679688, + "completions/mean_terminated_length": 626.8060302734375, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 2.4011661807580174, + "grad_norm": 0.12976829707622528, + "learning_rate": 1e-06, + "loss": -0.0068, + "num_tokens": 150975002.0, + "reward": 0.5926339626312256, + "reward_std": 0.2015921026468277, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 257 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2072.0, + "completions/mean_length": 613.458740234375, + "completions/mean_terminated_length": 558.1802978515625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.4104956268221573, + "grad_norm": 0.1474219262599945, + "learning_rate": 1e-06, + "loss": 0.0146, + "num_tokens": 151552221.0, + "reward": 0.6383928656578064, + "reward_std": 0.21327468752861023, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 258 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3121.0, + "completions/mean_length": 644.9910888671875, + "completions/mean_terminated_length": 574.241455078125, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.4198250728862973, + "grad_norm": 0.13549746572971344, + "learning_rate": 1e-06, + "loss": -0.0062, + "num_tokens": 152147941.0, + "reward": 0.6484375, + "reward_std": 0.18719924986362457, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3575.0, + "completions/mean_length": 712.3973388671875, + "completions/mean_terminated_length": 583.0127563476562, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.4291545189504373, + "grad_norm": 0.14896702766418457, + "learning_rate": 1e-06, + "loss": -0.0145, + "num_tokens": 152742713.0, + "reward": 0.5859375, + "reward_std": 0.22575145959854126, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 260 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3941.0, + "completions/mean_length": 727.5491333007812, + "completions/mean_terminated_length": 594.686767578125, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "epoch": 2.4384839650145773, + "grad_norm": 0.11539198458194733, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 153350613.0, + "reward": 0.6160714626312256, + "reward_std": 0.16044840216636658, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 261 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3950.0, + "completions/mean_length": 652.4910888671875, + "completions/mean_terminated_length": 565.8123168945312, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 2.4478134110787173, + "grad_norm": 0.14026306569576263, + "learning_rate": 1e-06, + "loss": 0.0047, + "num_tokens": 153933973.0, + "reward": 0.5792410969734192, + "reward_std": 0.19610463082790375, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 262 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3971.0, + "completions/mean_length": 673.318115234375, + "completions/mean_terminated_length": 566.9746704101562, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 2.4571428571428573, + "grad_norm": 0.13149487972259521, + "learning_rate": 1e-06, + "loss": 0.0023, + "num_tokens": 154511482.0, + "reward": 0.6149553656578064, + "reward_std": 0.15774421393871307, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 263 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2637.0, + "completions/mean_length": 708.6160888671875, + "completions/mean_terminated_length": 579.0869140625, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 2.466472303206997, + "grad_norm": 0.12381426990032196, + "learning_rate": 1e-06, + "loss": 0.006, + "num_tokens": 155105466.0, + "reward": 0.590401828289032, + "reward_std": 0.18670931458473206, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3791.0, + "completions/mean_length": 747.6719360351562, + "completions/mean_terminated_length": 611.5609741210938, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 2.4758017492711373, + "grad_norm": 0.1241414025425911, + "learning_rate": 1e-06, + "loss": -0.004, + "num_tokens": 155720764.0, + "reward": 0.5915178656578064, + "reward_std": 0.16950349509716034, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 265 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3594.0, + "completions/mean_length": 702.8370971679688, + "completions/mean_terminated_length": 569.0, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.485131195335277, + "grad_norm": 0.13266952335834503, + "learning_rate": 1e-06, + "loss": 0.0092, + "num_tokens": 156294586.0, + "reward": 0.65625, + "reward_std": 0.1885526329278946, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 266 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3111.0, + "completions/mean_length": 712.3895263671875, + "completions/mean_terminated_length": 599.2122192382812, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 2.494460641399417, + "grad_norm": 0.13610923290252686, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 156900311.0, + "reward": 0.6015625, + "reward_std": 0.20452313125133514, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 267 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4058.0, + "completions/mean_length": 845.966552734375, + "completions/mean_terminated_length": 657.9480590820312, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 2.503790087463557, + "grad_norm": 0.13023985922336578, + "learning_rate": 1e-06, + "loss": -0.0165, + "num_tokens": 157556569.0, + "reward": 0.5703125, + "reward_std": 0.20249363780021667, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 268 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3107.0, + "completions/mean_length": 696.1239013671875, + "completions/mean_terminated_length": 598.5384521484375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.513119533527697, + "grad_norm": 0.14194142818450928, + "learning_rate": 1e-06, + "loss": -0.0102, + "num_tokens": 158162736.0, + "reward": 0.645089328289032, + "reward_std": 0.19490069150924683, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 269 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2937.0, + "completions/mean_length": 694.4676513671875, + "completions/mean_terminated_length": 584.7407836914062, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 2.522448979591837, + "grad_norm": 0.127032071352005, + "learning_rate": 1e-06, + "loss": -0.0056, + "num_tokens": 158764067.0, + "reward": 0.5881696939468384, + "reward_std": 0.16969217360019684, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924396276473999, + "step": 270 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3898.0, + "completions/mean_length": 819.3303833007812, + "completions/mean_terminated_length": 658.1826171875, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 2.5317784256559768, + "grad_norm": 0.11828536540269852, + "learning_rate": 1e-06, + "loss": -0.0126, + "num_tokens": 159437155.0, + "reward": 0.5636160969734192, + "reward_std": 0.16713954508304596, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4081.0, + "completions/mean_length": 692.1060791015625, + "completions/mean_terminated_length": 586.3463745117188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.5411078717201168, + "grad_norm": 0.128693625330925, + "learning_rate": 1e-06, + "loss": 0.0013, + "num_tokens": 160041778.0, + "reward": 0.6127232313156128, + "reward_std": 0.1709274798631668, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 272 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2583.0, + "completions/mean_length": 720.4855346679688, + "completions/mean_terminated_length": 607.5789794921875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 2.5504373177842563, + "grad_norm": 0.13341568410396576, + "learning_rate": 1e-06, + "loss": -0.0138, + "num_tokens": 160650685.0, + "reward": 0.6350446939468384, + "reward_std": 0.21669355034828186, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2881.0, + "completions/mean_length": 834.2064819335938, + "completions/mean_terminated_length": 689.7447509765625, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 2.5597667638483967, + "grad_norm": 0.1141742542386055, + "learning_rate": 1e-06, + "loss": -0.0062, + "num_tokens": 161327294.0, + "reward": 0.5569196939468384, + "reward_std": 0.18318237364292145, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 274 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3884.0, + "completions/mean_length": 735.2232666015625, + "completions/mean_terminated_length": 618.799072265625, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.5690962099125363, + "grad_norm": 0.14005409181118011, + "learning_rate": 1e-06, + "loss": 0.0004, + "num_tokens": 161950254.0, + "reward": 0.5736607313156128, + "reward_std": 0.21793986856937408, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 275 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3661.0, + "completions/mean_length": 745.7188110351562, + "completions/mean_terminated_length": 629.658203125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 2.5784256559766763, + "grad_norm": 0.1346922665834427, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 162592810.0, + "reward": 0.6417410969734192, + "reward_std": 0.15308299660682678, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 276 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3345.0, + "completions/mean_length": 690.3192138671875, + "completions/mean_terminated_length": 596.5848388671875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.5877551020408163, + "grad_norm": 0.12718650698661804, + "learning_rate": 1e-06, + "loss": -0.0014, + "num_tokens": 163199968.0, + "reward": 0.5993303656578064, + "reward_std": 0.16612036526203156, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 277 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3958.0, + "completions/mean_length": 733.9542846679688, + "completions/mean_terminated_length": 617.4861450195312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.5970845481049563, + "grad_norm": 0.1426485925912857, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 163827303.0, + "reward": 0.566964328289032, + "reward_std": 0.227071613073349, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 278 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3851.0, + "completions/mean_length": 717.0826416015625, + "completions/mean_terminated_length": 595.9884033203125, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 2.6064139941690962, + "grad_norm": 0.12692730128765106, + "learning_rate": 1e-06, + "loss": -0.0129, + "num_tokens": 164429857.0, + "reward": 0.6473214626312256, + "reward_std": 0.18617936968803406, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 279 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3858.0, + "completions/mean_length": 691.7109985351562, + "completions/mean_terminated_length": 589.9735717773438, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.6157434402332362, + "grad_norm": 0.1253710389137268, + "learning_rate": 1e-06, + "loss": -0.0195, + "num_tokens": 165027518.0, + "reward": 0.5892857313156128, + "reward_std": 0.1799846738576889, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 280 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3746.0, + "completions/mean_length": 627.1495971679688, + "completions/mean_terminated_length": 551.9976806640625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.6250728862973762, + "grad_norm": 0.12407125532627106, + "learning_rate": 1e-06, + "loss": -0.0046, + "num_tokens": 165592540.0, + "reward": 0.6975446939468384, + "reward_std": 0.1655130386352539, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 281 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3987.0, + "completions/mean_length": 686.6830444335938, + "completions/mean_terminated_length": 600.864990234375, + "completions/min_length": 84.0, + "completions/min_terminated_length": 84.0, + "epoch": 2.6344023323615158, + "grad_norm": 0.13460440933704376, + "learning_rate": 1e-06, + "loss": 0.0021, + "num_tokens": 166202360.0, + "reward": 0.6741071939468384, + "reward_std": 0.21467113494873047, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 282 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3911.0, + "completions/mean_length": 712.7678833007812, + "completions/mean_terminated_length": 579.322509765625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 2.643731778425656, + "grad_norm": 0.13798287510871887, + "learning_rate": 1e-06, + "loss": 0.0041, + "num_tokens": 166782248.0, + "reward": 0.6439732313156128, + "reward_std": 0.18077561259269714, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3891.0, + "completions/mean_length": 624.8058471679688, + "completions/mean_terminated_length": 569.70751953125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 2.6530612244897958, + "grad_norm": 0.13172586262226105, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 167379986.0, + "reward": 0.6584821939468384, + "reward_std": 0.20380382239818573, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 284 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 731.4319458007812, + "completions/mean_terminated_length": 626.8941650390625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 2.6623906705539357, + "grad_norm": 0.13110998272895813, + "learning_rate": 1e-06, + "loss": 0.0097, + "num_tokens": 168014325.0, + "reward": 0.6049107313156128, + "reward_std": 0.20031006634235382, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3399.0, + "completions/mean_length": 836.4029541015625, + "completions/mean_terminated_length": 684.0852661132812, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 2.6717201166180757, + "grad_norm": 0.12780140340328217, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 168688862.0, + "reward": 0.5033482313156128, + "reward_std": 0.20929564535617828, + "rewards/verify_math_reward/mean": 0.5033482313156128, + "rewards/verify_math_reward/std": 0.5002680420875549, + "step": 286 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3452.0, + "completions/mean_length": 806.0279541015625, + "completions/mean_terminated_length": 640.1793212890625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.6810495626822157, + "grad_norm": 0.14262090623378754, + "learning_rate": 1e-06, + "loss": -0.0193, + "num_tokens": 169329551.0, + "reward": 0.5658482313156128, + "reward_std": 0.21023672819137573, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 287 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3893.0, + "completions/mean_length": 756.5324096679688, + "completions/mean_terminated_length": 656.732177734375, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 2.6903790087463557, + "grad_norm": 0.13744673132896423, + "learning_rate": 1e-06, + "loss": -0.0134, + "num_tokens": 169985756.0, + "reward": 0.5892857313156128, + "reward_std": 0.209860160946846, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 288 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3752.0, + "completions/mean_length": 791.9910888671875, + "completions/mean_terminated_length": 637.5980834960938, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 2.6997084548104957, + "grad_norm": 0.12104486674070358, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 170615828.0, + "reward": 0.5636160969734192, + "reward_std": 0.1975242644548416, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 289 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3303.0, + "completions/mean_length": 699.5926513671875, + "completions/mean_terminated_length": 585.9873046875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 2.7090379008746357, + "grad_norm": 0.13590072095394135, + "learning_rate": 1e-06, + "loss": -0.0094, + "num_tokens": 171211327.0, + "reward": 0.6428571939468384, + "reward_std": 0.21537834405899048, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 290 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3208.0, + "completions/mean_length": 694.810302734375, + "completions/mean_terminated_length": 617.1575317382812, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 2.7183673469387752, + "grad_norm": 0.12708239257335663, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 171842253.0, + "reward": 0.6383928656578064, + "reward_std": 0.18445271253585815, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 291 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 850.9721069335938, + "completions/mean_terminated_length": 634.6369018554688, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 2.7276967930029157, + "grad_norm": 0.15510284900665283, + "learning_rate": 1e-06, + "loss": -0.0266, + "num_tokens": 172477716.0, + "reward": 0.5267857313156128, + "reward_std": 0.22793109714984894, + "rewards/verify_math_reward/mean": 0.5267857313156128, + "rewards/verify_math_reward/std": 0.4995608627796173, + "step": 292 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4044.0, + "completions/mean_length": 826.5535888671875, + "completions/mean_terminated_length": 669.7730712890625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 2.7370262390670552, + "grad_norm": 0.12435416132211685, + "learning_rate": 1e-06, + "loss": -0.0161, + "num_tokens": 173141332.0, + "reward": 0.5859375, + "reward_std": 0.20125439763069153, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 293 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 755.4810791015625, + "completions/mean_terminated_length": 619.6875610351562, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 2.746355685131195, + "grad_norm": 0.13841715455055237, + "learning_rate": 1e-06, + "loss": -0.003, + "num_tokens": 173775715.0, + "reward": 0.5580357313156128, + "reward_std": 0.1816764622926712, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3861.0, + "completions/mean_length": 695.7176513671875, + "completions/mean_terminated_length": 614.11083984375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "epoch": 2.755685131195335, + "grad_norm": 0.11661636084318161, + "learning_rate": 1e-06, + "loss": -0.0054, + "num_tokens": 174398982.0, + "reward": 0.660714328289032, + "reward_std": 0.16758419573307037, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2765.0, + "completions/mean_length": 758.3739013671875, + "completions/mean_terminated_length": 630.7473754882812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 2.765014577259475, + "grad_norm": 0.12156891077756882, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 175031981.0, + "reward": 0.6149553656578064, + "reward_std": 0.18878085911273956, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 296 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3295.0, + "completions/mean_length": 734.0658569335938, + "completions/mean_terminated_length": 617.6016235351562, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 2.774344023323615, + "grad_norm": 0.119770847260952, + "learning_rate": 1e-06, + "loss": 0.0125, + "num_tokens": 175655112.0, + "reward": 0.5558035969734192, + "reward_std": 0.1534930318593979, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715372920036316, + "step": 297 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3443.0, + "completions/mean_length": 753.6842041015625, + "completions/mean_terminated_length": 645.8674926757812, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 2.783673469387755, + "grad_norm": 0.12274106591939926, + "learning_rate": 1e-06, + "loss": -0.0175, + "num_tokens": 176307997.0, + "reward": 0.5881696939468384, + "reward_std": 0.17292234301567078, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 298 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2867.0, + "completions/mean_length": 721.5748291015625, + "completions/mean_terminated_length": 616.730712890625, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 2.793002915451895, + "grad_norm": 0.13726380467414856, + "learning_rate": 1e-06, + "loss": 0.015, + "num_tokens": 176931768.0, + "reward": 0.5613839626312256, + "reward_std": 0.1970803141593933, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 299 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3998.0, + "completions/mean_length": 804.8471069335938, + "completions/mean_terminated_length": 663.086181640625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 2.8023323615160347, + "grad_norm": 0.12969516217708588, + "learning_rate": 1e-06, + "loss": -0.0093, + "num_tokens": 177585319.0, + "reward": 0.640625, + "reward_std": 0.18727383017539978, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 300 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3934.0, + "completions/mean_length": 726.7924194335938, + "completions/mean_terminated_length": 593.9002075195312, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 2.811661807580175, + "grad_norm": 0.14020316302776337, + "learning_rate": 1e-06, + "loss": 0.0162, + "num_tokens": 178177909.0, + "reward": 0.5993303656578064, + "reward_std": 0.19918397068977356, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 301 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3830.0, + "completions/mean_length": 809.2969360351562, + "completions/mean_terminated_length": 655.7125854492188, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 2.8209912536443147, + "grad_norm": 0.12125155329704285, + "learning_rate": 1e-06, + "loss": 0.0038, + "num_tokens": 178822887.0, + "reward": 0.574776828289032, + "reward_std": 0.18674279749393463, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3174.0, + "completions/mean_length": 877.5480346679688, + "completions/mean_terminated_length": 695.3714599609375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.8303206997084547, + "grad_norm": 0.11820695549249649, + "learning_rate": 1e-06, + "loss": 0.0074, + "num_tokens": 179502426.0, + "reward": 0.5848214626312256, + "reward_std": 0.1722055822610855, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 303 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3073.0, + "completions/mean_length": 800.5279541015625, + "completions/mean_terminated_length": 613.9917602539062, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.8396501457725947, + "grad_norm": 0.148163303732872, + "learning_rate": 1e-06, + "loss": -0.0133, + "num_tokens": 180114115.0, + "reward": 0.6584821939468384, + "reward_std": 0.20298148691654205, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 304 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3732.0, + "completions/mean_length": 752.1060791015625, + "completions/mean_terminated_length": 632.2670288085938, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.8489795918367347, + "grad_norm": 0.12780609726905823, + "learning_rate": 1e-06, + "loss": 0.0034, + "num_tokens": 180760290.0, + "reward": 0.5926339626312256, + "reward_std": 0.1936621367931366, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 305 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2813.0, + "completions/mean_length": 759.0078735351562, + "completions/mean_terminated_length": 590.7889404296875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.8583090379008746, + "grad_norm": 0.14202755689620972, + "learning_rate": 1e-06, + "loss": -0.0203, + "num_tokens": 181347369.0, + "reward": 0.6171875, + "reward_std": 0.19032178819179535, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 306 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 771.2701416015625, + "completions/mean_terminated_length": 611.8385620117188, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 2.8676384839650146, + "grad_norm": 0.12430532276630402, + "learning_rate": 1e-06, + "loss": -0.0189, + "num_tokens": 181954051.0, + "reward": 0.6037946939468384, + "reward_std": 0.18660052120685577, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 307 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3187.0, + "completions/mean_length": 709.0357666015625, + "completions/mean_terminated_length": 607.8161010742188, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 2.8769679300291546, + "grad_norm": 0.12658368051052094, + "learning_rate": 1e-06, + "loss": -0.0331, + "num_tokens": 182569659.0, + "reward": 0.6417410969734192, + "reward_std": 0.1687125563621521, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 308 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2317.0, + "completions/mean_length": 749.9420166015625, + "completions/mean_terminated_length": 597.6709594726562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 2.8862973760932946, + "grad_norm": 0.15306328237056732, + "learning_rate": 1e-06, + "loss": -0.0111, + "num_tokens": 183168999.0, + "reward": 0.5580357313156128, + "reward_std": 0.19185978174209595, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 309 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3690.0, + "completions/mean_length": 793.0424194335938, + "completions/mean_terminated_length": 682.5628662109375, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 2.8956268221574346, + "grad_norm": 0.1336117535829544, + "learning_rate": 1e-06, + "loss": 0.0022, + "num_tokens": 183846117.0, + "reward": 0.5658482313156128, + "reward_std": 0.2092161625623703, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 310 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3138.0, + "completions/mean_length": 654.21875, + "completions/mean_terminated_length": 571.615966796875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 2.904956268221574, + "grad_norm": 0.13915765285491943, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 184445465.0, + "reward": 0.6595982313156128, + "reward_std": 0.1670207679271698, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 311 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2999.0, + "completions/mean_length": 806.3750610351562, + "completions/mean_terminated_length": 616.0661010742188, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 2.914285714285714, + "grad_norm": 0.11749313771724701, + "learning_rate": 1e-06, + "loss": -0.0071, + "num_tokens": 185057617.0, + "reward": 0.6283482313156128, + "reward_std": 0.16330063343048096, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 312 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3283.0, + "completions/mean_length": 842.8047485351562, + "completions/mean_terminated_length": 654.603271484375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 2.923615160349854, + "grad_norm": 0.12479273229837418, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 185705722.0, + "reward": 0.5792410969734192, + "reward_std": 0.1788228303194046, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 313 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3960.0, + "completions/mean_length": 815.9564819335938, + "completions/mean_terminated_length": 678.65234375, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 2.932944606413994, + "grad_norm": 0.13630305230617523, + "learning_rate": 1e-06, + "loss": -0.0122, + "num_tokens": 186378835.0, + "reward": 0.5412946939468384, + "reward_std": 0.2210107296705246, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 314 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 789.427490234375, + "completions/mean_terminated_length": 642.9825439453125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 2.942274052478134, + "grad_norm": 0.1420001983642578, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 187025226.0, + "reward": 0.5613839626312256, + "reward_std": 0.2113226056098938, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 315 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3755.0, + "completions/mean_length": 719.5178833007812, + "completions/mean_terminated_length": 610.5990600585938, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 2.951603498542274, + "grad_norm": 0.13174320757389069, + "learning_rate": 1e-06, + "loss": 0.0008, + "num_tokens": 187650298.0, + "reward": 0.625, + "reward_std": 0.18167534470558167, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 316 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 897.2756958007812, + "completions/mean_terminated_length": 684.0274047851562, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 2.960932944606414, + "grad_norm": 0.14336876571178436, + "learning_rate": 1e-06, + "loss": -0.0131, + "num_tokens": 188312193.0, + "reward": 0.5848214626312256, + "reward_std": 0.22796174883842468, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 317 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3898.0, + "completions/mean_length": 705.3460083007812, + "completions/mean_terminated_length": 595.9700317382812, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 2.970262390670554, + "grad_norm": 0.137048602104187, + "learning_rate": 1e-06, + "loss": 0.015, + "num_tokens": 188918647.0, + "reward": 0.625, + "reward_std": 0.1834437996149063, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4074.0, + "completions/mean_length": 820.5960083007812, + "completions/mean_terminated_length": 627.01416015625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 2.979591836734694, + "grad_norm": 0.13644100725650787, + "learning_rate": 1e-06, + "loss": -0.0258, + "num_tokens": 189538885.0, + "reward": 0.59375, + "reward_std": 0.1806686669588089, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 319 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3684.0, + "completions/mean_length": 742.325927734375, + "completions/mean_terminated_length": 630.14990234375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 2.9889212827988336, + "grad_norm": 0.1360906958580017, + "learning_rate": 1e-06, + "loss": -0.0073, + "num_tokens": 190181929.0, + "reward": 0.6462053656578064, + "reward_std": 0.20564965903759003, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 320 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03409090909090906, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2206.0, + "completions/mean_length": 759.9290161132812, + "completions/mean_terminated_length": 642.185302734375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 2.9982507288629736, + "grad_norm": 0.1385015845298767, + "learning_rate": 1e-06, + "loss": -0.0007, + "num_tokens": 190826068.0, + "reward": 0.621651828289032, + "reward_std": 0.1921198070049286, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 321 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4041.0, + "completions/mean_length": 949.23779296875, + "completions/mean_terminated_length": 731.4427490234375, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 3.00932944606414, + "grad_norm": 0.11967211216688156, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 191529305.0, + "reward": 0.5167410969734192, + "reward_std": 0.18475790321826935, + "rewards/verify_math_reward/mean": 0.5167410969734192, + "rewards/verify_math_reward/std": 0.4999987483024597, + "step": 322 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3436.0, + "completions/mean_length": 883.6585083007812, + "completions/mean_terminated_length": 697.8204956054688, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 3.01865889212828, + "grad_norm": 0.1216469556093216, + "learning_rate": 1e-06, + "loss": -0.0085, + "num_tokens": 192202855.0, + "reward": 0.5993303656578064, + "reward_std": 0.18122071027755737, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3403.0, + "completions/mean_length": 802.2857666015625, + "completions/mean_terminated_length": 628.1174926757812, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.02798833819242, + "grad_norm": 0.14834488928318024, + "learning_rate": 1e-06, + "loss": -0.0246, + "num_tokens": 192827047.0, + "reward": 0.5792410969734192, + "reward_std": 0.18077422678470612, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3540.0, + "completions/mean_length": 762.044677734375, + "completions/mean_terminated_length": 614.386962890625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 3.03731778425656, + "grad_norm": 0.12265503406524658, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 193444783.0, + "reward": 0.6183035969734192, + "reward_std": 0.15834447741508484, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 325 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3284.0, + "completions/mean_length": 724.9910888671875, + "completions/mean_terminated_length": 592.02783203125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 3.0466472303206995, + "grad_norm": 0.1282912790775299, + "learning_rate": 1e-06, + "loss": -0.0114, + "num_tokens": 194035375.0, + "reward": 0.6194196939468384, + "reward_std": 0.1610461324453354, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 326 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3300.0, + "completions/mean_length": 807.2578735351562, + "completions/mean_terminated_length": 661.6026000976562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 3.0559766763848395, + "grad_norm": 0.13438823819160461, + "learning_rate": 1e-06, + "loss": -0.017, + "num_tokens": 194695182.0, + "reward": 0.5970982313156128, + "reward_std": 0.19633103907108307, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 327 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4056.0, + "completions/mean_length": 799.802490234375, + "completions/mean_terminated_length": 621.4199829101562, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 3.0653061224489795, + "grad_norm": 0.12980444729328156, + "learning_rate": 1e-06, + "loss": -0.0112, + "num_tokens": 195316141.0, + "reward": 0.6071428656578064, + "reward_std": 0.16435371339321136, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2957.0, + "completions/mean_length": 755.5502319335938, + "completions/mean_terminated_length": 591.2658081054688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.0746355685131195, + "grad_norm": 0.14052127301692963, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 195911890.0, + "reward": 0.6439732313156128, + "reward_std": 0.1638326495885849, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 329 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2740.0, + "completions/mean_length": 805.3292846679688, + "completions/mean_terminated_length": 667.5802001953125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 3.0839650145772595, + "grad_norm": 0.12807653844356537, + "learning_rate": 1e-06, + "loss": -0.0127, + "num_tokens": 196581745.0, + "reward": 0.543526828289032, + "reward_std": 0.19851389527320862, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 330 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2539.0, + "completions/mean_length": 757.7500610351562, + "completions/mean_terminated_length": 601.7570190429688, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 3.0932944606413995, + "grad_norm": 0.1517622321844101, + "learning_rate": 1e-06, + "loss": -0.0144, + "num_tokens": 197183625.0, + "reward": 0.6350446939468384, + "reward_std": 0.22100962698459625, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3366.0, + "completions/mean_length": 733.9185791015625, + "completions/mean_terminated_length": 621.4613647460938, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 3.1026239067055394, + "grad_norm": 0.1199645847082138, + "learning_rate": 1e-06, + "loss": -0.0018, + "num_tokens": 197809648.0, + "reward": 0.6584821939468384, + "reward_std": 0.15815789997577667, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 332 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3904.0, + "completions/mean_length": 678.7176513671875, + "completions/mean_terminated_length": 568.4827270507812, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 3.1119533527696794, + "grad_norm": 0.13154439628124237, + "learning_rate": 1e-06, + "loss": 0.0142, + "num_tokens": 198398419.0, + "reward": 0.6506696939468384, + "reward_std": 0.17104512453079224, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 333 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 814.3672485351562, + "completions/mean_terminated_length": 648.9390258789062, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.1212827988338194, + "grad_norm": 0.1308983713388443, + "learning_rate": 1e-06, + "loss": 0.0096, + "num_tokens": 199052796.0, + "reward": 0.5881696939468384, + "reward_std": 0.1871657818555832, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2258.0, + "completions/mean_length": 734.5279541015625, + "completions/mean_terminated_length": 569.2095947265625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.130612244897959, + "grad_norm": 0.13285918533802032, + "learning_rate": 1e-06, + "loss": -0.0104, + "num_tokens": 199620445.0, + "reward": 0.6930803656578064, + "reward_std": 0.1595052033662796, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3813.0, + "completions/mean_length": 834.1674194335938, + "completions/mean_terminated_length": 633.201416015625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.139941690962099, + "grad_norm": 0.14414021372795105, + "learning_rate": 1e-06, + "loss": -0.0133, + "num_tokens": 200245203.0, + "reward": 0.5558035969734192, + "reward_std": 0.21414414048194885, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715369939804077, + "step": 336 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2666.0, + "completions/mean_length": 717.2098388671875, + "completions/mean_terminated_length": 592.0694580078125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 3.149271137026239, + "grad_norm": 0.1350342333316803, + "learning_rate": 1e-06, + "loss": 0.013, + "num_tokens": 200844055.0, + "reward": 0.6886160969734192, + "reward_std": 0.1555229276418686, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 337 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2955.0, + "completions/mean_length": 1013.9542846679688, + "completions/mean_terminated_length": 703.4778442382812, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "epoch": 3.158600583090379, + "grad_norm": 0.13759945333003998, + "learning_rate": 1e-06, + "loss": -0.0421, + "num_tokens": 201531270.0, + "reward": 0.4843750298023224, + "reward_std": 0.2063218653202057, + "rewards/verify_math_reward/mean": 0.484375, + "rewards/verify_math_reward/std": 0.5000349283218384, + "step": 338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2865.0, + "completions/mean_length": 798.8047485351562, + "completions/mean_terminated_length": 652.7750854492188, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 3.167930029154519, + "grad_norm": 0.12224580347537994, + "learning_rate": 1e-06, + "loss": 0.0057, + "num_tokens": 202183879.0, + "reward": 0.6194196939468384, + "reward_std": 0.15172609686851501, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 339 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 839.232177734375, + "completions/mean_terminated_length": 646.7517700195312, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 3.177259475218659, + "grad_norm": 0.12868469953536987, + "learning_rate": 1e-06, + "loss": 0.0008, + "num_tokens": 202815927.0, + "reward": 0.6328125, + "reward_std": 0.16232240200042725, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 340 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 784.029052734375, + "completions/mean_terminated_length": 633.3092041015625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 3.186588921282799, + "grad_norm": 0.13547104597091675, + "learning_rate": 1e-06, + "loss": -0.0107, + "num_tokens": 203441345.0, + "reward": 0.5948660969734192, + "reward_std": 0.19956375658512115, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 341 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3756.0, + "completions/mean_length": 877.294677734375, + "completions/mean_terminated_length": 703.1058349609375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 3.195918367346939, + "grad_norm": 0.12017311155796051, + "learning_rate": 1e-06, + "loss": -0.02, + "num_tokens": 204132457.0, + "reward": 0.5703125, + "reward_std": 0.18701379001140594, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3038.0, + "completions/mean_length": 810.6585083007812, + "completions/mean_terminated_length": 640.9929809570312, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.205247813411079, + "grad_norm": 0.14066319167613983, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 204764687.0, + "reward": 0.6283482313156128, + "reward_std": 0.19855482876300812, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 343 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3736.0, + "completions/mean_length": 821.4386596679688, + "completions/mean_terminated_length": 640.161376953125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.2145772594752184, + "grad_norm": 0.13110028207302094, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 205396744.0, + "reward": 0.5725446939468384, + "reward_std": 0.17461301386356354, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 344 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3519.0, + "completions/mean_length": 894.1819458007812, + "completions/mean_terminated_length": 672.576416015625, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.2239067055393584, + "grad_norm": 0.1255418211221695, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 206055147.0, + "reward": 0.5714285969734192, + "reward_std": 0.16761739552021027, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 345 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3613.0, + "completions/mean_length": 761.8192138671875, + "completions/mean_terminated_length": 618.2048950195312, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.2332361516034984, + "grad_norm": 0.13259796798229218, + "learning_rate": 1e-06, + "loss": -0.0126, + "num_tokens": 206670137.0, + "reward": 0.6729910969734192, + "reward_std": 0.15240898728370667, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 346 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3766.0, + "completions/mean_length": 769.3906860351562, + "completions/mean_terminated_length": 638.1786499023438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.2425655976676384, + "grad_norm": 0.1300286501646042, + "learning_rate": 1e-06, + "loss": -0.0158, + "num_tokens": 207311479.0, + "reward": 0.606026828289032, + "reward_std": 0.1719466745853424, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 347 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3720.0, + "completions/mean_length": 715.3236694335938, + "completions/mean_terminated_length": 577.8977661132812, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 3.2518950437317784, + "grad_norm": 0.17092815041542053, + "learning_rate": 1e-06, + "loss": -0.0124, + "num_tokens": 207900249.0, + "reward": 0.6283482313156128, + "reward_std": 0.17604516446590424, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 348 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3926.0, + "completions/mean_length": 981.646240234375, + "completions/mean_terminated_length": 746.1068115234375, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 3.2612244897959184, + "grad_norm": 0.12846660614013672, + "learning_rate": 1e-06, + "loss": 0.0013, + "num_tokens": 208616604.0, + "reward": 0.5189732313156128, + "reward_std": 0.2100858986377716, + "rewards/verify_math_reward/mean": 0.5189732313156128, + "rewards/verify_math_reward/std": 0.49991893768310547, + "step": 349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3805.0, + "completions/mean_length": 798.0156860351562, + "completions/mean_terminated_length": 643.9041748046875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 3.2705539358600584, + "grad_norm": 0.12801715731620789, + "learning_rate": 1e-06, + "loss": -0.0081, + "num_tokens": 209260250.0, + "reward": 0.5647321939468384, + "reward_std": 0.17615465819835663, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 350 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 845.1674194335938, + "completions/mean_terminated_length": 648.9633178710938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 3.2798833819241984, + "grad_norm": 0.12162301689386368, + "learning_rate": 1e-06, + "loss": 0.0048, + "num_tokens": 209895280.0, + "reward": 0.5647321939468384, + "reward_std": 0.14481490850448608, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606895446777344, + "step": 351 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4081.0, + "completions/mean_length": 923.4342041015625, + "completions/mean_terminated_length": 715.95361328125, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 3.2892128279883384, + "grad_norm": 0.12108243256807327, + "learning_rate": 1e-06, + "loss": -0.0364, + "num_tokens": 210592045.0, + "reward": 0.5892857313156128, + "reward_std": 0.18948623538017273, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 352 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3635.0, + "completions/mean_length": 807.1127319335938, + "completions/mean_terminated_length": 625.0424194335938, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 3.298542274052478, + "grad_norm": 0.13322065770626068, + "learning_rate": 1e-06, + "loss": -0.0205, + "num_tokens": 211216226.0, + "reward": 0.6551339626312256, + "reward_std": 0.17345327138900757, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 353 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3189.0, + "completions/mean_length": 850.6451416015625, + "completions/mean_terminated_length": 666.9457397460938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 3.307871720116618, + "grad_norm": 0.1369079053401947, + "learning_rate": 1e-06, + "loss": -0.0038, + "num_tokens": 211872716.0, + "reward": 0.5714285969734192, + "reward_std": 0.20595093071460724, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514806270599365, + "step": 354 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3232.0, + "completions/mean_length": 659.6707763671875, + "completions/mean_terminated_length": 565.0928955078125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.317201166180758, + "grad_norm": 0.12828774750232697, + "learning_rate": 1e-06, + "loss": -0.0072, + "num_tokens": 212442845.0, + "reward": 0.699776828289032, + "reward_std": 0.14011836051940918, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 355 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3859.0, + "completions/mean_length": 855.2533569335938, + "completions/mean_terminated_length": 639.20361328125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 3.326530612244898, + "grad_norm": 0.1488858014345169, + "learning_rate": 1e-06, + "loss": -0.0146, + "num_tokens": 213070536.0, + "reward": 0.6350446939468384, + "reward_std": 0.19805879890918732, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 356 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3558.0, + "completions/mean_length": 906.4531860351562, + "completions/mean_terminated_length": 717.9456176757812, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 3.335860058309038, + "grad_norm": 0.12346269190311432, + "learning_rate": 1e-06, + "loss": -0.0187, + "num_tokens": 213775022.0, + "reward": 0.5334821939468384, + "reward_std": 0.19276244938373566, + "rewards/verify_math_reward/mean": 0.5334821343421936, + "rewards/verify_math_reward/std": 0.49915629625320435, + "step": 357 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3589.0, + "completions/mean_length": 749.1517944335938, + "completions/mean_terminated_length": 604.9918823242188, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 3.345189504373178, + "grad_norm": 0.13450050354003906, + "learning_rate": 1e-06, + "loss": 0.0006, + "num_tokens": 214378478.0, + "reward": 0.640625, + "reward_std": 0.18065868318080902, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3058.0, + "completions/mean_length": 925.1730346679688, + "completions/mean_terminated_length": 685.362548828125, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 3.354518950437318, + "grad_norm": 0.13113565742969513, + "learning_rate": 1e-06, + "loss": -0.0241, + "num_tokens": 215044105.0, + "reward": 0.5859375, + "reward_std": 0.1839662492275238, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 359 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2868.0, + "completions/mean_length": 796.208740234375, + "completions/mean_terminated_length": 617.6317138671875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 3.363848396501458, + "grad_norm": 0.13019555807113647, + "learning_rate": 1e-06, + "loss": -0.007, + "num_tokens": 215660004.0, + "reward": 0.6194196939468384, + "reward_std": 0.17130474746227264, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 360 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3120.0, + "completions/mean_length": 774.4866333007812, + "completions/mean_terminated_length": 586.4764404296875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.373177842565598, + "grad_norm": 0.13915102183818817, + "learning_rate": 1e-06, + "loss": -0.0244, + "num_tokens": 216258056.0, + "reward": 0.6417410969734192, + "reward_std": 0.1818615049123764, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 361 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3350.0, + "completions/mean_length": 713.1763916015625, + "completions/mean_terminated_length": 575.6631469726562, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 3.3825072886297374, + "grad_norm": 0.15761631727218628, + "learning_rate": 1e-06, + "loss": -0.0128, + "num_tokens": 216839854.0, + "reward": 0.6473214626312256, + "reward_std": 0.18945163488388062, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 362 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3488.0, + "completions/mean_length": 812.8873291015625, + "completions/mean_terminated_length": 647.384521484375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 3.3918367346938774, + "grad_norm": 0.13500329852104187, + "learning_rate": 1e-06, + "loss": -0.0293, + "num_tokens": 217482169.0, + "reward": 0.5535714626312256, + "reward_std": 0.19974736869335175, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973994791507721, + "step": 363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3229.0, + "completions/mean_length": 847.1160888671875, + "completions/mean_terminated_length": 626.393310546875, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 3.4011661807580174, + "grad_norm": 0.11885211616754532, + "learning_rate": 1e-06, + "loss": -0.0288, + "num_tokens": 218105561.0, + "reward": 0.5948660969734192, + "reward_std": 0.14935779571533203, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 364 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4084.0, + "completions/mean_length": 804.6730346679688, + "completions/mean_terminated_length": 634.6983642578125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 3.4104956268221573, + "grad_norm": 0.1255892515182495, + "learning_rate": 1e-06, + "loss": -0.0167, + "num_tokens": 218731284.0, + "reward": 0.6305803656578064, + "reward_std": 0.17435340583324432, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 365 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3231.0, + "completions/mean_length": 832.2734985351562, + "completions/mean_terminated_length": 667.7479248046875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 3.4198250728862973, + "grad_norm": 0.12889975309371948, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 219402873.0, + "reward": 0.5680803656578064, + "reward_std": 0.16055646538734436, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 366 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3358.0, + "completions/mean_length": 817.2254638671875, + "completions/mean_terminated_length": 647.8990478515625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 3.4291545189504373, + "grad_norm": 0.13108345866203308, + "learning_rate": 1e-06, + "loss": -0.0134, + "num_tokens": 220053091.0, + "reward": 0.5948660969734192, + "reward_std": 0.17051450908184052, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 367 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3030.0, + "completions/mean_length": 692.3795166015625, + "completions/mean_terminated_length": 586.6282958984375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 3.4384839650145773, + "grad_norm": 0.1256914734840393, + "learning_rate": 1e-06, + "loss": -0.0192, + "num_tokens": 220646951.0, + "reward": 0.691964328289032, + "reward_std": 0.15090195834636688, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 368 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3542.0, + "completions/mean_length": 782.6261596679688, + "completions/mean_terminated_length": 615.597900390625, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "epoch": 3.4478134110787173, + "grad_norm": 0.1355881541967392, + "learning_rate": 1e-06, + "loss": -0.0123, + "num_tokens": 221272408.0, + "reward": 0.5792410969734192, + "reward_std": 0.16450665891170502, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4041.0, + "completions/mean_length": 840.4788208007812, + "completions/mean_terminated_length": 660.255615234375, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 3.4571428571428573, + "grad_norm": 0.12165997922420502, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 221918829.0, + "reward": 0.5915178656578064, + "reward_std": 0.1488724797964096, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 370 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2862.0, + "completions/mean_length": 717.2433471679688, + "completions/mean_terminated_length": 592.1041870117188, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 3.466472303206997, + "grad_norm": 0.13266989588737488, + "learning_rate": 1e-06, + "loss": -0.0088, + "num_tokens": 222515615.0, + "reward": 0.6696428656578064, + "reward_std": 0.17378000915050507, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 371 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3329.0, + "completions/mean_length": 830.0156860351562, + "completions/mean_terminated_length": 624.680908203125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 3.4758017492711373, + "grad_norm": 0.14682939648628235, + "learning_rate": 1e-06, + "loss": -0.0279, + "num_tokens": 223138213.0, + "reward": 0.6116071939468384, + "reward_std": 0.20711460709571838, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 372 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3961.0, + "completions/mean_length": 798.1428833007812, + "completions/mean_terminated_length": 635.953125, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 3.485131195335277, + "grad_norm": 0.12675997614860535, + "learning_rate": 1e-06, + "loss": 0.0026, + "num_tokens": 223770685.0, + "reward": 0.6127232313156128, + "reward_std": 0.15563170611858368, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3948.0, + "completions/mean_length": 845.7422485351562, + "completions/mean_terminated_length": 641.396240234375, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 3.494460641399417, + "grad_norm": 0.1343097984790802, + "learning_rate": 1e-06, + "loss": -0.0044, + "num_tokens": 224408966.0, + "reward": 0.6149553656578064, + "reward_std": 0.16108709573745728, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 374 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 839.0658569335938, + "completions/mean_terminated_length": 626.0677490234375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 3.503790087463557, + "grad_norm": 0.10293253511190414, + "learning_rate": 1e-06, + "loss": -0.0117, + "num_tokens": 225013465.0, + "reward": 0.5613839626312256, + "reward_std": 0.12936869263648987, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 375 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3707.0, + "completions/mean_length": 812.8582763671875, + "completions/mean_terminated_length": 675.4244384765625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 3.513119533527697, + "grad_norm": 0.14238329231739044, + "learning_rate": 1e-06, + "loss": -0.0257, + "num_tokens": 225682746.0, + "reward": 0.645089328289032, + "reward_std": 0.2191762924194336, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 376 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3698.0, + "completions/mean_length": 848.1663208007812, + "completions/mean_terminated_length": 672.401123046875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "epoch": 3.522448979591837, + "grad_norm": 0.13714191317558289, + "learning_rate": 1e-06, + "loss": -0.0203, + "num_tokens": 226350583.0, + "reward": 0.5881696939468384, + "reward_std": 0.1800595223903656, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 377 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4060.0, + "completions/mean_length": 846.161865234375, + "completions/mean_terminated_length": 654.0910034179688, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 3.5317784256559768, + "grad_norm": 0.1421462744474411, + "learning_rate": 1e-06, + "loss": -0.0138, + "num_tokens": 226998800.0, + "reward": 0.6026785969734192, + "reward_std": 0.18501681089401245, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 378 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 789.0189819335938, + "completions/mean_terminated_length": 646.5762329101562, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 3.5411078717201168, + "grad_norm": 0.14839346706867218, + "learning_rate": 1e-06, + "loss": -0.0162, + "num_tokens": 227638849.0, + "reward": 0.6551339626312256, + "reward_std": 0.22469627857208252, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 379 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2788.0, + "completions/mean_length": 779.4944458007812, + "completions/mean_terminated_length": 608.219482421875, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 3.5504373177842563, + "grad_norm": 0.12876732647418976, + "learning_rate": 1e-06, + "loss": -0.0058, + "num_tokens": 228257132.0, + "reward": 0.6104910969734192, + "reward_std": 0.15747573971748352, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 380 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3033.0, + "completions/mean_length": 727.1563110351562, + "completions/mean_terminated_length": 594.2783813476562, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 3.5597667638483967, + "grad_norm": 0.1479600965976715, + "learning_rate": 1e-06, + "loss": -0.0056, + "num_tokens": 228852632.0, + "reward": 0.6696428656578064, + "reward_std": 0.20862093567848206, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3029.0, + "completions/mean_length": 746.0826416015625, + "completions/mean_terminated_length": 609.9070434570312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 3.5690962099125363, + "grad_norm": 0.12609915435314178, + "learning_rate": 1e-06, + "loss": -0.0304, + "num_tokens": 229468594.0, + "reward": 0.6238839626312256, + "reward_std": 0.1707048863172531, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 382 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2837.0, + "completions/mean_length": 788.0647583007812, + "completions/mean_terminated_length": 629.4385986328125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 3.5784256559766763, + "grad_norm": 0.14111444354057312, + "learning_rate": 1e-06, + "loss": -0.0161, + "num_tokens": 230102772.0, + "reward": 0.6049107313156128, + "reward_std": 0.2026451677083969, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 383 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2579.0, + "completions/mean_length": 793.0725708007812, + "completions/mean_terminated_length": 601.9940795898438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.5877551020408163, + "grad_norm": 0.14103786647319794, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 230706181.0, + "reward": 0.6584821939468384, + "reward_std": 0.16904956102371216, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 384 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2787.0, + "completions/mean_length": 818.7678833007812, + "completions/mean_terminated_length": 620.9703979492188, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 3.5970845481049563, + "grad_norm": 0.1310214400291443, + "learning_rate": 1e-06, + "loss": -0.0115, + "num_tokens": 231324173.0, + "reward": 0.6127232313156128, + "reward_std": 0.16522274911403656, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 385 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3210.0, + "completions/mean_length": 827.3973388671875, + "completions/mean_terminated_length": 650.5082397460938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 3.6064139941690962, + "grad_norm": 0.13954100012779236, + "learning_rate": 1e-06, + "loss": -0.0158, + "num_tokens": 231960393.0, + "reward": 0.598214328289032, + "reward_std": 0.17559054493904114, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 386 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2889.0, + "completions/mean_length": 873.4877319335938, + "completions/mean_terminated_length": 654.556640625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.6157434402332362, + "grad_norm": 0.11855709552764893, + "learning_rate": 1e-06, + "loss": -0.0227, + "num_tokens": 232599950.0, + "reward": 0.6439732313156128, + "reward_std": 0.17336632311344147, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 387 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4042.0, + "completions/mean_length": 856.0826416015625, + "completions/mean_terminated_length": 656.4668579101562, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 3.6250728862973762, + "grad_norm": 0.13760653138160706, + "learning_rate": 1e-06, + "loss": -0.0333, + "num_tokens": 233243344.0, + "reward": 0.5703125, + "reward_std": 0.22495870292186737, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2679.0, + "completions/mean_length": 674.271240234375, + "completions/mean_terminated_length": 572.0126342773438, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 3.6344023323615158, + "grad_norm": 0.14141114056110382, + "learning_rate": 1e-06, + "loss": -0.0108, + "num_tokens": 233830315.0, + "reward": 0.676339328289032, + "reward_std": 0.1742018610239029, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 389 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2845.0, + "completions/mean_length": 886.7589721679688, + "completions/mean_terminated_length": 664.6396484375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 3.643731778425656, + "grad_norm": 0.12106688320636749, + "learning_rate": 1e-06, + "loss": -0.0131, + "num_tokens": 234467515.0, + "reward": 0.5602678656578064, + "reward_std": 0.17682865262031555, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317415237427, + "step": 390 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4032.0, + "completions/mean_length": 880.0457763671875, + "completions/mean_terminated_length": 624.3192749023438, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.6530612244897958, + "grad_norm": 0.14808392524719238, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 235072244.0, + "reward": 0.6540178656578064, + "reward_std": 0.18426865339279175, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 391 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3671.0, + "completions/mean_length": 849.1986694335938, + "completions/mean_terminated_length": 636.8632202148438, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 3.6623906705539357, + "grad_norm": 0.1266641467809677, + "learning_rate": 1e-06, + "loss": -0.0286, + "num_tokens": 235706830.0, + "reward": 0.5892857313156128, + "reward_std": 0.15326662361621857, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 392 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3231.0, + "completions/mean_length": 690.5324096679688, + "completions/mean_terminated_length": 604.8111572265625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 3.6717201166180757, + "grad_norm": 0.1398337334394455, + "learning_rate": 1e-06, + "loss": -0.0025, + "num_tokens": 236330619.0, + "reward": 0.6830357313156128, + "reward_std": 0.1810370534658432, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 393 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3861.0, + "completions/mean_length": 816.9989013671875, + "completions/mean_terminated_length": 679.7383422851562, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.6810495626822157, + "grad_norm": 0.13178351521492004, + "learning_rate": 1e-06, + "loss": -0.0131, + "num_tokens": 237007722.0, + "reward": 0.6283482313156128, + "reward_std": 0.19787196815013885, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3567.0, + "completions/mean_length": 841.6819458007812, + "completions/mean_terminated_length": 649.3463134765625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 3.6903790087463557, + "grad_norm": 0.12276212126016617, + "learning_rate": 1e-06, + "loss": -0.0124, + "num_tokens": 237641293.0, + "reward": 0.5770089626312256, + "reward_std": 0.17292235791683197, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 395 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3292.0, + "completions/mean_length": 821.3092041015625, + "completions/mean_terminated_length": 623.6651000976562, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 3.6997084548104957, + "grad_norm": 0.12476027011871338, + "learning_rate": 1e-06, + "loss": -0.0234, + "num_tokens": 238257234.0, + "reward": 0.621651828289032, + "reward_std": 0.16777144372463226, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 396 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3490.0, + "completions/mean_length": 812.5424194335938, + "completions/mean_terminated_length": 589.4708251953125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.7090379008746357, + "grad_norm": 0.1332354098558426, + "learning_rate": 1e-06, + "loss": -0.0395, + "num_tokens": 238838160.0, + "reward": 0.6618303656578064, + "reward_std": 0.1938129961490631, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 397 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2642.0, + "completions/mean_length": 871.2969360351562, + "completions/mean_terminated_length": 684.7437744140625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 3.7183673469387752, + "grad_norm": 0.1296387016773224, + "learning_rate": 1e-06, + "loss": -0.0287, + "num_tokens": 239508754.0, + "reward": 0.551339328289032, + "reward_std": 0.2006898671388626, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4076.0, + "completions/mean_length": 721.4766235351562, + "completions/mean_terminated_length": 604.5762329101562, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 3.7276967930029157, + "grad_norm": 0.12461227923631668, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 240112509.0, + "reward": 0.699776828289032, + "reward_std": 0.1475597620010376, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 399 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3468.0, + "completions/mean_length": 819.8426513671875, + "completions/mean_terminated_length": 662.7403564453125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 3.7370262390670552, + "grad_norm": 0.13590675592422485, + "learning_rate": 1e-06, + "loss": -0.0113, + "num_tokens": 240759136.0, + "reward": 0.6640625, + "reward_std": 0.20594529807567596, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 400 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4030.0, + "completions/mean_length": 880.1172485351562, + "completions/mean_terminated_length": 641.0467529296875, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 3.746355685131195, + "grad_norm": 0.13403309881687164, + "learning_rate": 1e-06, + "loss": -0.0056, + "num_tokens": 241375433.0, + "reward": 0.6506696939468384, + "reward_std": 0.18081024289131165, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 401 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3179.0, + "completions/mean_length": 862.8594360351562, + "completions/mean_terminated_length": 671.775390625, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 3.755685131195335, + "grad_norm": 0.12232960760593414, + "learning_rate": 1e-06, + "loss": -0.0196, + "num_tokens": 242028955.0, + "reward": 0.6439732313156128, + "reward_std": 0.16383123397827148, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 402 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3985.0, + "completions/mean_length": 858.3392944335938, + "completions/mean_terminated_length": 666.9881591796875, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.765014577259475, + "grad_norm": 0.1501288115978241, + "learning_rate": 1e-06, + "loss": -0.0312, + "num_tokens": 242673971.0, + "reward": 0.609375, + "reward_std": 0.23015236854553223, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 403 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 917.1585083007812, + "completions/mean_terminated_length": 651.9347534179688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.774344023323615, + "grad_norm": 0.12584435939788818, + "learning_rate": 1e-06, + "loss": -0.0341, + "num_tokens": 243297705.0, + "reward": 0.6595982313156128, + "reward_std": 0.15510854125022888, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3786.0, + "completions/mean_length": 822.888427734375, + "completions/mean_terminated_length": 649.8096313476562, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 3.783673469387755, + "grad_norm": 0.12362376600503922, + "learning_rate": 1e-06, + "loss": -0.0382, + "num_tokens": 243952005.0, + "reward": 0.6238839626312256, + "reward_std": 0.18483206629753113, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 405 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 862.8694458007812, + "completions/mean_terminated_length": 643.2169189453125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 3.793002915451895, + "grad_norm": 0.12691886723041534, + "learning_rate": 1e-06, + "loss": -0.0182, + "num_tokens": 244571528.0, + "reward": 0.5848214626312256, + "reward_std": 0.16499200463294983, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 406 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 843.4620971679688, + "completions/mean_terminated_length": 667.4423217773438, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 3.8023323615160347, + "grad_norm": 0.13258963823318481, + "learning_rate": 1e-06, + "loss": -0.0145, + "num_tokens": 245220686.0, + "reward": 0.6551339626312256, + "reward_std": 0.19453050196170807, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 407 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 783.5736694335938, + "completions/mean_terminated_length": 608.4159545898438, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.811661807580175, + "grad_norm": 0.14538998901844025, + "learning_rate": 1e-06, + "loss": -0.0016, + "num_tokens": 245817520.0, + "reward": 0.6651785969734192, + "reward_std": 0.1835525780916214, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 408 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3520.0, + "completions/mean_length": 938.19873046875, + "completions/mean_terminated_length": 711.5621948242188, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 3.8209912536443147, + "grad_norm": 0.11863500624895096, + "learning_rate": 1e-06, + "loss": -0.0252, + "num_tokens": 246498850.0, + "reward": 0.546875, + "reward_std": 0.18553857505321503, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 409 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3933.0, + "completions/mean_length": 931.2545166015625, + "completions/mean_terminated_length": 679.6000366210938, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 3.8303206997084547, + "grad_norm": 0.1227620542049408, + "learning_rate": 1e-06, + "loss": -0.0128, + "num_tokens": 247154438.0, + "reward": 0.559151828289032, + "reward_std": 0.12088630348443985, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 410 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2384.0, + "completions/mean_length": 831.5201416015625, + "completions/mean_terminated_length": 609.73779296875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 3.8396501457725947, + "grad_norm": 0.13331255316734314, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 247750824.0, + "reward": 0.637276828289032, + "reward_std": 0.16499128937721252, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 411 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3892.0, + "completions/mean_length": 783.7064819335938, + "completions/mean_terminated_length": 645.0523071289062, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 3.8489795918367347, + "grad_norm": 0.13118019700050354, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 248390145.0, + "reward": 0.6171875, + "reward_std": 0.1855827420949936, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 412 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2644.0, + "completions/mean_length": 742.5011596679688, + "completions/mean_terminated_length": 573.4501342773438, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 3.8583090379008746, + "grad_norm": 0.1382608711719513, + "learning_rate": 1e-06, + "loss": -0.0076, + "num_tokens": 248958578.0, + "reward": 0.7120535969734192, + "reward_std": 0.1714577078819275, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2904.0, + "completions/mean_length": 803.9855346679688, + "completions/mean_terminated_length": 666.1802368164062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 3.8676384839650146, + "grad_norm": 0.12190508842468262, + "learning_rate": 1e-06, + "loss": -0.031, + "num_tokens": 249627165.0, + "reward": 0.6595982313156128, + "reward_std": 0.1710362583398819, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 414 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3626.0, + "completions/mean_length": 772.9542846679688, + "completions/mean_terminated_length": 597.2349853515625, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 3.8769679300291546, + "grad_norm": 0.16220395267009735, + "learning_rate": 1e-06, + "loss": -0.0022, + "num_tokens": 250219372.0, + "reward": 0.6741071939468384, + "reward_std": 0.19756954908370972, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 415 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3990.0, + "completions/mean_length": 957.0167846679688, + "completions/mean_terminated_length": 691.001220703125, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 3.8862973760932946, + "grad_norm": 0.1392662674188614, + "learning_rate": 1e-06, + "loss": -0.0076, + "num_tokens": 250886243.0, + "reward": 0.578125, + "reward_std": 0.18772776424884796, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 416 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3744.0, + "completions/mean_length": 912.2489013671875, + "completions/mean_terminated_length": 646.615478515625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.8956268221574346, + "grad_norm": 0.13970129191875458, + "learning_rate": 1e-06, + "loss": -0.0291, + "num_tokens": 251514202.0, + "reward": 0.606026828289032, + "reward_std": 0.18039445579051971, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2449.0, + "completions/mean_length": 761.3069458007812, + "completions/mean_terminated_length": 609.5531005859375, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 3.904956268221574, + "grad_norm": 0.15438564121723175, + "learning_rate": 1e-06, + "loss": -0.0018, + "num_tokens": 252128101.0, + "reward": 0.6517857313156128, + "reward_std": 0.17033086717128754, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 418 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3985.0, + "completions/mean_length": 900.3839721679688, + "completions/mean_terminated_length": 679.2076416015625, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 3.914285714285714, + "grad_norm": 0.18476028740406036, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 252782485.0, + "reward": 0.6171875, + "reward_std": 0.21704013645648956, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 419 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4027.0, + "completions/mean_length": 769.7835083007812, + "completions/mean_terminated_length": 606.1990356445312, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 3.923615160349854, + "grad_norm": 0.143397718667984, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 253388091.0, + "reward": 0.668526828289032, + "reward_std": 0.1959933340549469, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 420 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3937.0, + "completions/mean_length": 874.9285888671875, + "completions/mean_terminated_length": 664.27587890625, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 3.932944606413994, + "grad_norm": 0.13676884770393372, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 254040011.0, + "reward": 0.6071428656578064, + "reward_std": 0.18193607032299042, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 421 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3382.0, + "completions/mean_length": 883.9676513671875, + "completions/mean_terminated_length": 624.370361328125, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 3.942274052478134, + "grad_norm": 0.13674820959568024, + "learning_rate": 1e-06, + "loss": -0.0259, + "num_tokens": 254643622.0, + "reward": 0.6238839626312256, + "reward_std": 0.19047150015830994, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 422 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3193.0, + "completions/mean_length": 794.0301513671875, + "completions/mean_terminated_length": 659.8037109375, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 3.951603498542274, + "grad_norm": 0.11584319174289703, + "learning_rate": 1e-06, + "loss": -0.0067, + "num_tokens": 255302873.0, + "reward": 0.582589328289032, + "reward_std": 0.1438377946615219, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2948.0, + "completions/mean_length": 754.4085083007812, + "completions/mean_terminated_length": 581.8380126953125, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 3.960932944606414, + "grad_norm": 0.1446135938167572, + "learning_rate": 1e-06, + "loss": -0.0315, + "num_tokens": 255881111.0, + "reward": 0.6908482313156128, + "reward_std": 0.1897144913673401, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 424 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2790.0, + "completions/mean_length": 831.1183471679688, + "completions/mean_terminated_length": 654.4305419921875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 3.970262390670554, + "grad_norm": 0.12886013090610504, + "learning_rate": 1e-06, + "loss": -0.0167, + "num_tokens": 256525969.0, + "reward": 0.6484375, + "reward_std": 0.156466543674469, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 425 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3336.0, + "completions/mean_length": 778.1417846679688, + "completions/mean_terminated_length": 586.1995239257812, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.979591836734694, + "grad_norm": 0.12887297570705414, + "learning_rate": 1e-06, + "loss": 0.0028, + "num_tokens": 257112760.0, + "reward": 0.6618303656578064, + "reward_std": 0.13943690061569214, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 811.6295166015625, + "completions/mean_terminated_length": 625.7216796875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 3.9889212827988336, + "grad_norm": 0.13565848767757416, + "learning_rate": 1e-06, + "loss": -0.0305, + "num_tokens": 257733740.0, + "reward": 0.676339328289032, + "reward_std": 0.17942197620868683, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 427 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.051136363636363646, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3522.0, + "completions/mean_length": 743.42333984375, + "completions/mean_terminated_length": 562.7455444335938, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.9982507288629736, + "grad_norm": 0.13699331879615784, + "learning_rate": 1e-06, + "loss": -0.0245, + "num_tokens": 258324824.0, + "reward": 0.6863839626312256, + "reward_std": 0.17325864732265472, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 428 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2986.0, + "completions/mean_length": 823.0145263671875, + "completions/mean_terminated_length": 600.6543579101562, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 4.0093294460641395, + "grad_norm": 0.15516826510429382, + "learning_rate": 1e-06, + "loss": -0.0199, + "num_tokens": 258922869.0, + "reward": 0.6640625, + "reward_std": 0.16334158182144165, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 429 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3512.0, + "completions/mean_length": 763.3850708007812, + "completions/mean_terminated_length": 611.7257690429688, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 4.01865889212828, + "grad_norm": 0.13731922209262848, + "learning_rate": 1e-06, + "loss": -0.0185, + "num_tokens": 259531854.0, + "reward": 0.6383928656578064, + "reward_std": 0.18088509142398834, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 430 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4009.0, + "completions/mean_length": 845.7511596679688, + "completions/mean_terminated_length": 633.1902465820312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 4.0279883381924195, + "grad_norm": 0.1281159520149231, + "learning_rate": 1e-06, + "loss": -0.0212, + "num_tokens": 260142759.0, + "reward": 0.6283482313156128, + "reward_std": 0.1677047610282898, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 431 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3337.0, + "completions/mean_length": 865.1027221679688, + "completions/mean_terminated_length": 653.807373046875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 4.03731778425656, + "grad_norm": 0.14043253660202026, + "learning_rate": 1e-06, + "loss": -0.051, + "num_tokens": 260778595.0, + "reward": 0.6584821939468384, + "reward_std": 0.2066260725259781, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 432 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3463.0, + "completions/mean_length": 844.7835083007812, + "completions/mean_terminated_length": 660.7523803710938, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.0466472303206995, + "grad_norm": 0.12678323686122894, + "learning_rate": 1e-06, + "loss": -0.011, + "num_tokens": 261423969.0, + "reward": 0.6127232313156128, + "reward_std": 0.18411852419376373, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 433 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3335.0, + "completions/mean_length": 913.69091796875, + "completions/mean_terminated_length": 652.341796875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.05597667638484, + "grad_norm": 0.15005941689014435, + "learning_rate": 1e-06, + "loss": -0.0302, + "num_tokens": 262063980.0, + "reward": 0.5837053656578064, + "reward_std": 0.212483748793602, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 434 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4043.0, + "completions/mean_length": 752.2299194335938, + "completions/mean_terminated_length": 567.121337890625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 4.0653061224489795, + "grad_norm": 0.16207073628902435, + "learning_rate": 1e-06, + "loss": -0.0346, + "num_tokens": 262629466.0, + "reward": 0.6662946939468384, + "reward_std": 0.18693682551383972, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 435 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2566.0, + "completions/mean_length": 846.0279541015625, + "completions/mean_terminated_length": 608.6047973632812, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.07463556851312, + "grad_norm": 0.12854279577732086, + "learning_rate": 1e-06, + "loss": -0.0021, + "num_tokens": 263227955.0, + "reward": 0.652901828289032, + "reward_std": 0.1669120043516159, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 436 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2942.0, + "completions/mean_length": 897.193115234375, + "completions/mean_terminated_length": 638.6646728515625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.0839650145772595, + "grad_norm": 0.1507236361503601, + "learning_rate": 1e-06, + "loss": -0.0159, + "num_tokens": 263847064.0, + "reward": 0.5524553656578064, + "reward_std": 0.19117872416973114, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 437 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3083.0, + "completions/mean_length": 884.9732666015625, + "completions/mean_terminated_length": 674.9774169921875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 4.093294460641399, + "grad_norm": 0.13922719657421112, + "learning_rate": 1e-06, + "loss": -0.0216, + "num_tokens": 264495928.0, + "reward": 0.5859375, + "reward_std": 0.1878383606672287, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 438 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3660.0, + "completions/mean_length": 901.583740234375, + "completions/mean_terminated_length": 680.490478515625, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 4.1026239067055394, + "grad_norm": 0.13839367032051086, + "learning_rate": 1e-06, + "loss": -0.0171, + "num_tokens": 265156779.0, + "reward": 0.5959821939468384, + "reward_std": 0.19400165975093842, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 439 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3390.0, + "completions/mean_length": 879.5256958007812, + "completions/mean_terminated_length": 713.4166870117188, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.111953352769679, + "grad_norm": 0.12010248750448227, + "learning_rate": 1e-06, + "loss": -0.009, + "num_tokens": 265846698.0, + "reward": 0.5703125, + "reward_std": 0.18603530526161194, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 440 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3972.0, + "completions/mean_length": 844.404052734375, + "completions/mean_terminated_length": 664.3981323242188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 4.121282798833819, + "grad_norm": 0.1367029845714569, + "learning_rate": 1e-06, + "loss": -0.0024, + "num_tokens": 266494492.0, + "reward": 0.640625, + "reward_std": 0.17949752509593964, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 441 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 829.685302734375, + "completions/mean_terminated_length": 665.029296875, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 4.130612244897959, + "grad_norm": 0.12880828976631165, + "learning_rate": 1e-06, + "loss": -0.0171, + "num_tokens": 267152058.0, + "reward": 0.6015625, + "reward_std": 0.15138980746269226, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 442 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3688.0, + "completions/mean_length": 837.7767944335938, + "completions/mean_terminated_length": 649.2845458984375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.139941690962099, + "grad_norm": 0.11601654440164566, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 267800362.0, + "reward": 0.606026828289032, + "reward_std": 0.14879760146141052, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 443 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2430.0, + "completions/mean_length": 815.0022583007812, + "completions/mean_terminated_length": 600.430419921875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 4.149271137026239, + "grad_norm": 0.1244354099035263, + "learning_rate": 1e-06, + "loss": -0.0191, + "num_tokens": 268402428.0, + "reward": 0.6283482313156128, + "reward_std": 0.1474064141511917, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3838.0, + "completions/mean_length": 830.6272583007812, + "completions/mean_terminated_length": 629.4431762695312, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 4.158600583090379, + "grad_norm": 0.14222611486911774, + "learning_rate": 1e-06, + "loss": -0.0375, + "num_tokens": 269024686.0, + "reward": 0.6383928656578064, + "reward_std": 0.20801502466201782, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3478.0, + "completions/mean_length": 834.474365234375, + "completions/mean_terminated_length": 621.1759643554688, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 4.167930029154519, + "grad_norm": 0.12419307231903076, + "learning_rate": 1e-06, + "loss": 0.0026, + "num_tokens": 269633119.0, + "reward": 0.6495535969734192, + "reward_std": 0.17171591520309448, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2567.0, + "completions/mean_length": 831.0926513671875, + "completions/mean_terminated_length": 629.937255859375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 4.1772594752186585, + "grad_norm": 0.1301281750202179, + "learning_rate": 1e-06, + "loss": -0.0228, + "num_tokens": 270255194.0, + "reward": 0.6316964626312256, + "reward_std": 0.16799968481063843, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 447 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2505.0, + "completions/mean_length": 800.5167846679688, + "completions/mean_terminated_length": 634.390380859375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.186588921282799, + "grad_norm": 0.1374504715204239, + "learning_rate": 1e-06, + "loss": 0.0017, + "num_tokens": 270892433.0, + "reward": 0.6283482313156128, + "reward_std": 0.19219790399074554, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 448 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3817.0, + "completions/mean_length": 844.9631958007812, + "completions/mean_terminated_length": 652.821533203125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 4.1959183673469385, + "grad_norm": 0.1437891274690628, + "learning_rate": 1e-06, + "loss": -0.021, + "num_tokens": 271533840.0, + "reward": 0.6350446939468384, + "reward_std": 0.1975356638431549, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 449 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3789.0, + "completions/mean_length": 829.6373291015625, + "completions/mean_terminated_length": 628.3922119140625, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 4.205247813411079, + "grad_norm": 0.1272605061531067, + "learning_rate": 1e-06, + "loss": -0.0113, + "num_tokens": 272152539.0, + "reward": 0.613839328289032, + "reward_std": 0.16803036630153656, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 450 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3121.0, + "completions/mean_length": 842.2813110351562, + "completions/mean_terminated_length": 654.049560546875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.214577259475218, + "grad_norm": 0.13536153733730316, + "learning_rate": 1e-06, + "loss": -0.0118, + "num_tokens": 272798039.0, + "reward": 0.6350446939468384, + "reward_std": 0.19001756608486176, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 451 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2135.0, + "completions/mean_length": 781.0123291015625, + "completions/mean_terminated_length": 593.3714599609375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.223906705539359, + "grad_norm": 0.14610785245895386, + "learning_rate": 1e-06, + "loss": -0.0159, + "num_tokens": 273388186.0, + "reward": 0.621651828289032, + "reward_std": 0.15624040365219116, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 452 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3502.0, + "completions/mean_length": 885.3917846679688, + "completions/mean_terminated_length": 687.581787109375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 4.233236151603498, + "grad_norm": 0.1331864595413208, + "learning_rate": 1e-06, + "loss": -0.019, + "num_tokens": 274062385.0, + "reward": 0.5926339626312256, + "reward_std": 0.18370410799980164, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 453 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3417.0, + "completions/mean_length": 886.0748291015625, + "completions/mean_terminated_length": 630.8276977539062, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 4.242565597667639, + "grad_norm": 0.13325031101703644, + "learning_rate": 1e-06, + "loss": -0.0462, + "num_tokens": 274682740.0, + "reward": 0.6238839626312256, + "reward_std": 0.19117873907089233, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 454 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3515.0, + "completions/mean_length": 809.3828735351562, + "completions/mean_terminated_length": 639.6514282226562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.251895043731778, + "grad_norm": 0.13051557540893555, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 275312707.0, + "reward": 0.6674107313156128, + "reward_std": 0.14759115874767303, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 455 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4074.0, + "completions/mean_length": 854.1998291015625, + "completions/mean_terminated_length": 633.958251953125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 4.261224489795918, + "grad_norm": 0.12850071489810944, + "learning_rate": 1e-06, + "loss": -0.0237, + "num_tokens": 275929430.0, + "reward": 0.6238839626312256, + "reward_std": 0.15300628542900085, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 456 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3807.0, + "completions/mean_length": 911.8092041015625, + "completions/mean_terminated_length": 641.9625244140625, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 4.270553935860058, + "grad_norm": 0.13285867869853973, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 276545323.0, + "reward": 0.5970982313156128, + "reward_std": 0.17543786764144897, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.4907552897930145, + "step": 457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3820.0, + "completions/mean_length": 845.5736694335938, + "completions/mean_terminated_length": 653.4680786132812, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 4.279883381924198, + "grad_norm": 0.13263507187366486, + "learning_rate": 1e-06, + "loss": -0.0333, + "num_tokens": 277183829.0, + "reward": 0.606026828289032, + "reward_std": 0.17709895968437195, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 458 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3264.0, + "completions/mean_length": 894.3281860351562, + "completions/mean_terminated_length": 623.0, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 4.289212827988338, + "grad_norm": 0.1293635070323944, + "learning_rate": 1e-06, + "loss": -0.0382, + "num_tokens": 277787203.0, + "reward": 0.629464328289032, + "reward_std": 0.16247648000717163, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 459 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4066.0, + "completions/mean_length": 908.9788208007812, + "completions/mean_terminated_length": 696.5107421875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.298542274052478, + "grad_norm": 0.12947046756744385, + "learning_rate": 1e-06, + "loss": -0.0058, + "num_tokens": 278472496.0, + "reward": 0.5703125, + "reward_std": 0.19418390095233917, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 460 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3326.0, + "completions/mean_length": 873.3292846679688, + "completions/mean_terminated_length": 658.4845581054688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 4.307871720116618, + "grad_norm": 0.15107649564743042, + "learning_rate": 1e-06, + "loss": -0.0215, + "num_tokens": 279116975.0, + "reward": 0.59375, + "reward_std": 0.18776236474514008, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4015.0, + "completions/mean_length": 769.9486694335938, + "completions/mean_terminated_length": 598.1807861328125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.317201166180758, + "grad_norm": 0.13895389437675476, + "learning_rate": 1e-06, + "loss": -0.018, + "num_tokens": 279721281.0, + "reward": 0.6651785969734192, + "reward_std": 0.15849560499191284, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 462 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3963.0, + "completions/mean_length": 913.25341796875, + "completions/mean_terminated_length": 622.5030517578125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 4.326530612244898, + "grad_norm": 0.12598152458667755, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 280322884.0, + "reward": 0.6004464626312256, + "reward_std": 0.1537216752767563, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 463 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2918.0, + "completions/mean_length": 777.5267944335938, + "completions/mean_terminated_length": 650.6326293945312, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 4.335860058309038, + "grad_norm": 0.11926654726266861, + "learning_rate": 1e-06, + "loss": -0.0157, + "num_tokens": 280971196.0, + "reward": 0.6651785969734192, + "reward_std": 0.16423355042934418, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 464 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3446.0, + "completions/mean_length": 753.302490234375, + "completions/mean_terminated_length": 564.0931396484375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 4.345189504373177, + "grad_norm": 0.12969191372394562, + "learning_rate": 1e-06, + "loss": -0.018, + "num_tokens": 281536491.0, + "reward": 0.6875000596046448, + "reward_std": 0.13549628853797913, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 465 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4053.0, + "completions/mean_length": 792.3471069335938, + "completions/mean_terminated_length": 605.347900390625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 4.354518950437318, + "grad_norm": 0.13331782817840576, + "learning_rate": 1e-06, + "loss": -0.0055, + "num_tokens": 282137474.0, + "reward": 0.6830357313156128, + "reward_std": 0.15571656823158264, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2542.0, + "completions/mean_length": 755.9285888671875, + "completions/mean_terminated_length": 591.6627197265625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.363848396501457, + "grad_norm": 0.13045676052570343, + "learning_rate": 1e-06, + "loss": -0.0226, + "num_tokens": 282730218.0, + "reward": 0.6339285969734192, + "reward_std": 0.17232321202754974, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4032.0, + "completions/mean_length": 810.1127319335938, + "completions/mean_terminated_length": 640.4190063476562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 4.373177842565598, + "grad_norm": 0.12199151515960693, + "learning_rate": 1e-06, + "loss": -0.014, + "num_tokens": 283361479.0, + "reward": 0.6674107313156128, + "reward_std": 0.14549003541469574, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140389680862427, + "step": 468 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3275.0, + "completions/mean_length": 842.6585083007812, + "completions/mean_terminated_length": 654.4486083984375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 4.382507288629737, + "grad_norm": 0.1311843991279602, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 284005717.0, + "reward": 0.6171875, + "reward_std": 0.16037212312221527, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 469 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3587.0, + "completions/mean_length": 937.1808471679688, + "completions/mean_terminated_length": 652.8102416992188, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 4.391836734693878, + "grad_norm": 0.14349058270454407, + "learning_rate": 1e-06, + "loss": -0.0111, + "num_tokens": 284638343.0, + "reward": 0.5736607313156128, + "reward_std": 0.14879831671714783, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 470 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2885.0, + "completions/mean_length": 918.982177734375, + "completions/mean_terminated_length": 641.378662109375, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 4.401166180758017, + "grad_norm": 0.12580884993076324, + "learning_rate": 1e-06, + "loss": -0.0331, + "num_tokens": 285264471.0, + "reward": 0.6127232313156128, + "reward_std": 0.16217157244682312, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 471 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4043.0, + "completions/mean_length": 778.2545166015625, + "completions/mean_terminated_length": 606.91552734375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 4.410495626822158, + "grad_norm": 0.13940799236297607, + "learning_rate": 1e-06, + "loss": -0.0005, + "num_tokens": 285873011.0, + "reward": 0.6283482313156128, + "reward_std": 0.1662726104259491, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 472 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3661.0, + "completions/mean_length": 871.7176513671875, + "completions/mean_terminated_length": 602.7025756835938, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 4.419825072886297, + "grad_norm": 0.15017427504062653, + "learning_rate": 1e-06, + "loss": -0.0325, + "num_tokens": 286463270.0, + "reward": 0.6116071939468384, + "reward_std": 0.20688821375370026, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 473 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3859.0, + "completions/mean_length": 858.1875610351562, + "completions/mean_terminated_length": 646.43994140625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 4.429154518950437, + "grad_norm": 0.14438757300376892, + "learning_rate": 1e-06, + "loss": -0.0186, + "num_tokens": 287097894.0, + "reward": 0.6551339626312256, + "reward_std": 0.1825454831123352, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 474 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4050.0, + "completions/mean_length": 959.6607666015625, + "completions/mean_terminated_length": 673.1498413085938, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 4.438483965014577, + "grad_norm": 0.12239904701709747, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 287741798.0, + "reward": 0.5714285969734192, + "reward_std": 0.13816626369953156, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514806270599365, + "step": 475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3663.0, + "completions/mean_length": 794.3817138671875, + "completions/mean_terminated_length": 599.2506103515625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 4.447813411078717, + "grad_norm": 0.13859078288078308, + "learning_rate": 1e-06, + "loss": -0.002, + "num_tokens": 288337772.0, + "reward": 0.6517857313156128, + "reward_std": 0.17310599982738495, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 476 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2752.0, + "completions/mean_length": 817.7221069335938, + "completions/mean_terminated_length": 595.0023803710938, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 4.457142857142857, + "grad_norm": 0.15692879259586334, + "learning_rate": 1e-06, + "loss": -0.0319, + "num_tokens": 288922819.0, + "reward": 0.6707589626312256, + "reward_std": 0.20023773610591888, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 477 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4039.0, + "completions/mean_length": 800.9163208007812, + "completions/mean_terminated_length": 658.9860229492188, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.466472303206997, + "grad_norm": 0.13497665524482727, + "learning_rate": 1e-06, + "loss": -0.0228, + "num_tokens": 289576384.0, + "reward": 0.6361607313156128, + "reward_std": 0.1835193783044815, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 478 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3026.0, + "completions/mean_length": 891.0670166015625, + "completions/mean_terminated_length": 606.789794921875, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 4.475801749271137, + "grad_norm": 0.14471465349197388, + "learning_rate": 1e-06, + "loss": -0.0246, + "num_tokens": 290166556.0, + "reward": 0.652901828289032, + "reward_std": 0.17415766417980194, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 479 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3387.0, + "completions/mean_length": 972.4263916015625, + "completions/mean_terminated_length": 661.9852905273438, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 4.485131195335277, + "grad_norm": 0.1347092241048813, + "learning_rate": 1e-06, + "loss": -0.0435, + "num_tokens": 290799898.0, + "reward": 0.59375, + "reward_std": 0.17303043603897095, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 480 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3770.0, + "completions/mean_length": 843.364990234375, + "completions/mean_terminated_length": 584.7216796875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 4.494460641399417, + "grad_norm": 0.13890087604522705, + "learning_rate": 1e-06, + "loss": -0.0413, + "num_tokens": 291371993.0, + "reward": 0.6194196939468384, + "reward_std": 0.1665322184562683, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 481 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4062.0, + "completions/mean_length": 952.92529296875, + "completions/mean_terminated_length": 636.3009643554688, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 4.503790087463557, + "grad_norm": 0.1366848349571228, + "learning_rate": 1e-06, + "loss": -0.0119, + "num_tokens": 291982182.0, + "reward": 0.6238839626312256, + "reward_std": 0.16931098699569702, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 482 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2581.0, + "completions/mean_length": 802.8694458007812, + "completions/mean_terminated_length": 558.056396484375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 4.513119533527696, + "grad_norm": 0.1445130556821823, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 292527769.0, + "reward": 0.6941964626312256, + "reward_std": 0.14673562347888947, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 483 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3284.0, + "completions/mean_length": 871.6986694335938, + "completions/mean_terminated_length": 660.834716796875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 4.522448979591837, + "grad_norm": 0.12720076739788055, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 293165827.0, + "reward": 0.6305803656578064, + "reward_std": 0.16476628184318542, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3721.0, + "completions/mean_length": 1004.193115234375, + "completions/mean_terminated_length": 663.213134765625, + "completions/min_length": 196.0, + "completions/min_terminated_length": 196.0, + "epoch": 4.531778425655976, + "grad_norm": 0.12677869200706482, + "learning_rate": 1e-06, + "loss": -0.0289, + "num_tokens": 293793920.0, + "reward": 0.566964328289032, + "reward_std": 0.15751849114894867, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3881.0, + "completions/mean_length": 774.974365234375, + "completions/mean_terminated_length": 603.4659423828125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 4.541107871720117, + "grad_norm": 0.12845060229301453, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 294386425.0, + "reward": 0.6595982313156128, + "reward_std": 0.1436467170715332, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 486 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3193.0, + "completions/mean_length": 922.3795166015625, + "completions/mean_terminated_length": 636.6763916015625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.550437317784256, + "grad_norm": 0.12636502087116241, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 294995781.0, + "reward": 0.6618303656578064, + "reward_std": 0.14692038297653198, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.47335079312324524, + "step": 487 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2357.0, + "completions/mean_length": 782.0826416015625, + "completions/mean_terminated_length": 573.7343139648438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 4.559766763848397, + "grad_norm": 0.13962702453136444, + "learning_rate": 1e-06, + "loss": -0.0223, + "num_tokens": 295565335.0, + "reward": 0.738839328289032, + "reward_std": 0.16878741979599, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 488 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3860.0, + "completions/mean_length": 947.0770263671875, + "completions/mean_terminated_length": 646.81298828125, + "completions/min_length": 84.0, + "completions/min_terminated_length": 84.0, + "epoch": 4.569096209912536, + "grad_norm": 0.132938951253891, + "learning_rate": 1e-06, + "loss": -0.0378, + "num_tokens": 296187868.0, + "reward": 0.6350446939468384, + "reward_std": 0.15600895881652832, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 489 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3187.0, + "completions/mean_length": 886.2623291015625, + "completions/mean_terminated_length": 647.648681640625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.578425655976677, + "grad_norm": 0.1473442018032074, + "learning_rate": 1e-06, + "loss": -0.037, + "num_tokens": 296825559.0, + "reward": 0.5892857313156128, + "reward_std": 0.20305675268173218, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 490 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2786.0, + "completions/mean_length": 818.0067138671875, + "completions/mean_terminated_length": 603.6314086914062, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 4.587755102040816, + "grad_norm": 0.12897878885269165, + "learning_rate": 1e-06, + "loss": -0.0177, + "num_tokens": 297419053.0, + "reward": 0.6852678656578064, + "reward_std": 0.14741073548793793, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 491 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3134.0, + "completions/mean_length": 774.5335083007812, + "completions/mean_terminated_length": 619.3247680664062, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 4.597084548104956, + "grad_norm": 0.14938555657863617, + "learning_rate": 1e-06, + "loss": 0.0102, + "num_tokens": 298037499.0, + "reward": 0.6361607313156128, + "reward_std": 0.1947152316570282, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 492 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3953.0, + "completions/mean_length": 858.6283569335938, + "completions/mean_terminated_length": 592.7572631835938, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 4.606413994169096, + "grad_norm": 0.147346630692482, + "learning_rate": 1e-06, + "loss": -0.0278, + "num_tokens": 298614094.0, + "reward": 0.6049107313156128, + "reward_std": 0.14522789418697357, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 493 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3926.0, + "completions/mean_length": 864.3370971679688, + "completions/mean_terminated_length": 632.3994750976562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.615743440233236, + "grad_norm": 0.14167658984661102, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 299222164.0, + "reward": 0.6674107313156128, + "reward_std": 0.183292955160141, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 494 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2834.0, + "completions/mean_length": 816.4955444335938, + "completions/mean_terminated_length": 581.1243896484375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 4.625072886297376, + "grad_norm": 0.1331305205821991, + "learning_rate": 1e-06, + "loss": -0.0076, + "num_tokens": 299793616.0, + "reward": 0.7020089626312256, + "reward_std": 0.14985774457454681, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 495 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 894.37841796875, + "completions/mean_terminated_length": 627.2539672851562, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 4.634402332361516, + "grad_norm": 0.14743056893348694, + "learning_rate": 1e-06, + "loss": -0.0133, + "num_tokens": 300404115.0, + "reward": 0.5870535969734192, + "reward_std": 0.18847593665122986, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 496 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2680.0, + "completions/mean_length": 707.8605346679688, + "completions/mean_terminated_length": 578.3024291992188, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 4.643731778425656, + "grad_norm": 0.1283249855041504, + "learning_rate": 1e-06, + "loss": -0.0289, + "num_tokens": 300986022.0, + "reward": 0.6796875596046448, + "reward_std": 0.15368029475212097, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 497 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2971.0, + "completions/mean_length": 973.654052734375, + "completions/mean_terminated_length": 663.3349609375, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 4.653061224489796, + "grad_norm": 0.1258484274148941, + "learning_rate": 1e-06, + "loss": -0.0395, + "num_tokens": 301616176.0, + "reward": 0.6015625, + "reward_std": 0.18159796297550201, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 498 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3528.0, + "completions/mean_length": 887.72216796875, + "completions/mean_terminated_length": 594.6394653320312, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 4.662390670553936, + "grad_norm": 0.1510901153087616, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 302195407.0, + "reward": 0.676339328289032, + "reward_std": 0.1686376929283142, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 499 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3967.0, + "completions/mean_length": 866.3348388671875, + "completions/mean_terminated_length": 601.0966186523438, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 4.671720116618076, + "grad_norm": 0.15007464587688446, + "learning_rate": 1e-06, + "loss": -0.0443, + "num_tokens": 302784259.0, + "reward": 0.6283482313156128, + "reward_std": 0.16394002735614777, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 500 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2864.0, + "completions/mean_length": 903.3817138671875, + "completions/mean_terminated_length": 641.1859741210938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 4.681049562682215, + "grad_norm": 0.12818759679794312, + "learning_rate": 1e-06, + "loss": -0.036, + "num_tokens": 303410361.0, + "reward": 0.6183035969734192, + "reward_std": 0.16732317209243774, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 501 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3536.0, + "completions/mean_length": 868.4710083007812, + "completions/mean_terminated_length": 640.9629516601562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 4.690379008746356, + "grad_norm": 0.1186971366405487, + "learning_rate": 1e-06, + "loss": -0.0143, + "num_tokens": 304045423.0, + "reward": 0.566964328289032, + "reward_std": 0.1632988154888153, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 965.685302734375, + "completions/mean_terminated_length": 629.0506591796875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 4.699708454810495, + "grad_norm": 0.15227293968200684, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 304648605.0, + "reward": 0.5948660969734192, + "reward_std": 0.20474673807621002, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3013.0, + "completions/mean_length": 987.505615234375, + "completions/mean_terminated_length": 661.707763671875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.709037900874636, + "grad_norm": 0.1331615000963211, + "learning_rate": 1e-06, + "loss": -0.0357, + "num_tokens": 305276842.0, + "reward": 0.6238839626312256, + "reward_std": 0.16461291909217834, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 504 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2731.0, + "completions/mean_length": 899.0592041015625, + "completions/mean_terminated_length": 611.2567138671875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 4.718367346938775, + "grad_norm": 0.12163786590099335, + "learning_rate": 1e-06, + "loss": -0.0246, + "num_tokens": 305874063.0, + "reward": 0.6171875, + "reward_std": 0.14027062058448792, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 849.5324096679688, + "completions/mean_terminated_length": 633.1011962890625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 4.727696793002916, + "grad_norm": 0.13531754910945892, + "learning_rate": 1e-06, + "loss": -0.0191, + "num_tokens": 306495380.0, + "reward": 0.6004464626312256, + "reward_std": 0.14173416793346405, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 506 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2874.0, + "completions/mean_length": 798.4933471679688, + "completions/mean_terminated_length": 591.1767578125, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 4.737026239067055, + "grad_norm": 0.15166935324668884, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 307091854.0, + "reward": 0.6696428656578064, + "reward_std": 0.17446286976337433, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 507 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3327.0, + "completions/mean_length": 1000.7957763671875, + "completions/mean_terminated_length": 672.169189453125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 4.746355685131196, + "grad_norm": 0.12078892439603806, + "learning_rate": 1e-06, + "loss": -0.0146, + "num_tokens": 307731023.0, + "reward": 0.551339328289032, + "reward_std": 0.14522789418697357, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 508 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 817.6585083007812, + "completions/mean_terminated_length": 594.9344482421875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 4.755685131195335, + "grad_norm": 0.15145684778690338, + "learning_rate": 1e-06, + "loss": -0.0267, + "num_tokens": 308323797.0, + "reward": 0.6540178656578064, + "reward_std": 0.17115573585033417, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 509 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3160.0, + "completions/mean_length": 801.3114013671875, + "completions/mean_terminated_length": 631.1631469726562, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 4.765014577259475, + "grad_norm": 0.13937285542488098, + "learning_rate": 1e-06, + "loss": -0.0264, + "num_tokens": 308946148.0, + "reward": 0.7209821939468384, + "reward_std": 0.18295523524284363, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 510 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4023.0, + "completions/mean_length": 898.5592041015625, + "completions/mean_terminated_length": 664.9736938476562, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.774344023323615, + "grad_norm": 0.1371222287416458, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 309604385.0, + "reward": 0.6127232313156128, + "reward_std": 0.19328516721725464, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 511 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2417.0, + "completions/mean_length": 759.6551513671875, + "completions/mean_terminated_length": 549.8967895507812, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.783673469387755, + "grad_norm": 0.1425262838602066, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 310152996.0, + "reward": 0.6718750596046448, + "reward_std": 0.1539812833070755, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 512 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3742.0, + "completions/mean_length": 854.3560791015625, + "completions/mean_terminated_length": 617.5413208007812, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.793002915451895, + "grad_norm": 0.127227321267128, + "learning_rate": 1e-06, + "loss": -0.0182, + "num_tokens": 310760955.0, + "reward": 0.6071428656578064, + "reward_std": 0.14597856998443604, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 513 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3814.0, + "completions/mean_length": 845.8772583007812, + "completions/mean_terminated_length": 595.8677978515625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.802332361516035, + "grad_norm": 0.15406261384487152, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 311337605.0, + "reward": 0.6305803656578064, + "reward_std": 0.18859288096427917, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 514 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3212.0, + "completions/mean_length": 929.5881958007812, + "completions/mean_terminated_length": 631.891357421875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.811661807580175, + "grad_norm": 0.12791962921619415, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 311951900.0, + "reward": 0.6116071939468384, + "reward_std": 0.14661939442157745, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 515 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3723.0, + "completions/mean_length": 839.4810791015625, + "completions/mean_terminated_length": 630.6306762695312, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 4.820991253644315, + "grad_norm": 0.1630844622850418, + "learning_rate": 1e-06, + "loss": -0.0208, + "num_tokens": 312576987.0, + "reward": 0.609375, + "reward_std": 0.20154792070388794, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 516 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3576.0, + "completions/mean_length": 826.0379638671875, + "completions/mean_terminated_length": 632.7777709960938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.830320699708455, + "grad_norm": 0.1412864774465561, + "learning_rate": 1e-06, + "loss": -0.0176, + "num_tokens": 313201189.0, + "reward": 0.6116071939468384, + "reward_std": 0.19366033375263214, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 517 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3066.0, + "completions/mean_length": 941.6607666015625, + "completions/mean_terminated_length": 632.4118041992188, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 4.839650145772595, + "grad_norm": 0.11668366938829422, + "learning_rate": 1e-06, + "loss": -0.02, + "num_tokens": 313799549.0, + "reward": 0.6227678656578064, + "reward_std": 0.13034509122371674, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3242.0, + "completions/mean_length": 943.4453735351562, + "completions/mean_terminated_length": 621.5977783203125, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 4.848979591836734, + "grad_norm": 0.11190957576036453, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 314388012.0, + "reward": 0.609375, + "reward_std": 0.11933468282222748, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 519 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3943.0, + "completions/mean_length": 914.3326416015625, + "completions/mean_terminated_length": 640.516357421875, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 4.858309037900875, + "grad_norm": 0.1547592580318451, + "learning_rate": 1e-06, + "loss": -0.0216, + "num_tokens": 315004270.0, + "reward": 0.578125, + "reward_std": 0.18404366075992584, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 520 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3586.0, + "completions/mean_length": 858.6663208007812, + "completions/mean_terminated_length": 626.32177734375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 4.867638483965014, + "grad_norm": 0.12735597789287567, + "learning_rate": 1e-06, + "loss": -0.0312, + "num_tokens": 315620763.0, + "reward": 0.6350446939468384, + "reward_std": 0.13703903555870056, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 521 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3393.0, + "completions/mean_length": 754.9654541015625, + "completions/mean_terminated_length": 561.682373046875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 4.876967930029155, + "grad_norm": 0.1571730077266693, + "learning_rate": 1e-06, + "loss": -0.0215, + "num_tokens": 316181476.0, + "reward": 0.6986607313156128, + "reward_std": 0.16953922808170319, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960144996643, + "step": 522 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2997.0, + "completions/mean_length": 811.7846069335938, + "completions/mean_terminated_length": 617.6820068359375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 4.886297376093294, + "grad_norm": 0.14319612085819244, + "learning_rate": 1e-06, + "loss": -0.0175, + "num_tokens": 316800435.0, + "reward": 0.6104910969734192, + "reward_std": 0.1746155321598053, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 523 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2788.0, + "completions/mean_length": 829.2176513671875, + "completions/mean_terminated_length": 573.6931762695312, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 4.895626822157435, + "grad_norm": 0.13439147174358368, + "learning_rate": 1e-06, + "loss": -0.0453, + "num_tokens": 317368478.0, + "reward": 0.6305803656578064, + "reward_std": 0.15672755241394043, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 524 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3346.0, + "completions/mean_length": 1004.0335083007812, + "completions/mean_terminated_length": 700.8995361328125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 4.904956268221574, + "grad_norm": 0.1192203015089035, + "learning_rate": 1e-06, + "loss": -0.0383, + "num_tokens": 318036588.0, + "reward": 0.5502232313156128, + "reward_std": 0.17450454831123352, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 525 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3314.0, + "completions/mean_length": 963.818115234375, + "completions/mean_terminated_length": 639.7992553710938, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 4.914285714285715, + "grad_norm": 0.16149286925792694, + "learning_rate": 1e-06, + "loss": -0.0275, + "num_tokens": 318647921.0, + "reward": 0.598214328289032, + "reward_std": 0.19576901197433472, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 526 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4050.0, + "completions/mean_length": 878.7277221679688, + "completions/mean_terminated_length": 593.3560180664062, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 4.923615160349854, + "grad_norm": 0.1332666277885437, + "learning_rate": 1e-06, + "loss": -0.0402, + "num_tokens": 319225717.0, + "reward": 0.59375, + "reward_std": 0.14962315559387207, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 527 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3086.0, + "completions/mean_length": 866.7199096679688, + "completions/mean_terminated_length": 639.08837890625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 4.932944606413994, + "grad_norm": 0.1298314929008484, + "learning_rate": 1e-06, + "loss": -0.0247, + "num_tokens": 319849666.0, + "reward": 0.6473214626312256, + "reward_std": 0.1675507128238678, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 528 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2833.0, + "completions/mean_length": 797.5045166015625, + "completions/mean_terminated_length": 610.7971801757812, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 4.942274052478134, + "grad_norm": 0.13810941576957703, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 320447734.0, + "reward": 0.660714328289032, + "reward_std": 0.1637527495622635, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 529 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 856.8292846679688, + "completions/mean_terminated_length": 632.638427734375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.9516034985422746, + "grad_norm": 0.13064926862716675, + "learning_rate": 1e-06, + "loss": -0.0174, + "num_tokens": 321067693.0, + "reward": 0.6662946939468384, + "reward_std": 0.1661524474620819, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 530 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3872.0, + "completions/mean_length": 811.2098388671875, + "completions/mean_terminated_length": 575.4592895507812, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 4.960932944606414, + "grad_norm": 0.11666761338710785, + "learning_rate": 1e-06, + "loss": -0.0182, + "num_tokens": 321633025.0, + "reward": 0.7488839626312256, + "reward_std": 0.10840277373790741, + "rewards/verify_math_reward/mean": 0.7488839030265808, + "rewards/verify_math_reward/std": 0.43389734625816345, + "step": 531 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3707.0, + "completions/mean_length": 854.0569458007812, + "completions/mean_terminated_length": 621.3815307617188, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 4.970262390670554, + "grad_norm": 0.14021816849708557, + "learning_rate": 1e-06, + "loss": -0.0322, + "num_tokens": 322232188.0, + "reward": 0.652901828289032, + "reward_std": 0.18610312044620514, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 532 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3761.0, + "completions/mean_length": 874.6272583007812, + "completions/mean_terminated_length": 626.829345703125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 4.979591836734694, + "grad_norm": 0.1198548674583435, + "learning_rate": 1e-06, + "loss": -0.011, + "num_tokens": 322846414.0, + "reward": 0.6171875, + "reward_std": 0.12723183631896973, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3739.0, + "completions/mean_length": 937.794677734375, + "completions/mean_terminated_length": 645.0829467773438, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 4.988921282798834, + "grad_norm": 0.13322417438030243, + "learning_rate": 1e-06, + "loss": -0.037, + "num_tokens": 323467590.0, + "reward": 0.6183035969734192, + "reward_std": 0.1519557535648346, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.05965909090909094, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2171.0, + "completions/mean_length": 783.6079711914062, + "completions/mean_terminated_length": 573.4561767578125, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 4.998250728862974, + "grad_norm": 0.12898936867713928, + "learning_rate": 1e-06, + "loss": -0.021, + "num_tokens": 324074522.0, + "reward": 0.6261160969734192, + "reward_std": 0.14034590125083923, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 535 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3500.0, + "completions/mean_length": 813.0547485351562, + "completions/mean_terminated_length": 602.509521484375, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 5.0093294460641395, + "grad_norm": 0.14552631974220276, + "learning_rate": 1e-06, + "loss": -0.0386, + "num_tokens": 324659427.0, + "reward": 0.7008928656578064, + "reward_std": 0.17246659100055695, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 536 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3827.0, + "completions/mean_length": 904.4531860351562, + "completions/mean_terminated_length": 667.1918334960938, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 5.01865889212828, + "grad_norm": 0.13958227634429932, + "learning_rate": 1e-06, + "loss": -0.0179, + "num_tokens": 325297585.0, + "reward": 0.5859375, + "reward_std": 0.19350989162921906, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 537 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3601.0, + "completions/mean_length": 997.4420166015625, + "completions/mean_terminated_length": 701.98046875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.0279883381924195, + "grad_norm": 0.12478427588939667, + "learning_rate": 1e-06, + "loss": -0.0258, + "num_tokens": 325962165.0, + "reward": 0.5959821939468384, + "reward_std": 0.17002525925636292, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 538 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3311.0, + "completions/mean_length": 898.1585083007812, + "completions/mean_terminated_length": 627.1549682617188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.03731778425656, + "grad_norm": 0.15451836585998535, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 326572803.0, + "reward": 0.6238839626312256, + "reward_std": 0.16623163223266602, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3492.0, + "completions/mean_length": 816.6373291015625, + "completions/mean_terminated_length": 643.2279663085938, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 5.0466472303206995, + "grad_norm": 0.13921381533145905, + "learning_rate": 1e-06, + "loss": -0.0279, + "num_tokens": 327203302.0, + "reward": 0.684151828289032, + "reward_std": 0.18637412786483765, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 540 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2842.0, + "completions/mean_length": 900.4531860351562, + "completions/mean_terminated_length": 629.6441040039062, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.05597667638484, + "grad_norm": 0.14924168586730957, + "learning_rate": 1e-06, + "loss": -0.034, + "num_tokens": 327806316.0, + "reward": 0.6595982313156128, + "reward_std": 0.17288917303085327, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 541 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3337.0, + "completions/mean_length": 865.7210083007812, + "completions/mean_terminated_length": 633.8827514648438, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 5.0653061224489795, + "grad_norm": 0.14320100843906403, + "learning_rate": 1e-06, + "loss": -0.0221, + "num_tokens": 328418778.0, + "reward": 0.637276828289032, + "reward_std": 0.17341090738773346, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 542 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2601.0, + "completions/mean_length": 881.5636596679688, + "completions/mean_terminated_length": 613.3700561523438, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 5.07463556851312, + "grad_norm": 0.14697082340717316, + "learning_rate": 1e-06, + "loss": -0.0147, + "num_tokens": 329013483.0, + "reward": 0.625, + "reward_std": 0.16363763809204102, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 543 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3615.0, + "completions/mean_length": 800.9096069335938, + "completions/mean_terminated_length": 634.8030395507812, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 5.0839650145772595, + "grad_norm": 0.13694144785404205, + "learning_rate": 1e-06, + "loss": 0.0026, + "num_tokens": 329648474.0, + "reward": 0.6395089626312256, + "reward_std": 0.1633773148059845, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 544 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3268.0, + "completions/mean_length": 883.349365234375, + "completions/mean_terminated_length": 640.375732421875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 5.093294460641399, + "grad_norm": 0.1366586685180664, + "learning_rate": 1e-06, + "loss": -0.0298, + "num_tokens": 330281163.0, + "reward": 0.5345982313156128, + "reward_std": 0.15372127294540405, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 839.396240234375, + "completions/mean_terminated_length": 584.6679077148438, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 5.1026239067055394, + "grad_norm": 0.14498327672481537, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 330849990.0, + "reward": 0.6629464626312256, + "reward_std": 0.14672675728797913, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 546 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3762.0, + "completions/mean_length": 821.1953735351562, + "completions/mean_terminated_length": 648.0270385742188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.111953352769679, + "grad_norm": 0.12947578728199005, + "learning_rate": 1e-06, + "loss": -0.0013, + "num_tokens": 331491477.0, + "reward": 0.6729910969734192, + "reward_std": 0.16314978897571564, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3571.0, + "completions/mean_length": 809.1517944335938, + "completions/mean_terminated_length": 614.8936157226562, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 5.121282798833819, + "grad_norm": 0.14195102453231812, + "learning_rate": 1e-06, + "loss": -0.0051, + "num_tokens": 332093645.0, + "reward": 0.7198660969734192, + "reward_std": 0.1514332890510559, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 548 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4011.0, + "completions/mean_length": 916.3482666015625, + "completions/mean_terminated_length": 671.7596435546875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 5.130612244897959, + "grad_norm": 0.11762266606092453, + "learning_rate": 1e-06, + "loss": -0.0217, + "num_tokens": 332743541.0, + "reward": 0.6540178656578064, + "reward_std": 0.13771232962608337, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 549 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4011.0, + "completions/mean_length": 954.9063110351562, + "completions/mean_terminated_length": 617.1124877929688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.139941690962099, + "grad_norm": 0.1373228281736374, + "learning_rate": 1e-06, + "loss": -0.0344, + "num_tokens": 333342041.0, + "reward": 0.5234375, + "reward_std": 0.1512710452079773, + "rewards/verify_math_reward/mean": 0.5234375, + "rewards/verify_math_reward/std": 0.49972933530807495, + "step": 550 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2820.0, + "completions/mean_length": 838.5658569335938, + "completions/mean_terminated_length": 604.7786865234375, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 5.149271137026239, + "grad_norm": 0.1289171576499939, + "learning_rate": 1e-06, + "loss": -0.0293, + "num_tokens": 333935972.0, + "reward": 0.6618303656578064, + "reward_std": 0.14420942962169647, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 551 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3400.0, + "completions/mean_length": 900.5803833007812, + "completions/mean_terminated_length": 695.6484985351562, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 5.158600583090379, + "grad_norm": 0.13979865610599518, + "learning_rate": 1e-06, + "loss": -0.0259, + "num_tokens": 334608988.0, + "reward": 0.6674107313156128, + "reward_std": 0.19239293038845062, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 552 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4005.0, + "completions/mean_length": 890.7154541015625, + "completions/mean_terminated_length": 623.285400390625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 5.167930029154519, + "grad_norm": 0.14315195381641388, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 335210669.0, + "reward": 0.645089328289032, + "reward_std": 0.16093555092811584, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 553 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3909.0, + "completions/mean_length": 943.89404296875, + "completions/mean_terminated_length": 672.621826171875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 5.1772594752186585, + "grad_norm": 0.1337359994649887, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 335852998.0, + "reward": 0.6171875, + "reward_std": 0.1865263730287552, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2980.0, + "completions/mean_length": 887.0313110351562, + "completions/mean_terminated_length": 652.6035766601562, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 5.186588921282799, + "grad_norm": 0.1405474841594696, + "learning_rate": 1e-06, + "loss": -0.016, + "num_tokens": 336483826.0, + "reward": 0.6540178656578064, + "reward_std": 0.17167635262012482, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 555 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3039.0, + "completions/mean_length": 836.8828735351562, + "completions/mean_terminated_length": 631.9798583984375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 5.1959183673469385, + "grad_norm": 0.15516424179077148, + "learning_rate": 1e-06, + "loss": -0.0264, + "num_tokens": 337107697.0, + "reward": 0.6774553656578064, + "reward_std": 0.1729990392923355, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 556 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3043.0, + "completions/mean_length": 1003.779052734375, + "completions/mean_terminated_length": 667.00244140625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.205247813411079, + "grad_norm": 0.131962850689888, + "learning_rate": 1e-06, + "loss": -0.0397, + "num_tokens": 337735603.0, + "reward": 0.6127232313156128, + "reward_std": 0.16142340004444122, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 557 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3519.0, + "completions/mean_length": 904.9141235351562, + "completions/mean_terminated_length": 626.081298828125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 5.214577259475218, + "grad_norm": 0.14526274800300598, + "learning_rate": 1e-06, + "loss": -0.0273, + "num_tokens": 338347022.0, + "reward": 0.6595982313156128, + "reward_std": 0.1642010509967804, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 558 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3911.0, + "completions/mean_length": 981.40966796875, + "completions/mean_terminated_length": 625.013671875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 5.223906705539359, + "grad_norm": 0.14314597845077515, + "learning_rate": 1e-06, + "loss": -0.0604, + "num_tokens": 338934845.0, + "reward": 0.645089328289032, + "reward_std": 0.17119666934013367, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 559 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3593.0, + "completions/mean_length": 874.1897583007812, + "completions/mean_terminated_length": 626.3582153320312, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 5.233236151603498, + "grad_norm": 0.13769236207008362, + "learning_rate": 1e-06, + "loss": -0.0309, + "num_tokens": 339545695.0, + "reward": 0.5758928656578064, + "reward_std": 0.1521807461977005, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 560 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3652.0, + "completions/mean_length": 975.404052734375, + "completions/mean_terminated_length": 609.6483764648438, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 5.242565597667639, + "grad_norm": 0.13476085662841797, + "learning_rate": 1e-06, + "loss": -0.0275, + "num_tokens": 340125761.0, + "reward": 0.668526828289032, + "reward_std": 0.15153953433036804, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 561 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2965.0, + "completions/mean_length": 905.052490234375, + "completions/mean_terminated_length": 642.9939575195312, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.251895043731778, + "grad_norm": 0.14517644047737122, + "learning_rate": 1e-06, + "loss": -0.0269, + "num_tokens": 340743480.0, + "reward": 0.6439732313156128, + "reward_std": 0.15518662333488464, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3463.0, + "completions/mean_length": 908.62841796875, + "completions/mean_terminated_length": 642.69287109375, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 5.261224489795918, + "grad_norm": 0.14808538556098938, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 341360259.0, + "reward": 0.6171875, + "reward_std": 0.17972436547279358, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 563 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3911.0, + "completions/mean_length": 990.5904541015625, + "completions/mean_terminated_length": 706.905029296875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 5.270553935860058, + "grad_norm": 0.13488610088825226, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 342040276.0, + "reward": 0.5580357313156128, + "reward_std": 0.18047957122325897, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 564 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2525.0, + "completions/mean_length": 842.6641235351562, + "completions/mean_terminated_length": 609.1710205078125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 5.279883381924198, + "grad_norm": 0.1330379992723465, + "learning_rate": 1e-06, + "loss": -0.0255, + "num_tokens": 342640551.0, + "reward": 0.6339285969734192, + "reward_std": 0.16209810972213745, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 565 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3831.0, + "completions/mean_length": 857.114990234375, + "completions/mean_terminated_length": 624.6590576171875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 5.289212827988338, + "grad_norm": 0.12528836727142334, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 343248670.0, + "reward": 0.6573660969734192, + "reward_std": 0.13763242959976196, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 566 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3858.0, + "completions/mean_length": 933.4922485351562, + "completions/mean_terminated_length": 640.3817138671875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.298542274052478, + "grad_norm": 0.14895907044410706, + "learning_rate": 1e-06, + "loss": -0.0397, + "num_tokens": 343859247.0, + "reward": 0.6339285969734192, + "reward_std": 0.19527794420719147, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 567 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3634.0, + "completions/mean_length": 914.67529296875, + "completions/mean_terminated_length": 628.278564453125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 5.307871720116618, + "grad_norm": 0.1415090411901474, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 344462452.0, + "reward": 0.6328125, + "reward_std": 0.15210728347301483, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 568 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3712.0, + "completions/mean_length": 923.8873291015625, + "completions/mean_terminated_length": 671.64697265625, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 5.317201166180758, + "grad_norm": 0.1356671005487442, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 345112175.0, + "reward": 0.6383928656578064, + "reward_std": 0.15808121860027313, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 569 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 884.427490234375, + "completions/mean_terminated_length": 637.3834228515625, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 5.326530612244898, + "grad_norm": 0.1242910847067833, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 345727446.0, + "reward": 0.6272321939468384, + "reward_std": 0.14841853082180023, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 570 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3432.0, + "completions/mean_length": 917.950927734375, + "completions/mean_terminated_length": 627.6297607421875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 5.335860058309038, + "grad_norm": 0.1463022381067276, + "learning_rate": 1e-06, + "loss": -0.0339, + "num_tokens": 346324954.0, + "reward": 0.6473214626312256, + "reward_std": 0.1641671359539032, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 571 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3445.0, + "completions/mean_length": 877.927490234375, + "completions/mean_terminated_length": 592.48486328125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "epoch": 5.345189504373177, + "grad_norm": 0.12898695468902588, + "learning_rate": 1e-06, + "loss": -0.0132, + "num_tokens": 346905257.0, + "reward": 0.5881696939468384, + "reward_std": 0.13903389871120453, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 572 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 897.3761596679688, + "completions/mean_terminated_length": 617.8847045898438, + "completions/min_length": 82.0, + "completions/min_terminated_length": 82.0, + "epoch": 5.354518950437318, + "grad_norm": 0.13803651928901672, + "learning_rate": 1e-06, + "loss": -0.0225, + "num_tokens": 347508970.0, + "reward": 0.660714328289032, + "reward_std": 0.15026256442070007, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 573 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3207.0, + "completions/mean_length": 823.3761596679688, + "completions/mean_terminated_length": 567.3947143554688, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 5.363848396501457, + "grad_norm": 0.16021214425563812, + "learning_rate": 1e-06, + "loss": -0.0438, + "num_tokens": 348071027.0, + "reward": 0.625, + "reward_std": 0.1699499785900116, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 574 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3627.0, + "completions/mean_length": 974.5279541015625, + "completions/mean_terminated_length": 676.8814697265625, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 5.373177842565598, + "grad_norm": 0.13691644370555878, + "learning_rate": 1e-06, + "loss": -0.0438, + "num_tokens": 348717196.0, + "reward": 0.6082589626312256, + "reward_std": 0.19058279693126678, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841193318367004, + "step": 575 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3303.0, + "completions/mean_length": 861.9475708007812, + "completions/mean_terminated_length": 613.17431640625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 5.382507288629737, + "grad_norm": 0.12933017313480377, + "learning_rate": 1e-06, + "loss": -0.0419, + "num_tokens": 349323901.0, + "reward": 0.6506696939468384, + "reward_std": 0.14774522185325623, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 576 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4060.0, + "completions/mean_length": 879.1105346679688, + "completions/mean_terminated_length": 664.6512451171875, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 5.391836734693878, + "grad_norm": 0.11132606118917465, + "learning_rate": 1e-06, + "loss": -0.0085, + "num_tokens": 349968784.0, + "reward": 0.6897321939468384, + "reward_std": 0.12734243273735046, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2866.0, + "completions/mean_length": 889.68310546875, + "completions/mean_terminated_length": 609.5194091796875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.401166180758017, + "grad_norm": 0.12581051886081696, + "learning_rate": 1e-06, + "loss": -0.0282, + "num_tokens": 350566652.0, + "reward": 0.6439732313156128, + "reward_std": 0.1265924572944641, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 578 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3481.0, + "completions/mean_length": 957.44873046875, + "completions/mean_terminated_length": 645.51904296875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.410495626822158, + "grad_norm": 0.13605797290802002, + "learning_rate": 1e-06, + "loss": -0.0069, + "num_tokens": 351186790.0, + "reward": 0.5848214626312256, + "reward_std": 0.16224753856658936, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 579 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3499.0, + "completions/mean_length": 866.6964721679688, + "completions/mean_terminated_length": 609.908447265625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.419825072886297, + "grad_norm": 0.15027391910552979, + "learning_rate": 1e-06, + "loss": -0.0311, + "num_tokens": 351778750.0, + "reward": 0.652901828289032, + "reward_std": 0.1741601824760437, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 580 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3479.0, + "completions/mean_length": 886.44091796875, + "completions/mean_terminated_length": 593.2412109375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 5.429154518950437, + "grad_norm": 0.1291462481021881, + "learning_rate": 1e-06, + "loss": -0.0175, + "num_tokens": 352352025.0, + "reward": 0.6316964626312256, + "reward_std": 0.14361028373241425, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 581 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3480.0, + "completions/mean_length": 926.9219360351562, + "completions/mean_terminated_length": 624.7359619140625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 5.438483965014577, + "grad_norm": 0.13395194709300995, + "learning_rate": 1e-06, + "loss": -0.0157, + "num_tokens": 352944587.0, + "reward": 0.6629464626312256, + "reward_std": 0.15480755269527435, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 582 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3692.0, + "completions/mean_length": 854.021240234375, + "completions/mean_terminated_length": 596.225341796875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 5.447813411078717, + "grad_norm": 0.1539541482925415, + "learning_rate": 1e-06, + "loss": -0.048, + "num_tokens": 353522342.0, + "reward": 0.6674107313156128, + "reward_std": 0.18044115602970123, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 583 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4027.0, + "completions/mean_length": 858.6685791015625, + "completions/mean_terminated_length": 601.2421875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 5.457142857142857, + "grad_norm": 0.1502072811126709, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 354106013.0, + "reward": 0.629464328289032, + "reward_std": 0.19148364663124084, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 584 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2871.0, + "completions/mean_length": 1015.3906860351562, + "completions/mean_terminated_length": 637.0701904296875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.466472303206997, + "grad_norm": 0.13293175399303436, + "learning_rate": 1e-06, + "loss": -0.0464, + "num_tokens": 354711363.0, + "reward": 0.5915178656578064, + "reward_std": 0.146052747964859, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 585 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2175.0, + "completions/mean_length": 865.8917846679688, + "completions/mean_terminated_length": 592.15380859375, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 5.475801749271137, + "grad_norm": 0.15534572303295135, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 355288218.0, + "reward": 0.6350446939468384, + "reward_std": 0.1977938711643219, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 586 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3632.0, + "completions/mean_length": 995.3605346679688, + "completions/mean_terminated_length": 661.9171752929688, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 5.485131195335277, + "grad_norm": 0.14128495752811432, + "learning_rate": 1e-06, + "loss": -0.0276, + "num_tokens": 355922325.0, + "reward": 0.559151828289032, + "reward_std": 0.16059856116771698, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 587 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3981.0, + "completions/mean_length": 1025.6551513671875, + "completions/mean_terminated_length": 644.269775390625, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 5.494460641399417, + "grad_norm": 0.1482999175786972, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 356533480.0, + "reward": 0.6004464626312256, + "reward_std": 0.17017750442028046, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 588 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3994.0, + "completions/mean_length": 882.4766235351562, + "completions/mean_terminated_length": 639.4369506835938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.503790087463557, + "grad_norm": 0.1332821547985077, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 357151787.0, + "reward": 0.6964285969734192, + "reward_std": 0.16755323112010956, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 589 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2597.0, + "completions/mean_length": 783.0313110351562, + "completions/mean_terminated_length": 566.3685913085938, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 5.513119533527696, + "grad_norm": 0.15688472986221313, + "learning_rate": 1e-06, + "loss": -0.0181, + "num_tokens": 357711823.0, + "reward": 0.6662946939468384, + "reward_std": 0.14586912095546722, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 590 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4051.0, + "completions/mean_length": 1026.529052734375, + "completions/mean_terminated_length": 632.214111328125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 5.522448979591837, + "grad_norm": 0.14070656895637512, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 358309881.0, + "reward": 0.625, + "reward_std": 0.16717232763767242, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 591 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3189.0, + "completions/mean_length": 821.7332763671875, + "completions/mean_terminated_length": 603.4488525390625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.531778425655976, + "grad_norm": 0.13638702034950256, + "learning_rate": 1e-06, + "loss": -0.0094, + "num_tokens": 358918010.0, + "reward": 0.6395089626312256, + "reward_std": 0.13711318373680115, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 592 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3686.0, + "completions/mean_length": 915.25341796875, + "completions/mean_terminated_length": 594.8341674804688, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 5.541107871720117, + "grad_norm": 0.16951455175876617, + "learning_rate": 1e-06, + "loss": -0.0345, + "num_tokens": 359498525.0, + "reward": 0.6462053656578064, + "reward_std": 0.16679435968399048, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 593 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2943.0, + "completions/mean_length": 1003.8616333007812, + "completions/mean_terminated_length": 683.9852294921875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 5.550437317784256, + "grad_norm": 0.15062110126018524, + "learning_rate": 1e-06, + "loss": -0.0574, + "num_tokens": 360138497.0, + "reward": 0.606026828289032, + "reward_std": 0.2011367678642273, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 594 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3001.0, + "completions/mean_length": 1000.88623046875, + "completions/mean_terminated_length": 646.7188720703125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 5.559766763848397, + "grad_norm": 0.1648234874010086, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 360765499.0, + "reward": 0.543526828289032, + "reward_std": 0.1934778094291687, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 595 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3845.0, + "completions/mean_length": 939.7199096679688, + "completions/mean_terminated_length": 626.0281982421875, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 5.569096209912536, + "grad_norm": 0.15192818641662598, + "learning_rate": 1e-06, + "loss": -0.0487, + "num_tokens": 361351512.0, + "reward": 0.645089328289032, + "reward_std": 0.17844374477863312, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 596 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3842.0, + "completions/mean_length": 783.2377319335938, + "completions/mean_terminated_length": 591.59033203125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.578425655976677, + "grad_norm": 0.13535654544830322, + "learning_rate": 1e-06, + "loss": -0.0336, + "num_tokens": 361930173.0, + "reward": 0.684151828289032, + "reward_std": 0.13087642192840576, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 597 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2758.0, + "completions/mean_length": 869.6785888671875, + "completions/mean_terminated_length": 625.6710205078125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 5.587755102040816, + "grad_norm": 0.1371794492006302, + "learning_rate": 1e-06, + "loss": -0.0134, + "num_tokens": 362535869.0, + "reward": 0.6026785969734192, + "reward_std": 0.16575013101100922, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 598 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3988.0, + "completions/mean_length": 941.3906860351562, + "completions/mean_terminated_length": 644.803466796875, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 5.597084548104956, + "grad_norm": 0.12543779611587524, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 363155555.0, + "reward": 0.566964328289032, + "reward_std": 0.1598842889070511, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 599 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3489.0, + "completions/mean_length": 1007.9342041015625, + "completions/mean_terminated_length": 633.0375366210938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 5.606413994169096, + "grad_norm": 0.15942542254924774, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 363748664.0, + "reward": 0.6272321939468384, + "reward_std": 0.17585155367851257, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 600 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3927.0, + "completions/mean_length": 839.8013916015625, + "completions/mean_terminated_length": 643.2733764648438, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 5.615743440233236, + "grad_norm": 0.13871051371097565, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 364380670.0, + "reward": 0.6651785969734192, + "reward_std": 0.19208098948001862, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 601 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 842.1016235351562, + "completions/mean_terminated_length": 616.8914184570312, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 5.625072886297376, + "grad_norm": 0.1313745081424713, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 364977225.0, + "reward": 0.6662946939468384, + "reward_std": 0.16563068330287933, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 602 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3004.0, + "completions/mean_length": 925.3449096679688, + "completions/mean_terminated_length": 644.108154296875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 5.634402332361516, + "grad_norm": 0.16670598089694977, + "learning_rate": 1e-06, + "loss": -0.0206, + "num_tokens": 365602102.0, + "reward": 0.6651785969734192, + "reward_std": 0.20249293744564056, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 603 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3642.0, + "completions/mean_length": 904.255615234375, + "completions/mean_terminated_length": 662.8630981445312, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.643731778425656, + "grad_norm": 0.14878563582897186, + "learning_rate": 1e-06, + "loss": -0.0494, + "num_tokens": 366240707.0, + "reward": 0.6439732313156128, + "reward_std": 0.19475801289081573, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 604 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3581.0, + "completions/mean_length": 888.8348388671875, + "completions/mean_terminated_length": 629.630859375, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 5.653061224489796, + "grad_norm": 0.1483723521232605, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 366852159.0, + "reward": 0.640625, + "reward_std": 0.16923388838768005, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3789.0, + "completions/mean_length": 910.8527221679688, + "completions/mean_terminated_length": 611.3944091796875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 5.662390670553936, + "grad_norm": 0.15718428790569305, + "learning_rate": 1e-06, + "loss": -0.046, + "num_tokens": 367440075.0, + "reward": 0.6808035969734192, + "reward_std": 0.19385577738285065, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 606 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3415.0, + "completions/mean_length": 916.1484985351562, + "completions/mean_terminated_length": 629.8843994140625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.671720116618076, + "grad_norm": 0.13625109195709229, + "learning_rate": 1e-06, + "loss": -0.0343, + "num_tokens": 368044632.0, + "reward": 0.6796875596046448, + "reward_std": 0.15680313110351562, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 607 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3278.0, + "completions/mean_length": 813.708740234375, + "completions/mean_terminated_length": 586.533447265625, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 5.681049562682215, + "grad_norm": 0.12798573076725006, + "learning_rate": 1e-06, + "loss": -0.0335, + "num_tokens": 368630211.0, + "reward": 0.6886160969734192, + "reward_std": 0.14744353294372559, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 608 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2993.0, + "completions/mean_length": 1024.04248046875, + "completions/mean_terminated_length": 668.261474609375, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 5.690379008746356, + "grad_norm": 0.13185234367847443, + "learning_rate": 1e-06, + "loss": -0.0526, + "num_tokens": 369256209.0, + "reward": 0.613839328289032, + "reward_std": 0.16453805565834045, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 609 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2891.0, + "completions/mean_length": 845.8170166015625, + "completions/mean_terminated_length": 548.90625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 5.699708454810495, + "grad_norm": 0.14759176969528198, + "learning_rate": 1e-06, + "loss": -0.0241, + "num_tokens": 369798349.0, + "reward": 0.6785714626312256, + "reward_std": 0.14330969750881195, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 610 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2774.0, + "completions/mean_length": 874.9922485351562, + "completions/mean_terminated_length": 618.8638916015625, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 5.709037900874636, + "grad_norm": 0.1345078945159912, + "learning_rate": 1e-06, + "loss": -0.034, + "num_tokens": 370407406.0, + "reward": 0.637276828289032, + "reward_std": 0.1731833815574646, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 611 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3951.0, + "completions/mean_length": 1105.0692138671875, + "completions/mean_terminated_length": 646.9987182617188, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 5.718367346938775, + "grad_norm": 0.18225589394569397, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 371009772.0, + "reward": 0.629464328289032, + "reward_std": 0.23840835690498352, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 612 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3014.0, + "completions/mean_length": 974.927490234375, + "completions/mean_terminated_length": 609.115966796875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.727696793002916, + "grad_norm": 0.15671861171722412, + "learning_rate": 1e-06, + "loss": -0.0249, + "num_tokens": 371587939.0, + "reward": 0.6428571939468384, + "reward_std": 0.14940449595451355, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 613 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 941.404052734375, + "completions/mean_terminated_length": 606.4716186523438, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 5.737026239067055, + "grad_norm": 0.15861521661281586, + "learning_rate": 1e-06, + "loss": -0.0469, + "num_tokens": 372172917.0, + "reward": 0.6506696939468384, + "reward_std": 0.13771162927150726, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 614 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3809.0, + "completions/mean_length": 953.4141235351562, + "completions/mean_terminated_length": 632.584228515625, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 5.746355685131196, + "grad_norm": 0.14253567159175873, + "learning_rate": 1e-06, + "loss": -0.0622, + "num_tokens": 372775704.0, + "reward": 0.6116071939468384, + "reward_std": 0.17926721274852753, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 615 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2692.0, + "completions/mean_length": 910.1082763671875, + "completions/mean_terminated_length": 602.0477294921875, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 5.755685131195335, + "grad_norm": 0.14117032289505005, + "learning_rate": 1e-06, + "loss": -0.0069, + "num_tokens": 373360409.0, + "reward": 0.6662946939468384, + "reward_std": 0.14951257407665253, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 616 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2807.0, + "completions/mean_length": 927.6886596679688, + "completions/mean_terminated_length": 608.5220947265625, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 5.765014577259475, + "grad_norm": 0.15116062760353088, + "learning_rate": 1e-06, + "loss": -0.0508, + "num_tokens": 373942082.0, + "reward": 0.6662946939468384, + "reward_std": 0.17111042141914368, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 617 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3099.0, + "completions/mean_length": 1008.1842041015625, + "completions/mean_terminated_length": 671.8873901367188, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 5.774344023323615, + "grad_norm": 0.12510289251804352, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 374569255.0, + "reward": 0.629464328289032, + "reward_std": 0.18006137013435364, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 618 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3745.0, + "completions/mean_length": 974.8125610351562, + "completions/mean_terminated_length": 600.2699584960938, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 5.783673469387755, + "grad_norm": 0.13947227597236633, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 375138975.0, + "reward": 0.6272321939468384, + "reward_std": 0.16311588883399963, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 619 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3447.0, + "completions/mean_length": 863.1886596679688, + "completions/mean_terminated_length": 589.2215576171875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 5.793002915451895, + "grad_norm": 0.15783898532390594, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 375722552.0, + "reward": 0.6640625, + "reward_std": 0.1863730102777481, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 620 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 905.7701416015625, + "completions/mean_terminated_length": 631.2169799804688, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.802332361516035, + "grad_norm": 0.1397719383239746, + "learning_rate": 1e-06, + "loss": -0.0232, + "num_tokens": 376330386.0, + "reward": 0.6395089626312256, + "reward_std": 0.14812499284744263, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111123085022, + "step": 621 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3536.0, + "completions/mean_length": 915.5569458007812, + "completions/mean_terminated_length": 641.8460693359375, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 5.811661807580175, + "grad_norm": 0.12052565068006516, + "learning_rate": 1e-06, + "loss": -0.0136, + "num_tokens": 376944117.0, + "reward": 0.6584821939468384, + "reward_std": 0.1407175064086914, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 622 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3911.0, + "completions/mean_length": 1081.75, + "completions/mean_terminated_length": 703.0753784179688, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 5.820991253644315, + "grad_norm": 0.14530135691165924, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 377604165.0, + "reward": 0.6149553656578064, + "reward_std": 0.16424313187599182, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3502.0, + "completions/mean_length": 1032.03466796875, + "completions/mean_terminated_length": 642.7760620117188, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 5.830320699708455, + "grad_norm": 0.13524962961673737, + "learning_rate": 1e-06, + "loss": -0.0535, + "num_tokens": 378216948.0, + "reward": 0.606026828289032, + "reward_std": 0.16584935784339905, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890191316604614, + "step": 624 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3559.0, + "completions/mean_length": 846.4107666015625, + "completions/mean_terminated_length": 592.2310791015625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 5.839650145772595, + "grad_norm": 0.14528243243694305, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 378788724.0, + "reward": 0.723214328289032, + "reward_std": 0.16186915338039398, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 625 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2959.0, + "completions/mean_length": 850.9609985351562, + "completions/mean_terminated_length": 575.9576416015625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.848979591836734, + "grad_norm": 0.1341572254896164, + "learning_rate": 1e-06, + "loss": -0.0272, + "num_tokens": 379344441.0, + "reward": 0.652901828289032, + "reward_std": 0.1303137093782425, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 626 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2837.0, + "completions/mean_length": 822.5301513671875, + "completions/mean_terminated_length": 620.84716796875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.858309037900875, + "grad_norm": 0.13909585773944855, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 379963396.0, + "reward": 0.6718750596046448, + "reward_std": 0.18490804731845856, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 627 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3902.0, + "completions/mean_length": 909.5045166015625, + "completions/mean_terminated_length": 631.0728149414062, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 5.867638483965014, + "grad_norm": 0.14758238196372986, + "learning_rate": 1e-06, + "loss": -0.0268, + "num_tokens": 380563704.0, + "reward": 0.6439732313156128, + "reward_std": 0.16259340941905975, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 628 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3701.0, + "completions/mean_length": 973.08154296875, + "completions/mean_terminated_length": 632.9616088867188, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 5.876967930029155, + "grad_norm": 0.13139067590236664, + "learning_rate": 1e-06, + "loss": -0.0476, + "num_tokens": 381173257.0, + "reward": 0.6875000596046448, + "reward_std": 0.1308029592037201, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 629 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2968.0, + "completions/mean_length": 920.4230346679688, + "completions/mean_terminated_length": 613.35986328125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 5.886297376093294, + "grad_norm": 0.15726451575756073, + "learning_rate": 1e-06, + "loss": -0.0129, + "num_tokens": 381766404.0, + "reward": 0.6707589626312256, + "reward_std": 0.17145811021327972, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 630 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2643.0, + "completions/mean_length": 918.8381958007812, + "completions/mean_terminated_length": 598.7800903320312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 5.895626822157435, + "grad_norm": 0.15146227180957794, + "learning_rate": 1e-06, + "loss": -0.0466, + "num_tokens": 382341499.0, + "reward": 0.6283482313156128, + "reward_std": 0.14846019446849823, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 631 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2318.0, + "completions/mean_length": 945.5870971679688, + "completions/mean_terminated_length": 636.7230834960938, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 5.904956268221574, + "grad_norm": 0.15846389532089233, + "learning_rate": 1e-06, + "loss": -0.0341, + "num_tokens": 382958017.0, + "reward": 0.582589328289032, + "reward_std": 0.18277119100093842, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.493407279253006, + "step": 632 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2992.0, + "completions/mean_length": 870.130615234375, + "completions/mean_terminated_length": 592.5103149414062, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 5.914285714285715, + "grad_norm": 0.1437930166721344, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 383527846.0, + "reward": 0.65625, + "reward_std": 0.1528201550245285, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 633 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4032.0, + "completions/mean_length": 842.7422485351562, + "completions/mean_terminated_length": 617.576416015625, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 5.923615160349854, + "grad_norm": 0.140001580119133, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 384137615.0, + "reward": 0.676339328289032, + "reward_std": 0.16029614210128784, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 634 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3173.0, + "completions/mean_length": 1009.1004638671875, + "completions/mean_terminated_length": 647.2943115234375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.932944606413994, + "grad_norm": 0.15843509137630463, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 384754145.0, + "reward": 0.629464328289032, + "reward_std": 0.19061671197414398, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 635 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2906.0, + "completions/mean_length": 987.6495971679688, + "completions/mean_terminated_length": 636.2708129882812, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 5.942274052478134, + "grad_norm": 0.14235179126262665, + "learning_rate": 1e-06, + "loss": -0.0396, + "num_tokens": 385353463.0, + "reward": 0.6383928656578064, + "reward_std": 0.15075111389160156, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4055.0, + "completions/mean_length": 1003.5859985351562, + "completions/mean_terminated_length": 658.2791748046875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 5.9516034985422746, + "grad_norm": 0.13037067651748657, + "learning_rate": 1e-06, + "loss": -0.0305, + "num_tokens": 385979764.0, + "reward": 0.5970982313156128, + "reward_std": 0.16300641000270844, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 637 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3693.0, + "completions/mean_length": 1024.9888916015625, + "completions/mean_terminated_length": 690.5222778320312, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.960932944606414, + "grad_norm": 0.1297571361064911, + "learning_rate": 1e-06, + "loss": -0.0262, + "num_tokens": 386633442.0, + "reward": 0.6227678656578064, + "reward_std": 0.15405938029289246, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644601345062, + "step": 638 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2324.0, + "completions/mean_length": 896.3527221679688, + "completions/mean_terminated_length": 595.5311279296875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 5.970262390670554, + "grad_norm": 0.13038305938243866, + "learning_rate": 1e-06, + "loss": -0.0515, + "num_tokens": 387212398.0, + "reward": 0.6462053656578064, + "reward_std": 0.14158262312412262, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 639 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3936.0, + "completions/mean_length": 919.154052734375, + "completions/mean_terminated_length": 564.4193725585938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 5.979591836734694, + "grad_norm": 0.1239788755774498, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 387747560.0, + "reward": 0.6551339626312256, + "reward_std": 0.12448880821466446, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 640 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2632.0, + "completions/mean_length": 1172.763427734375, + "completions/mean_terminated_length": 626.83447265625, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 5.988921282798834, + "grad_norm": 0.16144493222236633, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 388320788.0, + "reward": 0.5814732313156128, + "reward_std": 0.1644297093153, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 641 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.11931818181818177, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3491.0, + "completions/mean_length": 1079.23583984375, + "completions/mean_terminated_length": 670.5128784179688, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 5.998250728862974, + "grad_norm": 0.12611761689186096, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 388903151.0, + "reward": 0.676339328289032, + "reward_std": 0.1252797544002533, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 642 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3967.0, + "completions/mean_length": 978.6897583007812, + "completions/mean_terminated_length": 673.0711059570312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 6.0093294460641395, + "grad_norm": 0.14926742017269135, + "learning_rate": 1e-06, + "loss": -0.0228, + "num_tokens": 389557321.0, + "reward": 0.6316964626312256, + "reward_std": 0.17104442417621613, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 643 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3307.0, + "completions/mean_length": 850.5692138671875, + "completions/mean_terminated_length": 566.9878540039062, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 6.01865889212828, + "grad_norm": 0.15022827684879303, + "learning_rate": 1e-06, + "loss": -0.0366, + "num_tokens": 390102135.0, + "reward": 0.6674107313156128, + "reward_std": 0.1798299103975296, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140389680862427, + "step": 644 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3990.0, + "completions/mean_length": 1025.227783203125, + "completions/mean_terminated_length": 613.2000122070312, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 6.0279883381924195, + "grad_norm": 0.17383526265621185, + "learning_rate": 1e-06, + "loss": -0.0649, + "num_tokens": 390678091.0, + "reward": 0.637276828289032, + "reward_std": 0.2123749703168869, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 645 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2602.0, + "completions/mean_length": 1063.946533203125, + "completions/mean_terminated_length": 621.9334716796875, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 6.03731778425656, + "grad_norm": 0.15711233019828796, + "learning_rate": 1e-06, + "loss": -0.0362, + "num_tokens": 391265275.0, + "reward": 0.5736607313156128, + "reward_std": 0.16465681791305542, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 646 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4096.0, + "completions/mean_length": 1064.532470703125, + "completions/mean_terminated_length": 700.7562255859375, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 6.0466472303206995, + "grad_norm": 0.12954795360565186, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 391910216.0, + "reward": 0.629464328289032, + "reward_std": 0.1645815223455429, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 647 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4059.0, + "completions/mean_length": 922.4788208007812, + "completions/mean_terminated_length": 607.0736083984375, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 6.05597667638484, + "grad_norm": 0.1555488258600235, + "learning_rate": 1e-06, + "loss": -0.0375, + "num_tokens": 392492517.0, + "reward": 0.6551339626312256, + "reward_std": 0.18472330272197723, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 648 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3981.0, + "completions/mean_length": 1020.1685791015625, + "completions/mean_terminated_length": 638.1016235351562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 6.0653061224489795, + "grad_norm": 0.12056968361139297, + "learning_rate": 1e-06, + "loss": -0.0343, + "num_tokens": 393088804.0, + "reward": 0.6473214626312256, + "reward_std": 0.12651507556438446, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 649 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2589.0, + "completions/mean_length": 898.1864013671875, + "completions/mean_terminated_length": 584.67529296875, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 6.07463556851312, + "grad_norm": 0.134687140583992, + "learning_rate": 1e-06, + "loss": -0.033, + "num_tokens": 393657235.0, + "reward": 0.6473214626312256, + "reward_std": 0.14409995079040527, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 650 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2604.0, + "completions/mean_length": 807.2020263671875, + "completions/mean_terminated_length": 579.576416015625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 6.0839650145772595, + "grad_norm": 0.15552006661891937, + "learning_rate": 1e-06, + "loss": -0.031, + "num_tokens": 394233912.0, + "reward": 0.6395089626312256, + "reward_std": 0.15477727353572845, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111123085022, + "step": 651 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2422.0, + "completions/mean_length": 933.0078735351562, + "completions/mean_terminated_length": 579.820068359375, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 6.093294460641399, + "grad_norm": 0.15305151045322418, + "learning_rate": 1e-06, + "loss": -0.0353, + "num_tokens": 394789423.0, + "reward": 0.6741071939468384, + "reward_std": 0.14733406901359558, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 652 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3308.0, + "completions/mean_length": 905.1395263671875, + "completions/mean_terminated_length": 617.8843994140625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.1026239067055394, + "grad_norm": 0.15112105011940002, + "learning_rate": 1e-06, + "loss": -0.0407, + "num_tokens": 395388116.0, + "reward": 0.6417410969734192, + "reward_std": 0.16108711063861847, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 653 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3642.0, + "completions/mean_length": 1040.6239013671875, + "completions/mean_terminated_length": 669.6958618164062, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 6.111953352769679, + "grad_norm": 0.13102900981903076, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 396020419.0, + "reward": 0.625, + "reward_std": 0.15176916122436523, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 654 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3262.0, + "completions/mean_length": 975.3438110351562, + "completions/mean_terminated_length": 565.5606079101562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 6.121282798833819, + "grad_norm": 0.15055210888385773, + "learning_rate": 1e-06, + "loss": -0.0558, + "num_tokens": 396550487.0, + "reward": 0.6473214626312256, + "reward_std": 0.16112099587917328, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 655 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2855.0, + "completions/mean_length": 865.3549194335938, + "completions/mean_terminated_length": 574.5182495117188, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 6.130612244897959, + "grad_norm": 0.14336565136909485, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 397099973.0, + "reward": 0.7220982313156128, + "reward_std": 0.13493286073207855, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 656 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3133.0, + "completions/mean_length": 870.6897583007812, + "completions/mean_terminated_length": 618.4091796875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 6.139941690962099, + "grad_norm": 0.14243707060813904, + "learning_rate": 1e-06, + "loss": -0.0725, + "num_tokens": 397698871.0, + "reward": 0.6863839626312256, + "reward_std": 0.1721295863389969, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 657 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3554.0, + "completions/mean_length": 1019.55810546875, + "completions/mean_terminated_length": 584.5452270507812, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 6.149271137026239, + "grad_norm": 0.1588740199804306, + "learning_rate": 1e-06, + "loss": -0.0443, + "num_tokens": 398251547.0, + "reward": 0.629464328289032, + "reward_std": 0.15495699644088745, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179546356201, + "step": 658 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 866.716552734375, + "completions/mean_terminated_length": 580.2794799804688, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 6.158600583090379, + "grad_norm": 0.14313143491744995, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 398814189.0, + "reward": 0.7187500596046448, + "reward_std": 0.14158332347869873, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 659 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3080.0, + "completions/mean_length": 966.9006958007812, + "completions/mean_terminated_length": 595.7839965820312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 6.167930029154519, + "grad_norm": 0.13183774054050446, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 399377956.0, + "reward": 0.6495535969734192, + "reward_std": 0.1441013514995575, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 660 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3293.0, + "completions/mean_length": 1092.44873046875, + "completions/mean_terminated_length": 667.74267578125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.1772594752186585, + "grad_norm": 0.16232049465179443, + "learning_rate": 1e-06, + "loss": -0.0354, + "num_tokens": 400004414.0, + "reward": 0.5491071939468384, + "reward_std": 0.1810370534658432, + "rewards/verify_math_reward/mean": 0.5491071343421936, + "rewards/verify_math_reward/std": 0.49786055088043213, + "step": 661 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2978.0, + "completions/mean_length": 841.4174194335938, + "completions/mean_terminated_length": 607.8348999023438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.186588921282799, + "grad_norm": 0.12299291789531708, + "learning_rate": 1e-06, + "loss": -0.0285, + "num_tokens": 400595492.0, + "reward": 0.6808035969734192, + "reward_std": 0.14867813885211945, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 662 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3307.0, + "completions/mean_length": 972.4085083007812, + "completions/mean_terminated_length": 593.1990356445312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 6.1959183673469385, + "grad_norm": 0.1376301348209381, + "learning_rate": 1e-06, + "loss": -0.0653, + "num_tokens": 401162874.0, + "reward": 0.684151828289032, + "reward_std": 0.14545612037181854, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 663 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4010.0, + "completions/mean_length": 920.7176513671875, + "completions/mean_terminated_length": 592.2401123046875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 6.205247813411079, + "grad_norm": 0.135666623711586, + "learning_rate": 1e-06, + "loss": -0.0419, + "num_tokens": 401724901.0, + "reward": 0.6227678656578064, + "reward_std": 0.13587717711925507, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 664 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2843.0, + "completions/mean_length": 1083.359375, + "completions/mean_terminated_length": 648.5848999023438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 6.214577259475218, + "grad_norm": 0.13558167219161987, + "learning_rate": 1e-06, + "loss": -0.0453, + "num_tokens": 402331623.0, + "reward": 0.5502232313156128, + "reward_std": 0.15548720955848694, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 665 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4082.0, + "completions/mean_length": 913.6574096679688, + "completions/mean_terminated_length": 536.2259521484375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 6.223906705539359, + "grad_norm": 0.1738593429327011, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 402839516.0, + "reward": 0.6707589626312256, + "reward_std": 0.14436279237270355, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 666 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3839.0, + "completions/mean_length": 1033.844970703125, + "completions/mean_terminated_length": 653.476806640625, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 6.233236151603498, + "grad_norm": 0.14586031436920166, + "learning_rate": 1e-06, + "loss": -0.0617, + "num_tokens": 403442457.0, + "reward": 0.6517857313156128, + "reward_std": 0.1676594614982605, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 667 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3068.0, + "completions/mean_length": 1085.3717041015625, + "completions/mean_terminated_length": 655.2818603515625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.242565597667639, + "grad_norm": 0.14115239679813385, + "learning_rate": 1e-06, + "loss": -0.0776, + "num_tokens": 404051982.0, + "reward": 0.6004464626312256, + "reward_std": 0.164730966091156, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 668 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3320.0, + "completions/mean_length": 964.1250610351562, + "completions/mean_terminated_length": 610.0869750976562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.251895043731778, + "grad_norm": 0.1455235630273819, + "learning_rate": 1e-06, + "loss": -0.0387, + "num_tokens": 404634454.0, + "reward": 0.5926339626312256, + "reward_std": 0.1539819985628128, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 669 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3266.0, + "completions/mean_length": 918.7098388671875, + "completions/mean_terminated_length": 563.925537109375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 6.261224489795918, + "grad_norm": 0.13864287734031677, + "learning_rate": 1e-06, + "loss": -0.0369, + "num_tokens": 405179210.0, + "reward": 0.6875000596046448, + "reward_std": 0.12129746377468109, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 670 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2980.0, + "completions/mean_length": 953.2154541015625, + "completions/mean_terminated_length": 580.4756469726562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 6.270553935860058, + "grad_norm": 0.16936975717544556, + "learning_rate": 1e-06, + "loss": -0.0758, + "num_tokens": 405733211.0, + "reward": 0.6651785969734192, + "reward_std": 0.1598842889070511, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 671 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3677.0, + "completions/mean_length": 989.482177734375, + "completions/mean_terminated_length": 629.6986083984375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 6.279883381924198, + "grad_norm": 0.13950200378894806, + "learning_rate": 1e-06, + "loss": -0.0703, + "num_tokens": 406333163.0, + "reward": 0.6651785969734192, + "reward_std": 0.14707191288471222, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 672 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2629.0, + "completions/mean_length": 1002.18310546875, + "completions/mean_terminated_length": 600.3379516601562, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 6.289212827988338, + "grad_norm": 0.1564752608537674, + "learning_rate": 1e-06, + "loss": -0.0599, + "num_tokens": 406908959.0, + "reward": 0.6305803656578064, + "reward_std": 0.16938655078411102, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 673 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3836.0, + "completions/mean_length": 944.0279541015625, + "completions/mean_terminated_length": 639.2472534179688, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 6.298542274052478, + "grad_norm": 0.14450131356716156, + "learning_rate": 1e-06, + "loss": -0.0349, + "num_tokens": 407518064.0, + "reward": 0.6595982313156128, + "reward_std": 0.171762615442276, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3451.0, + "completions/mean_length": 1000.7801513671875, + "completions/mean_terminated_length": 633.681640625, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 6.307871720116618, + "grad_norm": 0.1488829255104065, + "learning_rate": 1e-06, + "loss": -0.0725, + "num_tokens": 408101387.0, + "reward": 0.660714328289032, + "reward_std": 0.15856976807117462, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 675 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3398.0, + "completions/mean_length": 1001.6875610351562, + "completions/mean_terminated_length": 643.3175659179688, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 6.317201166180758, + "grad_norm": 0.14910122752189636, + "learning_rate": 1e-06, + "loss": -0.0838, + "num_tokens": 408705075.0, + "reward": 0.6395089626312256, + "reward_std": 0.1907336562871933, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 676 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3551.0, + "completions/mean_length": 1056.5179443359375, + "completions/mean_terminated_length": 586.4948120117188, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 6.326530612244898, + "grad_norm": 0.1475907415151596, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 409251555.0, + "reward": 0.625, + "reward_std": 0.14691968262195587, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 677 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2654.0, + "completions/mean_length": 999.1428833007812, + "completions/mean_terminated_length": 579.1635131835938, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 6.335860058309038, + "grad_norm": 0.15115858614444733, + "learning_rate": 1e-06, + "loss": -0.0375, + "num_tokens": 409811563.0, + "reward": 0.6328125, + "reward_std": 0.1451544314622879, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 678 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4023.0, + "completions/mean_length": 890.2522583007812, + "completions/mean_terminated_length": 580.271728515625, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 6.345189504373177, + "grad_norm": 0.1378905177116394, + "learning_rate": 1e-06, + "loss": -0.0547, + "num_tokens": 410374141.0, + "reward": 0.6819196939468384, + "reward_std": 0.15826597809791565, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 679 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3437.0, + "completions/mean_length": 984.8359985351562, + "completions/mean_terminated_length": 654.5148315429688, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.354518950437318, + "grad_norm": 0.13695885241031647, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 410990162.0, + "reward": 0.6495535969734192, + "reward_std": 0.16825932264328003, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.47737622261047363, + "step": 680 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2914.0, + "completions/mean_length": 1057.704345703125, + "completions/mean_terminated_length": 636.8983154296875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 6.363848396501457, + "grad_norm": 0.1458936631679535, + "learning_rate": 1e-06, + "loss": -0.0445, + "num_tokens": 411582201.0, + "reward": 0.606026828289032, + "reward_std": 0.16619591414928436, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 681 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3007.0, + "completions/mean_length": 1068.0670166015625, + "completions/mean_terminated_length": 635.5050659179688, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 6.373177842565598, + "grad_norm": 0.15132257342338562, + "learning_rate": 1e-06, + "loss": -0.025, + "num_tokens": 412175781.0, + "reward": 0.6205357313156128, + "reward_std": 0.16258637607097626, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2926.0, + "completions/mean_length": 947.7422485351562, + "completions/mean_terminated_length": 591.8521728515625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 6.382507288629737, + "grad_norm": 0.1570195108652115, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 412744822.0, + "reward": 0.715401828289032, + "reward_std": 0.15390783548355103, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 683 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3387.0, + "completions/mean_length": 1046.8326416015625, + "completions/mean_terminated_length": 637.7037963867188, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 6.391836734693878, + "grad_norm": 0.1534552276134491, + "learning_rate": 1e-06, + "loss": -0.0673, + "num_tokens": 413338928.0, + "reward": 0.5870535969734192, + "reward_std": 0.15932045876979828, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 684 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3770.0, + "completions/mean_length": 948.1328735351562, + "completions/mean_terminated_length": 557.117919921875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.401166180758017, + "grad_norm": 0.1437825709581375, + "learning_rate": 1e-06, + "loss": -0.0601, + "num_tokens": 413874599.0, + "reward": 0.7198660969734192, + "reward_std": 0.1479741632938385, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 685 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3242.0, + "completions/mean_length": 1084.251220703125, + "completions/mean_terminated_length": 591.4194946289062, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 6.410495626822158, + "grad_norm": 0.14510203897953033, + "learning_rate": 1e-06, + "loss": -0.0607, + "num_tokens": 414417496.0, + "reward": 0.6272321939468384, + "reward_std": 0.14496827125549316, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 686 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3217.0, + "completions/mean_length": 971.7578735351562, + "completions/mean_terminated_length": 652.8007202148438, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.419825072886297, + "grad_norm": 0.14245858788490295, + "learning_rate": 1e-06, + "loss": -0.0302, + "num_tokens": 415032655.0, + "reward": 0.6718750596046448, + "reward_std": 0.16119766235351562, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 687 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3769.0, + "completions/mean_length": 1006.2567138671875, + "completions/mean_terminated_length": 635.4874877929688, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 6.429154518950437, + "grad_norm": 0.1387123465538025, + "learning_rate": 1e-06, + "loss": -0.0484, + "num_tokens": 415628597.0, + "reward": 0.629464328289032, + "reward_std": 0.14173558354377747, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 688 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3964.0, + "completions/mean_length": 1053.0379638671875, + "completions/mean_terminated_length": 622.7592163085938, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 6.438483965014577, + "grad_norm": 0.14167840778827667, + "learning_rate": 1e-06, + "loss": -0.0463, + "num_tokens": 416208327.0, + "reward": 0.6071428656578064, + "reward_std": 0.14515121281147003, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 689 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3639.0, + "completions/mean_length": 1001.5178833007812, + "completions/mean_terminated_length": 625.84228515625, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 6.447813411078717, + "grad_norm": 0.15804407000541687, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 416793207.0, + "reward": 0.668526828289032, + "reward_std": 0.170363649725914, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 690 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3890.0, + "completions/mean_length": 1071.2366943359375, + "completions/mean_terminated_length": 647.9236450195312, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 6.457142857142857, + "grad_norm": 0.1627172827720642, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 417390947.0, + "reward": 0.629464328289032, + "reward_std": 0.19772042334079742, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 691 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3897.0, + "completions/mean_length": 859.9263916015625, + "completions/mean_terminated_length": 581.4278564453125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 6.466472303206997, + "grad_norm": 0.13965976238250732, + "learning_rate": 1e-06, + "loss": -0.0327, + "num_tokens": 417960913.0, + "reward": 0.6729910969734192, + "reward_std": 0.13801473379135132, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 692 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3663.0, + "completions/mean_length": 971.9553833007812, + "completions/mean_terminated_length": 614.4776000976562, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 6.475801749271137, + "grad_norm": 0.15344823896884918, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 418537345.0, + "reward": 0.6729910969734192, + "reward_std": 0.17644678056240082, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 693 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3841.0, + "completions/mean_length": 921.2020263671875, + "completions/mean_terminated_length": 588.4549560546875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 6.485131195335277, + "grad_norm": 0.16213160753250122, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 419106150.0, + "reward": 0.6651785969734192, + "reward_std": 0.13459837436676025, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 694 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3944.0, + "completions/mean_length": 1001.28466796875, + "completions/mean_terminated_length": 603.7266845703125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 6.494460641399417, + "grad_norm": 0.1728551685810089, + "learning_rate": 1e-06, + "loss": -0.0817, + "num_tokens": 419672869.0, + "reward": 0.6361607313156128, + "reward_std": 0.172694131731987, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 695 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2525.0, + "completions/mean_length": 1050.458740234375, + "completions/mean_terminated_length": 606.4795532226562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 6.503790087463557, + "grad_norm": 0.16064096987247467, + "learning_rate": 1e-06, + "loss": -0.0312, + "num_tokens": 420240728.0, + "reward": 0.6272321939468384, + "reward_std": 0.15056565403938293, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 696 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4025.0, + "completions/mean_length": 1152.38623046875, + "completions/mean_terminated_length": 625.6342163085938, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 6.513119533527696, + "grad_norm": 0.17802776396274567, + "learning_rate": 1e-06, + "loss": -0.0692, + "num_tokens": 420812882.0, + "reward": 0.6205357313156128, + "reward_std": 0.19523699581623077, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3379.0, + "completions/mean_length": 965.9420166015625, + "completions/mean_terminated_length": 629.3349609375, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 6.522448979591837, + "grad_norm": 0.1359633207321167, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 421421766.0, + "reward": 0.6584821939468384, + "reward_std": 0.1796458661556244, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 698 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2912.0, + "completions/mean_length": 981.7522583007812, + "completions/mean_terminated_length": 559.4144287109375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 6.531778425655976, + "grad_norm": 0.13205242156982422, + "learning_rate": 1e-06, + "loss": -0.054, + "num_tokens": 421952872.0, + "reward": 0.6964285969734192, + "reward_std": 0.12749215960502625, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 699 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2405.0, + "completions/mean_length": 868.075927734375, + "completions/mean_terminated_length": 568.9024658203125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.541107871720117, + "grad_norm": 0.12569795548915863, + "learning_rate": 1e-06, + "loss": -0.0238, + "num_tokens": 422499556.0, + "reward": 0.6741071939468384, + "reward_std": 0.12595123052597046, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.46896928548812866, + "step": 700 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3071.0, + "completions/mean_length": 1055.8717041015625, + "completions/mean_terminated_length": 612.6815795898438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.550437317784256, + "grad_norm": 0.14089711010456085, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 423063809.0, + "reward": 0.6395089626312256, + "reward_std": 0.13827574253082275, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111123085022, + "step": 701 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3540.0, + "completions/mean_length": 937.9442138671875, + "completions/mean_terminated_length": 589.6580200195312, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.559766763848397, + "grad_norm": 0.13796377182006836, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 423623967.0, + "reward": 0.668526828289032, + "reward_std": 0.1167111024260521, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 702 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.15625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3082.0, + "completions/mean_length": 1203.165283203125, + "completions/mean_terminated_length": 667.4550170898438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 6.569096209912536, + "grad_norm": 0.17245352268218994, + "learning_rate": 1e-06, + "loss": -0.075, + "num_tokens": 424204907.0, + "reward": 0.6049107313156128, + "reward_std": 0.1447407603263855, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 703 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2938.0, + "completions/mean_length": 864.9241333007812, + "completions/mean_terminated_length": 582.5971069335938, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 6.578425655976677, + "grad_norm": 0.15715719759464264, + "learning_rate": 1e-06, + "loss": -0.0316, + "num_tokens": 424770519.0, + "reward": 0.6897321939468384, + "reward_std": 0.14887388050556183, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 704 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2681.0, + "completions/mean_length": 998.1808471679688, + "completions/mean_terminated_length": 652.2704467773438, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 6.587755102040816, + "grad_norm": 0.15077020227909088, + "learning_rate": 1e-06, + "loss": -0.0458, + "num_tokens": 425377409.0, + "reward": 0.6930803656578064, + "reward_std": 0.17261813580989838, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 705 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4042.0, + "completions/mean_length": 1056.735595703125, + "completions/mean_terminated_length": 640.1865234375, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 6.597084548104956, + "grad_norm": 0.15572671592235565, + "learning_rate": 1e-06, + "loss": -0.042, + "num_tokens": 425974932.0, + "reward": 0.6183035969734192, + "reward_std": 0.16217085719108582, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 706 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2756.0, + "completions/mean_length": 883.94091796875, + "completions/mean_terminated_length": 586.23779296875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 6.606413994169096, + "grad_norm": 0.15391968190670013, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 426547039.0, + "reward": 0.6462053656578064, + "reward_std": 0.14060692489147186, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 707 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3692.0, + "completions/mean_length": 897.9967041015625, + "completions/mean_terminated_length": 580.1582641601562, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 6.615743440233236, + "grad_norm": 0.14751707017421722, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 427109996.0, + "reward": 0.6707589626312256, + "reward_std": 0.19320568442344666, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 708 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3979.0, + "completions/mean_length": 1117.0223388671875, + "completions/mean_terminated_length": 634.04931640625, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 6.625072886297376, + "grad_norm": 0.16707062721252441, + "learning_rate": 1e-06, + "loss": -0.0811, + "num_tokens": 427688144.0, + "reward": 0.6171875, + "reward_std": 0.16394254565238953, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 709 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3331.0, + "completions/mean_length": 1118.5882568359375, + "completions/mean_terminated_length": 590.4007568359375, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.634402332361516, + "grad_norm": 0.1403144896030426, + "learning_rate": 1e-06, + "loss": -0.0588, + "num_tokens": 428237231.0, + "reward": 0.6439732313156128, + "reward_std": 0.13639962673187256, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 710 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3550.0, + "completions/mean_length": 981.6160888671875, + "completions/mean_terminated_length": 633.8560791015625, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 6.643731778425656, + "grad_norm": 0.12499608844518661, + "learning_rate": 1e-06, + "loss": -0.0318, + "num_tokens": 428831975.0, + "reward": 0.6272321939468384, + "reward_std": 0.12531296908855438, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 711 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3603.0, + "completions/mean_length": 1115.3226318359375, + "completions/mean_terminated_length": 586.5558471679688, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 6.653061224489796, + "grad_norm": 0.1809120774269104, + "learning_rate": 1e-06, + "loss": -0.0601, + "num_tokens": 429374712.0, + "reward": 0.613839328289032, + "reward_std": 0.18975867331027985, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 712 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2627.0, + "completions/mean_length": 1017.966552734375, + "completions/mean_terminated_length": 609.3780517578125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 6.662390670553936, + "grad_norm": 0.13838516175746918, + "learning_rate": 1e-06, + "loss": -0.0476, + "num_tokens": 429947746.0, + "reward": 0.6506696939468384, + "reward_std": 0.1327946037054062, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 713 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3475.0, + "completions/mean_length": 1117.8348388671875, + "completions/mean_terminated_length": 625.9921875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 6.671720116618076, + "grad_norm": 0.15369167923927307, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 430528246.0, + "reward": 0.6026785969734192, + "reward_std": 0.12092021107673645, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961684107780457, + "step": 714 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2491.0, + "completions/mean_length": 1041.997802734375, + "completions/mean_terminated_length": 619.0164794921875, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "epoch": 6.681049562682215, + "grad_norm": 0.17129403352737427, + "learning_rate": 1e-06, + "loss": -0.055, + "num_tokens": 431106148.0, + "reward": 0.6383928656578064, + "reward_std": 0.16551373898983002, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 715 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2946.0, + "completions/mean_length": 1083.171875, + "completions/mean_terminated_length": 635.1102905273438, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 6.690379008746356, + "grad_norm": 0.15595877170562744, + "learning_rate": 1e-06, + "loss": -0.0716, + "num_tokens": 431692286.0, + "reward": 0.590401828289032, + "reward_std": 0.16311517357826233, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 716 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 1120.594970703125, + "completions/mean_terminated_length": 656.0477905273438, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 6.699708454810495, + "grad_norm": 0.12979260087013245, + "learning_rate": 1e-06, + "loss": -0.0696, + "num_tokens": 432289731.0, + "reward": 0.6004464626312256, + "reward_std": 0.1328631341457367, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 717 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2711.0, + "completions/mean_length": 1017.5982666015625, + "completions/mean_terminated_length": 604.5468139648438, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 6.709037900874636, + "grad_norm": 0.17066724598407745, + "learning_rate": 1e-06, + "loss": -0.0623, + "num_tokens": 432851643.0, + "reward": 0.6718750596046448, + "reward_std": 0.19501832127571106, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 718 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3789.0, + "completions/mean_length": 976.0022583007812, + "completions/mean_terminated_length": 618.987548828125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.718367346938775, + "grad_norm": 0.14995244145393372, + "learning_rate": 1e-06, + "loss": -0.0536, + "num_tokens": 433448189.0, + "reward": 0.6395089626312256, + "reward_std": 0.15893998742103577, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 719 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2896.0, + "completions/mean_length": 1039.6328125, + "completions/mean_terminated_length": 607.4586181640625, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 6.727696793002916, + "grad_norm": 0.15274804830551147, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 434013964.0, + "reward": 0.6417410969734192, + "reward_std": 0.14012089371681213, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 720 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2835.0, + "completions/mean_length": 976.6607666015625, + "completions/mean_terminated_length": 615.3922729492188, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 6.737026239067055, + "grad_norm": 0.1447124183177948, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 434589044.0, + "reward": 0.6696428656578064, + "reward_std": 0.14518442749977112, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 721 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2406.0, + "completions/mean_length": 1010.3303833007812, + "completions/mean_terminated_length": 574.0127563476562, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 6.746355685131196, + "grad_norm": 0.144724503159523, + "learning_rate": 1e-06, + "loss": -0.0594, + "num_tokens": 435142420.0, + "reward": 0.6551339626312256, + "reward_std": 0.143646702170372, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 722 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2387.0, + "completions/mean_length": 1048.5592041015625, + "completions/mean_terminated_length": 590.8562622070312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 6.755685131195335, + "grad_norm": 0.14169110357761383, + "learning_rate": 1e-06, + "loss": -0.0401, + "num_tokens": 435690633.0, + "reward": 0.6261160969734192, + "reward_std": 0.1201702207326889, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 723 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2746.0, + "completions/mean_length": 954.693115234375, + "completions/mean_terminated_length": 590.8804321289062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 6.765014577259475, + "grad_norm": 0.15913435816764832, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 436246142.0, + "reward": 0.6774553656578064, + "reward_std": 0.15597687661647797, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 724 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3877.0, + "completions/mean_length": 1101.71435546875, + "completions/mean_terminated_length": 643.1300048828125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 6.774344023323615, + "grad_norm": 0.1231972724199295, + "learning_rate": 1e-06, + "loss": -0.0695, + "num_tokens": 436831502.0, + "reward": 0.6160714626312256, + "reward_std": 0.137190580368042, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 725 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3839.0, + "completions/mean_length": 1031.219970703125, + "completions/mean_terminated_length": 615.5906372070312, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.783673469387755, + "grad_norm": 0.16800829768180847, + "learning_rate": 1e-06, + "loss": -0.062, + "num_tokens": 437404651.0, + "reward": 0.6718750596046448, + "reward_std": 0.16950741410255432, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 726 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3820.0, + "completions/mean_length": 1120.3348388671875, + "completions/mean_terminated_length": 628.9050903320312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 6.793002915451895, + "grad_norm": 0.1576811969280243, + "learning_rate": 1e-06, + "loss": -0.0834, + "num_tokens": 437978679.0, + "reward": 0.5881696939468384, + "reward_std": 0.17866946756839752, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924396276473999, + "step": 727 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4070.0, + "completions/mean_length": 1077.74560546875, + "completions/mean_terminated_length": 650.9605102539062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 6.802332361516035, + "grad_norm": 0.15597330033779144, + "learning_rate": 1e-06, + "loss": -0.041, + "num_tokens": 438573691.0, + "reward": 0.582589328289032, + "reward_std": 0.16897287964820862, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.493407279253006, + "step": 728 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4094.0, + "completions/mean_length": 1027.2679443359375, + "completions/mean_terminated_length": 619.9140625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 6.811661807580175, + "grad_norm": 0.16009485721588135, + "learning_rate": 1e-06, + "loss": -0.0631, + "num_tokens": 439145419.0, + "reward": 0.6651785969734192, + "reward_std": 0.16799716651439667, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 729 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2769.0, + "completions/mean_length": 1080.5023193359375, + "completions/mean_terminated_length": 645.3154296875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.820991253644315, + "grad_norm": 0.1696934551000595, + "learning_rate": 1e-06, + "loss": -0.0607, + "num_tokens": 439727237.0, + "reward": 0.5993303656578064, + "reward_std": 0.18344563245773315, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 730 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1629464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3934.0, + "completions/mean_length": 1211.430908203125, + "completions/mean_terminated_length": 649.9013061523438, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 6.830320699708455, + "grad_norm": 0.15010182559490204, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 440303463.0, + "reward": 0.6082589626312256, + "reward_std": 0.14587052166461945, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 731 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3932.0, + "completions/mean_length": 1067.74560546875, + "completions/mean_terminated_length": 635.1377563476562, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "epoch": 6.839650145772595, + "grad_norm": 0.15665987133979797, + "learning_rate": 1e-06, + "loss": -0.0687, + "num_tokens": 440891995.0, + "reward": 0.6506696939468384, + "reward_std": 0.1796797662973404, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 732 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4080.0, + "completions/mean_length": 1021.5435791015625, + "completions/mean_terminated_length": 635.3052978515625, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 6.848979591836734, + "grad_norm": 0.14565272629261017, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 441498786.0, + "reward": 0.6484375, + "reward_std": 0.16078400611877441, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 733 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3868.0, + "completions/mean_length": 1053.6998291015625, + "completions/mean_terminated_length": 610.193115234375, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 6.858309037900875, + "grad_norm": 0.12248275429010391, + "learning_rate": 1e-06, + "loss": -0.0548, + "num_tokens": 442064669.0, + "reward": 0.6272321939468384, + "reward_std": 0.12467243522405624, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 734 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3630.0, + "completions/mean_length": 1226.4364013671875, + "completions/mean_terminated_length": 630.8665771484375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 6.867638483965014, + "grad_norm": 0.14588762819766998, + "learning_rate": 1e-06, + "loss": -0.0951, + "num_tokens": 442632556.0, + "reward": 0.6127232313156128, + "reward_std": 0.14331039786338806, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2815.0, + "completions/mean_length": 1014.5558471679688, + "completions/mean_terminated_length": 609.9216918945312, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 6.876967930029155, + "grad_norm": 0.15247777104377747, + "learning_rate": 1e-06, + "loss": -0.0661, + "num_tokens": 443215750.0, + "reward": 0.6741071939468384, + "reward_std": 0.1789322942495346, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 736 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2444.0, + "completions/mean_length": 921.61279296875, + "completions/mean_terminated_length": 580.2385864257812, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 6.886297376093294, + "grad_norm": 0.146469846367836, + "learning_rate": 1e-06, + "loss": -0.0492, + "num_tokens": 443784179.0, + "reward": 0.6875000596046448, + "reward_std": 0.13166987895965576, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 737 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3031.0, + "completions/mean_length": 1093.0546875, + "completions/mean_terminated_length": 672.795166015625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 6.895626822157435, + "grad_norm": 0.14227406680583954, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 444407068.0, + "reward": 0.613839328289032, + "reward_std": 0.14943519234657288, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 738 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3521.0, + "completions/mean_length": 967.3917846679688, + "completions/mean_terminated_length": 583.1766967773438, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 6.904956268221574, + "grad_norm": 0.161316379904747, + "learning_rate": 1e-06, + "loss": -0.034, + "num_tokens": 444953627.0, + "reward": 0.6584821939468384, + "reward_std": 0.14173416793346405, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 739 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2674.0, + "completions/mean_length": 1104.263427734375, + "completions/mean_terminated_length": 641.6237182617188, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.914285714285715, + "grad_norm": 0.1274370402097702, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 445545927.0, + "reward": 0.5948660969734192, + "reward_std": 0.12174323946237564, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 740 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2638.0, + "completions/mean_length": 977.2879638671875, + "completions/mean_terminated_length": 567.7600708007812, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.923615160349854, + "grad_norm": 0.16711221635341644, + "learning_rate": 1e-06, + "loss": -0.0888, + "num_tokens": 446076081.0, + "reward": 0.6830357313156128, + "reward_std": 0.1892266422510147, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 741 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3409.0, + "completions/mean_length": 1017.7935791015625, + "completions/mean_terminated_length": 609.1820678710938, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 6.932944606413994, + "grad_norm": 0.14523564279079437, + "learning_rate": 1e-06, + "loss": -0.0332, + "num_tokens": 446642448.0, + "reward": 0.637276828289032, + "reward_std": 0.13203758001327515, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 742 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3425.0, + "completions/mean_length": 903.7120971679688, + "completions/mean_terminated_length": 573.475341796875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 6.942274052478134, + "grad_norm": 0.12424454092979431, + "learning_rate": 1e-06, + "loss": -0.0435, + "num_tokens": 447203078.0, + "reward": 0.660714328289032, + "reward_std": 0.11599572002887726, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1741071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2782.0, + "completions/mean_length": 1237.685302734375, + "completions/mean_terminated_length": 635.1216430664062, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 6.9516034985422746, + "grad_norm": 0.1522192507982254, + "learning_rate": 1e-06, + "loss": -0.0862, + "num_tokens": 447759372.0, + "reward": 0.5714285969734192, + "reward_std": 0.14714859426021576, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 744 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3506.0, + "completions/mean_length": 1002.8092041015625, + "completions/mean_terminated_length": 635.9512939453125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 6.960932944606414, + "grad_norm": 0.16338354349136353, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 448359329.0, + "reward": 0.6540178656578064, + "reward_std": 0.16468890011310577, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 745 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3056.0, + "completions/mean_length": 1093.83154296875, + "completions/mean_terminated_length": 607.0985717773438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.970262390670554, + "grad_norm": 0.14463958144187927, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 448918322.0, + "reward": 0.6350446939468384, + "reward_std": 0.14920946955680847, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 746 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3299.0, + "completions/mean_length": 1028.368408203125, + "completions/mean_terminated_length": 590.1351928710938, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 6.979591836734694, + "grad_norm": 0.16860869526863098, + "learning_rate": 1e-06, + "loss": -0.0652, + "num_tokens": 449467500.0, + "reward": 0.6517857313156128, + "reward_std": 0.20166808366775513, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 747 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3624.0, + "completions/mean_length": 1013.8582763671875, + "completions/mean_terminated_length": 600.3050537109375, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 6.988921282798834, + "grad_norm": 0.14585554599761963, + "learning_rate": 1e-06, + "loss": -0.0649, + "num_tokens": 450022117.0, + "reward": 0.6941964626312256, + "reward_std": 0.14917558431625366, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.46100425720214844, + "step": 748 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.15340909090909094, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2098.0, + "completions/mean_length": 1125.6534423828125, + "completions/mean_terminated_length": 587.4027099609375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 6.998250728862974, + "grad_norm": 0.1666969656944275, + "learning_rate": 1e-06, + "loss": -0.1032, + "num_tokens": 450573106.0, + "reward": 0.621651828289032, + "reward_std": 0.17333422601222992, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.485245943069458, + "step": 749 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3491.0, + "completions/mean_length": 1101.943115234375, + "completions/mean_terminated_length": 598.37939453125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 7.0093294460641395, + "grad_norm": 0.1618053913116455, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 451127951.0, + "reward": 0.6205357313156128, + "reward_std": 0.16570919752120972, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 750 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2904.0, + "completions/mean_length": 990.2734985351562, + "completions/mean_terminated_length": 586.8814697265625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 7.01865889212828, + "grad_norm": 0.1463259607553482, + "learning_rate": 1e-06, + "loss": -0.0769, + "num_tokens": 451684532.0, + "reward": 0.6741071939468384, + "reward_std": 0.17277081310749054, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 751 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2101.0, + "completions/mean_length": 1067.78466796875, + "completions/mean_terminated_length": 599.5038452148438, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 7.0279883381924195, + "grad_norm": 0.17970505356788635, + "learning_rate": 1e-06, + "loss": -0.0691, + "num_tokens": 452234059.0, + "reward": 0.652901828289032, + "reward_std": 0.16720734536647797, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 752 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2992.0, + "completions/mean_length": 1139.446533203125, + "completions/mean_terminated_length": 642.1903686523438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.03731778425656, + "grad_norm": 0.13983535766601562, + "learning_rate": 1e-06, + "loss": -0.0671, + "num_tokens": 452812275.0, + "reward": 0.6350446939468384, + "reward_std": 0.15018658339977264, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 753 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2732.0, + "completions/mean_length": 1135.7366943359375, + "completions/mean_terminated_length": 573.561767578125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.0466472303206995, + "grad_norm": 0.1569301187992096, + "learning_rate": 1e-06, + "loss": -0.0821, + "num_tokens": 453331071.0, + "reward": 0.6462053656578064, + "reward_std": 0.14766854047775269, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 754 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3035.0, + "completions/mean_length": 1066.384033203125, + "completions/mean_terminated_length": 629.1596069335938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 7.05597667638484, + "grad_norm": 0.16864013671875, + "learning_rate": 1e-06, + "loss": -0.0377, + "num_tokens": 453924383.0, + "reward": 0.6227678656578064, + "reward_std": 0.17261996865272522, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 755 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3674.0, + "completions/mean_length": 1001.9922485351562, + "completions/mean_terminated_length": 652.2347412109375, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 7.0653061224489795, + "grad_norm": 0.1340419054031372, + "learning_rate": 1e-06, + "loss": -0.0543, + "num_tokens": 454538504.0, + "reward": 0.6540178656578064, + "reward_std": 0.15372055768966675, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2668.0, + "completions/mean_length": 938.55029296875, + "completions/mean_terminated_length": 572.8679809570312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.07463556851312, + "grad_norm": 0.13865657150745392, + "learning_rate": 1e-06, + "loss": -0.035, + "num_tokens": 455085757.0, + "reward": 0.6741071939468384, + "reward_std": 0.13136427104473114, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 757 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3534.0, + "completions/mean_length": 1068.599365234375, + "completions/mean_terminated_length": 586.87841796875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 7.0839650145772595, + "grad_norm": 0.1785573959350586, + "learning_rate": 1e-06, + "loss": -0.0782, + "num_tokens": 455627406.0, + "reward": 0.6473214626312256, + "reward_std": 0.17870336771011353, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 758 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2359.0, + "completions/mean_length": 1011.97216796875, + "completions/mean_terminated_length": 580.3651123046875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 7.093294460641399, + "grad_norm": 0.14673587679862976, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 456175773.0, + "reward": 0.6417410969734192, + "reward_std": 0.1665654182434082, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 759 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2938.0, + "completions/mean_length": 1018.4263916015625, + "completions/mean_terminated_length": 627.43896484375, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 7.1026239067055394, + "grad_norm": 0.1314193457365036, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 456766507.0, + "reward": 0.637276828289032, + "reward_std": 0.13005296885967255, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 760 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2992.0, + "completions/mean_length": 1100.9085693359375, + "completions/mean_terminated_length": 601.7265625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.111953352769679, + "grad_norm": 0.15485523641109467, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 457311361.0, + "reward": 0.6428571939468384, + "reward_std": 0.14271126687526703, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3791.0, + "completions/mean_length": 1136.161865234375, + "completions/mean_terminated_length": 656.2918090820312, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 7.121282798833819, + "grad_norm": 0.1420951932668686, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 457907706.0, + "reward": 0.6183035969734192, + "reward_std": 0.14774340391159058, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740602016449, + "step": 762 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3293.0, + "completions/mean_length": 1117.15966796875, + "completions/mean_terminated_length": 611.6123046875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.130612244897959, + "grad_norm": 0.17951548099517822, + "learning_rate": 1e-06, + "loss": -0.0696, + "num_tokens": 458470953.0, + "reward": 0.6171875, + "reward_std": 0.20053830742835999, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 763 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3280.0, + "completions/mean_length": 969.3047485351562, + "completions/mean_terminated_length": 633.059326171875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 7.139941690962099, + "grad_norm": 0.12861484289169312, + "learning_rate": 1e-06, + "loss": -0.0466, + "num_tokens": 459063298.0, + "reward": 0.6830357313156128, + "reward_std": 0.14045536518096924, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 764 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1819196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2980.0, + "completions/mean_length": 1254.390625, + "completions/mean_terminated_length": 622.4911499023438, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 7.149271137026239, + "grad_norm": 0.16698843240737915, + "learning_rate": 1e-06, + "loss": -0.082, + "num_tokens": 459619056.0, + "reward": 0.6183035969734192, + "reward_std": 0.1616862416267395, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2727.0, + "completions/mean_length": 1059.31591796875, + "completions/mean_terminated_length": 580.6654052734375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 7.158600583090379, + "grad_norm": 0.14558841288089752, + "learning_rate": 1e-06, + "loss": -0.0189, + "num_tokens": 460151443.0, + "reward": 0.668526828289032, + "reward_std": 0.12697336077690125, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 766 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2526.0, + "completions/mean_length": 1019.71435546875, + "completions/mean_terminated_length": 641.9248046875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.167930029154519, + "grad_norm": 0.143239825963974, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 460757083.0, + "reward": 0.6004464626312256, + "reward_std": 0.1641247719526291, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 767 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3296.0, + "completions/mean_length": 1001.1529541015625, + "completions/mean_terminated_length": 594.7588500976562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 7.1772594752186585, + "grad_norm": 0.16062027215957642, + "learning_rate": 1e-06, + "loss": -0.0539, + "num_tokens": 461314212.0, + "reward": 0.668526828289032, + "reward_std": 0.15402548015117645, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4030.0, + "completions/mean_length": 962.7076416015625, + "completions/mean_terminated_length": 634.3107299804688, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 7.186588921282799, + "grad_norm": 0.14385035634040833, + "learning_rate": 1e-06, + "loss": -0.0666, + "num_tokens": 461916678.0, + "reward": 0.6986607313156128, + "reward_std": 0.1645801067352295, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 769 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2688.0, + "completions/mean_length": 1010.232177734375, + "completions/mean_terminated_length": 605.0302734375, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 7.1959183673469385, + "grad_norm": 0.12803252041339874, + "learning_rate": 1e-06, + "loss": -0.0727, + "num_tokens": 462488382.0, + "reward": 0.6462053656578064, + "reward_std": 0.13034580647945404, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 770 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4007.0, + "completions/mean_length": 1106.634033203125, + "completions/mean_terminated_length": 653.2339477539062, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 7.205247813411079, + "grad_norm": 0.13519492745399475, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 463082030.0, + "reward": 0.6272321939468384, + "reward_std": 0.12549589574337006, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 771 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3068.0, + "completions/mean_length": 1142.110595703125, + "completions/mean_terminated_length": 618.0958862304688, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 7.214577259475218, + "grad_norm": 0.16670489311218262, + "learning_rate": 1e-06, + "loss": -0.0953, + "num_tokens": 463645793.0, + "reward": 0.6618303656578064, + "reward_std": 0.15788871049880981, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 772 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2454.0, + "completions/mean_length": 989.1641235351562, + "completions/mean_terminated_length": 545.330322265625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 7.223906705539359, + "grad_norm": 0.16454815864562988, + "learning_rate": 1e-06, + "loss": -0.05, + "num_tokens": 464168212.0, + "reward": 0.6194196939468384, + "reward_std": 0.1539064198732376, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 773 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3919.0, + "completions/mean_length": 913.5201416015625, + "completions/mean_terminated_length": 601.5122680664062, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.233236151603498, + "grad_norm": 0.14666207134723663, + "learning_rate": 1e-06, + "loss": -0.05, + "num_tokens": 464750862.0, + "reward": 0.6395089626312256, + "reward_std": 0.1650358885526657, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 774 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3768.0, + "completions/mean_length": 1094.6529541015625, + "completions/mean_terminated_length": 652.7131958007812, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 7.242565597667639, + "grad_norm": 0.14351041615009308, + "learning_rate": 1e-06, + "loss": -0.0703, + "num_tokens": 465351919.0, + "reward": 0.6551339626312256, + "reward_std": 0.16671767830848694, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 775 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3495.0, + "completions/mean_length": 1077.279052734375, + "completions/mean_terminated_length": 628.341064453125, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 7.251895043731778, + "grad_norm": 0.16652733087539673, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 465928009.0, + "reward": 0.6417410969734192, + "reward_std": 0.1670207679271698, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3992.0, + "completions/mean_length": 972.5301513671875, + "completions/mean_terminated_length": 575.7119140625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 7.261224489795918, + "grad_norm": 0.13777396082878113, + "learning_rate": 1e-06, + "loss": -0.0715, + "num_tokens": 466467964.0, + "reward": 0.699776828289032, + "reward_std": 0.14902400970458984, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 777 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3125.0, + "completions/mean_length": 970.71435546875, + "completions/mean_terminated_length": 626.0421142578125, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 7.270553935860058, + "grad_norm": 0.14167393743991852, + "learning_rate": 1e-06, + "loss": -0.0777, + "num_tokens": 467058652.0, + "reward": 0.65625, + "reward_std": 0.1788976937532425, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 778 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2588.0, + "completions/mean_length": 1097.6373291015625, + "completions/mean_terminated_length": 682.3621215820312, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 7.279883381924198, + "grad_norm": 0.16399122774600983, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 467688207.0, + "reward": 0.590401828289032, + "reward_std": 0.2153034806251526, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 779 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3370.0, + "completions/mean_length": 976.3660888671875, + "completions/mean_terminated_length": 606.3720092773438, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.289212827988338, + "grad_norm": 0.14657285809516907, + "learning_rate": 1e-06, + "loss": -0.0356, + "num_tokens": 468262135.0, + "reward": 0.6473214626312256, + "reward_std": 0.14725738763809204, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 780 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2954.0, + "completions/mean_length": 920.0870971679688, + "completions/mean_terminated_length": 604.4441528320312, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 7.298542274052478, + "grad_norm": 0.1364431381225586, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 468837061.0, + "reward": 0.6897321939468384, + "reward_std": 0.13293729722499847, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 781 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2792.0, + "completions/mean_length": 1103.671875, + "completions/mean_terminated_length": 618.5343627929688, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.307871720116618, + "grad_norm": 0.16352158784866333, + "learning_rate": 1e-06, + "loss": -0.0716, + "num_tokens": 469407359.0, + "reward": 0.6160714626312256, + "reward_std": 0.1749500334262848, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 782 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4046.0, + "completions/mean_length": 986.9967041015625, + "completions/mean_terminated_length": 639.8374633789062, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.317201166180758, + "grad_norm": 0.14241141080856323, + "learning_rate": 1e-06, + "loss": -0.0414, + "num_tokens": 470008412.0, + "reward": 0.6674107313156128, + "reward_std": 0.13113674521446228, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 783 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 936.0357666015625, + "completions/mean_terminated_length": 574.44775390625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 7.326530612244898, + "grad_norm": 0.16738606989383698, + "learning_rate": 1e-06, + "loss": -0.0721, + "num_tokens": 470560212.0, + "reward": 0.6930803656578064, + "reward_std": 0.18227308988571167, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 784 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2862.0, + "completions/mean_length": 939.0714721679688, + "completions/mean_terminated_length": 599.5747680664062, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 7.335860058309038, + "grad_norm": 0.1392066329717636, + "learning_rate": 1e-06, + "loss": -0.0387, + "num_tokens": 471126908.0, + "reward": 0.6819196939468384, + "reward_std": 0.13508623838424683, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 785 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2201.0, + "completions/mean_length": 1056.08935546875, + "completions/mean_terminated_length": 621.8162841796875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 7.345189504373177, + "grad_norm": 0.15498603880405426, + "learning_rate": 1e-06, + "loss": -0.0339, + "num_tokens": 471704396.0, + "reward": 0.6049107313156128, + "reward_std": 0.1473340541124344, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 786 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3692.0, + "completions/mean_length": 1047.2601318359375, + "completions/mean_terminated_length": 646.9204711914062, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 7.354518950437318, + "grad_norm": 0.16760165989398956, + "learning_rate": 1e-06, + "loss": -0.0739, + "num_tokens": 472297013.0, + "reward": 0.6517857313156128, + "reward_std": 0.15451332926750183, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 787 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3973.0, + "completions/mean_length": 1065.068115234375, + "completions/mean_terminated_length": 675.7039794921875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 7.363848396501457, + "grad_norm": 0.1438007950782776, + "learning_rate": 1e-06, + "loss": -0.074, + "num_tokens": 472929090.0, + "reward": 0.6573660969734192, + "reward_std": 0.17254258692264557, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 788 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3282.0, + "completions/mean_length": 944.6160888671875, + "completions/mean_terminated_length": 597.065673828125, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 7.373177842565598, + "grad_norm": 0.12902919948101044, + "learning_rate": 1e-06, + "loss": -0.0342, + "num_tokens": 473496762.0, + "reward": 0.7254464626312256, + "reward_std": 0.12192799896001816, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 789 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2843.0, + "completions/mean_length": 802.7020263671875, + "completions/mean_terminated_length": 574.7649536132812, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 7.382507288629737, + "grad_norm": 0.14225251972675323, + "learning_rate": 1e-06, + "loss": -0.0294, + "num_tokens": 474068415.0, + "reward": 0.7109375596046448, + "reward_std": 0.15672755241394043, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 790 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4019.0, + "completions/mean_length": 910.9810791015625, + "completions/mean_terminated_length": 620.023193359375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 7.391836734693878, + "grad_norm": 0.161042720079422, + "learning_rate": 1e-06, + "loss": -0.0297, + "num_tokens": 474668030.0, + "reward": 0.707589328289032, + "reward_std": 0.15315786004066467, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 791 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2867.0, + "completions/mean_length": 916.62841796875, + "completions/mean_terminated_length": 561.6116333007812, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 7.401166180758017, + "grad_norm": 0.17593781650066376, + "learning_rate": 1e-06, + "loss": -0.0183, + "num_tokens": 475218361.0, + "reward": 0.660714328289032, + "reward_std": 0.1504133939743042, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3796.0, + "completions/mean_length": 954.6239013671875, + "completions/mean_terminated_length": 586.4326782226562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 7.410495626822158, + "grad_norm": 0.1471102088689804, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 475777728.0, + "reward": 0.6808035969734192, + "reward_std": 0.1367352306842804, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 793 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3421.0, + "completions/mean_length": 957.4163208007812, + "completions/mean_terminated_length": 615.59033203125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 7.419825072886297, + "grad_norm": 0.14781427383422852, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 476360973.0, + "reward": 0.6863839626312256, + "reward_std": 0.15282082557678223, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 794 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 981.2801513671875, + "completions/mean_terminated_length": 572.2764892578125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 7.429154518950437, + "grad_norm": 0.14452864229679108, + "learning_rate": 1e-06, + "loss": -0.0645, + "num_tokens": 476894488.0, + "reward": 0.6886160969734192, + "reward_std": 0.13283173739910126, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 795 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2993.0, + "completions/mean_length": 1171.3348388671875, + "completions/mean_terminated_length": 710.3411254882812, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 7.438483965014577, + "grad_norm": 0.14694994688034058, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 477543908.0, + "reward": 0.566964328289032, + "reward_std": 0.1586029827594757, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 796 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2992.0, + "completions/mean_length": 1129.8460693359375, + "completions/mean_terminated_length": 648.9520263671875, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 7.447813411078717, + "grad_norm": 0.17439693212509155, + "learning_rate": 1e-06, + "loss": -0.1045, + "num_tokens": 478131186.0, + "reward": 0.609375, + "reward_std": 0.21080127358436584, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 797 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3194.0, + "completions/mean_length": 1074.9921875, + "completions/mean_terminated_length": 634.5895385742188, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 7.457142857142857, + "grad_norm": 0.17701904475688934, + "learning_rate": 1e-06, + "loss": -0.0729, + "num_tokens": 478723603.0, + "reward": 0.6350446939468384, + "reward_std": 0.17949292063713074, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 798 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3232.0, + "completions/mean_length": 1005.0592041015625, + "completions/mean_terminated_length": 647.0796508789062, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 7.466472303206997, + "grad_norm": 0.12809105217456818, + "learning_rate": 1e-06, + "loss": -0.0501, + "num_tokens": 479334864.0, + "reward": 0.6640625, + "reward_std": 0.13809029757976532, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 799 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 988.7756958007812, + "completions/mean_terminated_length": 654.6242065429688, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 7.475801749271137, + "grad_norm": 0.13909395039081573, + "learning_rate": 1e-06, + "loss": -0.0227, + "num_tokens": 479955687.0, + "reward": 0.6875000596046448, + "reward_std": 0.15395060181617737, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 800 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3158.0, + "completions/mean_length": 1103.4420166015625, + "completions/mean_terminated_length": 627.2651977539062, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 7.485131195335277, + "grad_norm": 0.16141991317272186, + "learning_rate": 1e-06, + "loss": -0.0741, + "num_tokens": 480523363.0, + "reward": 0.6283482313156128, + "reward_std": 0.1770554929971695, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 801 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1618303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2823.0, + "completions/mean_length": 1211.69873046875, + "completions/mean_terminated_length": 654.8095703125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.494460641399417, + "grad_norm": 0.15116246044635773, + "learning_rate": 1e-06, + "loss": -0.1158, + "num_tokens": 481100005.0, + "reward": 0.6350446939468384, + "reward_std": 0.18952761590480804, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 802 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3256.0, + "completions/mean_length": 1101.47998046875, + "completions/mean_terminated_length": 633.9484252929688, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 7.503790087463557, + "grad_norm": 0.1282787024974823, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 481685987.0, + "reward": 0.6640625, + "reward_std": 0.12035568058490753, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 803 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3192.0, + "completions/mean_length": 1075.8616943359375, + "completions/mean_terminated_length": 608.8298950195312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 7.513119533527696, + "grad_norm": 0.1601278930902481, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 482250279.0, + "reward": 0.6540178656578064, + "reward_std": 0.16762737929821014, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2814.0, + "completions/mean_length": 1049.9453125, + "completions/mean_terminated_length": 614.7946166992188, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 7.522448979591837, + "grad_norm": 0.1626991629600525, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 482817430.0, + "reward": 0.652901828289032, + "reward_std": 0.17415907979011536, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 805 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3337.0, + "completions/mean_length": 1100.029052734375, + "completions/mean_terminated_length": 632.27099609375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 7.531778425655976, + "grad_norm": 0.14310038089752197, + "learning_rate": 1e-06, + "loss": -0.0512, + "num_tokens": 483386728.0, + "reward": 0.640625, + "reward_std": 0.135238915681839, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 806 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.15625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3505.0, + "completions/mean_length": 1171.40185546875, + "completions/mean_terminated_length": 629.8095092773438, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 7.541107871720117, + "grad_norm": 0.15656399726867676, + "learning_rate": 1e-06, + "loss": -0.0603, + "num_tokens": 483950504.0, + "reward": 0.6049107313156128, + "reward_std": 0.1259523630142212, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 807 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4045.0, + "completions/mean_length": 1061.6685791015625, + "completions/mean_terminated_length": 710.2453002929688, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 7.550437317784256, + "grad_norm": 0.1365361213684082, + "learning_rate": 1e-06, + "loss": -0.0346, + "num_tokens": 484600223.0, + "reward": 0.6194196939468384, + "reward_std": 0.14417481422424316, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 808 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3320.0, + "completions/mean_length": 1017.114990234375, + "completions/mean_terminated_length": 660.53173828125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.559766763848397, + "grad_norm": 0.15237265825271606, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 485218454.0, + "reward": 0.629464328289032, + "reward_std": 0.18550536036491394, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 809 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2910.0, + "completions/mean_length": 1116.6373291015625, + "completions/mean_terminated_length": 578.8603515625, + "completions/min_length": 199.0, + "completions/min_terminated_length": 199.0, + "epoch": 7.569096209912536, + "grad_norm": 0.1521894782781601, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 485759849.0, + "reward": 0.6272321939468384, + "reward_std": 0.13737350702285767, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 810 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3790.0, + "completions/mean_length": 1101.5123291015625, + "completions/mean_terminated_length": 678.087890625, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 7.578425655976677, + "grad_norm": 0.1567058563232422, + "learning_rate": 1e-06, + "loss": -0.0888, + "num_tokens": 486377700.0, + "reward": 0.6428571939468384, + "reward_std": 0.1518009603023529, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 811 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3786.0, + "completions/mean_length": 1089.810302734375, + "completions/mean_terminated_length": 647.157470703125, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 7.587755102040816, + "grad_norm": 0.13738732039928436, + "learning_rate": 1e-06, + "loss": -0.0375, + "num_tokens": 486976106.0, + "reward": 0.6350446939468384, + "reward_std": 0.1594706028699875, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 812 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 1118.40625, + "completions/mean_terminated_length": 640.139892578125, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 7.597084548104956, + "grad_norm": 0.13973119854927063, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 487557750.0, + "reward": 0.6674107313156128, + "reward_std": 0.13655118644237518, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 813 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4021.0, + "completions/mean_length": 840.1529541015625, + "completions/mean_terminated_length": 555.6614379882812, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 7.606413994169096, + "grad_norm": 0.16010016202926636, + "learning_rate": 1e-06, + "loss": -0.0327, + "num_tokens": 488104791.0, + "reward": 0.7566964626312256, + "reward_std": 0.16171550750732422, + "rewards/verify_math_reward/mean": 0.7566964030265808, + "rewards/verify_math_reward/std": 0.4293164908885956, + "step": 814 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3129.0, + "completions/mean_length": 1131.6920166015625, + "completions/mean_terminated_length": 610.409423828125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.615743440233236, + "grad_norm": 0.1457391381263733, + "learning_rate": 1e-06, + "loss": -0.0684, + "num_tokens": 488655459.0, + "reward": 0.6462053656578064, + "reward_std": 0.13629086315631866, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3884.0, + "completions/mean_length": 1082.7801513671875, + "completions/mean_terminated_length": 630.2169799804688, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 7.625072886297376, + "grad_norm": 0.16214509308338165, + "learning_rate": 1e-06, + "loss": -0.0956, + "num_tokens": 489232550.0, + "reward": 0.6082589626312256, + "reward_std": 0.1884777694940567, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.4884119927883148, + "step": 816 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2217.0, + "completions/mean_length": 1073.6004638671875, + "completions/mean_terminated_length": 606.2190551757812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.634402332361516, + "grad_norm": 0.11986333876848221, + "learning_rate": 1e-06, + "loss": -0.0671, + "num_tokens": 489788608.0, + "reward": 0.6238839626312256, + "reward_std": 0.12256810069084167, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 817 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2619.0, + "completions/mean_length": 1052.2132568359375, + "completions/mean_terminated_length": 612.9437866210938, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 7.643731778425656, + "grad_norm": 0.18690215051174164, + "learning_rate": 1e-06, + "loss": -0.0586, + "num_tokens": 490359959.0, + "reward": 0.6361607313156128, + "reward_std": 0.16296251118183136, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3396.0, + "completions/mean_length": 1057.0570068359375, + "completions/mean_terminated_length": 653.6574096679688, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 7.653061224489796, + "grad_norm": 0.14397084712982178, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 490974570.0, + "reward": 0.6417410969734192, + "reward_std": 0.17325754463672638, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 882.8359985351562, + "completions/mean_terminated_length": 580.7435913085938, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 7.662390670553936, + "grad_norm": 0.13318616151809692, + "learning_rate": 1e-06, + "loss": -0.0379, + "num_tokens": 491532767.0, + "reward": 0.7209821939468384, + "reward_std": 0.13016104698181152, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 820 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 1025.2645263671875, + "completions/mean_terminated_length": 665.3529052734375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 7.671720116618076, + "grad_norm": 0.16061414778232574, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 492145468.0, + "reward": 0.6328125, + "reward_std": 0.17408238351345062, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3687.0, + "completions/mean_length": 1062.6328125, + "completions/mean_terminated_length": 659.9734497070312, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 7.681049562682215, + "grad_norm": 0.15363216400146484, + "learning_rate": 1e-06, + "loss": -0.0589, + "num_tokens": 492763803.0, + "reward": 0.59375, + "reward_std": 0.17867128551006317, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 822 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3889.0, + "completions/mean_length": 1009.19873046875, + "completions/mean_terminated_length": 563.7215576171875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 7.690379008746356, + "grad_norm": 0.14927946031093597, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 493291309.0, + "reward": 0.6796875596046448, + "reward_std": 0.13718876242637634, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 823 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3626.0, + "completions/mean_length": 1106.888427734375, + "completions/mean_terminated_length": 585.8505859375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 7.699708454810495, + "grad_norm": 0.1388225555419922, + "learning_rate": 1e-06, + "loss": -0.0705, + "num_tokens": 493825257.0, + "reward": 0.6261160969734192, + "reward_std": 0.12993532419204712, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 824 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3871.0, + "completions/mean_length": 1013.1551513671875, + "completions/mean_terminated_length": 625.863037109375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.709037900874636, + "grad_norm": 0.13892771303653717, + "learning_rate": 1e-06, + "loss": -0.0733, + "num_tokens": 494418492.0, + "reward": 0.640625, + "reward_std": 0.1447407603263855, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 825 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3915.0, + "completions/mean_length": 1076.2109375, + "completions/mean_terminated_length": 622.6611328125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 7.718367346938775, + "grad_norm": 0.1543859988451004, + "learning_rate": 1e-06, + "loss": -0.0235, + "num_tokens": 494997657.0, + "reward": 0.6316964626312256, + "reward_std": 0.137190580368042, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 826 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3279.0, + "completions/mean_length": 1004.036865234375, + "completions/mean_terminated_length": 598.021484375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 7.727696793002916, + "grad_norm": 0.1700589656829834, + "learning_rate": 1e-06, + "loss": -0.0427, + "num_tokens": 495566890.0, + "reward": 0.6506696939468384, + "reward_std": 0.17941875755786896, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 827 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3340.0, + "completions/mean_length": 1039.23779296875, + "completions/mean_terminated_length": 655.2223510742188, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 7.737026239067055, + "grad_norm": 0.12745653092861176, + "learning_rate": 1e-06, + "loss": -0.0527, + "num_tokens": 496181191.0, + "reward": 0.6205357313156128, + "reward_std": 0.13034649193286896, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 828 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3624.0, + "completions/mean_length": 1017.3482666015625, + "completions/mean_terminated_length": 568.5421752929688, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "epoch": 7.746355685131196, + "grad_norm": 0.16452503204345703, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 496715935.0, + "reward": 0.6875000596046448, + "reward_std": 0.14650669693946838, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 829 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2865.0, + "completions/mean_length": 1011.9386596679688, + "completions/mean_terminated_length": 602.5499877929688, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 7.755685131195335, + "grad_norm": 0.1416017860174179, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 497285368.0, + "reward": 0.6696428656578064, + "reward_std": 0.14083515107631683, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 830 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3841.0, + "completions/mean_length": 1045.4453125, + "completions/mean_terminated_length": 627.3489990234375, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.765014577259475, + "grad_norm": 0.1301318258047104, + "learning_rate": 1e-06, + "loss": -0.0757, + "num_tokens": 497866455.0, + "reward": 0.6629464626312256, + "reward_std": 0.14706940948963165, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 831 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4053.0, + "completions/mean_length": 1156.813720703125, + "completions/mean_terminated_length": 680.2918090820312, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 7.774344023323615, + "grad_norm": 0.15623739361763, + "learning_rate": 1e-06, + "loss": -0.0463, + "num_tokens": 498472504.0, + "reward": 0.5613839626312256, + "reward_std": 0.14553281664848328, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 832 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3800.0, + "completions/mean_length": 868.8761596679688, + "completions/mean_terminated_length": 603.8466186523438, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 7.783673469387755, + "grad_norm": 0.15868671238422394, + "learning_rate": 1e-06, + "loss": -0.0309, + "num_tokens": 499060681.0, + "reward": 0.6484375, + "reward_std": 0.155263289809227, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 833 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3271.0, + "completions/mean_length": 891.9051513671875, + "completions/mean_terminated_length": 560.447021484375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 7.793002915451895, + "grad_norm": 0.1387527734041214, + "learning_rate": 1e-06, + "loss": -0.0251, + "num_tokens": 499612860.0, + "reward": 0.7209821939468384, + "reward_std": 0.130574032664299, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 834 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2885.0, + "completions/mean_length": 1093.3951416015625, + "completions/mean_terminated_length": 620.1162719726562, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 7.802332361516035, + "grad_norm": 0.1586790531873703, + "learning_rate": 1e-06, + "loss": -0.0589, + "num_tokens": 500185470.0, + "reward": 0.6305803656578064, + "reward_std": 0.14894986152648926, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 835 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3037.0, + "completions/mean_length": 906.935302734375, + "completions/mean_terminated_length": 581.3604125976562, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 7.811661807580175, + "grad_norm": 0.15984417498111725, + "learning_rate": 1e-06, + "loss": -0.0342, + "num_tokens": 500758788.0, + "reward": 0.6551339626312256, + "reward_std": 0.14984887838363647, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 836 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3331.0, + "completions/mean_length": 1044.630615234375, + "completions/mean_terminated_length": 635.2063598632812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.820991253644315, + "grad_norm": 0.12488831579685211, + "learning_rate": 1e-06, + "loss": -0.041, + "num_tokens": 501346505.0, + "reward": 0.6674107313156128, + "reward_std": 0.11840107291936874, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 837 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2381.0, + "completions/mean_length": 1007.013427734375, + "completions/mean_terminated_length": 588.1013793945312, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.830320699708455, + "grad_norm": 0.17905659973621368, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 501897661.0, + "reward": 0.660714328289032, + "reward_std": 0.1406497061252594, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2812.0, + "completions/mean_length": 898.8906860351562, + "completions/mean_terminated_length": 576.8230590820312, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 7.839650145772595, + "grad_norm": 0.16044704616069794, + "learning_rate": 1e-06, + "loss": -0.063, + "num_tokens": 502462475.0, + "reward": 0.6729910969734192, + "reward_std": 0.17261813580989838, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.4693825840950012, + "step": 839 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2658.0, + "completions/mean_length": 1028.896240234375, + "completions/mean_terminated_length": 599.6577758789062, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 7.848979591836734, + "grad_norm": 0.15445226430892944, + "learning_rate": 1e-06, + "loss": -0.0782, + "num_tokens": 503018686.0, + "reward": 0.629464328289032, + "reward_std": 0.1610877960920334, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 840 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3370.0, + "completions/mean_length": 860.9207763671875, + "completions/mean_terminated_length": 552.4413452148438, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 7.858309037900875, + "grad_norm": 0.14018657803535461, + "learning_rate": 1e-06, + "loss": -0.0622, + "num_tokens": 503567983.0, + "reward": 0.7287946939468384, + "reward_std": 0.14071708917617798, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3226.0, + "completions/mean_length": 1017.0692138671875, + "completions/mean_terminated_length": 625.909423828125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 7.867638483965014, + "grad_norm": 0.1535613238811493, + "learning_rate": 1e-06, + "loss": -0.0634, + "num_tokens": 504150149.0, + "reward": 0.6316964626312256, + "reward_std": 0.18208833038806915, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 842 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3406.0, + "completions/mean_length": 904.5803833007812, + "completions/mean_terminated_length": 561.37451171875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 7.876967930029155, + "grad_norm": 0.15810492634773254, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 504693469.0, + "reward": 0.6674107313156128, + "reward_std": 0.16104431450366974, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 843 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3386.0, + "completions/mean_length": 894.1406860351562, + "completions/mean_terminated_length": 532.1912841796875, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.886297376093294, + "grad_norm": 0.16374319791793823, + "learning_rate": 1e-06, + "loss": -0.0407, + "num_tokens": 505205035.0, + "reward": 0.6908482313156128, + "reward_std": 0.1348598301410675, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 844 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3366.0, + "completions/mean_length": 873.029052734375, + "completions/mean_terminated_length": 612.5476684570312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 7.895626822157435, + "grad_norm": 0.15619264543056488, + "learning_rate": 1e-06, + "loss": -0.0604, + "num_tokens": 505810189.0, + "reward": 0.652901828289032, + "reward_std": 0.17810674011707306, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 845 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3342.0, + "completions/mean_length": 931.4397583007812, + "completions/mean_terminated_length": 604.0714111328125, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 7.904956268221574, + "grad_norm": 0.134511798620224, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 506397359.0, + "reward": 0.6729910969734192, + "reward_std": 0.13601987063884735, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 846 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3867.0, + "completions/mean_length": 972.419677734375, + "completions/mean_terminated_length": 610.6600341796875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 7.914285714285715, + "grad_norm": 0.1518789678812027, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 506988551.0, + "reward": 0.684151828289032, + "reward_std": 0.16086068749427795, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 847 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2585.0, + "completions/mean_length": 1005.5926513671875, + "completions/mean_terminated_length": 573.0928955078125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.923615160349854, + "grad_norm": 0.16868479549884796, + "learning_rate": 1e-06, + "loss": -0.0771, + "num_tokens": 507530530.0, + "reward": 0.645089328289032, + "reward_std": 0.15349414944648743, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 848 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3804.0, + "completions/mean_length": 977.0402221679688, + "completions/mean_terminated_length": 598.3930053710938, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 7.932944606413994, + "grad_norm": 0.14959383010864258, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 508099878.0, + "reward": 0.660714328289032, + "reward_std": 0.14455640316009521, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 849 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1662946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 1207.18310546875, + "completions/mean_terminated_length": 630.966552734375, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 7.942274052478134, + "grad_norm": 0.17509475350379944, + "learning_rate": 1e-06, + "loss": -0.0818, + "num_tokens": 508665986.0, + "reward": 0.5948660969734192, + "reward_std": 0.19227345287799835, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 850 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3551.0, + "completions/mean_length": 965.7210083007812, + "completions/mean_terminated_length": 594.4644165039062, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 7.9516034985422746, + "grad_norm": 0.14221110939979553, + "learning_rate": 1e-06, + "loss": -0.0728, + "num_tokens": 509228304.0, + "reward": 0.637276828289032, + "reward_std": 0.15416745841503143, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 851 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3503.0, + "completions/mean_length": 1157.989990234375, + "completions/mean_terminated_length": 645.8598022460938, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 7.960932944606414, + "grad_norm": 0.17198315262794495, + "learning_rate": 1e-06, + "loss": -0.0704, + "num_tokens": 509809911.0, + "reward": 0.6004464626312256, + "reward_std": 0.174875870347023, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 852 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2859.0, + "completions/mean_length": 910.46435546875, + "completions/mean_terminated_length": 602.4381713867188, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 7.970262390670554, + "grad_norm": 0.15752212703227997, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 510382975.0, + "reward": 0.7120535969734192, + "reward_std": 0.18223915994167328, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 853 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3717.0, + "completions/mean_length": 999.8147583007812, + "completions/mean_terminated_length": 579.926513671875, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "epoch": 7.979591836734694, + "grad_norm": 0.12533758580684662, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 510925985.0, + "reward": 0.6953125596046448, + "reward_std": 0.12606042623519897, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3150.0, + "completions/mean_length": 1086.65966796875, + "completions/mean_terminated_length": 607.8124389648438, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 7.988921282798834, + "grad_norm": 0.15512913465499878, + "learning_rate": 1e-06, + "loss": -0.0732, + "num_tokens": 511496936.0, + "reward": 0.5959821939468384, + "reward_std": 0.16296502947807312, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 855 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.12784090909090906, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2544.0, + "completions/mean_length": 1129.3636474609375, + "completions/mean_terminated_length": 694.5146484375, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 7.998250728862974, + "grad_norm": 0.1386326402425766, + "learning_rate": 1e-06, + "loss": -0.0732, + "num_tokens": 512048994.0, + "reward": 0.676339328289032, + "reward_std": 0.1393595188856125, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 856 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2956.0, + "completions/mean_length": 1145.4866943359375, + "completions/mean_terminated_length": 622.0709228515625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.00932944606414, + "grad_norm": 0.15093453228473663, + "learning_rate": 1e-06, + "loss": -0.0654, + "num_tokens": 512616806.0, + "reward": 0.6261160969734192, + "reward_std": 0.1510535031557083, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 857 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3265.0, + "completions/mean_length": 1038.0859375, + "completions/mean_terminated_length": 592.3030395507812, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 8.018658892128279, + "grad_norm": 0.1357707977294922, + "learning_rate": 1e-06, + "loss": -0.0674, + "num_tokens": 513175843.0, + "reward": 0.6696428656578064, + "reward_std": 0.14995834231376648, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 858 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3095.0, + "completions/mean_length": 1149.2879638671875, + "completions/mean_terminated_length": 617.4044799804688, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 8.02798833819242, + "grad_norm": 0.14728151261806488, + "learning_rate": 1e-06, + "loss": -0.0561, + "num_tokens": 513729045.0, + "reward": 0.6305803656578064, + "reward_std": 0.1385032832622528, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.48291724920272827, + "step": 859 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3655.0, + "completions/mean_length": 1050.0201416015625, + "completions/mean_terminated_length": 654.388427734375, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 8.03731778425656, + "grad_norm": 0.14978253841400146, + "learning_rate": 1e-06, + "loss": -0.0675, + "num_tokens": 514328543.0, + "reward": 0.6495535969734192, + "reward_std": 0.16330133378505707, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 860 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3406.0, + "completions/mean_length": 977.4107666015625, + "completions/mean_terminated_length": 603.1799926757812, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 8.0466472303207, + "grad_norm": 0.15136602520942688, + "learning_rate": 1e-06, + "loss": -0.062, + "num_tokens": 514896135.0, + "reward": 0.6752232313156128, + "reward_std": 0.14951257407665253, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 861 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4056.0, + "completions/mean_length": 1060.4888916015625, + "completions/mean_terminated_length": 609.0538940429688, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 8.055976676384839, + "grad_norm": 0.15994176268577576, + "learning_rate": 1e-06, + "loss": -0.0818, + "num_tokens": 515452629.0, + "reward": 0.6830357313156128, + "reward_std": 0.1632232367992401, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 862 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3785.0, + "completions/mean_length": 1120.1529541015625, + "completions/mean_terminated_length": 610.5634155273438, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 8.06530612244898, + "grad_norm": 0.16492946445941925, + "learning_rate": 1e-06, + "loss": -0.0716, + "num_tokens": 516024046.0, + "reward": 0.6328125, + "reward_std": 0.15375559031963348, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 863 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3592.0, + "completions/mean_length": 970.93310546875, + "completions/mean_terminated_length": 626.2850341796875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 8.07463556851312, + "grad_norm": 0.12416037172079086, + "learning_rate": 1e-06, + "loss": -0.0377, + "num_tokens": 516615346.0, + "reward": 0.6707589626312256, + "reward_std": 0.12358727306127548, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 864 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2338.0, + "completions/mean_length": 1082.2254638671875, + "completions/mean_terminated_length": 575.3455200195312, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 8.08396501457726, + "grad_norm": 0.16226497292518616, + "learning_rate": 1e-06, + "loss": -0.0778, + "num_tokens": 517154060.0, + "reward": 0.640625, + "reward_std": 0.14011767506599426, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2840.0, + "completions/mean_length": 902.01123046875, + "completions/mean_terminated_length": 584.57177734375, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 8.093294460641399, + "grad_norm": 0.1480923444032669, + "learning_rate": 1e-06, + "loss": -0.0391, + "num_tokens": 517716126.0, + "reward": 0.6651785969734192, + "reward_std": 0.16394072771072388, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 866 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3843.0, + "completions/mean_length": 1153.10498046875, + "completions/mean_terminated_length": 662.6224365234375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 8.102623906705539, + "grad_norm": 0.144929900765419, + "learning_rate": 1e-06, + "loss": -0.0731, + "num_tokens": 518305724.0, + "reward": 0.6071428656578064, + "reward_std": 0.18463844060897827, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865824937820435, + "step": 867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3757.0, + "completions/mean_length": 1043.1328125, + "completions/mean_terminated_length": 633.5076293945312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.11195335276968, + "grad_norm": 0.1655925214290619, + "learning_rate": 1e-06, + "loss": -0.0397, + "num_tokens": 518903995.0, + "reward": 0.6517857313156128, + "reward_std": 0.17119666934013367, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 868 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3576.0, + "completions/mean_length": 1001.661865234375, + "completions/mean_terminated_length": 608.5446166992188, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 8.12128279883382, + "grad_norm": 0.16748148202896118, + "learning_rate": 1e-06, + "loss": -0.069, + "num_tokens": 519469948.0, + "reward": 0.6640625, + "reward_std": 0.16461403667926788, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 869 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3593.0, + "completions/mean_length": 1038.7310791015625, + "completions/mean_terminated_length": 632.8988647460938, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 8.130612244897959, + "grad_norm": 0.126932755112648, + "learning_rate": 1e-06, + "loss": -0.0776, + "num_tokens": 520055955.0, + "reward": 0.6339285969734192, + "reward_std": 0.12730783224105835, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 870 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3318.0, + "completions/mean_length": 1051.7734375, + "completions/mean_terminated_length": 607.9859008789062, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 8.139941690962099, + "grad_norm": 0.13293921947479248, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 520626712.0, + "reward": 0.6674107313156128, + "reward_std": 0.10780475288629532, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 871 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2910.0, + "completions/mean_length": 1010.3281860351562, + "completions/mean_terminated_length": 591.8656616210938, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 8.14927113702624, + "grad_norm": 0.1483665108680725, + "learning_rate": 1e-06, + "loss": -0.0801, + "num_tokens": 521190094.0, + "reward": 0.6640625, + "reward_std": 0.15431830286979675, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 872 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2800.0, + "completions/mean_length": 1017.9319458007812, + "completions/mean_terminated_length": 596.064697265625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 8.15860058309038, + "grad_norm": 0.16868089139461517, + "learning_rate": 1e-06, + "loss": -0.0918, + "num_tokens": 521752505.0, + "reward": 0.7220982313156128, + "reward_std": 0.18532174825668335, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 873 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1584821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 1199.4420166015625, + "completions/mean_terminated_length": 653.9363403320312, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "epoch": 8.167930029154519, + "grad_norm": 0.16264265775680542, + "learning_rate": 1e-06, + "loss": -0.1025, + "num_tokens": 522338133.0, + "reward": 0.6272321939468384, + "reward_std": 0.1818275898694992, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 874 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3431.0, + "completions/mean_length": 1028.091552734375, + "completions/mean_terminated_length": 580.8516845703125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 8.177259475218658, + "grad_norm": 0.17777729034423828, + "learning_rate": 1e-06, + "loss": -0.0396, + "num_tokens": 522872855.0, + "reward": 0.6774553656578064, + "reward_std": 0.161088228225708, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3197.0, + "completions/mean_length": 919.97998046875, + "completions/mean_terminated_length": 543.29833984375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 8.186588921282798, + "grad_norm": 0.13674885034561157, + "learning_rate": 1e-06, + "loss": -0.0438, + "num_tokens": 523400565.0, + "reward": 0.7254464626312256, + "reward_std": 0.10554774850606918, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 876 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2790.0, + "completions/mean_length": 1002.333740234375, + "completions/mean_terminated_length": 600.5081787109375, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 8.19591836734694, + "grad_norm": 0.15885168313980103, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 523976488.0, + "reward": 0.6796875596046448, + "reward_std": 0.15849490463733673, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 877 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3481.0, + "completions/mean_length": 1154.37060546875, + "completions/mean_terminated_length": 595.7344360351562, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 8.205247813411079, + "grad_norm": 0.14257389307022095, + "learning_rate": 1e-06, + "loss": -0.0828, + "num_tokens": 524518396.0, + "reward": 0.6462053656578064, + "reward_std": 0.13327428698539734, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 878 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3330.0, + "completions/mean_length": 1001.3984985351562, + "completions/mean_terminated_length": 612.62939453125, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 8.214577259475218, + "grad_norm": 0.1312379390001297, + "learning_rate": 1e-06, + "loss": -0.0166, + "num_tokens": 525098425.0, + "reward": 0.660714328289032, + "reward_std": 0.11152489483356476, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4025.0, + "completions/mean_length": 1136.8426513671875, + "completions/mean_terminated_length": 652.6168823242188, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 8.223906705539358, + "grad_norm": 0.16868945956230164, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 525690092.0, + "reward": 0.59375, + "reward_std": 0.17479778826236725, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 880 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2779.0, + "completions/mean_length": 1014.15185546875, + "completions/mean_terminated_length": 613.8612670898438, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 8.2332361516035, + "grad_norm": 0.14295226335525513, + "learning_rate": 1e-06, + "loss": -0.0539, + "num_tokens": 526272308.0, + "reward": 0.6339285969734192, + "reward_std": 0.14913460612297058, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 881 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1551339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2826.0, + "completions/mean_length": 1160.45654296875, + "completions/mean_terminated_length": 621.4332885742188, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 8.242565597667639, + "grad_norm": 0.16172069311141968, + "learning_rate": 1e-06, + "loss": -0.0829, + "num_tokens": 526846245.0, + "reward": 0.6071428656578064, + "reward_std": 0.1732582449913025, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865824937820435, + "step": 882 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3266.0, + "completions/mean_length": 1042.44091796875, + "completions/mean_terminated_length": 592.8130493164062, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 8.251895043731778, + "grad_norm": 0.14926446974277496, + "learning_rate": 1e-06, + "loss": -0.0562, + "num_tokens": 527398944.0, + "reward": 0.637276828289032, + "reward_std": 0.12358727306127548, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3236.0, + "completions/mean_length": 1076.0413818359375, + "completions/mean_terminated_length": 609.037353515625, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 8.261224489795918, + "grad_norm": 0.14860747754573822, + "learning_rate": 1e-06, + "loss": -0.0434, + "num_tokens": 527954357.0, + "reward": 0.637276828289032, + "reward_std": 0.13978207111358643, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 884 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3859.0, + "completions/mean_length": 1049.3248291015625, + "completions/mean_terminated_length": 614.08544921875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 8.270553935860057, + "grad_norm": 0.1655399203300476, + "learning_rate": 1e-06, + "loss": -0.0925, + "num_tokens": 528533240.0, + "reward": 0.6629464626312256, + "reward_std": 0.16671697795391083, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 885 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2722.0, + "completions/mean_length": 1051.6998291015625, + "completions/mean_terminated_length": 603.4353637695312, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 8.279883381924199, + "grad_norm": 0.17304687201976776, + "learning_rate": 1e-06, + "loss": -0.0455, + "num_tokens": 529095307.0, + "reward": 0.6428571939468384, + "reward_std": 0.1406843066215515, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.47942501306533813, + "step": 886 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2505.0, + "completions/mean_length": 964.4609985351562, + "completions/mean_terminated_length": 614.7853393554688, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 8.289212827988338, + "grad_norm": 0.14617525041103363, + "learning_rate": 1e-06, + "loss": -0.072, + "num_tokens": 529675360.0, + "reward": 0.6964285969734192, + "reward_std": 0.15852628648281097, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 887 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3792.0, + "completions/mean_length": 994.591552734375, + "completions/mean_terminated_length": 626.759033203125, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 8.298542274052478, + "grad_norm": 0.1307070255279541, + "learning_rate": 1e-06, + "loss": -0.046, + "num_tokens": 530267394.0, + "reward": 0.6495535969734192, + "reward_std": 0.13659213483333588, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 888 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3075.0, + "completions/mean_length": 1053.7679443359375, + "completions/mean_terminated_length": 601.3333740234375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 8.307871720116617, + "grad_norm": 0.15937061607837677, + "learning_rate": 1e-06, + "loss": -0.0932, + "num_tokens": 530822922.0, + "reward": 0.7120535969734192, + "reward_std": 0.16262732446193695, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 889 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3863.0, + "completions/mean_length": 1050.3348388671875, + "completions/mean_terminated_length": 606.3375854492188, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.317201166180759, + "grad_norm": 0.15355893969535828, + "learning_rate": 1e-06, + "loss": -0.0407, + "num_tokens": 531390670.0, + "reward": 0.6830357313156128, + "reward_std": 0.13929282128810883, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 890 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3655.0, + "completions/mean_length": 1113.3270263671875, + "completions/mean_terminated_length": 647.6452026367188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 8.326530612244898, + "grad_norm": 0.15006402134895325, + "learning_rate": 1e-06, + "loss": -0.057, + "num_tokens": 531988539.0, + "reward": 0.640625, + "reward_std": 0.15120504796504974, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 891 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3029.0, + "completions/mean_length": 1022.8739013671875, + "completions/mean_terminated_length": 561.313232421875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 8.335860058309038, + "grad_norm": 0.1519111841917038, + "learning_rate": 1e-06, + "loss": -0.0573, + "num_tokens": 532513786.0, + "reward": 0.625, + "reward_std": 0.13624808192253113, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 892 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2708.0, + "completions/mean_length": 1004.310302734375, + "completions/mean_terminated_length": 620.2734985351562, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 8.345189504373177, + "grad_norm": 0.15414708852767944, + "learning_rate": 1e-06, + "loss": -0.0728, + "num_tokens": 533089280.0, + "reward": 0.660714328289032, + "reward_std": 0.17187067866325378, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 893 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3471.0, + "completions/mean_length": 1069.9810791015625, + "completions/mean_terminated_length": 579.38134765625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 8.354518950437317, + "grad_norm": 0.15470200777053833, + "learning_rate": 1e-06, + "loss": -0.0646, + "num_tokens": 533629479.0, + "reward": 0.6584821939468384, + "reward_std": 0.12463894486427307, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 894 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1618303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2539.0, + "completions/mean_length": 1188.857177734375, + "completions/mean_terminated_length": 627.5579223632812, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.363848396501458, + "grad_norm": 0.17332008481025696, + "learning_rate": 1e-06, + "loss": -0.0934, + "num_tokens": 534184247.0, + "reward": 0.606026828289032, + "reward_std": 0.1813383400440216, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 895 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 1083.4989013671875, + "completions/mean_terminated_length": 661.9020385742188, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 8.373177842565598, + "grad_norm": 0.14186617732048035, + "learning_rate": 1e-06, + "loss": -0.081, + "num_tokens": 534787886.0, + "reward": 0.6830357313156128, + "reward_std": 0.14278654754161835, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 896 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3524.0, + "completions/mean_length": 936.771240234375, + "completions/mean_terminated_length": 544.3450317382812, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 8.382507288629737, + "grad_norm": 0.16455881297588348, + "learning_rate": 1e-06, + "loss": -0.0608, + "num_tokens": 535322449.0, + "reward": 0.668526828289032, + "reward_std": 0.1612725555896759, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 897 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3181.0, + "completions/mean_length": 1029.3348388671875, + "completions/mean_terminated_length": 600.1577758789062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.391836734693877, + "grad_norm": 0.15738701820373535, + "learning_rate": 1e-06, + "loss": -0.0831, + "num_tokens": 535871189.0, + "reward": 0.7008928656578064, + "reward_std": 0.15770143270492554, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 898 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4035.0, + "completions/mean_length": 971.5000610351562, + "completions/mean_terminated_length": 543.26904296875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 8.401166180758018, + "grad_norm": 0.15441343188285828, + "learning_rate": 1e-06, + "loss": -0.052, + "num_tokens": 536391389.0, + "reward": 0.6953125596046448, + "reward_std": 0.12219971418380737, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 899 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2442.0, + "completions/mean_length": 933.7131958007812, + "completions/mean_terminated_length": 571.8594360351562, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 8.410495626822158, + "grad_norm": 0.15590281784534454, + "learning_rate": 1e-06, + "loss": -0.0799, + "num_tokens": 536931564.0, + "reward": 0.7321428656578064, + "reward_std": 0.1756240278482437, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 900 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3387.0, + "completions/mean_length": 1023.6685791015625, + "completions/mean_terminated_length": 637.6972045898438, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.419825072886297, + "grad_norm": 0.15883347392082214, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 537522291.0, + "reward": 0.6975446939468384, + "reward_std": 0.15751782059669495, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 901 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3234.0, + "completions/mean_length": 913.35498046875, + "completions/mean_terminated_length": 579.7854614257812, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 8.429154518950437, + "grad_norm": 0.10371751338243484, + "learning_rate": 1e-06, + "loss": -0.0248, + "num_tokens": 538073625.0, + "reward": 0.7455357313156128, + "reward_std": 0.07792925834655762, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2465.0, + "completions/mean_length": 942.7388916015625, + "completions/mean_terminated_length": 581.9179077148438, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 8.438483965014576, + "grad_norm": 0.1574653536081314, + "learning_rate": 1e-06, + "loss": -0.0578, + "num_tokens": 538628919.0, + "reward": 0.6830357313156128, + "reward_std": 0.14117145538330078, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 903 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3728.0, + "completions/mean_length": 974.6808471679688, + "completions/mean_terminated_length": 639.0136108398438, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 8.447813411078718, + "grad_norm": 0.14363057911396027, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 539242681.0, + "reward": 0.6741071939468384, + "reward_std": 0.1500004082918167, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3631.0, + "completions/mean_length": 1052.607177734375, + "completions/mean_terminated_length": 581.9793701171875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 8.457142857142857, + "grad_norm": 0.1549978256225586, + "learning_rate": 1e-06, + "loss": -0.0794, + "num_tokens": 539790161.0, + "reward": 0.6618303656578064, + "reward_std": 0.17194415628910065, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3962.0, + "completions/mean_length": 954.2355346679688, + "completions/mean_terminated_length": 590.369873046875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 8.466472303206997, + "grad_norm": 0.14460058510303497, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 540350148.0, + "reward": 0.7109375596046448, + "reward_std": 0.14444582164287567, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 906 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3915.0, + "completions/mean_length": 1096.669677734375, + "completions/mean_terminated_length": 632.8556518554688, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 8.475801749271136, + "grad_norm": 0.1491176187992096, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 540926692.0, + "reward": 0.645089328289032, + "reward_std": 0.1572154015302658, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 907 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1662946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3849.0, + "completions/mean_length": 1207.3695068359375, + "completions/mean_terminated_length": 631.1900634765625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 8.485131195335278, + "grad_norm": 0.16436566412448883, + "learning_rate": 1e-06, + "loss": -0.0692, + "num_tokens": 541481423.0, + "reward": 0.6227678656578064, + "reward_std": 0.14015229046344757, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 908 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2280.0, + "completions/mean_length": 948.1172485351562, + "completions/mean_terminated_length": 600.9529418945312, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 8.494460641399417, + "grad_norm": 0.15124177932739258, + "learning_rate": 1e-06, + "loss": -0.0574, + "num_tokens": 542052448.0, + "reward": 0.6417410969734192, + "reward_std": 0.14730015397071838, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3658.0, + "completions/mean_length": 1066.1160888671875, + "completions/mean_terminated_length": 642.0864868164062, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 8.503790087463557, + "grad_norm": 0.1336049735546112, + "learning_rate": 1e-06, + "loss": -0.0556, + "num_tokens": 542657192.0, + "reward": 0.645089328289032, + "reward_std": 0.13380561769008636, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 910 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3747.0, + "completions/mean_length": 774.1574096679688, + "completions/mean_terminated_length": 548.4779663085938, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 8.513119533527696, + "grad_norm": 0.1571814864873886, + "learning_rate": 1e-06, + "loss": -0.0308, + "num_tokens": 543206221.0, + "reward": 0.7142857313156128, + "reward_std": 0.13305744528770447, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 911 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4056.0, + "completions/mean_length": 1178.599365234375, + "completions/mean_terminated_length": 665.5656127929688, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 8.522448979591836, + "grad_norm": 0.1574980467557907, + "learning_rate": 1e-06, + "loss": -0.0697, + "num_tokens": 543807582.0, + "reward": 0.5446428656578064, + "reward_std": 0.14650921523571014, + "rewards/verify_math_reward/mean": 0.5446428656578064, + "rewards/verify_math_reward/std": 0.4982811510562897, + "step": 912 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3956.0, + "completions/mean_length": 1108.7232666015625, + "completions/mean_terminated_length": 646.773193359375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 8.531778425655977, + "grad_norm": 0.15376946330070496, + "learning_rate": 1e-06, + "loss": -0.0676, + "num_tokens": 544394766.0, + "reward": 0.645089328289032, + "reward_std": 0.16070912778377533, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 913 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2992.0, + "completions/mean_length": 1027.4476318359375, + "completions/mean_terminated_length": 552.9290771484375, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 8.541107871720117, + "grad_norm": 0.17366181313991547, + "learning_rate": 1e-06, + "loss": -0.0805, + "num_tokens": 544918103.0, + "reward": 0.6808035969734192, + "reward_std": 0.14808222651481628, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 914 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2780.0, + "completions/mean_length": 882.185302734375, + "completions/mean_terminated_length": 597.1203002929688, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 8.550437317784256, + "grad_norm": 0.1341852843761444, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 545498109.0, + "reward": 0.7120535969734192, + "reward_std": 0.12422595918178558, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 915 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3485.0, + "completions/mean_length": 1174.6395263671875, + "completions/mean_terminated_length": 727.2239379882812, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 8.559766763848396, + "grad_norm": 0.1317211389541626, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 546162178.0, + "reward": 0.578125, + "reward_std": 0.1328285187482834, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3681.0, + "completions/mean_length": 1050.578125, + "completions/mean_terminated_length": 633.1852416992188, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.569096209912537, + "grad_norm": 0.144380122423172, + "learning_rate": 1e-06, + "loss": -0.032, + "num_tokens": 546749536.0, + "reward": 0.621651828289032, + "reward_std": 0.13493286073207855, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 917 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2272.0, + "completions/mean_length": 1031.766845703125, + "completions/mean_terminated_length": 598.4802856445312, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 8.578425655976677, + "grad_norm": 0.14323492348194122, + "learning_rate": 1e-06, + "loss": -0.0574, + "num_tokens": 547313319.0, + "reward": 0.6640625, + "reward_std": 0.1482747346162796, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 918 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3186.0, + "completions/mean_length": 831.3850708007812, + "completions/mean_terminated_length": 576.0300903320312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 8.587755102040816, + "grad_norm": 0.14809049665927887, + "learning_rate": 1e-06, + "loss": -0.0708, + "num_tokens": 547887520.0, + "reward": 0.7142857313156128, + "reward_std": 0.16491642594337463, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 919 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4047.0, + "completions/mean_length": 1038.96875, + "completions/mean_terminated_length": 659.2371215820312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 8.597084548104956, + "grad_norm": 0.15314710140228271, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 548495660.0, + "reward": 0.6205357313156128, + "reward_std": 0.15875521302223206, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 920 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3705.0, + "completions/mean_length": 964.9464721679688, + "completions/mean_terminated_length": 580.4310913085938, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 8.606413994169095, + "grad_norm": 0.15882423520088196, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 549041524.0, + "reward": 0.6964285969734192, + "reward_std": 0.13763564825057983, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 921 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3518.0, + "completions/mean_length": 926.443115234375, + "completions/mean_terminated_length": 653.6690673828125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.615743440233237, + "grad_norm": 0.14339464902877808, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 549664913.0, + "reward": 0.660714328289032, + "reward_std": 0.15074971318244934, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 922 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 1056.84375, + "completions/mean_terminated_length": 609.3367309570312, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 8.625072886297376, + "grad_norm": 0.14878612756729126, + "learning_rate": 1e-06, + "loss": -0.0796, + "num_tokens": 550236989.0, + "reward": 0.6875000596046448, + "reward_std": 0.14042216539382935, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 923 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3077.0, + "completions/mean_length": 918.9710083007812, + "completions/mean_terminated_length": 564.2158813476562, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 8.634402332361516, + "grad_norm": 0.17830529808998108, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 550784067.0, + "reward": 0.6819196939468384, + "reward_std": 0.1710130274295807, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3374.0, + "completions/mean_length": 1028.72998046875, + "completions/mean_terminated_length": 621.5701904296875, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 8.643731778425655, + "grad_norm": 0.15259338915348053, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 551367705.0, + "reward": 0.645089328289032, + "reward_std": 0.14774592220783234, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 925 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3203.0, + "completions/mean_length": 1023.1250610351562, + "completions/mean_terminated_length": 610.815185546875, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 8.653061224489797, + "grad_norm": 0.1758723109960556, + "learning_rate": 1e-06, + "loss": -0.0876, + "num_tokens": 551938129.0, + "reward": 0.6651785969734192, + "reward_std": 0.192087322473526, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 926 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2504.0, + "completions/mean_length": 967.193115234375, + "completions/mean_terminated_length": 556.3396606445312, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 8.662390670553936, + "grad_norm": 0.13319288194179535, + "learning_rate": 1e-06, + "loss": -0.0461, + "num_tokens": 552469966.0, + "reward": 0.6584821939468384, + "reward_std": 0.10795699059963226, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 927 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4091.0, + "completions/mean_length": 1093.1484375, + "completions/mean_terminated_length": 637.7030639648438, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 8.671720116618076, + "grad_norm": 0.13909879326820374, + "learning_rate": 1e-06, + "loss": -0.0639, + "num_tokens": 553055203.0, + "reward": 0.6171875, + "reward_std": 0.14838463068008423, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 928 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2764.0, + "completions/mean_length": 1066.759033203125, + "completions/mean_terminated_length": 607.31103515625, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 8.681049562682215, + "grad_norm": 0.1701621115207672, + "learning_rate": 1e-06, + "loss": -0.0785, + "num_tokens": 553616387.0, + "reward": 0.6595982313156128, + "reward_std": 0.15138980746269226, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 929 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3890.0, + "completions/mean_length": 1081.673095703125, + "completions/mean_terminated_length": 588.4194946289062, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 8.690379008746355, + "grad_norm": 0.16358621418476105, + "learning_rate": 1e-06, + "loss": -0.0669, + "num_tokens": 554155782.0, + "reward": 0.6618303656578064, + "reward_std": 0.15424412488937378, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 930 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3764.0, + "completions/mean_length": 1188.7545166015625, + "completions/mean_terminated_length": 668.5105590820312, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 8.699708454810496, + "grad_norm": 0.16409288346767426, + "learning_rate": 1e-06, + "loss": -0.0642, + "num_tokens": 554757290.0, + "reward": 0.6037946939468384, + "reward_std": 0.17434313893318176, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3769.0, + "completions/mean_length": 973.2489013671875, + "completions/mean_terminated_length": 641.6975708007812, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 8.709037900874636, + "grad_norm": 0.1480489820241928, + "learning_rate": 1e-06, + "loss": -0.0453, + "num_tokens": 555370673.0, + "reward": 0.6160714626312256, + "reward_std": 0.13958704471588135, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 932 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2983.0, + "completions/mean_length": 1122.9520263671875, + "completions/mean_terminated_length": 627.4440307617188, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 8.718367346938775, + "grad_norm": 0.17100459337234497, + "learning_rate": 1e-06, + "loss": -0.0514, + "num_tokens": 555941238.0, + "reward": 0.668526828289032, + "reward_std": 0.13440336287021637, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 933 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3000.0, + "completions/mean_length": 1114.6239013671875, + "completions/mean_terminated_length": 658.0167236328125, + "completions/min_length": 199.0, + "completions/min_terminated_length": 199.0, + "epoch": 8.727696793002915, + "grad_norm": 0.14564810693264008, + "learning_rate": 1e-06, + "loss": -0.0715, + "num_tokens": 556538685.0, + "reward": 0.6283482313156128, + "reward_std": 0.14954577386379242, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 934 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3009.0, + "completions/mean_length": 1065.2645263671875, + "completions/mean_terminated_length": 578.46240234375, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 8.737026239067056, + "grad_norm": 0.14742593467235565, + "learning_rate": 1e-06, + "loss": -0.0631, + "num_tokens": 557064722.0, + "reward": 0.6741071939468384, + "reward_std": 0.12106993794441223, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 935 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3937.0, + "completions/mean_length": 992.5870971679688, + "completions/mean_terminated_length": 598.3169555664062, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 8.746355685131196, + "grad_norm": 0.1486990749835968, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 557628496.0, + "reward": 0.684151828289032, + "reward_std": 0.1407923698425293, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 936 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3360.0, + "completions/mean_length": 1003.29248046875, + "completions/mean_terminated_length": 597.1793212890625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 8.755685131195335, + "grad_norm": 0.14442980289459229, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 558194678.0, + "reward": 0.6941964626312256, + "reward_std": 0.14207187294960022, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 937 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3625.0, + "completions/mean_length": 1034.5546875, + "completions/mean_terminated_length": 649.9509887695312, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 8.765014577259475, + "grad_norm": 0.2729840874671936, + "learning_rate": 1e-06, + "loss": -0.073, + "num_tokens": 558798711.0, + "reward": 0.6941964626312256, + "reward_std": 0.1557832509279251, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 938 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4035.0, + "completions/mean_length": 788.6864013671875, + "completions/mean_terminated_length": 538.5534057617188, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 8.774344023323614, + "grad_norm": 0.15766826272010803, + "learning_rate": 1e-06, + "loss": -0.0529, + "num_tokens": 559331302.0, + "reward": 0.7287946939468384, + "reward_std": 0.1349007785320282, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 939 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3821.0, + "completions/mean_length": 1134.8895263671875, + "completions/mean_terminated_length": 623.2840576171875, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 8.783673469387756, + "grad_norm": 0.17000079154968262, + "learning_rate": 1e-06, + "loss": -0.0793, + "num_tokens": 559901171.0, + "reward": 0.6283482313156128, + "reward_std": 0.14304685592651367, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 940 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2511.0, + "completions/mean_length": 901.47998046875, + "completions/mean_terminated_length": 592.5850830078125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 8.793002915451895, + "grad_norm": 0.12940526008605957, + "learning_rate": 1e-06, + "loss": -0.0569, + "num_tokens": 560464377.0, + "reward": 0.6986607313156128, + "reward_std": 0.12388965487480164, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960144996643, + "step": 941 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3554.0, + "completions/mean_length": 1056.4554443359375, + "completions/mean_terminated_length": 604.4205322265625, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 8.802332361516035, + "grad_norm": 0.15664970874786377, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 561020921.0, + "reward": 0.6696428656578064, + "reward_std": 0.12787306308746338, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 942 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3071.0, + "completions/mean_length": 944.68310546875, + "completions/mean_terminated_length": 592.7990112304688, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 8.811661807580174, + "grad_norm": 0.15528330206871033, + "learning_rate": 1e-06, + "loss": -0.0734, + "num_tokens": 561588853.0, + "reward": 0.6953125596046448, + "reward_std": 0.14420802891254425, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 943 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3023.0, + "completions/mean_length": 1003.5670166015625, + "completions/mean_terminated_length": 649.7064819335938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 8.820991253644316, + "grad_norm": 0.13627398014068604, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 562196809.0, + "reward": 0.6696428656578064, + "reward_std": 0.15409894287586212, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 944 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3902.0, + "completions/mean_length": 1060.505615234375, + "completions/mean_terminated_length": 617.9910278320312, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 8.830320699708455, + "grad_norm": 0.17659050226211548, + "learning_rate": 1e-06, + "loss": -0.0691, + "num_tokens": 562763950.0, + "reward": 0.6428571939468384, + "reward_std": 0.17017610371112823, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 945 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4088.0, + "completions/mean_length": 975.5535888671875, + "completions/mean_terminated_length": 592.3408203125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.839650145772595, + "grad_norm": 0.17134946584701538, + "learning_rate": 1e-06, + "loss": -0.085, + "num_tokens": 563320382.0, + "reward": 0.6796875596046448, + "reward_std": 0.17934459447860718, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 946 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3398.0, + "completions/mean_length": 986.0156860351562, + "completions/mean_terminated_length": 617.166015625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.848979591836734, + "grad_norm": 0.1389453113079071, + "learning_rate": 1e-06, + "loss": -0.0232, + "num_tokens": 563897252.0, + "reward": 0.640625, + "reward_std": 0.13616888225078583, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 947 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3182.0, + "completions/mean_length": 1040.96435546875, + "completions/mean_terminated_length": 678.6317138671875, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.858309037900874, + "grad_norm": 0.14301376044750214, + "learning_rate": 1e-06, + "loss": -0.0496, + "num_tokens": 564530956.0, + "reward": 0.6584821939468384, + "reward_std": 0.16833347082138062, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 948 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3543.0, + "completions/mean_length": 997.01904296875, + "completions/mean_terminated_length": 594.503173828125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 8.867638483965015, + "grad_norm": 0.14853087067604065, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 565089165.0, + "reward": 0.660714328289032, + "reward_std": 0.14545543491840363, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 949 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2576.0, + "completions/mean_length": 1098.884033203125, + "completions/mean_terminated_length": 590.2349853515625, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 8.876967930029155, + "grad_norm": 0.15760570764541626, + "learning_rate": 1e-06, + "loss": -0.0413, + "num_tokens": 565631341.0, + "reward": 0.6573660969734192, + "reward_std": 0.13298258185386658, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 950 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3224.0, + "completions/mean_length": 1056.919677734375, + "completions/mean_terminated_length": 613.88232421875, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 8.886297376093294, + "grad_norm": 0.13932481408119202, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 566205437.0, + "reward": 0.6428571939468384, + "reward_std": 0.10874612629413605, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4067.0, + "completions/mean_length": 973.489990234375, + "completions/mean_terminated_length": 646.22314453125, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 8.895626822157434, + "grad_norm": 0.15475697815418243, + "learning_rate": 1e-06, + "loss": -0.0285, + "num_tokens": 566819572.0, + "reward": 0.6774553656578064, + "reward_std": 0.14936032891273499, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 952 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3537.0, + "completions/mean_length": 1029.4866943359375, + "completions/mean_terminated_length": 568.9191284179688, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 8.904956268221575, + "grad_norm": 0.1805417388677597, + "learning_rate": 1e-06, + "loss": -0.0705, + "num_tokens": 567353936.0, + "reward": 0.645089328289032, + "reward_std": 0.16266193985939026, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 953 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2703.0, + "completions/mean_length": 983.0469360351562, + "completions/mean_terminated_length": 587.5647583007812, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 8.914285714285715, + "grad_norm": 0.17471978068351746, + "learning_rate": 1e-06, + "loss": -0.084, + "num_tokens": 567912282.0, + "reward": 0.6908482313156128, + "reward_std": 0.16540679335594177, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 954 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3157.0, + "completions/mean_length": 1032.2601318359375, + "completions/mean_terminated_length": 625.5689086914062, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 8.923615160349854, + "grad_norm": 0.17706899344921112, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 568490827.0, + "reward": 0.6774553656578064, + "reward_std": 0.1782590001821518, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 955 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3187.0, + "completions/mean_length": 1100.747802734375, + "completions/mean_terminated_length": 637.5643920898438, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 8.932944606413994, + "grad_norm": 0.14129015803337097, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 569075809.0, + "reward": 0.6495535969734192, + "reward_std": 0.12918394804000854, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 956 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1685267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3967.0, + "completions/mean_length": 1210.188720703125, + "completions/mean_terminated_length": 625.2792358398438, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.942274052478133, + "grad_norm": 0.15340207517147064, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 569633322.0, + "reward": 0.5770089626312256, + "reward_std": 0.14481531083583832, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 957 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2680.0, + "completions/mean_length": 1033.3504638671875, + "completions/mean_terminated_length": 657.235595703125, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 8.951603498542275, + "grad_norm": 0.1456962674856186, + "learning_rate": 1e-06, + "loss": -0.0621, + "num_tokens": 570253044.0, + "reward": 0.6328125, + "reward_std": 0.17209498584270477, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 958 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3812.0, + "completions/mean_length": 1101.59375, + "completions/mean_terminated_length": 597.9713134765625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 8.960932944606414, + "grad_norm": 0.18818694353103638, + "learning_rate": 1e-06, + "loss": -0.0774, + "num_tokens": 570802896.0, + "reward": 0.6930803656578064, + "reward_std": 0.135540172457695, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 959 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3949.0, + "completions/mean_length": 1114.310302734375, + "completions/mean_terminated_length": 644.3281860351562, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 8.970262390670554, + "grad_norm": 0.16497798264026642, + "learning_rate": 1e-06, + "loss": -0.0693, + "num_tokens": 571398326.0, + "reward": 0.6785714626312256, + "reward_std": 0.16570806503295898, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 960 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2544.0, + "completions/mean_length": 932.7745971679688, + "completions/mean_terminated_length": 579.560791015625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 8.979591836734693, + "grad_norm": 0.14131681621074677, + "learning_rate": 1e-06, + "loss": -0.0436, + "num_tokens": 571954012.0, + "reward": 0.7332589626312256, + "reward_std": 0.12351170927286148, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425029158592224, + "step": 961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4087.0, + "completions/mean_length": 1039.92529296875, + "completions/mean_terminated_length": 625.4765625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 8.988921282798835, + "grad_norm": 0.165950745344162, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 572544785.0, + "reward": 0.6484375, + "reward_std": 0.176642507314682, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 962 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.13920454545454541, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2209.0, + "completions/mean_length": 1126.8948974609375, + "completions/mean_terminated_length": 646.7426147460938, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 8.998250728862974, + "grad_norm": 0.15967780351638794, + "learning_rate": 1e-06, + "loss": -0.0775, + "num_tokens": 573171529.0, + "reward": 0.5892857313156128, + "reward_std": 0.17990799248218536, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 963 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4012.0, + "completions/mean_length": 1102.005615234375, + "completions/mean_terminated_length": 678.6509399414062, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 9.00932944606414, + "grad_norm": 0.16359856724739075, + "learning_rate": 1e-06, + "loss": -0.0725, + "num_tokens": 573796478.0, + "reward": 0.640625, + "reward_std": 0.19925881922245026, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 964 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4030.0, + "completions/mean_length": 991.5245971679688, + "completions/mean_terminated_length": 566.0380859375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.018658892128279, + "grad_norm": 0.1571071892976761, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 574328500.0, + "reward": 0.7020089626312256, + "reward_std": 0.1351197212934494, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 965 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3454.0, + "completions/mean_length": 953.536865234375, + "completions/mean_terminated_length": 589.5902709960938, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 9.02798833819242, + "grad_norm": 0.15737630426883698, + "learning_rate": 1e-06, + "loss": -0.0655, + "num_tokens": 574898613.0, + "reward": 0.6863839626312256, + "reward_std": 0.14931826293468475, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.4642214775085449, + "step": 966 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2875.0, + "completions/mean_length": 1113.2366943359375, + "completions/mean_terminated_length": 620.6345825195312, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 9.03731778425656, + "grad_norm": 0.1779172569513321, + "learning_rate": 1e-06, + "loss": -0.0645, + "num_tokens": 575466281.0, + "reward": 0.6272321939468384, + "reward_std": 0.14511913061141968, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 967 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3624.0, + "completions/mean_length": 1069.927490234375, + "completions/mean_terminated_length": 646.4312744140625, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.0466472303207, + "grad_norm": 0.17144325375556946, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 576080224.0, + "reward": 0.6071428656578064, + "reward_std": 0.14650921523571014, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 968 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3978.0, + "completions/mean_length": 974.1953735351562, + "completions/mean_terminated_length": 616.973876953125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 9.055976676384839, + "grad_norm": 0.14389780163764954, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 576660911.0, + "reward": 0.6986607313156128, + "reward_std": 0.1230994164943695, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 969 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3304.0, + "completions/mean_length": 925.0725708007812, + "completions/mean_terminated_length": 557.829345703125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 9.06530612244898, + "grad_norm": 0.16744780540466309, + "learning_rate": 1e-06, + "loss": -0.0357, + "num_tokens": 577193376.0, + "reward": 0.7020089626312256, + "reward_std": 0.14000959694385529, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 970 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2841.0, + "completions/mean_length": 1115.60498046875, + "completions/mean_terminated_length": 605.2366333007812, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 9.07463556851312, + "grad_norm": 0.17163673043251038, + "learning_rate": 1e-06, + "loss": -0.0742, + "num_tokens": 577748790.0, + "reward": 0.652901828289032, + "reward_std": 0.17093636095523834, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 971 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3152.0, + "completions/mean_length": 952.560302734375, + "completions/mean_terminated_length": 597.2149047851562, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 9.08396501457726, + "grad_norm": 0.14103926718235016, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 578312468.0, + "reward": 0.6975446939468384, + "reward_std": 0.14515303075313568, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2963.0, + "completions/mean_length": 976.27685546875, + "completions/mean_terminated_length": 627.9205932617188, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.093294460641399, + "grad_norm": 0.1296180635690689, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 578906596.0, + "reward": 0.6796875596046448, + "reward_std": 0.13906599581241608, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 973 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3989.0, + "completions/mean_length": 1201.048095703125, + "completions/mean_terminated_length": 678.5072631835938, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 9.102623906705539, + "grad_norm": 0.15290963649749756, + "learning_rate": 1e-06, + "loss": -0.1021, + "num_tokens": 579501695.0, + "reward": 0.6116071939468384, + "reward_std": 0.15466415882110596, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 974 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1752232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2475.0, + "completions/mean_length": 1211.4263916015625, + "completions/mean_terminated_length": 598.6007690429688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 9.11195335276968, + "grad_norm": 0.1727268248796463, + "learning_rate": 1e-06, + "loss": -0.0662, + "num_tokens": 580049573.0, + "reward": 0.5870535969734192, + "reward_std": 0.1580837368965149, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3260.0, + "completions/mean_length": 1119.37060546875, + "completions/mean_terminated_length": 586.7105712890625, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 9.12128279883382, + "grad_norm": 0.16328230500221252, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 580587649.0, + "reward": 0.606026828289032, + "reward_std": 0.16409479081630707, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 976 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3929.0, + "completions/mean_length": 1173.6607666015625, + "completions/mean_terminated_length": 668.75390625, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 9.130612244897959, + "grad_norm": 0.13676023483276367, + "learning_rate": 1e-06, + "loss": -0.0791, + "num_tokens": 581191529.0, + "reward": 0.6417410969734192, + "reward_std": 0.14740893244743347, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 977 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2863.0, + "completions/mean_length": 1130.85498046875, + "completions/mean_terminated_length": 613.9947509765625, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 9.139941690962099, + "grad_norm": 0.13505032658576965, + "learning_rate": 1e-06, + "loss": -0.0621, + "num_tokens": 581748775.0, + "reward": 0.6741071939468384, + "reward_std": 0.1155492514371872, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3442.0, + "completions/mean_length": 1083.33935546875, + "completions/mean_terminated_length": 621.9407958984375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 9.14927113702624, + "grad_norm": 0.16368120908737183, + "learning_rate": 1e-06, + "loss": -0.0812, + "num_tokens": 582326815.0, + "reward": 0.6651785969734192, + "reward_std": 0.14853617548942566, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2927.0, + "completions/mean_length": 1160.578125, + "completions/mean_terminated_length": 635.2921142578125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 9.15860058309038, + "grad_norm": 0.15550637245178223, + "learning_rate": 1e-06, + "loss": -0.096, + "num_tokens": 582900837.0, + "reward": 0.6584821939468384, + "reward_std": 0.18340173363685608, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 980 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3596.0, + "completions/mean_length": 1033.9765625, + "completions/mean_terminated_length": 605.4490966796875, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 9.167930029154519, + "grad_norm": 0.1382673978805542, + "learning_rate": 1e-06, + "loss": -0.0688, + "num_tokens": 583474144.0, + "reward": 0.6707589626312256, + "reward_std": 0.1295209527015686, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 981 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3985.0, + "completions/mean_length": 888.8114013671875, + "completions/mean_terminated_length": 617.0157470703125, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.177259475218658, + "grad_norm": 0.12372458726167679, + "learning_rate": 1e-06, + "loss": -0.0455, + "num_tokens": 584063255.0, + "reward": 0.7087053656578064, + "reward_std": 0.11016871780157089, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 982 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3634.0, + "completions/mean_length": 821.0078735351562, + "completions/mean_terminated_length": 606.8287963867188, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.186588921282798, + "grad_norm": 0.14347834885120392, + "learning_rate": 1e-06, + "loss": -0.0548, + "num_tokens": 584658950.0, + "reward": 0.7645089626312256, + "reward_std": 0.1400442123413086, + "rewards/verify_math_reward/mean": 0.7645089030265808, + "rewards/verify_math_reward/std": 0.42454230785369873, + "step": 983 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4046.0, + "completions/mean_length": 1115.6429443359375, + "completions/mean_terminated_length": 623.438232421875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 9.19591836734694, + "grad_norm": 0.15761616826057434, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 585228126.0, + "reward": 0.65625, + "reward_std": 0.13793987035751343, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 984 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3777.0, + "completions/mean_length": 876.5625610351562, + "completions/mean_terminated_length": 578.1755981445312, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 9.205247813411079, + "grad_norm": 0.14765629172325134, + "learning_rate": 1e-06, + "loss": -0.0527, + "num_tokens": 585773902.0, + "reward": 0.7276785969734192, + "reward_std": 0.12140623480081558, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2398.0, + "completions/mean_length": 843.2288208007812, + "completions/mean_terminated_length": 584.57470703125, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 9.214577259475218, + "grad_norm": 0.14843350648880005, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 586344683.0, + "reward": 0.7198660969734192, + "reward_std": 0.1426684856414795, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 986 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1707589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3150.0, + "completions/mean_length": 1231.532470703125, + "completions/mean_terminated_length": 641.6756591796875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.223906705539358, + "grad_norm": 0.13903850317001343, + "learning_rate": 1e-06, + "loss": -0.0881, + "num_tokens": 586920904.0, + "reward": 0.6283482313156128, + "reward_std": 0.12110385298728943, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 987 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3596.0, + "completions/mean_length": 1199.48779296875, + "completions/mean_terminated_length": 707.9125366210938, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 9.2332361516035, + "grad_norm": 0.16974543035030365, + "learning_rate": 1e-06, + "loss": -0.0869, + "num_tokens": 587563045.0, + "reward": 0.609375, + "reward_std": 0.17092426121234894, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 988 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3065.0, + "completions/mean_length": 921.8638916015625, + "completions/mean_terminated_length": 571.80419921875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.242565597667639, + "grad_norm": 0.1629951149225235, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 588108123.0, + "reward": 0.7064732313156128, + "reward_std": 0.14515303075313568, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 989 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3483.0, + "completions/mean_length": 1007.0614013671875, + "completions/mean_terminated_length": 601.4431762695312, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 9.251895043731778, + "grad_norm": 0.18752329051494598, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 588681850.0, + "reward": 0.6283482313156128, + "reward_std": 0.1760030835866928, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 990 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2605.0, + "completions/mean_length": 895.9207763671875, + "completions/mean_terminated_length": 612.0741577148438, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 9.261224489795918, + "grad_norm": 0.15491515398025513, + "learning_rate": 1e-06, + "loss": -0.0272, + "num_tokens": 589287715.0, + "reward": 0.625, + "reward_std": 0.13752686977386475, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 991 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3858.0, + "completions/mean_length": 1092.0435791015625, + "completions/mean_terminated_length": 645.3013305664062, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 9.270553935860057, + "grad_norm": 0.14504826068878174, + "learning_rate": 1e-06, + "loss": -0.0637, + "num_tokens": 589877386.0, + "reward": 0.613839328289032, + "reward_std": 0.14124701917171478, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 992 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3680.0, + "completions/mean_length": 1050.2991943359375, + "completions/mean_terminated_length": 637.2572631835938, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.279883381924199, + "grad_norm": 0.14347554743289948, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 590479390.0, + "reward": 0.6718750596046448, + "reward_std": 0.13282990455627441, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 993 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4040.0, + "completions/mean_length": 1054.1585693359375, + "completions/mean_terminated_length": 632.8614501953125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.289212827988338, + "grad_norm": 0.1658715307712555, + "learning_rate": 1e-06, + "loss": -0.0838, + "num_tokens": 591060628.0, + "reward": 0.6941964626312256, + "reward_std": 0.17453815042972565, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3667.0, + "completions/mean_length": 986.4922485351562, + "completions/mean_terminated_length": 613.3512573242188, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 9.298542274052478, + "grad_norm": 0.14837861061096191, + "learning_rate": 1e-06, + "loss": -0.053, + "num_tokens": 591643853.0, + "reward": 0.6941964626312256, + "reward_std": 0.15627038478851318, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 995 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3867.0, + "completions/mean_length": 1003.747802734375, + "completions/mean_terminated_length": 597.6944580078125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 9.307871720116617, + "grad_norm": 0.15311138331890106, + "learning_rate": 1e-06, + "loss": -0.0612, + "num_tokens": 592209667.0, + "reward": 0.6852678656578064, + "reward_std": 0.11239181458950043, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.4646684527397156, + "step": 996 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3719.0, + "completions/mean_length": 988.87841796875, + "completions/mean_terminated_length": 616.0237426757812, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.317201166180759, + "grad_norm": 0.18904991447925568, + "learning_rate": 1e-06, + "loss": -0.07, + "num_tokens": 592789814.0, + "reward": 0.6774553656578064, + "reward_std": 0.19340254366397858, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 997 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2956.0, + "completions/mean_length": 894.3058471679688, + "completions/mean_terminated_length": 614.546142578125, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 9.326530612244898, + "grad_norm": 0.15684527158737183, + "learning_rate": 1e-06, + "loss": -0.0193, + "num_tokens": 593373912.0, + "reward": 0.6819196939468384, + "reward_std": 0.15488353371620178, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 998 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3115.0, + "completions/mean_length": 933.6920166015625, + "completions/mean_terminated_length": 554.2149658203125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 9.335860058309038, + "grad_norm": 0.17012563347816467, + "learning_rate": 1e-06, + "loss": -0.0685, + "num_tokens": 593908660.0, + "reward": 0.6696428656578064, + "reward_std": 0.14278656244277954, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 999 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3339.0, + "completions/mean_length": 1029.54248046875, + "completions/mean_terminated_length": 622.4905395507812, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 9.345189504373177, + "grad_norm": 0.15460209548473358, + "learning_rate": 1e-06, + "loss": -0.0543, + "num_tokens": 594495770.0, + "reward": 0.6819196939468384, + "reward_std": 0.16195543110370636, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1000 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2648.0, + "completions/mean_length": 1158.1920166015625, + "completions/mean_terminated_length": 600.2815551757812, + "completions/min_length": 192.0, + "completions/min_terminated_length": 192.0, + "epoch": 9.354518950437317, + "grad_norm": 0.17640163004398346, + "learning_rate": 1e-06, + "loss": -0.0804, + "num_tokens": 595041630.0, + "reward": 0.6484375, + "reward_std": 0.16792230308055878, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3498.0, + "completions/mean_length": 1013.5803833007812, + "completions/mean_terminated_length": 550.6239013671875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 9.363848396501458, + "grad_norm": 0.13645939528942108, + "learning_rate": 1e-06, + "loss": -0.0608, + "num_tokens": 595553270.0, + "reward": 0.6495535969734192, + "reward_std": 0.10674984753131866, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1002 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3853.0, + "completions/mean_length": 1057.7310791015625, + "completions/mean_terminated_length": 628.1159057617188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 9.373177842565598, + "grad_norm": 0.15502622723579407, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 596134029.0, + "reward": 0.660714328289032, + "reward_std": 0.15533816814422607, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1003 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2907.0, + "completions/mean_length": 1050.0491943359375, + "completions/mean_terminated_length": 619.3477783203125, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 9.382507288629737, + "grad_norm": 0.14409136772155762, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 596711673.0, + "reward": 0.6350446939468384, + "reward_std": 0.1320810616016388, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 1004 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 1131.26123046875, + "completions/mean_terminated_length": 609.9028930664062, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 9.391836734693877, + "grad_norm": 0.18565106391906738, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 597261451.0, + "reward": 0.6640625, + "reward_std": 0.12854525446891785, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1005 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 1149.421875, + "completions/mean_terminated_length": 644.844482421875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 9.401166180758018, + "grad_norm": 0.14761878550052643, + "learning_rate": 1e-06, + "loss": -0.0688, + "num_tokens": 597843365.0, + "reward": 0.6328125, + "reward_std": 0.14530277252197266, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1006 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3793.0, + "completions/mean_length": 1111.69873046875, + "completions/mean_terminated_length": 586.9002685546875, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 9.410495626822158, + "grad_norm": 0.1614135056734085, + "learning_rate": 1e-06, + "loss": -0.0922, + "num_tokens": 598379031.0, + "reward": 0.6674107313156128, + "reward_std": 0.1623997986316681, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1007 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3697.0, + "completions/mean_length": 1044.341552734375, + "completions/mean_terminated_length": 626.0938720703125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.419825072886297, + "grad_norm": 0.1610298752784729, + "learning_rate": 1e-06, + "loss": -0.0442, + "num_tokens": 598960977.0, + "reward": 0.6272321939468384, + "reward_std": 0.15575116872787476, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111400604248, + "step": 1008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3849.0, + "completions/mean_length": 1042.302490234375, + "completions/mean_terminated_length": 658.672119140625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.429154518950437, + "grad_norm": 0.14980989694595337, + "learning_rate": 1e-06, + "loss": -0.0793, + "num_tokens": 599573112.0, + "reward": 0.6752232313156128, + "reward_std": 0.15762656927108765, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1009 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3379.0, + "completions/mean_length": 1104.771240234375, + "completions/mean_terminated_length": 583.3643798828125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.438483965014576, + "grad_norm": 0.1702858805656433, + "learning_rate": 1e-06, + "loss": -0.0968, + "num_tokens": 600115499.0, + "reward": 0.6618303656578064, + "reward_std": 0.13275253772735596, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1010 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3687.0, + "completions/mean_length": 1051.2366943359375, + "completions/mean_terminated_length": 598.4256591796875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.447813411078718, + "grad_norm": 0.14281457662582397, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 600672759.0, + "reward": 0.7198660969734192, + "reward_std": 0.13165919482707977, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2736.0, + "completions/mean_length": 1057.391845703125, + "completions/mean_terminated_length": 560.1649169921875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 9.457142857142857, + "grad_norm": 0.16530536115169525, + "learning_rate": 1e-06, + "loss": -0.0675, + "num_tokens": 601193822.0, + "reward": 0.6875000596046448, + "reward_std": 0.138991117477417, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1012 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2627.0, + "completions/mean_length": 1027.8248291015625, + "completions/mean_terminated_length": 607.3134155273438, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.466472303206997, + "grad_norm": 0.1515965461730957, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 601779225.0, + "reward": 0.6261160969734192, + "reward_std": 0.14034590125083923, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 1013 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3631.0, + "completions/mean_length": 1181.1585693359375, + "completions/mean_terminated_length": 682.0156860351562, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 9.475801749271136, + "grad_norm": 0.17264242470264435, + "learning_rate": 1e-06, + "loss": -0.0579, + "num_tokens": 602384447.0, + "reward": 0.6328125, + "reward_std": 0.15173636376857758, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1014 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1741071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3050.0, + "completions/mean_length": 1286.0614013671875, + "completions/mean_terminated_length": 693.6959838867188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.485131195335278, + "grad_norm": 0.14848469197750092, + "learning_rate": 1e-06, + "loss": -0.0745, + "num_tokens": 602987870.0, + "reward": 0.566964328289032, + "reward_std": 0.17021210491657257, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 1015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3524.0, + "completions/mean_length": 978.919677734375, + "completions/mean_terminated_length": 560.678466796875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 9.494460641399417, + "grad_norm": 0.16232028603553772, + "learning_rate": 1e-06, + "loss": -0.0506, + "num_tokens": 603510878.0, + "reward": 0.7098214626312256, + "reward_std": 0.12456366419792175, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1016 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3087.0, + "completions/mean_length": 1042.9007568359375, + "completions/mean_terminated_length": 597.8197021484375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 9.503790087463557, + "grad_norm": 0.14280246198177338, + "learning_rate": 1e-06, + "loss": -0.09, + "num_tokens": 604060685.0, + "reward": 0.6941964626312256, + "reward_std": 0.1500004231929779, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.46100425720214844, + "step": 1017 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3067.0, + "completions/mean_length": 1077.083740234375, + "completions/mean_terminated_length": 619.2017822265625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 9.513119533527696, + "grad_norm": 0.18426713347434998, + "learning_rate": 1e-06, + "loss": -0.0962, + "num_tokens": 604635816.0, + "reward": 0.6863839626312256, + "reward_std": 0.19317500293254852, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1018 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3407.0, + "completions/mean_length": 1048.7935791015625, + "completions/mean_terminated_length": 631.1560668945312, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 9.522448979591836, + "grad_norm": 0.1455802619457245, + "learning_rate": 1e-06, + "loss": -0.1028, + "num_tokens": 605214367.0, + "reward": 0.7265625596046448, + "reward_std": 0.15544624626636505, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1019 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1674107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3383.0, + "completions/mean_length": 1210.009033203125, + "completions/mean_terminated_length": 629.7158203125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 9.531778425655977, + "grad_norm": 0.16423115134239197, + "learning_rate": 1e-06, + "loss": -0.0934, + "num_tokens": 605768983.0, + "reward": 0.637276828289032, + "reward_std": 0.1583772599697113, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1020 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3906.0, + "completions/mean_length": 1084.4453125, + "completions/mean_terminated_length": 600.7240600585938, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 9.541107871720117, + "grad_norm": 0.1674438863992691, + "learning_rate": 1e-06, + "loss": -0.061, + "num_tokens": 606320534.0, + "reward": 0.6361607313156128, + "reward_std": 0.14966411888599396, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1021 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2727.0, + "completions/mean_length": 1050.6875, + "completions/mean_terminated_length": 602.2739868164062, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 9.550437317784256, + "grad_norm": 0.1553095281124115, + "learning_rate": 1e-06, + "loss": -0.0832, + "num_tokens": 606879062.0, + "reward": 0.6863839626312256, + "reward_std": 0.14794068038463593, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1022 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3367.0, + "completions/mean_length": 1199.6015625, + "completions/mean_terminated_length": 712.4628295898438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.559766763848396, + "grad_norm": 0.21782627701759338, + "learning_rate": 1e-06, + "loss": -0.083, + "num_tokens": 607526665.0, + "reward": 0.6149553656578064, + "reward_std": 0.1810377687215805, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 1023 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3037.0, + "completions/mean_length": 1039.290283203125, + "completions/mean_terminated_length": 607.0675048828125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 9.569096209912537, + "grad_norm": 0.16398470103740692, + "learning_rate": 1e-06, + "loss": -0.0618, + "num_tokens": 608102581.0, + "reward": 0.684151828289032, + "reward_std": 0.1660016030073166, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1024 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3717.0, + "completions/mean_length": 1067.0982666015625, + "completions/mean_terminated_length": 634.39794921875, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 9.578425655976677, + "grad_norm": 0.15100041031837463, + "learning_rate": 1e-06, + "loss": -0.081, + "num_tokens": 608683053.0, + "reward": 0.6819196939468384, + "reward_std": 0.16345147788524628, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1025 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3825.0, + "completions/mean_length": 1058.5067138671875, + "completions/mean_terminated_length": 556.8660278320312, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 9.587755102040816, + "grad_norm": 0.15612345933914185, + "learning_rate": 1e-06, + "loss": -0.0596, + "num_tokens": 609208563.0, + "reward": 0.6830357313156128, + "reward_std": 0.10941985249519348, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1026 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2794.0, + "completions/mean_length": 992.0067138671875, + "completions/mean_terminated_length": 579.9722290039062, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 9.597084548104956, + "grad_norm": 0.12874767184257507, + "learning_rate": 1e-06, + "loss": -0.0399, + "num_tokens": 609753305.0, + "reward": 0.7064732313156128, + "reward_std": 0.09198792278766632, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1027 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2869.0, + "completions/mean_length": 1163.48779296875, + "completions/mean_terminated_length": 683.6220703125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 9.606413994169095, + "grad_norm": 0.15402068197727203, + "learning_rate": 1e-06, + "loss": -0.0521, + "num_tokens": 610374126.0, + "reward": 0.6026785969734192, + "reward_std": 0.16322575509548187, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 1028 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2769.0, + "completions/mean_length": 997.380615234375, + "completions/mean_terminated_length": 563.7315673828125, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 9.615743440233237, + "grad_norm": 0.1549675464630127, + "learning_rate": 1e-06, + "loss": -0.0643, + "num_tokens": 610911987.0, + "reward": 0.6718750596046448, + "reward_std": 0.12598443031311035, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1029 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2942.0, + "completions/mean_length": 893.5960083007812, + "completions/mean_terminated_length": 553.5877075195312, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 9.625072886297376, + "grad_norm": 0.13310593366622925, + "learning_rate": 1e-06, + "loss": -0.0605, + "num_tokens": 611445745.0, + "reward": 0.7410714626312256, + "reward_std": 0.12392427772283554, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1030 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3800.0, + "completions/mean_length": 1061.96875, + "completions/mean_terminated_length": 632.952880859375, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 9.634402332361516, + "grad_norm": 0.13584552705287933, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 612032277.0, + "reward": 0.6495535969734192, + "reward_std": 0.13087712228298187, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1031 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3988.0, + "completions/mean_length": 1138.1976318359375, + "completions/mean_terminated_length": 663.110107421875, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 9.643731778425655, + "grad_norm": 0.14727799594402313, + "learning_rate": 1e-06, + "loss": -0.0772, + "num_tokens": 612638438.0, + "reward": 0.6417410969734192, + "reward_std": 0.15541306138038635, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1032 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2904.0, + "completions/mean_length": 1021.9542846679688, + "completions/mean_terminated_length": 582.8048095703125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.653061224489797, + "grad_norm": 0.1584702730178833, + "learning_rate": 1e-06, + "loss": -0.0471, + "num_tokens": 613179301.0, + "reward": 0.6975446939468384, + "reward_std": 0.13898929953575134, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1033 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3012.0, + "completions/mean_length": 1061.3382568359375, + "completions/mean_terminated_length": 632.233154296875, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 9.662390670553936, + "grad_norm": 0.13369698822498322, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 613765228.0, + "reward": 0.6674107313156128, + "reward_std": 0.11678596585988998, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1034 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3916.0, + "completions/mean_length": 961.5881958007812, + "completions/mean_terminated_length": 598.5740966796875, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 9.671720116618076, + "grad_norm": 0.22063378989696503, + "learning_rate": 1e-06, + "loss": -0.0745, + "num_tokens": 614330995.0, + "reward": 0.6573660969734192, + "reward_std": 0.17472319304943085, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1035 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3534.0, + "completions/mean_length": 973.200927734375, + "completions/mean_terminated_length": 585.2998657226562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 9.681049562682215, + "grad_norm": 0.16207484900951385, + "learning_rate": 1e-06, + "loss": -0.0896, + "num_tokens": 614886471.0, + "reward": 0.691964328289032, + "reward_std": 0.16322465240955353, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1036 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3137.0, + "completions/mean_length": 994.1082763671875, + "completions/mean_terminated_length": 613.1741943359375, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 9.690379008746355, + "grad_norm": 0.1522195190191269, + "learning_rate": 1e-06, + "loss": -0.0446, + "num_tokens": 615455056.0, + "reward": 0.7131696939468384, + "reward_std": 0.13594317436218262, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1037 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2599.0, + "completions/mean_length": 1052.138427734375, + "completions/mean_terminated_length": 656.7818603515625, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 9.699708454810496, + "grad_norm": 0.15572939813137054, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 616059548.0, + "reward": 0.6361607313156128, + "reward_std": 0.17472107708454132, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1038 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3026.0, + "completions/mean_length": 975.255615234375, + "completions/mean_terminated_length": 600.7662353515625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 9.709037900874636, + "grad_norm": 0.16639718413352966, + "learning_rate": 1e-06, + "loss": -0.048, + "num_tokens": 616629473.0, + "reward": 0.6662946939468384, + "reward_std": 0.16142340004444122, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1039 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1618303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3305.0, + "completions/mean_length": 1189.6629638671875, + "completions/mean_terminated_length": 628.519287109375, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.718367346938775, + "grad_norm": 0.16171219944953918, + "learning_rate": 1e-06, + "loss": -0.0977, + "num_tokens": 617186651.0, + "reward": 0.6205357313156128, + "reward_std": 0.15300559997558594, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1040 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3864.0, + "completions/mean_length": 905.7723388671875, + "completions/mean_terminated_length": 567.0568237304688, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.727696793002915, + "grad_norm": 0.17202991247177124, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 617736527.0, + "reward": 0.707589328289032, + "reward_std": 0.15364499390125275, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2718.0, + "completions/mean_length": 917.5982666015625, + "completions/mean_terminated_length": 601.7079467773438, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.737026239067056, + "grad_norm": 0.13633190095424652, + "learning_rate": 1e-06, + "loss": -0.0727, + "num_tokens": 618309303.0, + "reward": 0.7399553656578064, + "reward_std": 0.12828421592712402, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 1042 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3423.0, + "completions/mean_length": 1003.286865234375, + "completions/mean_terminated_length": 597.1729736328125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 9.746355685131196, + "grad_norm": 0.1549861878156662, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 618869192.0, + "reward": 0.6551339626312256, + "reward_std": 0.11802449822425842, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 1043 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3606.0, + "completions/mean_length": 1103.618408203125, + "completions/mean_terminated_length": 618.4721069335938, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.755685131195335, + "grad_norm": 0.1599438339471817, + "learning_rate": 1e-06, + "loss": -0.0673, + "num_tokens": 619431762.0, + "reward": 0.6796875596046448, + "reward_std": 0.14135761559009552, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1044 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3534.0, + "completions/mean_length": 937.62060546875, + "completions/mean_terminated_length": 597.9678344726562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 9.765014577259475, + "grad_norm": 0.1297610104084015, + "learning_rate": 1e-06, + "loss": -0.0539, + "num_tokens": 619990486.0, + "reward": 0.7187500596046448, + "reward_std": 0.11633063107728958, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1045 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3036.0, + "completions/mean_length": 1094.70654296875, + "completions/mean_terminated_length": 692.0012817382812, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 9.774344023323614, + "grad_norm": 0.14316801726818085, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 620631463.0, + "reward": 0.6272321939468384, + "reward_std": 0.15158483386039734, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 1046 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2983.0, + "completions/mean_length": 1160.3671875, + "completions/mean_terminated_length": 648.6514282226562, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 9.783673469387756, + "grad_norm": 0.14425425231456757, + "learning_rate": 1e-06, + "loss": -0.0733, + "num_tokens": 621223792.0, + "reward": 0.613839328289032, + "reward_std": 0.1328285187482834, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 1047 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3578.0, + "completions/mean_length": 1061.8538818359375, + "completions/mean_terminated_length": 623.9757080078125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.793002915451895, + "grad_norm": 0.1334461271762848, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 621805789.0, + "reward": 0.668526828289032, + "reward_std": 0.11396414041519165, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1048 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4072.0, + "completions/mean_length": 974.6652221679688, + "completions/mean_terminated_length": 630.4287719726562, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 9.802332361516035, + "grad_norm": 0.17007985711097717, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 622402385.0, + "reward": 0.645089328289032, + "reward_std": 0.16999317705631256, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1049 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2899.0, + "completions/mean_length": 965.6842041015625, + "completions/mean_terminated_length": 594.4232177734375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.811661807580174, + "grad_norm": 0.16194379329681396, + "learning_rate": 1e-06, + "loss": -0.0961, + "num_tokens": 622966358.0, + "reward": 0.7220982313156128, + "reward_std": 0.16258594393730164, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1050 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3583.0, + "completions/mean_length": 1048.3148193359375, + "completions/mean_terminated_length": 643.7547607421875, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 9.820991253644316, + "grad_norm": 0.1667070984840393, + "learning_rate": 1e-06, + "loss": -0.076, + "num_tokens": 623556136.0, + "reward": 0.6852678656578064, + "reward_std": 0.16488321125507355, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1051 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3350.0, + "completions/mean_length": 1208.4632568359375, + "completions/mean_terminated_length": 687.2608642578125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 9.830320699708455, + "grad_norm": 0.14641916751861572, + "learning_rate": 1e-06, + "loss": -0.0464, + "num_tokens": 624172015.0, + "reward": 0.613839328289032, + "reward_std": 0.12463782727718353, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 1052 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3434.0, + "completions/mean_length": 1022.7756958007812, + "completions/mean_terminated_length": 614.8255615234375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 9.839650145772595, + "grad_norm": 0.17525076866149902, + "learning_rate": 1e-06, + "loss": -0.0656, + "num_tokens": 624741182.0, + "reward": 0.6662946939468384, + "reward_std": 0.1623249500989914, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1053 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 1020.2098388671875, + "completions/mean_terminated_length": 620.7061767578125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 9.848979591836734, + "grad_norm": 0.15761888027191162, + "learning_rate": 1e-06, + "loss": -0.0201, + "num_tokens": 625323458.0, + "reward": 0.625, + "reward_std": 0.1371150016784668, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 1054 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2466.0, + "completions/mean_length": 1079.146240234375, + "completions/mean_terminated_length": 580.9141845703125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 9.858309037900874, + "grad_norm": 0.1381913125514984, + "learning_rate": 1e-06, + "loss": -0.0913, + "num_tokens": 625872333.0, + "reward": 0.6696428656578064, + "reward_std": 0.12065806984901428, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1055 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3856.0, + "completions/mean_length": 1147.857177734375, + "completions/mean_terminated_length": 687.5664672851562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 9.867638483965015, + "grad_norm": 0.15661345422267914, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 626499405.0, + "reward": 0.6160714626312256, + "reward_std": 0.16660849750041962, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 1056 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3863.0, + "completions/mean_length": 1091.1640625, + "completions/mean_terminated_length": 617.5336303710938, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.876967930029155, + "grad_norm": 0.1464555263519287, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 627064480.0, + "reward": 0.6439732313156128, + "reward_std": 0.13659143447875977, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1057 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3533.0, + "completions/mean_length": 1068.14404296875, + "completions/mean_terminated_length": 626.7429809570312, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 9.886297376093294, + "grad_norm": 0.1781105101108551, + "learning_rate": 1e-06, + "loss": -0.0653, + "num_tokens": 627648441.0, + "reward": 0.645089328289032, + "reward_std": 0.18809868395328522, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1058 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3707.0, + "completions/mean_length": 1056.8482666015625, + "completions/mean_terminated_length": 657.7677001953125, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 9.895626822157434, + "grad_norm": 0.17310470342636108, + "learning_rate": 1e-06, + "loss": -0.0306, + "num_tokens": 628255905.0, + "reward": 0.6205357313156128, + "reward_std": 0.17442826926708221, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1059 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3609.0, + "completions/mean_length": 933.810302734375, + "completions/mean_terminated_length": 567.5790405273438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.904956268221575, + "grad_norm": 0.15535572171211243, + "learning_rate": 1e-06, + "loss": -0.0256, + "num_tokens": 628797855.0, + "reward": 0.6819196939468384, + "reward_std": 0.10765069723129272, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1060 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 802.2623291015625, + "completions/mean_terminated_length": 518.8012084960938, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.914285714285715, + "grad_norm": 0.15084369480609894, + "learning_rate": 1e-06, + "loss": -0.0349, + "num_tokens": 629306490.0, + "reward": 0.746651828289032, + "reward_std": 0.11866138130426407, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1061 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3356.0, + "completions/mean_length": 957.2991333007812, + "completions/mean_terminated_length": 585.043701171875, + "completions/min_length": 201.0, + "completions/min_terminated_length": 201.0, + "epoch": 9.923615160349854, + "grad_norm": 0.15866389870643616, + "learning_rate": 1e-06, + "loss": -0.0201, + "num_tokens": 629863966.0, + "reward": 0.684151828289032, + "reward_std": 0.14334996044635773, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1062 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3786.0, + "completions/mean_length": 965.552490234375, + "completions/mean_terminated_length": 545.5177001953125, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 9.932944606413994, + "grad_norm": 0.1854659765958786, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 630390525.0, + "reward": 0.6707589626312256, + "reward_std": 0.12159238755702972, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1063 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3194.0, + "completions/mean_length": 1046.087158203125, + "completions/mean_terminated_length": 680.0974731445312, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 9.942274052478133, + "grad_norm": 0.16189289093017578, + "learning_rate": 1e-06, + "loss": -0.0637, + "num_tokens": 631014443.0, + "reward": 0.6986607313156128, + "reward_std": 0.14425332844257355, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1064 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2940.0, + "completions/mean_length": 1046.8192138671875, + "completions/mean_terminated_length": 620.0890502929688, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.951603498542275, + "grad_norm": 0.12819170951843262, + "learning_rate": 1e-06, + "loss": -0.0706, + "num_tokens": 631593737.0, + "reward": 0.684151828289032, + "reward_std": 0.12114479392766953, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1065 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2902.0, + "completions/mean_length": 885.2366333007812, + "completions/mean_terminated_length": 617.3494873046875, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 9.960932944606414, + "grad_norm": 0.15327103435993195, + "learning_rate": 1e-06, + "loss": -0.0325, + "num_tokens": 632195461.0, + "reward": 0.6729910969734192, + "reward_std": 0.13478201627731323, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1066 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3972.0, + "completions/mean_length": 1029.7734375, + "completions/mean_terminated_length": 596.205078125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 9.970262390670554, + "grad_norm": 0.17385685443878174, + "learning_rate": 1e-06, + "loss": -0.0797, + "num_tokens": 632756266.0, + "reward": 0.6774553656578064, + "reward_std": 0.16517673432826996, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1067 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2610.0, + "completions/mean_length": 902.9777221679688, + "completions/mean_terminated_length": 585.6343383789062, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 9.979591836734693, + "grad_norm": 0.13518302142620087, + "learning_rate": 1e-06, + "loss": -0.0504, + "num_tokens": 633312502.0, + "reward": 0.7265625596046448, + "reward_std": 0.1310618817806244, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1068 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3777.0, + "completions/mean_length": 1088.716552734375, + "completions/mean_terminated_length": 641.4794921875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 9.988921282798835, + "grad_norm": 0.16097402572631836, + "learning_rate": 1e-06, + "loss": -0.0265, + "num_tokens": 633916944.0, + "reward": 0.609375, + "reward_std": 0.13778649270534515, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 1069 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.20170454545454541, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2982.0, + "completions/mean_length": 1306.32958984375, + "completions/mean_terminated_length": 601.4661865234375, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 9.998250728862974, + "grad_norm": 0.15865516662597656, + "learning_rate": 1e-06, + "loss": -0.0683, + "num_tokens": 634462382.0, + "reward": 0.6986607313156128, + "reward_std": 0.14049884676933289, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1070 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3132.0, + "completions/mean_length": 1106.2723388671875, + "completions/mean_terminated_length": 589.7225341796875, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 10.00932944606414, + "grad_norm": 0.16653719544410706, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 635005434.0, + "reward": 0.637276828289032, + "reward_std": 0.12805670499801636, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1071 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3649.0, + "completions/mean_length": 1048.40966796875, + "completions/mean_terminated_length": 608.59130859375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 10.018658892128279, + "grad_norm": 0.16584089398384094, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 635575369.0, + "reward": 0.6227678656578064, + "reward_std": 0.1292620301246643, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 1072 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2698.0, + "completions/mean_length": 1080.5513916015625, + "completions/mean_terminated_length": 591.6653442382812, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 10.02798833819242, + "grad_norm": 0.13904087245464325, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 636122471.0, + "reward": 0.6551339626312256, + "reward_std": 0.1209183931350708, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 1073 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2482.0, + "completions/mean_length": 1050.6763916015625, + "completions/mean_terminated_length": 597.7820434570312, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 10.03731778425656, + "grad_norm": 0.17358961701393127, + "learning_rate": 1e-06, + "loss": -0.0882, + "num_tokens": 636674693.0, + "reward": 0.691964328289032, + "reward_std": 0.15431898832321167, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1074 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2720.0, + "completions/mean_length": 1021.5770263671875, + "completions/mean_terminated_length": 635.3429565429688, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 10.0466472303207, + "grad_norm": 0.1577582210302353, + "learning_rate": 1e-06, + "loss": -0.0744, + "num_tokens": 637275858.0, + "reward": 0.660714328289032, + "reward_std": 0.1511615812778473, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1075 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3989.0, + "completions/mean_length": 890.2266235351562, + "completions/mean_terminated_length": 593.1060791015625, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 10.055976676384839, + "grad_norm": 0.16038629412651062, + "learning_rate": 1e-06, + "loss": -0.0682, + "num_tokens": 637838781.0, + "reward": 0.7511160969734192, + "reward_std": 0.1537223756313324, + "rewards/verify_math_reward/mean": 0.7511160969734192, + "rewards/verify_math_reward/std": 0.43260788917541504, + "step": 1076 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3541.0, + "completions/mean_length": 1025.3170166015625, + "completions/mean_terminated_length": 600.025390625, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 10.06530612244898, + "grad_norm": 0.13877595961093903, + "learning_rate": 1e-06, + "loss": -0.0627, + "num_tokens": 638403345.0, + "reward": 0.6729910969734192, + "reward_std": 0.11501792073249817, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1077 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3632.0, + "completions/mean_length": 1067.9029541015625, + "completions/mean_terminated_length": 639.7261352539062, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 10.07463556851312, + "grad_norm": 0.18216530978679657, + "learning_rate": 1e-06, + "loss": -0.0635, + "num_tokens": 638993010.0, + "reward": 0.6350446939468384, + "reward_std": 0.19501762092113495, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 1078 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3097.0, + "completions/mean_length": 1112.5670166015625, + "completions/mean_terminated_length": 619.8543701171875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "epoch": 10.08396501457726, + "grad_norm": 0.1633872091770172, + "learning_rate": 1e-06, + "loss": -0.0675, + "num_tokens": 639572462.0, + "reward": 0.6082589626312256, + "reward_std": 0.16435188055038452, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.4884119927883148, + "step": 1079 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3079.0, + "completions/mean_length": 1025.5592041015625, + "completions/mean_terminated_length": 644.161865234375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 10.093294460641399, + "grad_norm": 0.14698049426078796, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 640187731.0, + "reward": 0.6819196939468384, + "reward_std": 0.1504882574081421, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1080 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3598.0, + "completions/mean_length": 1014.6506958007812, + "completions/mean_terminated_length": 596.7744140625, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 10.102623906705539, + "grad_norm": 0.1668751984834671, + "learning_rate": 1e-06, + "loss": -0.0765, + "num_tokens": 640751250.0, + "reward": 0.6473214626312256, + "reward_std": 0.1709691435098648, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 1081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3775.0, + "completions/mean_length": 932.28466796875, + "completions/mean_terminated_length": 596.3839721679688, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 10.11195335276968, + "grad_norm": 0.15423645079135895, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 641329593.0, + "reward": 0.6718750596046448, + "reward_std": 0.14853869378566742, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1082 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2921.0, + "completions/mean_length": 969.3984985351562, + "completions/mean_terminated_length": 633.1631469726562, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 10.12128279883382, + "grad_norm": 0.19824273884296417, + "learning_rate": 1e-06, + "loss": -0.0364, + "num_tokens": 641941838.0, + "reward": 0.6651785969734192, + "reward_std": 0.13598595559597015, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1083 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2640.0, + "completions/mean_length": 1022.349365234375, + "completions/mean_terminated_length": 649.2027587890625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 10.130612244897959, + "grad_norm": 0.13495387136936188, + "learning_rate": 1e-06, + "loss": -0.0444, + "num_tokens": 642554991.0, + "reward": 0.6272321939468384, + "reward_std": 0.1317005753517151, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111400604248, + "step": 1084 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2905.0, + "completions/mean_length": 1028.685302734375, + "completions/mean_terminated_length": 643.34423828125, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 10.139941690962099, + "grad_norm": 0.1342616230249405, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 643150741.0, + "reward": 0.6484375, + "reward_std": 0.12828563153743744, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1085 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2998.0, + "completions/mean_length": 894.5692138671875, + "completions/mean_terminated_length": 519.3391723632812, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 10.14927113702624, + "grad_norm": 0.15185034275054932, + "learning_rate": 1e-06, + "loss": -0.059, + "num_tokens": 643650803.0, + "reward": 0.746651828289032, + "reward_std": 0.1302691102027893, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1086 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 1003.83935546875, + "completions/mean_terminated_length": 566.6038208007812, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 10.15860058309038, + "grad_norm": 0.14555291831493378, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 644190675.0, + "reward": 0.6908482313156128, + "reward_std": 0.09405060112476349, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1087 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2786.0, + "completions/mean_length": 1041.8035888671875, + "completions/mean_terminated_length": 605.4898071289062, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 10.167930029154519, + "grad_norm": 0.16963058710098267, + "learning_rate": 1e-06, + "loss": -0.0707, + "num_tokens": 644750403.0, + "reward": 0.707589328289032, + "reward_std": 0.16266053915023804, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3891.0, + "completions/mean_length": 1129.2957763671875, + "completions/mean_terminated_length": 661.6757202148438, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 10.177259475218658, + "grad_norm": 0.14441253244876862, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 645354100.0, + "reward": 0.629464328289032, + "reward_std": 0.11765290051698685, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 1089 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3521.0, + "completions/mean_length": 977.8449096679688, + "completions/mean_terminated_length": 550.4835205078125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 10.186588921282798, + "grad_norm": 0.16496771574020386, + "learning_rate": 1e-06, + "loss": -0.082, + "num_tokens": 645873305.0, + "reward": 0.7265625596046448, + "reward_std": 0.1352691650390625, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1090 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2541.0, + "completions/mean_length": 884.4777221679688, + "completions/mean_terminated_length": 539.1099853515625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.19591836734694, + "grad_norm": 0.15475855767726898, + "learning_rate": 1e-06, + "loss": -0.0388, + "num_tokens": 646393597.0, + "reward": 0.7421875596046448, + "reward_std": 0.13778719305992126, + "rewards/verify_math_reward/mean": 0.7421875, + "rewards/verify_math_reward/std": 0.43767455220222473, + "step": 1091 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.15625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3091.0, + "completions/mean_length": 1162.6640625, + "completions/mean_terminated_length": 619.4536743164062, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 10.205247813411079, + "grad_norm": 0.1447315216064453, + "learning_rate": 1e-06, + "loss": -0.0757, + "num_tokens": 646953800.0, + "reward": 0.6160714626312256, + "reward_std": 0.15139050781726837, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 1092 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3445.0, + "completions/mean_length": 1009.17529296875, + "completions/mean_terminated_length": 612.6309814453125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 10.214577259475218, + "grad_norm": 0.1439095288515091, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 647523421.0, + "reward": 0.6741071939468384, + "reward_std": 0.13662534952163696, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1093 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3994.0, + "completions/mean_length": 1105.2913818359375, + "completions/mean_terminated_length": 633.8876342773438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 10.223906705539358, + "grad_norm": 0.1669865995645523, + "learning_rate": 1e-06, + "loss": -0.0582, + "num_tokens": 648110634.0, + "reward": 0.637276828289032, + "reward_std": 0.16404879093170166, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1094 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2817.0, + "completions/mean_length": 954.7969360351562, + "completions/mean_terminated_length": 590.9962158203125, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 10.2332361516035, + "grad_norm": 0.16492746770381927, + "learning_rate": 1e-06, + "loss": -0.0515, + "num_tokens": 648669180.0, + "reward": 0.6595982313156128, + "reward_std": 0.15451402962207794, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3004.0, + "completions/mean_length": 1175.524658203125, + "completions/mean_terminated_length": 630.1112670898438, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 10.242565597667639, + "grad_norm": 0.15489059686660767, + "learning_rate": 1e-06, + "loss": -0.0853, + "num_tokens": 649235674.0, + "reward": 0.645089328289032, + "reward_std": 0.15364499390125275, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1096 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3364.0, + "completions/mean_length": 928.1116333007812, + "completions/mean_terminated_length": 570.0025024414062, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 10.251895043731778, + "grad_norm": 0.14972880482673645, + "learning_rate": 1e-06, + "loss": -0.0615, + "num_tokens": 649781006.0, + "reward": 0.7142857313156128, + "reward_std": 0.12159170210361481, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1097 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3936.0, + "completions/mean_length": 995.771240234375, + "completions/mean_terminated_length": 623.7437133789062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.261224489795918, + "grad_norm": 0.160934180021286, + "learning_rate": 1e-06, + "loss": -0.0506, + "num_tokens": 650366169.0, + "reward": 0.6662946939468384, + "reward_std": 0.1384686678647995, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1098 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2483.0, + "completions/mean_length": 956.0547485351562, + "completions/mean_terminated_length": 548.2181396484375, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.270553935860057, + "grad_norm": 0.13288824260234833, + "learning_rate": 1e-06, + "loss": -0.0556, + "num_tokens": 650894226.0, + "reward": 0.6729910969734192, + "reward_std": 0.1174592673778534, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1099 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4079.0, + "completions/mean_length": 1054.505615234375, + "completions/mean_terminated_length": 588.6911010742188, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 10.279883381924199, + "grad_norm": 0.17229117453098297, + "learning_rate": 1e-06, + "loss": -0.0707, + "num_tokens": 651453095.0, + "reward": 0.6183035969734192, + "reward_std": 0.15634846687316895, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 1100 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2727.0, + "completions/mean_length": 864.1250610351562, + "completions/mean_terminated_length": 573.1776123046875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 10.289212827988338, + "grad_norm": 0.1550755798816681, + "learning_rate": 1e-06, + "loss": -0.0309, + "num_tokens": 652014183.0, + "reward": 0.6852678656578064, + "reward_std": 0.13711389899253845, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1101 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3396.0, + "completions/mean_length": 1015.310302734375, + "completions/mean_terminated_length": 601.951904296875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 10.298542274052478, + "grad_norm": 0.14750759303569794, + "learning_rate": 1e-06, + "loss": -0.0716, + "num_tokens": 652582373.0, + "reward": 0.6473214626312256, + "reward_std": 0.14315634965896606, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 1102 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3603.0, + "completions/mean_length": 1084.325927734375, + "completions/mean_terminated_length": 600.58544921875, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.307871720116617, + "grad_norm": 0.15466561913490295, + "learning_rate": 1e-06, + "loss": -0.0515, + "num_tokens": 653139041.0, + "reward": 0.6506696939468384, + "reward_std": 0.12192869931459427, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1103 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3120.0, + "completions/mean_length": 1058.1373291015625, + "completions/mean_terminated_length": 641.7804565429688, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 10.317201166180759, + "grad_norm": 0.16274167597293854, + "learning_rate": 1e-06, + "loss": -0.0653, + "num_tokens": 653729108.0, + "reward": 0.6540178656578064, + "reward_std": 0.143612802028656, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3069.0, + "completions/mean_length": 982.8694458007812, + "completions/mean_terminated_length": 609.2937622070312, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 10.326530612244898, + "grad_norm": 0.15139319002628326, + "learning_rate": 1e-06, + "loss": -0.0676, + "num_tokens": 654301231.0, + "reward": 0.676339328289032, + "reward_std": 0.15680241584777832, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 1105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3535.0, + "completions/mean_length": 1012.7042846679688, + "completions/mean_terminated_length": 607.8270263671875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 10.335860058309038, + "grad_norm": 0.1497790366411209, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 654867406.0, + "reward": 0.6718750596046448, + "reward_std": 0.15364639461040497, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1106 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3005.0, + "completions/mean_length": 828.5245971679688, + "completions/mean_terminated_length": 534.3722534179688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.345189504373177, + "grad_norm": 0.18536463379859924, + "learning_rate": 1e-06, + "loss": -0.0568, + "num_tokens": 655393244.0, + "reward": 0.7031250596046448, + "reward_std": 0.14740821719169617, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2854.0, + "completions/mean_length": 915.2467041015625, + "completions/mean_terminated_length": 590.520263671875, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 10.354518950437317, + "grad_norm": 0.1505555510520935, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 655957249.0, + "reward": 0.699776828289032, + "reward_std": 0.14184364676475525, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3287.0, + "completions/mean_length": 875.2266235351562, + "completions/mean_terminated_length": 563.7931518554688, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "epoch": 10.363848396501458, + "grad_norm": 0.1695970594882965, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 656505260.0, + "reward": 0.7064732313156128, + "reward_std": 0.14680273830890656, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1109 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3396.0, + "completions/mean_length": 1052.5625, + "completions/mean_terminated_length": 586.450439453125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 10.373177842565598, + "grad_norm": 0.15126630663871765, + "learning_rate": 1e-06, + "loss": -0.0744, + "num_tokens": 657046028.0, + "reward": 0.6439732313156128, + "reward_std": 0.1327090561389923, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1110 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3211.0, + "completions/mean_length": 1058.46435546875, + "completions/mean_terminated_length": 593.256103515625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 10.382507288629737, + "grad_norm": 0.161564439535141, + "learning_rate": 1e-06, + "loss": -0.0309, + "num_tokens": 657580740.0, + "reward": 0.6573660969734192, + "reward_std": 0.12297855317592621, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3370.0, + "completions/mean_length": 937.779052734375, + "completions/mean_terminated_length": 563.20849609375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.391836734693877, + "grad_norm": 0.18076066672801971, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 658117390.0, + "reward": 0.7366071939468384, + "reward_std": 0.1413569152355194, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1796875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 1251.134033203125, + "completions/mean_terminated_length": 627.9727783203125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 10.401166180758018, + "grad_norm": 0.1642141193151474, + "learning_rate": 1e-06, + "loss": -0.1078, + "num_tokens": 658678702.0, + "reward": 0.6049107313156128, + "reward_std": 0.13989154994487762, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 1113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2833.0, + "completions/mean_length": 943.44873046875, + "completions/mean_terminated_length": 573.9476318359375, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.410495626822158, + "grad_norm": 0.1734590381383896, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 659233376.0, + "reward": 0.7031250596046448, + "reward_std": 0.13309067487716675, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1114 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3993.0, + "completions/mean_length": 1058.83935546875, + "completions/mean_terminated_length": 611.6261596679688, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 10.419825072886297, + "grad_norm": 0.15424303710460663, + "learning_rate": 1e-06, + "loss": -0.1034, + "num_tokens": 659790352.0, + "reward": 0.6863839626312256, + "reward_std": 0.1413230150938034, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.4642214775085449, + "step": 1115 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3394.0, + "completions/mean_length": 1113.982177734375, + "completions/mean_terminated_length": 603.3359375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 10.429154518950437, + "grad_norm": 0.16236351430416107, + "learning_rate": 1e-06, + "loss": -0.0853, + "num_tokens": 660344160.0, + "reward": 0.6428571939468384, + "reward_std": 0.16394071280956268, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1116 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2838.0, + "completions/mean_length": 960.5982666015625, + "completions/mean_terminated_length": 610.4913330078125, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 10.438483965014576, + "grad_norm": 0.17198672890663147, + "learning_rate": 1e-06, + "loss": -0.0646, + "num_tokens": 660916104.0, + "reward": 0.7008928656578064, + "reward_std": 0.15788941085338593, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 1117 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3317.0, + "completions/mean_length": 1087.7991943359375, + "completions/mean_terminated_length": 635.98974609375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 10.447813411078718, + "grad_norm": 0.1357061117887497, + "learning_rate": 1e-06, + "loss": -0.0705, + "num_tokens": 661502812.0, + "reward": 0.6540178656578064, + "reward_std": 0.12764301896095276, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2802.0, + "completions/mean_length": 1091.060302734375, + "completions/mean_terminated_length": 617.4134521484375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 10.457142857142857, + "grad_norm": 0.14067105948925018, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 662073842.0, + "reward": 0.6261160969734192, + "reward_std": 0.12215623259544373, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 1119 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1584821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2393.0, + "completions/mean_length": 1134.614990234375, + "completions/mean_terminated_length": 576.9005126953125, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 10.466472303206997, + "grad_norm": 0.15287235379219055, + "learning_rate": 1e-06, + "loss": -0.0683, + "num_tokens": 662600201.0, + "reward": 0.6796875596046448, + "reward_std": 0.1327204555273056, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1120 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3066.0, + "completions/mean_length": 1100.3226318359375, + "completions/mean_terminated_length": 614.6420288085938, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 10.475801749271136, + "grad_norm": 0.15658532083034515, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 663165066.0, + "reward": 0.6417410969734192, + "reward_std": 0.13711319863796234, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1121 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2467.0, + "completions/mean_length": 1028.891845703125, + "completions/mean_terminated_length": 527.0012817382812, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 10.485131195335278, + "grad_norm": 0.18758775293827057, + "learning_rate": 1e-06, + "loss": -0.0863, + "num_tokens": 663651489.0, + "reward": 0.7243303656578064, + "reward_std": 0.1420711725950241, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1122 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3326.0, + "completions/mean_length": 1023.7745971679688, + "completions/mean_terminated_length": 589.3579711914062, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 10.494460641399417, + "grad_norm": 0.15295983850955963, + "learning_rate": 1e-06, + "loss": -0.0891, + "num_tokens": 664198231.0, + "reward": 0.6707589626312256, + "reward_std": 0.12535572052001953, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1123 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3910.0, + "completions/mean_length": 996.1629638671875, + "completions/mean_terminated_length": 597.9470825195312, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 10.503790087463557, + "grad_norm": 0.15623658895492554, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 664758905.0, + "reward": 0.6517857313156128, + "reward_std": 0.1357584297657013, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1124 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3711.0, + "completions/mean_length": 940.65966796875, + "completions/mean_terminated_length": 548.7164306640625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 10.513119533527696, + "grad_norm": 0.17176030576229095, + "learning_rate": 1e-06, + "loss": -0.0639, + "num_tokens": 665276456.0, + "reward": 0.7053571939468384, + "reward_std": 0.15477405488491058, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1125 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3550.0, + "completions/mean_length": 1006.9029541015625, + "completions/mean_terminated_length": 547.4987182617188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 10.522448979591836, + "grad_norm": 0.17296266555786133, + "learning_rate": 1e-06, + "loss": -0.0639, + "num_tokens": 665809289.0, + "reward": 0.6819196939468384, + "reward_std": 0.14696383476257324, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1126 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2798.0, + "completions/mean_length": 958.154052734375, + "completions/mean_terminated_length": 537.1265869140625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 10.531778425655977, + "grad_norm": 0.17006491124629974, + "learning_rate": 1e-06, + "loss": -0.0691, + "num_tokens": 666325459.0, + "reward": 0.6964285969734192, + "reward_std": 0.13978277146816254, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1127 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3639.0, + "completions/mean_length": 1009.08935546875, + "completions/mean_terminated_length": 550.01025390625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 10.541107871720117, + "grad_norm": 0.15393473207950592, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 666855955.0, + "reward": 0.6774553656578064, + "reward_std": 0.14586800336837769, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3568.0, + "completions/mean_length": 943.86279296875, + "completions/mean_terminated_length": 561.187744140625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.550437317784256, + "grad_norm": 0.18221965432167053, + "learning_rate": 1e-06, + "loss": -0.05, + "num_tokens": 667393128.0, + "reward": 0.6718750596046448, + "reward_std": 0.15559779107570648, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1129 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3907.0, + "completions/mean_length": 951.65966796875, + "completions/mean_terminated_length": 609.2066650390625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 10.559766763848396, + "grad_norm": 0.14804306626319885, + "learning_rate": 1e-06, + "loss": -0.0375, + "num_tokens": 667972279.0, + "reward": 0.7176339626312256, + "reward_std": 0.13293799757957458, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1130 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2337.0, + "completions/mean_length": 951.716552734375, + "completions/mean_terminated_length": 552.2540893554688, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 10.569096209912537, + "grad_norm": 0.151942178606987, + "learning_rate": 1e-06, + "loss": -0.0671, + "num_tokens": 668491785.0, + "reward": 0.6741071939468384, + "reward_std": 0.13824184238910675, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1131 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3668.0, + "completions/mean_length": 1134.501220703125, + "completions/mean_terminated_length": 645.4109497070312, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 10.578425655976677, + "grad_norm": 0.16631515324115753, + "learning_rate": 1e-06, + "loss": -0.0788, + "num_tokens": 669081954.0, + "reward": 0.6495535969734192, + "reward_std": 0.14984957873821259, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1132 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4031.0, + "completions/mean_length": 906.86279296875, + "completions/mean_terminated_length": 607.029296875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 10.587755102040816, + "grad_norm": 0.1585167497396469, + "learning_rate": 1e-06, + "loss": -0.0352, + "num_tokens": 669678031.0, + "reward": 0.6819196939468384, + "reward_std": 0.11314209550619125, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1133 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2717.0, + "completions/mean_length": 1051.25341796875, + "completions/mean_terminated_length": 616.28955078125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 10.597084548104956, + "grad_norm": 0.16445694863796234, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 670251034.0, + "reward": 0.6171875, + "reward_std": 0.15785551071166992, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 1134 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2990.0, + "completions/mean_length": 987.7835083007812, + "completions/mean_terminated_length": 557.2935180664062, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 10.606413994169095, + "grad_norm": 0.16716906428337097, + "learning_rate": 1e-06, + "loss": -0.0771, + "num_tokens": 670774832.0, + "reward": 0.7020089626312256, + "reward_std": 0.15526077151298523, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763102173805237, + "step": 1135 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2645.0, + "completions/mean_length": 1102.7578125, + "completions/mean_terminated_length": 576.3871459960938, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 10.615743440233237, + "grad_norm": 0.15244048833847046, + "learning_rate": 1e-06, + "loss": -0.0766, + "num_tokens": 671301767.0, + "reward": 0.6819196939468384, + "reward_std": 0.1349007785320282, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1136 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2719.0, + "completions/mean_length": 1099.9296875, + "completions/mean_terminated_length": 573.0617065429688, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 10.625072886297376, + "grad_norm": 0.17223677039146423, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 671835240.0, + "reward": 0.6662946939468384, + "reward_std": 0.13527168333530426, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1137 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3753.0, + "completions/mean_length": 963.0692138671875, + "completions/mean_terminated_length": 591.4981079101562, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "epoch": 10.634402332361516, + "grad_norm": 0.1511462777853012, + "learning_rate": 1e-06, + "loss": -0.0558, + "num_tokens": 672400966.0, + "reward": 0.6975446939468384, + "reward_std": 0.14666074514389038, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1138 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3794.0, + "completions/mean_length": 916.2444458007812, + "completions/mean_terminated_length": 595.925048828125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 10.643731778425655, + "grad_norm": 0.14917875826358795, + "learning_rate": 1e-06, + "loss": -0.0389, + "num_tokens": 672963185.0, + "reward": 0.715401828289032, + "reward_std": 0.12910906970500946, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1139 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2999.0, + "completions/mean_length": 1131.966552734375, + "completions/mean_terminated_length": 615.3001708984375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 10.653061224489797, + "grad_norm": 0.15106496214866638, + "learning_rate": 1e-06, + "loss": -0.0701, + "num_tokens": 673531235.0, + "reward": 0.6618303656578064, + "reward_std": 0.11922773718833923, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1140 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 1167.3795166015625, + "completions/mean_terminated_length": 638.7615356445312, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 10.662390670553936, + "grad_norm": 0.14597447216510773, + "learning_rate": 1e-06, + "loss": -0.0917, + "num_tokens": 674099991.0, + "reward": 0.6517857313156128, + "reward_std": 0.142221599817276, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1141 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3597.0, + "completions/mean_length": 1002.5926513671875, + "completions/mean_terminated_length": 613.9736328125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 10.671720116618076, + "grad_norm": 0.12379582226276398, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 674683306.0, + "reward": 0.7053571939468384, + "reward_std": 0.10810784995555878, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1142 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2491.0, + "completions/mean_length": 1064.8873291015625, + "completions/mean_terminated_length": 527.1734619140625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 10.681049562682215, + "grad_norm": 0.15036804974079132, + "learning_rate": 1e-06, + "loss": -0.0702, + "num_tokens": 675183357.0, + "reward": 0.6953125596046448, + "reward_std": 0.11399875581264496, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1143 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3383.0, + "completions/mean_length": 991.0714721679688, + "completions/mean_terminated_length": 578.9127807617188, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 10.690379008746355, + "grad_norm": 0.16261744499206543, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 675730933.0, + "reward": 0.6339285969734192, + "reward_std": 0.13703972101211548, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1144 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3813.0, + "completions/mean_length": 1006.8426513671875, + "completions/mean_terminated_length": 587.907470703125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 10.699708454810496, + "grad_norm": 0.14799970388412476, + "learning_rate": 1e-06, + "loss": -0.0774, + "num_tokens": 676283848.0, + "reward": 0.6674107313156128, + "reward_std": 0.1342620700597763, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1145 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 967.732177734375, + "completions/mean_terminated_length": 601.0773315429688, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 10.709037900874636, + "grad_norm": 0.1432398110628128, + "learning_rate": 1e-06, + "loss": -0.0316, + "num_tokens": 676847608.0, + "reward": 0.6383928656578064, + "reward_std": 0.13080044090747833, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 1146 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3812.0, + "completions/mean_length": 1041.482177734375, + "completions/mean_terminated_length": 600.6641235351562, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 10.718367346938775, + "grad_norm": 0.1304580718278885, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 677407608.0, + "reward": 0.6964285969734192, + "reward_std": 0.13511762022972107, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3806.0, + "completions/mean_length": 977.01123046875, + "completions/mean_terminated_length": 620.1119384765625, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 10.727696793002915, + "grad_norm": 0.15389768779277802, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 677991866.0, + "reward": 0.6707589626312256, + "reward_std": 0.12459757179021835, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1148 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3330.0, + "completions/mean_length": 926.8192138671875, + "completions/mean_terminated_length": 624.6234741210938, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 10.737026239067056, + "grad_norm": 0.1426958441734314, + "learning_rate": 1e-06, + "loss": -0.067, + "num_tokens": 678588336.0, + "reward": 0.7433035969734192, + "reward_std": 0.13778789341449738, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1149 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2925.0, + "completions/mean_length": 1102.0335693359375, + "completions/mean_terminated_length": 621.1372680664062, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 10.746355685131196, + "grad_norm": 0.1610199213027954, + "learning_rate": 1e-06, + "loss": -0.1098, + "num_tokens": 679159134.0, + "reward": 0.6149553656578064, + "reward_std": 0.1904703974723816, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 1150 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2655.0, + "completions/mean_length": 1085.9320068359375, + "completions/mean_terminated_length": 606.9689331054688, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 10.755685131195335, + "grad_norm": 0.14890161156654358, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 679718841.0, + "reward": 0.6238839626312256, + "reward_std": 0.14966341853141785, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 1151 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1752232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2878.0, + "completions/mean_length": 1217.3695068359375, + "completions/mean_terminated_length": 605.8064575195312, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 10.765014577259475, + "grad_norm": 0.162293940782547, + "learning_rate": 1e-06, + "loss": -0.1113, + "num_tokens": 680260452.0, + "reward": 0.6283482313156128, + "reward_std": 0.1647316813468933, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 1152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3522.0, + "completions/mean_length": 959.9676513671875, + "completions/mean_terminated_length": 592.4027709960938, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 10.774344023323614, + "grad_norm": 0.1541632115840912, + "learning_rate": 1e-06, + "loss": -0.0745, + "num_tokens": 680826975.0, + "reward": 0.699776828289032, + "reward_std": 0.1325695961713791, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3916.0, + "completions/mean_length": 957.4609985351562, + "completions/mean_terminated_length": 593.9688720703125, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 10.783673469387756, + "grad_norm": 0.14813178777694702, + "learning_rate": 1e-06, + "loss": -0.0635, + "num_tokens": 681392332.0, + "reward": 0.7042410969734192, + "reward_std": 0.13876289129257202, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1154 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2717.0, + "completions/mean_length": 925.3638916015625, + "completions/mean_terminated_length": 588.7283935546875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.793002915451895, + "grad_norm": 0.14858478307724, + "learning_rate": 1e-06, + "loss": -0.0287, + "num_tokens": 681956506.0, + "reward": 0.6361607313156128, + "reward_std": 0.14045536518096924, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2685.0, + "completions/mean_length": 1002.50341796875, + "completions/mean_terminated_length": 596.28662109375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 10.802332361516035, + "grad_norm": 0.14092624187469482, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 682519629.0, + "reward": 0.6595982313156128, + "reward_std": 0.12681996822357178, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1156 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 993.2656860351562, + "completions/mean_terminated_length": 572.4891967773438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 10.811661807580174, + "grad_norm": 0.16126419603824615, + "learning_rate": 1e-06, + "loss": -0.0352, + "num_tokens": 683069843.0, + "reward": 0.6741071939468384, + "reward_std": 0.1329386979341507, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2845.0, + "completions/mean_length": 977.83935546875, + "completions/mean_terminated_length": 586.1105346679688, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 10.820991253644316, + "grad_norm": 0.1482193022966385, + "learning_rate": 1e-06, + "loss": -0.0494, + "num_tokens": 683626171.0, + "reward": 0.6852678656578064, + "reward_std": 0.11783835291862488, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1158 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 794.7991333007812, + "completions/mean_terminated_length": 549.3861083984375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 10.830320699708455, + "grad_norm": 0.152037113904953, + "learning_rate": 1e-06, + "loss": -0.0661, + "num_tokens": 684169511.0, + "reward": 0.7734375596046448, + "reward_std": 0.1475943922996521, + "rewards/verify_math_reward/mean": 0.7734375, + "rewards/verify_math_reward/std": 0.4188409447669983, + "step": 1159 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3332.0, + "completions/mean_length": 1098.657470703125, + "completions/mean_terminated_length": 589.969970703125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 10.839650145772595, + "grad_norm": 0.16387182474136353, + "learning_rate": 1e-06, + "loss": -0.0781, + "num_tokens": 684715236.0, + "reward": 0.652901828289032, + "reward_std": 0.16558973491191864, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 1160 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3907.0, + "completions/mean_length": 898.0881958007812, + "completions/mean_terminated_length": 567.2697143554688, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 10.848979591836734, + "grad_norm": 0.16958512365818024, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 685257907.0, + "reward": 0.7142857313156128, + "reward_std": 0.14556558430194855, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1161 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4016.0, + "completions/mean_length": 902.2332763671875, + "completions/mean_terminated_length": 610.4762573242188, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 10.858309037900874, + "grad_norm": 0.14481863379478455, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 685846732.0, + "reward": 0.6886160969734192, + "reward_std": 0.14730015397071838, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1162 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3430.0, + "completions/mean_length": 819.0145263671875, + "completions/mean_terminated_length": 554.1676635742188, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 10.867638483965015, + "grad_norm": 0.1555619090795517, + "learning_rate": 1e-06, + "loss": -0.0589, + "num_tokens": 686395817.0, + "reward": 0.7220982313156128, + "reward_std": 0.14966341853141785, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1163 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3898.0, + "completions/mean_length": 1159.421875, + "completions/mean_terminated_length": 611.0013427734375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.876967930029155, + "grad_norm": 0.15758249163627625, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 686939459.0, + "reward": 0.6127232313156128, + "reward_std": 0.1626587212085724, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 1164 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2749.0, + "completions/mean_length": 1000.9598388671875, + "completions/mean_terminated_length": 549.7647094726562, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.886297376093294, + "grad_norm": 0.15486952662467957, + "learning_rate": 1e-06, + "loss": -0.0675, + "num_tokens": 687454591.0, + "reward": 0.7131696939468384, + "reward_std": 0.11032096296548843, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3211.0, + "completions/mean_length": 1010.0670166015625, + "completions/mean_terminated_length": 582.6632690429688, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 10.895626822157434, + "grad_norm": 0.18233704566955566, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 688001619.0, + "reward": 0.731026828289032, + "reward_std": 0.1368447095155716, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1166 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2702.0, + "completions/mean_length": 1012.4386596679688, + "completions/mean_terminated_length": 567.4290771484375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 10.904956268221575, + "grad_norm": 0.14485375583171844, + "learning_rate": 1e-06, + "loss": -0.0476, + "num_tokens": 688533764.0, + "reward": 0.7209821939468384, + "reward_std": 0.10926900058984756, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 1167 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2201.0, + "completions/mean_length": 978.8995971679688, + "completions/mean_terminated_length": 501.50579833984375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 10.914285714285715, + "grad_norm": 0.1333228200674057, + "learning_rate": 1e-06, + "loss": -0.0523, + "num_tokens": 689015186.0, + "reward": 0.7098214626312256, + "reward_std": 0.0942763164639473, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3510.0, + "completions/mean_length": 1227.2523193359375, + "completions/mean_terminated_length": 682.4568481445312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 10.923615160349854, + "grad_norm": 0.15539956092834473, + "learning_rate": 1e-06, + "loss": -0.0836, + "num_tokens": 689615172.0, + "reward": 0.6071428656578064, + "reward_std": 0.17874544858932495, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 1169 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2996.0, + "completions/mean_length": 964.0714721679688, + "completions/mean_terminated_length": 557.2761840820312, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 10.932944606413994, + "grad_norm": 0.1751089245080948, + "learning_rate": 1e-06, + "loss": -0.0787, + "num_tokens": 690148340.0, + "reward": 0.6819196939468384, + "reward_std": 0.17464762926101685, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1170 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3278.0, + "completions/mean_length": 1067.4788818359375, + "completions/mean_terminated_length": 603.6512451171875, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "epoch": 10.942274052478133, + "grad_norm": 0.153433158993721, + "learning_rate": 1e-06, + "loss": -0.0681, + "num_tokens": 690713801.0, + "reward": 0.5859375, + "reward_std": 0.16037100553512573, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 1171 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3570.0, + "completions/mean_length": 1132.790283203125, + "completions/mean_terminated_length": 661.2833251953125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 10.951603498542275, + "grad_norm": 0.13235421478748322, + "learning_rate": 1e-06, + "loss": -0.0647, + "num_tokens": 691315357.0, + "reward": 0.6205357313156128, + "reward_std": 0.13659071922302246, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1172 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3669.0, + "completions/mean_length": 1080.087158203125, + "completions/mean_terminated_length": 658.0126953125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 10.960932944606414, + "grad_norm": 0.17141176760196686, + "learning_rate": 1e-06, + "loss": -0.068, + "num_tokens": 691914083.0, + "reward": 0.6517857313156128, + "reward_std": 0.1681494265794754, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1173 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3734.0, + "completions/mean_length": 1016.8449096679688, + "completions/mean_terminated_length": 616.9041748046875, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 10.970262390670554, + "grad_norm": 0.15656226873397827, + "learning_rate": 1e-06, + "loss": -0.0512, + "num_tokens": 692488704.0, + "reward": 0.6473214626312256, + "reward_std": 0.12294856458902359, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 1174 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3043.0, + "completions/mean_length": 865.8516235351562, + "completions/mean_terminated_length": 592.1101684570312, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 10.979591836734693, + "grad_norm": 0.14476443827152252, + "learning_rate": 1e-06, + "loss": -0.0536, + "num_tokens": 693070787.0, + "reward": 0.707589328289032, + "reward_std": 0.1454133540391922, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1175 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3835.0, + "completions/mean_length": 962.5402221679688, + "completions/mean_terminated_length": 608.322998046875, + "completions/min_length": 198.0, + "completions/min_terminated_length": 198.0, + "epoch": 10.988921282798835, + "grad_norm": 0.1381600797176361, + "learning_rate": 1e-06, + "loss": -0.0728, + "num_tokens": 693650367.0, + "reward": 0.7131696939468384, + "reward_std": 0.14553029835224152, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1176 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.11647727272727271, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2684.0, + "completions/mean_length": 1063.375, + "completions/mean_terminated_length": 663.5755615234375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 10.998250728862974, + "grad_norm": 0.14036379754543304, + "learning_rate": 1e-06, + "loss": -0.0363, + "num_tokens": 694221257.0, + "reward": 0.6194196939468384, + "reward_std": 0.11836080253124237, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 1177 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3844.0, + "completions/mean_length": 1020.8326416015625, + "completions/mean_terminated_length": 647.5018920898438, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 11.00932944606414, + "grad_norm": 0.15750440955162048, + "learning_rate": 1e-06, + "loss": -0.0578, + "num_tokens": 694832307.0, + "reward": 0.684151828289032, + "reward_std": 0.14876440167427063, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1178 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3029.0, + "completions/mean_length": 1144.9989013671875, + "completions/mean_terminated_length": 688.6585083007812, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 11.018658892128279, + "grad_norm": 0.14614486694335938, + "learning_rate": 1e-06, + "loss": -0.0434, + "num_tokens": 695458178.0, + "reward": 0.6350446939468384, + "reward_std": 0.1377551257610321, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 1179 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2308.0, + "completions/mean_length": 1016.6685791015625, + "completions/mean_terminated_length": 563.245849609375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 11.02798833819242, + "grad_norm": 0.17202715575695038, + "learning_rate": 1e-06, + "loss": -0.0939, + "num_tokens": 695979441.0, + "reward": 0.6975446939468384, + "reward_std": 0.1751040816307068, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1180 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3305.0, + "completions/mean_length": 1011.4085083007812, + "completions/mean_terminated_length": 579.72265625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 11.03731778425656, + "grad_norm": 0.1744828224182129, + "learning_rate": 1e-06, + "loss": -0.0676, + "num_tokens": 696521255.0, + "reward": 0.6774553656578064, + "reward_std": 0.19182203710079193, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1181 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2443.0, + "completions/mean_length": 1013.9542846679688, + "completions/mean_terminated_length": 626.7625732421875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.0466472303207, + "grad_norm": 0.16108720004558563, + "learning_rate": 1e-06, + "loss": -0.0679, + "num_tokens": 697106622.0, + "reward": 0.6752232313156128, + "reward_std": 0.1451530158519745, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1182 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3895.0, + "completions/mean_length": 939.6183471679688, + "completions/mean_terminated_length": 613.0960693359375, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 11.055976676384839, + "grad_norm": 0.12822826206684113, + "learning_rate": 1e-06, + "loss": -0.0257, + "num_tokens": 697693840.0, + "reward": 0.6964285969734192, + "reward_std": 0.11381399631500244, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1183 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2811.0, + "completions/mean_length": 1046.766845703125, + "completions/mean_terminated_length": 628.8515014648438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.06530612244898, + "grad_norm": 0.14602269232273102, + "learning_rate": 1e-06, + "loss": -0.0626, + "num_tokens": 698285311.0, + "reward": 0.6618303656578064, + "reward_std": 0.14771203696727753, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1184 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3158.0, + "completions/mean_length": 1098.110595703125, + "completions/mean_terminated_length": 607.5467529296875, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "epoch": 11.07463556851312, + "grad_norm": 0.1648271530866623, + "learning_rate": 1e-06, + "loss": -0.0892, + "num_tokens": 698842674.0, + "reward": 0.621651828289032, + "reward_std": 0.1681487113237381, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 1185 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2911.0, + "completions/mean_length": 1121.610595703125, + "completions/mean_terminated_length": 625.87890625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 11.08396501457726, + "grad_norm": 0.16297124326229095, + "learning_rate": 1e-06, + "loss": -0.0617, + "num_tokens": 699418277.0, + "reward": 0.613839328289032, + "reward_std": 0.13568425178527832, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 1186 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 1003.6842041015625, + "completions/mean_terminated_length": 570.9172973632812, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 11.093294460641399, + "grad_norm": 0.14411231875419617, + "learning_rate": 1e-06, + "loss": -0.0643, + "num_tokens": 699960090.0, + "reward": 0.6662946939468384, + "reward_std": 0.14176806807518005, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1187 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2534.0, + "completions/mean_length": 1062.0145263671875, + "completions/mean_terminated_length": 628.5880126953125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.102623906705539, + "grad_norm": 0.15576989948749542, + "learning_rate": 1e-06, + "loss": -0.066, + "num_tokens": 700552911.0, + "reward": 0.6305803656578064, + "reward_std": 0.15331050753593445, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1188 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2376.0, + "completions/mean_length": 963.1495971679688, + "completions/mean_terminated_length": 556.2345581054688, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.11195335276968, + "grad_norm": 0.13737371563911438, + "learning_rate": 1e-06, + "loss": -0.045, + "num_tokens": 701069349.0, + "reward": 0.699776828289032, + "reward_std": 0.11426975578069687, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1189 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1852678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3818.0, + "completions/mean_length": 1250.52685546875, + "completions/mean_terminated_length": 603.4739990234375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 11.12128279883382, + "grad_norm": 0.1470501869916916, + "learning_rate": 1e-06, + "loss": -0.0791, + "num_tokens": 701604661.0, + "reward": 0.6462053656578064, + "reward_std": 0.1184028759598732, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1190 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3439.0, + "completions/mean_length": 1024.1060791015625, + "completions/mean_terminated_length": 567.26025390625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 11.130612244897959, + "grad_norm": 0.146786168217659, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 702124196.0, + "reward": 0.7098214626312256, + "reward_std": 0.09950529038906097, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1191 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 1142.96875, + "completions/mean_terminated_length": 664.2023315429688, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.139941690962099, + "grad_norm": 0.135858952999115, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 702739328.0, + "reward": 0.574776828289032, + "reward_std": 0.1335441768169403, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 1192 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2813.0, + "completions/mean_length": 1016.091552734375, + "completions/mean_terminated_length": 602.8379516601562, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 11.14927113702624, + "grad_norm": 0.1672326922416687, + "learning_rate": 1e-06, + "loss": -0.0453, + "num_tokens": 703299594.0, + "reward": 0.6830357313156128, + "reward_std": 0.16044698655605316, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1193 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2147.0, + "completions/mean_length": 964.2545166015625, + "completions/mean_terminated_length": 561.9395141601562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.15860058309038, + "grad_norm": 0.15635710954666138, + "learning_rate": 1e-06, + "loss": -0.0866, + "num_tokens": 703830270.0, + "reward": 0.7433035969734192, + "reward_std": 0.15094542503356934, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1194 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3841.0, + "completions/mean_length": 989.029052734375, + "completions/mean_terminated_length": 581.0429077148438, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 11.167930029154519, + "grad_norm": 0.16447122395038605, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 704392424.0, + "reward": 0.6741071939468384, + "reward_std": 0.1384693682193756, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.46896928548812866, + "step": 1195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3838.0, + "completions/mean_length": 1072.443115234375, + "completions/mean_terminated_length": 609.3757934570312, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 11.177259475218658, + "grad_norm": 0.18577629327774048, + "learning_rate": 1e-06, + "loss": -0.0665, + "num_tokens": 704944653.0, + "reward": 0.676339328289032, + "reward_std": 0.16022199392318726, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2636.0, + "completions/mean_length": 865.2355346679688, + "completions/mean_terminated_length": 557.16748046875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 11.186588921282798, + "grad_norm": 0.17470259964466095, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 705487072.0, + "reward": 0.7198660969734192, + "reward_std": 0.14887316524982452, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1197 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2983.0, + "completions/mean_length": 998.9453735351562, + "completions/mean_terminated_length": 592.2613525390625, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 11.19591836734694, + "grad_norm": 0.1412266492843628, + "learning_rate": 1e-06, + "loss": -0.0332, + "num_tokens": 706040559.0, + "reward": 0.6908482313156128, + "reward_std": 0.11265213787555695, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3100.0, + "completions/mean_length": 1081.765625, + "completions/mean_terminated_length": 565.6026611328125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.205247813411079, + "grad_norm": 0.14826829731464386, + "learning_rate": 1e-06, + "loss": -0.0735, + "num_tokens": 706568957.0, + "reward": 0.6741071939468384, + "reward_std": 0.1247912049293518, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1199 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3496.0, + "completions/mean_length": 993.5592041015625, + "completions/mean_terminated_length": 621.2662353515625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 11.214577259475218, + "grad_norm": 0.16074644029140472, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 707161178.0, + "reward": 0.6986607313156128, + "reward_std": 0.15480755269527435, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1200 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3452.0, + "completions/mean_length": 983.92529296875, + "completions/mean_terminated_length": 597.3563232421875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 11.223906705539358, + "grad_norm": 0.14656543731689453, + "learning_rate": 1e-06, + "loss": -0.0421, + "num_tokens": 707719967.0, + "reward": 0.6651785969734192, + "reward_std": 0.128467857837677, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1201 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3800.0, + "completions/mean_length": 962.7879638671875, + "completions/mean_terminated_length": 621.5469970703125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 11.2332361516035, + "grad_norm": 0.13340365886688232, + "learning_rate": 1e-06, + "loss": -0.0279, + "num_tokens": 708309265.0, + "reward": 0.7354910969734192, + "reward_std": 0.10626452416181564, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131773710250854, + "step": 1202 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2833.0, + "completions/mean_length": 1011.8348388671875, + "completions/mean_terminated_length": 557.7003784179688, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 11.242565597667639, + "grad_norm": 0.1607799530029297, + "learning_rate": 1e-06, + "loss": -0.0645, + "num_tokens": 708831261.0, + "reward": 0.723214328289032, + "reward_std": 0.1385025829076767, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2409.0, + "completions/mean_length": 912.1975708007812, + "completions/mean_terminated_length": 582.8386840820312, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.251895043731778, + "grad_norm": 0.14539699256420135, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 709397678.0, + "reward": 0.731026828289032, + "reward_std": 0.13752618432044983, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1204 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3790.0, + "completions/mean_length": 1107.640625, + "completions/mean_terminated_length": 627.6450805664062, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 11.261224489795918, + "grad_norm": 0.14058206975460052, + "learning_rate": 1e-06, + "loss": -0.0318, + "num_tokens": 709991780.0, + "reward": 0.6071428656578064, + "reward_std": 0.11945343762636185, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 1205 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3225.0, + "completions/mean_length": 902.318115234375, + "completions/mean_terminated_length": 519.0762329101562, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 11.270553935860057, + "grad_norm": 0.15452228486537933, + "learning_rate": 1e-06, + "loss": -0.0465, + "num_tokens": 710497977.0, + "reward": 0.6897321939468384, + "reward_std": 0.13316510617733002, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1206 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3253.0, + "completions/mean_length": 898.708740234375, + "completions/mean_terminated_length": 602.3743896484375, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 11.279883381924199, + "grad_norm": 0.1408669352531433, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 711078052.0, + "reward": 0.6897321939468384, + "reward_std": 0.149252250790596, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 1207 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2679.0, + "completions/mean_length": 984.239990234375, + "completions/mean_terminated_length": 602.093994140625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.289212827988338, + "grad_norm": 0.13973675668239594, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 711651491.0, + "reward": 0.6595982313156128, + "reward_std": 0.1270350068807602, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3328.0, + "completions/mean_length": 916.49560546875, + "completions/mean_terminated_length": 557.0733032226562, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 11.298542274052478, + "grad_norm": 0.12469842284917831, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 712187623.0, + "reward": 0.6852678656578064, + "reward_std": 0.1112217977643013, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3896.0, + "completions/mean_length": 850.0089721679688, + "completions/mean_terminated_length": 540.489013671875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 11.307871720116617, + "grad_norm": 0.1457255333662033, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 712708999.0, + "reward": 0.7712053656578064, + "reward_std": 0.11295635998249054, + "rewards/verify_math_reward/mean": 0.7712053656578064, + "rewards/verify_math_reward/std": 0.42029133439064026, + "step": 1210 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3227.0, + "completions/mean_length": 1164.90625, + "completions/mean_terminated_length": 640.394775390625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 11.317201166180759, + "grad_norm": 0.15248100459575653, + "learning_rate": 1e-06, + "loss": -0.0623, + "num_tokens": 713295915.0, + "reward": 0.6305803656578064, + "reward_std": 0.1587570458650589, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3702.0, + "completions/mean_length": 993.23779296875, + "completions/mean_terminated_length": 567.9860229492188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.326530612244898, + "grad_norm": 0.17671267688274384, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 713829992.0, + "reward": 0.6796875596046448, + "reward_std": 0.13767842948436737, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1212 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2544.0, + "completions/mean_length": 941.552490234375, + "completions/mean_terminated_length": 540.7987060546875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 11.335860058309038, + "grad_norm": 0.13897483050823212, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 714337255.0, + "reward": 0.7332589626312256, + "reward_std": 0.09912622720003128, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425029158592224, + "step": 1213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2753.0, + "completions/mean_length": 1007.0714721679688, + "completions/mean_terminated_length": 556.7672729492188, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 11.345189504373177, + "grad_norm": 0.1880517154932022, + "learning_rate": 1e-06, + "loss": -0.0812, + "num_tokens": 714860863.0, + "reward": 0.6439732313156128, + "reward_std": 0.13831782341003418, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1214 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2321.0, + "completions/mean_length": 981.19091796875, + "completions/mean_terminated_length": 616.1134643554688, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 11.354518950437317, + "grad_norm": 0.16923686861991882, + "learning_rate": 1e-06, + "loss": -0.0457, + "num_tokens": 715449442.0, + "reward": 0.7042410969734192, + "reward_std": 0.15661627054214478, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3942.0, + "completions/mean_length": 1144.2467041015625, + "completions/mean_terminated_length": 611.4532470703125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 11.363848396501458, + "grad_norm": 0.15719284117221832, + "learning_rate": 1e-06, + "loss": -0.0258, + "num_tokens": 715995359.0, + "reward": 0.6439732313156128, + "reward_std": 0.13406704366207123, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1216 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3627.0, + "completions/mean_length": 1046.896240234375, + "completions/mean_terminated_length": 575.3853149414062, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 11.373177842565598, + "grad_norm": 0.1701359897851944, + "learning_rate": 1e-06, + "loss": -0.0586, + "num_tokens": 716542106.0, + "reward": 0.6808035969734192, + "reward_std": 0.14199630916118622, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1217 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2889.0, + "completions/mean_length": 984.2723388671875, + "completions/mean_terminated_length": 597.7465209960938, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 11.382507288629737, + "grad_norm": 0.1496564745903015, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 717113526.0, + "reward": 0.7064732313156128, + "reward_std": 0.13106076419353485, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1218 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2849.0, + "completions/mean_length": 911.59716796875, + "completions/mean_terminated_length": 564.7809448242188, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 11.391836734693877, + "grad_norm": 0.16429300606250763, + "learning_rate": 1e-06, + "loss": -0.0346, + "num_tokens": 717659893.0, + "reward": 0.6941964626312256, + "reward_std": 0.1475936770439148, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3114.0, + "completions/mean_length": 969.0714721679688, + "completions/mean_terminated_length": 585.0626831054688, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 11.401166180758018, + "grad_norm": 0.15202312171459198, + "learning_rate": 1e-06, + "loss": -0.0619, + "num_tokens": 718214677.0, + "reward": 0.684151828289032, + "reward_std": 0.1448908895254135, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1220 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3235.0, + "completions/mean_length": 1104.579345703125, + "completions/mean_terminated_length": 619.5888671875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 11.410495626822158, + "grad_norm": 0.1416786164045334, + "learning_rate": 1e-06, + "loss": -0.0598, + "num_tokens": 718774780.0, + "reward": 0.6272321939468384, + "reward_std": 0.12287048995494843, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 1221 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2610.0, + "completions/mean_length": 966.2545166015625, + "completions/mean_terminated_length": 625.3910522460938, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 11.419825072886297, + "grad_norm": 0.15622742474079132, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 719367488.0, + "reward": 0.7064732313156128, + "reward_std": 0.15300628542900085, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1222 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3642.0, + "completions/mean_length": 1031.6373291015625, + "completions/mean_terminated_length": 668.198486328125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 11.429154518950437, + "grad_norm": 0.1766078770160675, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 719982819.0, + "reward": 0.6696428656578064, + "reward_std": 0.15905943512916565, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2454.0, + "completions/mean_length": 873.4710083007812, + "completions/mean_terminated_length": 548.8427124023438, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 11.438483965014576, + "grad_norm": 0.1429998129606247, + "learning_rate": 1e-06, + "loss": -0.0462, + "num_tokens": 720513721.0, + "reward": 0.738839328289032, + "reward_std": 0.1049504205584526, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1224 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3157.0, + "completions/mean_length": 1134.4285888671875, + "completions/mean_terminated_length": 636.3285522460938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.447813411078718, + "grad_norm": 0.1509985327720642, + "learning_rate": 1e-06, + "loss": -0.0458, + "num_tokens": 721093745.0, + "reward": 0.6417410969734192, + "reward_std": 0.131517231464386, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1225 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3144.0, + "completions/mean_length": 1034.74560546875, + "completions/mean_terminated_length": 592.9552612304688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 11.457142857142857, + "grad_norm": 0.14548541605472565, + "learning_rate": 1e-06, + "loss": -0.0839, + "num_tokens": 721634445.0, + "reward": 0.6774553656578064, + "reward_std": 0.1414317786693573, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1226 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3006.0, + "completions/mean_length": 994.5313110351562, + "completions/mean_terminated_length": 555.9796142578125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.466472303206997, + "grad_norm": 0.17421141266822815, + "learning_rate": 1e-06, + "loss": -0.0666, + "num_tokens": 722152897.0, + "reward": 0.6696428656578064, + "reward_std": 0.15469878911972046, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1227 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3726.0, + "completions/mean_length": 1047.0234375, + "completions/mean_terminated_length": 589.0899047851562, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.475801749271136, + "grad_norm": 0.17102757096290588, + "learning_rate": 1e-06, + "loss": -0.0728, + "num_tokens": 722696398.0, + "reward": 0.6830357313156128, + "reward_std": 0.16277816891670227, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1228 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2746.0, + "completions/mean_length": 990.255615234375, + "completions/mean_terminated_length": 546.5778198242188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 11.485131195335278, + "grad_norm": 0.14747671782970428, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 723196987.0, + "reward": 0.7131696939468384, + "reward_std": 0.12471381574869156, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1229 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3971.0, + "completions/mean_length": 1067.454345703125, + "completions/mean_terminated_length": 621.5095825195312, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 11.494460641399417, + "grad_norm": 0.17116190493106842, + "learning_rate": 1e-06, + "loss": -0.0868, + "num_tokens": 723774242.0, + "reward": 0.6964285969734192, + "reward_std": 0.1933586299419403, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1230 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3791.0, + "completions/mean_length": 1072.5570068359375, + "completions/mean_terminated_length": 600.5097045898438, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 11.503790087463557, + "grad_norm": 0.16753843426704407, + "learning_rate": 1e-06, + "loss": -0.085, + "num_tokens": 724332901.0, + "reward": 0.660714328289032, + "reward_std": 0.1647641658782959, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1231 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3600.0, + "completions/mean_length": 923.2344360351562, + "completions/mean_terminated_length": 603.619140625, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 11.513119533527696, + "grad_norm": 0.14656352996826172, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 724915543.0, + "reward": 0.6752232313156128, + "reward_std": 0.11866389214992523, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1232 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3952.0, + "completions/mean_length": 997.29248046875, + "completions/mean_terminated_length": 585.9595947265625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 11.522448979591836, + "grad_norm": 0.13994833827018738, + "learning_rate": 1e-06, + "loss": -0.059, + "num_tokens": 725461477.0, + "reward": 0.6930803656578064, + "reward_std": 0.12918464839458466, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 1233 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3499.0, + "completions/mean_length": 1067.76123046875, + "completions/mean_terminated_length": 630.735595703125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 11.531778425655977, + "grad_norm": 0.14088860154151917, + "learning_rate": 1e-06, + "loss": -0.0723, + "num_tokens": 726046823.0, + "reward": 0.6930803656578064, + "reward_std": 0.13185352087020874, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 1234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1551339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3166.0, + "completions/mean_length": 1185.0234375, + "completions/mean_terminated_length": 650.51123046875, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 11.541107871720117, + "grad_norm": 0.15262410044670105, + "learning_rate": 1e-06, + "loss": -0.0804, + "num_tokens": 726627388.0, + "reward": 0.6439732313156128, + "reward_std": 0.15916889905929565, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909069061279297, + "step": 1235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3667.0, + "completions/mean_length": 1134.5513916015625, + "completions/mean_terminated_length": 618.3355102539062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.550437317784256, + "grad_norm": 0.16433238983154297, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 727183410.0, + "reward": 0.6897321939468384, + "reward_std": 0.12538963556289673, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1236 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1674107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3244.0, + "completions/mean_length": 1191.3739013671875, + "completions/mean_terminated_length": 607.3338012695312, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 11.559766763848396, + "grad_norm": 0.18279282748699188, + "learning_rate": 1e-06, + "loss": -0.0829, + "num_tokens": 727741529.0, + "reward": 0.6417410969734192, + "reward_std": 0.17021138966083527, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1237 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2878.0, + "completions/mean_length": 1037.607177734375, + "completions/mean_terminated_length": 644.71533203125, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 11.569096209912537, + "grad_norm": 0.1803968995809555, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 728341633.0, + "reward": 0.6830357313156128, + "reward_std": 0.1692771017551422, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1238 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4041.0, + "completions/mean_length": 928.6016235351562, + "completions/mean_terminated_length": 583.6373901367188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.578425655976677, + "grad_norm": 0.11708416789770126, + "learning_rate": 1e-06, + "loss": -0.0519, + "num_tokens": 728900660.0, + "reward": 0.7566964626312256, + "reward_std": 0.09754250943660736, + "rewards/verify_math_reward/mean": 0.7566964030265808, + "rewards/verify_math_reward/std": 0.4293164908885956, + "step": 1239 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2527.0, + "completions/mean_length": 1066.9654541015625, + "completions/mean_terminated_length": 562.1263427734375, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 11.587755102040816, + "grad_norm": 0.15569400787353516, + "learning_rate": 1e-06, + "loss": -0.0642, + "num_tokens": 729426725.0, + "reward": 0.6595982313156128, + "reward_std": 0.13023591041564941, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1240 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2353.0, + "completions/mean_length": 1094.4129638671875, + "completions/mean_terminated_length": 585.0052490234375, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 11.597084548104956, + "grad_norm": 0.18812379240989685, + "learning_rate": 1e-06, + "loss": -0.0909, + "num_tokens": 729980007.0, + "reward": 0.6428571939468384, + "reward_std": 0.16255316138267517, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.47942501306533813, + "step": 1241 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3338.0, + "completions/mean_length": 1021.4263916015625, + "completions/mean_terminated_length": 626.4558715820312, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.606413994169095, + "grad_norm": 0.17050611972808838, + "learning_rate": 1e-06, + "loss": -0.0378, + "num_tokens": 730579893.0, + "reward": 0.6082589626312256, + "reward_std": 0.14578990638256073, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 1242 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2892.0, + "completions/mean_length": 1010.458740234375, + "completions/mean_terminated_length": 609.6885375976562, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 11.615743440233237, + "grad_norm": 0.15167243778705597, + "learning_rate": 1e-06, + "loss": -0.0231, + "num_tokens": 731151936.0, + "reward": 0.6707589626312256, + "reward_std": 0.11622254550457001, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1243 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2661.0, + "completions/mean_length": 921.7857666015625, + "completions/mean_terminated_length": 576.0792236328125, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 11.625072886297376, + "grad_norm": 0.15372061729431152, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 731705616.0, + "reward": 0.7712053656578064, + "reward_std": 0.11080951988697052, + "rewards/verify_math_reward/mean": 0.7712053656578064, + "rewards/verify_math_reward/std": 0.42029133439064026, + "step": 1244 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3713.0, + "completions/mean_length": 1012.03466796875, + "completions/mean_terminated_length": 607.0694580078125, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "epoch": 11.634402332361516, + "grad_norm": 0.16829776763916016, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 732274087.0, + "reward": 0.6506696939468384, + "reward_std": 0.15988247096538544, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1245 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3484.0, + "completions/mean_length": 1112.146240234375, + "completions/mean_terminated_length": 601.1856689453125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 11.643731778425655, + "grad_norm": 0.1765514463186264, + "learning_rate": 1e-06, + "loss": -0.1198, + "num_tokens": 732822338.0, + "reward": 0.699776828289032, + "reward_std": 0.17330214381217957, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1246 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3251.0, + "completions/mean_length": 1039.438720703125, + "completions/mean_terminated_length": 566.7744750976562, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.653061224489797, + "grad_norm": 0.16878652572631836, + "learning_rate": 1e-06, + "loss": -0.1062, + "num_tokens": 733351339.0, + "reward": 0.691964328289032, + "reward_std": 0.15319105982780457, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1247 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2482.0, + "completions/mean_length": 1099.396240234375, + "completions/mean_terminated_length": 581.6583862304688, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.662390670553936, + "grad_norm": 0.15502163767814636, + "learning_rate": 1e-06, + "loss": -0.0705, + "num_tokens": 733875838.0, + "reward": 0.6908482313156128, + "reward_std": 0.11956290900707245, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1248 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3285.0, + "completions/mean_length": 1128.11279296875, + "completions/mean_terminated_length": 606.2008056640625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.671720116618076, + "grad_norm": 0.16884633898735046, + "learning_rate": 1e-06, + "loss": -0.0851, + "num_tokens": 734422347.0, + "reward": 0.691964328289032, + "reward_std": 0.18445299565792084, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1249 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2906.0, + "completions/mean_length": 1043.91748046875, + "completions/mean_terminated_length": 590.0179443359375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 11.681049562682215, + "grad_norm": 0.1654636710882187, + "learning_rate": 1e-06, + "loss": -0.0715, + "num_tokens": 734974233.0, + "reward": 0.6618303656578064, + "reward_std": 0.1471799910068512, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1250 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3587.0, + "completions/mean_length": 1092.6395263671875, + "completions/mean_terminated_length": 663.5880126953125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 11.690379008746355, + "grad_norm": 0.1492314338684082, + "learning_rate": 1e-06, + "loss": -0.066, + "num_tokens": 735586478.0, + "reward": 0.6339285969734192, + "reward_std": 0.14943771064281464, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1251 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2584.0, + "completions/mean_length": 954.161865234375, + "completions/mean_terminated_length": 568.322021484375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.699708454810496, + "grad_norm": 0.14329566061496735, + "learning_rate": 1e-06, + "loss": -0.0527, + "num_tokens": 736136439.0, + "reward": 0.7131696939468384, + "reward_std": 0.11712156236171722, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1252 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3859.0, + "completions/mean_length": 1067.9163818359375, + "completions/mean_terminated_length": 630.9131469726562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 11.709037900874636, + "grad_norm": 0.1440056562423706, + "learning_rate": 1e-06, + "loss": -0.0696, + "num_tokens": 736717372.0, + "reward": 0.6953125596046448, + "reward_std": 0.13084320724010468, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1253 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2638.0, + "completions/mean_length": 1112.9085693359375, + "completions/mean_terminated_length": 592.9200439453125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 11.718367346938775, + "grad_norm": 0.1666734665632248, + "learning_rate": 1e-06, + "loss": -0.0846, + "num_tokens": 737259362.0, + "reward": 0.645089328289032, + "reward_std": 0.15056565403938293, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1254 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1819196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 1244.3951416015625, + "completions/mean_terminated_length": 610.2728271484375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 11.727696793002915, + "grad_norm": 0.1752176135778427, + "learning_rate": 1e-06, + "loss": -0.0734, + "num_tokens": 737801804.0, + "reward": 0.645089328289032, + "reward_std": 0.16213765740394592, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1255 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2908.0, + "completions/mean_length": 947.9475708007812, + "completions/mean_terminated_length": 592.0807495117188, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.737026239067056, + "grad_norm": 0.14963215589523315, + "learning_rate": 1e-06, + "loss": -0.0457, + "num_tokens": 738370461.0, + "reward": 0.6941964626312256, + "reward_std": 0.11727311462163925, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1256 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3767.0, + "completions/mean_length": 1003.6038208007812, + "completions/mean_terminated_length": 628.1814575195312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 11.746355685131196, + "grad_norm": 0.14419062435626984, + "learning_rate": 1e-06, + "loss": -0.0545, + "num_tokens": 738963226.0, + "reward": 0.668526828289032, + "reward_std": 0.1432323157787323, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1257 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1662946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3549.0, + "completions/mean_length": 1175.9285888671875, + "completions/mean_terminated_length": 593.4779052734375, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 11.755685131195335, + "grad_norm": 0.17012688517570496, + "learning_rate": 1e-06, + "loss": -0.0809, + "num_tokens": 739501490.0, + "reward": 0.6049107313156128, + "reward_std": 0.15811441838741302, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 1258 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 1041.57373046875, + "completions/mean_terminated_length": 622.9467163085938, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 11.765014577259475, + "grad_norm": 0.1758948713541031, + "learning_rate": 1e-06, + "loss": -0.0659, + "num_tokens": 740083804.0, + "reward": 0.6439732313156128, + "reward_std": 0.18001748621463776, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2793.0, + "completions/mean_length": 877.2745971679688, + "completions/mean_terminated_length": 566.0391845703125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 11.774344023323614, + "grad_norm": 0.14424894750118256, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 740630498.0, + "reward": 0.7176339626312256, + "reward_std": 0.11242461204528809, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1260 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2387.0, + "completions/mean_length": 936.068115234375, + "completions/mean_terminated_length": 521.1275024414062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.783673469387756, + "grad_norm": 0.15792350471019745, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 741126287.0, + "reward": 0.7165178656578064, + "reward_std": 0.10100274533033371, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1261 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 1040.5982666015625, + "completions/mean_terminated_length": 549.8341674804688, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 11.793002915451895, + "grad_norm": 0.1770157814025879, + "learning_rate": 1e-06, + "loss": -0.0635, + "num_tokens": 741637063.0, + "reward": 0.7220982313156128, + "reward_std": 0.14301617443561554, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1262 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3975.0, + "completions/mean_length": 1029.35498046875, + "completions/mean_terminated_length": 648.4290771484375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.802332361516035, + "grad_norm": 0.14042982459068298, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 742239581.0, + "reward": 0.6796875596046448, + "reward_std": 0.12543241679668427, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1263 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2887.0, + "completions/mean_length": 1012.4051513671875, + "completions/mean_terminated_length": 598.656982421875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.811661807580174, + "grad_norm": 0.1545601189136505, + "learning_rate": 1e-06, + "loss": -0.0788, + "num_tokens": 742799600.0, + "reward": 0.7299107313156128, + "reward_std": 0.128467857837677, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3547.0, + "completions/mean_length": 1138.8170166015625, + "completions/mean_terminated_length": 632.4235229492188, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 11.820991253644316, + "grad_norm": 0.16730423271656036, + "learning_rate": 1e-06, + "loss": -0.0943, + "num_tokens": 743372068.0, + "reward": 0.6517857313156128, + "reward_std": 0.14766784012317657, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1265 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3845.0, + "completions/mean_length": 1004.2489013671875, + "completions/mean_terminated_length": 650.4664306640625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 11.830320699708455, + "grad_norm": 0.1589493453502655, + "learning_rate": 1e-06, + "loss": -0.0782, + "num_tokens": 743978083.0, + "reward": 0.6462053656578064, + "reward_std": 0.16273538768291473, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1266 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3884.0, + "completions/mean_length": 1031.3671875, + "completions/mean_terminated_length": 606.91357421875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 11.839650145772595, + "grad_norm": 0.16469897329807281, + "learning_rate": 1e-06, + "loss": -0.0556, + "num_tokens": 744548116.0, + "reward": 0.676339328289032, + "reward_std": 0.12756884098052979, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1267 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3804.0, + "completions/mean_length": 950.56591796875, + "completions/mean_terminated_length": 590.6405639648438, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 11.848979591836734, + "grad_norm": 0.1767946034669876, + "learning_rate": 1e-06, + "loss": -0.0558, + "num_tokens": 745126039.0, + "reward": 0.6718750596046448, + "reward_std": 0.17900757491588593, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1268 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3578.0, + "completions/mean_length": 945.72216796875, + "completions/mean_terminated_length": 563.2728271484375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.858309037900874, + "grad_norm": 0.16459469497203827, + "learning_rate": 1e-06, + "loss": -0.0556, + "num_tokens": 745664062.0, + "reward": 0.691964328289032, + "reward_std": 0.14586980640888214, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1269 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2285.0, + "completions/mean_length": 889.2210083007812, + "completions/mean_terminated_length": 600.5328369140625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 11.867638483965015, + "grad_norm": 0.18522396683692932, + "learning_rate": 1e-06, + "loss": -0.0486, + "num_tokens": 746249236.0, + "reward": 0.684151828289032, + "reward_std": 0.16179178655147552, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1270 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2763.0, + "completions/mean_length": 931.286865234375, + "completions/mean_terminated_length": 582.2664184570312, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 11.876967930029155, + "grad_norm": 0.15078574419021606, + "learning_rate": 1e-06, + "loss": -0.0683, + "num_tokens": 746815525.0, + "reward": 0.7287946939468384, + "reward_std": 0.12245932221412659, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3625.0, + "completions/mean_length": 953.3995971679688, + "completions/mean_terminated_length": 589.4370727539062, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 11.886297376093294, + "grad_norm": 0.16365188360214233, + "learning_rate": 1e-06, + "loss": -0.0768, + "num_tokens": 747383939.0, + "reward": 0.6808035969734192, + "reward_std": 0.15781132876873016, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1272 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2704.0, + "completions/mean_length": 1130.805908203125, + "completions/mean_terminated_length": 627.5744018554688, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 11.895626822157434, + "grad_norm": 0.15552617609500885, + "learning_rate": 1e-06, + "loss": -0.0909, + "num_tokens": 747953341.0, + "reward": 0.6395089626312256, + "reward_std": 0.13996753096580505, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 1273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 1100.610595703125, + "completions/mean_terminated_length": 655.142333984375, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 11.904956268221575, + "grad_norm": 0.16665425896644592, + "learning_rate": 1e-06, + "loss": -0.0535, + "num_tokens": 748553680.0, + "reward": 0.645089328289032, + "reward_std": 0.15297240018844604, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1274 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1506696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4083.0, + "completions/mean_length": 1183.44873046875, + "completions/mean_terminated_length": 666.7673950195312, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 11.914285714285715, + "grad_norm": 0.16371269524097443, + "learning_rate": 1e-06, + "loss": -0.0844, + "num_tokens": 749152266.0, + "reward": 0.6640625, + "reward_std": 0.15608564019203186, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1275 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3141.0, + "completions/mean_length": 1021.65966796875, + "completions/mean_terminated_length": 595.8614501953125, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 11.923615160349854, + "grad_norm": 0.18976180255413055, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 749712889.0, + "reward": 0.6462053656578064, + "reward_std": 0.17836888134479523, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1276 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2850.0, + "completions/mean_length": 947.0201416015625, + "completions/mean_terminated_length": 604.0618896484375, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 11.932944606413994, + "grad_norm": 0.18226487934589386, + "learning_rate": 1e-06, + "loss": -0.0773, + "num_tokens": 750291443.0, + "reward": 0.7031250596046448, + "reward_std": 0.17795519530773163, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1277 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2764.0, + "completions/mean_length": 1127.5770263671875, + "completions/mean_terminated_length": 650.78369140625, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 11.942274052478133, + "grad_norm": 0.1681506484746933, + "learning_rate": 1e-06, + "loss": -0.0914, + "num_tokens": 750877864.0, + "reward": 0.6238839626312256, + "reward_std": 0.164840430021286, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 1278 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3972.0, + "completions/mean_length": 1046.89404296875, + "completions/mean_terminated_length": 615.7465209960938, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.951603498542275, + "grad_norm": 0.15831658244132996, + "learning_rate": 1e-06, + "loss": -0.06, + "num_tokens": 751442841.0, + "reward": 0.668526828289032, + "reward_std": 0.1277204006910324, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1279 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3195.0, + "completions/mean_length": 899.732177734375, + "completions/mean_terminated_length": 582.0662841796875, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 11.960932944606414, + "grad_norm": 0.17508095502853394, + "learning_rate": 1e-06, + "loss": -0.0801, + "num_tokens": 752011009.0, + "reward": 0.715401828289032, + "reward_std": 0.16491824388504028, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1280 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2870.0, + "completions/mean_length": 947.3136596679688, + "completions/mean_terminated_length": 587.0161743164062, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 11.970262390670554, + "grad_norm": 0.16519255936145782, + "learning_rate": 1e-06, + "loss": -0.07, + "num_tokens": 752575138.0, + "reward": 0.7220982313156128, + "reward_std": 0.1510535031557083, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1281 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2909.0, + "completions/mean_length": 1011.6629638671875, + "completions/mean_terminated_length": 628.5394897460938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.979591836734693, + "grad_norm": 0.1419551968574524, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 753169612.0, + "reward": 0.7020089626312256, + "reward_std": 0.141472727060318, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1282 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3548.0, + "completions/mean_length": 996.5859985351562, + "completions/mean_terminated_length": 633.31298828125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 11.988921282798835, + "grad_norm": 0.1388169378042221, + "learning_rate": 1e-06, + "loss": -0.0646, + "num_tokens": 753770721.0, + "reward": 0.6718750596046448, + "reward_std": 0.1377211958169937, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.13636363636363635, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2563.0, + "completions/mean_length": 1066.3409423828125, + "completions/mean_terminated_length": 587.9736938476562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 11.998250728862974, + "grad_norm": 0.15468640625476837, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 754336241.0, + "reward": 0.715401828289032, + "reward_std": 0.14571574330329895, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1284 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4078.0, + "completions/mean_length": 999.4319458007812, + "completions/mean_terminated_length": 561.5732421875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 12.00932944606414, + "grad_norm": 0.16018709540367126, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 754864572.0, + "reward": 0.652901828289032, + "reward_std": 0.1332090049982071, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 1285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3843.0, + "completions/mean_length": 1053.4554443359375, + "completions/mean_terminated_length": 596.48779296875, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 12.018658892128279, + "grad_norm": 0.14922991394996643, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 755417316.0, + "reward": 0.6640625, + "reward_std": 0.12343642860651016, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1286 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2513.0, + "completions/mean_length": 981.5167846679688, + "completions/mean_terminated_length": 581.4193725585938, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 12.02798833819242, + "grad_norm": 0.17613713443279266, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 755970331.0, + "reward": 0.7243303656578064, + "reward_std": 0.1424509584903717, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1287 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3031.0, + "completions/mean_length": 913.63623046875, + "completions/mean_terminated_length": 597.3521728515625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 12.03731778425656, + "grad_norm": 0.14524930715560913, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 756545493.0, + "reward": 0.7142857313156128, + "reward_std": 0.1368108093738556, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1288 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3689.0, + "completions/mean_length": 906.9185791015625, + "completions/mean_terminated_length": 581.3419189453125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 12.0466472303207, + "grad_norm": 0.15215623378753662, + "learning_rate": 1e-06, + "loss": -0.0603, + "num_tokens": 757096060.0, + "reward": 0.7444196939468384, + "reward_std": 0.13110284507274628, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1289 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3803.0, + "completions/mean_length": 1000.6964721679688, + "completions/mean_terminated_length": 598.6582641601562, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 12.055976676384839, + "grad_norm": 0.1543106883764267, + "learning_rate": 1e-06, + "loss": -0.0506, + "num_tokens": 757657388.0, + "reward": 0.7053571939468384, + "reward_std": 0.13203828036785126, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 1290 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3465.0, + "completions/mean_length": 944.0636596679688, + "completions/mean_terminated_length": 583.394287109375, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 12.06530612244898, + "grad_norm": 0.14733773469924927, + "learning_rate": 1e-06, + "loss": -0.0334, + "num_tokens": 758217045.0, + "reward": 0.6808035969734192, + "reward_std": 0.11611196398735046, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1291 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2882.0, + "completions/mean_length": 986.3605346679688, + "completions/mean_terminated_length": 586.8853759765625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 12.07463556851312, + "grad_norm": 0.1351013332605362, + "learning_rate": 1e-06, + "loss": -0.0473, + "num_tokens": 758777568.0, + "reward": 0.6808035969734192, + "reward_std": 0.12459687143564224, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1292 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3057.0, + "completions/mean_length": 1018.24560546875, + "completions/mean_terminated_length": 635.9397583007812, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 12.08396501457726, + "grad_norm": 0.15032874047756195, + "learning_rate": 1e-06, + "loss": -0.045, + "num_tokens": 759374668.0, + "reward": 0.6796875596046448, + "reward_std": 0.13437014818191528, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1293 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2206.0, + "completions/mean_length": 1071.5201416015625, + "completions/mean_terminated_length": 621.7256469726562, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 12.093294460641399, + "grad_norm": 0.1823098212480545, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 759954062.0, + "reward": 0.6584821939468384, + "reward_std": 0.1744614690542221, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3477.0, + "completions/mean_length": 1049.22216796875, + "completions/mean_terminated_length": 631.6433715820312, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 12.102623906705539, + "grad_norm": 0.1675415188074112, + "learning_rate": 1e-06, + "loss": -0.0632, + "num_tokens": 760544501.0, + "reward": 0.660714328289032, + "reward_std": 0.1587563306093216, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3533.0, + "completions/mean_length": 987.1719360351562, + "completions/mean_terminated_length": 583.3770751953125, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 12.11195335276968, + "grad_norm": 0.13941673934459686, + "learning_rate": 1e-06, + "loss": -0.0675, + "num_tokens": 761090199.0, + "reward": 0.6975446939468384, + "reward_std": 0.10934386402368546, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1296 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2820.0, + "completions/mean_length": 914.568115234375, + "completions/mean_terminated_length": 615.4591064453125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 12.12128279883382, + "grad_norm": 0.15313217043876648, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 761680364.0, + "reward": 0.7377232313156128, + "reward_std": 0.1338074505329132, + "rewards/verify_math_reward/mean": 0.7377232313156128, + "rewards/verify_math_reward/std": 0.4401180148124695, + "step": 1297 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3182.0, + "completions/mean_length": 1076.3348388671875, + "completions/mean_terminated_length": 622.8035888671875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 12.130612244897959, + "grad_norm": 0.15633031725883484, + "learning_rate": 1e-06, + "loss": -0.0652, + "num_tokens": 762259064.0, + "reward": 0.637276828289032, + "reward_std": 0.11757691204547882, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1298 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3224.0, + "completions/mean_length": 823.0033569335938, + "completions/mean_terminated_length": 562.740966796875, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 12.139941690962099, + "grad_norm": 0.18286950886249542, + "learning_rate": 1e-06, + "loss": -0.0631, + "num_tokens": 762811459.0, + "reward": 0.738839328289032, + "reward_std": 0.1469210684299469, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1299 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2416.0, + "completions/mean_length": 1119.755615234375, + "completions/mean_terminated_length": 632.7337646484375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 12.14927113702624, + "grad_norm": 0.16507063806056976, + "learning_rate": 1e-06, + "loss": -0.1039, + "num_tokens": 763398656.0, + "reward": 0.6852678656578064, + "reward_std": 0.15698787569999695, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1300 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2443.0, + "completions/mean_length": 856.9955444335938, + "completions/mean_terminated_length": 561.1060180664062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.15860058309038, + "grad_norm": 0.14617304503917694, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 763943852.0, + "reward": 0.7477678656578064, + "reward_std": 0.11565802246332169, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.4345363676548004, + "step": 1301 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2792.0, + "completions/mean_length": 930.5491333007812, + "completions/mean_terminated_length": 590.135986328125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 12.167930029154519, + "grad_norm": 0.1530526876449585, + "learning_rate": 1e-06, + "loss": -0.055, + "num_tokens": 764511712.0, + "reward": 0.7008928656578064, + "reward_std": 0.13087712228298187, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 1302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2997.0, + "completions/mean_length": 995.6484985351562, + "completions/mean_terminated_length": 561.7570190429688, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 12.177259475218658, + "grad_norm": 0.16424821317195892, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 765035965.0, + "reward": 0.7142857313156128, + "reward_std": 0.13508442044258118, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1303 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2650.0, + "completions/mean_length": 901.7020263671875, + "completions/mean_terminated_length": 531.7521362304688, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.186588921282798, + "grad_norm": 0.14962927997112274, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 765542642.0, + "reward": 0.7299107313156128, + "reward_std": 0.0907084122300148, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1304 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3175.0, + "completions/mean_length": 986.279052734375, + "completions/mean_terminated_length": 639.0396728515625, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 12.19591836734694, + "grad_norm": 0.1443362981081009, + "learning_rate": 1e-06, + "loss": -0.0407, + "num_tokens": 766148404.0, + "reward": 0.6897321939468384, + "reward_std": 0.11873988062143326, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1305 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2882.0, + "completions/mean_length": 820.3092041015625, + "completions/mean_terminated_length": 572.5678100585938, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 12.205247813411079, + "grad_norm": 0.15166939795017242, + "learning_rate": 1e-06, + "loss": -0.0266, + "num_tokens": 766711721.0, + "reward": 0.7109375596046448, + "reward_std": 0.13778719305992126, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1306 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3649.0, + "completions/mean_length": 1095.3795166015625, + "completions/mean_terminated_length": 595.2760620117188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 12.214577259475218, + "grad_norm": 0.1697177141904831, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 767257541.0, + "reward": 0.6796875596046448, + "reward_std": 0.13575884699821472, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1307 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3567.0, + "completions/mean_length": 1081.71875, + "completions/mean_terminated_length": 620.0720825195312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.223906705539358, + "grad_norm": 0.1579607129096985, + "learning_rate": 1e-06, + "loss": -0.05, + "num_tokens": 767825225.0, + "reward": 0.6696428656578064, + "reward_std": 0.13098590075969696, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1308 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4007.0, + "completions/mean_length": 1048.03125, + "completions/mean_terminated_length": 612.6071166992188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.2332361516035, + "grad_norm": 0.1728459596633911, + "learning_rate": 1e-06, + "loss": -0.0487, + "num_tokens": 768392317.0, + "reward": 0.6517857313156128, + "reward_std": 0.16349564492702484, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1309 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2244.0, + "completions/mean_length": 893.5881958007812, + "completions/mean_terminated_length": 601.0414428710938, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 12.242565597667639, + "grad_norm": 0.14819501340389252, + "learning_rate": 1e-06, + "loss": -0.0317, + "num_tokens": 768973044.0, + "reward": 0.7131696939468384, + "reward_std": 0.1465405821800232, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1310 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3865.0, + "completions/mean_length": 918.0904541015625, + "completions/mean_terminated_length": 550.03857421875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 12.251895043731778, + "grad_norm": 0.18584251403808594, + "learning_rate": 1e-06, + "loss": -0.0763, + "num_tokens": 769495069.0, + "reward": 0.6986607313156128, + "reward_std": 0.16499200463294983, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1311 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2679.0, + "completions/mean_length": 1109.23779296875, + "completions/mean_terminated_length": 584.0065307617188, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 12.261224489795918, + "grad_norm": 0.16517263650894165, + "learning_rate": 1e-06, + "loss": -0.0918, + "num_tokens": 770032426.0, + "reward": 0.6417410969734192, + "reward_std": 0.16525229811668396, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 1312 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3460.0, + "completions/mean_length": 1016.3951416015625, + "completions/mean_terminated_length": 638.197998046875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.270553935860057, + "grad_norm": 0.14833539724349976, + "learning_rate": 1e-06, + "loss": -0.0656, + "num_tokens": 770627332.0, + "reward": 0.6618303656578064, + "reward_std": 0.12700431048870087, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1313 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2367.0, + "completions/mean_length": 876.6596069335938, + "completions/mean_terminated_length": 565.36474609375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 12.279883381924199, + "grad_norm": 0.1626843512058258, + "learning_rate": 1e-06, + "loss": -0.0239, + "num_tokens": 771180267.0, + "reward": 0.6975446939468384, + "reward_std": 0.13110394775867462, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1314 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3596.0, + "completions/mean_length": 810.099365234375, + "completions/mean_terminated_length": 531.6331787109375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 12.289212827988338, + "grad_norm": 0.16345950961112976, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 771699612.0, + "reward": 0.7477678656578064, + "reward_std": 0.13602055609226227, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.4345363676548004, + "step": 1315 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3488.0, + "completions/mean_length": 973.94091796875, + "completions/mean_terminated_length": 563.9735107421875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 12.298542274052478, + "grad_norm": 0.14781834185123444, + "learning_rate": 1e-06, + "loss": -0.0631, + "num_tokens": 772238951.0, + "reward": 0.6774553656578064, + "reward_std": 0.13182030618190765, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1316 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3104.0, + "completions/mean_length": 845.9085083007812, + "completions/mean_terminated_length": 612.6483154296875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.307871720116617, + "grad_norm": 0.1481163501739502, + "learning_rate": 1e-06, + "loss": -0.0474, + "num_tokens": 772838269.0, + "reward": 0.7165178656578064, + "reward_std": 0.1413978636264801, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1317 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2480.0, + "completions/mean_length": 1088.271240234375, + "completions/mean_terminated_length": 627.6280517578125, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 12.317201166180759, + "grad_norm": 0.1655196249485016, + "learning_rate": 1e-06, + "loss": -0.0757, + "num_tokens": 773425488.0, + "reward": 0.6462053656578064, + "reward_std": 0.15010850131511688, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 877.349365234375, + "completions/mean_terminated_length": 553.1117553710938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 12.326530612244898, + "grad_norm": 0.1632184386253357, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 773959009.0, + "reward": 0.7243303656578064, + "reward_std": 0.12872818112373352, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1319 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3812.0, + "completions/mean_length": 1046.8170166015625, + "completions/mean_terminated_length": 650.7692260742188, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 12.335860058309038, + "grad_norm": 0.12957333028316498, + "learning_rate": 1e-06, + "loss": -0.0609, + "num_tokens": 774561469.0, + "reward": 0.6662946939468384, + "reward_std": 0.0958935096859932, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1320 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 1166.2254638671875, + "completions/mean_terminated_length": 669.0052490234375, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 12.345189504373177, + "grad_norm": 0.18731240928173065, + "learning_rate": 1e-06, + "loss": -0.0601, + "num_tokens": 775158551.0, + "reward": 0.621651828289032, + "reward_std": 0.17438660562038422, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.485245943069458, + "step": 1321 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2411.0, + "completions/mean_length": 1006.2756958007812, + "completions/mean_terminated_length": 555.85546875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 12.354518950437317, + "grad_norm": 0.14855942130088806, + "learning_rate": 1e-06, + "loss": -0.0415, + "num_tokens": 775680038.0, + "reward": 0.6785714626312256, + "reward_std": 0.10494930297136307, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1322 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2882.0, + "completions/mean_length": 996.2779541015625, + "completions/mean_terminated_length": 602.4767456054688, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 12.363848396501458, + "grad_norm": 0.1447145640850067, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 776246287.0, + "reward": 0.6707589626312256, + "reward_std": 0.11712156236171722, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2814.0, + "completions/mean_length": 1053.1429443359375, + "completions/mean_terminated_length": 636.1015014648438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.373177842565598, + "grad_norm": 0.17528286576271057, + "learning_rate": 1e-06, + "loss": -0.0644, + "num_tokens": 776843631.0, + "reward": 0.6283482313156128, + "reward_std": 0.17464621365070343, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 1324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3029.0, + "completions/mean_length": 1116.794677734375, + "completions/mean_terminated_length": 647.2041625976562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 12.382507288629737, + "grad_norm": 0.1563570499420166, + "learning_rate": 1e-06, + "loss": -0.0626, + "num_tokens": 777430359.0, + "reward": 0.6261160969734192, + "reward_std": 0.13429346680641174, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 1325 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3545.0, + "completions/mean_length": 1183.0982666015625, + "completions/mean_terminated_length": 679.822021484375, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 12.391836734693877, + "grad_norm": 0.15334337949752808, + "learning_rate": 1e-06, + "loss": -0.0786, + "num_tokens": 778042703.0, + "reward": 0.582589328289032, + "reward_std": 0.16206209361553192, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 1326 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2700.0, + "completions/mean_length": 963.64404296875, + "completions/mean_terminated_length": 570.1318969726562, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 12.401166180758018, + "grad_norm": 0.13570497930049896, + "learning_rate": 1e-06, + "loss": -0.0611, + "num_tokens": 778575152.0, + "reward": 0.6897321939468384, + "reward_std": 0.12343572080135345, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 1327 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2765.0, + "completions/mean_length": 945.27685546875, + "completions/mean_terminated_length": 589.1080932617188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.410495626822158, + "grad_norm": 0.171030193567276, + "learning_rate": 1e-06, + "loss": -0.0637, + "num_tokens": 779138576.0, + "reward": 0.7109375596046448, + "reward_std": 0.1475188136100769, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2268.0, + "completions/mean_length": 979.0547485351562, + "completions/mean_terminated_length": 542.8409423828125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 12.419825072886297, + "grad_norm": 0.16472549736499786, + "learning_rate": 1e-06, + "loss": -0.0565, + "num_tokens": 779655433.0, + "reward": 0.7042410969734192, + "reward_std": 0.13527169823646545, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1329 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2762.0, + "completions/mean_length": 1003.1920166015625, + "completions/mean_terminated_length": 588.2075805664062, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 12.429154518950437, + "grad_norm": 0.19368857145309448, + "learning_rate": 1e-06, + "loss": -0.0818, + "num_tokens": 780221221.0, + "reward": 0.6886160969734192, + "reward_std": 0.15123644471168518, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1330 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2555.0, + "completions/mean_length": 977.9375610351562, + "completions/mean_terminated_length": 550.5888061523438, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 12.438483965014576, + "grad_norm": 0.16014425456523895, + "learning_rate": 1e-06, + "loss": -0.038, + "num_tokens": 780742637.0, + "reward": 0.7142857313156128, + "reward_std": 0.1294117569923401, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2326.0, + "completions/mean_length": 784.0100708007812, + "completions/mean_terminated_length": 542.0562744140625, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 12.447813411078718, + "grad_norm": 0.1632382571697235, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 781281278.0, + "reward": 0.7522321939468384, + "reward_std": 0.15375375747680664, + "rewards/verify_math_reward/mean": 0.7522321343421936, + "rewards/verify_math_reward/std": 0.4319573938846588, + "step": 1332 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2669.0, + "completions/mean_length": 1023.6328735351562, + "completions/mean_terminated_length": 602.5469360351562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 12.457142857142857, + "grad_norm": 0.1574413776397705, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 781857085.0, + "reward": 0.6462053656578064, + "reward_std": 0.12914256751537323, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1333 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3184.0, + "completions/mean_length": 1004.7042846679688, + "completions/mean_terminated_length": 589.9227905273438, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 12.466472303206997, + "grad_norm": 0.17971403896808624, + "learning_rate": 1e-06, + "loss": -0.0724, + "num_tokens": 782415188.0, + "reward": 0.6953125596046448, + "reward_std": 0.15533748269081116, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3374.0, + "completions/mean_length": 752.9642944335938, + "completions/mean_terminated_length": 542.7852783203125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 12.475801749271136, + "grad_norm": 0.15812382102012634, + "learning_rate": 1e-06, + "loss": -0.0347, + "num_tokens": 782953372.0, + "reward": 0.7433035969734192, + "reward_std": 0.13958844542503357, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2359.0, + "completions/mean_length": 909.7053833007812, + "completions/mean_terminated_length": 584.4132690429688, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 12.485131195335278, + "grad_norm": 0.150274857878685, + "learning_rate": 1e-06, + "loss": -0.0534, + "num_tokens": 783523268.0, + "reward": 0.7131696939468384, + "reward_std": 0.13583439588546753, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1336 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2436.0, + "completions/mean_length": 914.2902221679688, + "completions/mean_terminated_length": 567.767333984375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 12.494460641399417, + "grad_norm": 0.17220300436019897, + "learning_rate": 1e-06, + "loss": -0.0245, + "num_tokens": 784074376.0, + "reward": 0.7042410969734192, + "reward_std": 0.14496758580207825, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663806796073914, + "step": 1337 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2966.0, + "completions/mean_length": 978.8225708007812, + "completions/mean_terminated_length": 609.1198120117188, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 12.503790087463557, + "grad_norm": 0.16410626471042633, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 784652057.0, + "reward": 0.6584821939468384, + "reward_std": 0.14455640316009521, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3562.0, + "completions/mean_length": 1074.0179443359375, + "completions/mean_terminated_length": 629.0396728515625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 12.513119533527696, + "grad_norm": 0.1667211502790451, + "learning_rate": 1e-06, + "loss": -0.0372, + "num_tokens": 785244305.0, + "reward": 0.6383928656578064, + "reward_std": 0.14628027379512787, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1339 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2260.0, + "completions/mean_length": 1019.26123046875, + "completions/mean_terminated_length": 619.63427734375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 12.522448979591836, + "grad_norm": 0.17799793183803558, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 785825643.0, + "reward": 0.7109375596046448, + "reward_std": 0.15285544097423553, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1340 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2771.0, + "completions/mean_length": 1017.0926513671875, + "completions/mean_terminated_length": 595.1104125976562, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 12.531778425655977, + "grad_norm": 0.16019681096076965, + "learning_rate": 1e-06, + "loss": -0.063, + "num_tokens": 786377886.0, + "reward": 0.6886160969734192, + "reward_std": 0.1312127262353897, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1341 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2515.0, + "completions/mean_length": 988.1038208007812, + "completions/mean_terminated_length": 593.26416015625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 12.541107871720117, + "grad_norm": 0.14082224667072296, + "learning_rate": 1e-06, + "loss": -0.0582, + "num_tokens": 786938163.0, + "reward": 0.6707589626312256, + "reward_std": 0.12554006278514862, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 998.9777221679688, + "completions/mean_terminated_length": 614.278564453125, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 12.550437317784256, + "grad_norm": 0.15896563231945038, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 787510191.0, + "reward": 0.6875000596046448, + "reward_std": 0.13740670680999756, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1343 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4010.0, + "completions/mean_length": 873.7176513671875, + "completions/mean_terminated_length": 566.4584350585938, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.559766763848396, + "grad_norm": 0.1577504724264145, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 788054018.0, + "reward": 0.7321428656578064, + "reward_std": 0.13793987035751343, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 1344 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3851.0, + "completions/mean_length": 998.9297485351562, + "completions/mean_terminated_length": 596.6620483398438, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 12.569096209912537, + "grad_norm": 0.18820682168006897, + "learning_rate": 1e-06, + "loss": -0.0465, + "num_tokens": 788620579.0, + "reward": 0.6674107313156128, + "reward_std": 0.16078217327594757, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1345 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2542.0, + "completions/mean_length": 783.5156860351562, + "completions/mean_terminated_length": 554.2506103515625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 12.578425655976677, + "grad_norm": 0.1534918248653412, + "learning_rate": 1e-06, + "loss": -0.0345, + "num_tokens": 789179353.0, + "reward": 0.7332589626312256, + "reward_std": 0.12538781762123108, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425029158592224, + "step": 1346 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3036.0, + "completions/mean_length": 1004.1975708007812, + "completions/mean_terminated_length": 557.9987182617188, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 12.587755102040816, + "grad_norm": 0.15490977466106415, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 789706594.0, + "reward": 0.6808035969734192, + "reward_std": 0.11840106546878815, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1347 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3798.0, + "completions/mean_length": 943.6663208007812, + "completions/mean_terminated_length": 578.5765991210938, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 12.597084548104956, + "grad_norm": 0.13104921579360962, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 790259151.0, + "reward": 0.7366071939468384, + "reward_std": 0.11178451776504517, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1348 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3620.0, + "completions/mean_length": 1013.0324096679688, + "completions/mean_terminated_length": 634.4223022460938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 12.606413994169095, + "grad_norm": 0.1768091917037964, + "learning_rate": 1e-06, + "loss": -0.056, + "num_tokens": 790863548.0, + "reward": 0.6439732313156128, + "reward_std": 0.1536777913570404, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2641.0, + "completions/mean_length": 927.7857666015625, + "completions/mean_terminated_length": 538.706787109375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 12.615743440233237, + "grad_norm": 0.16212619841098785, + "learning_rate": 1e-06, + "loss": -0.0422, + "num_tokens": 791386804.0, + "reward": 0.6897321939468384, + "reward_std": 0.14459103345870972, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1350 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2949.0, + "completions/mean_length": 1030.421875, + "completions/mean_terminated_length": 588.0076293945312, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 12.625072886297376, + "grad_norm": 0.20744355022907257, + "learning_rate": 1e-06, + "loss": -0.0801, + "num_tokens": 791936710.0, + "reward": 0.6238839626312256, + "reward_std": 0.18445341289043427, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 1351 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3062.0, + "completions/mean_length": 1043.9754638671875, + "completions/mean_terminated_length": 590.0846557617188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.634402332361516, + "grad_norm": 0.15848352015018463, + "learning_rate": 1e-06, + "loss": -0.0934, + "num_tokens": 792484888.0, + "reward": 0.6662946939468384, + "reward_std": 0.1431235522031784, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1352 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3645.0, + "completions/mean_length": 997.1417846679688, + "completions/mean_terminated_length": 572.4251098632812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.643731778425655, + "grad_norm": 0.16984209418296814, + "learning_rate": 1e-06, + "loss": -0.047, + "num_tokens": 793029991.0, + "reward": 0.7020089626312256, + "reward_std": 0.13767842948436737, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1353 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1860.0, + "completions/mean_length": 825.1239013671875, + "completions/mean_terminated_length": 500.0429382324219, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 12.653061224489797, + "grad_norm": 0.14401152729988098, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 793514862.0, + "reward": 0.7433035969734192, + "reward_std": 0.09175968915224075, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705445528030396, + "step": 1354 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2660.0, + "completions/mean_length": 887.7645263671875, + "completions/mean_terminated_length": 594.685791015625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 12.662390670553936, + "grad_norm": 0.1445140391588211, + "learning_rate": 1e-06, + "loss": -0.054, + "num_tokens": 794096259.0, + "reward": 0.7109375596046448, + "reward_std": 0.12234944105148315, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1355 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2784.0, + "completions/mean_length": 1125.9342041015625, + "completions/mean_terminated_length": 617.3346557617188, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 12.671720116618076, + "grad_norm": 0.17599262297153473, + "learning_rate": 1e-06, + "loss": -0.0697, + "num_tokens": 794670192.0, + "reward": 0.637276828289032, + "reward_std": 0.14124631881713867, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1356 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4049.0, + "completions/mean_length": 1081.80810546875, + "completions/mean_terminated_length": 620.175048828125, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 12.681049562682215, + "grad_norm": 0.15625596046447754, + "learning_rate": 1e-06, + "loss": -0.0775, + "num_tokens": 795238852.0, + "reward": 0.6796875596046448, + "reward_std": 0.14744988083839417, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1357 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3389.0, + "completions/mean_length": 896.3225708007812, + "completions/mean_terminated_length": 534.619873046875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.690379008746355, + "grad_norm": 0.1739022582769394, + "learning_rate": 1e-06, + "loss": -0.0355, + "num_tokens": 795755453.0, + "reward": 0.7321428656578064, + "reward_std": 0.13154971599578857, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 1358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2699.0, + "completions/mean_length": 921.630615234375, + "completions/mean_terminated_length": 580.2583618164062, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 12.699708454810496, + "grad_norm": 0.17853571474552155, + "learning_rate": 1e-06, + "loss": -0.0663, + "num_tokens": 796314938.0, + "reward": 0.7433035969734192, + "reward_std": 0.17179258167743683, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705445528030396, + "step": 1359 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2950.0, + "completions/mean_length": 961.7801513671875, + "completions/mean_terminated_length": 598.7882690429688, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 12.709037900874636, + "grad_norm": 0.1860802322626114, + "learning_rate": 1e-06, + "loss": -0.0474, + "num_tokens": 796879045.0, + "reward": 0.6640625, + "reward_std": 0.17615535855293274, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1360 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2350.0, + "completions/mean_length": 894.5357666015625, + "completions/mean_terminated_length": 593.5433959960938, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 12.718367346938775, + "grad_norm": 0.15276645123958588, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 797446453.0, + "reward": 0.6819196939468384, + "reward_std": 0.13309136033058167, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1361 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3274.0, + "completions/mean_length": 1128.1451416015625, + "completions/mean_terminated_length": 664.7767944335938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 12.727696793002915, + "grad_norm": 0.17742541432380676, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 798052919.0, + "reward": 0.5658482313156128, + "reward_std": 0.15357013046741486, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 1362 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2856.0, + "completions/mean_length": 1035.185302734375, + "completions/mean_terminated_length": 611.2604370117188, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 12.737026239067056, + "grad_norm": 0.15881547331809998, + "learning_rate": 1e-06, + "loss": -0.0489, + "num_tokens": 798627301.0, + "reward": 0.6573660969734192, + "reward_std": 0.15274415910243988, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485536336898804, + "step": 1363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3933.0, + "completions/mean_length": 975.9888916015625, + "completions/mean_terminated_length": 605.9500732421875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 12.746355685131196, + "grad_norm": 0.16645316779613495, + "learning_rate": 1e-06, + "loss": -0.0626, + "num_tokens": 799213939.0, + "reward": 0.6417410969734192, + "reward_std": 0.13778719305992126, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 1364 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2626.0, + "completions/mean_length": 923.2667846679688, + "completions/mean_terminated_length": 555.814453125, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 12.755685131195335, + "grad_norm": 0.16511110961437225, + "learning_rate": 1e-06, + "loss": -0.0699, + "num_tokens": 799740338.0, + "reward": 0.715401828289032, + "reward_std": 0.14533847570419312, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1365 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4016.0, + "completions/mean_length": 1069.1195068359375, + "completions/mean_terminated_length": 618.9679565429688, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 12.765014577259475, + "grad_norm": 0.16087523102760315, + "learning_rate": 1e-06, + "loss": -0.0535, + "num_tokens": 800313589.0, + "reward": 0.660714328289032, + "reward_std": 0.15165898203849792, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1366 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1584821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3205.0, + "completions/mean_length": 1139.23779296875, + "completions/mean_terminated_length": 582.3938598632812, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.774344023323614, + "grad_norm": 0.21401703357696533, + "learning_rate": 1e-06, + "loss": -0.0811, + "num_tokens": 800838802.0, + "reward": 0.6729910969734192, + "reward_std": 0.1496659368276596, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1367 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2582.0, + "completions/mean_length": 942.247802734375, + "completions/mean_terminated_length": 554.9448852539062, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 12.783673469387756, + "grad_norm": 0.17008671164512634, + "learning_rate": 1e-06, + "loss": -0.0377, + "num_tokens": 801367792.0, + "reward": 0.7042410969734192, + "reward_std": 0.11486568301916122, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1368 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3100.0, + "completions/mean_length": 971.75341796875, + "completions/mean_terminated_length": 609.9165649414062, + "completions/min_length": 196.0, + "completions/min_terminated_length": 196.0, + "epoch": 12.793002915451895, + "grad_norm": 0.15616938471794128, + "learning_rate": 1e-06, + "loss": -0.0448, + "num_tokens": 801959363.0, + "reward": 0.7031250596046448, + "reward_std": 0.13380561769008636, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2847.0, + "completions/mean_length": 997.4520263671875, + "completions/mean_terminated_length": 581.6974487304688, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.802332361516035, + "grad_norm": 0.15678907930850983, + "learning_rate": 1e-06, + "loss": -0.0601, + "num_tokens": 802509664.0, + "reward": 0.7131696939468384, + "reward_std": 0.13478201627731323, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1370 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3806.0, + "completions/mean_length": 1153.521240234375, + "completions/mean_terminated_length": 626.972412109375, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 12.811661807580174, + "grad_norm": 0.17155136168003082, + "learning_rate": 1e-06, + "loss": -0.1146, + "num_tokens": 803089171.0, + "reward": 0.6171875, + "reward_std": 0.16134923696517944, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 1371 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 1043.649658203125, + "completions/mean_terminated_length": 638.4703369140625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 12.820991253644316, + "grad_norm": 0.15405990183353424, + "learning_rate": 1e-06, + "loss": -0.0421, + "num_tokens": 803687601.0, + "reward": 0.6316964626312256, + "reward_std": 0.13444501161575317, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1372 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3640.0, + "completions/mean_length": 1055.359375, + "completions/mean_terminated_length": 616.5440673828125, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 12.830320699708455, + "grad_norm": 0.1485147476196289, + "learning_rate": 1e-06, + "loss": -0.0515, + "num_tokens": 804267067.0, + "reward": 0.7008928656578064, + "reward_std": 0.13500885665416718, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3295.0, + "completions/mean_length": 980.2645263671875, + "completions/mean_terminated_length": 588.8404541015625, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 12.839650145772595, + "grad_norm": 0.1513684242963791, + "learning_rate": 1e-06, + "loss": -0.0642, + "num_tokens": 804822800.0, + "reward": 0.6852678656578064, + "reward_std": 0.13185282051563263, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1374 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2991.0, + "completions/mean_length": 966.8582763671875, + "completions/mean_terminated_length": 533.4700927734375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 12.848979591836734, + "grad_norm": 0.175306037068367, + "learning_rate": 1e-06, + "loss": -0.0625, + "num_tokens": 805334625.0, + "reward": 0.7433035969734192, + "reward_std": 0.12219788879156113, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1375 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4048.0, + "completions/mean_length": 944.0826416015625, + "completions/mean_terminated_length": 626.5675659179688, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 12.858309037900874, + "grad_norm": 0.1883779615163803, + "learning_rate": 1e-06, + "loss": -0.0553, + "num_tokens": 805932691.0, + "reward": 0.6808035969734192, + "reward_std": 0.16409297287464142, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1376 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2510.0, + "completions/mean_length": 1091.3348388671875, + "completions/mean_terminated_length": 648.9065551757812, + "completions/min_length": 201.0, + "completions/min_terminated_length": 201.0, + "epoch": 12.867638483965015, + "grad_norm": 0.1939994990825653, + "learning_rate": 1e-06, + "loss": -0.0448, + "num_tokens": 806524471.0, + "reward": 0.6752232313156128, + "reward_std": 0.17096956074237823, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1377 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3295.0, + "completions/mean_length": 1127.0201416015625, + "completions/mean_terminated_length": 698.5465698242188, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 12.876967930029155, + "grad_norm": 0.15887637436389923, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 807174489.0, + "reward": 0.6082589626312256, + "reward_std": 0.14590120315551758, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.4884119927883148, + "step": 1378 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2641.0, + "completions/mean_length": 972.5670166015625, + "completions/mean_terminated_length": 606.4788208007812, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.886297376093294, + "grad_norm": 0.16199131309986115, + "learning_rate": 1e-06, + "loss": -0.0479, + "num_tokens": 807757877.0, + "reward": 0.6986607313156128, + "reward_std": 0.13767912983894348, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1379 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4094.0, + "completions/mean_length": 1061.1629638671875, + "completions/mean_terminated_length": 666.9785766601562, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.895626822157434, + "grad_norm": 0.15144431591033936, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 808373279.0, + "reward": 0.6796875596046448, + "reward_std": 0.1423000991344452, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1380 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4003.0, + "completions/mean_length": 1103.798095703125, + "completions/mean_terminated_length": 636.62841796875, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 12.904956268221575, + "grad_norm": 0.1728334277868271, + "learning_rate": 1e-06, + "loss": -0.0717, + "num_tokens": 808964178.0, + "reward": 0.6852678656578064, + "reward_std": 0.1537972241640091, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3250.0, + "completions/mean_length": 943.4844360351562, + "completions/mean_terminated_length": 587.113037109375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 12.914285714285715, + "grad_norm": 0.16759993135929108, + "learning_rate": 1e-06, + "loss": -0.066, + "num_tokens": 809517348.0, + "reward": 0.7176339626312256, + "reward_std": 0.14699594676494598, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1382 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3710.0, + "completions/mean_length": 899.9944458007812, + "completions/mean_terminated_length": 565.024658203125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 12.923615160349854, + "grad_norm": 0.15316635370254517, + "learning_rate": 1e-06, + "loss": -0.077, + "num_tokens": 810073471.0, + "reward": 0.7589285969734192, + "reward_std": 0.1394015997648239, + "rewards/verify_math_reward/mean": 0.7589285969734192, + "rewards/verify_math_reward/std": 0.4279724657535553, + "step": 1383 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3765.0, + "completions/mean_length": 992.0592041015625, + "completions/mean_terminated_length": 593.3161010742188, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 12.932944606413994, + "grad_norm": 0.15738971531391144, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 810636988.0, + "reward": 0.6830357313156128, + "reward_std": 0.14105269312858582, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1384 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3928.0, + "completions/mean_length": 1029.4598388671875, + "completions/mean_terminated_length": 609.1725463867188, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 12.942274052478133, + "grad_norm": 0.17590700089931488, + "learning_rate": 1e-06, + "loss": -0.0543, + "num_tokens": 811206224.0, + "reward": 0.7020089626312256, + "reward_std": 0.15980690717697144, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1385 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3679.0, + "completions/mean_length": 1005.1551513671875, + "completions/mean_terminated_length": 629.921142578125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 12.951603498542275, + "grad_norm": 0.17235314846038818, + "learning_rate": 1e-06, + "loss": -0.0671, + "num_tokens": 811795747.0, + "reward": 0.6752232313156128, + "reward_std": 0.15480685234069824, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1386 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2306.0, + "completions/mean_length": 989.07373046875, + "completions/mean_terminated_length": 540.6921997070312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 12.960932944606414, + "grad_norm": 0.15123394131660461, + "learning_rate": 1e-06, + "loss": -0.0534, + "num_tokens": 812315693.0, + "reward": 0.6752232313156128, + "reward_std": 0.11370452493429184, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1387 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3801.0, + "completions/mean_length": 974.4051513671875, + "completions/mean_terminated_length": 586.6536865234375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 12.970262390670554, + "grad_norm": 0.19044129550457, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 812870472.0, + "reward": 0.6540178656578064, + "reward_std": 0.1327211558818817, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 1037.0592041015625, + "completions/mean_terminated_length": 573.106689453125, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 12.979591836734693, + "grad_norm": 0.18398486077785492, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 813407565.0, + "reward": 0.684151828289032, + "reward_std": 0.13226650655269623, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1389 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3294.0, + "completions/mean_length": 1155.37841796875, + "completions/mean_terminated_length": 642.79296875, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 12.988921282798835, + "grad_norm": 0.14590129256248474, + "learning_rate": 1e-06, + "loss": -0.0106, + "num_tokens": 813986584.0, + "reward": 0.660714328289032, + "reward_std": 0.10521142929792404, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1390 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1335227272727273, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2960.0, + "completions/mean_length": 1141.0738525390625, + "completions/mean_terminated_length": 685.724609375, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 12.998250728862974, + "grad_norm": 0.14408764243125916, + "learning_rate": 1e-06, + "loss": -0.0364, + "num_tokens": 814566743.0, + "reward": 0.6629464626312256, + "reward_std": 0.11077672243118286, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1391 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3209.0, + "completions/mean_length": 1020.1239013671875, + "completions/mean_terminated_length": 598.5570678710938, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 13.00932944606414, + "grad_norm": 0.17730410397052765, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 815127054.0, + "reward": 0.6808035969734192, + "reward_std": 0.15300559997558594, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1392 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2503.0, + "completions/mean_length": 1016.7667846679688, + "completions/mean_terminated_length": 612.4229736328125, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 13.018658892128279, + "grad_norm": 0.13917987048625946, + "learning_rate": 1e-06, + "loss": -0.0904, + "num_tokens": 815691461.0, + "reward": 0.7410714626312256, + "reward_std": 0.14838533103466034, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1393 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3556.0, + "completions/mean_length": 1146.7523193359375, + "completions/mean_terminated_length": 664.1480102539062, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 13.02798833819242, + "grad_norm": 0.17217010259628296, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 816293383.0, + "reward": 0.6026785969734192, + "reward_std": 0.16251108050346375, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 1394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2426.0, + "completions/mean_length": 1027.30029296875, + "completions/mean_terminated_length": 561.8676147460938, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 13.03731778425656, + "grad_norm": 0.17021389305591583, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 816818044.0, + "reward": 0.6774553656578064, + "reward_std": 0.11768428236246109, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1395 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2936.0, + "completions/mean_length": 979.47998046875, + "completions/mean_terminated_length": 601.12890625, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 13.0466472303207, + "grad_norm": 0.14810553193092346, + "learning_rate": 1e-06, + "loss": -0.0726, + "num_tokens": 817390386.0, + "reward": 0.7120535969734192, + "reward_std": 0.13978277146816254, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1396 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3384.0, + "completions/mean_length": 907.607177734375, + "completions/mean_terminated_length": 573.4352416992188, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 13.055976676384839, + "grad_norm": 0.1687489151954651, + "learning_rate": 1e-06, + "loss": -0.0771, + "num_tokens": 817948394.0, + "reward": 0.7377232313156128, + "reward_std": 0.141063392162323, + "rewards/verify_math_reward/mean": 0.7377232313156128, + "rewards/verify_math_reward/std": 0.4401180148124695, + "step": 1397 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3775.0, + "completions/mean_length": 881.33154296875, + "completions/mean_terminated_length": 561.8367919921875, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 13.06530612244898, + "grad_norm": 0.15694163739681244, + "learning_rate": 1e-06, + "loss": -0.0438, + "num_tokens": 818484179.0, + "reward": 0.7656250596046448, + "reward_std": 0.1230994164943695, + "rewards/verify_math_reward/mean": 0.765625, + "rewards/verify_math_reward/std": 0.4238441288471222, + "step": 1398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 1042.997802734375, + "completions/mean_terminated_length": 620.1549682617188, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 13.07463556851312, + "grad_norm": 0.16730216145515442, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 819060665.0, + "reward": 0.7243303656578064, + "reward_std": 0.1439918875694275, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1399 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3591.0, + "completions/mean_length": 1065.825927734375, + "completions/mean_terminated_length": 632.94384765625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.08396501457726, + "grad_norm": 0.14252914488315582, + "learning_rate": 1e-06, + "loss": -0.0569, + "num_tokens": 819653829.0, + "reward": 0.6897321939468384, + "reward_std": 0.12531296908855438, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1400 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2812.0, + "completions/mean_length": 1029.62060546875, + "completions/mean_terminated_length": 622.5790405273438, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 13.093294460641399, + "grad_norm": 0.1579921394586563, + "learning_rate": 1e-06, + "loss": -0.0617, + "num_tokens": 820228121.0, + "reward": 0.699776828289032, + "reward_std": 0.12433655560016632, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1401 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3508.0, + "completions/mean_length": 1014.5324096679688, + "completions/mean_terminated_length": 618.67626953125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 13.102623906705539, + "grad_norm": 0.16034074127674103, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 820811574.0, + "reward": 0.7176339626312256, + "reward_std": 0.1167111024260521, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1402 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3382.0, + "completions/mean_length": 1077.075927734375, + "completions/mean_terminated_length": 592.1709594726562, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 13.11195335276968, + "grad_norm": 0.15379659831523895, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 821364090.0, + "reward": 0.6272321939468384, + "reward_std": 0.12952163815498352, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 1403 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3994.0, + "completions/mean_length": 1103.0826416015625, + "completions/mean_terminated_length": 613.3324584960938, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 13.12128279883382, + "grad_norm": 0.18127885460853577, + "learning_rate": 1e-06, + "loss": -0.0307, + "num_tokens": 821927876.0, + "reward": 0.6540178656578064, + "reward_std": 0.1534927487373352, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1741071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3765.0, + "completions/mean_length": 1308.747802734375, + "completions/mean_terminated_length": 721.1648559570312, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 13.130612244897959, + "grad_norm": 0.15667860209941864, + "learning_rate": 1e-06, + "loss": -0.0814, + "num_tokens": 822557458.0, + "reward": 0.5959821939468384, + "reward_std": 0.15409007668495178, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 1405 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3071.0, + "completions/mean_length": 1049.6373291015625, + "completions/mean_terminated_length": 658.2908935546875, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 13.139941690962099, + "grad_norm": 0.15307126939296722, + "learning_rate": 1e-06, + "loss": -0.0633, + "num_tokens": 823174285.0, + "reward": 0.6852678656578064, + "reward_std": 0.15015265345573425, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1406 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3257.0, + "completions/mean_length": 1007.5379638671875, + "completions/mean_terminated_length": 588.6970825195312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 13.14927113702624, + "grad_norm": 0.16991931200027466, + "learning_rate": 1e-06, + "loss": -0.0854, + "num_tokens": 823719647.0, + "reward": 0.7198660969734192, + "reward_std": 0.1423753798007965, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1407 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3353.0, + "completions/mean_length": 1148.9453125, + "completions/mean_terminated_length": 639.768310546875, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 13.15860058309038, + "grad_norm": 0.16819636523723602, + "learning_rate": 1e-06, + "loss": -0.04, + "num_tokens": 824302334.0, + "reward": 0.6473214626312256, + "reward_std": 0.1327543556690216, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 1408 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2709.0, + "completions/mean_length": 1028.9207763671875, + "completions/mean_terminated_length": 586.2899169921875, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 13.167930029154519, + "grad_norm": 0.13427141308784485, + "learning_rate": 1e-06, + "loss": -0.0236, + "num_tokens": 824867991.0, + "reward": 0.606026828289032, + "reward_std": 0.08691229671239853, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 1409 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3997.0, + "completions/mean_length": 855.1674194335938, + "completions/mean_terminated_length": 589.0120849609375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 13.177259475218658, + "grad_norm": 0.14154063165187836, + "learning_rate": 1e-06, + "loss": -0.0307, + "num_tokens": 825449893.0, + "reward": 0.7868303656578064, + "reward_std": 0.12185201048851013, + "rewards/verify_math_reward/mean": 0.7868303656578064, + "rewards/verify_math_reward/std": 0.4097752273082733, + "step": 1410 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3054.0, + "completions/mean_length": 944.44873046875, + "completions/mean_terminated_length": 570.6691284179688, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 13.186588921282798, + "grad_norm": 0.17151287198066711, + "learning_rate": 1e-06, + "loss": -0.0734, + "num_tokens": 825991383.0, + "reward": 0.7566964626312256, + "reward_std": 0.1796049177646637, + "rewards/verify_math_reward/mean": 0.7566964030265808, + "rewards/verify_math_reward/std": 0.4293164908885956, + "step": 1411 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3276.0, + "completions/mean_length": 1087.4754638671875, + "completions/mean_terminated_length": 657.6862182617188, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 13.19591836734694, + "grad_norm": 0.17296727001667023, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 826601857.0, + "reward": 0.6584821939468384, + "reward_std": 0.16462108492851257, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1412 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2910.0, + "completions/mean_length": 997.4810791015625, + "completions/mean_terminated_length": 586.1732177734375, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 13.205247813411079, + "grad_norm": 0.16448399424552917, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 827154792.0, + "reward": 0.6986607313156128, + "reward_std": 0.12509429454803467, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3051.0, + "completions/mean_length": 1113.55029296875, + "completions/mean_terminated_length": 602.830078125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 13.214577259475218, + "grad_norm": 0.16226400434970856, + "learning_rate": 1e-06, + "loss": -0.0734, + "num_tokens": 827704653.0, + "reward": 0.6629464626312256, + "reward_std": 0.13891445100307465, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1414 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3549.0, + "completions/mean_length": 1144.03466796875, + "completions/mean_terminated_length": 615.7882080078125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.223906705539358, + "grad_norm": 0.17290173470973969, + "learning_rate": 1e-06, + "loss": -0.0697, + "num_tokens": 828269148.0, + "reward": 0.6897321939468384, + "reward_std": 0.1392175406217575, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1415 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2882.0, + "completions/mean_length": 984.1529541015625, + "completions/mean_terminated_length": 562.1406860351562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 13.2332361516035, + "grad_norm": 0.15680143237113953, + "learning_rate": 1e-06, + "loss": -0.0664, + "num_tokens": 828808541.0, + "reward": 0.6718750596046448, + "reward_std": 0.11963889747858047, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1416 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3495.0, + "completions/mean_length": 1050.0023193359375, + "completions/mean_terminated_length": 658.7027587890625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 13.242565597667639, + "grad_norm": 0.1522519737482071, + "learning_rate": 1e-06, + "loss": -0.0701, + "num_tokens": 829411791.0, + "reward": 0.707589328289032, + "reward_std": 0.14466266334056854, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 1018.5535888671875, + "completions/mean_terminated_length": 662.136962890625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 13.251895043731778, + "grad_norm": 0.16142745316028595, + "learning_rate": 1e-06, + "loss": -0.0717, + "num_tokens": 830036711.0, + "reward": 0.7109375596046448, + "reward_std": 0.1527452915906906, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1418 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2828.0, + "completions/mean_length": 1015.529052734375, + "completions/mean_terminated_length": 675.7992553710938, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 13.261224489795918, + "grad_norm": 0.13795799016952515, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 830670905.0, + "reward": 0.6863839626312256, + "reward_std": 0.1266666203737259, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1419 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3457.0, + "completions/mean_length": 1000.4766235351562, + "completions/mean_terminated_length": 576.2169799804688, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 13.270553935860057, + "grad_norm": 0.15413826704025269, + "learning_rate": 1e-06, + "loss": -0.0933, + "num_tokens": 831219844.0, + "reward": 0.7254464626312256, + "reward_std": 0.1308746039867401, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1420 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3217.0, + "completions/mean_length": 956.7756958007812, + "completions/mean_terminated_length": 571.2568969726562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 13.279883381924199, + "grad_norm": 0.16560636460781097, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 831756555.0, + "reward": 0.6774553656578064, + "reward_std": 0.11678525805473328, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1421 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2800.0, + "completions/mean_length": 964.3594360351562, + "completions/mean_terminated_length": 601.666259765625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 13.289212827988338, + "grad_norm": 0.15646150708198547, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 832331765.0, + "reward": 0.7299107313156128, + "reward_std": 0.15879730880260468, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1422 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2770.0, + "completions/mean_length": 917.2756958007812, + "completions/mean_terminated_length": 618.4212646484375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 13.298542274052478, + "grad_norm": 0.14893360435962677, + "learning_rate": 1e-06, + "loss": -0.0359, + "num_tokens": 832927580.0, + "reward": 0.7187500596046448, + "reward_std": 0.1252797544002533, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 1110.5123291015625, + "completions/mean_terminated_length": 635.4605712890625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.307871720116617, + "grad_norm": 0.1834036409854889, + "learning_rate": 1e-06, + "loss": -0.0647, + "num_tokens": 833512215.0, + "reward": 0.6316964626312256, + "reward_std": 0.15146607160568237, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1424 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3907.0, + "completions/mean_length": 1120.1741943359375, + "completions/mean_terminated_length": 637.7120361328125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 13.317201166180759, + "grad_norm": 0.16921481490135193, + "learning_rate": 1e-06, + "loss": -0.051, + "num_tokens": 834084115.0, + "reward": 0.6462053656578064, + "reward_std": 0.16146548092365265, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1425 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3087.0, + "completions/mean_length": 1055.59375, + "completions/mean_terminated_length": 638.8883056640625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 13.326530612244898, + "grad_norm": 0.14250235259532928, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 834672135.0, + "reward": 0.7087053656578064, + "reward_std": 0.1195308193564415, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3801.0, + "completions/mean_length": 1217.7388916015625, + "completions/mean_terminated_length": 671.1367797851562, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 13.335860058309038, + "grad_norm": 0.1670420616865158, + "learning_rate": 1e-06, + "loss": -0.1058, + "num_tokens": 835264365.0, + "reward": 0.6629464626312256, + "reward_std": 0.155109241604805, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296735644340515, + "step": 1427 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3048.0, + "completions/mean_length": 1066.810302734375, + "completions/mean_terminated_length": 634.06884765625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 13.345189504373177, + "grad_norm": 0.16828158497810364, + "learning_rate": 1e-06, + "loss": -0.0657, + "num_tokens": 835849755.0, + "reward": 0.684151828289032, + "reward_std": 0.142863929271698, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1428 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2771.0, + "completions/mean_length": 951.6685791015625, + "completions/mean_terminated_length": 574.3487548828125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 13.354518950437317, + "grad_norm": 0.15304379165172577, + "learning_rate": 1e-06, + "loss": -0.0651, + "num_tokens": 836391170.0, + "reward": 0.7756696939468384, + "reward_std": 0.10494860261678696, + "rewards/verify_math_reward/mean": 0.7756696343421936, + "rewards/verify_math_reward/std": 0.41737356781959534, + "step": 1429 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2904.0, + "completions/mean_length": 1015.2444458007812, + "completions/mean_terminated_length": 615.0958251953125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 13.363848396501458, + "grad_norm": 0.16295866668224335, + "learning_rate": 1e-06, + "loss": -0.0293, + "num_tokens": 836959005.0, + "reward": 0.6953125596046448, + "reward_std": 0.133317768573761, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1430 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2946.0, + "completions/mean_length": 922.7824096679688, + "completions/mean_terminated_length": 568.4528198242188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.373177842565598, + "grad_norm": 0.15209509432315826, + "learning_rate": 1e-06, + "loss": -0.0646, + "num_tokens": 837512274.0, + "reward": 0.7477678656578064, + "reward_std": 0.12869791686534882, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.4345363676548004, + "step": 1431 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3202.0, + "completions/mean_length": 1049.3582763671875, + "completions/mean_terminated_length": 649.294189453125, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 13.382507288629737, + "grad_norm": 0.11591311544179916, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 838118523.0, + "reward": 0.6852678656578064, + "reward_std": 0.10926789045333862, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1432 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2802.0, + "completions/mean_length": 1029.0, + "completions/mean_terminated_length": 613.0697021484375, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 13.391836734693877, + "grad_norm": 0.14781732857227325, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 838696707.0, + "reward": 0.7165178656578064, + "reward_std": 0.13621236383914948, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1433 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 993.5201416015625, + "completions/mean_terminated_length": 599.3685302734375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 13.401166180758018, + "grad_norm": 0.14631949365139008, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 839265581.0, + "reward": 0.7366071939468384, + "reward_std": 0.11963889747858047, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1434 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3483.0, + "completions/mean_length": 985.9710083007812, + "completions/mean_terminated_length": 625.7808227539062, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 13.410495626822158, + "grad_norm": 0.1488763839006424, + "learning_rate": 1e-06, + "loss": -0.0655, + "num_tokens": 839847291.0, + "reward": 0.7444196939468384, + "reward_std": 0.1312882900238037, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1435 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3858.0, + "completions/mean_length": 1044.7288818359375, + "completions/mean_terminated_length": 622.1257934570312, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.419825072886297, + "grad_norm": 0.15852481126785278, + "learning_rate": 1e-06, + "loss": -0.0832, + "num_tokens": 840435592.0, + "reward": 0.7020089626312256, + "reward_std": 0.13872967660427094, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763102173805237, + "step": 1436 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3144.0, + "completions/mean_length": 1010.3594360351562, + "completions/mean_terminated_length": 670.0595092773438, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 13.429154518950437, + "grad_norm": 0.1600950062274933, + "learning_rate": 1e-06, + "loss": -0.0787, + "num_tokens": 841063146.0, + "reward": 0.6875000596046448, + "reward_std": 0.1607832908630371, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1437 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2799.0, + "completions/mean_length": 963.05810546875, + "completions/mean_terminated_length": 613.225830078125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.438483965014576, + "grad_norm": 0.1528957635164261, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 841648974.0, + "reward": 0.7254464626312256, + "reward_std": 0.12230778485536575, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1438 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1841517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3588.0, + "completions/mean_length": 1295.3895263671875, + "completions/mean_terminated_length": 663.2407836914062, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.447813411078718, + "grad_norm": 0.18343259394168854, + "learning_rate": 1e-06, + "loss": -0.0611, + "num_tokens": 842229339.0, + "reward": 0.606026828289032, + "reward_std": 0.1523655205965042, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 1439 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3159.0, + "completions/mean_length": 1078.43310546875, + "completions/mean_terminated_length": 656.1272583007812, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.457142857142857, + "grad_norm": 0.15937013924121857, + "learning_rate": 1e-06, + "loss": -0.0705, + "num_tokens": 842828407.0, + "reward": 0.7265625596046448, + "reward_std": 0.1368100941181183, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1440 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3625.0, + "completions/mean_length": 1174.8270263671875, + "completions/mean_terminated_length": 661.1299438476562, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 13.466472303206997, + "grad_norm": 0.14537642896175385, + "learning_rate": 1e-06, + "loss": -0.0878, + "num_tokens": 843423892.0, + "reward": 0.65625, + "reward_std": 0.11283829808235168, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1441 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 1126.46435546875, + "completions/mean_terminated_length": 671.6705322265625, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 13.475801749271136, + "grad_norm": 0.13487359881401062, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 844032516.0, + "reward": 0.7087053656578064, + "reward_std": 0.10716354846954346, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1442 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2837.0, + "completions/mean_length": 1031.4598388671875, + "completions/mean_terminated_length": 624.6624755859375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.485131195335278, + "grad_norm": 0.16721223294734955, + "learning_rate": 1e-06, + "loss": -0.0757, + "num_tokens": 844626440.0, + "reward": 0.6897321939468384, + "reward_std": 0.1381341814994812, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1443 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2995.0, + "completions/mean_length": 1031.7020263671875, + "completions/mean_terminated_length": 642.4012451171875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.494460641399417, + "grad_norm": 0.1554444581270218, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 845225237.0, + "reward": 0.7421875596046448, + "reward_std": 0.14045608043670654, + "rewards/verify_math_reward/mean": 0.7421875, + "rewards/verify_math_reward/std": 0.43767455220222473, + "step": 1444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3389.0, + "completions/mean_length": 1072.3092041015625, + "completions/mean_terminated_length": 644.7554321289062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 13.503790087463557, + "grad_norm": 0.13712087273597717, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 845822314.0, + "reward": 0.6975446939468384, + "reward_std": 0.09818120300769806, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3868.0, + "completions/mean_length": 1154.4320068359375, + "completions/mean_terminated_length": 655.210205078125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.513119533527696, + "grad_norm": 0.15192073583602905, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 846408069.0, + "reward": 0.6328125, + "reward_std": 0.126701220870018, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1540178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3258.0, + "completions/mean_length": 1176.661865234375, + "completions/mean_terminated_length": 645.1728515625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 13.522448979591836, + "grad_norm": 0.17441591620445251, + "learning_rate": 1e-06, + "loss": -0.0833, + "num_tokens": 846999054.0, + "reward": 0.6506696939468384, + "reward_std": 0.134780615568161, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1447 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1696428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3028.0, + "completions/mean_length": 1196.786865234375, + "completions/mean_terminated_length": 604.4744873046875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.531778425655977, + "grad_norm": 0.16436070203781128, + "learning_rate": 1e-06, + "loss": -0.073, + "num_tokens": 847529783.0, + "reward": 0.668526828289032, + "reward_std": 0.11603710055351257, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1448 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3672.0, + "completions/mean_length": 961.1964721679688, + "completions/mean_terminated_length": 636.9063720703125, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 13.541107871720117, + "grad_norm": 0.16785281896591187, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 848136631.0, + "reward": 0.7142857313156128, + "reward_std": 0.13970720767974854, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1449 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3163.0, + "completions/mean_length": 1086.58935546875, + "completions/mean_terminated_length": 621.2164916992188, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 13.550437317784256, + "grad_norm": 0.15177418291568756, + "learning_rate": 1e-06, + "loss": -0.0672, + "num_tokens": 848719183.0, + "reward": 0.684151828289032, + "reward_std": 0.12257696688175201, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1450 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2932.0, + "completions/mean_length": 1105.665283203125, + "completions/mean_terminated_length": 643.2422485351562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 13.559766763848396, + "grad_norm": 0.16427406668663025, + "learning_rate": 1e-06, + "loss": -0.093, + "num_tokens": 849298603.0, + "reward": 0.7131696939468384, + "reward_std": 0.1504141092300415, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 1451 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2985.0, + "completions/mean_length": 1098.6707763671875, + "completions/mean_terminated_length": 617.2344360351562, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 13.569096209912537, + "grad_norm": 0.15991352498531342, + "learning_rate": 1e-06, + "loss": -0.0188, + "num_tokens": 849858340.0, + "reward": 0.676339328289032, + "reward_std": 0.12204564362764359, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1452 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3240.0, + "completions/mean_length": 984.9989013671875, + "completions/mean_terminated_length": 629.013671875, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 13.578425655976677, + "grad_norm": 0.1567736268043518, + "learning_rate": 1e-06, + "loss": -0.0654, + "num_tokens": 850450315.0, + "reward": 0.7265625596046448, + "reward_std": 0.16096805036067963, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1453 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3973.0, + "completions/mean_length": 1160.5491943359375, + "completions/mean_terminated_length": 644.3411865234375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 13.587755102040816, + "grad_norm": 0.17545689642429352, + "learning_rate": 1e-06, + "loss": -0.0674, + "num_tokens": 851034815.0, + "reward": 0.668526828289032, + "reward_std": 0.17017818987369537, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1454 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3393.0, + "completions/mean_length": 1112.888427734375, + "completions/mean_terminated_length": 647.1380615234375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.597084548104956, + "grad_norm": 0.14614875614643097, + "learning_rate": 1e-06, + "loss": -0.0718, + "num_tokens": 851632619.0, + "reward": 0.6796875596046448, + "reward_std": 0.142819344997406, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1455 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3292.0, + "completions/mean_length": 1167.3382568359375, + "completions/mean_terminated_length": 652.3241577148438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 13.606413994169095, + "grad_norm": 0.1602410078048706, + "learning_rate": 1e-06, + "loss": -0.0858, + "num_tokens": 852226066.0, + "reward": 0.660714328289032, + "reward_std": 0.15454654395580292, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1456 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1540178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3457.0, + "completions/mean_length": 1190.6171875, + "completions/mean_terminated_length": 661.6688842773438, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 13.615743440233237, + "grad_norm": 0.16519398987293243, + "learning_rate": 1e-06, + "loss": -0.0839, + "num_tokens": 852814107.0, + "reward": 0.6707589626312256, + "reward_std": 0.16150008141994476, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3444.0, + "completions/mean_length": 883.60498046875, + "completions/mean_terminated_length": 585.8707275390625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.625072886297376, + "grad_norm": 0.13896900415420532, + "learning_rate": 1e-06, + "loss": -0.0679, + "num_tokens": 853383617.0, + "reward": 0.7488839626312256, + "reward_std": 0.1244862899184227, + "rewards/verify_math_reward/mean": 0.7488839030265808, + "rewards/verify_math_reward/std": 0.43389734625816345, + "step": 1458 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1662946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3897.0, + "completions/mean_length": 1274.8013916015625, + "completions/mean_terminated_length": 712.072265625, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 13.634402332361516, + "grad_norm": 0.1437322199344635, + "learning_rate": 1e-06, + "loss": -0.1042, + "num_tokens": 854011367.0, + "reward": 0.6640625, + "reward_std": 0.14687690138816833, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1459 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1640625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3815.0, + "completions/mean_length": 1254.2467041015625, + "completions/mean_terminated_length": 696.5194091796875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 13.643731778425655, + "grad_norm": 0.15712018311023712, + "learning_rate": 1e-06, + "loss": -0.0497, + "num_tokens": 854627884.0, + "reward": 0.6640625, + "reward_std": 0.1254630982875824, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1460 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3077.0, + "completions/mean_length": 1101.5614013671875, + "completions/mean_terminated_length": 704.069580078125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.653061224489797, + "grad_norm": 0.16194351017475128, + "learning_rate": 1e-06, + "loss": -0.0608, + "num_tokens": 855265315.0, + "reward": 0.7187500596046448, + "reward_std": 0.1452290266752243, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3717.0, + "completions/mean_length": 1066.774658203125, + "completions/mean_terminated_length": 647.224853515625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 13.662390670553936, + "grad_norm": 0.15248115360736847, + "learning_rate": 1e-06, + "loss": -0.0998, + "num_tokens": 855866009.0, + "reward": 0.7198660969734192, + "reward_std": 0.17502740025520325, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1462 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1808035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2933.0, + "completions/mean_length": 1301.790283203125, + "completions/mean_terminated_length": 685.08447265625, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 13.671720116618076, + "grad_norm": 0.1724405586719513, + "learning_rate": 1e-06, + "loss": -0.108, + "num_tokens": 856456301.0, + "reward": 0.629464328289032, + "reward_std": 0.1612718552350998, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 1463 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3559.0, + "completions/mean_length": 1219.5234375, + "completions/mean_terminated_length": 682.3271484375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 13.681049562682215, + "grad_norm": 0.17135068774223328, + "learning_rate": 1e-06, + "loss": -0.0718, + "num_tokens": 857056146.0, + "reward": 0.6339285969734192, + "reward_std": 0.14188753068447113, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199838399887085, + "step": 1464 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3272.0, + "completions/mean_length": 1029.54248046875, + "completions/mean_terminated_length": 644.3090209960938, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 13.690379008746355, + "grad_norm": 0.18565014004707336, + "learning_rate": 1e-06, + "loss": -0.0923, + "num_tokens": 857665456.0, + "reward": 0.6752232313156128, + "reward_std": 0.195464089512825, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1465 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3641.0, + "completions/mean_length": 1109.1373291015625, + "completions/mean_terminated_length": 615.8582763671875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 13.699708454810496, + "grad_norm": 0.1436578333377838, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 858217419.0, + "reward": 0.7198660969734192, + "reward_std": 0.11712157726287842, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931530952453613, + "step": 1466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3830.0, + "completions/mean_length": 1113.279052734375, + "completions/mean_terminated_length": 674.0819702148438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 13.709037900874636, + "grad_norm": 0.1680949181318283, + "learning_rate": 1e-06, + "loss": -0.0769, + "num_tokens": 858827861.0, + "reward": 0.6808035969734192, + "reward_std": 0.16319003701210022, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3761.0, + "completions/mean_length": 1127.1507568359375, + "completions/mean_terminated_length": 618.759521484375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 13.718367346938775, + "grad_norm": 0.16500741243362427, + "learning_rate": 1e-06, + "loss": -0.0526, + "num_tokens": 859402732.0, + "reward": 0.6428571939468384, + "reward_std": 0.12486536800861359, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1468 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2987.0, + "completions/mean_length": 1211.954345703125, + "completions/mean_terminated_length": 740.0194702148438, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 13.727696793002915, + "grad_norm": 0.16611486673355103, + "learning_rate": 1e-06, + "loss": -0.0643, + "num_tokens": 860063155.0, + "reward": 0.629464328289032, + "reward_std": 0.1650354564189911, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 1469 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3826.0, + "completions/mean_length": 1116.453125, + "completions/mean_terminated_length": 668.9473876953125, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 13.737026239067056, + "grad_norm": 0.13615849614143372, + "learning_rate": 1e-06, + "loss": -0.0675, + "num_tokens": 860663977.0, + "reward": 0.7109375596046448, + "reward_std": 0.13440224528312683, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1470 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 1056.774658203125, + "completions/mean_terminated_length": 627.0242309570312, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 13.746355685131196, + "grad_norm": 0.18000447750091553, + "learning_rate": 1e-06, + "loss": -0.069, + "num_tokens": 861244527.0, + "reward": 0.6953125596046448, + "reward_std": 0.15563170611858368, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1471 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3203.0, + "completions/mean_length": 1155.7913818359375, + "completions/mean_terminated_length": 687.9443969726562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 13.755685131195335, + "grad_norm": 0.1400819569826126, + "learning_rate": 1e-06, + "loss": -0.0578, + "num_tokens": 861875052.0, + "reward": 0.6718750596046448, + "reward_std": 0.13218912482261658, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1472 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3928.0, + "completions/mean_length": 1268.798095703125, + "completions/mean_terminated_length": 682.0202026367188, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 13.765014577259475, + "grad_norm": 0.15747207403182983, + "learning_rate": 1e-06, + "loss": -0.0641, + "num_tokens": 862465103.0, + "reward": 0.6350446939468384, + "reward_std": 0.14181266725063324, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 1473 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3793.0, + "completions/mean_length": 1195.829345703125, + "completions/mean_terminated_length": 654.2079467773438, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.774344023323614, + "grad_norm": 0.17201298475265503, + "learning_rate": 1e-06, + "loss": -0.0648, + "num_tokens": 863056934.0, + "reward": 0.6428571939468384, + "reward_std": 0.14034800231456757, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1474 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4030.0, + "completions/mean_length": 1254.54248046875, + "completions/mean_terminated_length": 741.657470703125, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 13.783673469387756, + "grad_norm": 0.14561475813388824, + "learning_rate": 1e-06, + "loss": -0.0767, + "num_tokens": 863713596.0, + "reward": 0.6261160969734192, + "reward_std": 0.15244106948375702, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 1475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2693.0, + "completions/mean_length": 1020.22998046875, + "completions/mean_terminated_length": 594.2337646484375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 13.793002915451895, + "grad_norm": 0.14293834567070007, + "learning_rate": 1e-06, + "loss": -0.0823, + "num_tokens": 864277786.0, + "reward": 0.7064732313156128, + "reward_std": 0.1345556080341339, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1476 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1863839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4043.0, + "completions/mean_length": 1327.759033203125, + "completions/mean_terminated_length": 693.607666015625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 13.802332361516035, + "grad_norm": 0.1671881228685379, + "learning_rate": 1e-06, + "loss": -0.0536, + "num_tokens": 864878738.0, + "reward": 0.652901828289032, + "reward_std": 0.1285017728805542, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 1477 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3433.0, + "completions/mean_length": 1096.251220703125, + "completions/mean_terminated_length": 641.2763671875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.811661807580174, + "grad_norm": 0.16302862763404846, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 865456091.0, + "reward": 0.7220982313156128, + "reward_std": 0.1344122290611267, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 1478 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3938.0, + "completions/mean_length": 1222.8929443359375, + "completions/mean_terminated_length": 686.3258666992188, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 13.820991253644316, + "grad_norm": 0.1538170427083969, + "learning_rate": 1e-06, + "loss": -0.0712, + "num_tokens": 866061803.0, + "reward": 0.6316964626312256, + "reward_std": 0.14086836576461792, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1479 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3209.0, + "completions/mean_length": 1122.040283203125, + "completions/mean_terminated_length": 603.6435546875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.830320699708455, + "grad_norm": 0.18746376037597656, + "learning_rate": 1e-06, + "loss": -0.108, + "num_tokens": 866606791.0, + "reward": 0.7444196939468384, + "reward_std": 0.10562513023614883, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1480 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3653.0, + "completions/mean_length": 1001.7232666015625, + "completions/mean_terminated_length": 656.2084350585938, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 13.839650145772595, + "grad_norm": 0.13892732560634613, + "learning_rate": 1e-06, + "loss": -0.0658, + "num_tokens": 867218487.0, + "reward": 0.7265625596046448, + "reward_std": 0.12707959115505219, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1481 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.2444196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4038.0, + "completions/mean_length": 1583.6663818359375, + "completions/mean_terminated_length": 770.9616088867188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.848979591836734, + "grad_norm": 0.1427098512649536, + "learning_rate": 1e-06, + "loss": -0.0763, + "num_tokens": 867840212.0, + "reward": 0.5569196939468384, + "reward_std": 0.11298984289169312, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.49702703952789307, + "step": 1482 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3875.0, + "completions/mean_length": 1114.67529296875, + "completions/mean_terminated_length": 658.075927734375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 13.858309037900874, + "grad_norm": 0.16994047164916992, + "learning_rate": 1e-06, + "loss": -0.0451, + "num_tokens": 868444633.0, + "reward": 0.6417410969734192, + "reward_std": 0.14820027351379395, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1483 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1618303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2932.0, + "completions/mean_length": 1167.53125, + "completions/mean_terminated_length": 602.114501953125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 13.867638483965015, + "grad_norm": 0.15329888463020325, + "learning_rate": 1e-06, + "loss": -0.1178, + "num_tokens": 868986389.0, + "reward": 0.6863839626312256, + "reward_std": 0.1310618817806244, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3082.0, + "completions/mean_length": 1032.5491943359375, + "completions/mean_terminated_length": 599.3732299804688, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 13.876967930029155, + "grad_norm": 0.1547268033027649, + "learning_rate": 1e-06, + "loss": -0.0597, + "num_tokens": 869538297.0, + "reward": 0.6830357313156128, + "reward_std": 0.13613680005073547, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1930803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3757.0, + "completions/mean_length": 1314.49560546875, + "completions/mean_terminated_length": 648.9349975585938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.886297376093294, + "grad_norm": 0.16535474359989166, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 870094389.0, + "reward": 0.6316964626312256, + "reward_std": 0.11802379786968231, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1486 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3750.0, + "completions/mean_length": 1063.477783203125, + "completions/mean_terminated_length": 616.947509765625, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 13.895626822157434, + "grad_norm": 0.1738092452287674, + "learning_rate": 1e-06, + "loss": -0.0681, + "num_tokens": 870665593.0, + "reward": 0.6852678656578064, + "reward_std": 0.14515121281147003, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1487 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3243.0, + "completions/mean_length": 964.1116333007812, + "completions/mean_terminated_length": 610.072021484375, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 13.904956268221575, + "grad_norm": 0.17861664295196533, + "learning_rate": 1e-06, + "loss": -0.0621, + "num_tokens": 871234437.0, + "reward": 0.7488839626312256, + "reward_std": 0.10803297907114029, + "rewards/verify_math_reward/mean": 0.7488839030265808, + "rewards/verify_math_reward/std": 0.43389734625816345, + "step": 1488 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3292.0, + "completions/mean_length": 1137.7098388671875, + "completions/mean_terminated_length": 622.0445556640625, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 13.914285714285715, + "grad_norm": 0.17084911465644836, + "learning_rate": 1e-06, + "loss": -0.1002, + "num_tokens": 871803969.0, + "reward": 0.6741071939468384, + "reward_std": 0.15390713512897491, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 1489 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1696428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3374.0, + "completions/mean_length": 1193.786865234375, + "completions/mean_terminated_length": 600.861572265625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 13.923615160349854, + "grad_norm": 0.16532692313194275, + "learning_rate": 1e-06, + "loss": -0.0845, + "num_tokens": 872339410.0, + "reward": 0.7008928656578064, + "reward_std": 0.13080044090747833, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1490 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.15625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2919.0, + "completions/mean_length": 1206.5491943359375, + "completions/mean_terminated_length": 671.465576171875, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 13.932944606413994, + "grad_norm": 0.1595967411994934, + "learning_rate": 1e-06, + "loss": -0.0833, + "num_tokens": 872934286.0, + "reward": 0.6328125, + "reward_std": 0.1573658436536789, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 1491 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3025.0, + "completions/mean_length": 1162.65185546875, + "completions/mean_terminated_length": 651.334228515625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.942274052478133, + "grad_norm": 0.17703810334205627, + "learning_rate": 1e-06, + "loss": -0.083, + "num_tokens": 873520422.0, + "reward": 0.6361607313156128, + "reward_std": 0.13038857281208038, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1492 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2937.0, + "completions/mean_length": 973.6842041015625, + "completions/mean_terminated_length": 625.0384521484375, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.951603498542275, + "grad_norm": 0.1360289752483368, + "learning_rate": 1e-06, + "loss": -0.0457, + "num_tokens": 874111507.0, + "reward": 0.7176339626312256, + "reward_std": 0.10179509967565536, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1493 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 1146.157470703125, + "completions/mean_terminated_length": 711.8015747070312, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 13.960932944606414, + "grad_norm": 0.1361832320690155, + "learning_rate": 1e-06, + "loss": -0.0827, + "num_tokens": 874763688.0, + "reward": 0.6852678656578064, + "reward_std": 0.1318553388118744, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1494 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3439.0, + "completions/mean_length": 1221.90185546875, + "completions/mean_terminated_length": 676.09033203125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 13.970262390670554, + "grad_norm": 0.15795740485191345, + "learning_rate": 1e-06, + "loss": -0.0838, + "num_tokens": 875353872.0, + "reward": 0.6339285969734192, + "reward_std": 0.13685175776481628, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1495 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3827.0, + "completions/mean_length": 1052.6484375, + "completions/mean_terminated_length": 613.44189453125, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 13.979591836734693, + "grad_norm": 0.13383625447750092, + "learning_rate": 1e-06, + "loss": -0.0295, + "num_tokens": 875919397.0, + "reward": 0.7020089626312256, + "reward_std": 0.09979882091283798, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1496 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3822.0, + "completions/mean_length": 1228.857177734375, + "completions/mean_terminated_length": 724.6614379882812, + "completions/min_length": 192.0, + "completions/min_terminated_length": 192.0, + "epoch": 13.988921282798835, + "grad_norm": 0.15809251368045807, + "learning_rate": 1e-06, + "loss": -0.0791, + "num_tokens": 876564637.0, + "reward": 0.6104910969734192, + "reward_std": 0.15582603216171265, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 1497 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.12784090909090906, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3002.0, + "completions/mean_length": 1121.2471923828125, + "completions/mean_terminated_length": 685.2084350585938, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 13.998250728862974, + "grad_norm": 0.18763110041618347, + "learning_rate": 1e-06, + "loss": -0.0662, + "num_tokens": 877167928.0, + "reward": 0.699776828289032, + "reward_std": 0.12001937627792358, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1498 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3320.0, + "completions/mean_length": 1155.9989013671875, + "completions/mean_terminated_length": 643.5216674804688, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 14.00932944606414, + "grad_norm": 0.14190790057182312, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 877746847.0, + "reward": 0.6540178656578064, + "reward_std": 0.12399843335151672, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1499 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3132.0, + "completions/mean_length": 998.3917846679688, + "completions/mean_terminated_length": 643.9390258789062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 14.018658892128279, + "grad_norm": 0.15186084806919098, + "learning_rate": 1e-06, + "loss": -0.0681, + "num_tokens": 878360262.0, + "reward": 0.7366071939468384, + "reward_std": 0.13180933892726898, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1500 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3048.0, + "completions/mean_length": 1165.5457763671875, + "completions/mean_terminated_length": 708.0167846679688, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 14.02798833819242, + "grad_norm": 0.14011088013648987, + "learning_rate": 1e-06, + "loss": -0.0238, + "num_tokens": 879007503.0, + "reward": 0.6283482313156128, + "reward_std": 0.10960489511489868, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 1501 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3400.0, + "completions/mean_length": 1183.9185791015625, + "completions/mean_terminated_length": 671.8201904296875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 14.03731778425656, + "grad_norm": 0.17413397133350372, + "learning_rate": 1e-06, + "loss": -0.1176, + "num_tokens": 879600558.0, + "reward": 0.6729910969734192, + "reward_std": 0.17036226391792297, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2363.0, + "completions/mean_length": 938.8683471679688, + "completions/mean_terminated_length": 633.5887451171875, + "completions/min_length": 192.0, + "completions/min_terminated_length": 192.0, + "epoch": 14.0466472303207, + "grad_norm": 0.15621769428253174, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 880208280.0, + "reward": 0.738839328289032, + "reward_std": 0.14350402355194092, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3159.0, + "completions/mean_length": 1037.560302734375, + "completions/mean_terminated_length": 644.6624145507812, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.055976676384839, + "grad_norm": 0.12551946938037872, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 880805302.0, + "reward": 0.6785714626312256, + "reward_std": 0.1083688735961914, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1504 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4058.0, + "completions/mean_length": 1039.7366943359375, + "completions/mean_terminated_length": 672.9849853515625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.06530612244898, + "grad_norm": 0.149945929646492, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 881431130.0, + "reward": 0.6875000596046448, + "reward_std": 0.15067441761493683, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3811.0, + "completions/mean_length": 1150.9498291015625, + "completions/mean_terminated_length": 651.1370849609375, + "completions/min_length": 203.0, + "completions/min_terminated_length": 203.0, + "epoch": 14.07463556851312, + "grad_norm": 0.15435577929019928, + "learning_rate": 1e-06, + "loss": -0.0657, + "num_tokens": 882024669.0, + "reward": 0.6886160969734192, + "reward_std": 0.12741729617118835, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1506 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1752232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3842.0, + "completions/mean_length": 1268.344970703125, + "completions/mean_terminated_length": 667.6116333007812, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 14.08396501457726, + "grad_norm": 0.18073588609695435, + "learning_rate": 1e-06, + "loss": -0.1257, + "num_tokens": 882606210.0, + "reward": 0.6718750596046448, + "reward_std": 0.15326520800590515, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1507 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 1084.6060791015625, + "completions/mean_terminated_length": 623.4015502929688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 14.093294460641399, + "grad_norm": 0.18910661339759827, + "learning_rate": 1e-06, + "loss": -0.07, + "num_tokens": 883178497.0, + "reward": 0.699776828289032, + "reward_std": 0.15064053237438202, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1508 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3268.0, + "completions/mean_length": 1090.8426513671875, + "completions/mean_terminated_length": 630.5933227539062, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 14.102623906705539, + "grad_norm": 0.13664594292640686, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 883759900.0, + "reward": 0.7287946939468384, + "reward_std": 0.1169707253575325, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483014941215515, + "step": 1509 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2645.0, + "completions/mean_length": 953.68310546875, + "completions/mean_terminated_length": 632.8806762695312, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.11195335276968, + "grad_norm": 0.13045363128185272, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 884356664.0, + "reward": 0.7600446939468384, + "reward_std": 0.12692874670028687, + "rewards/verify_math_reward/mean": 0.7600446343421936, + "rewards/verify_math_reward/std": 0.42729446291923523, + "step": 1510 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1607142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3490.0, + "completions/mean_length": 1236.2410888671875, + "completions/mean_terminated_length": 688.6276245117188, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 14.12128279883382, + "grad_norm": 0.13698236644268036, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 884958064.0, + "reward": 0.6729910969734192, + "reward_std": 0.11568940430879593, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1511 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3509.0, + "completions/mean_length": 842.9788208007812, + "completions/mean_terminated_length": 596.9519653320312, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 14.130612244897959, + "grad_norm": 0.15260770916938782, + "learning_rate": 1e-06, + "loss": -0.0272, + "num_tokens": 885535645.0, + "reward": 0.7611607313156128, + "reward_std": 0.13162529468536377, + "rewards/verify_math_reward/mean": 0.7611607313156128, + "rewards/verify_math_reward/std": 0.4266124963760376, + "step": 1512 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3809.0, + "completions/mean_length": 996.4989013671875, + "completions/mean_terminated_length": 667.4160766601562, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 14.139941690962099, + "grad_norm": 0.15160422027111053, + "learning_rate": 1e-06, + "loss": -0.0562, + "num_tokens": 886170100.0, + "reward": 0.676339328289032, + "reward_std": 0.148830384016037, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1513 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 1182.021240234375, + "completions/mean_terminated_length": 700.7789306640625, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.14927113702624, + "grad_norm": 0.17777971923351288, + "learning_rate": 1e-06, + "loss": -0.0726, + "num_tokens": 886803167.0, + "reward": 0.6629464626312256, + "reward_std": 0.15315786004066467, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1514 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3778.0, + "completions/mean_length": 1243.0614013671875, + "completions/mean_terminated_length": 732.5355834960938, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 14.15860058309038, + "grad_norm": 0.1419702023267746, + "learning_rate": 1e-06, + "loss": -0.0693, + "num_tokens": 887450334.0, + "reward": 0.6696428656578064, + "reward_std": 0.12464035302400589, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1515 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2721.0, + "completions/mean_length": 1073.8270263671875, + "completions/mean_terminated_length": 615.451171875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 14.167930029154519, + "grad_norm": 0.1759144365787506, + "learning_rate": 1e-06, + "loss": -0.0858, + "num_tokens": 888018419.0, + "reward": 0.731026828289032, + "reward_std": 0.12035568058490753, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1516 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 989.396240234375, + "completions/mean_terminated_length": 638.2149047851562, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 14.177259475218658, + "grad_norm": 0.16178756952285767, + "learning_rate": 1e-06, + "loss": -0.0988, + "num_tokens": 888620374.0, + "reward": 0.7299107313156128, + "reward_std": 0.15398018062114716, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1517 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4038.0, + "completions/mean_length": 1195.029052734375, + "completions/mean_terminated_length": 675.9078979492188, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 14.186588921282798, + "grad_norm": 0.14347553253173828, + "learning_rate": 1e-06, + "loss": -0.0778, + "num_tokens": 889219504.0, + "reward": 0.6629464626312256, + "reward_std": 0.12535391747951508, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4055.0, + "completions/mean_length": 1126.6507568359375, + "completions/mean_terminated_length": 627.2424926757812, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 14.19591836734694, + "grad_norm": 0.1469012051820755, + "learning_rate": 1e-06, + "loss": -0.0764, + "num_tokens": 889793591.0, + "reward": 0.7020089626312256, + "reward_std": 0.1195308193564415, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1519 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3181.0, + "completions/mean_length": 923.2589721679688, + "completions/mean_terminated_length": 620.7237548828125, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 14.205247813411079, + "grad_norm": 0.16731294989585876, + "learning_rate": 1e-06, + "loss": -0.046, + "num_tokens": 890390655.0, + "reward": 0.7399553656578064, + "reward_std": 0.1554141491651535, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890368938446045, + "step": 1520 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3113.0, + "completions/mean_length": 1113.1217041015625, + "completions/mean_terminated_length": 625.0142822265625, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 14.214577259475218, + "grad_norm": 0.15868867933750153, + "learning_rate": 1e-06, + "loss": -0.0827, + "num_tokens": 890957252.0, + "reward": 0.7031250596046448, + "reward_std": 0.14395050704479218, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1521 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3600.0, + "completions/mean_length": 1179.649658203125, + "completions/mean_terminated_length": 711.22021484375, + "completions/min_length": 198.0, + "completions/min_terminated_length": 198.0, + "epoch": 14.223906705539358, + "grad_norm": 0.1729939728975296, + "learning_rate": 1e-06, + "loss": -0.054, + "num_tokens": 891587850.0, + "reward": 0.684151828289032, + "reward_std": 0.1776525229215622, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1522 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3311.0, + "completions/mean_length": 1153.7623291015625, + "completions/mean_terminated_length": 724.8427124023438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 14.2332361516035, + "grad_norm": 0.15857946872711182, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 892242973.0, + "reward": 0.6383928656578064, + "reward_std": 0.14527249336242676, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 1523 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3138.0, + "completions/mean_length": 1093.036865234375, + "completions/mean_terminated_length": 668.4140014648438, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.242565597667639, + "grad_norm": 0.14814192056655884, + "learning_rate": 1e-06, + "loss": -0.0677, + "num_tokens": 892856134.0, + "reward": 0.691964328289032, + "reward_std": 0.1378631889820099, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1524 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3368.0, + "completions/mean_length": 1127.97216796875, + "completions/mean_terminated_length": 655.6986083984375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 14.251895043731778, + "grad_norm": 0.16072885692119598, + "learning_rate": 1e-06, + "loss": -0.0832, + "num_tokens": 893463869.0, + "reward": 0.6953125596046448, + "reward_std": 0.1269722282886505, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1525 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3451.0, + "completions/mean_length": 1105.708740234375, + "completions/mean_terminated_length": 656.5892333984375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 14.261224489795918, + "grad_norm": 0.17252424359321594, + "learning_rate": 1e-06, + "loss": -0.0955, + "num_tokens": 894065592.0, + "reward": 0.6830357313156128, + "reward_std": 0.16604548692703247, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1526 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3182.0, + "completions/mean_length": 1119.80810546875, + "completions/mean_terminated_length": 681.5723266601562, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 14.270553935860057, + "grad_norm": 0.18007849156856537, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 894695716.0, + "reward": 0.6852678656578064, + "reward_std": 0.18216247856616974, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1527 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1964285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3884.0, + "completions/mean_length": 1374.8717041015625, + "completions/mean_terminated_length": 709.7069702148438, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 14.279883381924199, + "grad_norm": 0.17875495553016663, + "learning_rate": 1e-06, + "loss": -0.0676, + "num_tokens": 895299417.0, + "reward": 0.5803571939468384, + "reward_std": 0.14327509701251984, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 1528 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3858.0, + "completions/mean_length": 1173.2879638671875, + "completions/mean_terminated_length": 712.6021118164062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 14.289212827988338, + "grad_norm": 0.16526354849338531, + "learning_rate": 1e-06, + "loss": -0.0797, + "num_tokens": 895931691.0, + "reward": 0.660714328289032, + "reward_std": 0.15210406482219696, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1529 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3692.0, + "completions/mean_length": 903.489990234375, + "completions/mean_terminated_length": 641.3031616210938, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 14.298542274052478, + "grad_norm": 0.13230258226394653, + "learning_rate": 1e-06, + "loss": -0.0823, + "num_tokens": 896535258.0, + "reward": 0.7912946939468384, + "reward_std": 0.11968054622411728, + "rewards/verify_math_reward/mean": 0.7912946343421936, + "rewards/verify_math_reward/std": 0.4066103398799896, + "step": 1530 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4001.0, + "completions/mean_length": 1059.4888916015625, + "completions/mean_terminated_length": 652.0582275390625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.307871720116617, + "grad_norm": 0.1585511863231659, + "learning_rate": 1e-06, + "loss": -0.0617, + "num_tokens": 897137888.0, + "reward": 0.6930803656578064, + "reward_std": 0.12990324199199677, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 1531 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3467.0, + "completions/mean_length": 1089.6082763671875, + "completions/mean_terminated_length": 681.8973388671875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 14.317201166180759, + "grad_norm": 0.16274017095565796, + "learning_rate": 1e-06, + "loss": -0.0227, + "num_tokens": 897769249.0, + "reward": 0.668526828289032, + "reward_std": 0.14387424290180206, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1532 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3812.0, + "completions/mean_length": 1075.634033203125, + "completions/mean_terminated_length": 635.3248291015625, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 14.326530612244898, + "grad_norm": 0.1630510687828064, + "learning_rate": 1e-06, + "loss": -0.0736, + "num_tokens": 898346137.0, + "reward": 0.6785714626312256, + "reward_std": 0.14353612065315247, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3887.0, + "completions/mean_length": 1200.6082763671875, + "completions/mean_terminated_length": 722.4356079101562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 14.335860058309038, + "grad_norm": 0.1703139692544937, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 898990034.0, + "reward": 0.645089328289032, + "reward_std": 0.16251856088638306, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2864.0, + "completions/mean_length": 1122.22998046875, + "completions/mean_terminated_length": 684.350830078125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 14.345189504373177, + "grad_norm": 0.13809487223625183, + "learning_rate": 1e-06, + "loss": -0.0574, + "num_tokens": 899615632.0, + "reward": 0.684151828289032, + "reward_std": 0.1373000591993332, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1535 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3517.0, + "completions/mean_length": 1062.8538818359375, + "completions/mean_terminated_length": 625.1200561523438, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 14.354518950437317, + "grad_norm": 0.1659601330757141, + "learning_rate": 1e-06, + "loss": -0.1104, + "num_tokens": 900188037.0, + "reward": 0.7176339626312256, + "reward_std": 0.1627788543701172, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1536 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3795.0, + "completions/mean_length": 1118.3482666015625, + "completions/mean_terminated_length": 748.476806640625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 14.363848396501458, + "grad_norm": 0.13885265588760376, + "learning_rate": 1e-06, + "loss": -0.0399, + "num_tokens": 900883805.0, + "reward": 0.629464328289032, + "reward_std": 0.14304757118225098, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 1537 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4037.0, + "completions/mean_length": 1039.6351318359375, + "completions/mean_terminated_length": 677.144775390625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 14.373177842565598, + "grad_norm": 0.17506875097751617, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 901506646.0, + "reward": 0.6852678656578064, + "reward_std": 0.17821483314037323, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1538 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2675.0, + "completions/mean_length": 1021.0592041015625, + "completions/mean_terminated_length": 656.3657836914062, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 14.382507288629737, + "grad_norm": 0.16278286278247833, + "learning_rate": 1e-06, + "loss": -0.0558, + "num_tokens": 902116971.0, + "reward": 0.6852678656578064, + "reward_std": 0.16228215396404266, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3287.0, + "completions/mean_length": 986.5826416015625, + "completions/mean_terminated_length": 582.7112426757812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 14.391836734693877, + "grad_norm": 0.16569572687149048, + "learning_rate": 1e-06, + "loss": -0.0727, + "num_tokens": 902661661.0, + "reward": 0.7220982313156128, + "reward_std": 0.13737532496452332, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 1540 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2450.0, + "completions/mean_length": 1127.771240234375, + "completions/mean_terminated_length": 725.2357177734375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.401166180758018, + "grad_norm": 0.15166299045085907, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 903328824.0, + "reward": 0.652901828289032, + "reward_std": 0.1588732749223709, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 1541 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3465.0, + "completions/mean_length": 988.3069458007812, + "completions/mean_terminated_length": 671.0393676757812, + "completions/min_length": 203.0, + "completions/min_terminated_length": 203.0, + "epoch": 14.410495626822158, + "grad_norm": 0.16228637099266052, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 903955059.0, + "reward": 0.7209821939468384, + "reward_std": 0.1727273166179657, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 1542 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3215.0, + "completions/mean_length": 1054.4732666015625, + "completions/mean_terminated_length": 668.0653686523438, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 14.419825072886297, + "grad_norm": 0.14531028270721436, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 904567835.0, + "reward": 0.7321428656578064, + "reward_std": 0.12779638171195984, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 1543 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1640625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3032.0, + "completions/mean_length": 1205.90185546875, + "completions/mean_terminated_length": 638.686279296875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 14.429154518950437, + "grad_norm": 0.1915040910243988, + "learning_rate": 1e-06, + "loss": -0.0711, + "num_tokens": 905148339.0, + "reward": 0.6808035969734192, + "reward_std": 0.14297199249267578, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1544 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3603.0, + "completions/mean_length": 1096.154052734375, + "completions/mean_terminated_length": 645.5994873046875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 14.438483965014576, + "grad_norm": 0.15866118669509888, + "learning_rate": 1e-06, + "loss": -0.0619, + "num_tokens": 905740325.0, + "reward": 0.6808035969734192, + "reward_std": 0.13865482807159424, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 1545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3115.0, + "completions/mean_length": 1175.4810791015625, + "completions/mean_terminated_length": 675.3660278320312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 14.447813411078718, + "grad_norm": 0.17744548618793488, + "learning_rate": 1e-06, + "loss": -0.0906, + "num_tokens": 906339860.0, + "reward": 0.6517857313156128, + "reward_std": 0.17513547837734222, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1546 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3579.0, + "completions/mean_length": 1018.7835083007812, + "completions/mean_terminated_length": 636.5445556640625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 14.457142857142857, + "grad_norm": 0.1561674028635025, + "learning_rate": 1e-06, + "loss": -0.049, + "num_tokens": 906941258.0, + "reward": 0.6439732313156128, + "reward_std": 0.1356835663318634, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3219.0, + "completions/mean_length": 1135.1998291015625, + "completions/mean_terminated_length": 703.5741577148438, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 14.466472303206997, + "grad_norm": 0.16115762293338776, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 907593909.0, + "reward": 0.6830357313156128, + "reward_std": 0.12843577563762665, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1548 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2937.0, + "completions/mean_length": 1016.700927734375, + "completions/mean_terminated_length": 625.4943237304688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 14.475801749271136, + "grad_norm": 0.1825391948223114, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 908172457.0, + "reward": 0.6886160969734192, + "reward_std": 0.15710480511188507, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1549 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3544.0, + "completions/mean_length": 941.62841796875, + "completions/mean_terminated_length": 640.8447875976562, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 14.485131195335278, + "grad_norm": 0.15126000344753265, + "learning_rate": 1e-06, + "loss": -0.0232, + "num_tokens": 908795748.0, + "reward": 0.6875000596046448, + "reward_std": 0.11952900886535645, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1550 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3508.0, + "completions/mean_length": 988.8750610351562, + "completions/mean_terminated_length": 663.220703125, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 14.494460641399417, + "grad_norm": 0.13762199878692627, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 909418764.0, + "reward": 0.7031250596046448, + "reward_std": 0.12403164058923721, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1551 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3610.0, + "completions/mean_length": 1139.841552734375, + "completions/mean_terminated_length": 647.1484375, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.503790087463557, + "grad_norm": 0.13642458617687225, + "learning_rate": 1e-06, + "loss": -0.0344, + "num_tokens": 910009790.0, + "reward": 0.6629464626312256, + "reward_std": 0.09299751371145248, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1552 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3898.0, + "completions/mean_length": 1104.102783203125, + "completions/mean_terminated_length": 659.1538696289062, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 14.513119533527696, + "grad_norm": 0.16526970267295837, + "learning_rate": 1e-06, + "loss": -0.0319, + "num_tokens": 910612594.0, + "reward": 0.7087053656578064, + "reward_std": 0.14060692489147186, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1553 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2662.0, + "completions/mean_length": 1108.985595703125, + "completions/mean_terminated_length": 638.1641235351562, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.522448979591836, + "grad_norm": 0.1656293272972107, + "learning_rate": 1e-06, + "loss": -0.1282, + "num_tokens": 911196997.0, + "reward": 0.6651785969734192, + "reward_std": 0.1597309112548828, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4039.0, + "completions/mean_length": 1031.0123291015625, + "completions/mean_terminated_length": 688.7680053710938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 14.531778425655977, + "grad_norm": 0.1355527937412262, + "learning_rate": 1e-06, + "loss": -0.0304, + "num_tokens": 911850672.0, + "reward": 0.6975446939468384, + "reward_std": 0.09915942698717117, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1555 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3543.0, + "completions/mean_length": 1136.9420166015625, + "completions/mean_terminated_length": 692.5134887695312, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 14.541107871720117, + "grad_norm": 0.15699976682662964, + "learning_rate": 1e-06, + "loss": -0.0319, + "num_tokens": 912485484.0, + "reward": 0.6540178656578064, + "reward_std": 0.13098448514938354, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1556 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3762.0, + "completions/mean_length": 922.0748291015625, + "completions/mean_terminated_length": 627.9061279296875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 14.550437317784256, + "grad_norm": 0.18139240145683289, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 913086815.0, + "reward": 0.7366071939468384, + "reward_std": 0.15416815876960754, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1557 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3797.0, + "completions/mean_length": 838.9319458007812, + "completions/mean_terminated_length": 584.1672973632812, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 14.559766763848396, + "grad_norm": 0.1686929315328598, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 913651666.0, + "reward": 0.7589285969734192, + "reward_std": 0.1356828659772873, + "rewards/verify_math_reward/mean": 0.7589285969734192, + "rewards/verify_math_reward/std": 0.4279724657535553, + "step": 1558 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2949.0, + "completions/mean_length": 924.357177734375, + "completions/mean_terminated_length": 600.5608520507812, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 14.569096209912537, + "grad_norm": 0.1644112467765808, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 914230282.0, + "reward": 0.707589328289032, + "reward_std": 0.1318553388118744, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1559 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2802.0, + "completions/mean_length": 1193.614990234375, + "completions/mean_terminated_length": 709.8841552734375, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 14.578425655976677, + "grad_norm": 0.14718376100063324, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 914863641.0, + "reward": 0.6361607313156128, + "reward_std": 0.12399844080209732, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1560 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2966.0, + "completions/mean_length": 920.216552734375, + "completions/mean_terminated_length": 583.0346069335938, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 14.587755102040816, + "grad_norm": 0.1438400149345398, + "learning_rate": 1e-06, + "loss": -0.0414, + "num_tokens": 915423219.0, + "reward": 0.7410714626312256, + "reward_std": 0.10675237327814102, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1561 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3831.0, + "completions/mean_length": 959.8906860351562, + "completions/mean_terminated_length": 648.2036743164062, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 14.597084548104956, + "grad_norm": 0.1563788652420044, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 916045481.0, + "reward": 0.6986607313156128, + "reward_std": 0.133991077542305, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3585.0, + "completions/mean_length": 930.41748046875, + "completions/mean_terminated_length": 615.8012084960938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 14.606413994169095, + "grad_norm": 0.13547132909297943, + "learning_rate": 1e-06, + "loss": -0.0519, + "num_tokens": 916632679.0, + "reward": 0.7332589626312256, + "reward_std": 0.11783905327320099, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425029158592224, + "step": 1563 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3101.0, + "completions/mean_length": 1052.4554443359375, + "completions/mean_terminated_length": 695.730712890625, + "completions/min_length": 202.0, + "completions/min_terminated_length": 202.0, + "epoch": 14.615743440233237, + "grad_norm": 0.14040714502334595, + "learning_rate": 1e-06, + "loss": -0.0616, + "num_tokens": 917280111.0, + "reward": 0.676339328289032, + "reward_std": 0.13177725672721863, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1564 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3343.0, + "completions/mean_length": 962.3460083007812, + "completions/mean_terminated_length": 629.6370849609375, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 14.625072886297376, + "grad_norm": 0.1533210426568985, + "learning_rate": 1e-06, + "loss": -0.0604, + "num_tokens": 917878677.0, + "reward": 0.7276785969734192, + "reward_std": 0.13764451444149017, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1565 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3241.0, + "completions/mean_length": 949.6094360351562, + "completions/mean_terminated_length": 619.8396606445312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.634402332361516, + "grad_norm": 0.1622307449579239, + "learning_rate": 1e-06, + "loss": -0.0378, + "num_tokens": 918466215.0, + "reward": 0.7120535969734192, + "reward_std": 0.12463782727718353, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530588984489441, + "step": 1566 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2442.0, + "completions/mean_length": 819.0960083007812, + "completions/mean_terminated_length": 583.9114379882812, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 14.643731778425655, + "grad_norm": 0.16707095503807068, + "learning_rate": 1e-06, + "loss": -0.027, + "num_tokens": 919050437.0, + "reward": 0.7433035969734192, + "reward_std": 0.1252797544002533, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1567 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3467.0, + "completions/mean_length": 974.5535888671875, + "completions/mean_terminated_length": 681.0842895507812, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 14.653061224489797, + "grad_norm": 0.1345309019088745, + "learning_rate": 1e-06, + "loss": -0.042, + "num_tokens": 919690853.0, + "reward": 0.7165178656578064, + "reward_std": 0.13357669115066528, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1568 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3878.0, + "completions/mean_length": 1104.671875, + "completions/mean_terminated_length": 703.3037719726562, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 14.662390670553936, + "grad_norm": 0.16132068634033203, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 920347439.0, + "reward": 0.6651785969734192, + "reward_std": 0.12508542835712433, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1569 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3737.0, + "completions/mean_length": 1070.46435546875, + "completions/mean_terminated_length": 647.0432739257812, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 14.671720116618076, + "grad_norm": 0.154660165309906, + "learning_rate": 1e-06, + "loss": -0.0342, + "num_tokens": 920943095.0, + "reward": 0.6473214626312256, + "reward_std": 0.11558134853839874, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 1570 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3164.0, + "completions/mean_length": 982.372802734375, + "completions/mean_terminated_length": 647.5327758789062, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 14.681049562682215, + "grad_norm": 0.16680721938610077, + "learning_rate": 1e-06, + "loss": -0.0504, + "num_tokens": 921559237.0, + "reward": 0.7042410969734192, + "reward_std": 0.1314416527748108, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1571 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3876.0, + "completions/mean_length": 999.083740234375, + "completions/mean_terminated_length": 605.6389770507812, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 14.690379008746355, + "grad_norm": 0.1560639590024948, + "learning_rate": 1e-06, + "loss": -0.0547, + "num_tokens": 922138216.0, + "reward": 0.7087053656578064, + "reward_std": 0.10878115892410278, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1572 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3804.0, + "completions/mean_length": 1218.380615234375, + "completions/mean_terminated_length": 730.0117797851562, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 14.699708454810496, + "grad_norm": 0.17366640269756317, + "learning_rate": 1e-06, + "loss": -0.0811, + "num_tokens": 922788645.0, + "reward": 0.578125, + "reward_std": 0.1615314483642578, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 1573 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3735.0, + "completions/mean_length": 954.8125610351562, + "completions/mean_terminated_length": 663.6780395507812, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 14.709037900874636, + "grad_norm": 0.17071594297885895, + "learning_rate": 1e-06, + "loss": -0.0512, + "num_tokens": 923420597.0, + "reward": 0.7421875596046448, + "reward_std": 0.1054396778345108, + "rewards/verify_math_reward/mean": 0.7421875, + "rewards/verify_math_reward/std": 0.43767455220222473, + "step": 1574 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3705.0, + "completions/mean_length": 948.7109985351562, + "completions/mean_terminated_length": 601.6121826171875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 14.718367346938775, + "grad_norm": 0.168731227517128, + "learning_rate": 1e-06, + "loss": -0.0193, + "num_tokens": 923987674.0, + "reward": 0.7332589626312256, + "reward_std": 0.13177795708179474, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425029158592224, + "step": 1575 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4053.0, + "completions/mean_length": 1063.0513916015625, + "completions/mean_terminated_length": 673.4281616210938, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 14.727696793002915, + "grad_norm": 0.16611315310001373, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 924629064.0, + "reward": 0.6886160969734192, + "reward_std": 0.1350441575050354, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1576 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3143.0, + "completions/mean_length": 951.5301513671875, + "completions/mean_terminated_length": 630.5079956054688, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 14.737026239067056, + "grad_norm": 0.13521364331245422, + "learning_rate": 1e-06, + "loss": -0.0699, + "num_tokens": 925240995.0, + "reward": 0.7198660969734192, + "reward_std": 0.1194516196846962, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931530952453613, + "step": 1577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3079.0, + "completions/mean_length": 771.6194458007812, + "completions/mean_terminated_length": 628.42724609375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 14.746355685131196, + "grad_norm": 0.15980949997901917, + "learning_rate": 1e-06, + "loss": -0.023, + "num_tokens": 925861366.0, + "reward": 0.7745535969734192, + "reward_std": 0.16322465240955353, + "rewards/verify_math_reward/mean": 0.7745535969734192, + "rewards/verify_math_reward/std": 0.41810935735702515, + "step": 1578 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2827.0, + "completions/mean_length": 876.5402221679688, + "completions/mean_terminated_length": 628.8894653320312, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 14.755685131195335, + "grad_norm": 0.1481754332780838, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 926479690.0, + "reward": 0.7678571939468384, + "reward_std": 0.13087712228298187, + "rewards/verify_math_reward/mean": 0.7678571343421936, + "rewards/verify_math_reward/std": 0.422435462474823, + "step": 1579 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2881.0, + "completions/mean_length": 963.833740234375, + "completions/mean_terminated_length": 635.5548706054688, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 14.765014577259475, + "grad_norm": 0.16474878787994385, + "learning_rate": 1e-06, + "loss": -0.0523, + "num_tokens": 927085917.0, + "reward": 0.7354910969734192, + "reward_std": 0.14373226463794708, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131770730018616, + "step": 1580 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4035.0, + "completions/mean_length": 1159.3326416015625, + "completions/mean_terminated_length": 722.5974731445312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.774344023323614, + "grad_norm": 0.15768881142139435, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 927751079.0, + "reward": 0.6160714626312256, + "reward_std": 0.1441766321659088, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 1581 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3881.0, + "completions/mean_length": 1042.9029541015625, + "completions/mean_terminated_length": 667.9611206054688, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 14.783673469387756, + "grad_norm": 0.15043064951896667, + "learning_rate": 1e-06, + "loss": -0.0093, + "num_tokens": 928376480.0, + "reward": 0.660714328289032, + "reward_std": 0.1129891499876976, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1582 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3446.0, + "completions/mean_length": 995.8225708007812, + "completions/mean_terminated_length": 610.7315063476562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 14.793002915451895, + "grad_norm": 0.1480284184217453, + "learning_rate": 1e-06, + "loss": -0.0543, + "num_tokens": 928948073.0, + "reward": 0.7433035969734192, + "reward_std": 0.1117856353521347, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705445528030396, + "step": 1583 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2481.0, + "completions/mean_length": 1005.3292846679688, + "completions/mean_terminated_length": 647.3810424804688, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 14.802332361516035, + "grad_norm": 0.15579748153686523, + "learning_rate": 1e-06, + "loss": -0.0339, + "num_tokens": 929553904.0, + "reward": 0.6718750596046448, + "reward_std": 0.14489158987998962, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1584 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3216.0, + "completions/mean_length": 1137.7991943359375, + "completions/mean_terminated_length": 710.8812255859375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 14.811661807580174, + "grad_norm": 0.15973977744579315, + "learning_rate": 1e-06, + "loss": -0.047, + "num_tokens": 930206516.0, + "reward": 0.6462053656578064, + "reward_std": 0.15939390659332275, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1585 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3993.0, + "completions/mean_length": 1041.3717041015625, + "completions/mean_terminated_length": 670.534423828125, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 14.820991253644316, + "grad_norm": 0.14964067935943604, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 930834929.0, + "reward": 0.7008928656578064, + "reward_std": 0.14079168438911438, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 1586 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3547.0, + "completions/mean_length": 1185.08935546875, + "completions/mean_terminated_length": 717.5336303710938, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 14.830320699708455, + "grad_norm": 0.17296187579631805, + "learning_rate": 1e-06, + "loss": -0.0627, + "num_tokens": 931474041.0, + "reward": 0.6439732313156128, + "reward_std": 0.14951257407665253, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1587 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3146.0, + "completions/mean_length": 911.0558471679688, + "completions/mean_terminated_length": 607.3569946289062, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 14.839650145772595, + "grad_norm": 0.17618399858474731, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 932063403.0, + "reward": 0.7254464626312256, + "reward_std": 0.14778690040111542, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1588 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4006.0, + "completions/mean_length": 1179.4632568359375, + "completions/mean_terminated_length": 737.1092529296875, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 14.848979591836734, + "grad_norm": 0.1581123024225235, + "learning_rate": 1e-06, + "loss": -0.0961, + "num_tokens": 932725682.0, + "reward": 0.65625, + "reward_std": 0.1491014063358307, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1589 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3432.0, + "completions/mean_length": 1119.5592041015625, + "completions/mean_terminated_length": 703.0089111328125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 14.858309037900874, + "grad_norm": 0.14995414018630981, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 933371031.0, + "reward": 0.640625, + "reward_std": 0.1327543705701828, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 1590 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3279.0, + "completions/mean_length": 1029.571533203125, + "completions/mean_terminated_length": 661.5999755859375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 14.867638483965015, + "grad_norm": 0.13666093349456787, + "learning_rate": 1e-06, + "loss": -0.0598, + "num_tokens": 933988527.0, + "reward": 0.7377232313156128, + "reward_std": 0.126776784658432, + "rewards/verify_math_reward/mean": 0.7377232313156128, + "rewards/verify_math_reward/std": 0.4401180148124695, + "step": 1591 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3389.0, + "completions/mean_length": 893.700927734375, + "completions/mean_terminated_length": 588.3472290039062, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 14.876967930029155, + "grad_norm": 0.17856274545192719, + "learning_rate": 1e-06, + "loss": -0.0443, + "num_tokens": 934557627.0, + "reward": 0.7488839626312256, + "reward_std": 0.12192869931459427, + "rewards/verify_math_reward/mean": 0.7488839030265808, + "rewards/verify_math_reward/std": 0.43389734625816345, + "step": 1592 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3577.0, + "completions/mean_length": 1117.649658203125, + "completions/mean_terminated_length": 696.5070190429688, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 14.886297376093294, + "grad_norm": 0.17614515125751495, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 935200617.0, + "reward": 0.6584821939468384, + "reward_std": 0.1531924605369568, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1593 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4033.0, + "completions/mean_length": 1109.4442138671875, + "completions/mean_terminated_length": 620.7350463867188, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 14.895626822157434, + "grad_norm": 0.12018298357725143, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 935770863.0, + "reward": 0.6517857313156128, + "reward_std": 0.09777005016803741, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 1594 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3482.0, + "completions/mean_length": 1059.3616943359375, + "completions/mean_terminated_length": 669.2644653320312, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.904956268221575, + "grad_norm": 0.13828334212303162, + "learning_rate": 1e-06, + "loss": -0.0612, + "num_tokens": 936392891.0, + "reward": 0.6741071939468384, + "reward_std": 0.1239563599228859, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 1595 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3029.0, + "completions/mean_length": 999.1529541015625, + "completions/mean_terminated_length": 712.1280517578125, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 14.914285714285715, + "grad_norm": 0.1584785431623459, + "learning_rate": 1e-06, + "loss": -0.038, + "num_tokens": 937062532.0, + "reward": 0.6863839626312256, + "reward_std": 0.16597090661525726, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1596 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2936.0, + "completions/mean_length": 1004.482177734375, + "completions/mean_terminated_length": 629.1664428710938, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 14.923615160349854, + "grad_norm": 0.19859588146209717, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 937646916.0, + "reward": 0.7287946939468384, + "reward_std": 0.1315924972295761, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1597 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 1086.69873046875, + "completions/mean_terminated_length": 682.9190063476562, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 14.932944606413994, + "grad_norm": 0.13819001615047455, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 938270406.0, + "reward": 0.7064732313156128, + "reward_std": 0.12241724878549576, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1598 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3063.0, + "completions/mean_length": 1041.5692138671875, + "completions/mean_terminated_length": 631.7341918945312, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 14.942274052478133, + "grad_norm": 0.15569481253623962, + "learning_rate": 1e-06, + "loss": -0.0342, + "num_tokens": 938861948.0, + "reward": 0.6941964626312256, + "reward_std": 0.11073214560747147, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.46100425720214844, + "step": 1599 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1741071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3242.0, + "completions/mean_length": 1301.2779541015625, + "completions/mean_terminated_length": 712.1203002929688, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 14.951603498542275, + "grad_norm": 0.14823003113269806, + "learning_rate": 1e-06, + "loss": -0.0556, + "num_tokens": 939489317.0, + "reward": 0.59375, + "reward_std": 0.12166837602853775, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 1600 + }, + { + "epoch": 14.951603498542275, + "step": 1600, + "total_flos": 0.0, + "train_loss": -0.039273733161653585, + "train_runtime": 125060.3323, + "train_samples_per_second": 11.463, + "train_steps_per_second": 0.013 + } + ], + "logging_steps": 1, + "max_steps": 1600, + "num_input_tokens_seen": 939489317, + "num_train_epochs": 15, + "save_steps": 160, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +}