Lansechen's picture
Model save
9876eaa verified
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.9936628643852978,
"eval_steps": 100,
"global_step": 196,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"clip_ratio": 0.0,
"completion_length": 544.7745819091797,
"epoch": 0.010139416983523447,
"grad_norm": 0.2070285975933075,
"learning_rate": 5e-08,
"loss": -0.1577,
"num_tokens": 636118.0,
"reward": -0.35065557062625885,
"reward_std": 0.656378798186779,
"rewards/cosine_scaled_reward": -0.35400376841425896,
"rewards/format_reward": 0.003348214435391128,
"step": 1
},
{
"clip_ratio": 0.0,
"completion_length": 549.2812843322754,
"epoch": 0.020278833967046894,
"grad_norm": 0.2276979684829712,
"learning_rate": 1e-07,
"loss": -0.1565,
"num_tokens": 1263330.0,
"reward": -0.3044062051922083,
"reward_std": 0.6904742792248726,
"rewards/cosine_scaled_reward": -0.31221868470311165,
"rewards/format_reward": 0.007812500349245965,
"step": 2
},
{
"clip_ratio": 0.0,
"completion_length": 588.2891006469727,
"epoch": 0.030418250950570342,
"grad_norm": 0.19815148413181305,
"learning_rate": 1.5e-07,
"loss": -0.1123,
"num_tokens": 1930557.0,
"reward": -0.2540861386805773,
"reward_std": 0.6807773411273956,
"rewards/cosine_scaled_reward": -0.2618986228480935,
"rewards/format_reward": 0.007812500232830644,
"step": 3
},
{
"clip_ratio": 0.0,
"completion_length": 552.0669898986816,
"epoch": 0.04055766793409379,
"grad_norm": 0.20235052704811096,
"learning_rate": 2e-07,
"loss": -0.1652,
"num_tokens": 2562161.0,
"reward": -0.3022355027496815,
"reward_std": 0.6755559742450714,
"rewards/cosine_scaled_reward": -0.3122801296412945,
"rewards/format_reward": 0.010044643189758062,
"step": 4
},
{
"clip_ratio": 0.0,
"completion_length": 571.6172142028809,
"epoch": 0.050697084917617236,
"grad_norm": 0.19000709056854248,
"learning_rate": 2.5e-07,
"loss": -0.1267,
"num_tokens": 3201114.0,
"reward": -0.2917697671800852,
"reward_std": 0.6964326798915863,
"rewards/cosine_scaled_reward": -0.2984661813825369,
"rewards/format_reward": 0.006696428870782256,
"step": 5
},
{
"clip_ratio": 0.0,
"completion_length": 561.6384201049805,
"epoch": 0.060836501901140684,
"grad_norm": 0.21064706146717072,
"learning_rate": 3e-07,
"loss": -0.1771,
"num_tokens": 3833534.0,
"reward": -0.28072800021618605,
"reward_std": 0.6883805021643639,
"rewards/cosine_scaled_reward": -0.2840762068517506,
"rewards/format_reward": 0.003348214435391128,
"step": 6
},
{
"clip_ratio": 0.0,
"completion_length": 542.1149749755859,
"epoch": 0.07097591888466413,
"grad_norm": 0.22860734164714813,
"learning_rate": 3.5e-07,
"loss": -0.131,
"num_tokens": 4443693.0,
"reward": -0.23171719023957849,
"reward_std": 0.7431022375822067,
"rewards/cosine_scaled_reward": -0.24287789640948176,
"rewards/format_reward": 0.011160714668221772,
"step": 7
},
{
"clip_ratio": 0.0,
"completion_length": 563.7890892028809,
"epoch": 0.08111533586818757,
"grad_norm": 0.19847826659679413,
"learning_rate": 4e-07,
"loss": -0.142,
"num_tokens": 5076400.0,
"reward": -0.3053848221898079,
"reward_std": 0.6786770969629288,
"rewards/cosine_scaled_reward": -0.3109651692211628,
"rewards/format_reward": 0.005580357392318547,
"step": 8
},
{
"clip_ratio": 0.0,
"completion_length": 585.7589569091797,
"epoch": 0.09125475285171103,
"grad_norm": 0.1838308870792389,
"learning_rate": 4.5e-07,
"loss": -0.1486,
"num_tokens": 5724568.0,
"reward": -0.26803811825811863,
"reward_std": 0.6804082244634628,
"rewards/cosine_scaled_reward": -0.26915417425334454,
"rewards/format_reward": 0.0011160714784637094,
"step": 9
},
{
"clip_ratio": 0.0,
"completion_length": 516.0636291503906,
"epoch": 0.10139416983523447,
"grad_norm": 0.6098268032073975,
"learning_rate": 5e-07,
"loss": -0.1271,
"num_tokens": 6313689.0,
"reward": -0.2813316825777292,
"reward_std": 0.7330739945173264,
"rewards/cosine_scaled_reward": -0.2880280986428261,
"rewards/format_reward": 0.006696428870782256,
"step": 10
},
{
"clip_ratio": 0.0,
"completion_length": 568.8984718322754,
"epoch": 0.11153358681875793,
"grad_norm": 0.5140683054924011,
"learning_rate": 5.5e-07,
"loss": -0.1441,
"num_tokens": 6960958.0,
"reward": -0.2723624687641859,
"reward_std": 0.6729818060994148,
"rewards/cosine_scaled_reward": -0.275710666552186,
"rewards/format_reward": 0.003348214435391128,
"step": 11
},
{
"clip_ratio": 0.0,
"completion_length": 584.2366409301758,
"epoch": 0.12167300380228137,
"grad_norm": 0.20556923747062683,
"learning_rate": 6e-07,
"loss": -0.1173,
"num_tokens": 7615746.0,
"reward": -0.21572664007544518,
"reward_std": 0.7041841298341751,
"rewards/cosine_scaled_reward": -0.2201909152790904,
"rewards/format_reward": 0.004464285913854837,
"step": 12
},
{
"clip_ratio": 0.0,
"completion_length": 598.1797065734863,
"epoch": 0.13181242078580482,
"grad_norm": 0.19161486625671387,
"learning_rate": 6.5e-07,
"loss": -0.1216,
"num_tokens": 8279899.0,
"reward": -0.23263627663254738,
"reward_std": 0.7194378077983856,
"rewards/cosine_scaled_reward": -0.23933268897235394,
"rewards/format_reward": 0.006696428870782256,
"step": 13
},
{
"clip_ratio": 0.0,
"completion_length": 539.4096183776855,
"epoch": 0.14195183776932827,
"grad_norm": 0.20079070329666138,
"learning_rate": 7e-07,
"loss": -0.1091,
"num_tokens": 8895386.0,
"reward": -0.21676091896370053,
"reward_std": 0.7302467301487923,
"rewards/cosine_scaled_reward": -0.22209045942872763,
"rewards/format_reward": 0.0011160714784637094,
"step": 14
},
{
"clip_ratio": 0.0,
"completion_length": 581.7645378112793,
"epoch": 0.1520912547528517,
"grad_norm": 0.21947209537029266,
"learning_rate": 7.5e-07,
"loss": -0.097,
"num_tokens": 9545087.0,
"reward": -0.10511145042255521,
"reward_std": 0.7483287081122398,
"rewards/cosine_scaled_reward": -0.10957573121413589,
"rewards/format_reward": 0.004464285797439516,
"step": 15
},
{
"clip_ratio": 0.0,
"completion_length": 587.3013725280762,
"epoch": 0.16223067173637515,
"grad_norm": 0.20107129216194153,
"learning_rate": 8e-07,
"loss": -0.0894,
"num_tokens": 10198237.0,
"reward": -0.14020115474704653,
"reward_std": 0.7795401588082314,
"rewards/cosine_scaled_reward": -0.14689757686574012,
"rewards/format_reward": 0.006696428870782256,
"step": 16
},
{
"clip_ratio": 0.0,
"completion_length": 616.6361923217773,
"epoch": 0.17237008871989862,
"grad_norm": 0.17595961689949036,
"learning_rate": 8.499999999999999e-07,
"loss": -0.1258,
"num_tokens": 10885255.0,
"reward": -0.16032503126189113,
"reward_std": 0.7151542007923126,
"rewards/cosine_scaled_reward": -0.16478930541779846,
"rewards/format_reward": 0.004464285913854837,
"step": 17
},
{
"clip_ratio": 0.0,
"completion_length": 614.3348388671875,
"epoch": 0.18250950570342206,
"grad_norm": 0.17326559126377106,
"learning_rate": 9e-07,
"loss": -0.1223,
"num_tokens": 11559043.0,
"reward": -0.05317553365603089,
"reward_std": 0.8043533340096474,
"rewards/cosine_scaled_reward": -0.055407675448805094,
"rewards/format_reward": 0.0022321429569274187,
"step": 18
},
{
"clip_ratio": 0.0,
"completion_length": 576.6774749755859,
"epoch": 0.1926489226869455,
"grad_norm": 0.19331927597522736,
"learning_rate": 9.499999999999999e-07,
"loss": -0.0965,
"num_tokens": 12207730.0,
"reward": -0.01400054944679141,
"reward_std": 0.8140848129987717,
"rewards/cosine_scaled_reward": -0.017348763067275286,
"rewards/format_reward": 0.003348214435391128,
"step": 19
},
{
"clip_ratio": 0.0,
"completion_length": 659.9911117553711,
"epoch": 0.20278833967046894,
"grad_norm": 0.16300565004348755,
"learning_rate": 1e-06,
"loss": -0.0491,
"num_tokens": 12919506.0,
"reward": 0.12035302398726344,
"reward_std": 0.7825085148215294,
"rewards/cosine_scaled_reward": 0.11588872922584414,
"rewards/format_reward": 0.004464285797439516,
"step": 20
},
{
"clip_ratio": 0.0,
"completion_length": 641.3013687133789,
"epoch": 0.21292775665399238,
"grad_norm": 0.24253493547439575,
"learning_rate": 9.999203468625015e-07,
"loss": -0.0523,
"num_tokens": 13615384.0,
"reward": 0.19580535404384136,
"reward_std": 0.8355192169547081,
"rewards/cosine_scaled_reward": 0.19357320480048656,
"rewards/format_reward": 0.0022321429569274187,
"step": 21
},
{
"clip_ratio": 0.0,
"completion_length": 694.412971496582,
"epoch": 0.22306717363751585,
"grad_norm": 0.22512537240982056,
"learning_rate": 9.99681412828496e-07,
"loss": -0.0494,
"num_tokens": 14370738.0,
"reward": 0.2461150388699025,
"reward_std": 0.7426658719778061,
"rewards/cosine_scaled_reward": 0.24388288520276546,
"rewards/format_reward": 0.0022321429569274187,
"step": 22
},
{
"clip_ratio": 0.0,
"completion_length": 675.4542617797852,
"epoch": 0.2332065906210393,
"grad_norm": 0.16492308676242828,
"learning_rate": 9.992832740253644e-07,
"loss": -0.0398,
"num_tokens": 15114193.0,
"reward": 0.2721172561869025,
"reward_std": 0.8025888651609421,
"rewards/cosine_scaled_reward": 0.2710011739982292,
"rewards/format_reward": 0.0011160714784637094,
"step": 23
},
{
"clip_ratio": 0.0,
"completion_length": 719.147346496582,
"epoch": 0.24334600760456274,
"grad_norm": 0.1536169797182083,
"learning_rate": 9.987260573051267e-07,
"loss": -0.0321,
"num_tokens": 15888093.0,
"reward": 0.30394730158150196,
"reward_std": 0.7235733941197395,
"rewards/cosine_scaled_reward": 0.30394728668034077,
"rewards/format_reward": 0.0,
"step": 24
},
{
"clip_ratio": 0.0,
"completion_length": 665.2589645385742,
"epoch": 0.2534854245880862,
"grad_norm": 0.15167948603630066,
"learning_rate": 9.98009940204023e-07,
"loss": -0.0196,
"num_tokens": 16615493.0,
"reward": 0.4243004210293293,
"reward_std": 0.784424401819706,
"rewards/cosine_scaled_reward": 0.42318433709442616,
"rewards/format_reward": 0.0011160714784637094,
"step": 25
},
{
"clip_ratio": 0.0,
"completion_length": 680.0814971923828,
"epoch": 0.26362484157160965,
"grad_norm": 0.1787332445383072,
"learning_rate": 9.971351508859486e-07,
"loss": -0.0554,
"num_tokens": 17343678.0,
"reward": 0.5126287564635277,
"reward_std": 0.7500302940607071,
"rewards/cosine_scaled_reward": 0.5126287154853344,
"rewards/format_reward": 0.0,
"step": 26
},
{
"clip_ratio": 0.0,
"completion_length": 691.2957992553711,
"epoch": 0.2737642585551331,
"grad_norm": 0.15422694385051727,
"learning_rate": 9.961019680697591e-07,
"loss": 0.0019,
"num_tokens": 18091255.0,
"reward": 0.478223854675889,
"reward_std": 0.7884277924895287,
"rewards/cosine_scaled_reward": 0.4771077586337924,
"rewards/format_reward": 0.0011160714784637094,
"step": 27
},
{
"clip_ratio": 0.0,
"completion_length": 724.2433319091797,
"epoch": 0.28390367553865653,
"grad_norm": 0.14453695714473724,
"learning_rate": 9.949107209404663e-07,
"loss": 0.0015,
"num_tokens": 18873817.0,
"reward": 0.6080499514937401,
"reward_std": 0.7710720151662827,
"rewards/cosine_scaled_reward": 0.6080499142408371,
"rewards/format_reward": 0.0,
"step": 28
},
{
"clip_ratio": 0.0,
"completion_length": 719.5837326049805,
"epoch": 0.29404309252217997,
"grad_norm": 0.14530107378959656,
"learning_rate": 9.935617890443554e-07,
"loss": 0.0026,
"num_tokens": 19659300.0,
"reward": 0.6268112398684025,
"reward_std": 0.6641118600964546,
"rewards/cosine_scaled_reward": 0.6268112063407898,
"rewards/format_reward": 0.0,
"step": 29
},
{
"clip_ratio": 0.0,
"completion_length": 793.0312805175781,
"epoch": 0.3041825095057034,
"grad_norm": 0.1358717679977417,
"learning_rate": 9.92055602168058e-07,
"loss": -0.0059,
"num_tokens": 20515344.0,
"reward": 0.6840809024870396,
"reward_std": 0.6523778513073921,
"rewards/cosine_scaled_reward": 0.6840808726847172,
"rewards/format_reward": 0.0,
"step": 30
},
{
"clip_ratio": 0.0,
"completion_length": 766.2176818847656,
"epoch": 0.31432192648922685,
"grad_norm": 0.13609299063682556,
"learning_rate": 9.90392640201615e-07,
"loss": 0.0048,
"num_tokens": 21329483.0,
"reward": 0.7440241314470768,
"reward_std": 0.6845037117600441,
"rewards/cosine_scaled_reward": 0.7440241128206253,
"rewards/format_reward": 0.0,
"step": 31
},
{
"clip_ratio": 0.0,
"completion_length": 760.3225784301758,
"epoch": 0.3244613434727503,
"grad_norm": 0.1422865241765976,
"learning_rate": 9.885734329855797e-07,
"loss": 0.0076,
"num_tokens": 22130476.0,
"reward": 0.8551607951521873,
"reward_std": 0.6532174497842789,
"rewards/cosine_scaled_reward": 0.8551607728004456,
"rewards/format_reward": 0.0,
"step": 32
},
{
"clip_ratio": 0.0,
"completion_length": 788.7120895385742,
"epoch": 0.33460076045627374,
"grad_norm": 0.13245096802711487,
"learning_rate": 9.865985601422017e-07,
"loss": -0.0077,
"num_tokens": 22971218.0,
"reward": 0.7257559858262539,
"reward_std": 0.6099549047648907,
"rewards/cosine_scaled_reward": 0.7302779853343964,
"rewards/format_reward": 0.0,
"step": 33
},
{
"clip_ratio": 0.0,
"completion_length": 802.2344055175781,
"epoch": 0.34474017743979724,
"grad_norm": 0.13606438040733337,
"learning_rate": 9.844686508907537e-07,
"loss": -0.0022,
"num_tokens": 23817508.0,
"reward": 0.9126424044370651,
"reward_std": 0.6410668790340424,
"rewards/cosine_scaled_reward": 0.9126423671841621,
"rewards/format_reward": 0.0,
"step": 34
},
{
"clip_ratio": 0.0,
"completion_length": 779.1975784301758,
"epoch": 0.3548795944233207,
"grad_norm": 0.13803985714912415,
"learning_rate": 9.821843838470534e-07,
"loss": -0.0014,
"num_tokens": 24658397.0,
"reward": 0.951663188636303,
"reward_std": 0.6128395721316338,
"rewards/cosine_scaled_reward": 0.9516631364822388,
"rewards/format_reward": 0.0,
"step": 35
},
{
"clip_ratio": 0.0,
"completion_length": 790.4141006469727,
"epoch": 0.3650190114068441,
"grad_norm": 0.15671393275260925,
"learning_rate": 9.797464868072486e-07,
"loss": 0.0435,
"num_tokens": 25494720.0,
"reward": 0.952189639210701,
"reward_std": 0.6134774163365364,
"rewards/cosine_scaled_reward": 0.9521896243095398,
"rewards/format_reward": 0.0,
"step": 36
},
{
"clip_ratio": 0.0,
"completion_length": 780.6808395385742,
"epoch": 0.37515842839036756,
"grad_norm": 0.13327902555465698,
"learning_rate": 9.771557365159319e-07,
"loss": -0.0051,
"num_tokens": 26319738.0,
"reward": 0.9673999920487404,
"reward_std": 0.5673549436032772,
"rewards/cosine_scaled_reward": 0.9673999547958374,
"rewards/format_reward": 0.0,
"step": 37
},
{
"clip_ratio": 0.0,
"completion_length": 797.1105194091797,
"epoch": 0.385297845373891,
"grad_norm": 0.14108316600322723,
"learning_rate": 9.744129584186597e-07,
"loss": -0.0108,
"num_tokens": 27160557.0,
"reward": 0.9532012119889259,
"reward_std": 0.5637913830578327,
"rewards/cosine_scaled_reward": 0.9532012045383453,
"rewards/format_reward": 0.0,
"step": 38
},
{
"clip_ratio": 0.0,
"completion_length": 857.9230270385742,
"epoch": 0.39543726235741444,
"grad_norm": 0.13035330176353455,
"learning_rate": 9.71519026398956e-07,
"loss": 0.0268,
"num_tokens": 28067040.0,
"reward": 0.9902723953127861,
"reward_std": 0.5246665701270103,
"rewards/cosine_scaled_reward": 0.9902723580598831,
"rewards/format_reward": 0.0,
"step": 39
},
{
"clip_ratio": 0.0,
"completion_length": 793.2377471923828,
"epoch": 0.4055766793409379,
"grad_norm": 0.13015364110469818,
"learning_rate": 9.68474862499881e-07,
"loss": 0.0268,
"num_tokens": 28897045.0,
"reward": 1.063771203160286,
"reward_std": 0.49807993322610855,
"rewards/cosine_scaled_reward": 1.0637711435556412,
"rewards/format_reward": 0.0,
"step": 40
},
{
"clip_ratio": 0.0,
"completion_length": 776.3783798217773,
"epoch": 0.4157160963244613,
"grad_norm": 0.12312573939561844,
"learning_rate": 9.652814366302568e-07,
"loss": 0.0136,
"num_tokens": 29718656.0,
"reward": 1.1480179950594902,
"reward_std": 0.5026725828647614,
"rewards/cosine_scaled_reward": 1.1480179578065872,
"rewards/format_reward": 0.0,
"step": 41
},
{
"clip_ratio": 0.0,
"completion_length": 782.9832916259766,
"epoch": 0.42585551330798477,
"grad_norm": 0.12757915258407593,
"learning_rate": 9.619397662556433e-07,
"loss": 0.0089,
"num_tokens": 30549177.0,
"reward": 1.153884269297123,
"reward_std": 0.43890631198883057,
"rewards/cosine_scaled_reward": 1.1538842469453812,
"rewards/format_reward": 0.0,
"step": 42
},
{
"clip_ratio": 0.0,
"completion_length": 855.8906631469727,
"epoch": 0.43599493029150826,
"grad_norm": 0.11601021140813828,
"learning_rate": 9.5845091607416e-07,
"loss": -0.0027,
"num_tokens": 31434759.0,
"reward": 1.1496833115816116,
"reward_std": 0.44448646157979965,
"rewards/cosine_scaled_reward": 1.1496832445263863,
"rewards/format_reward": 0.0,
"step": 43
},
{
"clip_ratio": 0.0,
"completion_length": 862.2332916259766,
"epoch": 0.4461343472750317,
"grad_norm": 0.11746807396411896,
"learning_rate": 9.548159976772592e-07,
"loss": 0.0197,
"num_tokens": 32337368.0,
"reward": 1.1307310312986374,
"reward_std": 0.4776446260511875,
"rewards/cosine_scaled_reward": 1.130730964243412,
"rewards/format_reward": 0.0,
"step": 44
},
{
"clip_ratio": 0.0,
"completion_length": 900.1975860595703,
"epoch": 0.45627376425855515,
"grad_norm": 0.1248885914683342,
"learning_rate": 9.510361691955606e-07,
"loss": 0.0165,
"num_tokens": 33266393.0,
"reward": 1.060933604836464,
"reward_std": 0.46989936381578445,
"rewards/cosine_scaled_reward": 1.060933567583561,
"rewards/format_reward": 0.0,
"step": 45
},
{
"clip_ratio": 0.0,
"completion_length": 815.2812881469727,
"epoch": 0.4664131812420786,
"grad_norm": 0.9569998979568481,
"learning_rate": 9.471126349298556e-07,
"loss": 0.0354,
"num_tokens": 34127101.0,
"reward": 1.151917465031147,
"reward_std": 0.5036191828548908,
"rewards/cosine_scaled_reward": 1.1519174352288246,
"rewards/format_reward": 0.0,
"step": 46
},
{
"clip_ratio": 0.0,
"completion_length": 833.6239242553711,
"epoch": 0.47655259822560203,
"grad_norm": 0.15996558964252472,
"learning_rate": 9.430466449674013e-07,
"loss": -0.0196,
"num_tokens": 35022380.0,
"reward": 1.0830270573496819,
"reward_std": 0.4072240814566612,
"rewards/cosine_scaled_reward": 1.0830269828438759,
"rewards/format_reward": 0.0,
"step": 47
},
{
"clip_ratio": 0.0,
"completion_length": 892.449821472168,
"epoch": 0.4866920152091255,
"grad_norm": 0.11413749307394028,
"learning_rate": 9.388394947836278e-07,
"loss": -0.0125,
"num_tokens": 35948751.0,
"reward": 1.0179572254419327,
"reward_std": 0.45433366298675537,
"rewards/cosine_scaled_reward": 1.0179571583867073,
"rewards/format_reward": 0.0,
"step": 48
},
{
"clip_ratio": 0.0,
"completion_length": 912.2645492553711,
"epoch": 0.4968314321926489,
"grad_norm": 0.11429373919963837,
"learning_rate": 9.344925248293835e-07,
"loss": -0.001,
"num_tokens": 36887148.0,
"reward": 1.1400989294052124,
"reward_std": 0.4401072785258293,
"rewards/cosine_scaled_reward": 1.1400988548994064,
"rewards/format_reward": 0.0,
"step": 49
},
{
"clip_ratio": 0.0,
"completion_length": 866.2511596679688,
"epoch": 0.5069708491761724,
"grad_norm": 0.11980035156011581,
"learning_rate": 9.300071201038501e-07,
"loss": 0.0215,
"num_tokens": 37789485.0,
"reward": 1.13294617831707,
"reward_std": 0.4442785568535328,
"rewards/cosine_scaled_reward": 1.1329461187124252,
"rewards/format_reward": 0.0,
"step": 50
},
{
"clip_ratio": 0.0,
"completion_length": 799.482177734375,
"epoch": 0.5171102661596958,
"grad_norm": 0.122144915163517,
"learning_rate": 9.253847097132655e-07,
"loss": 0.0044,
"num_tokens": 38636645.0,
"reward": 1.2101461067795753,
"reward_std": 0.41771573945879936,
"rewards/cosine_scaled_reward": 1.210146002471447,
"rewards/format_reward": 0.0,
"step": 51
},
{
"clip_ratio": 0.0,
"completion_length": 847.4330749511719,
"epoch": 0.5272496831432193,
"grad_norm": 0.1199701800942421,
"learning_rate": 9.206267664155906e-07,
"loss": 0.0082,
"num_tokens": 39516321.0,
"reward": 1.1085221618413925,
"reward_std": 0.4500999916344881,
"rewards/cosine_scaled_reward": 1.1085221245884895,
"rewards/format_reward": 0.0,
"step": 52
},
{
"clip_ratio": 0.0,
"completion_length": 873.8102951049805,
"epoch": 0.5373891001267427,
"grad_norm": 0.14303860068321228,
"learning_rate": 9.157348061512726e-07,
"loss": -0.0033,
"num_tokens": 40435335.0,
"reward": 1.1145585402846336,
"reward_std": 0.47080911323428154,
"rewards/cosine_scaled_reward": 1.1145584508776665,
"rewards/format_reward": 0.0,
"step": 53
},
{
"clip_ratio": 0.0,
"completion_length": 961.9029312133789,
"epoch": 0.5475285171102662,
"grad_norm": 0.11308898031711578,
"learning_rate": 9.107103875602458e-07,
"loss": -0.0123,
"num_tokens": 41415752.0,
"reward": 1.0546733513474464,
"reward_std": 0.41840433329343796,
"rewards/cosine_scaled_reward": 1.054673284292221,
"rewards/format_reward": 0.0,
"step": 54
},
{
"clip_ratio": 0.0,
"completion_length": 892.6484680175781,
"epoch": 0.5576679340937896,
"grad_norm": 0.12083946913480759,
"learning_rate": 9.055551114853295e-07,
"loss": 0.0043,
"num_tokens": 42345957.0,
"reward": 1.0607689917087555,
"reward_std": 0.4757756106555462,
"rewards/cosine_scaled_reward": 1.0607689544558525,
"rewards/format_reward": 0.0,
"step": 55
},
{
"clip_ratio": 0.0,
"completion_length": 823.7578506469727,
"epoch": 0.5678073510773131,
"grad_norm": 0.21660108864307404,
"learning_rate": 9.002706204621802e-07,
"loss": 0.0311,
"num_tokens": 43207724.0,
"reward": 1.1668334901332855,
"reward_std": 0.482263408601284,
"rewards/cosine_scaled_reward": 1.1668334528803825,
"rewards/format_reward": 0.0,
"step": 56
},
{
"clip_ratio": 0.0,
"completion_length": 844.7210159301758,
"epoch": 0.5779467680608364,
"grad_norm": 0.4439857602119446,
"learning_rate": 8.948585981959578e-07,
"loss": 0.0112,
"num_tokens": 44096178.0,
"reward": 1.1872114390134811,
"reward_std": 0.44185368344187737,
"rewards/cosine_scaled_reward": 1.1872113794088364,
"rewards/format_reward": 0.0,
"step": 57
},
{
"clip_ratio": 0.0,
"completion_length": 832.4553909301758,
"epoch": 0.5880861850443599,
"grad_norm": 0.12814383208751678,
"learning_rate": 8.893207690248775e-07,
"loss": -0.0022,
"num_tokens": 44963122.0,
"reward": 1.1495825573801994,
"reward_std": 0.43896420300006866,
"rewards/cosine_scaled_reward": 1.149582527577877,
"rewards/format_reward": 0.0,
"step": 58
},
{
"clip_ratio": 0.0,
"completion_length": 819.3661117553711,
"epoch": 0.5982256020278834,
"grad_norm": 0.12473784387111664,
"learning_rate": 8.836588973708128e-07,
"loss": -0.0152,
"num_tokens": 45838866.0,
"reward": 1.101597860455513,
"reward_std": 0.44026929885149,
"rewards/cosine_scaled_reward": 1.1015978381037712,
"rewards/format_reward": 0.0,
"step": 59
},
{
"clip_ratio": 0.0,
"completion_length": 883.5223617553711,
"epoch": 0.6083650190114068,
"grad_norm": 0.12360174208879471,
"learning_rate": 8.778747871771291e-07,
"loss": 0.0119,
"num_tokens": 46767750.0,
"reward": 1.1120553016662598,
"reward_std": 0.43890899419784546,
"rewards/cosine_scaled_reward": 1.1120552495121956,
"rewards/format_reward": 0.0,
"step": 60
},
{
"clip_ratio": 0.0,
"completion_length": 830.4174499511719,
"epoch": 0.6185044359949303,
"grad_norm": 0.12516207993030548,
"learning_rate": 8.719702813339247e-07,
"loss": 0.007,
"num_tokens": 47634148.0,
"reward": 1.1350341737270355,
"reward_std": 0.3932863809168339,
"rewards/cosine_scaled_reward": 1.1350341215729713,
"rewards/format_reward": 0.0,
"step": 61
},
{
"clip_ratio": 0.0,
"completion_length": 860.8895492553711,
"epoch": 0.6286438529784537,
"grad_norm": 0.11765479296445847,
"learning_rate": 8.659472610908627e-07,
"loss": -0.0127,
"num_tokens": 48534897.0,
"reward": 1.0959327295422554,
"reward_std": 0.43786268681287766,
"rewards/cosine_scaled_reward": 1.09593266248703,
"rewards/format_reward": 0.0,
"step": 62
},
{
"clip_ratio": 0.0,
"completion_length": 789.957633972168,
"epoch": 0.6387832699619772,
"grad_norm": 0.12538108229637146,
"learning_rate": 8.598076454577814e-07,
"loss": 0.0043,
"num_tokens": 49366995.0,
"reward": 1.2128012925386429,
"reward_std": 0.5132176093757153,
"rewards/cosine_scaled_reward": 1.2128012254834175,
"rewards/format_reward": 0.0,
"step": 63
},
{
"clip_ratio": 0.0,
"completion_length": 814.1540603637695,
"epoch": 0.6489226869455006,
"grad_norm": 0.12521201372146606,
"learning_rate": 8.535533905932737e-07,
"loss": 0.031,
"num_tokens": 50244605.0,
"reward": 1.2290623039007187,
"reward_std": 0.35417212545871735,
"rewards/cosine_scaled_reward": 1.2290622666478157,
"rewards/format_reward": 0.0,
"step": 64
},
{
"clip_ratio": 0.0,
"completion_length": 826.7444610595703,
"epoch": 0.6590621039290241,
"grad_norm": 0.1192706748843193,
"learning_rate": 8.471864891814304e-07,
"loss": 0.0187,
"num_tokens": 51104472.0,
"reward": 1.2224428355693817,
"reward_std": 0.4238445721566677,
"rewards/cosine_scaled_reward": 1.2224428057670593,
"rewards/format_reward": 0.0,
"step": 65
},
{
"clip_ratio": 0.0,
"completion_length": 890.8181228637695,
"epoch": 0.6692015209125475,
"grad_norm": 0.11663277447223663,
"learning_rate": 8.407089697969456e-07,
"loss": -0.0075,
"num_tokens": 52027509.0,
"reward": 1.1307330876588821,
"reward_std": 0.43055112659931183,
"rewards/cosine_scaled_reward": 1.130733035504818,
"rewards/format_reward": 0.0,
"step": 66
},
{
"clip_ratio": 0.0,
"completion_length": 896.5524978637695,
"epoch": 0.679340937896071,
"grad_norm": 0.11125318706035614,
"learning_rate": 8.341228962587881e-07,
"loss": -0.0128,
"num_tokens": 52966948.0,
"reward": 1.1304329186677933,
"reward_std": 0.38643651083111763,
"rewards/cosine_scaled_reward": 1.1304328739643097,
"rewards/format_reward": 0.0,
"step": 67
},
{
"clip_ratio": 0.0,
"completion_length": 805.1919937133789,
"epoch": 0.6894803548795945,
"grad_norm": 0.1318301111459732,
"learning_rate": 8.274303669726426e-07,
"loss": 0.0267,
"num_tokens": 53809720.0,
"reward": 1.2848119288682938,
"reward_std": 0.4024609997868538,
"rewards/cosine_scaled_reward": 1.2848118543624878,
"rewards/format_reward": 0.0,
"step": 68
},
{
"clip_ratio": 0.0,
"completion_length": 897.3783950805664,
"epoch": 0.6996197718631179,
"grad_norm": 0.11534889042377472,
"learning_rate": 8.206335142623304e-07,
"loss": -0.0213,
"num_tokens": 54743531.0,
"reward": 1.0870068296790123,
"reward_std": 0.396763913333416,
"rewards/cosine_scaled_reward": 1.0870067551732063,
"rewards/format_reward": 0.0,
"step": 69
},
{
"clip_ratio": 0.0,
"completion_length": 849.7935562133789,
"epoch": 0.7097591888466414,
"grad_norm": 0.11737702041864395,
"learning_rate": 8.137345036904259e-07,
"loss": 0.0049,
"num_tokens": 55624322.0,
"reward": 1.1217299401760101,
"reward_std": 0.4600865840911865,
"rewards/cosine_scaled_reward": 1.1217298731207848,
"rewards/format_reward": 0.0,
"step": 70
},
{
"clip_ratio": 0.0,
"completion_length": 843.4699020385742,
"epoch": 0.7198986058301647,
"grad_norm": 0.12105900794267654,
"learning_rate": 8.067355333682797e-07,
"loss": 0.0058,
"num_tokens": 56518839.0,
"reward": 1.128207340836525,
"reward_std": 0.39387310668826103,
"rewards/cosine_scaled_reward": 1.1282072588801384,
"rewards/format_reward": 0.0,
"step": 71
},
{
"clip_ratio": 0.0,
"completion_length": 854.9007110595703,
"epoch": 0.7300380228136882,
"grad_norm": 0.11754778772592545,
"learning_rate": 7.996388332556734e-07,
"loss": 0.0184,
"num_tokens": 57413550.0,
"reward": 1.1404274106025696,
"reward_std": 0.45089153945446014,
"rewards/cosine_scaled_reward": 1.1404273584485054,
"rewards/format_reward": 0.0,
"step": 72
},
{
"clip_ratio": 0.0,
"completion_length": 900.8973541259766,
"epoch": 0.7401774397972116,
"grad_norm": 0.11590257287025452,
"learning_rate": 7.924466644503264e-07,
"loss": -0.0059,
"num_tokens": 58353250.0,
"reward": 1.1319697871804237,
"reward_std": 0.4219101257622242,
"rewards/cosine_scaled_reward": 1.131969727575779,
"rewards/format_reward": 0.0,
"step": 73
},
{
"clip_ratio": 0.0,
"completion_length": 806.4665603637695,
"epoch": 0.7503168567807351,
"grad_norm": 0.12108156830072403,
"learning_rate": 7.85161318467482e-07,
"loss": -0.0001,
"num_tokens": 59211228.0,
"reward": 1.1962338984012604,
"reward_std": 0.45812737941741943,
"rewards/cosine_scaled_reward": 1.196233868598938,
"rewards/format_reward": 0.0,
"step": 74
},
{
"clip_ratio": 0.0,
"completion_length": 812.6629791259766,
"epoch": 0.7604562737642585,
"grad_norm": 0.12231780588626862,
"learning_rate": 7.777851165098011e-07,
"loss": 0.0223,
"num_tokens": 60070054.0,
"reward": 1.2705483883619308,
"reward_std": 0.4315441697835922,
"rewards/cosine_scaled_reward": 1.2705483138561249,
"rewards/format_reward": 0.0,
"step": 75
},
{
"clip_ratio": 0.0,
"completion_length": 792.5323944091797,
"epoch": 0.770595690747782,
"grad_norm": 0.13156548142433167,
"learning_rate": 7.703204087277988e-07,
"loss": 0.0054,
"num_tokens": 60907403.0,
"reward": 1.0341035649180412,
"reward_std": 0.4903283417224884,
"rewards/cosine_scaled_reward": 1.0341035276651382,
"rewards/format_reward": 0.0,
"step": 76
},
{
"clip_ratio": 0.0,
"completion_length": 843.8315200805664,
"epoch": 0.7807351077313055,
"grad_norm": 0.12101699411869049,
"learning_rate": 7.627695734710564e-07,
"loss": 0.0109,
"num_tokens": 61785700.0,
"reward": 1.226922646164894,
"reward_std": 0.38846197351813316,
"rewards/cosine_scaled_reward": 1.2269226014614105,
"rewards/format_reward": 0.0,
"step": 77
},
{
"clip_ratio": 0.0,
"completion_length": 859.2433395385742,
"epoch": 0.7908745247148289,
"grad_norm": 0.1146969199180603,
"learning_rate": 7.551350165304499e-07,
"loss": 0.0219,
"num_tokens": 62673358.0,
"reward": 1.2758738100528717,
"reward_std": 0.4126880206167698,
"rewards/cosine_scaled_reward": 1.275873750448227,
"rewards/format_reward": 0.0,
"step": 78
},
{
"clip_ratio": 0.0,
"completion_length": 847.4732513427734,
"epoch": 0.8010139416983524,
"grad_norm": 0.11725428700447083,
"learning_rate": 7.474191703716338e-07,
"loss": -0.0134,
"num_tokens": 63563254.0,
"reward": 1.1727145612239838,
"reward_std": 0.3684752322733402,
"rewards/cosine_scaled_reward": 1.1727144792675972,
"rewards/format_reward": 0.0,
"step": 79
},
{
"clip_ratio": 0.0,
"completion_length": 831.0089721679688,
"epoch": 0.8111533586818758,
"grad_norm": 0.1215280145406723,
"learning_rate": 7.396244933600284e-07,
"loss": 0.0269,
"num_tokens": 64432126.0,
"reward": 1.1777353882789612,
"reward_std": 0.420895554125309,
"rewards/cosine_scaled_reward": 1.1777353659272194,
"rewards/format_reward": 0.0,
"step": 80
},
{
"clip_ratio": 0.0,
"completion_length": 848.4598617553711,
"epoch": 0.8212927756653993,
"grad_norm": 0.13223014771938324,
"learning_rate": 7.317534689775527e-07,
"loss": 0.0072,
"num_tokens": 65312050.0,
"reward": 1.1698129922151566,
"reward_std": 0.4710509404540062,
"rewards/cosine_scaled_reward": 1.1698129400610924,
"rewards/format_reward": 0.0,
"step": 81
},
{
"clip_ratio": 0.0,
"completion_length": 811.4230346679688,
"epoch": 0.8314321926489227,
"grad_norm": 0.12015990167856216,
"learning_rate": 7.238086050313562e-07,
"loss": -0.0264,
"num_tokens": 66174341.0,
"reward": 1.124290645122528,
"reward_std": 0.40254683420062065,
"rewards/cosine_scaled_reward": 1.1242906004190445,
"rewards/format_reward": 0.0,
"step": 82
},
{
"clip_ratio": 0.0,
"completion_length": 819.8382110595703,
"epoch": 0.8415716096324461,
"grad_norm": 0.14578643441200256,
"learning_rate": 7.157924328548002e-07,
"loss": 0.026,
"num_tokens": 67042508.0,
"reward": 1.2035124003887177,
"reward_std": 0.4501505568623543,
"rewards/cosine_scaled_reward": 1.2035123333334923,
"rewards/format_reward": 0.0,
"step": 83
},
{
"clip_ratio": 0.0,
"completion_length": 832.864990234375,
"epoch": 0.8517110266159695,
"grad_norm": 0.11915823817253113,
"learning_rate": 7.077075065009433e-07,
"loss": 0.0419,
"num_tokens": 67922067.0,
"reward": 1.2239351123571396,
"reward_std": 0.4415438733994961,
"rewards/cosine_scaled_reward": 1.2239350602030754,
"rewards/format_reward": 0.0,
"step": 84
},
{
"clip_ratio": 0.0,
"completion_length": 794.1897659301758,
"epoch": 0.861850443599493,
"grad_norm": 0.12470311671495438,
"learning_rate": 6.995564019287869e-07,
"loss": 0.0085,
"num_tokens": 68763477.0,
"reward": 1.1386212185025215,
"reward_std": 0.4416712336242199,
"rewards/cosine_scaled_reward": 1.1386211588978767,
"rewards/format_reward": 0.0,
"step": 85
},
{
"clip_ratio": 0.0,
"completion_length": 865.2344207763672,
"epoch": 0.8719898605830165,
"grad_norm": 0.1167558878660202,
"learning_rate": 6.913417161825449e-07,
"loss": -0.0029,
"num_tokens": 69673639.0,
"reward": 1.1339842602610588,
"reward_std": 0.44360488280653954,
"rewards/cosine_scaled_reward": 1.1339841783046722,
"rewards/format_reward": 0.0,
"step": 86
},
{
"clip_ratio": 0.0,
"completion_length": 807.192008972168,
"epoch": 0.8821292775665399,
"grad_norm": 0.22383332252502441,
"learning_rate": 6.830660665641897e-07,
"loss": 0.0167,
"num_tokens": 70526603.0,
"reward": 1.1783367395401,
"reward_std": 0.44941645860671997,
"rewards/cosine_scaled_reward": 1.1783367022871971,
"rewards/format_reward": 0.0,
"step": 87
},
{
"clip_ratio": 0.0,
"completion_length": 842.9832916259766,
"epoch": 0.8922686945500634,
"grad_norm": 0.12214665859937668,
"learning_rate": 6.747320897995492e-07,
"loss": 0.0028,
"num_tokens": 71410908.0,
"reward": 1.1979937851428986,
"reward_std": 0.41686391085386276,
"rewards/cosine_scaled_reward": 1.1979937255382538,
"rewards/format_reward": 0.0,
"step": 88
},
{
"clip_ratio": 0.0,
"completion_length": 752.7991409301758,
"epoch": 0.9024081115335868,
"grad_norm": 0.12918852269649506,
"learning_rate": 6.66342441198212e-07,
"loss": -0.0199,
"num_tokens": 72209320.0,
"reward": 1.1556047648191452,
"reward_std": 0.39559851959347725,
"rewards/cosine_scaled_reward": 1.1556047424674034,
"rewards/format_reward": 0.0,
"step": 89
},
{
"clip_ratio": 0.0,
"completion_length": 788.5815048217773,
"epoch": 0.9125475285171103,
"grad_norm": 0.2962648272514343,
"learning_rate": 6.578997938075125e-07,
"loss": 0.0258,
"num_tokens": 73040361.0,
"reward": 1.1919841021299362,
"reward_std": 0.40032988227903843,
"rewards/cosine_scaled_reward": 1.1919840648770332,
"rewards/format_reward": 0.0,
"step": 90
},
{
"clip_ratio": 0.0,
"completion_length": 787.3214645385742,
"epoch": 0.9226869455006337,
"grad_norm": 0.15694542229175568,
"learning_rate": 6.494068375608646e-07,
"loss": -0.0147,
"num_tokens": 73880425.0,
"reward": 1.1829868927598,
"reward_std": 0.3718248065561056,
"rewards/cosine_scaled_reward": 1.1829868257045746,
"rewards/format_reward": 0.0,
"step": 91
},
{
"clip_ratio": 0.0,
"completion_length": 821.3850708007812,
"epoch": 0.9328263624841572,
"grad_norm": 0.11674854159355164,
"learning_rate": 6.408662784207149e-07,
"loss": 0.0095,
"num_tokens": 74755762.0,
"reward": 1.2243514209985733,
"reward_std": 0.4170701913535595,
"rewards/cosine_scaled_reward": 1.2243513762950897,
"rewards/format_reward": 0.0,
"step": 92
},
{
"clip_ratio": 0.0,
"completion_length": 761.3248138427734,
"epoch": 0.9429657794676806,
"grad_norm": 0.12954537570476532,
"learning_rate": 6.322808375163895e-07,
"loss": 0.0386,
"num_tokens": 75564789.0,
"reward": 1.1891243010759354,
"reward_std": 0.45389125496149063,
"rewards/cosine_scaled_reward": 1.1891242563724518,
"rewards/format_reward": 0.0,
"step": 93
},
{
"clip_ratio": 0.0,
"completion_length": 793.9643249511719,
"epoch": 0.9531051964512041,
"grad_norm": 0.12744048237800598,
"learning_rate": 6.236532502771077e-07,
"loss": 0.0271,
"num_tokens": 76409781.0,
"reward": 1.1992642730474472,
"reward_std": 0.4242793843150139,
"rewards/cosine_scaled_reward": 1.1992641612887383,
"rewards/format_reward": 0.0,
"step": 94
},
{
"clip_ratio": 0.0,
"completion_length": 789.4219207763672,
"epoch": 0.9632446134347274,
"grad_norm": 0.12654094398021698,
"learning_rate": 6.149862655604403e-07,
"loss": 0.0339,
"num_tokens": 77242791.0,
"reward": 1.2284981906414032,
"reward_std": 0.42066042125225067,
"rewards/cosine_scaled_reward": 1.2284981459379196,
"rewards/format_reward": 0.0,
"step": 95
},
{
"clip_ratio": 0.0,
"completion_length": 795.716552734375,
"epoch": 0.973384030418251,
"grad_norm": 0.12548381090164185,
"learning_rate": 6.062826447764883e-07,
"loss": 0.008,
"num_tokens": 78078497.0,
"reward": 1.2442717999219894,
"reward_std": 0.35396804101765156,
"rewards/cosine_scaled_reward": 1.2442717254161835,
"rewards/format_reward": 0.0,
"step": 96
},
{
"clip_ratio": 0.0,
"completion_length": 793.0692291259766,
"epoch": 0.9835234474017744,
"grad_norm": 0.12744414806365967,
"learning_rate": 5.975451610080642e-07,
"loss": -0.0194,
"num_tokens": 78916895.0,
"reward": 1.1445267572999,
"reward_std": 0.37857919186353683,
"rewards/cosine_scaled_reward": 1.1445267051458359,
"rewards/format_reward": 0.0,
"step": 97
},
{
"clip_ratio": 0.0,
"completion_length": 782.1171875,
"epoch": 0.9936628643852978,
"grad_norm": 0.12226420640945435,
"learning_rate": 5.887765981271517e-07,
"loss": 0.0066,
"num_tokens": 79784779.0,
"reward": 1.168272152543068,
"reward_std": 0.39856788516044617,
"rewards/cosine_scaled_reward": 1.1682721078395844,
"rewards/format_reward": 0.0,
"step": 98
},
{
"clip_ratio": 0.0,
"completion_length": 783.4386367797852,
"epoch": 1.0101394169835234,
"grad_norm": 0.13228166103363037,
"learning_rate": 5.7997974990793e-07,
"loss": 0.0145,
"num_tokens": 80615732.0,
"reward": 1.2336678951978683,
"reward_std": 0.38141428492963314,
"rewards/cosine_scaled_reward": 1.233667865395546,
"rewards/format_reward": 0.0,
"step": 99
},
{
"clip_ratio": 0.0,
"completion_length": 806.2355346679688,
"epoch": 1.020278833967047,
"grad_norm": 0.12670038640499115,
"learning_rate": 5.711574191366427e-07,
"loss": 0.0028,
"num_tokens": 81465495.0,
"reward": 1.1623064577579498,
"reward_std": 0.3932801876217127,
"rewards/cosine_scaled_reward": 1.1623064130544662,
"rewards/format_reward": 0.0,
"step": 100
},
{
"clip_ratio": 0.0,
"completion_length": 790.1205749511719,
"epoch": 1.0304182509505704,
"grad_norm": 0.12086564302444458,
"learning_rate": 5.623124167185929e-07,
"loss": 0.0063,
"num_tokens": 82295971.0,
"reward": 1.2574269622564316,
"reward_std": 0.3898525554686785,
"rewards/cosine_scaled_reward": 1.257426917552948,
"rewards/format_reward": 0.0,
"step": 101
},
{
"clip_ratio": 0.0,
"completion_length": 758.897346496582,
"epoch": 1.0405576679340938,
"grad_norm": 0.14097033441066742,
"learning_rate": 5.534475607825565e-07,
"loss": -0.0126,
"num_tokens": 83103231.0,
"reward": 1.2127035036683083,
"reward_std": 0.3951477538794279,
"rewards/cosine_scaled_reward": 1.2127034589648247,
"rewards/format_reward": 0.0,
"step": 102
},
{
"clip_ratio": 0.0,
"completion_length": 823.2377624511719,
"epoch": 1.0506970849176172,
"grad_norm": 0.11900278925895691,
"learning_rate": 5.445656757828879e-07,
"loss": -0.0341,
"num_tokens": 83972156.0,
"reward": 1.0786512047052383,
"reward_std": 0.386748855933547,
"rewards/cosine_scaled_reward": 1.078651174902916,
"rewards/format_reward": 0.0,
"step": 103
},
{
"clip_ratio": 0.0,
"completion_length": 778.1082992553711,
"epoch": 1.0608365019011408,
"grad_norm": 0.12485373765230179,
"learning_rate": 5.356695915996161e-07,
"loss": -0.0152,
"num_tokens": 84803701.0,
"reward": 1.1336441561579704,
"reward_std": 0.43268290534615517,
"rewards/cosine_scaled_reward": 1.1336440965533257,
"rewards/format_reward": 0.0,
"step": 104
},
{
"clip_ratio": 0.0,
"completion_length": 747.7790451049805,
"epoch": 1.0709759188846641,
"grad_norm": 0.14275644719600677,
"learning_rate": 5.267621426368075e-07,
"loss": 0.0402,
"num_tokens": 85592927.0,
"reward": 1.2611572295427322,
"reward_std": 0.4418743886053562,
"rewards/cosine_scaled_reward": 1.2611571997404099,
"rewards/format_reward": 0.0,
"step": 105
},
{
"clip_ratio": 0.0,
"completion_length": 686.8861770629883,
"epoch": 1.0811153358681875,
"grad_norm": 0.2392619252204895,
"learning_rate": 5.178461669194903e-07,
"loss": 0.0233,
"num_tokens": 86335585.0,
"reward": 1.2902618646621704,
"reward_std": 0.4318929500877857,
"rewards/cosine_scaled_reward": 1.2902617901563644,
"rewards/format_reward": 0.0,
"step": 106
},
{
"clip_ratio": 0.0,
"completion_length": 765.6752548217773,
"epoch": 1.091254752851711,
"grad_norm": 0.12888778746128082,
"learning_rate": 5.08924505189423e-07,
"loss": -0.0053,
"num_tokens": 87147590.0,
"reward": 1.2406404912471771,
"reward_std": 0.3950204625725746,
"rewards/cosine_scaled_reward": 1.2406404465436935,
"rewards/format_reward": 0.0,
"step": 107
},
{
"clip_ratio": 0.0,
"completion_length": 774.6596298217773,
"epoch": 1.1013941698352345,
"grad_norm": 0.16063496470451355,
"learning_rate": 5e-07,
"loss": 0.0407,
"num_tokens": 87965805.0,
"reward": 1.2785077691078186,
"reward_std": 0.39564304798841476,
"rewards/cosine_scaled_reward": 1.2785076946020126,
"rewards/format_reward": 0.0,
"step": 108
},
{
"clip_ratio": 0.0,
"completion_length": 800.3393249511719,
"epoch": 1.111533586818758,
"grad_norm": 0.23322370648384094,
"learning_rate": 4.91075494810577e-07,
"loss": 0.0181,
"num_tokens": 88800901.0,
"reward": 1.2374378964304924,
"reward_std": 0.39184754714369774,
"rewards/cosine_scaled_reward": 1.2374378442764282,
"rewards/format_reward": 0.0,
"step": 109
},
{
"clip_ratio": 0.0,
"completion_length": 802.3839721679688,
"epoch": 1.1216730038022813,
"grad_norm": 0.12637439370155334,
"learning_rate": 4.821538330805098e-07,
"loss": -0.0142,
"num_tokens": 89646357.0,
"reward": 1.2594422698020935,
"reward_std": 0.3456545788794756,
"rewards/cosine_scaled_reward": 1.2594422399997711,
"rewards/format_reward": 0.0,
"step": 110
},
{
"clip_ratio": 0.0,
"completion_length": 804.2109680175781,
"epoch": 1.131812420785805,
"grad_norm": 0.4338585138320923,
"learning_rate": 4.732378573631924e-07,
"loss": -0.0124,
"num_tokens": 90489658.0,
"reward": 1.204567864537239,
"reward_std": 0.44952966272830963,
"rewards/cosine_scaled_reward": 1.2045678049325943,
"rewards/format_reward": 0.0,
"step": 111
},
{
"clip_ratio": 0.0,
"completion_length": 814.7243728637695,
"epoch": 1.1419518377693283,
"grad_norm": 0.12191635370254517,
"learning_rate": 4.643304084003838e-07,
"loss": 0.0138,
"num_tokens": 91347099.0,
"reward": 1.1750361323356628,
"reward_std": 0.4308737814426422,
"rewards/cosine_scaled_reward": 1.1750360652804375,
"rewards/format_reward": 0.0,
"step": 112
},
{
"clip_ratio": 0.0,
"completion_length": 767.3794937133789,
"epoch": 1.1520912547528517,
"grad_norm": 0.13367751240730286,
"learning_rate": 4.55434324217112e-07,
"loss": -0.026,
"num_tokens": 92163911.0,
"reward": 1.0875676944851875,
"reward_std": 0.4285357743501663,
"rewards/cosine_scaled_reward": 1.0875676572322845,
"rewards/format_reward": 0.0,
"step": 113
},
{
"clip_ratio": 0.0,
"completion_length": 759.1138687133789,
"epoch": 1.162230671736375,
"grad_norm": 0.6044357419013977,
"learning_rate": 4.4655243921744367e-07,
"loss": -0.0034,
"num_tokens": 92969037.0,
"reward": 1.1716800779104233,
"reward_std": 0.40124501287937164,
"rewards/cosine_scaled_reward": 1.1716800406575203,
"rewards/format_reward": 0.0,
"step": 114
},
{
"clip_ratio": 0.0,
"completion_length": 742.3248138427734,
"epoch": 1.1723700887198987,
"grad_norm": 0.12603311240673065,
"learning_rate": 4.37687583281407e-07,
"loss": -0.0106,
"num_tokens": 93761376.0,
"reward": 1.2006227523088455,
"reward_std": 0.3855649037286639,
"rewards/cosine_scaled_reward": 1.2006226852536201,
"rewards/format_reward": 0.0,
"step": 115
},
{
"clip_ratio": 0.0,
"completion_length": 750.4765930175781,
"epoch": 1.182509505703422,
"grad_norm": 0.13189645111560822,
"learning_rate": 4.2884258086335745e-07,
"loss": 0.0114,
"num_tokens": 94573043.0,
"reward": 1.2511205673217773,
"reward_std": 0.408273559063673,
"rewards/cosine_scaled_reward": 1.2511205524206161,
"rewards/format_reward": 0.0,
"step": 116
},
{
"clip_ratio": 0.0,
"completion_length": 748.3750305175781,
"epoch": 1.1926489226869454,
"grad_norm": 0.13386943936347961,
"learning_rate": 4.2002025009206987e-07,
"loss": 0.0081,
"num_tokens": 95375539.0,
"reward": 1.1935091465711594,
"reward_std": 0.39292260631918907,
"rewards/cosine_scaled_reward": 1.1935090869665146,
"rewards/format_reward": 0.0,
"step": 117
},
{
"clip_ratio": 0.0,
"completion_length": 752.021240234375,
"epoch": 1.202788339670469,
"grad_norm": 0.13410834968090057,
"learning_rate": 4.1122340187284845e-07,
"loss": -0.0055,
"num_tokens": 96171638.0,
"reward": 1.215317651629448,
"reward_std": 0.4204988572746515,
"rewards/cosine_scaled_reward": 1.2153176292777061,
"rewards/format_reward": 0.0,
"step": 118
},
{
"clip_ratio": 0.0,
"completion_length": 741.1294937133789,
"epoch": 1.2129277566539924,
"grad_norm": 0.14352215826511383,
"learning_rate": 4.0245483899193586e-07,
"loss": -0.0021,
"num_tokens": 96963066.0,
"reward": 1.171108528971672,
"reward_std": 0.390108622610569,
"rewards/cosine_scaled_reward": 1.171108491718769,
"rewards/format_reward": 0.0,
"step": 119
},
{
"clip_ratio": 0.0,
"completion_length": 731.9676666259766,
"epoch": 1.2230671736375158,
"grad_norm": 0.3331538140773773,
"learning_rate": 3.937173552235116e-07,
"loss": -0.0142,
"num_tokens": 97751717.0,
"reward": 1.1743304282426834,
"reward_std": 0.3502536006271839,
"rewards/cosine_scaled_reward": 1.174330398440361,
"rewards/format_reward": 0.0,
"step": 120
},
{
"clip_ratio": 0.0,
"completion_length": 722.037971496582,
"epoch": 1.2332065906210392,
"grad_norm": 0.5637516975402832,
"learning_rate": 3.850137344395598e-07,
"loss": 0.0232,
"num_tokens": 98521879.0,
"reward": 1.261700451374054,
"reward_std": 0.34792254120111465,
"rewards/cosine_scaled_reward": 1.2617004364728928,
"rewards/format_reward": 0.0,
"step": 121
},
{
"clip_ratio": 0.0,
"completion_length": 810.3984832763672,
"epoch": 1.2433460076045628,
"grad_norm": 0.12412155419588089,
"learning_rate": 3.763467497228922e-07,
"loss": -0.0175,
"num_tokens": 99386028.0,
"reward": 1.1121523007750511,
"reward_std": 0.4185631051659584,
"rewards/cosine_scaled_reward": 1.11215228587389,
"rewards/format_reward": 0.0,
"step": 122
},
{
"clip_ratio": 0.0,
"completion_length": 762.1272659301758,
"epoch": 1.2534854245880862,
"grad_norm": 0.5518940091133118,
"learning_rate": 3.677191624836106e-07,
"loss": -0.0156,
"num_tokens": 100190606.0,
"reward": 1.2098966389894485,
"reward_std": 0.42287516221404076,
"rewards/cosine_scaled_reward": 1.2098965868353844,
"rewards/format_reward": 0.0,
"step": 123
},
{
"clip_ratio": 0.0,
"completion_length": 797.1741409301758,
"epoch": 1.2636248415716096,
"grad_norm": 0.13090519607067108,
"learning_rate": 3.591337215792851e-07,
"loss": -0.0049,
"num_tokens": 101033042.0,
"reward": 1.2026162892580032,
"reward_std": 0.38727976381778717,
"rewards/cosine_scaled_reward": 1.2026162147521973,
"rewards/format_reward": 0.0,
"step": 124
},
{
"clip_ratio": 0.0,
"completion_length": 755.0368728637695,
"epoch": 1.2737642585551332,
"grad_norm": 0.13797762989997864,
"learning_rate": 3.505931624391355e-07,
"loss": 0.0129,
"num_tokens": 101827819.0,
"reward": 1.1926257461309433,
"reward_std": 0.44687049090862274,
"rewards/cosine_scaled_reward": 1.1926256865262985,
"rewards/format_reward": 0.0,
"step": 125
},
{
"clip_ratio": 0.0,
"completion_length": 741.3024978637695,
"epoch": 1.2839036755386566,
"grad_norm": 0.18938252329826355,
"learning_rate": 3.421002061924876e-07,
"loss": 0.0201,
"num_tokens": 102618890.0,
"reward": 1.263178527355194,
"reward_std": 0.3987303748726845,
"rewards/cosine_scaled_reward": 1.2631784602999687,
"rewards/format_reward": 0.0,
"step": 126
},
{
"clip_ratio": 0.0,
"completion_length": 747.1696701049805,
"epoch": 1.29404309252218,
"grad_norm": 0.13063876330852509,
"learning_rate": 3.3365755880178807e-07,
"loss": 0.0312,
"num_tokens": 103410058.0,
"reward": 1.2827118784189224,
"reward_std": 0.3673780020326376,
"rewards/cosine_scaled_reward": 1.2827118039131165,
"rewards/format_reward": 0.0,
"step": 127
},
{
"clip_ratio": 0.0,
"completion_length": 796.0893173217773,
"epoch": 1.3041825095057034,
"grad_norm": 0.12410367280244827,
"learning_rate": 3.2526791020045087e-07,
"loss": 0.0153,
"num_tokens": 104255954.0,
"reward": 1.13737141340971,
"reward_std": 0.43270402774214745,
"rewards/cosine_scaled_reward": 1.1373713836073875,
"rewards/format_reward": 0.0,
"step": 128
},
{
"clip_ratio": 0.0,
"completion_length": 709.1953353881836,
"epoch": 1.3143219264892267,
"grad_norm": 0.13791009783744812,
"learning_rate": 3.169339334358104e-07,
"loss": 0.0012,
"num_tokens": 105024737.0,
"reward": 1.2293454632163048,
"reward_std": 0.36772353760898113,
"rewards/cosine_scaled_reward": 1.2293454110622406,
"rewards/format_reward": 0.0,
"step": 129
},
{
"clip_ratio": 0.0,
"completion_length": 754.825927734375,
"epoch": 1.3244613434727504,
"grad_norm": 0.13509666919708252,
"learning_rate": 3.086582838174551e-07,
"loss": -0.0033,
"num_tokens": 105833549.0,
"reward": 1.1948751211166382,
"reward_std": 0.37703990191221237,
"rewards/cosine_scaled_reward": 1.1948750466108322,
"rewards/format_reward": 0.0,
"step": 130
},
{
"clip_ratio": 0.0,
"completion_length": 752.9855194091797,
"epoch": 1.3346007604562737,
"grad_norm": 0.13409507274627686,
"learning_rate": 3.004435980712129e-07,
"loss": -0.0299,
"num_tokens": 106631688.0,
"reward": 1.2489657402038574,
"reward_std": 0.3884260356426239,
"rewards/cosine_scaled_reward": 1.248965710401535,
"rewards/format_reward": 0.0,
"step": 131
},
{
"clip_ratio": 0.0,
"completion_length": 728.0178909301758,
"epoch": 1.3447401774397973,
"grad_norm": 0.14096473157405853,
"learning_rate": 2.922924934990568e-07,
"loss": 0.0421,
"num_tokens": 107409144.0,
"reward": 1.2679341062903404,
"reward_std": 0.4113157168030739,
"rewards/cosine_scaled_reward": 1.2679340541362762,
"rewards/format_reward": 0.0,
"step": 132
},
{
"clip_ratio": 0.0,
"completion_length": 708.2924423217773,
"epoch": 1.3548795944233207,
"grad_norm": 0.13527466356754303,
"learning_rate": 2.8420756714519954e-07,
"loss": 0.0099,
"num_tokens": 108188078.0,
"reward": 1.2067532986402512,
"reward_std": 0.4394964314997196,
"rewards/cosine_scaled_reward": 1.2067532688379288,
"rewards/format_reward": 0.0,
"step": 133
},
{
"clip_ratio": 0.0,
"completion_length": 745.264533996582,
"epoch": 1.3650190114068441,
"grad_norm": 0.12941937148571014,
"learning_rate": 2.7619139496864376e-07,
"loss": -0.0117,
"num_tokens": 108991819.0,
"reward": 1.1342971697449684,
"reward_std": 0.35589409433305264,
"rewards/cosine_scaled_reward": 1.1342971250414848,
"rewards/format_reward": 0.0,
"step": 134
},
{
"clip_ratio": 0.0,
"completion_length": 690.6328506469727,
"epoch": 1.3751584283903675,
"grad_norm": 0.1342969536781311,
"learning_rate": 2.6824653102244727e-07,
"loss": 0.0057,
"num_tokens": 109744698.0,
"reward": 1.20839624106884,
"reward_std": 0.44357743114233017,
"rewards/cosine_scaled_reward": 1.208396166563034,
"rewards/format_reward": 0.0,
"step": 135
},
{
"clip_ratio": 0.0,
"completion_length": 739.3605270385742,
"epoch": 1.385297845373891,
"grad_norm": 0.22325177490711212,
"learning_rate": 2.603755066399718e-07,
"loss": 0.0066,
"num_tokens": 110537245.0,
"reward": 1.1991283893585205,
"reward_std": 0.4422716274857521,
"rewards/cosine_scaled_reward": 1.1991283744573593,
"rewards/format_reward": 0.0,
"step": 136
},
{
"clip_ratio": 0.0,
"completion_length": 771.5045013427734,
"epoch": 1.3954372623574145,
"grad_norm": 0.131967693567276,
"learning_rate": 2.5258082962836614e-07,
"loss": -0.0609,
"num_tokens": 111363073.0,
"reward": 1.0822726637125015,
"reward_std": 0.31478939671069384,
"rewards/cosine_scaled_reward": 1.082272619009018,
"rewards/format_reward": 0.0,
"step": 137
},
{
"clip_ratio": 0.0,
"completion_length": 745.6908798217773,
"epoch": 1.4055766793409379,
"grad_norm": 0.13431350886821747,
"learning_rate": 2.4486498346955023e-07,
"loss": 0.0393,
"num_tokens": 112154044.0,
"reward": 1.2083490192890167,
"reward_std": 0.4234139323234558,
"rewards/cosine_scaled_reward": 1.2083489745855331,
"rewards/format_reward": 0.0,
"step": 138
},
{
"clip_ratio": 0.0,
"completion_length": 730.8303909301758,
"epoch": 1.4157160963244613,
"grad_norm": 0.37255144119262695,
"learning_rate": 2.372304265289436e-07,
"loss": 0.0129,
"num_tokens": 112933164.0,
"reward": 1.14224873483181,
"reward_std": 0.47245585173368454,
"rewards/cosine_scaled_reward": 1.142248660326004,
"rewards/format_reward": 0.0,
"step": 139
},
{
"clip_ratio": 0.0,
"completion_length": 723.6975708007812,
"epoch": 1.4258555133079849,
"grad_norm": 0.13437899947166443,
"learning_rate": 2.2967959127220137e-07,
"loss": 0.0149,
"num_tokens": 113703557.0,
"reward": 1.2409879192709923,
"reward_std": 0.39572571590542793,
"rewards/cosine_scaled_reward": 1.240987867116928,
"rewards/format_reward": 0.0,
"step": 140
},
{
"clip_ratio": 0.0,
"completion_length": 715.0223617553711,
"epoch": 1.4359949302915083,
"grad_norm": 0.5516566634178162,
"learning_rate": 2.2221488349019902e-07,
"loss": -0.0049,
"num_tokens": 114492577.0,
"reward": 1.1441184803843498,
"reward_std": 0.4257240351289511,
"rewards/cosine_scaled_reward": 1.1441184356808662,
"rewards/format_reward": 0.0,
"step": 141
},
{
"clip_ratio": 0.0,
"completion_length": 683.0033798217773,
"epoch": 1.4461343472750317,
"grad_norm": 0.135258749127388,
"learning_rate": 2.1483868153251788e-07,
"loss": 0.0098,
"num_tokens": 115236092.0,
"reward": 1.2539073526859283,
"reward_std": 0.42142050713300705,
"rewards/cosine_scaled_reward": 1.2539073079824448,
"rewards/format_reward": 0.0,
"step": 142
},
{
"clip_ratio": 0.0,
"completion_length": 783.0201263427734,
"epoch": 1.456273764258555,
"grad_norm": 0.17022480070590973,
"learning_rate": 2.0755333554967346e-07,
"loss": -0.0005,
"num_tokens": 116074878.0,
"reward": 1.0983483716845512,
"reward_std": 0.42650508880615234,
"rewards/cosine_scaled_reward": 1.0983483269810677,
"rewards/format_reward": 0.0,
"step": 143
},
{
"clip_ratio": 0.0,
"completion_length": 774.2790603637695,
"epoch": 1.4664131812420786,
"grad_norm": 0.1283206343650818,
"learning_rate": 2.0036116674432652e-07,
"loss": -0.0156,
"num_tokens": 116905080.0,
"reward": 1.1725050881505013,
"reward_std": 0.35209081321954727,
"rewards/cosine_scaled_reward": 1.172505035996437,
"rewards/format_reward": 0.0,
"step": 144
},
{
"clip_ratio": 0.0,
"completion_length": 658.7600746154785,
"epoch": 1.476552598225602,
"grad_norm": 0.14014765620231628,
"learning_rate": 1.9326446663172035e-07,
"loss": 0.0474,
"num_tokens": 117626809.0,
"reward": 1.341113954782486,
"reward_std": 0.4142299462109804,
"rewards/cosine_scaled_reward": 1.3411139100790024,
"rewards/format_reward": 0.0,
"step": 145
},
{
"clip_ratio": 0.0,
"completion_length": 679.3192367553711,
"epoch": 1.4866920152091254,
"grad_norm": 0.15331801772117615,
"learning_rate": 1.8626549630957395e-07,
"loss": 0.0136,
"num_tokens": 118363623.0,
"reward": 1.19945827126503,
"reward_std": 0.4244145043194294,
"rewards/cosine_scaled_reward": 1.1994582116603851,
"rewards/format_reward": 0.0,
"step": 146
},
{
"clip_ratio": 0.0,
"completion_length": 752.2890930175781,
"epoch": 1.496831432192649,
"grad_norm": 0.13101238012313843,
"learning_rate": 1.7936648573766954e-07,
"loss": -0.0056,
"num_tokens": 119174386.0,
"reward": 1.2080383449792862,
"reward_std": 0.4013860188424587,
"rewards/cosine_scaled_reward": 1.2080383002758026,
"rewards/format_reward": 0.0,
"step": 147
},
{
"clip_ratio": 0.0,
"completion_length": 762.8594131469727,
"epoch": 1.5069708491761724,
"grad_norm": 0.12902143597602844,
"learning_rate": 1.725696330273575e-07,
"loss": -0.0383,
"num_tokens": 119988252.0,
"reward": 1.169223740696907,
"reward_std": 0.39561279490590096,
"rewards/cosine_scaled_reward": 1.169223703444004,
"rewards/format_reward": 0.0,
"step": 148
},
{
"clip_ratio": 0.0,
"completion_length": 700.4531478881836,
"epoch": 1.5171102661596958,
"grad_norm": 0.1390618234872818,
"learning_rate": 1.65877103741212e-07,
"loss": 0.0519,
"num_tokens": 120740586.0,
"reward": 1.37008798122406,
"reward_std": 0.3735492154955864,
"rewards/cosine_scaled_reward": 1.3700879365205765,
"rewards/format_reward": 0.0,
"step": 149
},
{
"clip_ratio": 0.0,
"completion_length": 706.7366485595703,
"epoch": 1.5272496831432192,
"grad_norm": 0.13693000376224518,
"learning_rate": 1.592910302030544e-07,
"loss": 0.0186,
"num_tokens": 121493126.0,
"reward": 1.1771222800016403,
"reward_std": 0.4693867526948452,
"rewards/cosine_scaled_reward": 1.1771221980452538,
"rewards/format_reward": 0.0,
"step": 150
},
{
"clip_ratio": 0.0,
"completion_length": 725.279052734375,
"epoch": 1.5373891001267426,
"grad_norm": 0.16313283145427704,
"learning_rate": 1.5281351081856976e-07,
"loss": 0.0023,
"num_tokens": 122274232.0,
"reward": 1.203133448958397,
"reward_std": 0.41216023825109005,
"rewards/cosine_scaled_reward": 1.2031333893537521,
"rewards/format_reward": 0.0,
"step": 151
},
{
"clip_ratio": 0.0,
"completion_length": 703.8917770385742,
"epoch": 1.5475285171102662,
"grad_norm": 0.16185523569583893,
"learning_rate": 1.4644660940672627e-07,
"loss": -0.0126,
"num_tokens": 123030511.0,
"reward": 1.2789395153522491,
"reward_std": 0.382132101804018,
"rewards/cosine_scaled_reward": 1.2789394706487656,
"rewards/format_reward": 0.0,
"step": 152
},
{
"clip_ratio": 0.0,
"completion_length": 690.9888610839844,
"epoch": 1.5576679340937896,
"grad_norm": 0.6896323561668396,
"learning_rate": 1.4019235454221856e-07,
"loss": -0.0251,
"num_tokens": 123783709.0,
"reward": 1.230302333831787,
"reward_std": 0.3682190030813217,
"rewards/cosine_scaled_reward": 1.2303023040294647,
"rewards/format_reward": 0.0,
"step": 153
},
{
"clip_ratio": 0.0,
"completion_length": 702.8527069091797,
"epoch": 1.5678073510773132,
"grad_norm": 0.14129234850406647,
"learning_rate": 1.3405273890913737e-07,
"loss": 0.0005,
"num_tokens": 124535481.0,
"reward": 1.2051593363285065,
"reward_std": 0.4200758896768093,
"rewards/cosine_scaled_reward": 1.205159269273281,
"rewards/format_reward": 0.0,
"step": 154
},
{
"clip_ratio": 0.0,
"completion_length": 710.2265853881836,
"epoch": 1.5779467680608366,
"grad_norm": 0.1381485015153885,
"learning_rate": 1.280297186660752e-07,
"loss": 0.0181,
"num_tokens": 125290220.0,
"reward": 1.2739061638712883,
"reward_std": 0.4085957519710064,
"rewards/cosine_scaled_reward": 1.2739060744643211,
"rewards/format_reward": 0.0,
"step": 155
},
{
"clip_ratio": 0.0,
"completion_length": 808.0145492553711,
"epoch": 1.58808618504436,
"grad_norm": 0.13849686086177826,
"learning_rate": 1.2212521282287093e-07,
"loss": -0.005,
"num_tokens": 126134129.0,
"reward": 1.175368033349514,
"reward_std": 0.3809916824102402,
"rewards/cosine_scaled_reward": 1.1753679811954498,
"rewards/format_reward": 0.0,
"step": 156
},
{
"clip_ratio": 0.0,
"completion_length": 754.8638763427734,
"epoch": 1.5982256020278833,
"grad_norm": 0.14109836518764496,
"learning_rate": 1.1634110262918717e-07,
"loss": -0.0008,
"num_tokens": 126954495.0,
"reward": 1.2158635556697845,
"reward_std": 0.34099204279482365,
"rewards/cosine_scaled_reward": 1.2158635258674622,
"rewards/format_reward": 0.0,
"step": 157
},
{
"clip_ratio": 0.0,
"completion_length": 775.9598541259766,
"epoch": 1.6083650190114067,
"grad_norm": 0.12850748002529144,
"learning_rate": 1.1067923097512255e-07,
"loss": -0.0005,
"num_tokens": 127780467.0,
"reward": 1.1476298496127129,
"reward_std": 0.4081256799399853,
"rewards/cosine_scaled_reward": 1.1476298123598099,
"rewards/format_reward": 0.0,
"step": 158
},
{
"clip_ratio": 0.0,
"completion_length": 753.6797256469727,
"epoch": 1.6185044359949303,
"grad_norm": 0.36824139952659607,
"learning_rate": 1.0514140180404202e-07,
"loss": 0.0054,
"num_tokens": 128584820.0,
"reward": 1.0943407788872719,
"reward_std": 0.41071633249521255,
"rewards/cosine_scaled_reward": 1.0943407267332077,
"rewards/format_reward": 0.0,
"step": 159
},
{
"clip_ratio": 0.0,
"completion_length": 821.4553909301758,
"epoch": 1.6286438529784537,
"grad_norm": 0.15713122487068176,
"learning_rate": 9.972937953781984e-08,
"loss": 0.0185,
"num_tokens": 129467812.0,
"reward": 1.086461715400219,
"reward_std": 0.4641183950006962,
"rewards/cosine_scaled_reward": 1.0864616632461548,
"rewards/format_reward": 0.0,
"step": 160
},
{
"clip_ratio": 0.0,
"completion_length": 705.2288208007812,
"epoch": 1.6387832699619773,
"grad_norm": 0.13201545178890228,
"learning_rate": 9.444488851467041e-08,
"loss": 0.0098,
"num_tokens": 130217369.0,
"reward": 1.2702905237674713,
"reward_std": 0.3846561126410961,
"rewards/cosine_scaled_reward": 1.2702905088663101,
"rewards/format_reward": 0.0,
"step": 161
},
{
"clip_ratio": 0.0,
"completion_length": 732.4542770385742,
"epoch": 1.6489226869455007,
"grad_norm": 0.12551482021808624,
"learning_rate": 8.928961243975436e-08,
"loss": 0.0025,
"num_tokens": 131006504.0,
"reward": 1.1392531916499138,
"reward_std": 0.4202544465661049,
"rewards/cosine_scaled_reward": 1.150568701326847,
"rewards/format_reward": 0.0,
"step": 162
},
{
"clip_ratio": 0.0,
"completion_length": 690.4821739196777,
"epoch": 1.659062103929024,
"grad_norm": 0.14846356213092804,
"learning_rate": 8.426519384872732e-08,
"loss": 0.0151,
"num_tokens": 131747312.0,
"reward": 1.2898453325033188,
"reward_std": 0.3518759645521641,
"rewards/cosine_scaled_reward": 1.2898452877998352,
"rewards/format_reward": 0.0,
"step": 163
},
{
"clip_ratio": 0.0,
"completion_length": 749.0937881469727,
"epoch": 1.6692015209125475,
"grad_norm": 0.13933837413787842,
"learning_rate": 7.937323358440934e-08,
"loss": 0.028,
"num_tokens": 132549532.0,
"reward": 1.1977392584085464,
"reward_std": 0.43290739692747593,
"rewards/cosine_scaled_reward": 1.1977391764521599,
"rewards/format_reward": 0.0,
"step": 164
},
{
"clip_ratio": 0.0,
"completion_length": 750.6361999511719,
"epoch": 1.6793409378960709,
"grad_norm": 0.1263691633939743,
"learning_rate": 7.461529028673463e-08,
"loss": 0.0414,
"num_tokens": 133354910.0,
"reward": 1.2933627367019653,
"reward_std": 0.41812827065587044,
"rewards/cosine_scaled_reward": 1.2933626621961594,
"rewards/format_reward": 0.0,
"step": 165
},
{
"clip_ratio": 0.0,
"completion_length": 748.4531555175781,
"epoch": 1.6894803548795945,
"grad_norm": 0.1337745487689972,
"learning_rate": 6.999287989614971e-08,
"loss": 0.0003,
"num_tokens": 134157964.0,
"reward": 1.1261718794703484,
"reward_std": 0.43879349157214165,
"rewards/cosine_scaled_reward": 1.1261718571186066,
"rewards/format_reward": 0.0,
"step": 166
},
{
"clip_ratio": 0.0,
"completion_length": 684.155158996582,
"epoch": 1.6996197718631179,
"grad_norm": 0.14675858616828918,
"learning_rate": 6.550747517061656e-08,
"loss": 0.0135,
"num_tokens": 134897511.0,
"reward": 1.31067955493927,
"reward_std": 0.4250149428844452,
"rewards/cosine_scaled_reward": 1.3106794953346252,
"rewards/format_reward": 0.0,
"step": 167
},
{
"clip_ratio": 0.0,
"completion_length": 708.9955825805664,
"epoch": 1.7097591888466415,
"grad_norm": 0.1361108422279358,
"learning_rate": 6.116050521637218e-08,
"loss": 0.0403,
"num_tokens": 135661099.0,
"reward": 1.2582149505615234,
"reward_std": 0.4354391284286976,
"rewards/cosine_scaled_reward": 1.2582148760557175,
"rewards/format_reward": 0.0,
"step": 168
},
{
"clip_ratio": 0.0,
"completion_length": 708.2154312133789,
"epoch": 1.7198986058301649,
"grad_norm": 0.13875119388103485,
"learning_rate": 5.6953355032598795e-08,
"loss": -0.0346,
"num_tokens": 136432732.0,
"reward": 1.1267788708209991,
"reward_std": 0.43002479895949364,
"rewards/cosine_scaled_reward": 1.126778818666935,
"rewards/format_reward": 0.0,
"step": 169
},
{
"clip_ratio": 0.0,
"completion_length": 729.7221374511719,
"epoch": 1.7300380228136882,
"grad_norm": 0.136752188205719,
"learning_rate": 5.288736507014435e-08,
"loss": -0.0081,
"num_tokens": 137210563.0,
"reward": 1.251162275671959,
"reward_std": 0.36924222111701965,
"rewards/cosine_scaled_reward": 1.2511622309684753,
"rewards/format_reward": 0.0,
"step": 170
},
{
"clip_ratio": 0.0,
"completion_length": 707.0078430175781,
"epoch": 1.7401774397972116,
"grad_norm": 0.132461279630661,
"learning_rate": 4.896383080443933e-08,
"loss": 0.0231,
"num_tokens": 137968794.0,
"reward": 1.2026441097259521,
"reward_std": 0.42860128730535507,
"rewards/cosine_scaled_reward": 1.202644057571888,
"rewards/format_reward": 0.0,
"step": 171
},
{
"clip_ratio": 0.0,
"completion_length": 690.8460159301758,
"epoch": 1.750316856780735,
"grad_norm": 0.14912405610084534,
"learning_rate": 4.518400232274078e-08,
"loss": 0.0096,
"num_tokens": 138718296.0,
"reward": 1.188114494085312,
"reward_std": 0.39624364115297794,
"rewards/cosine_scaled_reward": 1.2003108784556389,
"rewards/format_reward": 0.0,
"step": 172
},
{
"clip_ratio": 0.0,
"completion_length": 684.9765930175781,
"epoch": 1.7604562737642584,
"grad_norm": 0.14246663451194763,
"learning_rate": 4.1549083925840165e-08,
"loss": 0.0125,
"num_tokens": 139449483.0,
"reward": 1.2953269332647324,
"reward_std": 0.4603438973426819,
"rewards/cosine_scaled_reward": 1.29532690346241,
"rewards/format_reward": 0.0,
"step": 173
},
{
"clip_ratio": 0.0,
"completion_length": 685.9721298217773,
"epoch": 1.770595690747782,
"grad_norm": 0.1809549778699875,
"learning_rate": 3.806023374435663e-08,
"loss": -0.0066,
"num_tokens": 140192642.0,
"reward": 1.18511962890625,
"reward_std": 0.365385384298861,
"rewards/cosine_scaled_reward": 1.1851196065545082,
"rewards/format_reward": 0.0,
"step": 174
},
{
"clip_ratio": 0.0,
"completion_length": 718.2578353881836,
"epoch": 1.7807351077313056,
"grad_norm": 2.9148614406585693,
"learning_rate": 3.4718563369743213e-08,
"loss": 0.0077,
"num_tokens": 140969529.0,
"reward": 1.1897177621722221,
"reward_std": 0.44410305470228195,
"rewards/cosine_scaled_reward": 1.1897177323698997,
"rewards/format_reward": 0.0,
"step": 175
},
{
"clip_ratio": 0.0,
"completion_length": 672.2578392028809,
"epoch": 1.790874524714829,
"grad_norm": 0.14049698412418365,
"learning_rate": 3.15251375001192e-08,
"loss": 0.0167,
"num_tokens": 141711976.0,
"reward": 1.2624182030558586,
"reward_std": 0.45537084713578224,
"rewards/cosine_scaled_reward": 1.2624181509017944,
"rewards/format_reward": 0.0,
"step": 176
},
{
"clip_ratio": 0.0,
"completion_length": 697.0145378112793,
"epoch": 1.8010139416983524,
"grad_norm": 0.22702006995677948,
"learning_rate": 2.8480973601043955e-08,
"loss": -0.01,
"num_tokens": 142471309.0,
"reward": 1.249974675476551,
"reward_std": 0.3576275184750557,
"rewards/cosine_scaled_reward": 1.2499746307730675,
"rewards/format_reward": 0.0,
"step": 177
},
{
"clip_ratio": 0.0,
"completion_length": 744.4810409545898,
"epoch": 1.8111533586818758,
"grad_norm": 0.12813909351825714,
"learning_rate": 2.558704158134023e-08,
"loss": -0.0078,
"num_tokens": 143278388.0,
"reward": 1.1319350376725197,
"reward_std": 0.41413314267992973,
"rewards/cosine_scaled_reward": 1.1319349706172943,
"rewards/format_reward": 0.0,
"step": 178
},
{
"clip_ratio": 0.0,
"completion_length": 750.3560638427734,
"epoch": 1.8212927756653992,
"grad_norm": 0.13594254851341248,
"learning_rate": 2.2844263484068093e-08,
"loss": 0.0003,
"num_tokens": 144077707.0,
"reward": 1.2112557888031006,
"reward_std": 0.3645152598619461,
"rewards/cosine_scaled_reward": 1.2112557291984558,
"rewards/format_reward": 0.0,
"step": 179
},
{
"clip_ratio": 0.0,
"completion_length": 682.1417694091797,
"epoch": 1.8314321926489225,
"grad_norm": 0.15148992836475372,
"learning_rate": 2.025351319275137e-08,
"loss": -0.0027,
"num_tokens": 144819874.0,
"reward": 1.2360253632068634,
"reward_std": 0.43914780393242836,
"rewards/cosine_scaled_reward": 1.2360252812504768,
"rewards/format_reward": 0.0,
"step": 180
},
{
"clip_ratio": 0.0,
"completion_length": 724.106071472168,
"epoch": 1.8415716096324461,
"grad_norm": 0.14325113594532013,
"learning_rate": 1.781561615294652e-08,
"loss": 0.0052,
"num_tokens": 145603137.0,
"reward": 1.226213201880455,
"reward_std": 0.4393893778324127,
"rewards/cosine_scaled_reward": 1.2262131571769714,
"rewards/format_reward": 0.0,
"step": 181
},
{
"clip_ratio": 0.0,
"completion_length": 782.7399749755859,
"epoch": 1.8517110266159695,
"grad_norm": 0.1272156685590744,
"learning_rate": 1.553134910924636e-08,
"loss": -0.0251,
"num_tokens": 146440576.0,
"reward": 1.0694795548915863,
"reward_std": 0.3888841047883034,
"rewards/cosine_scaled_reward": 1.0694795101881027,
"rewards/format_reward": 0.0,
"step": 182
},
{
"clip_ratio": 0.0,
"completion_length": 677.8895492553711,
"epoch": 1.8618504435994931,
"grad_norm": 0.14028111100196838,
"learning_rate": 1.340143985779829e-08,
"loss": 0.0087,
"num_tokens": 147174725.0,
"reward": 1.2450221478939056,
"reward_std": 0.3928499836474657,
"rewards/cosine_scaled_reward": 1.2450221329927444,
"rewards/format_reward": 0.0,
"step": 183
},
{
"clip_ratio": 0.0,
"completion_length": 698.4096298217773,
"epoch": 1.8719898605830165,
"grad_norm": 0.1386420577764511,
"learning_rate": 1.1426567014420297e-08,
"loss": -0.001,
"num_tokens": 147927204.0,
"reward": 1.2862959504127502,
"reward_std": 0.39938454143702984,
"rewards/cosine_scaled_reward": 1.2862958759069443,
"rewards/format_reward": 0.0,
"step": 184
},
{
"clip_ratio": 0.0,
"completion_length": 749.4877471923828,
"epoch": 1.88212927756654,
"grad_norm": 0.26094314455986023,
"learning_rate": 9.607359798384784e-09,
"loss": 0.0099,
"num_tokens": 148714225.0,
"reward": 1.15062565356493,
"reward_std": 0.49165763705968857,
"rewards/cosine_scaled_reward": 1.150625616312027,
"rewards/format_reward": 0.0,
"step": 185
},
{
"clip_ratio": 0.0,
"completion_length": 701.1071624755859,
"epoch": 1.8922686945500633,
"grad_norm": 0.14008425176143646,
"learning_rate": 7.944397831941951e-09,
"loss": 0.0034,
"num_tokens": 149488913.0,
"reward": 1.181966707110405,
"reward_std": 0.4252682514488697,
"rewards/cosine_scaled_reward": 1.181966632604599,
"rewards/format_reward": 0.0,
"step": 186
},
{
"clip_ratio": 0.0,
"completion_length": 779.9766082763672,
"epoch": 1.9024081115335867,
"grad_norm": 0.16207562386989594,
"learning_rate": 6.438210955644452e-09,
"loss": -0.0089,
"num_tokens": 150330196.0,
"reward": 1.0697861537337303,
"reward_std": 0.3918624483048916,
"rewards/cosine_scaled_reward": 1.0697861164808273,
"rewards/format_reward": 0.0,
"step": 187
},
{
"clip_ratio": 0.0,
"completion_length": 725.6529235839844,
"epoch": 1.9125475285171103,
"grad_norm": 0.1371779590845108,
"learning_rate": 5.0892790595336575e-09,
"loss": 0.0236,
"num_tokens": 151117397.0,
"reward": 1.2266732975840569,
"reward_std": 0.4197930619120598,
"rewards/cosine_scaled_reward": 1.226673237979412,
"rewards/format_reward": 0.0,
"step": 188
},
{
"clip_ratio": 0.0,
"completion_length": 735.1306228637695,
"epoch": 1.9226869455006337,
"grad_norm": 0.14734813570976257,
"learning_rate": 3.898031930240797e-09,
"loss": -0.013,
"num_tokens": 151901074.0,
"reward": 1.1564877182245255,
"reward_std": 0.4156847130507231,
"rewards/cosine_scaled_reward": 1.1564876809716225,
"rewards/format_reward": 0.0,
"step": 189
},
{
"clip_ratio": 0.0,
"completion_length": 774.4129943847656,
"epoch": 1.9328263624841573,
"grad_norm": 0.12307488918304443,
"learning_rate": 2.8648491140513264e-09,
"loss": -0.0197,
"num_tokens": 152721396.0,
"reward": 1.2314427196979523,
"reward_std": 0.39738379418849945,
"rewards/cosine_scaled_reward": 1.2314426600933075,
"rewards/format_reward": 0.0,
"step": 190
},
{
"clip_ratio": 0.0,
"completion_length": 639.6127433776855,
"epoch": 1.9429657794676807,
"grad_norm": 0.15224085748195648,
"learning_rate": 1.9900597959770505e-09,
"loss": 0.0385,
"num_tokens": 153419617.0,
"reward": 1.3687696307897568,
"reward_std": 0.3240409158170223,
"rewards/cosine_scaled_reward": 1.3687695413827896,
"rewards/format_reward": 0.0,
"step": 191
},
{
"clip_ratio": 0.0,
"completion_length": 727.8538208007812,
"epoch": 1.953105196451204,
"grad_norm": 0.12927082180976868,
"learning_rate": 1.2739426948732424e-09,
"loss": 0.002,
"num_tokens": 154205742.0,
"reward": 1.2560684382915497,
"reward_std": 0.3877686746418476,
"rewards/cosine_scaled_reward": 1.256068378686905,
"rewards/format_reward": 0.0,
"step": 192
},
{
"clip_ratio": 0.0,
"completion_length": 720.2232513427734,
"epoch": 1.9632446134347274,
"grad_norm": 0.5261313915252686,
"learning_rate": 7.16725974635568e-10,
"loss": 0.0203,
"num_tokens": 154976902.0,
"reward": 1.1852427944540977,
"reward_std": 0.4709174670279026,
"rewards/cosine_scaled_reward": 1.1852427273988724,
"rewards/format_reward": 0.0,
"step": 193
},
{
"clip_ratio": 0.0,
"completion_length": 744.1283721923828,
"epoch": 1.9733840304182508,
"grad_norm": 0.13184396922588348,
"learning_rate": 3.185871715041255e-10,
"loss": -0.0006,
"num_tokens": 155776489.0,
"reward": 1.153962902724743,
"reward_std": 0.35575347393751144,
"rewards/cosine_scaled_reward": 1.1539628580212593,
"rewards/format_reward": 0.0,
"step": 194
},
{
"clip_ratio": 0.0,
"completion_length": 681.7768096923828,
"epoch": 1.9835234474017744,
"grad_norm": 0.14284567534923553,
"learning_rate": 7.96531374983589e-11,
"loss": 0.0074,
"num_tokens": 156507841.0,
"reward": 1.263557255268097,
"reward_std": 0.42133441381156445,
"rewards/cosine_scaled_reward": 1.263557218015194,
"rewards/format_reward": 0.0,
"step": 195
},
{
"clip_ratio": 0.0,
"completion_length": 712.19921875,
"epoch": 1.9936628643852978,
"grad_norm": 0.12404794245958328,
"learning_rate": 0.0,
"loss": -0.0205,
"num_tokens": 157287049.0,
"reward": 1.1777000427246094,
"reward_std": 0.4066210947930813,
"rewards/cosine_scaled_reward": 1.1776999980211258,
"rewards/format_reward": 0.0,
"step": 196
},
{
"epoch": 1.9936628643852978,
"step": 196,
"total_flos": 0.0,
"train_loss": -0.009652293185873921,
"train_runtime": 33755.1654,
"train_samples_per_second": 0.654,
"train_steps_per_second": 0.006
}
],
"logging_steps": 1,
"max_steps": 196,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 500,
"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": 16,
"trial_name": null,
"trial_params": null
}