{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 14.597084548104956,
  "eval_steps": 500,
  "global_step": 960,
  "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": 0.014334542410714302,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4089.0,
      "completions/mean_length": 600.2800903320312,
      "completions/mean_terminated_length": 549.4417724609375,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.009329446064139942,
      "grad_norm": 0.15390798449516296,
      "learning_rate": 1e-06,
      "loss": 0.0326,
      "num_tokens": 36741188.0,
      "reward": 0.4951869547367096,
      "reward_std": 0.26015907526016235,
      "rewards/simpleverify_reward/mean": 0.4951869547367096,
      "rewards/simpleverify_reward/std": 0.49998119473457336,
      "step": 1
    },
    {
      "clip_ratio/high_max": 0.002040916406258475,
      "clip_ratio/high_mean": 0.0008884637609298807,
      "clip_ratio/low_mean": 0.000776253073127009,
      "clip_ratio/low_min": 9.176790354104014e-05,
      "clip_ratio/region_mean": 0.001664716844970826,
      "epoch": 0.018658892128279883,
      "grad_norm": 0.16035835444927216,
      "learning_rate": 1e-06,
      "loss": 0.0402,
      "step": 2
    },
    {
      "clip_ratio/high_max": 0.0026521540712565184,
      "clip_ratio/high_mean": 0.0011930029759241734,
      "clip_ratio/low_mean": 0.0007639474333700491,
      "clip_ratio/low_min": 5.708100434276275e-05,
      "clip_ratio/region_mean": 0.0019569503856473602,
      "epoch": 0.027988338192419825,
      "grad_norm": 0.13767586648464203,
      "learning_rate": 1e-06,
      "loss": -0.0356,
      "step": 3
    },
    {
      "clip_ratio/high_max": 0.0025948614711523987,
      "clip_ratio/high_mean": 0.0011277109006186947,
      "clip_ratio/low_mean": 0.0007378015725407749,
      "clip_ratio/low_min": 0.00012079647240170743,
      "clip_ratio/region_mean": 0.0018655124586075544,
      "epoch": 0.037317784256559766,
      "grad_norm": 0.12275255471467972,
      "learning_rate": 1e-06,
      "loss": -0.0086,
      "step": 4
    },
    {
      "clip_ratio/high_max": 0.002557778687332757,
      "clip_ratio/high_mean": 0.001121524917834904,
      "clip_ratio/low_mean": 0.0008878749722498469,
      "clip_ratio/low_min": 0.00013076019331492716,
      "clip_ratio/region_mean": 0.002009399853704963,
      "epoch": 0.04664723032069971,
      "grad_norm": 0.1217966079711914,
      "learning_rate": 1e-06,
      "loss": 0.0253,
      "step": 5
    },
    {
      "clip_ratio/high_max": 0.0031994826713344082,
      "clip_ratio/high_mean": 0.0015321520622819662,
      "clip_ratio/low_mean": 0.0011226907390664564,
      "clip_ratio/low_min": 8.586031071899924e-05,
      "clip_ratio/region_mean": 0.002654842821357306,
      "epoch": 0.05597667638483965,
      "grad_norm": 0.1310003250837326,
      "learning_rate": 1e-06,
      "loss": -0.0679,
      "step": 6
    },
    {
      "clip_ratio/high_max": 0.00303863592853304,
      "clip_ratio/high_mean": 0.001447087044653017,
      "clip_ratio/low_mean": 0.001268966148927575,
      "clip_ratio/low_min": 0.0001895853083624388,
      "clip_ratio/region_mean": 0.0027160532044945285,
      "epoch": 0.0653061224489796,
      "grad_norm": 0.13644307851791382,
      "learning_rate": 1e-06,
      "loss": -0.0298,
      "step": 7
    },
    {
      "clip_ratio/high_max": 0.002758739014097955,
      "clip_ratio/high_mean": 0.001347529236227274,
      "clip_ratio/low_mean": 0.0014581874274881557,
      "clip_ratio/low_min": 0.00022467970302386675,
      "clip_ratio/region_mean": 0.0028057167073711753,
      "epoch": 0.07463556851311953,
      "grad_norm": 0.14762556552886963,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 8
    },
    {
      "clip_ratio/high_max": 0.0028129097700002603,
      "clip_ratio/high_mean": 0.0013104464815114625,
      "clip_ratio/low_mean": 0.0014768756373086944,
      "clip_ratio/low_min": 0.00033178317517013056,
      "clip_ratio/region_mean": 0.002787322082440369,
      "epoch": 0.08396501457725948,
      "grad_norm": 0.13718098402023315,
      "learning_rate": 1e-06,
      "loss": 0.0285,
      "step": 9
    },
    {
      "clip_ratio/high_max": 0.002485598270141054,
      "clip_ratio/high_mean": 0.001291550697715138,
      "clip_ratio/low_mean": 0.0014255444948503282,
      "clip_ratio/low_min": 0.00017188102083309786,
      "clip_ratio/region_mean": 0.002717095223488286,
      "epoch": 0.09329446064139942,
      "grad_norm": 0.12989400327205658,
      "learning_rate": 1e-06,
      "loss": -0.0213,
      "step": 10
    },
    {
      "clip_ratio/high_max": 0.0028909092070534825,
      "clip_ratio/high_mean": 0.001177898364403518,
      "clip_ratio/low_mean": 0.0013606542852357961,
      "clip_ratio/low_min": 0.00022655712837149622,
      "clip_ratio/region_mean": 0.002538552609621547,
      "epoch": 0.10262390670553936,
      "grad_norm": 0.13184423744678497,
      "learning_rate": 1e-06,
      "loss": -0.0268,
      "step": 11
    },
    {
      "clip_ratio/high_max": 0.002287557836098131,
      "clip_ratio/high_mean": 0.0010311130045010941,
      "clip_ratio/low_mean": 0.0012395995436236262,
      "clip_ratio/low_min": 0.00029447417091432726,
      "clip_ratio/region_mean": 0.002270712524477858,
      "epoch": 0.1119533527696793,
      "grad_norm": 0.12579387426376343,
      "learning_rate": 1e-06,
      "loss": 0.0185,
      "step": 12
    },
    {
      "clip_ratio/high_max": 0.0024642881980980746,
      "clip_ratio/high_mean": 0.00115407105840859,
      "clip_ratio/low_mean": 0.0013097346891299821,
      "clip_ratio/low_min": 0.00022087641173129668,
      "clip_ratio/region_mean": 0.0024638057366246358,
      "epoch": 0.12128279883381925,
      "grad_norm": 0.13448752462863922,
      "learning_rate": 1e-06,
      "loss": 0.0292,
      "step": 13
    },
    {
      "clip_ratio/high_max": 0.0028096670503146015,
      "clip_ratio/high_mean": 0.0012107493930670898,
      "clip_ratio/low_mean": 0.001211751383380033,
      "clip_ratio/low_min": 0.00017167441728815902,
      "clip_ratio/region_mean": 0.0024225007655331865,
      "epoch": 0.1306122448979592,
      "grad_norm": 0.12716948986053467,
      "learning_rate": 1e-06,
      "loss": 0.0234,
      "step": 14
    },
    {
      "clip_ratio/high_max": 0.002425587757898029,
      "clip_ratio/high_mean": 0.00104737462970661,
      "clip_ratio/low_mean": 0.0009895519833662547,
      "clip_ratio/low_min": 8.760188029555138e-05,
      "clip_ratio/region_mean": 0.0020369266494526528,
      "epoch": 0.13994169096209913,
      "grad_norm": 0.12380632013082504,
      "learning_rate": 1e-06,
      "loss": 0.0069,
      "step": 15
    },
    {
      "clip_ratio/high_max": 0.0025905574584612623,
      "clip_ratio/high_mean": 0.0012410984381858725,
      "clip_ratio/low_mean": 0.0010446769429108826,
      "clip_ratio/low_min": 0.0001288526018470293,
      "clip_ratio/region_mean": 0.0022857753647258505,
      "epoch": 0.14927113702623906,
      "grad_norm": 0.13263782858848572,
      "learning_rate": 1e-06,
      "loss": 0.0004,
      "step": 16
    },
    {
      "clip_ratio/high_max": 0.002322110165550839,
      "clip_ratio/high_mean": 0.0010622955705912318,
      "clip_ratio/low_mean": 0.0009947660964826355,
      "clip_ratio/low_min": 0.00019454384892014787,
      "clip_ratio/region_mean": 0.0020570616325130686,
      "epoch": 0.158600583090379,
      "grad_norm": 0.13527724146842957,
      "learning_rate": 1e-06,
      "loss": 0.0222,
      "step": 17
    },
    {
      "clip_ratio/high_max": 0.0030840404360787943,
      "clip_ratio/high_mean": 0.001330941988271661,
      "clip_ratio/low_mean": 0.001035875771776773,
      "clip_ratio/low_min": 0.0001651506663620239,
      "clip_ratio/region_mean": 0.0023668177746003494,
      "epoch": 0.16793002915451896,
      "grad_norm": 0.14462633430957794,
      "learning_rate": 1e-06,
      "loss": -0.0272,
      "step": 18
    },
    {
      "clip_ratio/high_max": 0.0024542947721784003,
      "clip_ratio/high_mean": 0.0012647232906601857,
      "clip_ratio/low_mean": 0.0010453857139509637,
      "clip_ratio/low_min": 0.00011290672682662262,
      "clip_ratio/region_mean": 0.0023101089682313614,
      "epoch": 0.1772594752186589,
      "grad_norm": 0.12958800792694092,
      "learning_rate": 1e-06,
      "loss": -0.0264,
      "step": 19
    },
    {
      "clip_ratio/high_max": 0.002545717667089775,
      "clip_ratio/high_mean": 0.0010949806928692851,
      "clip_ratio/low_mean": 0.0010164689556404483,
      "clip_ratio/low_min": 9.848788249655627e-05,
      "clip_ratio/region_mean": 0.0021114496994414367,
      "epoch": 0.18658892128279883,
      "grad_norm": 0.11860532313585281,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 20
    },
    {
      "clip_ratio/high_max": 0.0023502288095187396,
      "clip_ratio/high_mean": 0.0010851689585251734,
      "clip_ratio/low_mean": 0.0010610801327857189,
      "clip_ratio/low_min": 0.00015702244400017662,
      "clip_ratio/region_mean": 0.002146249076758977,
      "epoch": 0.19591836734693877,
      "grad_norm": 0.12849994003772736,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 21
    },
    {
      "clip_ratio/high_max": 0.0024970949816633947,
      "clip_ratio/high_mean": 0.0011011122587660793,
      "clip_ratio/low_mean": 0.0010701754727051593,
      "clip_ratio/low_min": 0.00012273165157239418,
      "clip_ratio/region_mean": 0.0021712877860409208,
      "epoch": 0.20524781341107873,
      "grad_norm": 0.12600012123584747,
      "learning_rate": 1e-06,
      "loss": -0.0336,
      "step": 22
    },
    {
      "clip_ratio/high_max": 0.0028476907100412063,
      "clip_ratio/high_mean": 0.001330400391452713,
      "clip_ratio/low_mean": 0.0010206206570728682,
      "clip_ratio/low_min": 0.00012209860324219335,
      "clip_ratio/region_mean": 0.0023510210594395176,
      "epoch": 0.21457725947521866,
      "grad_norm": 0.12997882068157196,
      "learning_rate": 1e-06,
      "loss": -0.039,
      "step": 23
    },
    {
      "clip_ratio/high_max": 0.002843989888788201,
      "clip_ratio/high_mean": 0.0012584266005433165,
      "clip_ratio/low_mean": 0.001166309833934065,
      "clip_ratio/low_min": 0.00014769538393011317,
      "clip_ratio/region_mean": 0.002424736405373551,
      "epoch": 0.2239067055393586,
      "grad_norm": 0.1239733025431633,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 24
    },
    {
      "clip_ratio/high_max": 0.0025600656226743013,
      "clip_ratio/high_mean": 0.0012093647965230048,
      "clip_ratio/low_mean": 0.0012888945675513241,
      "clip_ratio/low_min": 0.00022058806007407838,
      "clip_ratio/region_mean": 0.002498259309504647,
      "epoch": 0.23323615160349853,
      "grad_norm": 0.12248584628105164,
      "learning_rate": 1e-06,
      "loss": 0.0448,
      "step": 25
    },
    {
      "clip_ratio/high_max": 0.003117875530733727,
      "clip_ratio/high_mean": 0.0014248310253606178,
      "clip_ratio/low_mean": 0.0012216954382893164,
      "clip_ratio/low_min": 0.0002501699418644421,
      "clip_ratio/region_mean": 0.0026465264527359977,
      "epoch": 0.2425655976676385,
      "grad_norm": 0.14297905564308167,
      "learning_rate": 1e-06,
      "loss": -0.0336,
      "step": 26
    },
    {
      "clip_ratio/high_max": 0.0024331261156476103,
      "clip_ratio/high_mean": 0.0011360391254129354,
      "clip_ratio/low_mean": 0.0012286992132430896,
      "clip_ratio/low_min": 0.00022854195776744746,
      "clip_ratio/region_mean": 0.0023647383277420886,
      "epoch": 0.2518950437317784,
      "grad_norm": 0.12130679935216904,
      "learning_rate": 1e-06,
      "loss": 0.0061,
      "step": 27
    },
    {
      "clip_ratio/high_max": 0.002945994412584696,
      "clip_ratio/high_mean": 0.0012949165575264487,
      "clip_ratio/low_mean": 0.001161579024483217,
      "clip_ratio/low_min": 0.00013133108677720884,
      "clip_ratio/region_mean": 0.002456495611113496,
      "epoch": 0.2612244897959184,
      "grad_norm": 0.12663106620311737,
      "learning_rate": 1e-06,
      "loss": 0.013,
      "step": 28
    },
    {
      "clip_ratio/high_max": 0.002999761069077067,
      "clip_ratio/high_mean": 0.0014481958205578849,
      "clip_ratio/low_mean": 0.001349988884612685,
      "clip_ratio/low_min": 0.00022052016811358044,
      "clip_ratio/region_mean": 0.0027981847088085487,
      "epoch": 0.2705539358600583,
      "grad_norm": 0.12327197194099426,
      "learning_rate": 1e-06,
      "loss": -0.0069,
      "step": 29
    },
    {
      "clip_ratio/high_max": 0.0026856460244744085,
      "clip_ratio/high_mean": 0.001247929823875893,
      "clip_ratio/low_mean": 0.0012160168189438991,
      "clip_ratio/low_min": 0.00017969683722185437,
      "clip_ratio/region_mean": 0.002463946628267877,
      "epoch": 0.27988338192419826,
      "grad_norm": 0.11699430644512177,
      "learning_rate": 1e-06,
      "loss": -0.0276,
      "step": 30
    },
    {
      "clip_ratio/high_max": 0.002770919840259012,
      "clip_ratio/high_mean": 0.001350192105746828,
      "clip_ratio/low_mean": 0.001140667860454414,
      "clip_ratio/low_min": 0.00015164834076131228,
      "clip_ratio/region_mean": 0.002490859966201242,
      "epoch": 0.2892128279883382,
      "grad_norm": 0.11804838478565216,
      "learning_rate": 1e-06,
      "loss": -0.0527,
      "step": 31
    },
    {
      "clip_ratio/high_max": 0.0028033702255925164,
      "clip_ratio/high_mean": 0.001267817300686147,
      "clip_ratio/low_mean": 0.0012801564880646765,
      "clip_ratio/low_min": 0.00023108240202418528,
      "clip_ratio/region_mean": 0.0025479737741989084,
      "epoch": 0.29854227405247813,
      "grad_norm": 0.12966667115688324,
      "learning_rate": 1e-06,
      "loss": -0.0081,
      "step": 32
    },
    {
      "clip_ratio/high_max": 0.002397582560661249,
      "clip_ratio/high_mean": 0.0011626613413682207,
      "clip_ratio/low_mean": 0.0013735218126385007,
      "clip_ratio/low_min": 0.00026043377147288993,
      "clip_ratio/region_mean": 0.0025361831576447003,
      "epoch": 0.30787172011661806,
      "grad_norm": 0.1260962039232254,
      "learning_rate": 1e-06,
      "loss": 0.0082,
      "step": 33
    },
    {
      "clip_ratio/high_max": 0.002426173727144487,
      "clip_ratio/high_mean": 0.0012151251939940266,
      "clip_ratio/low_mean": 0.001366448705084622,
      "clip_ratio/low_min": 0.00022134304435894592,
      "clip_ratio/region_mean": 0.0025815738626988605,
      "epoch": 0.317201166180758,
      "grad_norm": 0.1335056573152542,
      "learning_rate": 1e-06,
      "loss": 0.0099,
      "step": 34
    },
    {
      "clip_ratio/high_max": 0.002810610079905018,
      "clip_ratio/high_mean": 0.0012503327416197862,
      "clip_ratio/low_mean": 0.0013840347965015098,
      "clip_ratio/low_min": 0.00017831472086982103,
      "clip_ratio/region_mean": 0.0026343676072428934,
      "epoch": 0.32653061224489793,
      "grad_norm": 0.1279270052909851,
      "learning_rate": 1e-06,
      "loss": 0.0059,
      "step": 35
    },
    {
      "clip_ratio/high_max": 0.0025273438441217877,
      "clip_ratio/high_mean": 0.001138790245022392,
      "clip_ratio/low_mean": 0.001418263065716019,
      "clip_ratio/low_min": 0.00021497471152542857,
      "clip_ratio/region_mean": 0.0025570532961864956,
      "epoch": 0.3358600583090379,
      "grad_norm": 0.13317278027534485,
      "learning_rate": 1e-06,
      "loss": 0.0212,
      "step": 36
    },
    {
      "clip_ratio/high_max": 0.0025164432227029465,
      "clip_ratio/high_mean": 0.0011276436598564032,
      "clip_ratio/low_mean": 0.0013073582776996773,
      "clip_ratio/low_min": 0.00022627342241321458,
      "clip_ratio/region_mean": 0.0024350019739358686,
      "epoch": 0.34518950437317786,
      "grad_norm": 0.13060995936393738,
      "learning_rate": 1e-06,
      "loss": 0.0039,
      "step": 37
    },
    {
      "clip_ratio/high_max": 0.0028637529103434645,
      "clip_ratio/high_mean": 0.00132327246319619,
      "clip_ratio/low_mean": 0.001336863530013943,
      "clip_ratio/low_min": 0.00024114352345350198,
      "clip_ratio/region_mean": 0.002660136087797582,
      "epoch": 0.3545189504373178,
      "grad_norm": 0.1329226940870285,
      "learning_rate": 1e-06,
      "loss": 0.0076,
      "step": 38
    },
    {
      "clip_ratio/high_max": 0.0025575220279279165,
      "clip_ratio/high_mean": 0.0012403778309817426,
      "clip_ratio/low_mean": 0.001399258468154585,
      "clip_ratio/low_min": 0.0002571259856267716,
      "clip_ratio/region_mean": 0.0026396363173262216,
      "epoch": 0.3638483965014577,
      "grad_norm": 0.12120100110769272,
      "learning_rate": 1e-06,
      "loss": 0.0232,
      "step": 39
    },
    {
      "clip_ratio/high_max": 0.0024824743013596162,
      "clip_ratio/high_mean": 0.0011955045811191667,
      "clip_ratio/low_mean": 0.0012919128421344794,
      "clip_ratio/low_min": 0.00021298496631061425,
      "clip_ratio/region_mean": 0.0024874173614080064,
      "epoch": 0.37317784256559766,
      "grad_norm": 0.122945636510849,
      "learning_rate": 1e-06,
      "loss": -0.0113,
      "step": 40
    },
    {
      "clip_ratio/high_max": 0.002869045907573309,
      "clip_ratio/high_mean": 0.001285800724872388,
      "clip_ratio/low_mean": 0.0015269205105141737,
      "clip_ratio/low_min": 0.00028516309794213157,
      "clip_ratio/region_mean": 0.002812721228110604,
      "epoch": 0.3825072886297376,
      "grad_norm": 0.12388288229703903,
      "learning_rate": 1e-06,
      "loss": 0.0419,
      "step": 41
    },
    {
      "clip_ratio/high_max": 0.0029178798286011443,
      "clip_ratio/high_mean": 0.0012012958577543031,
      "clip_ratio/low_mean": 0.0012740831043629441,
      "clip_ratio/low_min": 5.371477254811907e-05,
      "clip_ratio/region_mean": 0.002475378933013417,
      "epoch": 0.39183673469387753,
      "grad_norm": 0.12117232382297516,
      "learning_rate": 1e-06,
      "loss": 0.009,
      "step": 42
    },
    {
      "clip_ratio/high_max": 0.002626564892125316,
      "clip_ratio/high_mean": 0.0012901496957056224,
      "clip_ratio/low_mean": 0.0014641547277278733,
      "clip_ratio/low_min": 0.00010231374562863493,
      "clip_ratio/region_mean": 0.0027543044125195593,
      "epoch": 0.40116618075801747,
      "grad_norm": 0.12484277039766312,
      "learning_rate": 1e-06,
      "loss": 0.0042,
      "step": 43
    },
    {
      "clip_ratio/high_max": 0.0026751656587293837,
      "clip_ratio/high_mean": 0.0013287226647662465,
      "clip_ratio/low_mean": 0.0013269794908410404,
      "clip_ratio/low_min": 0.0003367429753780016,
      "clip_ratio/region_mean": 0.002655702155607287,
      "epoch": 0.41049562682215746,
      "grad_norm": 0.12827810645103455,
      "learning_rate": 1e-06,
      "loss": -0.0553,
      "step": 44
    },
    {
      "clip_ratio/high_max": 0.002334374737984035,
      "clip_ratio/high_mean": 0.0010745539257186465,
      "clip_ratio/low_mean": 0.001300467600231059,
      "clip_ratio/low_min": 0.00024644834047649056,
      "clip_ratio/region_mean": 0.0023750215186737478,
      "epoch": 0.4198250728862974,
      "grad_norm": 0.13005675375461578,
      "learning_rate": 1e-06,
      "loss": 0.0297,
      "step": 45
    },
    {
      "clip_ratio/high_max": 0.002404369901341852,
      "clip_ratio/high_mean": 0.0011186322062712861,
      "clip_ratio/low_mean": 0.0014560700765287038,
      "clip_ratio/low_min": 0.00018331499268242624,
      "clip_ratio/region_mean": 0.0025747022591531277,
      "epoch": 0.4291545189504373,
      "grad_norm": 0.1254495084285736,
      "learning_rate": 1e-06,
      "loss": 0.0067,
      "step": 46
    },
    {
      "clip_ratio/high_max": 0.0028180109220556915,
      "clip_ratio/high_mean": 0.0012702910062216688,
      "clip_ratio/low_mean": 0.0012854832057200838,
      "clip_ratio/low_min": 0.00023281864105229033,
      "clip_ratio/region_mean": 0.0025557741901138797,
      "epoch": 0.43848396501457726,
      "grad_norm": 0.13360561430454254,
      "learning_rate": 1e-06,
      "loss": -0.0471,
      "step": 47
    },
    {
      "clip_ratio/high_max": 0.002505217395082582,
      "clip_ratio/high_mean": 0.001214497637192835,
      "clip_ratio/low_mean": 0.0014355284001794644,
      "clip_ratio/low_min": 0.00041710363802849315,
      "clip_ratio/region_mean": 0.0026500260792090558,
      "epoch": 0.4478134110787172,
      "grad_norm": 0.1364157795906067,
      "learning_rate": 1e-06,
      "loss": 0.0439,
      "step": 48
    },
    {
      "clip_ratio/high_max": 0.002508827055862639,
      "clip_ratio/high_mean": 0.0011414676773711108,
      "clip_ratio/low_mean": 0.0014058454689802602,
      "clip_ratio/low_min": 0.0002719927506404929,
      "clip_ratio/region_mean": 0.002547313124523498,
      "epoch": 0.45714285714285713,
      "grad_norm": 0.12717197835445404,
      "learning_rate": 1e-06,
      "loss": 0.0095,
      "step": 49
    },
    {
      "clip_ratio/high_max": 0.002623106040118728,
      "clip_ratio/high_mean": 0.0012397630598570686,
      "clip_ratio/low_mean": 0.0014246280334191397,
      "clip_ratio/low_min": 0.0001757342561177211,
      "clip_ratio/region_mean": 0.0026643910532584414,
      "epoch": 0.46647230320699706,
      "grad_norm": 0.13832078874111176,
      "learning_rate": 1e-06,
      "loss": -0.0098,
      "step": 50
    },
    {
      "clip_ratio/high_max": 0.0022491210911539383,
      "clip_ratio/high_mean": 0.0009918743453454226,
      "clip_ratio/low_mean": 0.0014475052776106168,
      "clip_ratio/low_min": 0.00035459636001178296,
      "clip_ratio/region_mean": 0.0024393795829382725,
      "epoch": 0.47580174927113705,
      "grad_norm": 0.13653312623500824,
      "learning_rate": 1e-06,
      "loss": 0.0559,
      "step": 51
    },
    {
      "clip_ratio/high_max": 0.002902555323089473,
      "clip_ratio/high_mean": 0.0014078625245019794,
      "clip_ratio/low_mean": 0.0015024737840576563,
      "clip_ratio/low_min": 0.00015843062828935217,
      "clip_ratio/region_mean": 0.0029103363849571906,
      "epoch": 0.485131195335277,
      "grad_norm": 0.135065495967865,
      "learning_rate": 1e-06,
      "loss": -0.0427,
      "step": 52
    },
    {
      "clip_ratio/high_max": 0.002970592620840762,
      "clip_ratio/high_mean": 0.001307408521824982,
      "clip_ratio/low_mean": 0.0016858880917425267,
      "clip_ratio/low_min": 0.00040222108418674907,
      "clip_ratio/region_mean": 0.0029932965408079326,
      "epoch": 0.4944606413994169,
      "grad_norm": 0.12439320981502533,
      "learning_rate": 1e-06,
      "loss": 0.0004,
      "step": 53
    },
    {
      "clip_ratio/high_max": 0.00270408335927641,
      "clip_ratio/high_mean": 0.0011514070811244892,
      "clip_ratio/low_mean": 0.001619567414309131,
      "clip_ratio/low_min": 0.00038420907822001027,
      "clip_ratio/region_mean": 0.002770974548184313,
      "epoch": 0.5037900874635568,
      "grad_norm": 0.11202271282672882,
      "learning_rate": 1e-06,
      "loss": 0.0129,
      "step": 54
    },
    {
      "clip_ratio/high_max": 0.0030225567606976256,
      "clip_ratio/high_mean": 0.0012184053775854409,
      "clip_ratio/low_mean": 0.0016841035976540297,
      "clip_ratio/low_min": 0.0003986812216680846,
      "clip_ratio/region_mean": 0.002902509004343301,
      "epoch": 0.5131195335276968,
      "grad_norm": 0.12663580477237701,
      "learning_rate": 1e-06,
      "loss": 0.022,
      "step": 55
    },
    {
      "clip_ratio/high_max": 0.0027979429287370294,
      "clip_ratio/high_mean": 0.0012466053067328176,
      "clip_ratio/low_mean": 0.0015858514307183214,
      "clip_ratio/low_min": 0.00025280885984102497,
      "clip_ratio/region_mean": 0.0028324567392701283,
      "epoch": 0.5224489795918368,
      "grad_norm": 0.12395124137401581,
      "learning_rate": 1e-06,
      "loss": 0.0161,
      "step": 56
    },
    {
      "clip_ratio/high_max": 0.0025067654423764907,
      "clip_ratio/high_mean": 0.001166765494417632,
      "clip_ratio/low_mean": 0.0014032908948138356,
      "clip_ratio/low_min": 0.000294826317258412,
      "clip_ratio/region_mean": 0.0025700564074213617,
      "epoch": 0.5317784256559767,
      "grad_norm": 0.12397922575473785,
      "learning_rate": 1e-06,
      "loss": -0.0104,
      "step": 57
    },
    {
      "clip_ratio/high_max": 0.002598407110781409,
      "clip_ratio/high_mean": 0.0012452545888663735,
      "clip_ratio/low_mean": 0.0013761853988398798,
      "clip_ratio/low_min": 8.24566468509147e-05,
      "clip_ratio/region_mean": 0.002621439976792317,
      "epoch": 0.5411078717201167,
      "grad_norm": 0.11927986145019531,
      "learning_rate": 1e-06,
      "loss": 0.0114,
      "step": 58
    },
    {
      "clip_ratio/high_max": 0.002735268135438673,
      "clip_ratio/high_mean": 0.001259966225916287,
      "clip_ratio/low_mean": 0.0016105202994367573,
      "clip_ratio/low_min": 0.0002903498934756499,
      "clip_ratio/region_mean": 0.0028704865399049595,
      "epoch": 0.5504373177842565,
      "grad_norm": 0.13499929010868073,
      "learning_rate": 1e-06,
      "loss": 0.0192,
      "step": 59
    },
    {
      "clip_ratio/high_max": 0.002663150749867782,
      "clip_ratio/high_mean": 0.0013437504931061994,
      "clip_ratio/low_mean": 0.0012987726768187713,
      "clip_ratio/low_min": 0.00014798885058553424,
      "clip_ratio/region_mean": 0.002642523162649013,
      "epoch": 0.5597667638483965,
      "grad_norm": 0.11873684078454971,
      "learning_rate": 1e-06,
      "loss": -0.0634,
      "step": 60
    },
    {
      "clip_ratio/high_max": 0.0028377935814205557,
      "clip_ratio/high_mean": 0.0012613039725692943,
      "clip_ratio/low_mean": 0.0013451987870212179,
      "clip_ratio/low_min": 7.511279727623332e-05,
      "clip_ratio/region_mean": 0.002606502835988067,
      "epoch": 0.5690962099125364,
      "grad_norm": 0.12681037187576294,
      "learning_rate": 1e-06,
      "loss": 0.0163,
      "step": 61
    },
    {
      "clip_ratio/high_max": 0.0026151974452659488,
      "clip_ratio/high_mean": 0.0012464570500014815,
      "clip_ratio/low_mean": 0.0014267960323195439,
      "clip_ratio/low_min": 0.00012743554179905914,
      "clip_ratio/region_mean": 0.002673253125976771,
      "epoch": 0.5784256559766764,
      "grad_norm": 0.13074587285518646,
      "learning_rate": 1e-06,
      "loss": 0.0177,
      "step": 62
    },
    {
      "clip_ratio/high_max": 0.0027048191550420597,
      "clip_ratio/high_mean": 0.001161007228802191,
      "clip_ratio/low_mean": 0.0015286516900232527,
      "clip_ratio/low_min": 0.00031732848583487794,
      "clip_ratio/region_mean": 0.002689658969757147,
      "epoch": 0.5877551020408164,
      "grad_norm": 0.1284467577934265,
      "learning_rate": 1e-06,
      "loss": 0.0316,
      "step": 63
    },
    {
      "clip_ratio/high_max": 0.0025533259758958593,
      "clip_ratio/high_mean": 0.001261223471374251,
      "clip_ratio/low_mean": 0.0014119726583885495,
      "clip_ratio/low_min": 0.0001766451823641546,
      "clip_ratio/region_mean": 0.0026731961042969488,
      "epoch": 0.5970845481049563,
      "grad_norm": 0.11957770586013794,
      "learning_rate": 1e-06,
      "loss": 0.0072,
      "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": 0.015223911830357095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4092.0,
      "completions/mean_length": 611.696044921875,
      "completions/mean_terminated_length": 557.831298828125,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 1.00932944606414,
      "grad_norm": 0.14397354423999786,
      "learning_rate": 1e-06,
      "loss": -0.008,
      "num_tokens": 73998447.0,
      "reward": 0.518798828125,
      "reward_std": 0.24383334815502167,
      "rewards/simpleverify_reward/mean": 0.518798828125,
      "rewards/simpleverify_reward/std": 0.499650776386261,
      "step": 65
    },
    {
      "clip_ratio/high_max": 0.0021691659931093454,
      "clip_ratio/high_mean": 0.000986040557108936,
      "clip_ratio/low_mean": 0.0006652650536125293,
      "clip_ratio/low_min": 2.287994448124664e-05,
      "clip_ratio/region_mean": 0.0016513056034455076,
      "epoch": 1.01865889212828,
      "grad_norm": 0.1515653133392334,
      "learning_rate": 1e-06,
      "loss": -0.0171,
      "step": 66
    },
    {
      "clip_ratio/high_max": 0.0018939887231681496,
      "clip_ratio/high_mean": 0.0009140307156485505,
      "clip_ratio/low_mean": 0.0006537802073580679,
      "clip_ratio/low_min": 6.565342209796654e-05,
      "clip_ratio/region_mean": 0.0015678109630243853,
      "epoch": 1.0279883381924197,
      "grad_norm": 0.13714466989040375,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 67
    },
    {
      "clip_ratio/high_max": 0.0023807338438928127,
      "clip_ratio/high_mean": 0.0009935587368090637,
      "clip_ratio/low_mean": 0.0007937938444229076,
      "clip_ratio/low_min": 1.8679020286072046e-05,
      "clip_ratio/region_mean": 0.0017873525648610666,
      "epoch": 1.0373177842565597,
      "grad_norm": 0.13353446125984192,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "step": 68
    },
    {
      "clip_ratio/high_max": 0.002437128023302648,
      "clip_ratio/high_mean": 0.0011547244575922377,
      "clip_ratio/low_mean": 0.0006415299794753082,
      "clip_ratio/low_min": 5.737626906920923e-05,
      "clip_ratio/region_mean": 0.0017962544807232916,
      "epoch": 1.0466472303206997,
      "grad_norm": 0.12103041261434555,
      "learning_rate": 1e-06,
      "loss": -0.0682,
      "step": 69
    },
    {
      "clip_ratio/high_max": 0.002175499241275247,
      "clip_ratio/high_mean": 0.0010054696431325283,
      "clip_ratio/low_mean": 0.0010076910075440537,
      "clip_ratio/low_min": 0.00017112510613515042,
      "clip_ratio/region_mean": 0.0020131606652284972,
      "epoch": 1.0559766763848397,
      "grad_norm": 0.1344999372959137,
      "learning_rate": 1e-06,
      "loss": 0.0129,
      "step": 70
    },
    {
      "clip_ratio/high_max": 0.002760676587058697,
      "clip_ratio/high_mean": 0.001276401770155644,
      "clip_ratio/low_mean": 0.0011189830875082407,
      "clip_ratio/low_min": 8.982975941762561e-05,
      "clip_ratio/region_mean": 0.002395384850387927,
      "epoch": 1.0653061224489795,
      "grad_norm": 0.12529118359088898,
      "learning_rate": 1e-06,
      "loss": -0.0272,
      "step": 71
    },
    {
      "clip_ratio/high_max": 0.002425005626719212,
      "clip_ratio/high_mean": 0.0012668097879213747,
      "clip_ratio/low_mean": 0.001233299874002114,
      "clip_ratio/low_min": 0.00016012470769055653,
      "clip_ratio/region_mean": 0.002500109672837425,
      "epoch": 1.0746355685131195,
      "grad_norm": 0.12796109914779663,
      "learning_rate": 1e-06,
      "loss": -0.0255,
      "step": 72
    },
    {
      "clip_ratio/high_max": 0.0027319155851728283,
      "clip_ratio/high_mean": 0.0012632101424969733,
      "clip_ratio/low_mean": 0.001098712942621205,
      "clip_ratio/low_min": 8.281663940579165e-05,
      "clip_ratio/region_mean": 0.0023619230996700935,
      "epoch": 1.0839650145772595,
      "grad_norm": 0.13395914435386658,
      "learning_rate": 1e-06,
      "loss": 0.0012,
      "step": 73
    },
    {
      "clip_ratio/high_max": 0.002780590672045946,
      "clip_ratio/high_mean": 0.0012711367890005931,
      "clip_ratio/low_mean": 0.001221397178596817,
      "clip_ratio/low_min": 0.00021613473745674128,
      "clip_ratio/region_mean": 0.0024925339603214525,
      "epoch": 1.0932944606413995,
      "grad_norm": 0.1335933655500412,
      "learning_rate": 1e-06,
      "loss": -0.0042,
      "step": 74
    },
    {
      "clip_ratio/high_max": 0.0024426456511719152,
      "clip_ratio/high_mean": 0.0010831852414412424,
      "clip_ratio/low_mean": 0.0012574343418236822,
      "clip_ratio/low_min": 0.00013535620473703602,
      "clip_ratio/region_mean": 0.0023406196196447127,
      "epoch": 1.1026239067055394,
      "grad_norm": 0.12600579857826233,
      "learning_rate": 1e-06,
      "loss": 0.0296,
      "step": 75
    },
    {
      "clip_ratio/high_max": 0.002680013974895701,
      "clip_ratio/high_mean": 0.0011871287460962776,
      "clip_ratio/low_mean": 0.0013076191280561034,
      "clip_ratio/low_min": 0.00019419438103795983,
      "clip_ratio/region_mean": 0.002494747932360042,
      "epoch": 1.1119533527696792,
      "grad_norm": 0.1342019885778427,
      "learning_rate": 1e-06,
      "loss": 0.0173,
      "step": 76
    },
    {
      "clip_ratio/high_max": 0.00274390618869802,
      "clip_ratio/high_mean": 0.0011277977500867564,
      "clip_ratio/low_mean": 0.001229974255693378,
      "clip_ratio/low_min": 0.00015567049013043288,
      "clip_ratio/region_mean": 0.0023577720057801344,
      "epoch": 1.1212827988338192,
      "grad_norm": 0.12711849808692932,
      "learning_rate": 1e-06,
      "loss": 0.0563,
      "step": 77
    },
    {
      "clip_ratio/high_max": 0.00250905611392227,
      "clip_ratio/high_mean": 0.0011992687723250128,
      "clip_ratio/low_mean": 0.0009937717040884309,
      "clip_ratio/low_min": 0.00012563654308905825,
      "clip_ratio/region_mean": 0.0021930404982413165,
      "epoch": 1.1306122448979592,
      "grad_norm": 0.12553954124450684,
      "learning_rate": 1e-06,
      "loss": 0.0198,
      "step": 78
    },
    {
      "clip_ratio/high_max": 0.002618203783640638,
      "clip_ratio/high_mean": 0.0011074682588514406,
      "clip_ratio/low_mean": 0.0010921286484517623,
      "clip_ratio/low_min": 0.0001569984806337743,
      "clip_ratio/region_mean": 0.002199596907303203,
      "epoch": 1.1399416909620992,
      "grad_norm": 0.12352953851222992,
      "learning_rate": 1e-06,
      "loss": 0.0117,
      "step": 79
    },
    {
      "clip_ratio/high_max": 0.0024745419214013964,
      "clip_ratio/high_mean": 0.001142636134318309,
      "clip_ratio/low_mean": 0.0010781606142700184,
      "clip_ratio/low_min": 0.00022882952453073813,
      "clip_ratio/region_mean": 0.0022207967122085392,
      "epoch": 1.149271137026239,
      "grad_norm": 0.13095960021018982,
      "learning_rate": 1e-06,
      "loss": 0.0198,
      "step": 80
    },
    {
      "clip_ratio/high_max": 0.0026096545552718453,
      "clip_ratio/high_mean": 0.0012076985840394627,
      "clip_ratio/low_mean": 0.0010498119991098065,
      "clip_ratio/low_min": 0.00014993723016232252,
      "clip_ratio/region_mean": 0.0022575106195290573,
      "epoch": 1.158600583090379,
      "grad_norm": 0.1313534677028656,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 81
    },
    {
      "clip_ratio/high_max": 0.0024646231322549284,
      "clip_ratio/high_mean": 0.0010702001382014714,
      "clip_ratio/low_mean": 0.0008362182961718645,
      "clip_ratio/low_min": 5.304603291733656e-05,
      "clip_ratio/region_mean": 0.00190641845256323,
      "epoch": 1.167930029154519,
      "grad_norm": 0.11646479368209839,
      "learning_rate": 1e-06,
      "loss": 0.006,
      "step": 82
    },
    {
      "clip_ratio/high_max": 0.0026362404896644875,
      "clip_ratio/high_mean": 0.0011965960220550187,
      "clip_ratio/low_mean": 0.0010191613037022762,
      "clip_ratio/low_min": 9.004955882119248e-05,
      "clip_ratio/region_mean": 0.0022157572966534644,
      "epoch": 1.177259475218659,
      "grad_norm": 0.1397159844636917,
      "learning_rate": 1e-06,
      "loss": 0.0168,
      "step": 83
    },
    {
      "clip_ratio/high_max": 0.002594729870907031,
      "clip_ratio/high_mean": 0.0011606750231294427,
      "clip_ratio/low_mean": 0.001007588787615532,
      "clip_ratio/low_min": 0.00011338917374814628,
      "clip_ratio/region_mean": 0.002168263803469017,
      "epoch": 1.186588921282799,
      "grad_norm": 0.116698257625103,
      "learning_rate": 1e-06,
      "loss": -0.0184,
      "step": 84
    },
    {
      "clip_ratio/high_max": 0.002944628540717531,
      "clip_ratio/high_mean": 0.0012051014819007833,
      "clip_ratio/low_mean": 0.0009211932047037408,
      "clip_ratio/low_min": 9.40364752750611e-05,
      "clip_ratio/region_mean": 0.002126294632034842,
      "epoch": 1.1959183673469387,
      "grad_norm": 0.12869130074977875,
      "learning_rate": 1e-06,
      "loss": -0.0503,
      "step": 85
    },
    {
      "clip_ratio/high_max": 0.003023597426363267,
      "clip_ratio/high_mean": 0.0012567652265715878,
      "clip_ratio/low_mean": 0.0010700837374315597,
      "clip_ratio/low_min": 0.00016499041339557152,
      "clip_ratio/region_mean": 0.0023268489676411264,
      "epoch": 1.2052478134110787,
      "grad_norm": 0.1386902779340744,
      "learning_rate": 1e-06,
      "loss": -0.0133,
      "step": 86
    },
    {
      "clip_ratio/high_max": 0.0020999868429498747,
      "clip_ratio/high_mean": 0.0010495151036593597,
      "clip_ratio/low_mean": 0.0013269814935483737,
      "clip_ratio/low_min": 0.00025912320597853977,
      "clip_ratio/region_mean": 0.0023764966026647016,
      "epoch": 1.2145772594752187,
      "grad_norm": 0.11641263216733932,
      "learning_rate": 1e-06,
      "loss": 0.0314,
      "step": 87
    },
    {
      "clip_ratio/high_max": 0.002789079662761651,
      "clip_ratio/high_mean": 0.001221463731781114,
      "clip_ratio/low_mean": 0.001243677568709245,
      "clip_ratio/low_min": 6.885257789690513e-05,
      "clip_ratio/region_mean": 0.0024651413405081257,
      "epoch": 1.2239067055393587,
      "grad_norm": 0.13545332849025726,
      "learning_rate": 1e-06,
      "loss": 0.0046,
      "step": 88
    },
    {
      "clip_ratio/high_max": 0.0023673701580264606,
      "clip_ratio/high_mean": 0.001076894306606846,
      "clip_ratio/low_mean": 0.0011390012023184681,
      "clip_ratio/low_min": 5.4253629059530795e-05,
      "clip_ratio/region_mean": 0.0022158954889164306,
      "epoch": 1.2332361516034984,
      "grad_norm": 0.12344934791326523,
      "learning_rate": 1e-06,
      "loss": -0.027,
      "step": 89
    },
    {
      "clip_ratio/high_max": 0.002265146693389397,
      "clip_ratio/high_mean": 0.0010651956836227328,
      "clip_ratio/low_mean": 0.0013935370188846719,
      "clip_ratio/low_min": 0.00024814777134452015,
      "clip_ratio/region_mean": 0.0024587327061453834,
      "epoch": 1.2425655976676384,
      "grad_norm": 0.1261015087366104,
      "learning_rate": 1e-06,
      "loss": 0.0214,
      "step": 90
    },
    {
      "clip_ratio/high_max": 0.0024361753421544563,
      "clip_ratio/high_mean": 0.0010475805756868795,
      "clip_ratio/low_mean": 0.0014298978239821736,
      "clip_ratio/low_min": 0.0002577286486484809,
      "clip_ratio/region_mean": 0.0024774784251349047,
      "epoch": 1.2518950437317784,
      "grad_norm": 0.11616481840610504,
      "learning_rate": 1e-06,
      "loss": 0.0076,
      "step": 91
    },
    {
      "clip_ratio/high_max": 0.0026217842314508744,
      "clip_ratio/high_mean": 0.0012180431076558307,
      "clip_ratio/low_mean": 0.0014414700308407191,
      "clip_ratio/low_min": 0.00017707774895825423,
      "clip_ratio/region_mean": 0.0026595131057547405,
      "epoch": 1.2612244897959184,
      "grad_norm": 0.12859784066677094,
      "learning_rate": 1e-06,
      "loss": -0.0164,
      "step": 92
    },
    {
      "clip_ratio/high_max": 0.002220765261881752,
      "clip_ratio/high_mean": 0.0010178859138250118,
      "clip_ratio/low_mean": 0.0013474979423335753,
      "clip_ratio/low_min": 0.00016744094500609208,
      "clip_ratio/region_mean": 0.002365383850701619,
      "epoch": 1.2705539358600584,
      "grad_norm": 0.1287926584482193,
      "learning_rate": 1e-06,
      "loss": 0.024,
      "step": 93
    },
    {
      "clip_ratio/high_max": 0.0025237340450985357,
      "clip_ratio/high_mean": 0.0011239920822845306,
      "clip_ratio/low_mean": 0.0013649618667841423,
      "clip_ratio/low_min": 0.00015807248564669862,
      "clip_ratio/region_mean": 0.002488953970896546,
      "epoch": 1.2798833819241984,
      "grad_norm": 0.13355687260627747,
      "learning_rate": 1e-06,
      "loss": 0.0121,
      "step": 94
    },
    {
      "clip_ratio/high_max": 0.002789656529785134,
      "clip_ratio/high_mean": 0.0011940061158384196,
      "clip_ratio/low_mean": 0.0012921697198180482,
      "clip_ratio/low_min": 0.00013409827261057217,
      "clip_ratio/region_mean": 0.0024861758211045526,
      "epoch": 1.2892128279883381,
      "grad_norm": 0.11790642887353897,
      "learning_rate": 1e-06,
      "loss": 0.0033,
      "step": 95
    },
    {
      "clip_ratio/high_max": 0.0025969174603233114,
      "clip_ratio/high_mean": 0.0011176285879628267,
      "clip_ratio/low_mean": 0.0013295766366354655,
      "clip_ratio/low_min": 0.00017343667241220828,
      "clip_ratio/region_mean": 0.0024472052027704194,
      "epoch": 1.2985422740524781,
      "grad_norm": 0.12013210356235504,
      "learning_rate": 1e-06,
      "loss": 0.0186,
      "step": 96
    },
    {
      "clip_ratio/high_max": 0.002730941509071272,
      "clip_ratio/high_mean": 0.0012301779170229565,
      "clip_ratio/low_mean": 0.0013543912064051256,
      "clip_ratio/low_min": 0.00025461651785008144,
      "clip_ratio/region_mean": 0.002584569127066061,
      "epoch": 1.3078717201166181,
      "grad_norm": 0.13513752818107605,
      "learning_rate": 1e-06,
      "loss": 0.0166,
      "step": 97
    },
    {
      "clip_ratio/high_max": 0.00276844174368307,
      "clip_ratio/high_mean": 0.0012111759460822213,
      "clip_ratio/low_mean": 0.0012890180332760792,
      "clip_ratio/low_min": 0.00014023223957337905,
      "clip_ratio/region_mean": 0.0025001940230140463,
      "epoch": 1.3172011661807579,
      "grad_norm": 0.12273237109184265,
      "learning_rate": 1e-06,
      "loss": -0.0035,
      "step": 98
    },
    {
      "clip_ratio/high_max": 0.0023963413841556758,
      "clip_ratio/high_mean": 0.0010976555058732629,
      "clip_ratio/low_mean": 0.0012034707215207163,
      "clip_ratio/low_min": 0.000233849114010809,
      "clip_ratio/region_mean": 0.002301126216480043,
      "epoch": 1.3265306122448979,
      "grad_norm": 0.11905274540185928,
      "learning_rate": 1e-06,
      "loss": 0.03,
      "step": 99
    },
    {
      "clip_ratio/high_max": 0.0026407210316392593,
      "clip_ratio/high_mean": 0.001146331040217774,
      "clip_ratio/low_mean": 0.0012838526636187453,
      "clip_ratio/low_min": 0.00028504238434834406,
      "clip_ratio/region_mean": 0.002430183667456731,
      "epoch": 1.3358600583090379,
      "grad_norm": 0.10808532685041428,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 100
    },
    {
      "clip_ratio/high_max": 0.002646134111273568,
      "clip_ratio/high_mean": 0.001193745136333746,
      "clip_ratio/low_mean": 0.0011777418985730037,
      "clip_ratio/low_min": 0.00012790255914296722,
      "clip_ratio/region_mean": 0.0023714870258118026,
      "epoch": 1.3451895043731779,
      "grad_norm": 0.1202731505036354,
      "learning_rate": 1e-06,
      "loss": -0.0177,
      "step": 101
    },
    {
      "clip_ratio/high_max": 0.002772345149423927,
      "clip_ratio/high_mean": 0.0012068417308910284,
      "clip_ratio/low_mean": 0.0013185823318053735,
      "clip_ratio/low_min": 0.0002663707937244908,
      "clip_ratio/region_mean": 0.0025254240608774126,
      "epoch": 1.3545189504373178,
      "grad_norm": 0.1238364428281784,
      "learning_rate": 1e-06,
      "loss": 0.0313,
      "step": 102
    },
    {
      "clip_ratio/high_max": 0.0025565283285686746,
      "clip_ratio/high_mean": 0.001128703574067913,
      "clip_ratio/low_mean": 0.001262762634723913,
      "clip_ratio/low_min": 0.00019348273417563178,
      "clip_ratio/region_mean": 0.002391466245171614,
      "epoch": 1.3638483965014578,
      "grad_norm": 0.12007968127727509,
      "learning_rate": 1e-06,
      "loss": 0.0374,
      "step": 103
    },
    {
      "clip_ratio/high_max": 0.002913759570219554,
      "clip_ratio/high_mean": 0.0012902501111966558,
      "clip_ratio/low_mean": 0.001323579916061135,
      "clip_ratio/low_min": 0.00014899642701493576,
      "clip_ratio/region_mean": 0.0026138300163438544,
      "epoch": 1.3731778425655976,
      "grad_norm": 0.11830270290374756,
      "learning_rate": 1e-06,
      "loss": -0.0538,
      "step": 104
    },
    {
      "clip_ratio/high_max": 0.002359229707508348,
      "clip_ratio/high_mean": 0.001132728015363682,
      "clip_ratio/low_mean": 0.0014911993275745772,
      "clip_ratio/low_min": 0.0002770377668639412,
      "clip_ratio/region_mean": 0.002623927350214217,
      "epoch": 1.3825072886297376,
      "grad_norm": 0.13317887485027313,
      "learning_rate": 1e-06,
      "loss": 0.0613,
      "step": 105
    },
    {
      "clip_ratio/high_max": 0.002849008917110041,
      "clip_ratio/high_mean": 0.0013902151185902767,
      "clip_ratio/low_mean": 0.0012934510014019907,
      "clip_ratio/low_min": 0.00018193295545643196,
      "clip_ratio/region_mean": 0.002683666090888437,
      "epoch": 1.3918367346938776,
      "grad_norm": 0.1259160339832306,
      "learning_rate": 1e-06,
      "loss": -0.0508,
      "step": 106
    },
    {
      "clip_ratio/high_max": 0.0025456000585108995,
      "clip_ratio/high_mean": 0.0010939477979263756,
      "clip_ratio/low_mean": 0.0012786947954737116,
      "clip_ratio/low_min": 0.00023171612610894954,
      "clip_ratio/region_mean": 0.0023726425642962568,
      "epoch": 1.4011661807580174,
      "grad_norm": 0.12226609140634537,
      "learning_rate": 1e-06,
      "loss": 0.0334,
      "step": 107
    },
    {
      "clip_ratio/high_max": 0.0025956555691664107,
      "clip_ratio/high_mean": 0.0012804821199097205,
      "clip_ratio/low_mean": 0.001096187344955979,
      "clip_ratio/low_min": 4.2086021494469605e-05,
      "clip_ratio/region_mean": 0.0023766695157974027,
      "epoch": 1.4104956268221573,
      "grad_norm": 0.12037496268749237,
      "learning_rate": 1e-06,
      "loss": -0.0331,
      "step": 108
    },
    {
      "clip_ratio/high_max": 0.002631728450069204,
      "clip_ratio/high_mean": 0.0013049101835349575,
      "clip_ratio/low_mean": 0.0012595389234775212,
      "clip_ratio/low_min": 0.00012353983856883133,
      "clip_ratio/region_mean": 0.0025644490160630085,
      "epoch": 1.4198250728862973,
      "grad_norm": 0.11882280558347702,
      "learning_rate": 1e-06,
      "loss": -0.0165,
      "step": 109
    },
    {
      "clip_ratio/high_max": 0.002339031118026469,
      "clip_ratio/high_mean": 0.0010034456208813936,
      "clip_ratio/low_mean": 0.0013370133092394099,
      "clip_ratio/low_min": 0.00019935806267312728,
      "clip_ratio/region_mean": 0.0023404588937410153,
      "epoch": 1.4291545189504373,
      "grad_norm": 0.11937922984361649,
      "learning_rate": 1e-06,
      "loss": 0.0289,
      "step": 110
    },
    {
      "clip_ratio/high_max": 0.0025980642822105438,
      "clip_ratio/high_mean": 0.001102815480408026,
      "clip_ratio/low_mean": 0.0013320943726284895,
      "clip_ratio/low_min": 0.00018518740307627013,
      "clip_ratio/region_mean": 0.0024349098675884306,
      "epoch": 1.4384839650145773,
      "grad_norm": 0.12355878204107285,
      "learning_rate": 1e-06,
      "loss": 0.0311,
      "step": 111
    },
    {
      "clip_ratio/high_max": 0.0027734357863664627,
      "clip_ratio/high_mean": 0.00125026136811357,
      "clip_ratio/low_mean": 0.0013873162351956125,
      "clip_ratio/low_min": 0.00022643257216259371,
      "clip_ratio/region_mean": 0.0026375776360509917,
      "epoch": 1.4478134110787173,
      "grad_norm": 0.13517841696739197,
      "learning_rate": 1e-06,
      "loss": -0.0237,
      "step": 112
    },
    {
      "clip_ratio/high_max": 0.0028571598086273298,
      "clip_ratio/high_mean": 0.0012286463352211285,
      "clip_ratio/low_mean": 0.0012820788542740047,
      "clip_ratio/low_min": 0.00010312809172319248,
      "clip_ratio/region_mean": 0.002510725178581197,
      "epoch": 1.457142857142857,
      "grad_norm": 0.13367316126823425,
      "learning_rate": 1e-06,
      "loss": -0.0258,
      "step": 113
    },
    {
      "clip_ratio/high_max": 0.0026264126936439425,
      "clip_ratio/high_mean": 0.0013097691626171581,
      "clip_ratio/low_mean": 0.0011364430793037172,
      "clip_ratio/low_min": 0.0001136418832174968,
      "clip_ratio/region_mean": 0.002446212231006939,
      "epoch": 1.466472303206997,
      "grad_norm": 0.12540413439273834,
      "learning_rate": 1e-06,
      "loss": -0.0729,
      "step": 114
    },
    {
      "clip_ratio/high_max": 0.0027379567764000967,
      "clip_ratio/high_mean": 0.001255874401977053,
      "clip_ratio/low_mean": 0.0014245363781810738,
      "clip_ratio/low_min": 0.00013933339687355328,
      "clip_ratio/region_mean": 0.0026804107474163175,
      "epoch": 1.475801749271137,
      "grad_norm": 0.12620235979557037,
      "learning_rate": 1e-06,
      "loss": -0.0501,
      "step": 115
    },
    {
      "clip_ratio/high_max": 0.0027128457950311713,
      "clip_ratio/high_mean": 0.001143619228969328,
      "clip_ratio/low_mean": 0.0015026697838038672,
      "clip_ratio/low_min": 0.0003423382295295596,
      "clip_ratio/region_mean": 0.0026462890236871317,
      "epoch": 1.485131195335277,
      "grad_norm": 0.13259916007518768,
      "learning_rate": 1e-06,
      "loss": 0.0598,
      "step": 116
    },
    {
      "clip_ratio/high_max": 0.002739972267590929,
      "clip_ratio/high_mean": 0.0011430823360569775,
      "clip_ratio/low_mean": 0.00152621894449112,
      "clip_ratio/low_min": 0.00020008663068438182,
      "clip_ratio/region_mean": 0.0026693012623582035,
      "epoch": 1.4944606413994168,
      "grad_norm": 0.1265629231929779,
      "learning_rate": 1e-06,
      "loss": 0.0165,
      "step": 117
    },
    {
      "clip_ratio/high_max": 0.0026464135371497832,
      "clip_ratio/high_mean": 0.0011411299274186604,
      "clip_ratio/low_mean": 0.0016658984532114118,
      "clip_ratio/low_min": 0.0003561044750313158,
      "clip_ratio/region_mean": 0.002807028402457945,
      "epoch": 1.5037900874635568,
      "grad_norm": 0.12272118777036667,
      "learning_rate": 1e-06,
      "loss": 0.0238,
      "step": 118
    },
    {
      "clip_ratio/high_max": 0.0023708494918537326,
      "clip_ratio/high_mean": 0.001126310646213824,
      "clip_ratio/low_mean": 0.001422655645001214,
      "clip_ratio/low_min": 0.0002803808511089301,
      "clip_ratio/region_mean": 0.0025489662803011015,
      "epoch": 1.5131195335276968,
      "grad_norm": 0.11832080781459808,
      "learning_rate": 1e-06,
      "loss": 0.0022,
      "step": 119
    },
    {
      "clip_ratio/high_max": 0.0027115305565530434,
      "clip_ratio/high_mean": 0.0011410775823605945,
      "clip_ratio/low_mean": 0.001546949762996519,
      "clip_ratio/low_min": 0.00015985599748091772,
      "clip_ratio/region_mean": 0.00268802738719387,
      "epoch": 1.5224489795918368,
      "grad_norm": 0.12926329672336578,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 120
    },
    {
      "clip_ratio/high_max": 0.0024924994722823612,
      "clip_ratio/high_mean": 0.0011002993305737618,
      "clip_ratio/low_mean": 0.0016539994103368372,
      "clip_ratio/low_min": 0.0002064968457489158,
      "clip_ratio/region_mean": 0.0027542986936168745,
      "epoch": 1.5317784256559768,
      "grad_norm": 0.12762393057346344,
      "learning_rate": 1e-06,
      "loss": 0.0508,
      "step": 121
    },
    {
      "clip_ratio/high_max": 0.0027188550302525982,
      "clip_ratio/high_mean": 0.0011645578233583365,
      "clip_ratio/low_mean": 0.001327578615018865,
      "clip_ratio/low_min": 0.00012605087704287143,
      "clip_ratio/region_mean": 0.002492136452929117,
      "epoch": 1.5411078717201168,
      "grad_norm": 0.11825843900442123,
      "learning_rate": 1e-06,
      "loss": -0.0314,
      "step": 122
    },
    {
      "clip_ratio/high_max": 0.0026440328801982105,
      "clip_ratio/high_mean": 0.001086208168999292,
      "clip_ratio/low_mean": 0.0014585714488930535,
      "clip_ratio/low_min": 0.00016493189650645945,
      "clip_ratio/region_mean": 0.002544779628806282,
      "epoch": 1.5504373177842565,
      "grad_norm": 0.11494413763284683,
      "learning_rate": 1e-06,
      "loss": 0.0197,
      "step": 123
    },
    {
      "clip_ratio/high_max": 0.0025432173133594915,
      "clip_ratio/high_mean": 0.001221496258949628,
      "clip_ratio/low_mean": 0.0014564773591700941,
      "clip_ratio/low_min": 0.00016185795720957685,
      "clip_ratio/region_mean": 0.0026779736144817434,
      "epoch": 1.5597667638483965,
      "grad_norm": 0.12496867775917053,
      "learning_rate": 1e-06,
      "loss": -0.0414,
      "step": 124
    },
    {
      "clip_ratio/high_max": 0.0027226682868786156,
      "clip_ratio/high_mean": 0.0011761972891690675,
      "clip_ratio/low_mean": 0.001594664470758289,
      "clip_ratio/low_min": 0.0003164536065014545,
      "clip_ratio/region_mean": 0.002770861829048954,
      "epoch": 1.5690962099125363,
      "grad_norm": 0.12851716578006744,
      "learning_rate": 1e-06,
      "loss": 0.0164,
      "step": 125
    },
    {
      "clip_ratio/high_max": 0.002813838582369499,
      "clip_ratio/high_mean": 0.0013134839355188888,
      "clip_ratio/low_mean": 0.001403712642058963,
      "clip_ratio/low_min": 0.00017359862067678478,
      "clip_ratio/region_mean": 0.0027171965411980636,
      "epoch": 1.5784256559766763,
      "grad_norm": 0.12595194578170776,
      "learning_rate": 1e-06,
      "loss": -0.0183,
      "step": 126
    },
    {
      "clip_ratio/high_max": 0.002468951170158107,
      "clip_ratio/high_mean": 0.0011238552433496807,
      "clip_ratio/low_mean": 0.0013457736095006112,
      "clip_ratio/low_min": 0.00025545547032379545,
      "clip_ratio/region_mean": 0.002469628867402207,
      "epoch": 1.5877551020408163,
      "grad_norm": 0.12416993081569672,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "step": 127
    },
    {
      "clip_ratio/high_max": 0.002700058714253828,
      "clip_ratio/high_mean": 0.001289540250581922,
      "clip_ratio/low_mean": 0.0016328505371348,
      "clip_ratio/low_min": 0.0003622167514549801,
      "clip_ratio/region_mean": 0.002922390791354701,
      "epoch": 1.5970845481049563,
      "grad_norm": 0.14150327444076538,
      "learning_rate": 1e-06,
      "loss": 0.0272,
      "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": 0.016688755580357095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4094.0,
      "completions/mean_length": 621.1257934570312,
      "completions/mean_terminated_length": 562.1502075195312,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 2.00932944606414,
      "grad_norm": 0.117433562874794,
      "learning_rate": 1e-06,
      "loss": 0.0143,
      "num_tokens": 111420498.0,
      "reward": 0.5467703938484192,
      "reward_std": 0.22544994950294495,
      "rewards/simpleverify_reward/mean": 0.5467703938484192,
      "rewards/simpleverify_reward/std": 0.49781206250190735,
      "step": 129
    },
    {
      "clip_ratio/high_max": 0.002604475354019087,
      "clip_ratio/high_mean": 0.0010097449339809828,
      "clip_ratio/low_mean": 0.0006020423088557436,
      "clip_ratio/low_min": 5.247698936727829e-05,
      "clip_ratio/region_mean": 0.001611787243746221,
      "epoch": 2.01865889212828,
      "grad_norm": 0.13862793147563934,
      "learning_rate": 1e-06,
      "loss": 0.0176,
      "step": 130
    },
    {
      "clip_ratio/high_max": 0.00264801752200583,
      "clip_ratio/high_mean": 0.0010852071100089233,
      "clip_ratio/low_mean": 0.0007682156174269039,
      "clip_ratio/low_min": 8.871639238350326e-05,
      "clip_ratio/region_mean": 0.001853422712883912,
      "epoch": 2.02798833819242,
      "grad_norm": 0.13071204721927643,
      "learning_rate": 1e-06,
      "loss": 0.0274,
      "step": 131
    },
    {
      "clip_ratio/high_max": 0.002122259458701592,
      "clip_ratio/high_mean": 0.0010517163027543575,
      "clip_ratio/low_mean": 0.0007108729969331762,
      "clip_ratio/low_min": 7.806723351677647e-05,
      "clip_ratio/region_mean": 0.0017625892942305654,
      "epoch": 2.03731778425656,
      "grad_norm": 0.12775462865829468,
      "learning_rate": 1e-06,
      "loss": -0.002,
      "step": 132
    },
    {
      "clip_ratio/high_max": 0.0023499272392655257,
      "clip_ratio/high_mean": 0.0011211607252334943,
      "clip_ratio/low_mean": 0.0006883806290716166,
      "clip_ratio/low_min": 0.00010209465199295664,
      "clip_ratio/region_mean": 0.0018095413834089413,
      "epoch": 2.0466472303206995,
      "grad_norm": 0.1324433982372284,
      "learning_rate": 1e-06,
      "loss": -0.0079,
      "step": 133
    },
    {
      "clip_ratio/high_max": 0.0022856025461805984,
      "clip_ratio/high_mean": 0.0010311819933122024,
      "clip_ratio/low_mean": 0.0008853242634359049,
      "clip_ratio/low_min": 3.5076616768492386e-05,
      "clip_ratio/region_mean": 0.001916506233101245,
      "epoch": 2.0559766763848395,
      "grad_norm": 0.11901295185089111,
      "learning_rate": 1e-06,
      "loss": 0.0205,
      "step": 134
    },
    {
      "clip_ratio/high_max": 0.0026263585314154625,
      "clip_ratio/high_mean": 0.0009817987520364113,
      "clip_ratio/low_mean": 0.0007800115417921916,
      "clip_ratio/low_min": 2.192982537962962e-05,
      "clip_ratio/region_mean": 0.0017618103011045605,
      "epoch": 2.0653061224489795,
      "grad_norm": 0.11027567833662033,
      "learning_rate": 1e-06,
      "loss": -0.0069,
      "step": 135
    },
    {
      "clip_ratio/high_max": 0.0024101691669784486,
      "clip_ratio/high_mean": 0.0011272243682469707,
      "clip_ratio/low_mean": 0.0010336852465115953,
      "clip_ratio/low_min": 0.00010650556123437127,
      "clip_ratio/region_mean": 0.0021609096438623965,
      "epoch": 2.0746355685131195,
      "grad_norm": 0.1345003843307495,
      "learning_rate": 1e-06,
      "loss": 0.0049,
      "step": 136
    },
    {
      "clip_ratio/high_max": 0.002551914469222538,
      "clip_ratio/high_mean": 0.0011151861217513215,
      "clip_ratio/low_mean": 0.0009470010681980057,
      "clip_ratio/low_min": 8.230858020397136e-05,
      "clip_ratio/region_mean": 0.00206218718085438,
      "epoch": 2.0839650145772595,
      "grad_norm": 0.12139299511909485,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 137
    },
    {
      "clip_ratio/high_max": 0.0024498684288118966,
      "clip_ratio/high_mean": 0.0011389465907996055,
      "clip_ratio/low_mean": 0.0008113535404845607,
      "clip_ratio/low_min": 7.063543580443365e-05,
      "clip_ratio/region_mean": 0.001950300153112039,
      "epoch": 2.0932944606413995,
      "grad_norm": 0.12825439870357513,
      "learning_rate": 1e-06,
      "loss": -0.0274,
      "step": 138
    },
    {
      "clip_ratio/high_max": 0.0026781300475704484,
      "clip_ratio/high_mean": 0.0011479368913569488,
      "clip_ratio/low_mean": 0.0009903935588226886,
      "clip_ratio/low_min": 9.716414569993503e-05,
      "clip_ratio/region_mean": 0.002138330506568309,
      "epoch": 2.1026239067055394,
      "grad_norm": 0.12423088401556015,
      "learning_rate": 1e-06,
      "loss": 0.0035,
      "step": 139
    },
    {
      "clip_ratio/high_max": 0.002209495934948791,
      "clip_ratio/high_mean": 0.001039613678585738,
      "clip_ratio/low_mean": 0.000837922811115277,
      "clip_ratio/low_min": 7.025053128018044e-05,
      "clip_ratio/region_mean": 0.001877536516985856,
      "epoch": 2.1119533527696794,
      "grad_norm": 0.11786050349473953,
      "learning_rate": 1e-06,
      "loss": 0.0093,
      "step": 140
    },
    {
      "clip_ratio/high_max": 0.002277497696923092,
      "clip_ratio/high_mean": 0.0010755934345070273,
      "clip_ratio/low_mean": 0.0009602748850738863,
      "clip_ratio/low_min": 0.00013899358418711927,
      "clip_ratio/region_mean": 0.002035868397797458,
      "epoch": 2.1212827988338194,
      "grad_norm": 0.10998106747865677,
      "learning_rate": 1e-06,
      "loss": 0.0057,
      "step": 141
    },
    {
      "clip_ratio/high_max": 0.0026260833037667908,
      "clip_ratio/high_mean": 0.0011495222570374608,
      "clip_ratio/low_mean": 0.0007736221250524977,
      "clip_ratio/low_min": 2.812830098264385e-05,
      "clip_ratio/region_mean": 0.0019231443584430963,
      "epoch": 2.130612244897959,
      "grad_norm": 0.11774852871894836,
      "learning_rate": 1e-06,
      "loss": -0.0475,
      "step": 142
    },
    {
      "clip_ratio/high_max": 0.002484799741068855,
      "clip_ratio/high_mean": 0.001073298950359458,
      "clip_ratio/low_mean": 0.000964514612860512,
      "clip_ratio/low_min": 8.693009112903383e-05,
      "clip_ratio/region_mean": 0.0020378135741339065,
      "epoch": 2.139941690962099,
      "grad_norm": 0.11981480568647385,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 143
    },
    {
      "clip_ratio/high_max": 0.0021002113207941875,
      "clip_ratio/high_mean": 0.0008897184779925738,
      "clip_ratio/low_mean": 0.0009667385347711388,
      "clip_ratio/low_min": 0.00014040692803973798,
      "clip_ratio/region_mean": 0.0018564570636954159,
      "epoch": 2.149271137026239,
      "grad_norm": 0.11887800693511963,
      "learning_rate": 1e-06,
      "loss": 0.0132,
      "step": 144
    },
    {
      "clip_ratio/high_max": 0.0024115627675200813,
      "clip_ratio/high_mean": 0.0010114257602253929,
      "clip_ratio/low_mean": 0.0009929460829880554,
      "clip_ratio/low_min": 7.398496745736338e-05,
      "clip_ratio/region_mean": 0.002004371846851427,
      "epoch": 2.158600583090379,
      "grad_norm": 0.11972886323928833,
      "learning_rate": 1e-06,
      "loss": 0.0076,
      "step": 145
    },
    {
      "clip_ratio/high_max": 0.0022257216442085337,
      "clip_ratio/high_mean": 0.0009576232823746977,
      "clip_ratio/low_mean": 0.0009820960804063361,
      "clip_ratio/low_min": 0.0001672759808570845,
      "clip_ratio/region_mean": 0.0019397193609620444,
      "epoch": 2.167930029154519,
      "grad_norm": 0.13610480725765228,
      "learning_rate": 1e-06,
      "loss": 0.0396,
      "step": 146
    },
    {
      "clip_ratio/high_max": 0.002450355350447353,
      "clip_ratio/high_mean": 0.001083731036487734,
      "clip_ratio/low_mean": 0.0010364038644183893,
      "clip_ratio/low_min": 7.660711708012968e-05,
      "clip_ratio/region_mean": 0.0021201349372859113,
      "epoch": 2.177259475218659,
      "grad_norm": 0.11820273846387863,
      "learning_rate": 1e-06,
      "loss": 0.0086,
      "step": 147
    },
    {
      "clip_ratio/high_max": 0.002981394987727981,
      "clip_ratio/high_mean": 0.0012385466507112142,
      "clip_ratio/low_mean": 0.0008997840905067278,
      "clip_ratio/low_min": 2.615828543639509e-05,
      "clip_ratio/region_mean": 0.0021383307830546983,
      "epoch": 2.186588921282799,
      "grad_norm": 0.12196576595306396,
      "learning_rate": 1e-06,
      "loss": -0.0092,
      "step": 148
    },
    {
      "clip_ratio/high_max": 0.002453926943417173,
      "clip_ratio/high_mean": 0.0010992179522872902,
      "clip_ratio/low_mean": 0.0010282481543981703,
      "clip_ratio/low_min": 2.8204158297739923e-05,
      "clip_ratio/region_mean": 0.0021274661121424288,
      "epoch": 2.195918367346939,
      "grad_norm": 0.11932481080293655,
      "learning_rate": 1e-06,
      "loss": -0.015,
      "step": 149
    },
    {
      "clip_ratio/high_max": 0.0022341832373058423,
      "clip_ratio/high_mean": 0.0010397633450338617,
      "clip_ratio/low_mean": 0.00110031434087432,
      "clip_ratio/low_min": 0.00010811330230353633,
      "clip_ratio/region_mean": 0.0021400777040980756,
      "epoch": 2.205247813411079,
      "grad_norm": 0.13031426072120667,
      "learning_rate": 1e-06,
      "loss": 0.0142,
      "step": 150
    },
    {
      "clip_ratio/high_max": 0.002100627498293761,
      "clip_ratio/high_mean": 0.0009728965815156698,
      "clip_ratio/low_mean": 0.001109271001041634,
      "clip_ratio/low_min": 0.00018426047336106421,
      "clip_ratio/region_mean": 0.0020821675861952826,
      "epoch": 2.2145772594752184,
      "grad_norm": 0.1201627254486084,
      "learning_rate": 1e-06,
      "loss": 0.0334,
      "step": 151
    },
    {
      "clip_ratio/high_max": 0.002272098499815911,
      "clip_ratio/high_mean": 0.0009493880716036074,
      "clip_ratio/low_mean": 0.0012250771833350882,
      "clip_ratio/low_min": 0.0001314482269663131,
      "clip_ratio/region_mean": 0.00217446521128295,
      "epoch": 2.2239067055393584,
      "grad_norm": 0.11559823900461197,
      "learning_rate": 1e-06,
      "loss": 0.0049,
      "step": 152
    },
    {
      "clip_ratio/high_max": 0.0023405511310556903,
      "clip_ratio/high_mean": 0.0010975574332405813,
      "clip_ratio/low_mean": 0.0011285115579084959,
      "clip_ratio/low_min": 0.00010950531759590376,
      "clip_ratio/region_mean": 0.0022260689511313103,
      "epoch": 2.2332361516034984,
      "grad_norm": 0.130247563123703,
      "learning_rate": 1e-06,
      "loss": 0.004,
      "step": 153
    },
    {
      "clip_ratio/high_max": 0.0025937355821952224,
      "clip_ratio/high_mean": 0.0010923339614237193,
      "clip_ratio/low_mean": 0.001415761093085166,
      "clip_ratio/low_min": 0.0001920558606798295,
      "clip_ratio/region_mean": 0.002508095058146864,
      "epoch": 2.2425655976676384,
      "grad_norm": 0.12192545086145401,
      "learning_rate": 1e-06,
      "loss": 0.0625,
      "step": 154
    },
    {
      "clip_ratio/high_max": 0.002497716595826205,
      "clip_ratio/high_mean": 0.0011746687414415646,
      "clip_ratio/low_mean": 0.0012749092384183314,
      "clip_ratio/low_min": 0.0001659430708969012,
      "clip_ratio/region_mean": 0.0024495779653079808,
      "epoch": 2.2518950437317784,
      "grad_norm": 0.1451082080602646,
      "learning_rate": 1e-06,
      "loss": 0.0093,
      "step": 155
    },
    {
      "clip_ratio/high_max": 0.002519707406463567,
      "clip_ratio/high_mean": 0.0011039472628908698,
      "clip_ratio/low_mean": 0.0011554806114872918,
      "clip_ratio/low_min": 6.777406906621763e-05,
      "clip_ratio/region_mean": 0.0022594279143959284,
      "epoch": 2.2612244897959184,
      "grad_norm": 0.12295163422822952,
      "learning_rate": 1e-06,
      "loss": -0.0061,
      "step": 156
    },
    {
      "clip_ratio/high_max": 0.0024124329356709495,
      "clip_ratio/high_mean": 0.0010858879431907553,
      "clip_ratio/low_mean": 0.0011228975854464807,
      "clip_ratio/low_min": 0.00012370092008495703,
      "clip_ratio/region_mean": 0.00220878554682713,
      "epoch": 2.2705539358600584,
      "grad_norm": 0.12332891672849655,
      "learning_rate": 1e-06,
      "loss": -0.0336,
      "step": 157
    },
    {
      "clip_ratio/high_max": 0.0023857227715780027,
      "clip_ratio/high_mean": 0.0010250295272271615,
      "clip_ratio/low_mean": 0.0011704927565006074,
      "clip_ratio/low_min": 0.0001437042319594184,
      "clip_ratio/region_mean": 0.002195522283727769,
      "epoch": 2.2798833819241984,
      "grad_norm": 0.12864260375499725,
      "learning_rate": 1e-06,
      "loss": 0.0402,
      "step": 158
    },
    {
      "clip_ratio/high_max": 0.0025151486552204005,
      "clip_ratio/high_mean": 0.0011685123172355816,
      "clip_ratio/low_mean": 0.0012465614381653722,
      "clip_ratio/low_min": 0.00018614440705277957,
      "clip_ratio/region_mean": 0.002415073788142763,
      "epoch": 2.2892128279883384,
      "grad_norm": 0.12569065392017365,
      "learning_rate": 1e-06,
      "loss": 0.0289,
      "step": 159
    },
    {
      "clip_ratio/high_max": 0.0025789209248614497,
      "clip_ratio/high_mean": 0.001108487436795258,
      "clip_ratio/low_mean": 0.0011688708073052112,
      "clip_ratio/low_min": 0.00015057267228257842,
      "clip_ratio/region_mean": 0.002277358216815628,
      "epoch": 2.298542274052478,
      "grad_norm": 0.12745057046413422,
      "learning_rate": 1e-06,
      "loss": -0.012,
      "step": 160
    },
    {
      "clip_ratio/high_max": 0.0026628526757122017,
      "clip_ratio/high_mean": 0.0010252606225549243,
      "clip_ratio/low_mean": 0.0010789855259645265,
      "clip_ratio/low_min": 0.00017235779523616657,
      "clip_ratio/region_mean": 0.0021042461157776415,
      "epoch": 2.307871720116618,
      "grad_norm": 0.11011394113302231,
      "learning_rate": 1e-06,
      "loss": -0.0114,
      "step": 161
    },
    {
      "clip_ratio/high_max": 0.0024281475925818086,
      "clip_ratio/high_mean": 0.0011097267633886077,
      "clip_ratio/low_mean": 0.001112452879169723,
      "clip_ratio/low_min": 7.253573585330741e-05,
      "clip_ratio/region_mean": 0.0022221796316443942,
      "epoch": 2.317201166180758,
      "grad_norm": 0.11440815031528473,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "step": 162
    },
    {
      "clip_ratio/high_max": 0.0027596873696893454,
      "clip_ratio/high_mean": 0.0012952804318047129,
      "clip_ratio/low_mean": 0.001032649219268933,
      "clip_ratio/low_min": 8.360037463717163e-05,
      "clip_ratio/region_mean": 0.0023279297311091796,
      "epoch": 2.326530612244898,
      "grad_norm": 0.13623954355716705,
      "learning_rate": 1e-06,
      "loss": -0.0408,
      "step": 163
    },
    {
      "clip_ratio/high_max": 0.0025571161750121973,
      "clip_ratio/high_mean": 0.0011863762665598188,
      "clip_ratio/low_mean": 0.0010855233413167298,
      "clip_ratio/low_min": 0.00010149566878681071,
      "clip_ratio/region_mean": 0.0022718996115145274,
      "epoch": 2.335860058309038,
      "grad_norm": 0.12146523594856262,
      "learning_rate": 1e-06,
      "loss": -0.009,
      "step": 164
    },
    {
      "clip_ratio/high_max": 0.0023731513720122166,
      "clip_ratio/high_mean": 0.0010484574231668375,
      "clip_ratio/low_mean": 0.0011893968248841702,
      "clip_ratio/low_min": 0.00021224252031970536,
      "clip_ratio/region_mean": 0.002237854241684545,
      "epoch": 2.345189504373178,
      "grad_norm": 0.12812286615371704,
      "learning_rate": 1e-06,
      "loss": 0.0263,
      "step": 165
    },
    {
      "clip_ratio/high_max": 0.002815638297761325,
      "clip_ratio/high_mean": 0.0012737084653053898,
      "clip_ratio/low_mean": 0.001155774309154367,
      "clip_ratio/low_min": 8.2797123468481e-05,
      "clip_ratio/region_mean": 0.00242948282539146,
      "epoch": 2.354518950437318,
      "grad_norm": 0.14721983671188354,
      "learning_rate": 1e-06,
      "loss": 0.0212,
      "step": 166
    },
    {
      "clip_ratio/high_max": 0.0026847347908187658,
      "clip_ratio/high_mean": 0.0011205359696759842,
      "clip_ratio/low_mean": 0.0013116085028741509,
      "clip_ratio/low_min": 0.00010896177991526201,
      "clip_ratio/region_mean": 0.002432144472550135,
      "epoch": 2.363848396501458,
      "grad_norm": 0.13157665729522705,
      "learning_rate": 1e-06,
      "loss": 0.0157,
      "step": 167
    },
    {
      "clip_ratio/high_max": 0.0032510058663319796,
      "clip_ratio/high_mean": 0.0013470837984641548,
      "clip_ratio/low_mean": 0.0011965052653977182,
      "clip_ratio/low_min": 0.00017041550654539606,
      "clip_ratio/region_mean": 0.0025435890711378306,
      "epoch": 2.373177842565598,
      "grad_norm": 0.12876376509666443,
      "learning_rate": 1e-06,
      "loss": -0.0324,
      "step": 168
    },
    {
      "clip_ratio/high_max": 0.002307305403519422,
      "clip_ratio/high_mean": 0.001048291931510903,
      "clip_ratio/low_mean": 0.001503323615906993,
      "clip_ratio/low_min": 0.0001809802015486639,
      "clip_ratio/region_mean": 0.002551615543779917,
      "epoch": 2.3825072886297374,
      "grad_norm": 0.11819365620613098,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "step": 169
    },
    {
      "clip_ratio/high_max": 0.0024908224149839953,
      "clip_ratio/high_mean": 0.001245201681740582,
      "clip_ratio/low_mean": 0.0011851935705635697,
      "clip_ratio/low_min": 6.285757990553975e-05,
      "clip_ratio/region_mean": 0.002430395245028194,
      "epoch": 2.3918367346938774,
      "grad_norm": 0.1131335124373436,
      "learning_rate": 1e-06,
      "loss": -0.0308,
      "step": 170
    },
    {
      "clip_ratio/high_max": 0.00256960164551856,
      "clip_ratio/high_mean": 0.0012161276208644267,
      "clip_ratio/low_mean": 0.0012400449450069573,
      "clip_ratio/low_min": 0.0001294164349019411,
      "clip_ratio/region_mean": 0.0024561725804232992,
      "epoch": 2.4011661807580174,
      "grad_norm": 0.13580158352851868,
      "learning_rate": 1e-06,
      "loss": -0.0075,
      "step": 171
    },
    {
      "clip_ratio/high_max": 0.0027758334763348103,
      "clip_ratio/high_mean": 0.001231379010278033,
      "clip_ratio/low_mean": 0.0015014320160844363,
      "clip_ratio/low_min": 0.00014884592201269697,
      "clip_ratio/region_mean": 0.002732811088208109,
      "epoch": 2.4104956268221573,
      "grad_norm": 0.12529368698596954,
      "learning_rate": 1e-06,
      "loss": 0.0067,
      "step": 172
    },
    {
      "clip_ratio/high_max": 0.002594049474282656,
      "clip_ratio/high_mean": 0.0012897534725198057,
      "clip_ratio/low_mean": 0.0014303950920293573,
      "clip_ratio/low_min": 0.00011217205792490859,
      "clip_ratio/region_mean": 0.0027201485718251206,
      "epoch": 2.4198250728862973,
      "grad_norm": 0.12434151768684387,
      "learning_rate": 1e-06,
      "loss": 0.021,
      "step": 173
    },
    {
      "clip_ratio/high_max": 0.002626042078190949,
      "clip_ratio/high_mean": 0.001195309447211912,
      "clip_ratio/low_mean": 0.0013356239687709603,
      "clip_ratio/low_min": 0.00014563699096470373,
      "clip_ratio/region_mean": 0.0025309334087069146,
      "epoch": 2.4291545189504373,
      "grad_norm": 0.12320226430892944,
      "learning_rate": 1e-06,
      "loss": -0.0107,
      "step": 174
    },
    {
      "clip_ratio/high_max": 0.0025353476012242027,
      "clip_ratio/high_mean": 0.0011494746577227488,
      "clip_ratio/low_mean": 0.0012201142781123053,
      "clip_ratio/low_min": 0.000194438808648556,
      "clip_ratio/region_mean": 0.0023695889103692025,
      "epoch": 2.4384839650145773,
      "grad_norm": 0.12133309245109558,
      "learning_rate": 1e-06,
      "loss": 0.0077,
      "step": 175
    },
    {
      "clip_ratio/high_max": 0.002551100442360621,
      "clip_ratio/high_mean": 0.0011013783114321996,
      "clip_ratio/low_mean": 0.001278181738598505,
      "clip_ratio/low_min": 0.00016299128856189782,
      "clip_ratio/region_mean": 0.0023795600136509165,
      "epoch": 2.4478134110787173,
      "grad_norm": 0.11697214841842651,
      "learning_rate": 1e-06,
      "loss": -0.011,
      "step": 176
    },
    {
      "clip_ratio/high_max": 0.002325334346096497,
      "clip_ratio/high_mean": 0.0011147658078698441,
      "clip_ratio/low_mean": 0.0010799937226693146,
      "clip_ratio/low_min": 4.906066351395566e-05,
      "clip_ratio/region_mean": 0.0021947595305391587,
      "epoch": 2.4571428571428573,
      "grad_norm": 0.12206520885229111,
      "learning_rate": 1e-06,
      "loss": -0.011,
      "step": 177
    },
    {
      "clip_ratio/high_max": 0.002367674111155793,
      "clip_ratio/high_mean": 0.001087657799871522,
      "clip_ratio/low_mean": 0.0012481089543143753,
      "clip_ratio/low_min": 0.0002468786715326132,
      "clip_ratio/region_mean": 0.002335766730539035,
      "epoch": 2.466472303206997,
      "grad_norm": 0.12972880899906158,
      "learning_rate": 1e-06,
      "loss": 0.0494,
      "step": 178
    },
    {
      "clip_ratio/high_max": 0.0028937013885297347,
      "clip_ratio/high_mean": 0.0012001742725260556,
      "clip_ratio/low_mean": 0.0011895055395143572,
      "clip_ratio/low_min": 0.00017757929708750453,
      "clip_ratio/region_mean": 0.0023896798520581797,
      "epoch": 2.4758017492711373,
      "grad_norm": 0.12528245151042938,
      "learning_rate": 1e-06,
      "loss": 0.0275,
      "step": 179
    },
    {
      "clip_ratio/high_max": 0.002909040755184833,
      "clip_ratio/high_mean": 0.0012443441919458564,
      "clip_ratio/low_mean": 0.0010223880271951202,
      "clip_ratio/low_min": 7.617511073476635e-05,
      "clip_ratio/region_mean": 0.00226673227007268,
      "epoch": 2.485131195335277,
      "grad_norm": 0.11673883348703384,
      "learning_rate": 1e-06,
      "loss": -0.0137,
      "step": 180
    },
    {
      "clip_ratio/high_max": 0.0027894804734387435,
      "clip_ratio/high_mean": 0.0013034097573836334,
      "clip_ratio/low_mean": 0.0012337905136519112,
      "clip_ratio/low_min": 0.00013576465244113933,
      "clip_ratio/region_mean": 0.002537200241931714,
      "epoch": 2.494460641399417,
      "grad_norm": 0.12992659211158752,
      "learning_rate": 1e-06,
      "loss": 0.0102,
      "step": 181
    },
    {
      "clip_ratio/high_max": 0.0029747471562586725,
      "clip_ratio/high_mean": 0.0013314384959812742,
      "clip_ratio/low_mean": 0.001018367001961451,
      "clip_ratio/low_min": 8.735746359889163e-05,
      "clip_ratio/region_mean": 0.0023498054724768735,
      "epoch": 2.503790087463557,
      "grad_norm": 0.12337908893823624,
      "learning_rate": 1e-06,
      "loss": -0.0315,
      "step": 182
    },
    {
      "clip_ratio/high_max": 0.002388678018178325,
      "clip_ratio/high_mean": 0.000977828683971893,
      "clip_ratio/low_mean": 0.0012462238591979258,
      "clip_ratio/low_min": 0.00012972266449651215,
      "clip_ratio/region_mean": 0.0022240525649976917,
      "epoch": 2.513119533527697,
      "grad_norm": 0.1161714568734169,
      "learning_rate": 1e-06,
      "loss": -0.0029,
      "step": 183
    },
    {
      "clip_ratio/high_max": 0.0027139371595694683,
      "clip_ratio/high_mean": 0.0011963951837969944,
      "clip_ratio/low_mean": 0.0014414133238460636,
      "clip_ratio/low_min": 0.0002481808433003607,
      "clip_ratio/region_mean": 0.0026378084148745984,
      "epoch": 2.522448979591837,
      "grad_norm": 0.12418004870414734,
      "learning_rate": 1e-06,
      "loss": 0.0223,
      "step": 184
    },
    {
      "clip_ratio/high_max": 0.002537755797675345,
      "clip_ratio/high_mean": 0.0011132990730402526,
      "clip_ratio/low_mean": 0.0012987907139176968,
      "clip_ratio/low_min": 0.00025258928053517593,
      "clip_ratio/region_mean": 0.0024120898087858222,
      "epoch": 2.5317784256559768,
      "grad_norm": 0.10932054370641708,
      "learning_rate": 1e-06,
      "loss": 0.0057,
      "step": 185
    },
    {
      "clip_ratio/high_max": 0.0025301558707724325,
      "clip_ratio/high_mean": 0.0010961110274365637,
      "clip_ratio/low_mean": 0.0012796354512829566,
      "clip_ratio/low_min": 9.25321837712545e-05,
      "clip_ratio/region_mean": 0.00237574648053851,
      "epoch": 2.5411078717201168,
      "grad_norm": 0.11127616465091705,
      "learning_rate": 1e-06,
      "loss": 0.0173,
      "step": 186
    },
    {
      "clip_ratio/high_max": 0.002816690379404463,
      "clip_ratio/high_mean": 0.00122662079957081,
      "clip_ratio/low_mean": 0.0013554430515796412,
      "clip_ratio/low_min": 0.00013329072771739447,
      "clip_ratio/region_mean": 0.0025820638693403453,
      "epoch": 2.5504373177842563,
      "grad_norm": 0.12166678160429001,
      "learning_rate": 1e-06,
      "loss": -0.0123,
      "step": 187
    },
    {
      "clip_ratio/high_max": 0.0027079341089120135,
      "clip_ratio/high_mean": 0.0012754563576891087,
      "clip_ratio/low_mean": 0.0013666823324456345,
      "clip_ratio/low_min": 0.00017096808824135223,
      "clip_ratio/region_mean": 0.002642138693772722,
      "epoch": 2.5597667638483967,
      "grad_norm": 0.12692616879940033,
      "learning_rate": 1e-06,
      "loss": -0.0342,
      "step": 188
    },
    {
      "clip_ratio/high_max": 0.002358461577387061,
      "clip_ratio/high_mean": 0.0011824131706816843,
      "clip_ratio/low_mean": 0.0011677085858536884,
      "clip_ratio/low_min": 7.566150998172816e-05,
      "clip_ratio/region_mean": 0.0023501217874581926,
      "epoch": 2.5690962099125363,
      "grad_norm": 0.12305018305778503,
      "learning_rate": 1e-06,
      "loss": -0.032,
      "step": 189
    },
    {
      "clip_ratio/high_max": 0.0023616322796442546,
      "clip_ratio/high_mean": 0.0011595011492318008,
      "clip_ratio/low_mean": 0.001299076937357313,
      "clip_ratio/low_min": 0.00024115154883475043,
      "clip_ratio/region_mean": 0.002458578055666294,
      "epoch": 2.5784256559766763,
      "grad_norm": 0.1243428960442543,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "step": 190
    },
    {
      "clip_ratio/high_max": 0.0028254646531422623,
      "clip_ratio/high_mean": 0.001123283529523178,
      "clip_ratio/low_mean": 0.001313371496507898,
      "clip_ratio/low_min": 0.00016924353076319676,
      "clip_ratio/region_mean": 0.0024366549914702773,
      "epoch": 2.5877551020408163,
      "grad_norm": 0.12276402860879898,
      "learning_rate": 1e-06,
      "loss": 0.0246,
      "step": 191
    },
    {
      "clip_ratio/high_max": 0.002761791678494774,
      "clip_ratio/high_mean": 0.0012255481524334755,
      "clip_ratio/low_mean": 0.0012073713296558708,
      "clip_ratio/low_min": 0.0001706942430246272,
      "clip_ratio/region_mean": 0.002432919565762859,
      "epoch": 2.5970845481049563,
      "grad_norm": 0.12500406801700592,
      "learning_rate": 1e-06,
      "loss": -0.0294,
      "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": 0.020106724330357095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4078.0,
      "completions/mean_length": 634.357177734375,
      "completions/mean_terminated_length": 563.32666015625,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 3.00932944606414,
      "grad_norm": 0.14026591181755066,
      "learning_rate": 1e-06,
      "loss": -0.0017,
      "num_tokens": 148897704.0,
      "reward": 0.5502232313156128,
      "reward_std": 0.21750041842460632,
      "rewards/simpleverify_reward/mean": 0.5502232313156128,
      "rewards/simpleverify_reward/std": 0.49747559428215027,
      "step": 193
    },
    {
      "clip_ratio/high_max": 0.002061748346022796,
      "clip_ratio/high_mean": 0.000871242256835103,
      "clip_ratio/low_mean": 0.0006570765872311313,
      "clip_ratio/low_min": 4.9448425670561846e-05,
      "clip_ratio/region_mean": 0.0015283188258763403,
      "epoch": 3.01865889212828,
      "grad_norm": 0.1298675686120987,
      "learning_rate": 1e-06,
      "loss": 0.069,
      "step": 194
    },
    {
      "clip_ratio/high_max": 0.0021520642694667913,
      "clip_ratio/high_mean": 0.0009898013740894385,
      "clip_ratio/low_mean": 0.0005682344035449205,
      "clip_ratio/low_min": 2.4069012397376355e-05,
      "clip_ratio/region_mean": 0.0015580357830913272,
      "epoch": 3.02798833819242,
      "grad_norm": 0.12610502541065216,
      "learning_rate": 1e-06,
      "loss": -0.0041,
      "step": 195
    },
    {
      "clip_ratio/high_max": 0.0024360017268918455,
      "clip_ratio/high_mean": 0.0010614840466587339,
      "clip_ratio/low_mean": 0.0006560986976182903,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017175827670143917,
      "epoch": 3.03731778425656,
      "grad_norm": 0.1263195127248764,
      "learning_rate": 1e-06,
      "loss": -0.0162,
      "step": 196
    },
    {
      "clip_ratio/high_max": 0.0024873623697203584,
      "clip_ratio/high_mean": 0.0010059822707262356,
      "clip_ratio/low_mean": 0.0006211861041265365,
      "clip_ratio/low_min": 5.088131820230046e-05,
      "clip_ratio/region_mean": 0.0016271683998638764,
      "epoch": 3.0466472303206995,
      "grad_norm": 0.12883612513542175,
      "learning_rate": 1e-06,
      "loss": 0.0115,
      "step": 197
    },
    {
      "clip_ratio/high_max": 0.0022831738024251536,
      "clip_ratio/high_mean": 0.0008958075559348799,
      "clip_ratio/low_mean": 0.0007660653791390359,
      "clip_ratio/low_min": 8.209856878238497e-05,
      "clip_ratio/region_mean": 0.0016618729350739159,
      "epoch": 3.0559766763848395,
      "grad_norm": 0.12781526148319244,
      "learning_rate": 1e-06,
      "loss": 0.1007,
      "step": 198
    },
    {
      "clip_ratio/high_max": 0.002296483340614941,
      "clip_ratio/high_mean": 0.0009581223021086771,
      "clip_ratio/low_mean": 0.0008079199051280739,
      "clip_ratio/low_min": 7.637055568920914e-05,
      "clip_ratio/region_mean": 0.0017660422236076556,
      "epoch": 3.0653061224489795,
      "grad_norm": 0.11599215865135193,
      "learning_rate": 1e-06,
      "loss": 0.0101,
      "step": 199
    },
    {
      "clip_ratio/high_max": 0.0023743426500004716,
      "clip_ratio/high_mean": 0.001073000101314392,
      "clip_ratio/low_mean": 0.000863245242726407,
      "clip_ratio/low_min": 5.384240830608178e-05,
      "clip_ratio/region_mean": 0.0019362453313078731,
      "epoch": 3.0746355685131195,
      "grad_norm": 0.1352672278881073,
      "learning_rate": 1e-06,
      "loss": 0.0064,
      "step": 200
    },
    {
      "clip_ratio/high_max": 0.002242562113679014,
      "clip_ratio/high_mean": 0.0009041177800099831,
      "clip_ratio/low_mean": 0.0008844915282679722,
      "clip_ratio/low_min": 7.005702718743123e-05,
      "clip_ratio/region_mean": 0.0017886093191918917,
      "epoch": 3.0839650145772595,
      "grad_norm": 0.11848468333482742,
      "learning_rate": 1e-06,
      "loss": 0.0065,
      "step": 201
    },
    {
      "clip_ratio/high_max": 0.0027339655498508364,
      "clip_ratio/high_mean": 0.0011076306145696435,
      "clip_ratio/low_mean": 0.0009033701608132105,
      "clip_ratio/low_min": 4.685640669777058e-05,
      "clip_ratio/region_mean": 0.002011000833590515,
      "epoch": 3.0932944606413995,
      "grad_norm": 0.11213834583759308,
      "learning_rate": 1e-06,
      "loss": -0.014,
      "step": 202
    },
    {
      "clip_ratio/high_max": 0.0023162214856711216,
      "clip_ratio/high_mean": 0.001049979320669081,
      "clip_ratio/low_mean": 0.0010175596817134647,
      "clip_ratio/low_min": 0.00015259493011399172,
      "clip_ratio/region_mean": 0.0020675390187534504,
      "epoch": 3.1026239067055394,
      "grad_norm": 0.12015374004840851,
      "learning_rate": 1e-06,
      "loss": 0.0422,
      "step": 203
    },
    {
      "clip_ratio/high_max": 0.0027470207278383896,
      "clip_ratio/high_mean": 0.001183155553007964,
      "clip_ratio/low_mean": 0.0008524330405634828,
      "clip_ratio/low_min": 6.070850304240594e-05,
      "clip_ratio/region_mean": 0.002035588629951235,
      "epoch": 3.1119533527696794,
      "grad_norm": 0.12526558339595795,
      "learning_rate": 1e-06,
      "loss": 0.0043,
      "step": 204
    },
    {
      "clip_ratio/high_max": 0.002381809303187765,
      "clip_ratio/high_mean": 0.0009171813799184747,
      "clip_ratio/low_mean": 0.0010008566132455599,
      "clip_ratio/low_min": 0.00011786398226831807,
      "clip_ratio/region_mean": 0.0019180379385943525,
      "epoch": 3.1212827988338194,
      "grad_norm": 0.11298627406358719,
      "learning_rate": 1e-06,
      "loss": 0.0409,
      "step": 205
    },
    {
      "clip_ratio/high_max": 0.0023509277816629037,
      "clip_ratio/high_mean": 0.0010854686042875983,
      "clip_ratio/low_mean": 0.0010403982960269786,
      "clip_ratio/low_min": 0.0001352437393506989,
      "clip_ratio/region_mean": 0.0021258669075905345,
      "epoch": 3.130612244897959,
      "grad_norm": 0.11448032408952713,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "step": 206
    },
    {
      "clip_ratio/high_max": 0.002645508779096417,
      "clip_ratio/high_mean": 0.0011063714482588693,
      "clip_ratio/low_mean": 0.0009564560987200821,
      "clip_ratio/low_min": 0.00013615141688205767,
      "clip_ratio/region_mean": 0.0020628275960916653,
      "epoch": 3.139941690962099,
      "grad_norm": 0.1367618441581726,
      "learning_rate": 1e-06,
      "loss": 0.0169,
      "step": 207
    },
    {
      "clip_ratio/high_max": 0.0025479232353973202,
      "clip_ratio/high_mean": 0.0012087199647794478,
      "clip_ratio/low_mean": 0.0009755447499628644,
      "clip_ratio/low_min": 0.00010132720126421191,
      "clip_ratio/region_mean": 0.0021842647329322062,
      "epoch": 3.149271137026239,
      "grad_norm": 0.1251148134469986,
      "learning_rate": 1e-06,
      "loss": -0.0055,
      "step": 208
    },
    {
      "clip_ratio/high_max": 0.0026119275571545586,
      "clip_ratio/high_mean": 0.0011381316871847957,
      "clip_ratio/low_mean": 0.0010236974267172627,
      "clip_ratio/low_min": 0.00011333390102663543,
      "clip_ratio/region_mean": 0.0021618290993501432,
      "epoch": 3.158600583090379,
      "grad_norm": 0.11783840507268906,
      "learning_rate": 1e-06,
      "loss": -0.0065,
      "step": 209
    },
    {
      "clip_ratio/high_max": 0.002686173000256531,
      "clip_ratio/high_mean": 0.0011166190597577952,
      "clip_ratio/low_mean": 0.0010319768443878274,
      "clip_ratio/low_min": 0.00015408775379910367,
      "clip_ratio/region_mean": 0.0021485959077836014,
      "epoch": 3.167930029154519,
      "grad_norm": 0.12133133411407471,
      "learning_rate": 1e-06,
      "loss": -0.0118,
      "step": 210
    },
    {
      "clip_ratio/high_max": 0.0021286107366904616,
      "clip_ratio/high_mean": 0.000993017772998428,
      "clip_ratio/low_mean": 0.0009988674282794818,
      "clip_ratio/low_min": 7.460783126589376e-05,
      "clip_ratio/region_mean": 0.0019918852412956767,
      "epoch": 3.177259475218659,
      "grad_norm": 0.13970410823822021,
      "learning_rate": 1e-06,
      "loss": 0.0438,
      "step": 211
    },
    {
      "clip_ratio/high_max": 0.0022555009491043165,
      "clip_ratio/high_mean": 0.0011183221104147378,
      "clip_ratio/low_mean": 0.0009317137592006475,
      "clip_ratio/low_min": 8.494884968968108e-05,
      "clip_ratio/region_mean": 0.0020500359023571946,
      "epoch": 3.186588921282799,
      "grad_norm": 0.11689542233943939,
      "learning_rate": 1e-06,
      "loss": -0.0057,
      "step": 212
    },
    {
      "clip_ratio/high_max": 0.0025477413946646266,
      "clip_ratio/high_mean": 0.0011465173229225911,
      "clip_ratio/low_mean": 0.0008799154056760017,
      "clip_ratio/low_min": 8.384949524042895e-05,
      "clip_ratio/region_mean": 0.002026432746788487,
      "epoch": 3.195918367346939,
      "grad_norm": 0.12078435719013214,
      "learning_rate": 1e-06,
      "loss": -0.0247,
      "step": 213
    },
    {
      "clip_ratio/high_max": 0.002567364979768172,
      "clip_ratio/high_mean": 0.0010069786167150596,
      "clip_ratio/low_mean": 0.0009920025204337435,
      "clip_ratio/low_min": 0.00011640755474218167,
      "clip_ratio/region_mean": 0.00199898117716657,
      "epoch": 3.205247813411079,
      "grad_norm": 0.12187445163726807,
      "learning_rate": 1e-06,
      "loss": 0.0415,
      "step": 214
    },
    {
      "clip_ratio/high_max": 0.002537337342801038,
      "clip_ratio/high_mean": 0.0011145730495627504,
      "clip_ratio/low_mean": 0.0008835631269903388,
      "clip_ratio/low_min": 5.051165771874366e-05,
      "clip_ratio/region_mean": 0.0019981361292593647,
      "epoch": 3.2145772594752184,
      "grad_norm": 0.1302180290222168,
      "learning_rate": 1e-06,
      "loss": -0.0072,
      "step": 215
    },
    {
      "clip_ratio/high_max": 0.002595059428131208,
      "clip_ratio/high_mean": 0.0011486775511002634,
      "clip_ratio/low_mean": 0.0009252007530449191,
      "clip_ratio/low_min": 5.344779128790833e-05,
      "clip_ratio/region_mean": 0.0020738783132401295,
      "epoch": 3.2239067055393584,
      "grad_norm": 0.11370299011468887,
      "learning_rate": 1e-06,
      "loss": -0.0113,
      "step": 216
    },
    {
      "clip_ratio/high_max": 0.0026351970300311223,
      "clip_ratio/high_mean": 0.0011887289365404285,
      "clip_ratio/low_mean": 0.0009616375009500189,
      "clip_ratio/low_min": 5.4788140005257446e-05,
      "clip_ratio/region_mean": 0.0021503664465853944,
      "epoch": 3.2332361516034984,
      "grad_norm": 0.12059260904788971,
      "learning_rate": 1e-06,
      "loss": -0.0097,
      "step": 217
    },
    {
      "clip_ratio/high_max": 0.0022073932123021223,
      "clip_ratio/high_mean": 0.0010123799802386202,
      "clip_ratio/low_mean": 0.0010926084069069475,
      "clip_ratio/low_min": 0.00010800341442518402,
      "clip_ratio/region_mean": 0.002104988401697483,
      "epoch": 3.2425655976676384,
      "grad_norm": 0.1268319934606552,
      "learning_rate": 1e-06,
      "loss": 0.0427,
      "step": 218
    },
    {
      "clip_ratio/high_max": 0.0024583201957284473,
      "clip_ratio/high_mean": 0.0012200099808978848,
      "clip_ratio/low_mean": 0.0009574005161994137,
      "clip_ratio/low_min": 8.215757043217309e-05,
      "clip_ratio/region_mean": 0.0021774104825453833,
      "epoch": 3.2518950437317784,
      "grad_norm": 0.12496983259916306,
      "learning_rate": 1e-06,
      "loss": -0.012,
      "step": 219
    },
    {
      "clip_ratio/high_max": 0.002703683465369977,
      "clip_ratio/high_mean": 0.0011089182080468163,
      "clip_ratio/low_mean": 0.0011269765527686104,
      "clip_ratio/low_min": 9.889045850286493e-05,
      "clip_ratio/region_mean": 0.0022358947971952148,
      "epoch": 3.2612244897959184,
      "grad_norm": 0.12382709234952927,
      "learning_rate": 1e-06,
      "loss": 0.0077,
      "step": 220
    },
    {
      "clip_ratio/high_max": 0.0023699304219917394,
      "clip_ratio/high_mean": 0.0010214009125775192,
      "clip_ratio/low_mean": 0.0012914672552142292,
      "clip_ratio/low_min": 0.00020172488621028606,
      "clip_ratio/region_mean": 0.0023128682223614305,
      "epoch": 3.2705539358600584,
      "grad_norm": 0.13156366348266602,
      "learning_rate": 1e-06,
      "loss": 0.0273,
      "step": 221
    },
    {
      "clip_ratio/high_max": 0.0021955612755846232,
      "clip_ratio/high_mean": 0.001025475967253442,
      "clip_ratio/low_mean": 0.0012421034098224482,
      "clip_ratio/low_min": 0.00019581201377150137,
      "clip_ratio/region_mean": 0.0022675793879898265,
      "epoch": 3.2798833819241984,
      "grad_norm": 0.1161755695939064,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "step": 222
    },
    {
      "clip_ratio/high_max": 0.0029818642142345197,
      "clip_ratio/high_mean": 0.001216008182382211,
      "clip_ratio/low_mean": 0.001189883601909969,
      "clip_ratio/low_min": 0.0001064189327735221,
      "clip_ratio/region_mean": 0.0024058917915681377,
      "epoch": 3.2892128279883384,
      "grad_norm": 0.12272651493549347,
      "learning_rate": 1e-06,
      "loss": -0.0349,
      "step": 223
    },
    {
      "clip_ratio/high_max": 0.002536017753300257,
      "clip_ratio/high_mean": 0.0010533328058954794,
      "clip_ratio/low_mean": 0.0012618183500308078,
      "clip_ratio/low_min": 6.540621598105645e-05,
      "clip_ratio/region_mean": 0.0023151511340984143,
      "epoch": 3.298542274052478,
      "grad_norm": 0.12759383022785187,
      "learning_rate": 1e-06,
      "loss": -0.0036,
      "step": 224
    },
    {
      "clip_ratio/high_max": 0.00248715727502713,
      "clip_ratio/high_mean": 0.0012340993707766756,
      "clip_ratio/low_mean": 0.001422618228389183,
      "clip_ratio/low_min": 0.0003430717397350236,
      "clip_ratio/region_mean": 0.002656717610079795,
      "epoch": 3.307871720116618,
      "grad_norm": 0.12538866698741913,
      "learning_rate": 1e-06,
      "loss": -0.015,
      "step": 225
    },
    {
      "clip_ratio/high_max": 0.002855788894521538,
      "clip_ratio/high_mean": 0.0011752312384487595,
      "clip_ratio/low_mean": 0.0013954976457171142,
      "clip_ratio/low_min": 0.00013497537202056265,
      "clip_ratio/region_mean": 0.0025707288732519373,
      "epoch": 3.317201166180758,
      "grad_norm": 0.2173781841993332,
      "learning_rate": 1e-06,
      "loss": 0.0119,
      "step": 226
    },
    {
      "clip_ratio/high_max": 0.0023392430739477277,
      "clip_ratio/high_mean": 0.001059344613167923,
      "clip_ratio/low_mean": 0.0013196719337429386,
      "clip_ratio/low_min": 0.00021052706506452523,
      "clip_ratio/region_mean": 0.0023790165141690522,
      "epoch": 3.326530612244898,
      "grad_norm": 0.12303262948989868,
      "learning_rate": 1e-06,
      "loss": 0.0181,
      "step": 227
    },
    {
      "clip_ratio/high_max": 0.002634392280015163,
      "clip_ratio/high_mean": 0.0011596684671530966,
      "clip_ratio/low_mean": 0.001220979800564237,
      "clip_ratio/low_min": 0.00014960934277041815,
      "clip_ratio/region_mean": 0.0023806482640793547,
      "epoch": 3.335860058309038,
      "grad_norm": 0.12350213527679443,
      "learning_rate": 1e-06,
      "loss": 0.0138,
      "step": 228
    },
    {
      "clip_ratio/high_max": 0.0027777489740401506,
      "clip_ratio/high_mean": 0.0012519868578237947,
      "clip_ratio/low_mean": 0.0010895985979004763,
      "clip_ratio/low_min": 0.00019499014160828665,
      "clip_ratio/region_mean": 0.002341585510293953,
      "epoch": 3.345189504373178,
      "grad_norm": 0.12810732424259186,
      "learning_rate": 1e-06,
      "loss": 0.0115,
      "step": 229
    },
    {
      "clip_ratio/high_max": 0.0023193202359834686,
      "clip_ratio/high_mean": 0.0009734930645208806,
      "clip_ratio/low_mean": 0.0009347590639663395,
      "clip_ratio/low_min": 9.29479674596223e-05,
      "clip_ratio/region_mean": 0.0019082521539530717,
      "epoch": 3.354518950437318,
      "grad_norm": 0.12010464072227478,
      "learning_rate": 1e-06,
      "loss": -0.0183,
      "step": 230
    },
    {
      "clip_ratio/high_max": 0.0024950333499873523,
      "clip_ratio/high_mean": 0.0011771378412959166,
      "clip_ratio/low_mean": 0.0011222091561648995,
      "clip_ratio/low_min": 0.00018253801681566983,
      "clip_ratio/region_mean": 0.002299347033840604,
      "epoch": 3.363848396501458,
      "grad_norm": 0.12164637446403503,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "step": 231
    },
    {
      "clip_ratio/high_max": 0.0025469671163591556,
      "clip_ratio/high_mean": 0.001141063494287664,
      "clip_ratio/low_mean": 0.0010985283624904696,
      "clip_ratio/low_min": 0.0001385923442285275,
      "clip_ratio/region_mean": 0.0022395918422262184,
      "epoch": 3.373177842565598,
      "grad_norm": 0.11820098012685776,
      "learning_rate": 1e-06,
      "loss": 0.0093,
      "step": 232
    },
    {
      "clip_ratio/high_max": 0.0025209588202415034,
      "clip_ratio/high_mean": 0.000971782978012925,
      "clip_ratio/low_mean": 0.0012674398058152292,
      "clip_ratio/low_min": 0.00016506315932929283,
      "clip_ratio/region_mean": 0.0022392228347598575,
      "epoch": 3.3825072886297374,
      "grad_norm": 0.1263270229101181,
      "learning_rate": 1e-06,
      "loss": 0.0484,
      "step": 233
    },
    {
      "clip_ratio/high_max": 0.0025999623903771862,
      "clip_ratio/high_mean": 0.0010336642844777089,
      "clip_ratio/low_mean": 0.0012233874804223888,
      "clip_ratio/low_min": 0.0001128667072407552,
      "clip_ratio/region_mean": 0.002257051761262119,
      "epoch": 3.3918367346938774,
      "grad_norm": 0.1158517450094223,
      "learning_rate": 1e-06,
      "loss": -0.0161,
      "step": 234
    },
    {
      "clip_ratio/high_max": 0.0027295962936477736,
      "clip_ratio/high_mean": 0.0011273406016698573,
      "clip_ratio/low_mean": 0.0012824184886994772,
      "clip_ratio/low_min": 5.478709499584511e-05,
      "clip_ratio/region_mean": 0.002409759115835186,
      "epoch": 3.4011661807580174,
      "grad_norm": 0.12060651928186417,
      "learning_rate": 1e-06,
      "loss": 0.006,
      "step": 235
    },
    {
      "clip_ratio/high_max": 0.002363648010941688,
      "clip_ratio/high_mean": 0.0010719535202952102,
      "clip_ratio/low_mean": 0.0012619421759154648,
      "clip_ratio/low_min": 0.00010520220712351147,
      "clip_ratio/region_mean": 0.00233389571076259,
      "epoch": 3.4104956268221573,
      "grad_norm": 0.11047002673149109,
      "learning_rate": 1e-06,
      "loss": 0.0145,
      "step": 236
    },
    {
      "clip_ratio/high_max": 0.0028598464632523246,
      "clip_ratio/high_mean": 0.0012210337044962216,
      "clip_ratio/low_mean": 0.0014538453215209302,
      "clip_ratio/low_min": 0.00013639724420499988,
      "clip_ratio/region_mean": 0.0026748790478450246,
      "epoch": 3.4198250728862973,
      "grad_norm": 0.12310261279344559,
      "learning_rate": 1e-06,
      "loss": -0.0097,
      "step": 237
    },
    {
      "clip_ratio/high_max": 0.003119135377346538,
      "clip_ratio/high_mean": 0.0013212138073868118,
      "clip_ratio/low_mean": 0.0012602354945556726,
      "clip_ratio/low_min": 4.994714799977373e-05,
      "clip_ratio/region_mean": 0.002581449312856421,
      "epoch": 3.4291545189504373,
      "grad_norm": 0.12614542245864868,
      "learning_rate": 1e-06,
      "loss": -0.0207,
      "step": 238
    },
    {
      "clip_ratio/high_max": 0.002636818950122688,
      "clip_ratio/high_mean": 0.0011825242509075906,
      "clip_ratio/low_mean": 0.0013021347840549424,
      "clip_ratio/low_min": 0.0001502478298789356,
      "clip_ratio/region_mean": 0.002484659016772639,
      "epoch": 3.4384839650145773,
      "grad_norm": 0.11874942481517792,
      "learning_rate": 1e-06,
      "loss": -0.0475,
      "step": 239
    },
    {
      "clip_ratio/high_max": 0.00306760219973512,
      "clip_ratio/high_mean": 0.0014245340571505949,
      "clip_ratio/low_mean": 0.0015436406611115672,
      "clip_ratio/low_min": 0.0002461949588905554,
      "clip_ratio/region_mean": 0.0029681747837457806,
      "epoch": 3.4478134110787173,
      "grad_norm": 0.1389058381319046,
      "learning_rate": 1e-06,
      "loss": -0.0446,
      "step": 240
    },
    {
      "clip_ratio/high_max": 0.002759681534371339,
      "clip_ratio/high_mean": 0.0011857538902404485,
      "clip_ratio/low_mean": 0.0013369897751545068,
      "clip_ratio/low_min": 0.00013312562623468693,
      "clip_ratio/region_mean": 0.0025227436490240507,
      "epoch": 3.4571428571428573,
      "grad_norm": 0.12311290204524994,
      "learning_rate": 1e-06,
      "loss": -0.0382,
      "step": 241
    },
    {
      "clip_ratio/high_max": 0.0029041526722721756,
      "clip_ratio/high_mean": 0.0012738214209093712,
      "clip_ratio/low_mean": 0.0013352936330193188,
      "clip_ratio/low_min": 0.00010003879469877575,
      "clip_ratio/region_mean": 0.002609115093946457,
      "epoch": 3.466472303206997,
      "grad_norm": 0.1270136684179306,
      "learning_rate": 1e-06,
      "loss": 0.0007,
      "step": 242
    },
    {
      "clip_ratio/high_max": 0.002884062720113434,
      "clip_ratio/high_mean": 0.0012906028241559397,
      "clip_ratio/low_mean": 0.0011787245639425237,
      "clip_ratio/low_min": 6.64215040160343e-05,
      "clip_ratio/region_mean": 0.002469327358994633,
      "epoch": 3.4758017492711373,
      "grad_norm": 0.11275164037942886,
      "learning_rate": 1e-06,
      "loss": -0.0278,
      "step": 243
    },
    {
      "clip_ratio/high_max": 0.0024617731032776646,
      "clip_ratio/high_mean": 0.0012163817264081445,
      "clip_ratio/low_mean": 0.0012092451579519548,
      "clip_ratio/low_min": 0.00017671631940174848,
      "clip_ratio/region_mean": 0.002425626909825951,
      "epoch": 3.485131195335277,
      "grad_norm": 0.12209966778755188,
      "learning_rate": 1e-06,
      "loss": -0.014,
      "step": 244
    },
    {
      "clip_ratio/high_max": 0.0024834484211169183,
      "clip_ratio/high_mean": 0.0011175874242326245,
      "clip_ratio/low_mean": 0.0012580749589687912,
      "clip_ratio/low_min": 0.00012665904250752646,
      "clip_ratio/region_mean": 0.002375662421400193,
      "epoch": 3.494460641399417,
      "grad_norm": 0.13395462930202484,
      "learning_rate": 1e-06,
      "loss": 0.0237,
      "step": 245
    },
    {
      "clip_ratio/high_max": 0.003192849879269488,
      "clip_ratio/high_mean": 0.0013443793941405602,
      "clip_ratio/low_mean": 0.0013164814190531615,
      "clip_ratio/low_min": 0.00011386331061657984,
      "clip_ratio/region_mean": 0.002660860765899997,
      "epoch": 3.503790087463557,
      "grad_norm": 0.1278659701347351,
      "learning_rate": 1e-06,
      "loss": -0.0339,
      "step": 246
    },
    {
      "clip_ratio/high_max": 0.002776773377263453,
      "clip_ratio/high_mean": 0.0012837477552238852,
      "clip_ratio/low_mean": 0.001044681812345516,
      "clip_ratio/low_min": 1.19434362204629e-05,
      "clip_ratio/region_mean": 0.002328429589397274,
      "epoch": 3.513119533527697,
      "grad_norm": 0.12169931828975677,
      "learning_rate": 1e-06,
      "loss": -0.0478,
      "step": 247
    },
    {
      "clip_ratio/high_max": 0.002933812269475311,
      "clip_ratio/high_mean": 0.00136197380561498,
      "clip_ratio/low_mean": 0.0010882375408982625,
      "clip_ratio/low_min": 8.942526983446442e-05,
      "clip_ratio/region_mean": 0.0024502113883499987,
      "epoch": 3.522448979591837,
      "grad_norm": 0.13563691079616547,
      "learning_rate": 1e-06,
      "loss": -0.0443,
      "step": 248
    },
    {
      "clip_ratio/high_max": 0.002817699161823839,
      "clip_ratio/high_mean": 0.0012116857615183108,
      "clip_ratio/low_mean": 0.0011869966874655802,
      "clip_ratio/low_min": 0.00013031674552621553,
      "clip_ratio/region_mean": 0.0023986824016901664,
      "epoch": 3.5317784256559768,
      "grad_norm": 0.11663589626550674,
      "learning_rate": 1e-06,
      "loss": -0.0086,
      "step": 249
    },
    {
      "clip_ratio/high_max": 0.0026597510805004276,
      "clip_ratio/high_mean": 0.001287872190005146,
      "clip_ratio/low_mean": 0.0011019548092008336,
      "clip_ratio/low_min": 9.463294645684073e-05,
      "clip_ratio/region_mean": 0.0023898269719211385,
      "epoch": 3.5411078717201168,
      "grad_norm": 0.1313488930463791,
      "learning_rate": 1e-06,
      "loss": -0.0176,
      "step": 250
    },
    {
      "clip_ratio/high_max": 0.0026479246371309273,
      "clip_ratio/high_mean": 0.001119159955123905,
      "clip_ratio/low_mean": 0.0012401145868352614,
      "clip_ratio/low_min": 0.0001741686228342587,
      "clip_ratio/region_mean": 0.0023592745055793785,
      "epoch": 3.5504373177842563,
      "grad_norm": 0.12081325799226761,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 251
    },
    {
      "clip_ratio/high_max": 0.002332132924493635,
      "clip_ratio/high_mean": 0.0010688346010283567,
      "clip_ratio/low_mean": 0.0015507310672546737,
      "clip_ratio/low_min": 5.5538073866046034e-05,
      "clip_ratio/region_mean": 0.0026195656246272847,
      "epoch": 3.5597667638483967,
      "grad_norm": 0.11730310320854187,
      "learning_rate": 1e-06,
      "loss": 0.0101,
      "step": 252
    },
    {
      "clip_ratio/high_max": 0.002911170980951283,
      "clip_ratio/high_mean": 0.0013587812572950497,
      "clip_ratio/low_mean": 0.001245073573954869,
      "clip_ratio/low_min": 8.216908281610813e-05,
      "clip_ratio/region_mean": 0.002603854867629707,
      "epoch": 3.5690962099125363,
      "grad_norm": 0.1244642436504364,
      "learning_rate": 1e-06,
      "loss": -0.0078,
      "step": 253
    },
    {
      "clip_ratio/high_max": 0.002901366307924036,
      "clip_ratio/high_mean": 0.0012334625025687274,
      "clip_ratio/low_mean": 0.0014136367390165105,
      "clip_ratio/low_min": 0.00012771209549100604,
      "clip_ratio/region_mean": 0.0026470992161193863,
      "epoch": 3.5784256559766763,
      "grad_norm": 0.12591688334941864,
      "learning_rate": 1e-06,
      "loss": 0.0308,
      "step": 254
    },
    {
      "clip_ratio/high_max": 0.0023122161510400474,
      "clip_ratio/high_mean": 0.001132892728492152,
      "clip_ratio/low_mean": 0.0013907535576436203,
      "clip_ratio/low_min": 0.00016566381236771122,
      "clip_ratio/region_mean": 0.0025236463407054543,
      "epoch": 3.5877551020408163,
      "grad_norm": 0.12184864282608032,
      "learning_rate": 1e-06,
      "loss": -0.0038,
      "step": 255
    },
    {
      "clip_ratio/high_max": 0.0026201951259281486,
      "clip_ratio/high_mean": 0.0011652265438897302,
      "clip_ratio/low_mean": 0.001409195323503809,
      "clip_ratio/low_min": 0.00011419090697017964,
      "clip_ratio/region_mean": 0.0025744218146428466,
      "epoch": 3.5970845481049563,
      "grad_norm": 0.11056038737297058,
      "learning_rate": 1e-06,
      "loss": 0.0087,
      "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": 0.0198451450892857,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4069.0,
      "completions/mean_length": 633.3449096679688,
      "completions/mean_terminated_length": 563.2366943359375,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 4.0093294460641395,
      "grad_norm": 0.12988856434822083,
      "learning_rate": 1e-06,
      "loss": 0.013,
      "num_tokens": 186267505.0,
      "reward": 0.5701032876968384,
      "reward_std": 0.20984646677970886,
      "rewards/simpleverify_reward/mean": 0.5701032280921936,
      "rewards/simpleverify_reward/std": 0.4950654208660126,
      "step": 257
    },
    {
      "clip_ratio/high_max": 0.0020908534424961545,
      "clip_ratio/high_mean": 0.0009248588939954061,
      "clip_ratio/low_mean": 0.0006118935616541421,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015367524756584316,
      "epoch": 4.01865889212828,
      "grad_norm": 0.11844545602798462,
      "learning_rate": 1e-06,
      "loss": -0.0163,
      "step": 258
    },
    {
      "clip_ratio/high_max": 0.0024169725002138875,
      "clip_ratio/high_mean": 0.0009596188428986352,
      "clip_ratio/low_mean": 0.0006339341798593523,
      "clip_ratio/low_min": 5.774290275439853e-05,
      "clip_ratio/region_mean": 0.0015935530100250617,
      "epoch": 4.0279883381924195,
      "grad_norm": 0.12983858585357666,
      "learning_rate": 1e-06,
      "loss": 0.0301,
      "step": 259
    },
    {
      "clip_ratio/high_max": 0.0021098739161971025,
      "clip_ratio/high_mean": 0.0009157013937510783,
      "clip_ratio/low_mean": 0.0005333022372724372,
      "clip_ratio/low_min": 2.7551244784262963e-05,
      "clip_ratio/region_mean": 0.0014490036192000844,
      "epoch": 4.03731778425656,
      "grad_norm": 0.1142122820019722,
      "learning_rate": 1e-06,
      "loss": 0.0201,
      "step": 260
    },
    {
      "clip_ratio/high_max": 0.0025062408021767624,
      "clip_ratio/high_mean": 0.0010443536357342964,
      "clip_ratio/low_mean": 0.0006596801667910768,
      "clip_ratio/low_min": 4.3925439968006685e-05,
      "clip_ratio/region_mean": 0.0017040337916114368,
      "epoch": 4.0466472303206995,
      "grad_norm": 0.11819064617156982,
      "learning_rate": 1e-06,
      "loss": 0.0198,
      "step": 261
    },
    {
      "clip_ratio/high_max": 0.002299812222190667,
      "clip_ratio/high_mean": 0.0008828037225612206,
      "clip_ratio/low_mean": 0.0006577786462003132,
      "clip_ratio/low_min": 1.6408506780862808e-05,
      "clip_ratio/region_mean": 0.0015405823396577034,
      "epoch": 4.05597667638484,
      "grad_norm": 0.10681001096963882,
      "learning_rate": 1e-06,
      "loss": 0.0042,
      "step": 262
    },
    {
      "clip_ratio/high_max": 0.002206483230111189,
      "clip_ratio/high_mean": 0.0009063853976840619,
      "clip_ratio/low_mean": 0.0007963443276821636,
      "clip_ratio/low_min": 4.1082725147134624e-05,
      "clip_ratio/region_mean": 0.0017027296926244162,
      "epoch": 4.0653061224489795,
      "grad_norm": 0.11958717554807663,
      "learning_rate": 1e-06,
      "loss": 0.0314,
      "step": 263
    },
    {
      "clip_ratio/high_max": 0.002453448105370626,
      "clip_ratio/high_mean": 0.000957541968091391,
      "clip_ratio/low_mean": 0.0006116921285865828,
      "clip_ratio/low_min": 1.8647438992047682e-05,
      "clip_ratio/region_mean": 0.0015692340821260586,
      "epoch": 4.07463556851312,
      "grad_norm": 0.11270119994878769,
      "learning_rate": 1e-06,
      "loss": -0.0197,
      "step": 264
    },
    {
      "clip_ratio/high_max": 0.0022682247654302046,
      "clip_ratio/high_mean": 0.0010995151060342323,
      "clip_ratio/low_mean": 0.0008586617605033098,
      "clip_ratio/low_min": 3.056699097214732e-05,
      "clip_ratio/region_mean": 0.0019581768137868494,
      "epoch": 4.0839650145772595,
      "grad_norm": 0.12431418895721436,
      "learning_rate": 1e-06,
      "loss": -0.0429,
      "step": 265
    },
    {
      "clip_ratio/high_max": 0.0021887957336730324,
      "clip_ratio/high_mean": 0.0008886759751476347,
      "clip_ratio/low_mean": 0.000879854835147853,
      "clip_ratio/low_min": 0.00010521688454900868,
      "clip_ratio/region_mean": 0.0017685308412183076,
      "epoch": 4.093294460641399,
      "grad_norm": 0.11466091871261597,
      "learning_rate": 1e-06,
      "loss": 0.0031,
      "step": 266
    },
    {
      "clip_ratio/high_max": 0.002351123606786132,
      "clip_ratio/high_mean": 0.0009354367393825669,
      "clip_ratio/low_mean": 0.0008512701897416264,
      "clip_ratio/low_min": 1.6979081919998862e-05,
      "clip_ratio/region_mean": 0.0017867069254862145,
      "epoch": 4.1026239067055394,
      "grad_norm": 0.1345750391483307,
      "learning_rate": 1e-06,
      "loss": 0.0144,
      "step": 267
    },
    {
      "clip_ratio/high_max": 0.00229811549797887,
      "clip_ratio/high_mean": 0.0010024150597018888,
      "clip_ratio/low_mean": 0.001148469036706956,
      "clip_ratio/low_min": 7.296910644072341e-05,
      "clip_ratio/region_mean": 0.002150884138245601,
      "epoch": 4.111953352769679,
      "grad_norm": 0.12159934639930725,
      "learning_rate": 1e-06,
      "loss": 0.0448,
      "step": 268
    },
    {
      "clip_ratio/high_max": 0.002365124732023105,
      "clip_ratio/high_mean": 0.0009946237587428186,
      "clip_ratio/low_mean": 0.000946328169447952,
      "clip_ratio/low_min": 1.065098877006676e-05,
      "clip_ratio/region_mean": 0.0019409519154578447,
      "epoch": 4.121282798833819,
      "grad_norm": 0.12911644577980042,
      "learning_rate": 1e-06,
      "loss": 0.0135,
      "step": 269
    },
    {
      "clip_ratio/high_max": 0.002485420598532073,
      "clip_ratio/high_mean": 0.0010724675994424615,
      "clip_ratio/low_mean": 0.000984783640888054,
      "clip_ratio/low_min": 4.1116952161246445e-05,
      "clip_ratio/region_mean": 0.0020572512185026426,
      "epoch": 4.130612244897959,
      "grad_norm": 0.12297946214675903,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 270
    },
    {
      "clip_ratio/high_max": 0.0020749176692334004,
      "clip_ratio/high_mean": 0.0009312258298450615,
      "clip_ratio/low_mean": 0.0009791130141820759,
      "clip_ratio/low_min": 4.269963847036706e-05,
      "clip_ratio/region_mean": 0.0019103388767689466,
      "epoch": 4.139941690962099,
      "grad_norm": 0.10888082534074783,
      "learning_rate": 1e-06,
      "loss": 0.0142,
      "step": 271
    },
    {
      "clip_ratio/high_max": 0.002122955716913566,
      "clip_ratio/high_mean": 0.0009231488220393658,
      "clip_ratio/low_mean": 0.0008404498512391001,
      "clip_ratio/low_min": 1.2605889423866756e-05,
      "clip_ratio/region_mean": 0.001763598651450593,
      "epoch": 4.149271137026239,
      "grad_norm": 0.11918072402477264,
      "learning_rate": 1e-06,
      "loss": -0.0408,
      "step": 272
    },
    {
      "clip_ratio/high_max": 0.002558688349381555,
      "clip_ratio/high_mean": 0.0010125796143256593,
      "clip_ratio/low_mean": 0.0010489000196685083,
      "clip_ratio/low_min": 4.142809302720707e-05,
      "clip_ratio/region_mean": 0.002061479608528316,
      "epoch": 4.158600583090379,
      "grad_norm": 0.12786640226840973,
      "learning_rate": 1e-06,
      "loss": -0.0022,
      "step": 273
    },
    {
      "clip_ratio/high_max": 0.002730152144067688,
      "clip_ratio/high_mean": 0.0010852359282580437,
      "clip_ratio/low_mean": 0.0011555486598808784,
      "clip_ratio/low_min": 9.449987373955082e-05,
      "clip_ratio/region_mean": 0.0022407845754059963,
      "epoch": 4.167930029154519,
      "grad_norm": 0.12328851222991943,
      "learning_rate": 1e-06,
      "loss": -0.0073,
      "step": 274
    },
    {
      "clip_ratio/high_max": 0.002337503588933032,
      "clip_ratio/high_mean": 0.0011032946422346868,
      "clip_ratio/low_mean": 0.0010695339278754545,
      "clip_ratio/low_min": 9.362400578538654e-05,
      "clip_ratio/region_mean": 0.002172828622860834,
      "epoch": 4.1772594752186585,
      "grad_norm": 0.1283477246761322,
      "learning_rate": 1e-06,
      "loss": -0.0113,
      "step": 275
    },
    {
      "clip_ratio/high_max": 0.0024091437298920937,
      "clip_ratio/high_mean": 0.0010472331050550565,
      "clip_ratio/low_mean": 0.001251158282684628,
      "clip_ratio/low_min": 9.006630898511503e-05,
      "clip_ratio/region_mean": 0.002298391453223303,
      "epoch": 4.186588921282799,
      "grad_norm": 0.12439129501581192,
      "learning_rate": 1e-06,
      "loss": 0.0253,
      "step": 276
    },
    {
      "clip_ratio/high_max": 0.002555749480961822,
      "clip_ratio/high_mean": 0.001136602833867073,
      "clip_ratio/low_mean": 0.00109997466825007,
      "clip_ratio/low_min": 0.00010164565719605889,
      "clip_ratio/region_mean": 0.0022365774857462384,
      "epoch": 4.1959183673469385,
      "grad_norm": 0.1284770518541336,
      "learning_rate": 1e-06,
      "loss": -0.0316,
      "step": 277
    },
    {
      "clip_ratio/high_max": 0.002383962339081336,
      "clip_ratio/high_mean": 0.0010050461278297007,
      "clip_ratio/low_mean": 0.0012296343520574737,
      "clip_ratio/low_min": 0.00013423965629044687,
      "clip_ratio/region_mean": 0.002234680468973238,
      "epoch": 4.205247813411079,
      "grad_norm": 0.1269519329071045,
      "learning_rate": 1e-06,
      "loss": 0.0315,
      "step": 278
    },
    {
      "clip_ratio/high_max": 0.0025625654016039334,
      "clip_ratio/high_mean": 0.0010282204348186497,
      "clip_ratio/low_mean": 0.0011528983050084207,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002181118747103028,
      "epoch": 4.214577259475218,
      "grad_norm": 0.11142841726541519,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 279
    },
    {
      "clip_ratio/high_max": 0.0028328740372671746,
      "clip_ratio/high_mean": 0.001188128771900665,
      "clip_ratio/low_mean": 0.001024641631374834,
      "clip_ratio/low_min": 0.00012055297520419117,
      "clip_ratio/region_mean": 0.002212770385085605,
      "epoch": 4.223906705539359,
      "grad_norm": 0.12641702592372894,
      "learning_rate": 1e-06,
      "loss": -0.0712,
      "step": 280
    },
    {
      "clip_ratio/high_max": 0.002491969891707413,
      "clip_ratio/high_mean": 0.0011266416040598415,
      "clip_ratio/low_mean": 0.0011804221103375312,
      "clip_ratio/low_min": 7.797056605340913e-05,
      "clip_ratio/region_mean": 0.002307063725311309,
      "epoch": 4.233236151603498,
      "grad_norm": 0.12074771523475647,
      "learning_rate": 1e-06,
      "loss": -0.0392,
      "step": 281
    },
    {
      "clip_ratio/high_max": 0.00244262828346109,
      "clip_ratio/high_mean": 0.0009119560554609052,
      "clip_ratio/low_mean": 0.0011744700714189094,
      "clip_ratio/low_min": 0.00022259577690419974,
      "clip_ratio/region_mean": 0.0020864261059614364,
      "epoch": 4.242565597667639,
      "grad_norm": 0.11982996016740799,
      "learning_rate": 1e-06,
      "loss": 0.0058,
      "step": 282
    },
    {
      "clip_ratio/high_max": 0.001973771104530897,
      "clip_ratio/high_mean": 0.0009037029594765045,
      "clip_ratio/low_mean": 0.0014119122315605637,
      "clip_ratio/low_min": 5.4030817409511656e-05,
      "clip_ratio/region_mean": 0.002315615158295259,
      "epoch": 4.251895043731778,
      "grad_norm": 0.11924701929092407,
      "learning_rate": 1e-06,
      "loss": 0.0379,
      "step": 283
    },
    {
      "clip_ratio/high_max": 0.002192493964685127,
      "clip_ratio/high_mean": 0.0009306118699896615,
      "clip_ratio/low_mean": 0.0010762799756776076,
      "clip_ratio/low_min": 0.00012505865015555173,
      "clip_ratio/region_mean": 0.002006891882047057,
      "epoch": 4.261224489795918,
      "grad_norm": 0.11580277979373932,
      "learning_rate": 1e-06,
      "loss": 0.026,
      "step": 284
    },
    {
      "clip_ratio/high_max": 0.002154302259441465,
      "clip_ratio/high_mean": 0.0010459450022608507,
      "clip_ratio/low_mean": 0.0012652925979637075,
      "clip_ratio/low_min": 0.000134850743052084,
      "clip_ratio/region_mean": 0.002311237607500516,
      "epoch": 4.270553935860058,
      "grad_norm": 0.13243380188941956,
      "learning_rate": 1e-06,
      "loss": 0.0234,
      "step": 285
    },
    {
      "clip_ratio/high_max": 0.0024635402733110823,
      "clip_ratio/high_mean": 0.000988963203781168,
      "clip_ratio/low_mean": 0.001163630215160083,
      "clip_ratio/low_min": 0.0001924317257362418,
      "clip_ratio/region_mean": 0.00215259339165641,
      "epoch": 4.279883381924198,
      "grad_norm": 0.11850841343402863,
      "learning_rate": 1e-06,
      "loss": -0.0112,
      "step": 286
    },
    {
      "clip_ratio/high_max": 0.002560369757702574,
      "clip_ratio/high_mean": 0.0011404558681533672,
      "clip_ratio/low_mean": 0.0013572694479080383,
      "clip_ratio/low_min": 0.00013896627842768794,
      "clip_ratio/region_mean": 0.0024977253124234267,
      "epoch": 4.289212827988338,
      "grad_norm": 0.1256835013628006,
      "learning_rate": 1e-06,
      "loss": 0.0401,
      "step": 287
    },
    {
      "clip_ratio/high_max": 0.0022343709933920763,
      "clip_ratio/high_mean": 0.0009108921967708739,
      "clip_ratio/low_mean": 0.0011752546743082348,
      "clip_ratio/low_min": 0.0001174202079710085,
      "clip_ratio/region_mean": 0.0020861468656221405,
      "epoch": 4.298542274052478,
      "grad_norm": 0.1245896965265274,
      "learning_rate": 1e-06,
      "loss": 0.0404,
      "step": 288
    },
    {
      "clip_ratio/high_max": 0.0026747331357910298,
      "clip_ratio/high_mean": 0.0010331952926208032,
      "clip_ratio/low_mean": 0.001311896885454189,
      "clip_ratio/low_min": 0.00014947346244298387,
      "clip_ratio/region_mean": 0.002345092194445897,
      "epoch": 4.307871720116618,
      "grad_norm": 0.11085493862628937,
      "learning_rate": 1e-06,
      "loss": 0.0267,
      "step": 289
    },
    {
      "clip_ratio/high_max": 0.0027068806084571406,
      "clip_ratio/high_mean": 0.0011260343235335313,
      "clip_ratio/low_mean": 0.00114196895447094,
      "clip_ratio/low_min": 0.00014054378698347136,
      "clip_ratio/region_mean": 0.0022680032925563864,
      "epoch": 4.317201166180758,
      "grad_norm": 0.12553079426288605,
      "learning_rate": 1e-06,
      "loss": 0.0117,
      "step": 290
    },
    {
      "clip_ratio/high_max": 0.002579201631306205,
      "clip_ratio/high_mean": 0.0010826847319549415,
      "clip_ratio/low_mean": 0.001319100076216273,
      "clip_ratio/low_min": 0.00016733756547182566,
      "clip_ratio/region_mean": 0.0024017847754294053,
      "epoch": 4.326530612244898,
      "grad_norm": 0.12764209508895874,
      "learning_rate": 1e-06,
      "loss": 0.0144,
      "step": 291
    },
    {
      "clip_ratio/high_max": 0.00249725476896856,
      "clip_ratio/high_mean": 0.0011542900429049041,
      "clip_ratio/low_mean": 0.00111367888712266,
      "clip_ratio/low_min": 2.815515017573489e-05,
      "clip_ratio/region_mean": 0.0022679689282085747,
      "epoch": 4.335860058309038,
      "grad_norm": 0.1246132105588913,
      "learning_rate": 1e-06,
      "loss": -0.0238,
      "step": 292
    },
    {
      "clip_ratio/high_max": 0.00241856378124794,
      "clip_ratio/high_mean": 0.0010271665178152034,
      "clip_ratio/low_mean": 0.0012715058073808905,
      "clip_ratio/low_min": 0.00011820547661045566,
      "clip_ratio/region_mean": 0.0022986723633948714,
      "epoch": 4.345189504373177,
      "grad_norm": 0.125278502702713,
      "learning_rate": 1e-06,
      "loss": 0.0223,
      "step": 293
    },
    {
      "clip_ratio/high_max": 0.003100218717008829,
      "clip_ratio/high_mean": 0.001345846161711961,
      "clip_ratio/low_mean": 0.0011336277548252838,
      "clip_ratio/low_min": 0.00016647075244691223,
      "clip_ratio/region_mean": 0.0024794739365461282,
      "epoch": 4.354518950437318,
      "grad_norm": 0.12681154906749725,
      "learning_rate": 1e-06,
      "loss": -0.0231,
      "step": 294
    },
    {
      "clip_ratio/high_max": 0.0025621378226787783,
      "clip_ratio/high_mean": 0.00110545907591586,
      "clip_ratio/low_mean": 0.0011410650822654134,
      "clip_ratio/low_min": 0.0001393528000335209,
      "clip_ratio/region_mean": 0.002246524156362284,
      "epoch": 4.363848396501457,
      "grad_norm": 0.11944369971752167,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 295
    },
    {
      "clip_ratio/high_max": 0.0024155409410013817,
      "clip_ratio/high_mean": 0.0011104911027359776,
      "clip_ratio/low_mean": 0.0011493445454107132,
      "clip_ratio/low_min": 9.195715210807975e-05,
      "clip_ratio/region_mean": 0.0022598356954404153,
      "epoch": 4.373177842565598,
      "grad_norm": 0.12992224097251892,
      "learning_rate": 1e-06,
      "loss": -0.0027,
      "step": 296
    },
    {
      "clip_ratio/high_max": 0.002720894022786524,
      "clip_ratio/high_mean": 0.0012110996212868486,
      "clip_ratio/low_mean": 0.0009991417209676001,
      "clip_ratio/low_min": 5.8841151258093305e-05,
      "clip_ratio/region_mean": 0.0022102413131506182,
      "epoch": 4.382507288629737,
      "grad_norm": 0.1187589094042778,
      "learning_rate": 1e-06,
      "loss": -0.0411,
      "step": 297
    },
    {
      "clip_ratio/high_max": 0.002440937743813265,
      "clip_ratio/high_mean": 0.0012230294705659617,
      "clip_ratio/low_mean": 0.0012636580413527554,
      "clip_ratio/low_min": 0.0001834376325859921,
      "clip_ratio/region_mean": 0.0024866875173756853,
      "epoch": 4.391836734693878,
      "grad_norm": 0.11927883327007294,
      "learning_rate": 1e-06,
      "loss": -0.0068,
      "step": 298
    },
    {
      "clip_ratio/high_max": 0.00268653185776202,
      "clip_ratio/high_mean": 0.0011927109226235189,
      "clip_ratio/low_mean": 0.001165936449979199,
      "clip_ratio/low_min": 0.00020812458114960464,
      "clip_ratio/region_mean": 0.002358647332584951,
      "epoch": 4.401166180758017,
      "grad_norm": 0.13587383925914764,
      "learning_rate": 1e-06,
      "loss": -0.0018,
      "step": 299
    },
    {
      "clip_ratio/high_max": 0.002681189842405729,
      "clip_ratio/high_mean": 0.0012017971057503019,
      "clip_ratio/low_mean": 0.0009524114539090078,
      "clip_ratio/low_min": 2.6411143153382e-05,
      "clip_ratio/region_mean": 0.0021542085451073945,
      "epoch": 4.410495626822158,
      "grad_norm": 0.11905869841575623,
      "learning_rate": 1e-06,
      "loss": -0.0027,
      "step": 300
    },
    {
      "clip_ratio/high_max": 0.0030099656178208534,
      "clip_ratio/high_mean": 0.0012513937235780759,
      "clip_ratio/low_mean": 0.0011389283727112343,
      "clip_ratio/low_min": 5.8096340580959804e-05,
      "clip_ratio/region_mean": 0.0023903220062493347,
      "epoch": 4.419825072886297,
      "grad_norm": 0.1261315494775772,
      "learning_rate": 1e-06,
      "loss": -0.0325,
      "step": 301
    },
    {
      "clip_ratio/high_max": 0.00295354258196312,
      "clip_ratio/high_mean": 0.0011922542544198222,
      "clip_ratio/low_mean": 0.0011407496240281034,
      "clip_ratio/low_min": 0.00012139445607317612,
      "clip_ratio/region_mean": 0.002333003889361862,
      "epoch": 4.429154518950437,
      "grad_norm": 0.1312561184167862,
      "learning_rate": 1e-06,
      "loss": 0.0067,
      "step": 302
    },
    {
      "clip_ratio/high_max": 0.0026947886726702563,
      "clip_ratio/high_mean": 0.001134440271925996,
      "clip_ratio/low_mean": 0.0011474858451947512,
      "clip_ratio/low_min": 3.3685639209579676e-05,
      "clip_ratio/region_mean": 0.0022819261357653886,
      "epoch": 4.438483965014577,
      "grad_norm": 0.13060961663722992,
      "learning_rate": 1e-06,
      "loss": 0.0052,
      "step": 303
    },
    {
      "clip_ratio/high_max": 0.0028368889543344267,
      "clip_ratio/high_mean": 0.0011986778699792922,
      "clip_ratio/low_mean": 0.0011177725564266439,
      "clip_ratio/low_min": 0.00013663903700944502,
      "clip_ratio/region_mean": 0.00231645046005724,
      "epoch": 4.447813411078717,
      "grad_norm": 0.12184076756238937,
      "learning_rate": 1e-06,
      "loss": 0.0294,
      "step": 304
    },
    {
      "clip_ratio/high_max": 0.0026461620291229337,
      "clip_ratio/high_mean": 0.0012459568715712521,
      "clip_ratio/low_mean": 0.001223670846229652,
      "clip_ratio/low_min": 0.00022749119671061635,
      "clip_ratio/region_mean": 0.0024696276741451584,
      "epoch": 4.457142857142857,
      "grad_norm": 0.1280045509338379,
      "learning_rate": 1e-06,
      "loss": 0.0149,
      "step": 305
    },
    {
      "clip_ratio/high_max": 0.0028626590574276634,
      "clip_ratio/high_mean": 0.0011490106480778195,
      "clip_ratio/low_mean": 0.0010497920302441344,
      "clip_ratio/low_min": 5.4598634960711934e-05,
      "clip_ratio/region_mean": 0.0021988027074257843,
      "epoch": 4.466472303206997,
      "grad_norm": 0.12246491760015488,
      "learning_rate": 1e-06,
      "loss": 0.0215,
      "step": 306
    },
    {
      "clip_ratio/high_max": 0.0027839219983434305,
      "clip_ratio/high_mean": 0.0011757337906601606,
      "clip_ratio/low_mean": 0.0010917840081674512,
      "clip_ratio/low_min": 7.153618389565963e-05,
      "clip_ratio/region_mean": 0.002267517826112453,
      "epoch": 4.475801749271137,
      "grad_norm": 0.11523416638374329,
      "learning_rate": 1e-06,
      "loss": -0.0143,
      "step": 307
    },
    {
      "clip_ratio/high_max": 0.0026096665096702054,
      "clip_ratio/high_mean": 0.0012803206191165373,
      "clip_ratio/low_mean": 0.001147410617704736,
      "clip_ratio/low_min": 4.01049528591102e-05,
      "clip_ratio/region_mean": 0.0024277312550111674,
      "epoch": 4.485131195335277,
      "grad_norm": 0.12913104891777039,
      "learning_rate": 1e-06,
      "loss": -0.0084,
      "step": 308
    },
    {
      "clip_ratio/high_max": 0.002587119917734526,
      "clip_ratio/high_mean": 0.0010348725227231625,
      "clip_ratio/low_mean": 0.0012732024624710903,
      "clip_ratio/low_min": 9.98653322312748e-05,
      "clip_ratio/region_mean": 0.0023080749670043588,
      "epoch": 4.494460641399417,
      "grad_norm": 0.13543091714382172,
      "learning_rate": 1e-06,
      "loss": 0.0552,
      "step": 309
    },
    {
      "clip_ratio/high_max": 0.002501116046914831,
      "clip_ratio/high_mean": 0.0011944292346015573,
      "clip_ratio/low_mean": 0.0012013216619379818,
      "clip_ratio/low_min": 0.00016099921867862577,
      "clip_ratio/region_mean": 0.002395750881987624,
      "epoch": 4.503790087463557,
      "grad_norm": 0.10411816090345383,
      "learning_rate": 1e-06,
      "loss": 0.0056,
      "step": 310
    },
    {
      "clip_ratio/high_max": 0.0026851797665585764,
      "clip_ratio/high_mean": 0.0010980216757161543,
      "clip_ratio/low_mean": 0.0015183269351837225,
      "clip_ratio/low_min": 0.0001243856950168265,
      "clip_ratio/region_mean": 0.002616348610899877,
      "epoch": 4.513119533527696,
      "grad_norm": 0.12977652251720428,
      "learning_rate": 1e-06,
      "loss": 0.0259,
      "step": 311
    },
    {
      "clip_ratio/high_max": 0.0024513979733455926,
      "clip_ratio/high_mean": 0.0010857978741114493,
      "clip_ratio/low_mean": 0.0014130021700111683,
      "clip_ratio/low_min": 0.0003086895030719461,
      "clip_ratio/region_mean": 0.0024987999931909144,
      "epoch": 4.522448979591837,
      "grad_norm": 0.11546915024518967,
      "learning_rate": 1e-06,
      "loss": 0.0296,
      "step": 312
    },
    {
      "clip_ratio/high_max": 0.002582550539955264,
      "clip_ratio/high_mean": 0.0011205352147953818,
      "clip_ratio/low_mean": 0.001067399003659375,
      "clip_ratio/low_min": 0.00013611125268653268,
      "clip_ratio/region_mean": 0.0021879342166357674,
      "epoch": 4.531778425655976,
      "grad_norm": 0.1222282350063324,
      "learning_rate": 1e-06,
      "loss": -0.0312,
      "step": 313
    },
    {
      "clip_ratio/high_max": 0.002486808065441437,
      "clip_ratio/high_mean": 0.001108541680878261,
      "clip_ratio/low_mean": 0.0013029424262640532,
      "clip_ratio/low_min": 0.00014830433246970642,
      "clip_ratio/region_mean": 0.0024114841216942295,
      "epoch": 4.541107871720117,
      "grad_norm": 0.1204584464430809,
      "learning_rate": 1e-06,
      "loss": -0.0196,
      "step": 314
    },
    {
      "clip_ratio/high_max": 0.002567245996033307,
      "clip_ratio/high_mean": 0.0011452670205471804,
      "clip_ratio/low_mean": 0.0012851643477915786,
      "clip_ratio/low_min": 0.00021749450661445735,
      "clip_ratio/region_mean": 0.0024304313483298756,
      "epoch": 4.550437317784256,
      "grad_norm": 0.12581069767475128,
      "learning_rate": 1e-06,
      "loss": -0.0084,
      "step": 315
    },
    {
      "clip_ratio/high_max": 0.0029545294819399714,
      "clip_ratio/high_mean": 0.0012033091225021053,
      "clip_ratio/low_mean": 0.0010612360765662743,
      "clip_ratio/low_min": 7.345522863033693e-05,
      "clip_ratio/region_mean": 0.00226454519724939,
      "epoch": 4.559766763848397,
      "grad_norm": 0.12477901577949524,
      "learning_rate": 1e-06,
      "loss": -0.0127,
      "step": 316
    },
    {
      "clip_ratio/high_max": 0.0026394765154691413,
      "clip_ratio/high_mean": 0.0012130733375670388,
      "clip_ratio/low_mean": 0.0010973194766847882,
      "clip_ratio/low_min": 8.088763570412993e-05,
      "clip_ratio/region_mean": 0.002310392759682145,
      "epoch": 4.569096209912536,
      "grad_norm": 0.11980269104242325,
      "learning_rate": 1e-06,
      "loss": -0.0294,
      "step": 317
    },
    {
      "clip_ratio/high_max": 0.0027670708805089816,
      "clip_ratio/high_mean": 0.0011913067246496212,
      "clip_ratio/low_mean": 0.0012311444261285942,
      "clip_ratio/low_min": 0.00015340284244302893,
      "clip_ratio/region_mean": 0.0024224511289503425,
      "epoch": 4.578425655976677,
      "grad_norm": 0.12332651019096375,
      "learning_rate": 1e-06,
      "loss": 0.0144,
      "step": 318
    },
    {
      "clip_ratio/high_max": 0.0026189246054855175,
      "clip_ratio/high_mean": 0.0011845795452245511,
      "clip_ratio/low_mean": 0.0012857424590038136,
      "clip_ratio/low_min": 4.9929793931369204e-05,
      "clip_ratio/region_mean": 0.0024703219169168733,
      "epoch": 4.587755102040816,
      "grad_norm": 0.11435230821371078,
      "learning_rate": 1e-06,
      "loss": -0.007,
      "step": 319
    },
    {
      "clip_ratio/high_max": 0.0026665675541153178,
      "clip_ratio/high_mean": 0.001226918548127287,
      "clip_ratio/low_mean": 0.0011941068987653125,
      "clip_ratio/low_min": 0.00016394666363339638,
      "clip_ratio/region_mean": 0.0024210253832279705,
      "epoch": 4.597084548104956,
      "grad_norm": 0.12260650098323822,
      "learning_rate": 1e-06,
      "loss": -0.0185,
      "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": 0.023297991071428603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4096.0,
      "completions/mean_length": 650.0001831054688,
      "completions/mean_terminated_length": 567.8002319335938,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 5.0093294460641395,
      "grad_norm": 0.1160479187965393,
      "learning_rate": 1e-06,
      "loss": -0.0212,
      "num_tokens": 223770075.0,
      "reward": 0.5726667642593384,
      "reward_std": 0.20052407681941986,
      "rewards/simpleverify_reward/mean": 0.5726667046546936,
      "rewards/simpleverify_reward/std": 0.4946956932544708,
      "step": 321
    },
    {
      "clip_ratio/high_max": 0.002331171781406738,
      "clip_ratio/high_mean": 0.0008940868410718394,
      "clip_ratio/low_mean": 0.0006400234624379664,
      "clip_ratio/low_min": 6.731693156325491e-05,
      "clip_ratio/region_mean": 0.00153411030623829,
      "epoch": 5.01865889212828,
      "grad_norm": 0.12034779787063599,
      "learning_rate": 1e-06,
      "loss": 0.0079,
      "step": 322
    },
    {
      "clip_ratio/high_max": 0.002272610225190874,
      "clip_ratio/high_mean": 0.0008834997934172861,
      "clip_ratio/low_mean": 0.000596067064179806,
      "clip_ratio/low_min": 7.751191424176795e-05,
      "clip_ratio/region_mean": 0.001479566897614859,
      "epoch": 5.0279883381924195,
      "grad_norm": 0.12023911625146866,
      "learning_rate": 1e-06,
      "loss": 0.0223,
      "step": 323
    },
    {
      "clip_ratio/high_max": 0.0020319217583164573,
      "clip_ratio/high_mean": 0.0008217888735089218,
      "clip_ratio/low_mean": 0.00048340864668716677,
      "clip_ratio/low_min": 8.909480129659642e-06,
      "clip_ratio/region_mean": 0.0013051975111011416,
      "epoch": 5.03731778425656,
      "grad_norm": 0.11286524683237076,
      "learning_rate": 1e-06,
      "loss": -0.0048,
      "step": 324
    },
    {
      "clip_ratio/high_max": 0.0018754447446553968,
      "clip_ratio/high_mean": 0.0008405982625845354,
      "clip_ratio/low_mean": 0.0006338683879221207,
      "clip_ratio/low_min": 7.990335507201962e-05,
      "clip_ratio/region_mean": 0.0014744666223123204,
      "epoch": 5.0466472303206995,
      "grad_norm": 0.11907852441072464,
      "learning_rate": 1e-06,
      "loss": 0.0325,
      "step": 325
    },
    {
      "clip_ratio/high_max": 0.0025481356206000783,
      "clip_ratio/high_mean": 0.0010352416102250572,
      "clip_ratio/low_mean": 0.0006705237265123287,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017057653240044601,
      "epoch": 5.05597667638484,
      "grad_norm": 0.12502989172935486,
      "learning_rate": 1e-06,
      "loss": -0.0346,
      "step": 326
    },
    {
      "clip_ratio/high_max": 0.0019633485935628414,
      "clip_ratio/high_mean": 0.0008725416064407909,
      "clip_ratio/low_mean": 0.0007412556515191682,
      "clip_ratio/low_min": 0.0001087099444703199,
      "clip_ratio/region_mean": 0.0016137972525029909,
      "epoch": 5.0653061224489795,
      "grad_norm": 0.1014496237039566,
      "learning_rate": 1e-06,
      "loss": 0.0234,
      "step": 327
    },
    {
      "clip_ratio/high_max": 0.0021971239148115274,
      "clip_ratio/high_mean": 0.0009192342222377192,
      "clip_ratio/low_mean": 0.000827921638119733,
      "clip_ratio/low_min": 8.556289049010957e-05,
      "clip_ratio/region_mean": 0.0017471558603574522,
      "epoch": 5.07463556851312,
      "grad_norm": 0.12265725433826447,
      "learning_rate": 1e-06,
      "loss": 0.058,
      "step": 328
    },
    {
      "clip_ratio/high_max": 0.002226817923656199,
      "clip_ratio/high_mean": 0.000960827879680437,
      "clip_ratio/low_mean": 0.0009769243661139626,
      "clip_ratio/low_min": 8.831449940771563e-05,
      "clip_ratio/region_mean": 0.001937752211233601,
      "epoch": 5.0839650145772595,
      "grad_norm": 0.11613408476114273,
      "learning_rate": 1e-06,
      "loss": 0.0296,
      "step": 329
    },
    {
      "clip_ratio/high_max": 0.0021740121810580604,
      "clip_ratio/high_mean": 0.0010172566544497386,
      "clip_ratio/low_mean": 0.001015359772281954,
      "clip_ratio/low_min": 7.959023332659854e-05,
      "clip_ratio/region_mean": 0.0020326164449215867,
      "epoch": 5.093294460641399,
      "grad_norm": 0.12618719041347504,
      "learning_rate": 1e-06,
      "loss": 0.006,
      "step": 330
    },
    {
      "clip_ratio/high_max": 0.002297082610311918,
      "clip_ratio/high_mean": 0.0009815042503760196,
      "clip_ratio/low_mean": 0.0007617335259055835,
      "clip_ratio/low_min": 3.755227771762293e-05,
      "clip_ratio/region_mean": 0.001743237764458172,
      "epoch": 5.1026239067055394,
      "grad_norm": 0.11878614127635956,
      "learning_rate": 1e-06,
      "loss": -0.0225,
      "step": 331
    },
    {
      "clip_ratio/high_max": 0.002167860948247835,
      "clip_ratio/high_mean": 0.0009389580663992092,
      "clip_ratio/low_mean": 0.0008977128345577512,
      "clip_ratio/low_min": 5.93311578995781e-05,
      "clip_ratio/region_mean": 0.0018366708682151511,
      "epoch": 5.111953352769679,
      "grad_norm": 0.12522326409816742,
      "learning_rate": 1e-06,
      "loss": 0.0203,
      "step": 332
    },
    {
      "clip_ratio/high_max": 0.0024267619373858906,
      "clip_ratio/high_mean": 0.0010387013389845379,
      "clip_ratio/low_mean": 0.0009657479022280313,
      "clip_ratio/low_min": 6.526062134071253e-05,
      "clip_ratio/region_mean": 0.0020044491757289506,
      "epoch": 5.121282798833819,
      "grad_norm": 0.13469408452510834,
      "learning_rate": 1e-06,
      "loss": 0.0223,
      "step": 333
    },
    {
      "clip_ratio/high_max": 0.002821952301019337,
      "clip_ratio/high_mean": 0.001127877689214074,
      "clip_ratio/low_mean": 0.0008927829876483884,
      "clip_ratio/low_min": 6.936312092875596e-05,
      "clip_ratio/region_mean": 0.002020660656853579,
      "epoch": 5.130612244897959,
      "grad_norm": 0.11073664575815201,
      "learning_rate": 1e-06,
      "loss": -0.0158,
      "step": 334
    },
    {
      "clip_ratio/high_max": 0.0023295316786970943,
      "clip_ratio/high_mean": 0.000995270282146521,
      "clip_ratio/low_mean": 0.0009571791997586843,
      "clip_ratio/low_min": 6.559574740094831e-05,
      "clip_ratio/region_mean": 0.0019524495146470144,
      "epoch": 5.139941690962099,
      "grad_norm": 0.12611344456672668,
      "learning_rate": 1e-06,
      "loss": 0.0282,
      "step": 335
    },
    {
      "clip_ratio/high_max": 0.002461226442392217,
      "clip_ratio/high_mean": 0.0010669805851648562,
      "clip_ratio/low_mean": 0.0008438949716946809,
      "clip_ratio/low_min": 0.00011409677699703025,
      "clip_ratio/region_mean": 0.0019108755222987384,
      "epoch": 5.149271137026239,
      "grad_norm": 0.12872008979320526,
      "learning_rate": 1e-06,
      "loss": 0.0103,
      "step": 336
    },
    {
      "clip_ratio/high_max": 0.0028701744886348024,
      "clip_ratio/high_mean": 0.0011309242145216558,
      "clip_ratio/low_mean": 0.0009735184194141766,
      "clip_ratio/low_min": 6.777114595024614e-05,
      "clip_ratio/region_mean": 0.00210444267213461,
      "epoch": 5.158600583090379,
      "grad_norm": 0.12863174080848694,
      "learning_rate": 1e-06,
      "loss": -0.0052,
      "step": 337
    },
    {
      "clip_ratio/high_max": 0.0024400284673902206,
      "clip_ratio/high_mean": 0.0011732769817172084,
      "clip_ratio/low_mean": 0.0008540326525690034,
      "clip_ratio/low_min": 9.653192682890221e-05,
      "clip_ratio/region_mean": 0.002027309630648233,
      "epoch": 5.167930029154519,
      "grad_norm": 0.16117098927497864,
      "learning_rate": 1e-06,
      "loss": 0.0118,
      "step": 338
    },
    {
      "clip_ratio/high_max": 0.0023087248409865424,
      "clip_ratio/high_mean": 0.0010590554120426532,
      "clip_ratio/low_mean": 0.0008177639774658019,
      "clip_ratio/low_min": 8.208412145904731e-05,
      "clip_ratio/region_mean": 0.0018768193549476564,
      "epoch": 5.1772594752186585,
      "grad_norm": 0.12683363258838654,
      "learning_rate": 1e-06,
      "loss": -0.0036,
      "step": 339
    },
    {
      "clip_ratio/high_max": 0.0023404003150062636,
      "clip_ratio/high_mean": 0.0010037991087301634,
      "clip_ratio/low_mean": 0.0009337233095720876,
      "clip_ratio/low_min": 3.5834959817293566e-05,
      "clip_ratio/region_mean": 0.0019375223855604418,
      "epoch": 5.186588921282799,
      "grad_norm": 0.11845511943101883,
      "learning_rate": 1e-06,
      "loss": -0.0148,
      "step": 340
    },
    {
      "clip_ratio/high_max": 0.0024044037927524187,
      "clip_ratio/high_mean": 0.0010416822224215139,
      "clip_ratio/low_mean": 0.0008880656714609358,
      "clip_ratio/low_min": 1.4420857951336075e-05,
      "clip_ratio/region_mean": 0.0019297479229862802,
      "epoch": 5.1959183673469385,
      "grad_norm": 0.11200569570064545,
      "learning_rate": 1e-06,
      "loss": -0.0225,
      "step": 341
    },
    {
      "clip_ratio/high_max": 0.002634903801663313,
      "clip_ratio/high_mean": 0.0011278366218903102,
      "clip_ratio/low_mean": 0.0009324102247774135,
      "clip_ratio/low_min": 3.589794141589664e-05,
      "clip_ratio/region_mean": 0.0020602468866854906,
      "epoch": 5.205247813411079,
      "grad_norm": 0.12052994966506958,
      "learning_rate": 1e-06,
      "loss": -0.0287,
      "step": 342
    },
    {
      "clip_ratio/high_max": 0.0022138115018606186,
      "clip_ratio/high_mean": 0.0009726383177621756,
      "clip_ratio/low_mean": 0.0011030281657440355,
      "clip_ratio/low_min": 5.18448341608746e-05,
      "clip_ratio/region_mean": 0.0020756665180670097,
      "epoch": 5.214577259475218,
      "grad_norm": 0.12434986978769302,
      "learning_rate": 1e-06,
      "loss": 0.0144,
      "step": 343
    },
    {
      "clip_ratio/high_max": 0.0021740799566032365,
      "clip_ratio/high_mean": 0.0010007506134570576,
      "clip_ratio/low_mean": 0.0012497354691731744,
      "clip_ratio/low_min": 0.00015783339495101245,
      "clip_ratio/region_mean": 0.0022504861117340624,
      "epoch": 5.223906705539359,
      "grad_norm": 0.11278295516967773,
      "learning_rate": 1e-06,
      "loss": 0.0057,
      "step": 344
    },
    {
      "clip_ratio/high_max": 0.0024882980651455,
      "clip_ratio/high_mean": 0.0011950697225984186,
      "clip_ratio/low_mean": 0.0010738543496700004,
      "clip_ratio/low_min": 3.2618742807244416e-05,
      "clip_ratio/region_mean": 0.0022689240795443766,
      "epoch": 5.233236151603498,
      "grad_norm": 0.12787608802318573,
      "learning_rate": 1e-06,
      "loss": -0.0455,
      "step": 345
    },
    {
      "clip_ratio/high_max": 0.0026640323412721045,
      "clip_ratio/high_mean": 0.0011983844451606274,
      "clip_ratio/low_mean": 0.0011817977501777932,
      "clip_ratio/low_min": 0.0001532142487121746,
      "clip_ratio/region_mean": 0.0023801822171662934,
      "epoch": 5.242565597667639,
      "grad_norm": 0.12050221115350723,
      "learning_rate": 1e-06,
      "loss": 0.0163,
      "step": 346
    },
    {
      "clip_ratio/high_max": 0.002212466177297756,
      "clip_ratio/high_mean": 0.0010118559766851831,
      "clip_ratio/low_mean": 0.001352729981590528,
      "clip_ratio/low_min": 0.00010361557269789046,
      "clip_ratio/region_mean": 0.0023645859546377324,
      "epoch": 5.251895043731778,
      "grad_norm": 0.11172395944595337,
      "learning_rate": 1e-06,
      "loss": 0.0195,
      "step": 347
    },
    {
      "clip_ratio/high_max": 0.0026730827521532774,
      "clip_ratio/high_mean": 0.0011719388348865323,
      "clip_ratio/low_mean": 0.0010778136183944298,
      "clip_ratio/low_min": 8.645639081805712e-05,
      "clip_ratio/region_mean": 0.0022497524769278243,
      "epoch": 5.261224489795918,
      "grad_norm": 0.12630048394203186,
      "learning_rate": 1e-06,
      "loss": -0.0338,
      "step": 348
    },
    {
      "clip_ratio/high_max": 0.0025160069053526968,
      "clip_ratio/high_mean": 0.0011735569787560962,
      "clip_ratio/low_mean": 0.0011306305132166017,
      "clip_ratio/low_min": 6.0883179685333744e-05,
      "clip_ratio/region_mean": 0.0023041874592308886,
      "epoch": 5.270553935860058,
      "grad_norm": 0.1273338794708252,
      "learning_rate": 1e-06,
      "loss": -0.022,
      "step": 349
    },
    {
      "clip_ratio/high_max": 0.0023636617697775364,
      "clip_ratio/high_mean": 0.0010016886990342755,
      "clip_ratio/low_mean": 0.0012023055060126353,
      "clip_ratio/low_min": 0.0001444820827600779,
      "clip_ratio/region_mean": 0.002203994197770953,
      "epoch": 5.279883381924198,
      "grad_norm": 0.12244521081447601,
      "learning_rate": 1e-06,
      "loss": 0.0098,
      "step": 350
    },
    {
      "clip_ratio/high_max": 0.002760966177447699,
      "clip_ratio/high_mean": 0.0010318204731447622,
      "clip_ratio/low_mean": 0.001118751186368172,
      "clip_ratio/low_min": 4.047739548695972e-05,
      "clip_ratio/region_mean": 0.002150571694073733,
      "epoch": 5.289212827988338,
      "grad_norm": 0.11932908743619919,
      "learning_rate": 1e-06,
      "loss": 0.006,
      "step": 351
    },
    {
      "clip_ratio/high_max": 0.002449590181640815,
      "clip_ratio/high_mean": 0.0009278356610593619,
      "clip_ratio/low_mean": 0.001007508508337196,
      "clip_ratio/low_min": 0.00011873193670908222,
      "clip_ratio/region_mean": 0.0019353441166458651,
      "epoch": 5.298542274052478,
      "grad_norm": 0.12832795083522797,
      "learning_rate": 1e-06,
      "loss": -0.0164,
      "step": 352
    },
    {
      "clip_ratio/high_max": 0.002966827465570532,
      "clip_ratio/high_mean": 0.001323889613559004,
      "clip_ratio/low_mean": 0.0009575075691827806,
      "clip_ratio/low_min": 7.740714681858663e-05,
      "clip_ratio/region_mean": 0.0022813971736468375,
      "epoch": 5.307871720116618,
      "grad_norm": 0.1216970905661583,
      "learning_rate": 1e-06,
      "loss": -0.0673,
      "step": 353
    },
    {
      "clip_ratio/high_max": 0.0026493128534639254,
      "clip_ratio/high_mean": 0.0011015911186404992,
      "clip_ratio/low_mean": 0.0011613713577389717,
      "clip_ratio/low_min": 0.00019017464182979893,
      "clip_ratio/region_mean": 0.0022629625163972378,
      "epoch": 5.317201166180758,
      "grad_norm": 0.12599064409732819,
      "learning_rate": 1e-06,
      "loss": 0.019,
      "step": 354
    },
    {
      "clip_ratio/high_max": 0.0023974219511728734,
      "clip_ratio/high_mean": 0.0010399304883321747,
      "clip_ratio/low_mean": 0.0012875821412308142,
      "clip_ratio/low_min": 0.00022142051602713764,
      "clip_ratio/region_mean": 0.0023275126222870313,
      "epoch": 5.326530612244898,
      "grad_norm": 0.12695588171482086,
      "learning_rate": 1e-06,
      "loss": -0.0154,
      "step": 355
    },
    {
      "clip_ratio/high_max": 0.0025081900967052206,
      "clip_ratio/high_mean": 0.0010517009268369293,
      "clip_ratio/low_mean": 0.00108551873563556,
      "clip_ratio/low_min": 3.621984251367394e-05,
      "clip_ratio/region_mean": 0.0021372197006712668,
      "epoch": 5.335860058309038,
      "grad_norm": 0.11755160242319107,
      "learning_rate": 1e-06,
      "loss": -0.003,
      "step": 356
    },
    {
      "clip_ratio/high_max": 0.0024951930026873015,
      "clip_ratio/high_mean": 0.0010354282185289776,
      "clip_ratio/low_mean": 0.0011063976344303228,
      "clip_ratio/low_min": 8.091726158454549e-05,
      "clip_ratio/region_mean": 0.0021418258838821203,
      "epoch": 5.345189504373177,
      "grad_norm": 0.1337786763906479,
      "learning_rate": 1e-06,
      "loss": 0.0198,
      "step": 357
    },
    {
      "clip_ratio/high_max": 0.0024985221825772896,
      "clip_ratio/high_mean": 0.0010636583574523684,
      "clip_ratio/low_mean": 0.0010912470243056305,
      "clip_ratio/low_min": 1.3997760106576607e-05,
      "clip_ratio/region_mean": 0.0021549054436036386,
      "epoch": 5.354518950437318,
      "grad_norm": 0.1356169581413269,
      "learning_rate": 1e-06,
      "loss": -0.0059,
      "step": 358
    },
    {
      "clip_ratio/high_max": 0.0029880207948735915,
      "clip_ratio/high_mean": 0.0011652917928586248,
      "clip_ratio/low_mean": 0.0011996547182206996,
      "clip_ratio/low_min": 8.742724821786396e-05,
      "clip_ratio/region_mean": 0.002364946558373049,
      "epoch": 5.363848396501457,
      "grad_norm": 0.11533825099468231,
      "learning_rate": 1e-06,
      "loss": 0.0052,
      "step": 359
    },
    {
      "clip_ratio/high_max": 0.0027105355038656853,
      "clip_ratio/high_mean": 0.0011398502610973082,
      "clip_ratio/low_mean": 0.0011601633377722465,
      "clip_ratio/low_min": 9.1689289547503e-05,
      "clip_ratio/region_mean": 0.00230001361342147,
      "epoch": 5.373177842565598,
      "grad_norm": 0.11489439755678177,
      "learning_rate": 1e-06,
      "loss": 0.0121,
      "step": 360
    },
    {
      "clip_ratio/high_max": 0.002577749670308549,
      "clip_ratio/high_mean": 0.0011227032646274893,
      "clip_ratio/low_mean": 0.0010684164280974073,
      "clip_ratio/low_min": 5.337702987162629e-05,
      "clip_ratio/region_mean": 0.002191119630879257,
      "epoch": 5.382507288629737,
      "grad_norm": 0.11034545302391052,
      "learning_rate": 1e-06,
      "loss": 0.0443,
      "step": 361
    },
    {
      "clip_ratio/high_max": 0.002486003162630368,
      "clip_ratio/high_mean": 0.0010863362404052168,
      "clip_ratio/low_mean": 0.0011133703737868927,
      "clip_ratio/low_min": 6.0286936786724254e-05,
      "clip_ratio/region_mean": 0.002199706606916152,
      "epoch": 5.391836734693878,
      "grad_norm": 0.11492704600095749,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 362
    },
    {
      "clip_ratio/high_max": 0.002338725797017105,
      "clip_ratio/high_mean": 0.0009308470598625718,
      "clip_ratio/low_mean": 0.0010402001280453987,
      "clip_ratio/low_min": 0.00010227440270682564,
      "clip_ratio/region_mean": 0.001971047226106748,
      "epoch": 5.401166180758017,
      "grad_norm": 0.11579523235559464,
      "learning_rate": 1e-06,
      "loss": 0.0383,
      "step": 363
    },
    {
      "clip_ratio/high_max": 0.002818047876644414,
      "clip_ratio/high_mean": 0.0011057969968533143,
      "clip_ratio/low_mean": 0.0010729788518801797,
      "clip_ratio/low_min": 6.820256567152683e-05,
      "clip_ratio/region_mean": 0.0021787758596474305,
      "epoch": 5.410495626822158,
      "grad_norm": 0.11623141914606094,
      "learning_rate": 1e-06,
      "loss": 0.0049,
      "step": 364
    },
    {
      "clip_ratio/high_max": 0.002724984889937332,
      "clip_ratio/high_mean": 0.00099569046869874,
      "clip_ratio/low_mean": 0.0013767164055025205,
      "clip_ratio/low_min": 0.00013524930636776844,
      "clip_ratio/region_mean": 0.002372406866925303,
      "epoch": 5.419825072886297,
      "grad_norm": 0.120187908411026,
      "learning_rate": 1e-06,
      "loss": 0.0306,
      "step": 365
    },
    {
      "clip_ratio/high_max": 0.0020549648943415377,
      "clip_ratio/high_mean": 0.0009662838820077013,
      "clip_ratio/low_mean": 0.0010244866098219063,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019907705354853533,
      "epoch": 5.429154518950437,
      "grad_norm": 0.1159253865480423,
      "learning_rate": 1e-06,
      "loss": 0.0056,
      "step": 366
    },
    {
      "clip_ratio/high_max": 0.002610755203932058,
      "clip_ratio/high_mean": 0.0012442871011444367,
      "clip_ratio/low_mean": 0.0011028297485609073,
      "clip_ratio/low_min": 0.0001869299112513545,
      "clip_ratio/region_mean": 0.0023471168824471533,
      "epoch": 5.438483965014577,
      "grad_norm": 0.13068360090255737,
      "learning_rate": 1e-06,
      "loss": -0.0493,
      "step": 367
    },
    {
      "clip_ratio/high_max": 0.0027408521928009577,
      "clip_ratio/high_mean": 0.001190292568935547,
      "clip_ratio/low_mean": 0.001175787670945283,
      "clip_ratio/low_min": 7.003932842053473e-05,
      "clip_ratio/region_mean": 0.0023660802689846605,
      "epoch": 5.447813411078717,
      "grad_norm": 0.12470521032810211,
      "learning_rate": 1e-06,
      "loss": 0.0069,
      "step": 368
    },
    {
      "clip_ratio/high_max": 0.0026157360480283387,
      "clip_ratio/high_mean": 0.0010111351512023248,
      "clip_ratio/low_mean": 0.0011441534406912979,
      "clip_ratio/low_min": 4.739906762551982e-05,
      "clip_ratio/region_mean": 0.002155288602807559,
      "epoch": 5.457142857142857,
      "grad_norm": 0.11756538599729538,
      "learning_rate": 1e-06,
      "loss": -0.0253,
      "step": 369
    },
    {
      "clip_ratio/high_max": 0.0024719072753214277,
      "clip_ratio/high_mean": 0.0010762463098217268,
      "clip_ratio/low_mean": 0.001172886106360238,
      "clip_ratio/low_min": 6.325067806756124e-05,
      "clip_ratio/region_mean": 0.002249132412543986,
      "epoch": 5.466472303206997,
      "grad_norm": 0.11395695060491562,
      "learning_rate": 1e-06,
      "loss": -0.0032,
      "step": 370
    },
    {
      "clip_ratio/high_max": 0.002748812308709603,
      "clip_ratio/high_mean": 0.0011688599443004932,
      "clip_ratio/low_mean": 0.0013253480719868094,
      "clip_ratio/low_min": 0.0001301289785260451,
      "clip_ratio/region_mean": 0.0024942079471657053,
      "epoch": 5.475801749271137,
      "grad_norm": 0.12123240530490875,
      "learning_rate": 1e-06,
      "loss": 0.0427,
      "step": 371
    },
    {
      "clip_ratio/high_max": 0.001981674569833558,
      "clip_ratio/high_mean": 0.000952392572798999,
      "clip_ratio/low_mean": 0.001176582720290753,
      "clip_ratio/low_min": 0.00018622579682414653,
      "clip_ratio/region_mean": 0.0021289753130986355,
      "epoch": 5.485131195335277,
      "grad_norm": 0.12081332504749298,
      "learning_rate": 1e-06,
      "loss": 0.0367,
      "step": 372
    },
    {
      "clip_ratio/high_max": 0.0022334548193612136,
      "clip_ratio/high_mean": 0.0010129321344720665,
      "clip_ratio/low_mean": 0.001158166785899084,
      "clip_ratio/low_min": 0.00011376764450687915,
      "clip_ratio/region_mean": 0.002171098931285087,
      "epoch": 5.494460641399417,
      "grad_norm": 0.12550733983516693,
      "learning_rate": 1e-06,
      "loss": 0.0019,
      "step": 373
    },
    {
      "clip_ratio/high_max": 0.0023999250479391776,
      "clip_ratio/high_mean": 0.001019960787743912,
      "clip_ratio/low_mean": 0.001106353285649675,
      "clip_ratio/low_min": 7.851872396713588e-05,
      "clip_ratio/region_mean": 0.002126314131601248,
      "epoch": 5.503790087463557,
      "grad_norm": 0.11755459755659103,
      "learning_rate": 1e-06,
      "loss": 0.0438,
      "step": 374
    },
    {
      "clip_ratio/high_max": 0.0026940009629470296,
      "clip_ratio/high_mean": 0.001118268981372239,
      "clip_ratio/low_mean": 0.0011399310315027833,
      "clip_ratio/low_min": 8.784370584180579e-05,
      "clip_ratio/region_mean": 0.0022581999946851283,
      "epoch": 5.513119533527696,
      "grad_norm": 0.11725718528032303,
      "learning_rate": 1e-06,
      "loss": -0.022,
      "step": 375
    },
    {
      "clip_ratio/high_max": 0.0025228596277884208,
      "clip_ratio/high_mean": 0.0011313870236335788,
      "clip_ratio/low_mean": 0.0011020937345165294,
      "clip_ratio/low_min": 0.00010748695331130875,
      "clip_ratio/region_mean": 0.00223348072177032,
      "epoch": 5.522448979591837,
      "grad_norm": 0.11317726224660873,
      "learning_rate": 1e-06,
      "loss": 0.0117,
      "step": 376
    },
    {
      "clip_ratio/high_max": 0.0027919566273340024,
      "clip_ratio/high_mean": 0.0010849928148672916,
      "clip_ratio/low_mean": 0.0012107754591852427,
      "clip_ratio/low_min": 8.374091157747898e-05,
      "clip_ratio/region_mean": 0.0022957682886044495,
      "epoch": 5.531778425655976,
      "grad_norm": 0.12786352634429932,
      "learning_rate": 1e-06,
      "loss": -0.0225,
      "step": 377
    },
    {
      "clip_ratio/high_max": 0.0025416793214390054,
      "clip_ratio/high_mean": 0.0011544046974449884,
      "clip_ratio/low_mean": 0.0010452705901116133,
      "clip_ratio/low_min": 7.583070419059368e-05,
      "clip_ratio/region_mean": 0.002199675283918623,
      "epoch": 5.541107871720117,
      "grad_norm": 0.12185269594192505,
      "learning_rate": 1e-06,
      "loss": 0.0075,
      "step": 378
    },
    {
      "clip_ratio/high_max": 0.002840211265720427,
      "clip_ratio/high_mean": 0.0012857738220191095,
      "clip_ratio/low_mean": 0.0010438932349643437,
      "clip_ratio/low_min": 6.664845750492532e-05,
      "clip_ratio/region_mean": 0.0023296670260606334,
      "epoch": 5.550437317784256,
      "grad_norm": 0.11564303934574127,
      "learning_rate": 1e-06,
      "loss": -0.0389,
      "step": 379
    },
    {
      "clip_ratio/high_max": 0.002514765234082006,
      "clip_ratio/high_mean": 0.001148274626757484,
      "clip_ratio/low_mean": 0.001131675027863821,
      "clip_ratio/low_min": 0.00013708053484151606,
      "clip_ratio/region_mean": 0.0022799496437073685,
      "epoch": 5.559766763848397,
      "grad_norm": 0.10961645841598511,
      "learning_rate": 1e-06,
      "loss": -0.0063,
      "step": 380
    },
    {
      "clip_ratio/high_max": 0.002365084528719308,
      "clip_ratio/high_mean": 0.0011176858915860066,
      "clip_ratio/low_mean": 0.0011285935797786806,
      "clip_ratio/low_min": 9.97857105176081e-05,
      "clip_ratio/region_mean": 0.0022462794513558038,
      "epoch": 5.569096209912536,
      "grad_norm": 0.1276141107082367,
      "learning_rate": 1e-06,
      "loss": 0.0246,
      "step": 381
    },
    {
      "clip_ratio/high_max": 0.0022333588131004944,
      "clip_ratio/high_mean": 0.0011299518828309374,
      "clip_ratio/low_mean": 0.0011652440462057712,
      "clip_ratio/low_min": 7.245335291372612e-05,
      "clip_ratio/region_mean": 0.002295195899932878,
      "epoch": 5.578425655976677,
      "grad_norm": 0.11832830309867859,
      "learning_rate": 1e-06,
      "loss": -0.0126,
      "step": 382
    },
    {
      "clip_ratio/high_max": 0.002499848931620363,
      "clip_ratio/high_mean": 0.0011582183506106958,
      "clip_ratio/low_mean": 0.0011567379515327048,
      "clip_ratio/low_min": 7.030448614386842e-05,
      "clip_ratio/region_mean": 0.002314956385816913,
      "epoch": 5.587755102040816,
      "grad_norm": 0.12387925386428833,
      "learning_rate": 1e-06,
      "loss": 0.0019,
      "step": 383
    },
    {
      "clip_ratio/high_max": 0.003059231021325104,
      "clip_ratio/high_mean": 0.0011537707687239163,
      "clip_ratio/low_mean": 0.0012203647420392372,
      "clip_ratio/low_min": 9.396314089826774e-05,
      "clip_ratio/region_mean": 0.002374135481659323,
      "epoch": 5.597084548104956,
      "grad_norm": 0.1167876198887825,
      "learning_rate": 1e-06,
      "loss": 0.007,
      "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": 0.026593889508928603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4091.0,
      "completions/mean_length": 657.3092041015625,
      "completions/mean_terminated_length": 563.3626708984375,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 6.0093294460641395,
      "grad_norm": 0.13324810564517975,
      "learning_rate": 1e-06,
      "loss": -0.0165,
      "num_tokens": 260918718.0,
      "reward": 0.5862165689468384,
      "reward_std": 0.19810469448566437,
      "rewards/simpleverify_reward/mean": 0.5862165093421936,
      "rewards/simpleverify_reward/std": 0.4925149381160736,
      "step": 385
    },
    {
      "clip_ratio/high_max": 0.0019730062267626636,
      "clip_ratio/high_mean": 0.0008151712245307863,
      "clip_ratio/low_mean": 0.0005951927241767407,
      "clip_ratio/low_min": 2.9049500881228596e-05,
      "clip_ratio/region_mean": 0.0014103639587119687,
      "epoch": 6.01865889212828,
      "grad_norm": 0.12404173612594604,
      "learning_rate": 1e-06,
      "loss": 0.027,
      "step": 386
    },
    {
      "clip_ratio/high_max": 0.0018699504144024104,
      "clip_ratio/high_mean": 0.0007915903934190283,
      "clip_ratio/low_mean": 0.0006663777894573286,
      "clip_ratio/low_min": 2.9187213840486947e-05,
      "clip_ratio/region_mean": 0.001457968191971304,
      "epoch": 6.0279883381924195,
      "grad_norm": 0.1162559986114502,
      "learning_rate": 1e-06,
      "loss": 0.0099,
      "step": 387
    },
    {
      "clip_ratio/high_max": 0.002215134503785521,
      "clip_ratio/high_mean": 0.000869595336553175,
      "clip_ratio/low_mean": 0.0007131926340662176,
      "clip_ratio/low_min": 3.1063616916071624e-05,
      "clip_ratio/region_mean": 0.0015827879433345515,
      "epoch": 6.03731778425656,
      "grad_norm": 0.13062424957752228,
      "learning_rate": 1e-06,
      "loss": 0.0404,
      "step": 388
    },
    {
      "clip_ratio/high_max": 0.002217354194726795,
      "clip_ratio/high_mean": 0.000930943337152712,
      "clip_ratio/low_mean": 0.0006345977781165857,
      "clip_ratio/low_min": 2.1999296222929843e-05,
      "clip_ratio/region_mean": 0.0015655411334591918,
      "epoch": 6.0466472303206995,
      "grad_norm": 0.1238541230559349,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 389
    },
    {
      "clip_ratio/high_max": 0.0020582657016348094,
      "clip_ratio/high_mean": 0.0009012167392938863,
      "clip_ratio/low_mean": 0.0008245883491326822,
      "clip_ratio/low_min": 3.943782940041274e-05,
      "clip_ratio/region_mean": 0.0017258051193493884,
      "epoch": 6.05597667638484,
      "grad_norm": 0.12203321605920792,
      "learning_rate": 1e-06,
      "loss": 0.0155,
      "step": 390
    },
    {
      "clip_ratio/high_max": 0.001856033326475881,
      "clip_ratio/high_mean": 0.0008223900895245606,
      "clip_ratio/low_mean": 0.0007892270768934395,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016116171536850743,
      "epoch": 6.0653061224489795,
      "grad_norm": 0.1219545230269432,
      "learning_rate": 1e-06,
      "loss": -0.0036,
      "step": 391
    },
    {
      "clip_ratio/high_max": 0.0019292017168481834,
      "clip_ratio/high_mean": 0.0008448498720099451,
      "clip_ratio/low_mean": 0.0008102485962808714,
      "clip_ratio/low_min": 3.42438888765173e-05,
      "clip_ratio/region_mean": 0.0016550984946661629,
      "epoch": 6.07463556851312,
      "grad_norm": 0.1314750462770462,
      "learning_rate": 1e-06,
      "loss": 0.0492,
      "step": 392
    },
    {
      "clip_ratio/high_max": 0.0020054995693499222,
      "clip_ratio/high_mean": 0.0009038643911480904,
      "clip_ratio/low_mean": 0.0008857637440087274,
      "clip_ratio/low_min": 4.695428560808068e-05,
      "clip_ratio/region_mean": 0.0017896280623972416,
      "epoch": 6.0839650145772595,
      "grad_norm": 0.11901679635047913,
      "learning_rate": 1e-06,
      "loss": -0.0027,
      "step": 393
    },
    {
      "clip_ratio/high_max": 0.0021379522295319475,
      "clip_ratio/high_mean": 0.0009289547087973915,
      "clip_ratio/low_mean": 0.0008250554019468836,
      "clip_ratio/low_min": 3.163425935781561e-05,
      "clip_ratio/region_mean": 0.0017540101398481056,
      "epoch": 6.093294460641399,
      "grad_norm": 0.1270415186882019,
      "learning_rate": 1e-06,
      "loss": 0.0566,
      "step": 394
    },
    {
      "clip_ratio/high_max": 0.002225495525635779,
      "clip_ratio/high_mean": 0.000858900921230088,
      "clip_ratio/low_mean": 0.0009488018295087386,
      "clip_ratio/low_min": 9.476575269218301e-05,
      "clip_ratio/region_mean": 0.0018077027561957948,
      "epoch": 6.1026239067055394,
      "grad_norm": 0.11928771436214447,
      "learning_rate": 1e-06,
      "loss": 0.0186,
      "step": 395
    },
    {
      "clip_ratio/high_max": 0.002256974628835451,
      "clip_ratio/high_mean": 0.0009545545817672973,
      "clip_ratio/low_mean": 0.0009311052472185111,
      "clip_ratio/low_min": 6.442429730668664e-05,
      "clip_ratio/region_mean": 0.0018856598253478296,
      "epoch": 6.111953352769679,
      "grad_norm": 0.12399211525917053,
      "learning_rate": 1e-06,
      "loss": 0.0461,
      "step": 396
    },
    {
      "clip_ratio/high_max": 0.0022681606715195812,
      "clip_ratio/high_mean": 0.0008418467168667121,
      "clip_ratio/low_mean": 0.0008683422320245882,
      "clip_ratio/low_min": 9.498564031673595e-05,
      "clip_ratio/region_mean": 0.0017101889461628161,
      "epoch": 6.121282798833819,
      "grad_norm": 0.12473439425230026,
      "learning_rate": 1e-06,
      "loss": -0.0112,
      "step": 397
    },
    {
      "clip_ratio/high_max": 0.002651781542226672,
      "clip_ratio/high_mean": 0.0010984279942931607,
      "clip_ratio/low_mean": 0.000939944467972964,
      "clip_ratio/low_min": 6.169328571559163e-05,
      "clip_ratio/region_mean": 0.002038372454990167,
      "epoch": 6.130612244897959,
      "grad_norm": 0.11931362003087997,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 398
    },
    {
      "clip_ratio/high_max": 0.0026233702956233174,
      "clip_ratio/high_mean": 0.0011405270706745796,
      "clip_ratio/low_mean": 0.0007884775477577932,
      "clip_ratio/low_min": 1.4172335795592517e-05,
      "clip_ratio/region_mean": 0.0019290045893285424,
      "epoch": 6.139941690962099,
      "grad_norm": 0.12697699666023254,
      "learning_rate": 1e-06,
      "loss": -0.0166,
      "step": 399
    },
    {
      "clip_ratio/high_max": 0.0026444002796779387,
      "clip_ratio/high_mean": 0.0010543804346525576,
      "clip_ratio/low_mean": 0.0007893318888818612,
      "clip_ratio/low_min": 5.639423034153879e-05,
      "clip_ratio/region_mean": 0.0018437123580952175,
      "epoch": 6.149271137026239,
      "grad_norm": 0.11611927300691605,
      "learning_rate": 1e-06,
      "loss": -0.0209,
      "step": 400
    },
    {
      "clip_ratio/high_max": 0.0024569192646595184,
      "clip_ratio/high_mean": 0.0009570927404638496,
      "clip_ratio/low_mean": 0.0009184736973111285,
      "clip_ratio/low_min": 2.44923039645073e-05,
      "clip_ratio/region_mean": 0.0018755664132186212,
      "epoch": 6.158600583090379,
      "grad_norm": 0.11482306569814682,
      "learning_rate": 1e-06,
      "loss": 0.0178,
      "step": 401
    },
    {
      "clip_ratio/high_max": 0.002464692723151529,
      "clip_ratio/high_mean": 0.000978390735326684,
      "clip_ratio/low_mean": 0.0007694177529629087,
      "clip_ratio/low_min": 6.687936092930613e-05,
      "clip_ratio/region_mean": 0.0017478084337199107,
      "epoch": 6.167930029154519,
      "grad_norm": 0.12404350936412811,
      "learning_rate": 1e-06,
      "loss": -0.0061,
      "step": 402
    },
    {
      "clip_ratio/high_max": 0.0025488484316156246,
      "clip_ratio/high_mean": 0.0010794915397127625,
      "clip_ratio/low_mean": 0.0009122902974922908,
      "clip_ratio/low_min": 5.434190825326368e-05,
      "clip_ratio/region_mean": 0.001991781849937979,
      "epoch": 6.1772594752186585,
      "grad_norm": 0.12119971960783005,
      "learning_rate": 1e-06,
      "loss": -0.017,
      "step": 403
    },
    {
      "clip_ratio/high_max": 0.0026436956832185388,
      "clip_ratio/high_mean": 0.0011907307889487129,
      "clip_ratio/low_mean": 0.0009121412922468153,
      "clip_ratio/low_min": 8.524704844603548e-05,
      "clip_ratio/region_mean": 0.0021028720948379487,
      "epoch": 6.186588921282799,
      "grad_norm": 0.1308881640434265,
      "learning_rate": 1e-06,
      "loss": -0.0215,
      "step": 404
    },
    {
      "clip_ratio/high_max": 0.002429325701086782,
      "clip_ratio/high_mean": 0.001083399933122564,
      "clip_ratio/low_mean": 0.001123972426285036,
      "clip_ratio/low_min": 6.25228767603403e-05,
      "clip_ratio/region_mean": 0.0022073723521316424,
      "epoch": 6.1959183673469385,
      "grad_norm": 0.1356833130121231,
      "learning_rate": 1e-06,
      "loss": 0.0154,
      "step": 405
    },
    {
      "clip_ratio/high_max": 0.002275712904520333,
      "clip_ratio/high_mean": 0.0009847798719420098,
      "clip_ratio/low_mean": 0.0008147502612700919,
      "clip_ratio/low_min": 3.122023645119043e-05,
      "clip_ratio/region_mean": 0.001799530153220985,
      "epoch": 6.205247813411079,
      "grad_norm": 0.11942654848098755,
      "learning_rate": 1e-06,
      "loss": -0.0519,
      "step": 406
    },
    {
      "clip_ratio/high_max": 0.0022498919279314578,
      "clip_ratio/high_mean": 0.0010205811140622245,
      "clip_ratio/low_mean": 0.001104500805013231,
      "clip_ratio/low_min": 9.362457240058575e-05,
      "clip_ratio/region_mean": 0.0021250819190754555,
      "epoch": 6.214577259475218,
      "grad_norm": 0.12094379961490631,
      "learning_rate": 1e-06,
      "loss": -0.0155,
      "step": 407
    },
    {
      "clip_ratio/high_max": 0.002205601616878994,
      "clip_ratio/high_mean": 0.0009364961479150224,
      "clip_ratio/low_mean": 0.0011847007408505306,
      "clip_ratio/low_min": 0.0001823055117711192,
      "clip_ratio/region_mean": 0.002121196906955447,
      "epoch": 6.223906705539359,
      "grad_norm": 0.12259498238563538,
      "learning_rate": 1e-06,
      "loss": 0.0451,
      "step": 408
    },
    {
      "clip_ratio/high_max": 0.0026138292960240506,
      "clip_ratio/high_mean": 0.001239790661202278,
      "clip_ratio/low_mean": 0.0012175131996627897,
      "clip_ratio/low_min": 0.00011217290102649713,
      "clip_ratio/region_mean": 0.002457303839037195,
      "epoch": 6.233236151603498,
      "grad_norm": 0.12996727228164673,
      "learning_rate": 1e-06,
      "loss": -0.029,
      "step": 409
    },
    {
      "clip_ratio/high_max": 0.00241059968539048,
      "clip_ratio/high_mean": 0.0010214820522378432,
      "clip_ratio/low_mean": 0.001148920736341097,
      "clip_ratio/low_min": 0.00026637545033736387,
      "clip_ratio/region_mean": 0.0021704028185922652,
      "epoch": 6.242565597667639,
      "grad_norm": 0.12623687088489532,
      "learning_rate": 1e-06,
      "loss": 0.0163,
      "step": 410
    },
    {
      "clip_ratio/high_max": 0.0021047641457698774,
      "clip_ratio/high_mean": 0.0008414694184466498,
      "clip_ratio/low_mean": 0.0011207731076865457,
      "clip_ratio/low_min": 0.00013071524972474435,
      "clip_ratio/region_mean": 0.001962242524314206,
      "epoch": 6.251895043731778,
      "grad_norm": 0.11063150316476822,
      "learning_rate": 1e-06,
      "loss": 0.0377,
      "step": 411
    },
    {
      "clip_ratio/high_max": 0.002761717958492227,
      "clip_ratio/high_mean": 0.0011835738114314154,
      "clip_ratio/low_mean": 0.001144818448665319,
      "clip_ratio/low_min": 0.00010819083581736777,
      "clip_ratio/region_mean": 0.002328392227354925,
      "epoch": 6.261224489795918,
      "grad_norm": 0.1341608613729477,
      "learning_rate": 1e-06,
      "loss": -0.0096,
      "step": 412
    },
    {
      "clip_ratio/high_max": 0.0023905458001536317,
      "clip_ratio/high_mean": 0.0010264829361403827,
      "clip_ratio/low_mean": 0.0011098397117166314,
      "clip_ratio/low_min": 1.5318628356908448e-05,
      "clip_ratio/region_mean": 0.0021363226769608445,
      "epoch": 6.270553935860058,
      "grad_norm": 0.10965313017368317,
      "learning_rate": 1e-06,
      "loss": 0.0478,
      "step": 413
    },
    {
      "clip_ratio/high_max": 0.0024112718856486026,
      "clip_ratio/high_mean": 0.0010530406416364713,
      "clip_ratio/low_mean": 0.0011679758717946243,
      "clip_ratio/low_min": 7.482237742806319e-05,
      "clip_ratio/region_mean": 0.0022210165043361485,
      "epoch": 6.279883381924198,
      "grad_norm": 0.12129051983356476,
      "learning_rate": 1e-06,
      "loss": -0.0074,
      "step": 414
    },
    {
      "clip_ratio/high_max": 0.0027557728462852538,
      "clip_ratio/high_mean": 0.001191336185002001,
      "clip_ratio/low_mean": 0.0011624247908912366,
      "clip_ratio/low_min": 8.862009144650074e-05,
      "clip_ratio/region_mean": 0.0023537609740742482,
      "epoch": 6.289212827988338,
      "grad_norm": 0.11545326560735703,
      "learning_rate": 1e-06,
      "loss": -0.0171,
      "step": 415
    },
    {
      "clip_ratio/high_max": 0.002751666062977165,
      "clip_ratio/high_mean": 0.0011539407842064975,
      "clip_ratio/low_mean": 0.0011127577636216301,
      "clip_ratio/low_min": 9.093844164453913e-05,
      "clip_ratio/region_mean": 0.002266698553285096,
      "epoch": 6.298542274052478,
      "grad_norm": 0.1203823983669281,
      "learning_rate": 1e-06,
      "loss": -0.0383,
      "step": 416
    },
    {
      "clip_ratio/high_max": 0.002638985257362947,
      "clip_ratio/high_mean": 0.0011841504056064878,
      "clip_ratio/low_mean": 0.0010946346446871758,
      "clip_ratio/low_min": 7.127127719286364e-05,
      "clip_ratio/region_mean": 0.0022787850175518543,
      "epoch": 6.307871720116618,
      "grad_norm": 0.12271475046873093,
      "learning_rate": 1e-06,
      "loss": -0.0433,
      "step": 417
    },
    {
      "clip_ratio/high_max": 0.0027284378520562313,
      "clip_ratio/high_mean": 0.0012587599994731136,
      "clip_ratio/low_mean": 0.0010464410588610917,
      "clip_ratio/low_min": 7.53785625420278e-05,
      "clip_ratio/region_mean": 0.0023052010001265444,
      "epoch": 6.317201166180758,
      "grad_norm": 0.12796494364738464,
      "learning_rate": 1e-06,
      "loss": -0.0071,
      "step": 418
    },
    {
      "clip_ratio/high_max": 0.0026533188938628882,
      "clip_ratio/high_mean": 0.0010761897137854248,
      "clip_ratio/low_mean": 0.001051126469974406,
      "clip_ratio/low_min": 5.620411866402719e-05,
      "clip_ratio/region_mean": 0.0021273161764838733,
      "epoch": 6.326530612244898,
      "grad_norm": 0.11977878212928772,
      "learning_rate": 1e-06,
      "loss": -0.012,
      "step": 419
    },
    {
      "clip_ratio/high_max": 0.002444995494442992,
      "clip_ratio/high_mean": 0.0010190890898229554,
      "clip_ratio/low_mean": 0.0011028128428733908,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021219019254203886,
      "epoch": 6.335860058309038,
      "grad_norm": 0.12461034208536148,
      "learning_rate": 1e-06,
      "loss": -0.0075,
      "step": 420
    },
    {
      "clip_ratio/high_max": 0.002507720455469098,
      "clip_ratio/high_mean": 0.0011358726369508076,
      "clip_ratio/low_mean": 0.0011133389180031372,
      "clip_ratio/low_min": 4.179173538432224e-05,
      "clip_ratio/region_mean": 0.0022492115676868707,
      "epoch": 6.345189504373177,
      "grad_norm": 0.12048563361167908,
      "learning_rate": 1e-06,
      "loss": -0.0056,
      "step": 421
    },
    {
      "clip_ratio/high_max": 0.002550481578509789,
      "clip_ratio/high_mean": 0.0011290810034552123,
      "clip_ratio/low_mean": 0.0012278427548153559,
      "clip_ratio/low_min": 0.00010310869220120367,
      "clip_ratio/region_mean": 0.002356923716433812,
      "epoch": 6.354518950437318,
      "grad_norm": 0.1191607341170311,
      "learning_rate": 1e-06,
      "loss": -0.0114,
      "step": 422
    },
    {
      "clip_ratio/high_max": 0.0022768615526729263,
      "clip_ratio/high_mean": 0.0010551393497735262,
      "clip_ratio/low_mean": 0.0012340135363047011,
      "clip_ratio/low_min": 0.000130217413243372,
      "clip_ratio/region_mean": 0.0022891528860782273,
      "epoch": 6.363848396501457,
      "grad_norm": 0.11609805375337601,
      "learning_rate": 1e-06,
      "loss": 0.0164,
      "step": 423
    },
    {
      "clip_ratio/high_max": 0.002574537407781463,
      "clip_ratio/high_mean": 0.0011344671584083699,
      "clip_ratio/low_mean": 0.001283480891288491,
      "clip_ratio/low_min": 5.9218235037405975e-05,
      "clip_ratio/region_mean": 0.0024179480387829244,
      "epoch": 6.373177842565598,
      "grad_norm": 0.12159954011440277,
      "learning_rate": 1e-06,
      "loss": 0.0054,
      "step": 424
    },
    {
      "clip_ratio/high_max": 0.00263553905824665,
      "clip_ratio/high_mean": 0.0010845652286661789,
      "clip_ratio/low_mean": 0.0012535936548374593,
      "clip_ratio/low_min": 4.384953444969142e-05,
      "clip_ratio/region_mean": 0.0023381588180200197,
      "epoch": 6.382507288629737,
      "grad_norm": 0.11567821353673935,
      "learning_rate": 1e-06,
      "loss": 0.0074,
      "step": 425
    },
    {
      "clip_ratio/high_max": 0.0024593555863248184,
      "clip_ratio/high_mean": 0.0011447414435679093,
      "clip_ratio/low_mean": 0.0012152215676906053,
      "clip_ratio/low_min": 8.845467073115287e-05,
      "clip_ratio/region_mean": 0.0023599630876560695,
      "epoch": 6.391836734693878,
      "grad_norm": 0.11477188020944595,
      "learning_rate": 1e-06,
      "loss": -0.013,
      "step": 426
    },
    {
      "clip_ratio/high_max": 0.002639504578837659,
      "clip_ratio/high_mean": 0.0012066481067449786,
      "clip_ratio/low_mean": 0.0010413934905955102,
      "clip_ratio/low_min": 4.2450270484550856e-05,
      "clip_ratio/region_mean": 0.002248041651910171,
      "epoch": 6.401166180758017,
      "grad_norm": 0.1149083599448204,
      "learning_rate": 1e-06,
      "loss": -0.0252,
      "step": 427
    },
    {
      "clip_ratio/high_max": 0.0025213523040292785,
      "clip_ratio/high_mean": 0.0011777847539633512,
      "clip_ratio/low_mean": 0.0012336117506492883,
      "clip_ratio/low_min": 0.00012792700636055088,
      "clip_ratio/region_mean": 0.0024113965046126395,
      "epoch": 6.410495626822158,
      "grad_norm": 0.1340465098619461,
      "learning_rate": 1e-06,
      "loss": 0.0161,
      "step": 428
    },
    {
      "clip_ratio/high_max": 0.002063682539301226,
      "clip_ratio/high_mean": 0.0009665437473813654,
      "clip_ratio/low_mean": 0.001161672826128779,
      "clip_ratio/low_min": 9.626071005186532e-05,
      "clip_ratio/region_mean": 0.0021282165398588404,
      "epoch": 6.419825072886297,
      "grad_norm": 0.12817728519439697,
      "learning_rate": 1e-06,
      "loss": 0.063,
      "step": 429
    },
    {
      "clip_ratio/high_max": 0.0027733416500268504,
      "clip_ratio/high_mean": 0.0010656159720383584,
      "clip_ratio/low_mean": 0.0011418458598200232,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002207461810030509,
      "epoch": 6.429154518950437,
      "grad_norm": 0.1197596862912178,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "step": 430
    },
    {
      "clip_ratio/high_max": 0.002767742520518368,
      "clip_ratio/high_mean": 0.0011557379566511372,
      "clip_ratio/low_mean": 0.0011472212318039965,
      "clip_ratio/low_min": 8.432554022874683e-05,
      "clip_ratio/region_mean": 0.0023029592193779536,
      "epoch": 6.438483965014577,
      "grad_norm": 0.11540327221155167,
      "learning_rate": 1e-06,
      "loss": 0.016,
      "step": 431
    },
    {
      "clip_ratio/high_max": 0.0026349278050474823,
      "clip_ratio/high_mean": 0.0011248408754909178,
      "clip_ratio/low_mean": 0.0012321955964580411,
      "clip_ratio/low_min": 7.574339269922348e-05,
      "clip_ratio/region_mean": 0.002357036486500874,
      "epoch": 6.447813411078717,
      "grad_norm": 0.11919288337230682,
      "learning_rate": 1e-06,
      "loss": -0.0046,
      "step": 432
    },
    {
      "clip_ratio/high_max": 0.0024311832567036618,
      "clip_ratio/high_mean": 0.0010826660882230499,
      "clip_ratio/low_mean": 0.0011767278829211136,
      "clip_ratio/low_min": 4.409579651110107e-05,
      "clip_ratio/region_mean": 0.002259393957501743,
      "epoch": 6.457142857142857,
      "grad_norm": 0.12749038636684418,
      "learning_rate": 1e-06,
      "loss": -0.0073,
      "step": 433
    },
    {
      "clip_ratio/high_max": 0.002881565764255356,
      "clip_ratio/high_mean": 0.001180740146082826,
      "clip_ratio/low_mean": 0.0011400093189877225,
      "clip_ratio/low_min": 0.00011249742601648904,
      "clip_ratio/region_mean": 0.002320749481441453,
      "epoch": 6.466472303206997,
      "grad_norm": 0.12400262802839279,
      "learning_rate": 1e-06,
      "loss": -0.0078,
      "step": 434
    },
    {
      "clip_ratio/high_max": 0.002597883423732128,
      "clip_ratio/high_mean": 0.001204928175866371,
      "clip_ratio/low_mean": 0.0011466079904494109,
      "clip_ratio/low_min": 1.800633799575735e-05,
      "clip_ratio/region_mean": 0.0023515361026511528,
      "epoch": 6.475801749271137,
      "grad_norm": 0.12447620928287506,
      "learning_rate": 1e-06,
      "loss": -0.0586,
      "step": 435
    },
    {
      "clip_ratio/high_max": 0.002456329522829037,
      "clip_ratio/high_mean": 0.0011332140529702883,
      "clip_ratio/low_mean": 0.0011636990839178907,
      "clip_ratio/low_min": 0.00010544337419560179,
      "clip_ratio/region_mean": 0.0022969131605350412,
      "epoch": 6.485131195335277,
      "grad_norm": 0.12708230316638947,
      "learning_rate": 1e-06,
      "loss": -0.0181,
      "step": 436
    },
    {
      "clip_ratio/high_max": 0.002377366407017689,
      "clip_ratio/high_mean": 0.001035883195072529,
      "clip_ratio/low_mean": 0.0014171033435559366,
      "clip_ratio/low_min": 8.347125458385563e-05,
      "clip_ratio/region_mean": 0.0024529865258955397,
      "epoch": 6.494460641399417,
      "grad_norm": 0.12132944166660309,
      "learning_rate": 1e-06,
      "loss": -0.0088,
      "step": 437
    },
    {
      "clip_ratio/high_max": 0.0025897861196426675,
      "clip_ratio/high_mean": 0.0010773708763736067,
      "clip_ratio/low_mean": 0.0013784209841105621,
      "clip_ratio/low_min": 9.457713167648762e-05,
      "clip_ratio/region_mean": 0.0024557918368373066,
      "epoch": 6.503790087463557,
      "grad_norm": 0.13327062129974365,
      "learning_rate": 1e-06,
      "loss": 0.0301,
      "step": 438
    },
    {
      "clip_ratio/high_max": 0.002631640942126978,
      "clip_ratio/high_mean": 0.0011027808650396764,
      "clip_ratio/low_mean": 0.0013612366092274897,
      "clip_ratio/low_min": 7.40386876714183e-05,
      "clip_ratio/region_mean": 0.002464017437887378,
      "epoch": 6.513119533527696,
      "grad_norm": 0.12736468017101288,
      "learning_rate": 1e-06,
      "loss": 0.0417,
      "step": 439
    },
    {
      "clip_ratio/high_max": 0.002668326771527063,
      "clip_ratio/high_mean": 0.0011093713037553243,
      "clip_ratio/low_mean": 0.0013399616182141472,
      "clip_ratio/low_min": 7.074468885548413e-05,
      "clip_ratio/region_mean": 0.002449332874675747,
      "epoch": 6.522448979591837,
      "grad_norm": 0.12351478636264801,
      "learning_rate": 1e-06,
      "loss": 0.0189,
      "step": 440
    },
    {
      "clip_ratio/high_max": 0.0026762376073747873,
      "clip_ratio/high_mean": 0.0011525229892868083,
      "clip_ratio/low_mean": 0.0010786298189486843,
      "clip_ratio/low_min": 0.00013715065870201215,
      "clip_ratio/region_mean": 0.0022311528300633654,
      "epoch": 6.531778425655976,
      "grad_norm": 0.1278199851512909,
      "learning_rate": 1e-06,
      "loss": -0.0338,
      "step": 441
    },
    {
      "clip_ratio/high_max": 0.0030020528502063826,
      "clip_ratio/high_mean": 0.00113365893412265,
      "clip_ratio/low_mean": 0.0010324991271772888,
      "clip_ratio/low_min": 0.00012030499237880576,
      "clip_ratio/region_mean": 0.002166158090403769,
      "epoch": 6.541107871720117,
      "grad_norm": 0.1174987331032753,
      "learning_rate": 1e-06,
      "loss": -0.0177,
      "step": 442
    },
    {
      "clip_ratio/high_max": 0.0030522443776135333,
      "clip_ratio/high_mean": 0.001245684783498291,
      "clip_ratio/low_mean": 0.0011783944974013139,
      "clip_ratio/low_min": 2.3487411453970708e-05,
      "clip_ratio/region_mean": 0.0024240792627097107,
      "epoch": 6.550437317784256,
      "grad_norm": 0.11356744915246964,
      "learning_rate": 1e-06,
      "loss": -0.0356,
      "step": 443
    },
    {
      "clip_ratio/high_max": 0.0026246298439218663,
      "clip_ratio/high_mean": 0.0011240662479394814,
      "clip_ratio/low_mean": 0.0013824096095049754,
      "clip_ratio/low_min": 0.0001541881611046847,
      "clip_ratio/region_mean": 0.0025064759101951495,
      "epoch": 6.559766763848397,
      "grad_norm": 0.11744951456785202,
      "learning_rate": 1e-06,
      "loss": 0.0368,
      "step": 444
    },
    {
      "clip_ratio/high_max": 0.0026455251208972186,
      "clip_ratio/high_mean": 0.001171026855445234,
      "clip_ratio/low_mean": 0.0012133766613260377,
      "clip_ratio/low_min": 0.00010664136061677709,
      "clip_ratio/region_mean": 0.002384403494943399,
      "epoch": 6.569096209912536,
      "grad_norm": 0.12163163721561432,
      "learning_rate": 1e-06,
      "loss": -0.0115,
      "step": 445
    },
    {
      "clip_ratio/high_max": 0.002699668490095064,
      "clip_ratio/high_mean": 0.001236726289789658,
      "clip_ratio/low_mean": 0.0011418564463383518,
      "clip_ratio/low_min": 8.202681965485681e-05,
      "clip_ratio/region_mean": 0.0023785828016116284,
      "epoch": 6.578425655976677,
      "grad_norm": 0.12173604965209961,
      "learning_rate": 1e-06,
      "loss": -0.0022,
      "step": 446
    },
    {
      "clip_ratio/high_max": 0.002826439398631919,
      "clip_ratio/high_mean": 0.0011543520049599465,
      "clip_ratio/low_mean": 0.0011969719380431343,
      "clip_ratio/low_min": 4.8110874558915384e-05,
      "clip_ratio/region_mean": 0.002351323935727123,
      "epoch": 6.587755102040816,
      "grad_norm": 0.12113094329833984,
      "learning_rate": 1e-06,
      "loss": -0.0158,
      "step": 447
    },
    {
      "clip_ratio/high_max": 0.002635091499541886,
      "clip_ratio/high_mean": 0.0011047649331885623,
      "clip_ratio/low_mean": 0.0013531399727071403,
      "clip_ratio/low_min": 0.00018112690850102808,
      "clip_ratio/region_mean": 0.002457904862239957,
      "epoch": 6.597084548104956,
      "grad_norm": 0.11818385869264603,
      "learning_rate": 1e-06,
      "loss": -0.0143,
      "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": 0.028529575892857095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4091.0,
      "completions/mean_length": 667.4026489257812,
      "completions/mean_terminated_length": 566.7135620117188,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 7.0093294460641395,
      "grad_norm": 0.12843579053878784,
      "learning_rate": 1e-06,
      "loss": -0.0144,
      "num_tokens": 298188831.0,
      "reward": 0.5926164984703064,
      "reward_std": 0.19546960294246674,
      "rewards/simpleverify_reward/mean": 0.5926164984703064,
      "rewards/simpleverify_reward/std": 0.4913516044616699,
      "step": 449
    },
    {
      "clip_ratio/high_max": 0.0017942120612133294,
      "clip_ratio/high_mean": 0.0008397825422434835,
      "clip_ratio/low_mean": 0.0005558612620006897,
      "clip_ratio/low_min": 3.84697978006443e-05,
      "clip_ratio/region_mean": 0.0013956437978777103,
      "epoch": 7.01865889212828,
      "grad_norm": 0.12819872796535492,
      "learning_rate": 1e-06,
      "loss": 0.0123,
      "step": 450
    },
    {
      "clip_ratio/high_max": 0.0018787424378388096,
      "clip_ratio/high_mean": 0.0007833713934815023,
      "clip_ratio/low_mean": 0.0005890489737794269,
      "clip_ratio/low_min": 4.643387728719972e-05,
      "clip_ratio/region_mean": 0.0013724203672609292,
      "epoch": 7.0279883381924195,
      "grad_norm": 0.11226193606853485,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "step": 451
    },
    {
      "clip_ratio/high_max": 0.002255278632219415,
      "clip_ratio/high_mean": 0.0009132294890150661,
      "clip_ratio/low_mean": 0.000774587537307525,
      "clip_ratio/low_min": 9.933398905559443e-05,
      "clip_ratio/region_mean": 0.00168781699903775,
      "epoch": 7.03731778425656,
      "grad_norm": 0.13736703991889954,
      "learning_rate": 1e-06,
      "loss": 0.0151,
      "step": 452
    },
    {
      "clip_ratio/high_max": 0.0022158552310429513,
      "clip_ratio/high_mean": 0.000965459239523625,
      "clip_ratio/low_mean": 0.0005958840192761272,
      "clip_ratio/low_min": 1.1770244782383088e-05,
      "clip_ratio/region_mean": 0.001561343262437731,
      "epoch": 7.0466472303206995,
      "grad_norm": 0.11436965316534042,
      "learning_rate": 1e-06,
      "loss": 0.0041,
      "step": 453
    },
    {
      "clip_ratio/high_max": 0.0022080572525737807,
      "clip_ratio/high_mean": 0.0009584508152329363,
      "clip_ratio/low_mean": 0.0005942661882727407,
      "clip_ratio/low_min": 2.36720325119677e-05,
      "clip_ratio/region_mean": 0.0015527170253335498,
      "epoch": 7.05597667638484,
      "grad_norm": 0.10365674644708633,
      "learning_rate": 1e-06,
      "loss": -0.0316,
      "step": 454
    },
    {
      "clip_ratio/high_max": 0.0024227470566984266,
      "clip_ratio/high_mean": 0.0009210606094711693,
      "clip_ratio/low_mean": 0.0007715885913057718,
      "clip_ratio/low_min": 1.1220825399504974e-05,
      "clip_ratio/region_mean": 0.0016926492025959305,
      "epoch": 7.0653061224489795,
      "grad_norm": 0.13070574402809143,
      "learning_rate": 1e-06,
      "loss": 0.0291,
      "step": 455
    },
    {
      "clip_ratio/high_max": 0.002068601446808316,
      "clip_ratio/high_mean": 0.000858913339470746,
      "clip_ratio/low_mean": 0.0007577644046250498,
      "clip_ratio/low_min": 3.540270517987665e-05,
      "clip_ratio/region_mean": 0.001616677804122446,
      "epoch": 7.07463556851312,
      "grad_norm": 0.12022151052951813,
      "learning_rate": 1e-06,
      "loss": 0.0323,
      "step": 456
    },
    {
      "clip_ratio/high_max": 0.0020541218982543796,
      "clip_ratio/high_mean": 0.0009139200956269633,
      "clip_ratio/low_mean": 0.0008234897231886862,
      "clip_ratio/low_min": 0.00010106707668455783,
      "clip_ratio/region_mean": 0.0017374098533764482,
      "epoch": 7.0839650145772595,
      "grad_norm": 0.1268029659986496,
      "learning_rate": 1e-06,
      "loss": -0.0139,
      "step": 457
    },
    {
      "clip_ratio/high_max": 0.0017584330453246366,
      "clip_ratio/high_mean": 0.0008227400994655909,
      "clip_ratio/low_mean": 0.000893424346941174,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001716164442768786,
      "epoch": 7.093294460641399,
      "grad_norm": 0.12214168161153793,
      "learning_rate": 1e-06,
      "loss": 0.0126,
      "step": 458
    },
    {
      "clip_ratio/high_max": 0.0020603343364200555,
      "clip_ratio/high_mean": 0.0009572600629326189,
      "clip_ratio/low_mean": 0.0007520450217270991,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017093050919356756,
      "epoch": 7.1026239067055394,
      "grad_norm": 0.13558630645275116,
      "learning_rate": 1e-06,
      "loss": -0.0098,
      "step": 459
    },
    {
      "clip_ratio/high_max": 0.00220011221972527,
      "clip_ratio/high_mean": 0.0009486302678851644,
      "clip_ratio/low_mean": 0.0010386598223703913,
      "clip_ratio/low_min": 1.4434180229727644e-05,
      "clip_ratio/region_mean": 0.001987290110264439,
      "epoch": 7.111953352769679,
      "grad_norm": 0.11543698608875275,
      "learning_rate": 1e-06,
      "loss": 0.0081,
      "step": 460
    },
    {
      "clip_ratio/high_max": 0.002159191975806607,
      "clip_ratio/high_mean": 0.0010031382134911837,
      "clip_ratio/low_mean": 0.0009441756719752448,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019473139109322801,
      "epoch": 7.121282798833819,
      "grad_norm": 0.12990941107273102,
      "learning_rate": 1e-06,
      "loss": 0.0156,
      "step": 461
    },
    {
      "clip_ratio/high_max": 0.002285688191477675,
      "clip_ratio/high_mean": 0.0010069564887089655,
      "clip_ratio/low_mean": 0.0009788917122932617,
      "clip_ratio/low_min": 5.853668335475959e-05,
      "clip_ratio/region_mean": 0.0019858481900882907,
      "epoch": 7.130612244897959,
      "grad_norm": 0.12387382984161377,
      "learning_rate": 1e-06,
      "loss": 0.012,
      "step": 462
    },
    {
      "clip_ratio/high_max": 0.0023053531476762146,
      "clip_ratio/high_mean": 0.001027273538056761,
      "clip_ratio/low_mean": 0.0009512579526926856,
      "clip_ratio/low_min": 4.76672485092422e-05,
      "clip_ratio/region_mean": 0.001978531538043171,
      "epoch": 7.139941690962099,
      "grad_norm": 0.12219683080911636,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 463
    },
    {
      "clip_ratio/high_max": 0.0023550383702968247,
      "clip_ratio/high_mean": 0.0009040625118359458,
      "clip_ratio/low_mean": 0.0010384545057604555,
      "clip_ratio/low_min": 6.813026629970409e-05,
      "clip_ratio/region_mean": 0.0019425170175964013,
      "epoch": 7.149271137026239,
      "grad_norm": 0.1295541226863861,
      "learning_rate": 1e-06,
      "loss": 0.0464,
      "step": 464
    },
    {
      "clip_ratio/high_max": 0.0028461072870413773,
      "clip_ratio/high_mean": 0.0012132468764320947,
      "clip_ratio/low_mean": 0.0009434288076590747,
      "clip_ratio/low_min": 7.763068970234599e-05,
      "clip_ratio/region_mean": 0.0021566756477113813,
      "epoch": 7.158600583090379,
      "grad_norm": 0.12018097937107086,
      "learning_rate": 1e-06,
      "loss": -0.0533,
      "step": 465
    },
    {
      "clip_ratio/high_max": 0.002219119167421013,
      "clip_ratio/high_mean": 0.0009892565085465321,
      "clip_ratio/low_mean": 0.0009505505477136467,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019398070908209775,
      "epoch": 7.167930029154519,
      "grad_norm": 0.11691570281982422,
      "learning_rate": 1e-06,
      "loss": -0.0212,
      "step": 466
    },
    {
      "clip_ratio/high_max": 0.0024940534567576833,
      "clip_ratio/high_mean": 0.0009516580576018896,
      "clip_ratio/low_mean": 0.0009481899232923752,
      "clip_ratio/low_min": 2.4400567781412974e-05,
      "clip_ratio/region_mean": 0.001899847928143572,
      "epoch": 7.1772594752186585,
      "grad_norm": 0.12506066262722015,
      "learning_rate": 1e-06,
      "loss": 0.0465,
      "step": 467
    },
    {
      "clip_ratio/high_max": 0.002621080646349583,
      "clip_ratio/high_mean": 0.0011659690462693106,
      "clip_ratio/low_mean": 0.0008470480552205117,
      "clip_ratio/low_min": 2.7496033908391837e-05,
      "clip_ratio/region_mean": 0.0020130171105847694,
      "epoch": 7.186588921282799,
      "grad_norm": 0.1221906766295433,
      "learning_rate": 1e-06,
      "loss": -0.0348,
      "step": 468
    },
    {
      "clip_ratio/high_max": 0.0028215171914780512,
      "clip_ratio/high_mean": 0.001101287591154687,
      "clip_ratio/low_mean": 0.0010750167948572198,
      "clip_ratio/low_min": 5.850890011060983e-05,
      "clip_ratio/region_mean": 0.0021763044060207903,
      "epoch": 7.1959183673469385,
      "grad_norm": 0.12542860209941864,
      "learning_rate": 1e-06,
      "loss": -0.0076,
      "step": 469
    },
    {
      "clip_ratio/high_max": 0.002720331212913152,
      "clip_ratio/high_mean": 0.0012087243449059315,
      "clip_ratio/low_mean": 0.0007662919597350992,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001975016319192946,
      "epoch": 7.205247813411079,
      "grad_norm": 0.11854458600282669,
      "learning_rate": 1e-06,
      "loss": -0.078,
      "step": 470
    },
    {
      "clip_ratio/high_max": 0.0022921818599570543,
      "clip_ratio/high_mean": 0.0010176220603170805,
      "clip_ratio/low_mean": 0.0010260480739816558,
      "clip_ratio/low_min": 5.1201528549427167e-05,
      "clip_ratio/region_mean": 0.0020436701088328846,
      "epoch": 7.214577259475218,
      "grad_norm": 0.11927090585231781,
      "learning_rate": 1e-06,
      "loss": -0.0112,
      "step": 471
    },
    {
      "clip_ratio/high_max": 0.0028679999231826514,
      "clip_ratio/high_mean": 0.0012150522015872411,
      "clip_ratio/low_mean": 0.0009577400123816915,
      "clip_ratio/low_min": 4.991071546101011e-05,
      "clip_ratio/region_mean": 0.0021727922212448902,
      "epoch": 7.223906705539359,
      "grad_norm": 0.11966941505670547,
      "learning_rate": 1e-06,
      "loss": -0.0254,
      "step": 472
    },
    {
      "clip_ratio/high_max": 0.00237848923279671,
      "clip_ratio/high_mean": 0.000964981853030622,
      "clip_ratio/low_mean": 0.001007246806693729,
      "clip_ratio/low_min": 3.565316728781909e-05,
      "clip_ratio/region_mean": 0.0019722286670003086,
      "epoch": 7.233236151603498,
      "grad_norm": 0.11110007017850876,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 473
    },
    {
      "clip_ratio/high_max": 0.002525853124097921,
      "clip_ratio/high_mean": 0.0012086549195373664,
      "clip_ratio/low_mean": 0.0010972887157549849,
      "clip_ratio/low_min": 7.160093355196295e-05,
      "clip_ratio/region_mean": 0.0023059436280163936,
      "epoch": 7.242565597667639,
      "grad_norm": 0.14252908527851105,
      "learning_rate": 1e-06,
      "loss": 0.0043,
      "step": 474
    },
    {
      "clip_ratio/high_max": 0.0023088748566806316,
      "clip_ratio/high_mean": 0.001019692297631991,
      "clip_ratio/low_mean": 0.001102666345104808,
      "clip_ratio/low_min": 4.645337685360573e-05,
      "clip_ratio/region_mean": 0.0021223586081760004,
      "epoch": 7.251895043731778,
      "grad_norm": 0.11959218978881836,
      "learning_rate": 1e-06,
      "loss": 0.0049,
      "step": 475
    },
    {
      "clip_ratio/high_max": 0.002331751282326877,
      "clip_ratio/high_mean": 0.0010360079450038029,
      "clip_ratio/low_mean": 0.0011968514390900964,
      "clip_ratio/low_min": 0.0002064881800833973,
      "clip_ratio/region_mean": 0.002232859391369857,
      "epoch": 7.261224489795918,
      "grad_norm": 0.11894406378269196,
      "learning_rate": 1e-06,
      "loss": 0.0146,
      "step": 476
    },
    {
      "clip_ratio/high_max": 0.002595975158328656,
      "clip_ratio/high_mean": 0.0011571917748369742,
      "clip_ratio/low_mean": 0.0012506505263445433,
      "clip_ratio/low_min": 0.00010009974721469916,
      "clip_ratio/region_mean": 0.0024078422793536447,
      "epoch": 7.270553935860058,
      "grad_norm": 0.12460054457187653,
      "learning_rate": 1e-06,
      "loss": -0.0071,
      "step": 477
    },
    {
      "clip_ratio/high_max": 0.0028466668009059504,
      "clip_ratio/high_mean": 0.0011999497983197216,
      "clip_ratio/low_mean": 0.001092759595849202,
      "clip_ratio/low_min": 0.00011529072799021378,
      "clip_ratio/region_mean": 0.0022927094250917435,
      "epoch": 7.279883381924198,
      "grad_norm": 0.11852549016475677,
      "learning_rate": 1e-06,
      "loss": -0.0332,
      "step": 478
    },
    {
      "clip_ratio/high_max": 0.002350836235564202,
      "clip_ratio/high_mean": 0.0010577016655588523,
      "clip_ratio/low_mean": 0.0012045312541886233,
      "clip_ratio/low_min": 4.7819434257689863e-05,
      "clip_ratio/region_mean": 0.0022622328833676875,
      "epoch": 7.289212827988338,
      "grad_norm": 0.11749444901943207,
      "learning_rate": 1e-06,
      "loss": 0.0065,
      "step": 479
    },
    {
      "clip_ratio/high_max": 0.0023864907198003493,
      "clip_ratio/high_mean": 0.0010245456105621997,
      "clip_ratio/low_mean": 0.0011411018876970047,
      "clip_ratio/low_min": 7.487407674489077e-05,
      "clip_ratio/region_mean": 0.002165647536457982,
      "epoch": 7.298542274052478,
      "grad_norm": 0.11341959983110428,
      "learning_rate": 1e-06,
      "loss": -0.0399,
      "step": 480
    },
    {
      "clip_ratio/high_max": 0.00280179691981175,
      "clip_ratio/high_mean": 0.001034307469126361,
      "clip_ratio/low_mean": 0.0012337325679254718,
      "clip_ratio/low_min": 0.0002653695974004222,
      "clip_ratio/region_mean": 0.0022680400652461685,
      "epoch": 7.307871720116618,
      "grad_norm": 0.11593709141016006,
      "learning_rate": 1e-06,
      "loss": 0.0082,
      "step": 481
    },
    {
      "clip_ratio/high_max": 0.002634935633977875,
      "clip_ratio/high_mean": 0.0012052946658513974,
      "clip_ratio/low_mean": 0.0011131834107800387,
      "clip_ratio/low_min": 5.8690147852757946e-05,
      "clip_ratio/region_mean": 0.002318478080269415,
      "epoch": 7.317201166180758,
      "grad_norm": 0.15193410217761993,
      "learning_rate": 1e-06,
      "loss": -0.044,
      "step": 482
    },
    {
      "clip_ratio/high_max": 0.0024060449650278315,
      "clip_ratio/high_mean": 0.0011861616112582851,
      "clip_ratio/low_mean": 0.00122487666158122,
      "clip_ratio/low_min": 8.172049638233148e-05,
      "clip_ratio/region_mean": 0.0024110383092192933,
      "epoch": 7.326530612244898,
      "grad_norm": 0.12320085614919662,
      "learning_rate": 1e-06,
      "loss": -0.0204,
      "step": 483
    },
    {
      "clip_ratio/high_max": 0.002540133020374924,
      "clip_ratio/high_mean": 0.0009502205120952567,
      "clip_ratio/low_mean": 0.0013773989558103494,
      "clip_ratio/low_min": 0.00021999615364620695,
      "clip_ratio/region_mean": 0.0023276194187928922,
      "epoch": 7.335860058309038,
      "grad_norm": 0.11977312713861465,
      "learning_rate": 1e-06,
      "loss": 0.0406,
      "step": 484
    },
    {
      "clip_ratio/high_max": 0.002486824589141179,
      "clip_ratio/high_mean": 0.001049243288434809,
      "clip_ratio/low_mean": 0.0010668040977179771,
      "clip_ratio/low_min": 0.00011340638320689322,
      "clip_ratio/region_mean": 0.0021160473697818816,
      "epoch": 7.345189504373177,
      "grad_norm": 0.12062793225049973,
      "learning_rate": 1e-06,
      "loss": -0.0119,
      "step": 485
    },
    {
      "clip_ratio/high_max": 0.00300940345186973,
      "clip_ratio/high_mean": 0.001183753051009262,
      "clip_ratio/low_mean": 0.0011702056508511305,
      "clip_ratio/low_min": 4.2600540837156586e-05,
      "clip_ratio/region_mean": 0.002353958727326244,
      "epoch": 7.354518950437318,
      "grad_norm": 0.1330283284187317,
      "learning_rate": 1e-06,
      "loss": -0.03,
      "step": 486
    },
    {
      "clip_ratio/high_max": 0.0028180379085824825,
      "clip_ratio/high_mean": 0.0011440986272646114,
      "clip_ratio/low_mean": 0.0011010474700015038,
      "clip_ratio/low_min": 4.2307438889110927e-05,
      "clip_ratio/region_mean": 0.0022451460899901576,
      "epoch": 7.363848396501457,
      "grad_norm": 0.12328824400901794,
      "learning_rate": 1e-06,
      "loss": -0.0719,
      "step": 487
    },
    {
      "clip_ratio/high_max": 0.0022763478991691954,
      "clip_ratio/high_mean": 0.0009702771349111572,
      "clip_ratio/low_mean": 0.0012585149852384347,
      "clip_ratio/low_min": 0.00011112267930002417,
      "clip_ratio/region_mean": 0.0022287920728558674,
      "epoch": 7.373177842565598,
      "grad_norm": 0.11739755421876907,
      "learning_rate": 1e-06,
      "loss": 0.021,
      "step": 488
    },
    {
      "clip_ratio/high_max": 0.002318337545148097,
      "clip_ratio/high_mean": 0.0008995488115033368,
      "clip_ratio/low_mean": 0.0012490651024563704,
      "clip_ratio/low_min": 0.00013070694421912776,
      "clip_ratio/region_mean": 0.002148613908502739,
      "epoch": 7.382507288629737,
      "grad_norm": 0.12185444682836533,
      "learning_rate": 1e-06,
      "loss": 0.0232,
      "step": 489
    },
    {
      "clip_ratio/high_max": 0.002299207051692065,
      "clip_ratio/high_mean": 0.0010201399891229812,
      "clip_ratio/low_mean": 0.0012904405848530587,
      "clip_ratio/low_min": 7.079569604684366e-05,
      "clip_ratio/region_mean": 0.002310580624907743,
      "epoch": 7.391836734693878,
      "grad_norm": 0.13964992761611938,
      "learning_rate": 1e-06,
      "loss": 0.0224,
      "step": 490
    },
    {
      "clip_ratio/high_max": 0.0024046644393820316,
      "clip_ratio/high_mean": 0.0009953425542335026,
      "clip_ratio/low_mean": 0.0012848304431827273,
      "clip_ratio/low_min": 9.583777227817336e-05,
      "clip_ratio/region_mean": 0.002280172942846548,
      "epoch": 7.401166180758017,
      "grad_norm": 0.1222560703754425,
      "learning_rate": 1e-06,
      "loss": -0.0026,
      "step": 491
    },
    {
      "clip_ratio/high_max": 0.002676264019100927,
      "clip_ratio/high_mean": 0.0011157149474456673,
      "clip_ratio/low_mean": 0.00140094916059752,
      "clip_ratio/low_min": 9.902396141114878e-05,
      "clip_ratio/region_mean": 0.0025166641717078164,
      "epoch": 7.410495626822158,
      "grad_norm": 0.11028287559747696,
      "learning_rate": 1e-06,
      "loss": 0.0072,
      "step": 492
    },
    {
      "clip_ratio/high_max": 0.0022441234323196113,
      "clip_ratio/high_mean": 0.0008991280119516887,
      "clip_ratio/low_mean": 0.0012561939838633407,
      "clip_ratio/low_min": 7.770371576043544e-05,
      "clip_ratio/region_mean": 0.002155322006728966,
      "epoch": 7.419825072886297,
      "grad_norm": 0.1132696196436882,
      "learning_rate": 1e-06,
      "loss": 0.022,
      "step": 493
    },
    {
      "clip_ratio/high_max": 0.0025834246989688836,
      "clip_ratio/high_mean": 0.001081963720935164,
      "clip_ratio/low_mean": 0.0013424667704384774,
      "clip_ratio/low_min": 0.00011595202522585168,
      "clip_ratio/region_mean": 0.002424430502287578,
      "epoch": 7.429154518950437,
      "grad_norm": 0.11117098480463028,
      "learning_rate": 1e-06,
      "loss": 0.0305,
      "step": 494
    },
    {
      "clip_ratio/high_max": 0.002640688711835537,
      "clip_ratio/high_mean": 0.0011757747306546662,
      "clip_ratio/low_mean": 0.0010532915075600613,
      "clip_ratio/low_min": 6.40508187643718e-05,
      "clip_ratio/region_mean": 0.002229066281870473,
      "epoch": 7.438483965014577,
      "grad_norm": 0.12091293931007385,
      "learning_rate": 1e-06,
      "loss": -0.038,
      "step": 495
    },
    {
      "clip_ratio/high_max": 0.0028101488860556856,
      "clip_ratio/high_mean": 0.001151785887486767,
      "clip_ratio/low_mean": 0.0011464080089353956,
      "clip_ratio/low_min": 5.842746213602368e-05,
      "clip_ratio/region_mean": 0.002298193867318332,
      "epoch": 7.447813411078717,
      "grad_norm": 0.11562146991491318,
      "learning_rate": 1e-06,
      "loss": -0.0246,
      "step": 496
    },
    {
      "clip_ratio/high_max": 0.0026234120814478956,
      "clip_ratio/high_mean": 0.0010815629375429126,
      "clip_ratio/low_mean": 0.00119368602099712,
      "clip_ratio/low_min": 6.950084389245603e-05,
      "clip_ratio/region_mean": 0.002275248960359022,
      "epoch": 7.457142857142857,
      "grad_norm": 0.11595509201288223,
      "learning_rate": 1e-06,
      "loss": -0.0112,
      "step": 497
    },
    {
      "clip_ratio/high_max": 0.0029682919339393266,
      "clip_ratio/high_mean": 0.001281288899917854,
      "clip_ratio/low_mean": 0.0010473563852428924,
      "clip_ratio/low_min": 7.987911521922797e-05,
      "clip_ratio/region_mean": 0.0023286452633328736,
      "epoch": 7.466472303206997,
      "grad_norm": 0.13541395962238312,
      "learning_rate": 1e-06,
      "loss": 0.0254,
      "step": 498
    },
    {
      "clip_ratio/high_max": 0.00292497353075305,
      "clip_ratio/high_mean": 0.001249731027201051,
      "clip_ratio/low_mean": 0.0012320617788645905,
      "clip_ratio/low_min": 2.4895438400562853e-05,
      "clip_ratio/region_mean": 0.002481792827893514,
      "epoch": 7.475801749271137,
      "grad_norm": 0.12132713943719864,
      "learning_rate": 1e-06,
      "loss": 0.0039,
      "step": 499
    },
    {
      "clip_ratio/high_max": 0.0030203382921172306,
      "clip_ratio/high_mean": 0.0012031479163852055,
      "clip_ratio/low_mean": 0.0010051077952084597,
      "clip_ratio/low_min": 0.00012008448720735032,
      "clip_ratio/region_mean": 0.0022082556897657923,
      "epoch": 7.485131195335277,
      "grad_norm": 0.13366498053073883,
      "learning_rate": 1e-06,
      "loss": -0.0249,
      "step": 500
    },
    {
      "clip_ratio/high_max": 0.002366674401855562,
      "clip_ratio/high_mean": 0.0011580788050196134,
      "clip_ratio/low_mean": 0.0011073204150306992,
      "clip_ratio/low_min": 3.966103213315364e-05,
      "clip_ratio/region_mean": 0.0022653992564301006,
      "epoch": 7.494460641399417,
      "grad_norm": 0.1229439452290535,
      "learning_rate": 1e-06,
      "loss": -0.002,
      "step": 501
    },
    {
      "clip_ratio/high_max": 0.0026265027190675028,
      "clip_ratio/high_mean": 0.0010871550020965515,
      "clip_ratio/low_mean": 0.001378980334266089,
      "clip_ratio/low_min": 9.8238020655117e-05,
      "clip_ratio/region_mean": 0.002466135352733545,
      "epoch": 7.503790087463557,
      "grad_norm": 0.12062812596559525,
      "learning_rate": 1e-06,
      "loss": 0.0455,
      "step": 502
    },
    {
      "clip_ratio/high_max": 0.0027687020337907597,
      "clip_ratio/high_mean": 0.0010825806493812706,
      "clip_ratio/low_mean": 0.0011020002893928904,
      "clip_ratio/low_min": 5.1435548812150955e-05,
      "clip_ratio/region_mean": 0.0021845809460501187,
      "epoch": 7.513119533527696,
      "grad_norm": 0.11830519884824753,
      "learning_rate": 1e-06,
      "loss": -0.0148,
      "step": 503
    },
    {
      "clip_ratio/high_max": 0.0025952173164114356,
      "clip_ratio/high_mean": 0.001042579371642205,
      "clip_ratio/low_mean": 0.0011546709829417523,
      "clip_ratio/low_min": 7.714566891081631e-05,
      "clip_ratio/region_mean": 0.0021972503236611374,
      "epoch": 7.522448979591837,
      "grad_norm": 0.13077348470687866,
      "learning_rate": 1e-06,
      "loss": 0.0284,
      "step": 504
    },
    {
      "clip_ratio/high_max": 0.002429267842671834,
      "clip_ratio/high_mean": 0.0010484676458872855,
      "clip_ratio/low_mean": 0.0013895470146962907,
      "clip_ratio/low_min": 5.0701567488431465e-05,
      "clip_ratio/region_mean": 0.0024380146569455974,
      "epoch": 7.531778425655976,
      "grad_norm": 0.12644539773464203,
      "learning_rate": 1e-06,
      "loss": 0.0141,
      "step": 505
    },
    {
      "clip_ratio/high_max": 0.00278996788256336,
      "clip_ratio/high_mean": 0.0012562477604660671,
      "clip_ratio/low_mean": 0.0014506569168588612,
      "clip_ratio/low_min": 5.539823541766964e-05,
      "clip_ratio/region_mean": 0.0027069046191172674,
      "epoch": 7.541107871720117,
      "grad_norm": 0.13555505871772766,
      "learning_rate": 1e-06,
      "loss": 0.0127,
      "step": 506
    },
    {
      "clip_ratio/high_max": 0.00280632284557214,
      "clip_ratio/high_mean": 0.0011085463847848587,
      "clip_ratio/low_mean": 0.0013826628965034615,
      "clip_ratio/low_min": 0.00018379312132310588,
      "clip_ratio/region_mean": 0.0024912092194426805,
      "epoch": 7.550437317784256,
      "grad_norm": 0.12285593152046204,
      "learning_rate": 1e-06,
      "loss": -0.0052,
      "step": 507
    },
    {
      "clip_ratio/high_max": 0.0026030231747427024,
      "clip_ratio/high_mean": 0.0011438296223786892,
      "clip_ratio/low_mean": 0.0013101555305183865,
      "clip_ratio/low_min": 0.00020059679991391022,
      "clip_ratio/region_mean": 0.0024539851438021287,
      "epoch": 7.559766763848397,
      "grad_norm": 0.12937642633914948,
      "learning_rate": 1e-06,
      "loss": -0.0163,
      "step": 508
    },
    {
      "clip_ratio/high_max": 0.002509424688469153,
      "clip_ratio/high_mean": 0.0010707279761845712,
      "clip_ratio/low_mean": 0.0013914137198298704,
      "clip_ratio/low_min": 0.0002121368579537375,
      "clip_ratio/region_mean": 0.0024621416814625263,
      "epoch": 7.569096209912536,
      "grad_norm": 0.1315005123615265,
      "learning_rate": 1e-06,
      "loss": 0.0272,
      "step": 509
    },
    {
      "clip_ratio/high_max": 0.002701772638829425,
      "clip_ratio/high_mean": 0.0012115241843275726,
      "clip_ratio/low_mean": 0.0012911251669720514,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002502649374946486,
      "epoch": 7.578425655976677,
      "grad_norm": 0.12416743487119675,
      "learning_rate": 1e-06,
      "loss": -0.0084,
      "step": 510
    },
    {
      "clip_ratio/high_max": 0.002634147080243565,
      "clip_ratio/high_mean": 0.00117255356417445,
      "clip_ratio/low_mean": 0.0011262230727879796,
      "clip_ratio/low_min": 6.877614850964164e-05,
      "clip_ratio/region_mean": 0.002298776584211737,
      "epoch": 7.587755102040816,
      "grad_norm": 0.11074145138263702,
      "learning_rate": 1e-06,
      "loss": -0.0347,
      "step": 511
    },
    {
      "clip_ratio/high_max": 0.0025929158146027476,
      "clip_ratio/high_mean": 0.001219710407895036,
      "clip_ratio/low_mean": 0.001214678086398635,
      "clip_ratio/low_min": 0.00012000472815998364,
      "clip_ratio/region_mean": 0.0024343884870177135,
      "epoch": 7.597084548104956,
      "grad_norm": 0.11801495403051376,
      "learning_rate": 1e-06,
      "loss": -0.005,
      "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": 0.0298549107142857,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4090.0,
      "completions/mean_length": 664.7576904296875,
      "completions/mean_terminated_length": 559.1658325195312,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 8.00932944606414,
      "grad_norm": 0.1283450722694397,
      "learning_rate": 1e-06,
      "loss": -0.0177,
      "num_tokens": 335048009.0,
      "reward": 0.601318359375,
      "reward_std": 0.1860881894826889,
      "rewards/simpleverify_reward/mean": 0.601318359375,
      "rewards/simpleverify_reward/std": 0.48963120579719543,
      "step": 513
    },
    {
      "clip_ratio/high_max": 0.002308113791514188,
      "clip_ratio/high_mean": 0.0008192843579308828,
      "clip_ratio/low_mean": 0.0005590575974565581,
      "clip_ratio/low_min": 2.5047583221748937e-05,
      "clip_ratio/region_mean": 0.0013783419344690628,
      "epoch": 8.018658892128279,
      "grad_norm": 0.1203589215874672,
      "learning_rate": 1e-06,
      "loss": -0.0165,
      "step": 514
    },
    {
      "clip_ratio/high_max": 0.0021161512413527817,
      "clip_ratio/high_mean": 0.0008925096390157705,
      "clip_ratio/low_mean": 0.000621844644228986,
      "clip_ratio/low_min": 6.628180472034728e-05,
      "clip_ratio/region_mean": 0.0015143542623263784,
      "epoch": 8.02798833819242,
      "grad_norm": 0.126325324177742,
      "learning_rate": 1e-06,
      "loss": 0.0157,
      "step": 515
    },
    {
      "clip_ratio/high_max": 0.0024138694352586754,
      "clip_ratio/high_mean": 0.0009235054749296978,
      "clip_ratio/low_mean": 0.0005920641942793736,
      "clip_ratio/low_min": 3.052536521863658e-05,
      "clip_ratio/region_mean": 0.0015155696528381668,
      "epoch": 8.03731778425656,
      "grad_norm": 0.11950349062681198,
      "learning_rate": 1e-06,
      "loss": -0.0263,
      "step": 516
    },
    {
      "clip_ratio/high_max": 0.0024672703802934848,
      "clip_ratio/high_mean": 0.0009641834167268826,
      "clip_ratio/low_mean": 0.0006728258522343822,
      "clip_ratio/low_min": 2.986503386637196e-05,
      "clip_ratio/region_mean": 0.0016370092198485509,
      "epoch": 8.0466472303207,
      "grad_norm": 0.11088596284389496,
      "learning_rate": 1e-06,
      "loss": 0.0094,
      "step": 517
    },
    {
      "clip_ratio/high_max": 0.0023653750286030117,
      "clip_ratio/high_mean": 0.0008969077725851093,
      "clip_ratio/low_mean": 0.0006713878083246527,
      "clip_ratio/low_min": 6.067686717869947e-05,
      "clip_ratio/region_mean": 0.0015682956000091508,
      "epoch": 8.055976676384839,
      "grad_norm": 0.1267935335636139,
      "learning_rate": 1e-06,
      "loss": -0.0151,
      "step": 518
    },
    {
      "clip_ratio/high_max": 0.002127373474650085,
      "clip_ratio/high_mean": 0.0008364786281163106,
      "clip_ratio/low_mean": 0.000634467440249864,
      "clip_ratio/low_min": 6.502617270598421e-05,
      "clip_ratio/region_mean": 0.0014709460447193123,
      "epoch": 8.06530612244898,
      "grad_norm": 0.13391147553920746,
      "learning_rate": 1e-06,
      "loss": -0.0103,
      "step": 519
    },
    {
      "clip_ratio/high_max": 0.0022948138357605785,
      "clip_ratio/high_mean": 0.001019012230244698,
      "clip_ratio/low_mean": 0.000763603351515485,
      "clip_ratio/low_min": 3.301637480035424e-05,
      "clip_ratio/region_mean": 0.001782615545380395,
      "epoch": 8.07463556851312,
      "grad_norm": 0.12723638117313385,
      "learning_rate": 1e-06,
      "loss": -0.0179,
      "step": 520
    },
    {
      "clip_ratio/high_max": 0.0023748731291561853,
      "clip_ratio/high_mean": 0.0009810123592615128,
      "clip_ratio/low_mean": 0.0009004772473417688,
      "clip_ratio/low_min": 8.037665793381166e-05,
      "clip_ratio/region_mean": 0.0018814895738614723,
      "epoch": 8.08396501457726,
      "grad_norm": 0.1290731132030487,
      "learning_rate": 1e-06,
      "loss": 0.0121,
      "step": 521
    },
    {
      "clip_ratio/high_max": 0.002102063983329572,
      "clip_ratio/high_mean": 0.0009693754291220102,
      "clip_ratio/low_mean": 0.0007923240900709061,
      "clip_ratio/low_min": 4.505354900175007e-05,
      "clip_ratio/region_mean": 0.0017616995246498846,
      "epoch": 8.093294460641399,
      "grad_norm": 0.12019843608140945,
      "learning_rate": 1e-06,
      "loss": 0.0361,
      "step": 522
    },
    {
      "clip_ratio/high_max": 0.0025198195362463593,
      "clip_ratio/high_mean": 0.0010217487033514772,
      "clip_ratio/low_mean": 0.0008695691230968805,
      "clip_ratio/low_min": 6.385423421306768e-05,
      "clip_ratio/region_mean": 0.0018913177846116014,
      "epoch": 8.102623906705539,
      "grad_norm": 0.11757130920886993,
      "learning_rate": 1e-06,
      "loss": -0.0083,
      "step": 523
    },
    {
      "clip_ratio/high_max": 0.0026558830504654907,
      "clip_ratio/high_mean": 0.0010986463385052048,
      "clip_ratio/low_mean": 0.0007407835273625096,
      "clip_ratio/low_min": 5.301088822307065e-05,
      "clip_ratio/region_mean": 0.001839429882238619,
      "epoch": 8.11195335276968,
      "grad_norm": 0.1299556940793991,
      "learning_rate": 1e-06,
      "loss": -0.0375,
      "step": 524
    },
    {
      "clip_ratio/high_max": 0.0027868784964084625,
      "clip_ratio/high_mean": 0.0012555852845252957,
      "clip_ratio/low_mean": 0.0009408311379957013,
      "clip_ratio/low_min": 8.384651482629124e-05,
      "clip_ratio/region_mean": 0.002196416440710891,
      "epoch": 8.12128279883382,
      "grad_norm": 0.20711393654346466,
      "learning_rate": 1e-06,
      "loss": -0.0083,
      "step": 525
    },
    {
      "clip_ratio/high_max": 0.0025017392545123585,
      "clip_ratio/high_mean": 0.0010335601491533453,
      "clip_ratio/low_mean": 0.000903732990991557,
      "clip_ratio/low_min": 5.938757021795027e-05,
      "clip_ratio/region_mean": 0.0019372931201360188,
      "epoch": 8.130612244897959,
      "grad_norm": 0.13137264549732208,
      "learning_rate": 1e-06,
      "loss": -0.0095,
      "step": 526
    },
    {
      "clip_ratio/high_max": 0.002261149580590427,
      "clip_ratio/high_mean": 0.0009380212159157963,
      "clip_ratio/low_mean": 0.0009593696158844978,
      "clip_ratio/low_min": 0.00012619020617421484,
      "clip_ratio/region_mean": 0.0018973908154293895,
      "epoch": 8.139941690962099,
      "grad_norm": 0.12289122492074966,
      "learning_rate": 1e-06,
      "loss": -0.02,
      "step": 527
    },
    {
      "clip_ratio/high_max": 0.0023602232613484375,
      "clip_ratio/high_mean": 0.0009294587598560611,
      "clip_ratio/low_mean": 0.0008977753086583107,
      "clip_ratio/low_min": 6.272373320825864e-05,
      "clip_ratio/region_mean": 0.0018272340821567923,
      "epoch": 8.14927113702624,
      "grad_norm": 0.11843208223581314,
      "learning_rate": 1e-06,
      "loss": 0.0021,
      "step": 528
    },
    {
      "clip_ratio/high_max": 0.0027792146502179094,
      "clip_ratio/high_mean": 0.0011187990567123052,
      "clip_ratio/low_mean": 0.0010161497466469882,
      "clip_ratio/low_min": 8.624043675808934e-05,
      "clip_ratio/region_mean": 0.002134948816092219,
      "epoch": 8.15860058309038,
      "grad_norm": 0.14305076003074646,
      "learning_rate": 1e-06,
      "loss": 0.0006,
      "step": 529
    },
    {
      "clip_ratio/high_max": 0.002307569797267206,
      "clip_ratio/high_mean": 0.0009824995304370532,
      "clip_ratio/low_mean": 0.0009696361557871569,
      "clip_ratio/low_min": 8.904736841941485e-05,
      "clip_ratio/region_mean": 0.0019521356298355386,
      "epoch": 8.167930029154519,
      "grad_norm": 0.11763215065002441,
      "learning_rate": 1e-06,
      "loss": -0.0178,
      "step": 530
    },
    {
      "clip_ratio/high_max": 0.002417347248410806,
      "clip_ratio/high_mean": 0.0010127573004865553,
      "clip_ratio/low_mean": 0.0010786624043248594,
      "clip_ratio/low_min": 6.551867681992007e-05,
      "clip_ratio/region_mean": 0.002091419737553224,
      "epoch": 8.177259475218658,
      "grad_norm": 0.12113494426012039,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 531
    },
    {
      "clip_ratio/high_max": 0.0026721061367425136,
      "clip_ratio/high_mean": 0.001091474034183193,
      "clip_ratio/low_mean": 0.0009800746793189319,
      "clip_ratio/low_min": 6.606509850826114e-05,
      "clip_ratio/region_mean": 0.0020715487116831355,
      "epoch": 8.186588921282798,
      "grad_norm": 0.12066452950239182,
      "learning_rate": 1e-06,
      "loss": -0.0374,
      "step": 532
    },
    {
      "clip_ratio/high_max": 0.0024120587113429792,
      "clip_ratio/high_mean": 0.0009337450119346613,
      "clip_ratio/low_mean": 0.0010737554603110766,
      "clip_ratio/low_min": 0.00010615810242597945,
      "clip_ratio/region_mean": 0.0020075004940736108,
      "epoch": 8.19591836734694,
      "grad_norm": 0.12330883741378784,
      "learning_rate": 1e-06,
      "loss": 0.0088,
      "step": 533
    },
    {
      "clip_ratio/high_max": 0.0024619104879093356,
      "clip_ratio/high_mean": 0.0009373362427140819,
      "clip_ratio/low_mean": 0.001089069301087875,
      "clip_ratio/low_min": 0.00020098720142414095,
      "clip_ratio/region_mean": 0.0020264055565348826,
      "epoch": 8.205247813411079,
      "grad_norm": 0.12495143711566925,
      "learning_rate": 1e-06,
      "loss": 0.0387,
      "step": 534
    },
    {
      "clip_ratio/high_max": 0.0021820884139742702,
      "clip_ratio/high_mean": 0.001043860804202268,
      "clip_ratio/low_mean": 0.001224691775860265,
      "clip_ratio/low_min": 8.937551865528803e-05,
      "clip_ratio/region_mean": 0.0022685526782879606,
      "epoch": 8.214577259475218,
      "grad_norm": 0.12983602285385132,
      "learning_rate": 1e-06,
      "loss": -0.0295,
      "step": 535
    },
    {
      "clip_ratio/high_max": 0.0024420370391453616,
      "clip_ratio/high_mean": 0.0009962133681256091,
      "clip_ratio/low_mean": 0.0011644633668765891,
      "clip_ratio/low_min": 4.2788336031662766e-05,
      "clip_ratio/region_mean": 0.002160676784114912,
      "epoch": 8.223906705539358,
      "grad_norm": 0.12035165727138519,
      "learning_rate": 1e-06,
      "loss": 0.0106,
      "step": 536
    },
    {
      "clip_ratio/high_max": 0.0026347746461397037,
      "clip_ratio/high_mean": 0.0010284888685418991,
      "clip_ratio/low_mean": 0.001169606843177462,
      "clip_ratio/low_min": 6.066674359317403e-05,
      "clip_ratio/region_mean": 0.002198095760832075,
      "epoch": 8.2332361516035,
      "grad_norm": 0.11909116059541702,
      "learning_rate": 1e-06,
      "loss": -0.0095,
      "step": 537
    },
    {
      "clip_ratio/high_max": 0.002415731898508966,
      "clip_ratio/high_mean": 0.0009782180168258492,
      "clip_ratio/low_mean": 0.0012490935987443663,
      "clip_ratio/low_min": 7.025418744888157e-05,
      "clip_ratio/region_mean": 0.002227311626484152,
      "epoch": 8.242565597667639,
      "grad_norm": 0.11141186952590942,
      "learning_rate": 1e-06,
      "loss": 0.0441,
      "step": 538
    },
    {
      "clip_ratio/high_max": 0.003209587899618782,
      "clip_ratio/high_mean": 0.0011492324210848892,
      "clip_ratio/low_mean": 0.0010563824962446233,
      "clip_ratio/low_min": 7.435286534018815e-05,
      "clip_ratio/region_mean": 0.002205614946433343,
      "epoch": 8.251895043731778,
      "grad_norm": 0.12210160493850708,
      "learning_rate": 1e-06,
      "loss": -0.0027,
      "step": 539
    },
    {
      "clip_ratio/high_max": 0.0024030416971072555,
      "clip_ratio/high_mean": 0.0010791882777994033,
      "clip_ratio/low_mean": 0.0011722274102794472,
      "clip_ratio/low_min": 7.202471897471696e-05,
      "clip_ratio/region_mean": 0.002251415731734596,
      "epoch": 8.261224489795918,
      "grad_norm": 0.11297818273305893,
      "learning_rate": 1e-06,
      "loss": -0.002,
      "step": 540
    },
    {
      "clip_ratio/high_max": 0.0027120920858578756,
      "clip_ratio/high_mean": 0.0010439105390105397,
      "clip_ratio/low_mean": 0.0011526052039698698,
      "clip_ratio/low_min": 0.00010454573930473998,
      "clip_ratio/region_mean": 0.002196515735704452,
      "epoch": 8.270553935860057,
      "grad_norm": 0.1224856972694397,
      "learning_rate": 1e-06,
      "loss": 0.0208,
      "step": 541
    },
    {
      "clip_ratio/high_max": 0.0023931062205519993,
      "clip_ratio/high_mean": 0.0009649040293879807,
      "clip_ratio/low_mean": 0.001129512344050454,
      "clip_ratio/low_min": 7.973248648340814e-05,
      "clip_ratio/region_mean": 0.0020944163552485406,
      "epoch": 8.279883381924199,
      "grad_norm": 0.11838693171739578,
      "learning_rate": 1e-06,
      "loss": 0.0019,
      "step": 542
    },
    {
      "clip_ratio/high_max": 0.0024988237855723128,
      "clip_ratio/high_mean": 0.0010949193238047883,
      "clip_ratio/low_mean": 0.0010932493860309478,
      "clip_ratio/low_min": 4.4433726543502416e-05,
      "clip_ratio/region_mean": 0.002188168669817969,
      "epoch": 8.289212827988338,
      "grad_norm": 0.13246051967144012,
      "learning_rate": 1e-06,
      "loss": 0.0327,
      "step": 543
    },
    {
      "clip_ratio/high_max": 0.002739011964877136,
      "clip_ratio/high_mean": 0.0009972331954486435,
      "clip_ratio/low_mean": 0.0011938736060983501,
      "clip_ratio/low_min": 2.8269846552575473e-05,
      "clip_ratio/region_mean": 0.0021911067960900255,
      "epoch": 8.298542274052478,
      "grad_norm": 0.1231093630194664,
      "learning_rate": 1e-06,
      "loss": 0.0309,
      "step": 544
    },
    {
      "clip_ratio/high_max": 0.0027517258749867324,
      "clip_ratio/high_mean": 0.0012929725635331124,
      "clip_ratio/low_mean": 0.0010480324508534977,
      "clip_ratio/low_min": 8.57794439070858e-05,
      "clip_ratio/region_mean": 0.002341005005291663,
      "epoch": 8.307871720116617,
      "grad_norm": 0.1205020323395729,
      "learning_rate": 1e-06,
      "loss": -0.0242,
      "step": 545
    },
    {
      "clip_ratio/high_max": 0.00230911258404376,
      "clip_ratio/high_mean": 0.0010238617669529049,
      "clip_ratio/low_mean": 0.0012371332541079028,
      "clip_ratio/low_min": 0.00018028656995738856,
      "clip_ratio/region_mean": 0.0022609949519392103,
      "epoch": 8.317201166180759,
      "grad_norm": 0.12609079480171204,
      "learning_rate": 1e-06,
      "loss": 0.0361,
      "step": 546
    },
    {
      "clip_ratio/high_max": 0.0022343277887557633,
      "clip_ratio/high_mean": 0.0010184749244217528,
      "clip_ratio/low_mean": 0.0011197592430107761,
      "clip_ratio/low_min": 6.587505777133629e-05,
      "clip_ratio/region_mean": 0.0021382341801654547,
      "epoch": 8.326530612244898,
      "grad_norm": 0.1255784034729004,
      "learning_rate": 1e-06,
      "loss": -0.0124,
      "step": 547
    },
    {
      "clip_ratio/high_max": 0.0024673875923326705,
      "clip_ratio/high_mean": 0.0009784684116311837,
      "clip_ratio/low_mean": 0.0012355259750620462,
      "clip_ratio/low_min": 9.116447290580254e-05,
      "clip_ratio/region_mean": 0.002213994404883124,
      "epoch": 8.335860058309038,
      "grad_norm": 0.11374979466199875,
      "learning_rate": 1e-06,
      "loss": 0.0421,
      "step": 548
    },
    {
      "clip_ratio/high_max": 0.0028497118619270623,
      "clip_ratio/high_mean": 0.0011652501452772412,
      "clip_ratio/low_mean": 0.0012390206211421173,
      "clip_ratio/low_min": 4.902833552478114e-05,
      "clip_ratio/region_mean": 0.0024042707809712738,
      "epoch": 8.345189504373177,
      "grad_norm": 0.12544985115528107,
      "learning_rate": 1e-06,
      "loss": -0.004,
      "step": 549
    },
    {
      "clip_ratio/high_max": 0.002698931406484917,
      "clip_ratio/high_mean": 0.0012142670348112006,
      "clip_ratio/low_mean": 0.0011726453540177317,
      "clip_ratio/low_min": 1.7424030374968424e-05,
      "clip_ratio/region_mean": 0.002386912368820049,
      "epoch": 8.354518950437317,
      "grad_norm": 0.12661287188529968,
      "learning_rate": 1e-06,
      "loss": -0.0532,
      "step": 550
    },
    {
      "clip_ratio/high_max": 0.002647233435709495,
      "clip_ratio/high_mean": 0.000994792913843412,
      "clip_ratio/low_mean": 0.0011252933873038273,
      "clip_ratio/low_min": 4.500423347053584e-05,
      "clip_ratio/region_mean": 0.0021200862975092605,
      "epoch": 8.363848396501458,
      "grad_norm": 0.1246127262711525,
      "learning_rate": 1e-06,
      "loss": 0.03,
      "step": 551
    },
    {
      "clip_ratio/high_max": 0.0030342207101057284,
      "clip_ratio/high_mean": 0.0012242562188475858,
      "clip_ratio/low_mean": 0.0011163458912051283,
      "clip_ratio/low_min": 0.00012498951218731236,
      "clip_ratio/region_mean": 0.002340602077310905,
      "epoch": 8.373177842565598,
      "grad_norm": 0.12463736534118652,
      "learning_rate": 1e-06,
      "loss": -0.0148,
      "step": 552
    },
    {
      "clip_ratio/high_max": 0.0024178476305678487,
      "clip_ratio/high_mean": 0.0011164468160131946,
      "clip_ratio/low_mean": 0.0011468453139968915,
      "clip_ratio/low_min": 7.790819927322445e-05,
      "clip_ratio/region_mean": 0.002263292102725245,
      "epoch": 8.382507288629737,
      "grad_norm": 0.13012540340423584,
      "learning_rate": 1e-06,
      "loss": -0.0152,
      "step": 553
    },
    {
      "clip_ratio/high_max": 0.0026487481663934886,
      "clip_ratio/high_mean": 0.0011689358070725575,
      "clip_ratio/low_mean": 0.0013282163308758754,
      "clip_ratio/low_min": 0.00016393925579905044,
      "clip_ratio/region_mean": 0.002497152134310454,
      "epoch": 8.391836734693877,
      "grad_norm": 0.1218249574303627,
      "learning_rate": 1e-06,
      "loss": 0.0141,
      "step": 554
    },
    {
      "clip_ratio/high_max": 0.002430065687804017,
      "clip_ratio/high_mean": 0.001079453075362835,
      "clip_ratio/low_mean": 0.0011727480468834983,
      "clip_ratio/low_min": 8.196657290682197e-05,
      "clip_ratio/region_mean": 0.002252201098599471,
      "epoch": 8.401166180758018,
      "grad_norm": 0.12104137241840363,
      "learning_rate": 1e-06,
      "loss": 0.0165,
      "step": 555
    },
    {
      "clip_ratio/high_max": 0.0024578454613219947,
      "clip_ratio/high_mean": 0.0010770013213914353,
      "clip_ratio/low_mean": 0.001363682411465561,
      "clip_ratio/low_min": 0.00010510472839087015,
      "clip_ratio/region_mean": 0.0024406837910646573,
      "epoch": 8.410495626822158,
      "grad_norm": 0.1271679401397705,
      "learning_rate": 1e-06,
      "loss": 0.0297,
      "step": 556
    },
    {
      "clip_ratio/high_max": 0.0025669149981695227,
      "clip_ratio/high_mean": 0.0010316094776499085,
      "clip_ratio/low_mean": 0.0013937309922766872,
      "clip_ratio/low_min": 0.00010997794561262708,
      "clip_ratio/region_mean": 0.002425340462650638,
      "epoch": 8.419825072886297,
      "grad_norm": 0.11484405398368835,
      "learning_rate": 1e-06,
      "loss": 0.035,
      "step": 557
    },
    {
      "clip_ratio/high_max": 0.0028174765466246754,
      "clip_ratio/high_mean": 0.001194695392769063,
      "clip_ratio/low_mean": 0.0012002154344372684,
      "clip_ratio/low_min": 0.00014790886416449212,
      "clip_ratio/region_mean": 0.0023949107926455326,
      "epoch": 8.429154518950437,
      "grad_norm": 0.12489781528711319,
      "learning_rate": 1e-06,
      "loss": -0.0216,
      "step": 558
    },
    {
      "clip_ratio/high_max": 0.0020478023834584747,
      "clip_ratio/high_mean": 0.0009551603143336251,
      "clip_ratio/low_mean": 0.0013717133515456226,
      "clip_ratio/low_min": 0.0002706992481762427,
      "clip_ratio/region_mean": 0.002326873625861481,
      "epoch": 8.438483965014576,
      "grad_norm": 0.1298668384552002,
      "learning_rate": 1e-06,
      "loss": 0.0136,
      "step": 559
    },
    {
      "clip_ratio/high_max": 0.002336836994800251,
      "clip_ratio/high_mean": 0.001020310399326263,
      "clip_ratio/low_mean": 0.0012555011489894241,
      "clip_ratio/low_min": 0.00013547780417866306,
      "clip_ratio/region_mean": 0.0022758115374017507,
      "epoch": 8.447813411078718,
      "grad_norm": 0.11268424242734909,
      "learning_rate": 1e-06,
      "loss": -0.0434,
      "step": 560
    },
    {
      "clip_ratio/high_max": 0.0027585324132815003,
      "clip_ratio/high_mean": 0.0010997622666764073,
      "clip_ratio/low_mean": 0.0011211327837372664,
      "clip_ratio/low_min": 3.309066960355267e-05,
      "clip_ratio/region_mean": 0.0022208950758795254,
      "epoch": 8.457142857142857,
      "grad_norm": 0.11445292085409164,
      "learning_rate": 1e-06,
      "loss": -0.0114,
      "step": 561
    },
    {
      "clip_ratio/high_max": 0.0024568577791796997,
      "clip_ratio/high_mean": 0.0009768408581294352,
      "clip_ratio/low_mean": 0.001210312597322627,
      "clip_ratio/low_min": 0.00016917249013204128,
      "clip_ratio/region_mean": 0.0021871533899684437,
      "epoch": 8.466472303206997,
      "grad_norm": 0.12216100096702576,
      "learning_rate": 1e-06,
      "loss": -0.014,
      "step": 562
    },
    {
      "clip_ratio/high_max": 0.0026356438829679973,
      "clip_ratio/high_mean": 0.0011154105686728144,
      "clip_ratio/low_mean": 0.0012608977049239911,
      "clip_ratio/low_min": 8.81749529071385e-05,
      "clip_ratio/region_mean": 0.0023763082645018585,
      "epoch": 8.475801749271136,
      "grad_norm": 0.11466651409864426,
      "learning_rate": 1e-06,
      "loss": -0.0028,
      "step": 563
    },
    {
      "clip_ratio/high_max": 0.00265516534273047,
      "clip_ratio/high_mean": 0.001099496028473368,
      "clip_ratio/low_mean": 0.001142338413046673,
      "clip_ratio/low_min": 4.412595444591716e-05,
      "clip_ratio/region_mean": 0.0022418344815378077,
      "epoch": 8.485131195335278,
      "grad_norm": 0.12823975086212158,
      "learning_rate": 1e-06,
      "loss": -0.0133,
      "step": 564
    },
    {
      "clip_ratio/high_max": 0.002664219711732585,
      "clip_ratio/high_mean": 0.0011713347194017842,
      "clip_ratio/low_mean": 0.0010626559778756928,
      "clip_ratio/low_min": 6.0224479966564104e-05,
      "clip_ratio/region_mean": 0.002233990693639498,
      "epoch": 8.494460641399417,
      "grad_norm": 0.1139550507068634,
      "learning_rate": 1e-06,
      "loss": -0.023,
      "step": 565
    },
    {
      "clip_ratio/high_max": 0.0030304407991934568,
      "clip_ratio/high_mean": 0.0013030897425778676,
      "clip_ratio/low_mean": 0.0010486265764484415,
      "clip_ratio/low_min": 1.514050381956622e-05,
      "clip_ratio/region_mean": 0.002351716277189553,
      "epoch": 8.503790087463557,
      "grad_norm": 0.1180059164762497,
      "learning_rate": 1e-06,
      "loss": -0.0591,
      "step": 566
    },
    {
      "clip_ratio/high_max": 0.0029900972949690185,
      "clip_ratio/high_mean": 0.0011263616142969113,
      "clip_ratio/low_mean": 0.0011362304103386123,
      "clip_ratio/low_min": 5.1244685892015696e-05,
      "clip_ratio/region_mean": 0.002262592039187439,
      "epoch": 8.513119533527696,
      "grad_norm": 0.125900536775589,
      "learning_rate": 1e-06,
      "loss": -0.0057,
      "step": 567
    },
    {
      "clip_ratio/high_max": 0.0028118100890424103,
      "clip_ratio/high_mean": 0.0010995810516760685,
      "clip_ratio/low_mean": 0.0011372938515705755,
      "clip_ratio/low_min": 9.694679647509474e-05,
      "clip_ratio/region_mean": 0.0022368749632732943,
      "epoch": 8.522448979591836,
      "grad_norm": 0.12383028119802475,
      "learning_rate": 1e-06,
      "loss": -0.0127,
      "step": 568
    },
    {
      "clip_ratio/high_max": 0.002562097586633172,
      "clip_ratio/high_mean": 0.0011202461500943173,
      "clip_ratio/low_mean": 0.0011596060830925126,
      "clip_ratio/low_min": 3.097893568337895e-05,
      "clip_ratio/region_mean": 0.00227985223318683,
      "epoch": 8.531778425655977,
      "grad_norm": 0.11733996123075485,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "step": 569
    },
    {
      "clip_ratio/high_max": 0.0029617850013892166,
      "clip_ratio/high_mean": 0.0012063594731444027,
      "clip_ratio/low_mean": 0.001099165501727839,
      "clip_ratio/low_min": 0.00010354910409660079,
      "clip_ratio/region_mean": 0.002305525013071019,
      "epoch": 8.541107871720117,
      "grad_norm": 0.11968277394771576,
      "learning_rate": 1e-06,
      "loss": -0.0487,
      "step": 570
    },
    {
      "clip_ratio/high_max": 0.002759381903160829,
      "clip_ratio/high_mean": 0.001307042890402954,
      "clip_ratio/low_mean": 0.001245315230335109,
      "clip_ratio/low_min": 7.169775381044019e-05,
      "clip_ratio/region_mean": 0.0025523581716697663,
      "epoch": 8.550437317784256,
      "grad_norm": 0.12500333786010742,
      "learning_rate": 1e-06,
      "loss": 0.0054,
      "step": 571
    },
    {
      "clip_ratio/high_max": 0.002503732561308425,
      "clip_ratio/high_mean": 0.0010237629794573877,
      "clip_ratio/low_mean": 0.0013923473561590072,
      "clip_ratio/low_min": 6.724958439008333e-05,
      "clip_ratio/region_mean": 0.0024161103574442677,
      "epoch": 8.559766763848396,
      "grad_norm": 0.12822483479976654,
      "learning_rate": 1e-06,
      "loss": 0.0198,
      "step": 572
    },
    {
      "clip_ratio/high_max": 0.0029791240303893574,
      "clip_ratio/high_mean": 0.00127996172523126,
      "clip_ratio/low_mean": 0.0011200966910109855,
      "clip_ratio/low_min": 1.5089329281181563e-05,
      "clip_ratio/region_mean": 0.0024000583944143727,
      "epoch": 8.569096209912537,
      "grad_norm": 0.1319812536239624,
      "learning_rate": 1e-06,
      "loss": -0.0376,
      "step": 573
    },
    {
      "clip_ratio/high_max": 0.00259275332064135,
      "clip_ratio/high_mean": 0.0011809986899606884,
      "clip_ratio/low_mean": 0.0015884614949754905,
      "clip_ratio/low_min": 0.0001600734985913732,
      "clip_ratio/region_mean": 0.0027694601958501153,
      "epoch": 8.578425655976677,
      "grad_norm": 0.13094407320022583,
      "learning_rate": 1e-06,
      "loss": 0.0353,
      "step": 574
    },
    {
      "clip_ratio/high_max": 0.0023651821647945326,
      "clip_ratio/high_mean": 0.000938820505325566,
      "clip_ratio/low_mean": 0.001352846775262151,
      "clip_ratio/low_min": 6.29270625722711e-05,
      "clip_ratio/region_mean": 0.002291667216923088,
      "epoch": 8.587755102040816,
      "grad_norm": 0.11425367742776871,
      "learning_rate": 1e-06,
      "loss": 0.0165,
      "step": 575
    },
    {
      "clip_ratio/high_max": 0.002650808120961301,
      "clip_ratio/high_mean": 0.0010786241364257876,
      "clip_ratio/low_mean": 0.001199135702336207,
      "clip_ratio/low_min": 7.314833601412829e-05,
      "clip_ratio/region_mean": 0.0022777598787797615,
      "epoch": 8.597084548104956,
      "grad_norm": 0.11515863239765167,
      "learning_rate": 1e-06,
      "loss": 0.0034,
      "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": 0.0337437220982143,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4091.0,
      "completions/mean_length": 680.9325561523438,
      "completions/mean_terminated_length": 561.671142578125,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 9.00932944606414,
      "grad_norm": 0.13285647332668304,
      "learning_rate": 1e-06,
      "loss": -0.0147,
      "num_tokens": 371932110.0,
      "reward": 0.6011962890625,
      "reward_std": 0.18452098965644836,
      "rewards/simpleverify_reward/mean": 0.6011962890625,
      "rewards/simpleverify_reward/std": 0.4896564781665802,
      "step": 577
    },
    {
      "clip_ratio/high_max": 0.002095008887408767,
      "clip_ratio/high_mean": 0.0008718671524547972,
      "clip_ratio/low_mean": 0.0005618682589556556,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001433735407772474,
      "epoch": 9.018658892128279,
      "grad_norm": 0.13264068961143494,
      "learning_rate": 1e-06,
      "loss": 0.0151,
      "step": 578
    },
    {
      "clip_ratio/high_max": 0.0025229359016520903,
      "clip_ratio/high_mean": 0.0010117592555616284,
      "clip_ratio/low_mean": 0.0005826697615702869,
      "clip_ratio/low_min": 3.274198115832405e-05,
      "clip_ratio/region_mean": 0.0015944289916660637,
      "epoch": 9.02798833819242,
      "grad_norm": 0.12184672057628632,
      "learning_rate": 1e-06,
      "loss": -0.0197,
      "step": 579
    },
    {
      "clip_ratio/high_max": 0.002105564715748187,
      "clip_ratio/high_mean": 0.0008783109951764345,
      "clip_ratio/low_mean": 0.000537934067324386,
      "clip_ratio/low_min": 1.5927624190226197e-05,
      "clip_ratio/region_mean": 0.00141624506431981,
      "epoch": 9.03731778425656,
      "grad_norm": 0.12321247905492783,
      "learning_rate": 1e-06,
      "loss": -0.0258,
      "step": 580
    },
    {
      "clip_ratio/high_max": 0.0019943970619351603,
      "clip_ratio/high_mean": 0.0007671414259675657,
      "clip_ratio/low_mean": 0.0006085099275878747,
      "clip_ratio/low_min": 2.1079258658573963e-05,
      "clip_ratio/region_mean": 0.00137565133991302,
      "epoch": 9.0466472303207,
      "grad_norm": 0.11829300969839096,
      "learning_rate": 1e-06,
      "loss": 0.0153,
      "step": 581
    },
    {
      "clip_ratio/high_max": 0.001837521373090567,
      "clip_ratio/high_mean": 0.0007918200026324484,
      "clip_ratio/low_mean": 0.0007074921368257492,
      "clip_ratio/low_min": 7.013408139755484e-05,
      "clip_ratio/region_mean": 0.0014993121221777983,
      "epoch": 9.055976676384839,
      "grad_norm": 0.13303428888320923,
      "learning_rate": 1e-06,
      "loss": 0.0474,
      "step": 582
    },
    {
      "clip_ratio/high_max": 0.00226736001422978,
      "clip_ratio/high_mean": 0.0008523790684193955,
      "clip_ratio/low_mean": 0.0005297585084917955,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0013821375687257387,
      "epoch": 9.06530612244898,
      "grad_norm": 0.11975163966417313,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "step": 583
    },
    {
      "clip_ratio/high_max": 0.0024744837573962286,
      "clip_ratio/high_mean": 0.0009029937828017864,
      "clip_ratio/low_mean": 0.0006766312617401127,
      "clip_ratio/low_min": 3.183332773915026e-05,
      "clip_ratio/region_mean": 0.0015796250518178567,
      "epoch": 9.07463556851312,
      "grad_norm": 0.1264863908290863,
      "learning_rate": 1e-06,
      "loss": -0.0051,
      "step": 584
    },
    {
      "clip_ratio/high_max": 0.0024912997178034857,
      "clip_ratio/high_mean": 0.001027233785862336,
      "clip_ratio/low_mean": 0.0008041461042012088,
      "clip_ratio/low_min": 8.870780948200263e-05,
      "clip_ratio/region_mean": 0.0018313798718736507,
      "epoch": 9.08396501457726,
      "grad_norm": 0.13035285472869873,
      "learning_rate": 1e-06,
      "loss": -0.0025,
      "step": 585
    },
    {
      "clip_ratio/high_max": 0.002423805846774485,
      "clip_ratio/high_mean": 0.0010285060652677203,
      "clip_ratio/low_mean": 0.0006853322192910127,
      "clip_ratio/low_min": 2.946752283605747e-05,
      "clip_ratio/region_mean": 0.0017138382681878284,
      "epoch": 9.093294460641399,
      "grad_norm": 0.12404130399227142,
      "learning_rate": 1e-06,
      "loss": -0.0057,
      "step": 586
    },
    {
      "clip_ratio/high_max": 0.0024937425114330836,
      "clip_ratio/high_mean": 0.0009585611333022825,
      "clip_ratio/low_mean": 0.0007391421513602836,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001697703315585386,
      "epoch": 9.102623906705539,
      "grad_norm": 0.12626686692237854,
      "learning_rate": 1e-06,
      "loss": 0.0064,
      "step": 587
    },
    {
      "clip_ratio/high_max": 0.0019201044015062507,
      "clip_ratio/high_mean": 0.0008419741425313987,
      "clip_ratio/low_mean": 0.0008389076047023991,
      "clip_ratio/low_min": 4.0875421291275416e-05,
      "clip_ratio/region_mean": 0.001680881716310978,
      "epoch": 9.11195335276968,
      "grad_norm": 0.1348470002412796,
      "learning_rate": 1e-06,
      "loss": 0.0212,
      "step": 588
    },
    {
      "clip_ratio/high_max": 0.0023662338862777688,
      "clip_ratio/high_mean": 0.0009226638903783169,
      "clip_ratio/low_mean": 0.0008935322985053062,
      "clip_ratio/low_min": 8.200370484701125e-05,
      "clip_ratio/region_mean": 0.0018161961852456443,
      "epoch": 9.12128279883382,
      "grad_norm": 0.11456402391195297,
      "learning_rate": 1e-06,
      "loss": 0.0051,
      "step": 589
    },
    {
      "clip_ratio/high_max": 0.002215280426753452,
      "clip_ratio/high_mean": 0.0009347568302473519,
      "clip_ratio/low_mean": 0.0009179181470244657,
      "clip_ratio/low_min": 1.3326225598575547e-05,
      "clip_ratio/region_mean": 0.0018526749663578812,
      "epoch": 9.130612244897959,
      "grad_norm": 0.12748394906520844,
      "learning_rate": 1e-06,
      "loss": 0.0289,
      "step": 590
    },
    {
      "clip_ratio/high_max": 0.0024388823949266225,
      "clip_ratio/high_mean": 0.0010755289495136822,
      "clip_ratio/low_mean": 0.0009498595791228581,
      "clip_ratio/low_min": 2.879520798160229e-05,
      "clip_ratio/region_mean": 0.0020253884940757416,
      "epoch": 9.139941690962099,
      "grad_norm": 0.12352927029132843,
      "learning_rate": 1e-06,
      "loss": -0.0126,
      "step": 591
    },
    {
      "clip_ratio/high_max": 0.0024118531509884633,
      "clip_ratio/high_mean": 0.0010310264515283052,
      "clip_ratio/low_mean": 0.0008847234876157017,
      "clip_ratio/low_min": 3.9131849916884676e-05,
      "clip_ratio/region_mean": 0.001915749948238954,
      "epoch": 9.14927113702624,
      "grad_norm": 0.12940713763237,
      "learning_rate": 1e-06,
      "loss": -0.032,
      "step": 592
    },
    {
      "clip_ratio/high_max": 0.002132603374775499,
      "clip_ratio/high_mean": 0.0010033632170234341,
      "clip_ratio/low_mean": 0.0009103004704229534,
      "clip_ratio/low_min": 0.0001242578064193367,
      "clip_ratio/region_mean": 0.001913663676532451,
      "epoch": 9.15860058309038,
      "grad_norm": 0.11409001797437668,
      "learning_rate": 1e-06,
      "loss": -0.0087,
      "step": 593
    },
    {
      "clip_ratio/high_max": 0.00252408941014437,
      "clip_ratio/high_mean": 0.0010411773146188352,
      "clip_ratio/low_mean": 0.0008968510146587505,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001938028355652932,
      "epoch": 9.167930029154519,
      "grad_norm": 0.11580094695091248,
      "learning_rate": 1e-06,
      "loss": -0.0194,
      "step": 594
    },
    {
      "clip_ratio/high_max": 0.002397456963080913,
      "clip_ratio/high_mean": 0.0010751511490525445,
      "clip_ratio/low_mean": 0.0009203229292324977,
      "clip_ratio/low_min": 3.9574621951032896e-05,
      "clip_ratio/region_mean": 0.0019954740128014237,
      "epoch": 9.177259475218658,
      "grad_norm": 0.12042679637670517,
      "learning_rate": 1e-06,
      "loss": -0.0321,
      "step": 595
    },
    {
      "clip_ratio/high_max": 0.0026028020729427226,
      "clip_ratio/high_mean": 0.0009402143041370437,
      "clip_ratio/low_mean": 0.001097114090953255,
      "clip_ratio/low_min": 4.034861194668338e-05,
      "clip_ratio/region_mean": 0.0020373283623484895,
      "epoch": 9.186588921282798,
      "grad_norm": 0.11380136013031006,
      "learning_rate": 1e-06,
      "loss": 0.0082,
      "step": 596
    },
    {
      "clip_ratio/high_max": 0.0025542153452988714,
      "clip_ratio/high_mean": 0.0010863021307159215,
      "clip_ratio/low_mean": 0.0010616341896820813,
      "clip_ratio/low_min": 2.504006442904938e-05,
      "clip_ratio/region_mean": 0.002147936334949918,
      "epoch": 9.19591836734694,
      "grad_norm": 0.13990618288516998,
      "learning_rate": 1e-06,
      "loss": -0.0257,
      "step": 597
    },
    {
      "clip_ratio/high_max": 0.0024183921777876094,
      "clip_ratio/high_mean": 0.0010112084855791181,
      "clip_ratio/low_mean": 0.0010370101445005275,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002048218622803688,
      "epoch": 9.205247813411079,
      "grad_norm": 0.1352010816335678,
      "learning_rate": 1e-06,
      "loss": 0.0106,
      "step": 598
    },
    {
      "clip_ratio/high_max": 0.002286081580678001,
      "clip_ratio/high_mean": 0.00101827048274572,
      "clip_ratio/low_mean": 0.0010397060759714805,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020579765550792217,
      "epoch": 9.214577259475218,
      "grad_norm": 0.12656739354133606,
      "learning_rate": 1e-06,
      "loss": -0.0157,
      "step": 599
    },
    {
      "clip_ratio/high_max": 0.002966120846394915,
      "clip_ratio/high_mean": 0.00115546835695568,
      "clip_ratio/low_mean": 0.0009242282285413239,
      "clip_ratio/low_min": 5.58867359359283e-05,
      "clip_ratio/region_mean": 0.0020796966055058874,
      "epoch": 9.223906705539358,
      "grad_norm": 0.13320180773735046,
      "learning_rate": 1e-06,
      "loss": -0.0322,
      "step": 600
    },
    {
      "clip_ratio/high_max": 0.002570431104686577,
      "clip_ratio/high_mean": 0.001044825570716057,
      "clip_ratio/low_mean": 0.001054975482475129,
      "clip_ratio/low_min": 4.5328949454415124e-05,
      "clip_ratio/region_mean": 0.0020998011023039,
      "epoch": 9.2332361516035,
      "grad_norm": 0.12351791560649872,
      "learning_rate": 1e-06,
      "loss": -0.0275,
      "step": 601
    },
    {
      "clip_ratio/high_max": 0.002675545278179925,
      "clip_ratio/high_mean": 0.0010400641695014201,
      "clip_ratio/low_mean": 0.0011567278088477906,
      "clip_ratio/low_min": 7.891396489867475e-05,
      "clip_ratio/region_mean": 0.002196792025642935,
      "epoch": 9.242565597667639,
      "grad_norm": 0.1305057853460312,
      "learning_rate": 1e-06,
      "loss": 0.0043,
      "step": 602
    },
    {
      "clip_ratio/high_max": 0.0023576595849590376,
      "clip_ratio/high_mean": 0.0009686773337307386,
      "clip_ratio/low_mean": 0.0011383059954823693,
      "clip_ratio/low_min": 0.00010036851017503068,
      "clip_ratio/region_mean": 0.0021069833310320973,
      "epoch": 9.251895043731778,
      "grad_norm": 0.12467783689498901,
      "learning_rate": 1e-06,
      "loss": 0.0067,
      "step": 603
    },
    {
      "clip_ratio/high_max": 0.002176429348764941,
      "clip_ratio/high_mean": 0.0010041748682851903,
      "clip_ratio/low_mean": 0.00103925911753322,
      "clip_ratio/low_min": 7.47635403968161e-05,
      "clip_ratio/region_mean": 0.002043434033112135,
      "epoch": 9.261224489795918,
      "grad_norm": 0.12222645431756973,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "step": 604
    },
    {
      "clip_ratio/high_max": 0.0023096971854101866,
      "clip_ratio/high_mean": 0.0011084528614446754,
      "clip_ratio/low_mean": 0.0008928074948926223,
      "clip_ratio/low_min": 5.180273365112953e-05,
      "clip_ratio/region_mean": 0.0020012603272334673,
      "epoch": 9.270553935860057,
      "grad_norm": 0.12752173840999603,
      "learning_rate": 1e-06,
      "loss": -0.0462,
      "step": 605
    },
    {
      "clip_ratio/high_max": 0.002885396752390079,
      "clip_ratio/high_mean": 0.001113216520025162,
      "clip_ratio/low_mean": 0.0010250577224724111,
      "clip_ratio/low_min": 5.2374300139490515e-05,
      "clip_ratio/region_mean": 0.0021382742197602056,
      "epoch": 9.279883381924199,
      "grad_norm": 0.1273031085729599,
      "learning_rate": 1e-06,
      "loss": -0.035,
      "step": 606
    },
    {
      "clip_ratio/high_max": 0.0024283317252411507,
      "clip_ratio/high_mean": 0.0011151989856443834,
      "clip_ratio/low_mean": 0.0010842994015547447,
      "clip_ratio/low_min": 4.774637272930704e-05,
      "clip_ratio/region_mean": 0.0021994983835611492,
      "epoch": 9.289212827988338,
      "grad_norm": 0.12082916498184204,
      "learning_rate": 1e-06,
      "loss": -0.0084,
      "step": 607
    },
    {
      "clip_ratio/high_max": 0.0022554691677214578,
      "clip_ratio/high_mean": 0.00090830139015452,
      "clip_ratio/low_mean": 0.0011348240113875363,
      "clip_ratio/low_min": 9.679474896984175e-05,
      "clip_ratio/region_mean": 0.0020431253578863107,
      "epoch": 9.298542274052478,
      "grad_norm": 0.11502255499362946,
      "learning_rate": 1e-06,
      "loss": 0.0456,
      "step": 608
    },
    {
      "clip_ratio/high_max": 0.002281930428580381,
      "clip_ratio/high_mean": 0.0010584541523712687,
      "clip_ratio/low_mean": 0.001140832609962672,
      "clip_ratio/low_min": 0.0001009505576803349,
      "clip_ratio/region_mean": 0.002199286813265644,
      "epoch": 9.307871720116617,
      "grad_norm": 0.13609622418880463,
      "learning_rate": 1e-06,
      "loss": -0.0035,
      "step": 609
    },
    {
      "clip_ratio/high_max": 0.0022955050953896716,
      "clip_ratio/high_mean": 0.0009464829636272043,
      "clip_ratio/low_mean": 0.0012318073968344834,
      "clip_ratio/low_min": 4.103742685401812e-05,
      "clip_ratio/region_mean": 0.0021782904223073274,
      "epoch": 9.317201166180759,
      "grad_norm": 0.11110639572143555,
      "learning_rate": 1e-06,
      "loss": 0.0354,
      "step": 610
    },
    {
      "clip_ratio/high_max": 0.0026898417563643306,
      "clip_ratio/high_mean": 0.0011072184024669696,
      "clip_ratio/low_mean": 0.0012886570548289455,
      "clip_ratio/low_min": 0.00011487503797980025,
      "clip_ratio/region_mean": 0.002395875475485809,
      "epoch": 9.326530612244898,
      "grad_norm": 0.12480517476797104,
      "learning_rate": 1e-06,
      "loss": -0.0055,
      "step": 611
    },
    {
      "clip_ratio/high_max": 0.002011195130762644,
      "clip_ratio/high_mean": 0.0010132121260539861,
      "clip_ratio/low_mean": 0.001125232516642427,
      "clip_ratio/low_min": 0.00010078738341690041,
      "clip_ratio/region_mean": 0.0021384446299634874,
      "epoch": 9.335860058309038,
      "grad_norm": 0.10422385483980179,
      "learning_rate": 1e-06,
      "loss": 0.0158,
      "step": 612
    },
    {
      "clip_ratio/high_max": 0.0025812284147832543,
      "clip_ratio/high_mean": 0.0011202832974959165,
      "clip_ratio/low_mean": 0.0011645415434031747,
      "clip_ratio/low_min": 0.00012398919170664158,
      "clip_ratio/region_mean": 0.0022848248700029217,
      "epoch": 9.345189504373177,
      "grad_norm": 0.12176833301782608,
      "learning_rate": 1e-06,
      "loss": -0.0198,
      "step": 613
    },
    {
      "clip_ratio/high_max": 0.002298432886163937,
      "clip_ratio/high_mean": 0.0010270744842273416,
      "clip_ratio/low_mean": 0.0011916998264496215,
      "clip_ratio/low_min": 1.1743705726985354e-05,
      "clip_ratio/region_mean": 0.002218774359789677,
      "epoch": 9.354518950437317,
      "grad_norm": 0.12416286766529083,
      "learning_rate": 1e-06,
      "loss": -0.0283,
      "step": 614
    },
    {
      "clip_ratio/high_max": 0.0024842171660566237,
      "clip_ratio/high_mean": 0.0010853604235308012,
      "clip_ratio/low_mean": 0.0011723396019078791,
      "clip_ratio/low_min": 8.367688496946357e-05,
      "clip_ratio/region_mean": 0.0022577000709134154,
      "epoch": 9.363848396501458,
      "grad_norm": 0.12261302024126053,
      "learning_rate": 1e-06,
      "loss": -0.02,
      "step": 615
    },
    {
      "clip_ratio/high_max": 0.002542728660046123,
      "clip_ratio/high_mean": 0.0010006554966821568,
      "clip_ratio/low_mean": 0.0012999935970583465,
      "clip_ratio/low_min": 0.00010261273564537987,
      "clip_ratio/region_mean": 0.00230064907373162,
      "epoch": 9.373177842565598,
      "grad_norm": 0.12616237998008728,
      "learning_rate": 1e-06,
      "loss": -0.0085,
      "step": 616
    },
    {
      "clip_ratio/high_max": 0.0025822238967521116,
      "clip_ratio/high_mean": 0.0011296734628558625,
      "clip_ratio/low_mean": 0.001227458327775821,
      "clip_ratio/low_min": 5.39282946192543e-05,
      "clip_ratio/region_mean": 0.0023571317942696624,
      "epoch": 9.382507288629737,
      "grad_norm": 0.1137118861079216,
      "learning_rate": 1e-06,
      "loss": 0.0039,
      "step": 617
    },
    {
      "clip_ratio/high_max": 0.0022787383568356745,
      "clip_ratio/high_mean": 0.0010230451371171512,
      "clip_ratio/low_mean": 0.0011172243212058675,
      "clip_ratio/low_min": 8.193383473553695e-05,
      "clip_ratio/region_mean": 0.0021402694765129127,
      "epoch": 9.391836734693877,
      "grad_norm": 0.1310858279466629,
      "learning_rate": 1e-06,
      "loss": -0.0114,
      "step": 618
    },
    {
      "clip_ratio/high_max": 0.0028333611408015713,
      "clip_ratio/high_mean": 0.0010583788407529937,
      "clip_ratio/low_mean": 0.0013321573205757886,
      "clip_ratio/low_min": 0.0001295913098147139,
      "clip_ratio/region_mean": 0.0023905362031655386,
      "epoch": 9.401166180758018,
      "grad_norm": 0.13438986241817474,
      "learning_rate": 1e-06,
      "loss": 0.0471,
      "step": 619
    },
    {
      "clip_ratio/high_max": 0.00254293757097912,
      "clip_ratio/high_mean": 0.0011987317921011709,
      "clip_ratio/low_mean": 0.0011739305682567647,
      "clip_ratio/low_min": 7.440601257258095e-05,
      "clip_ratio/region_mean": 0.0023726623912807554,
      "epoch": 9.410495626822158,
      "grad_norm": 0.1205848976969719,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "step": 620
    },
    {
      "clip_ratio/high_max": 0.0026614868984324858,
      "clip_ratio/high_mean": 0.0009962680505850585,
      "clip_ratio/low_mean": 0.0012723621366603766,
      "clip_ratio/low_min": 0.0002207628476753598,
      "clip_ratio/region_mean": 0.00226863023272017,
      "epoch": 9.419825072886297,
      "grad_norm": 0.11237507313489914,
      "learning_rate": 1e-06,
      "loss": 0.0184,
      "step": 621
    },
    {
      "clip_ratio/high_max": 0.0026837940749828704,
      "clip_ratio/high_mean": 0.0011231761309318244,
      "clip_ratio/low_mean": 0.0011201181878277566,
      "clip_ratio/low_min": 1.602975135028828e-05,
      "clip_ratio/region_mean": 0.0022432943151216023,
      "epoch": 9.429154518950437,
      "grad_norm": 0.11535617709159851,
      "learning_rate": 1e-06,
      "loss": 0.0047,
      "step": 622
    },
    {
      "clip_ratio/high_max": 0.002706679268158041,
      "clip_ratio/high_mean": 0.0011671201209537685,
      "clip_ratio/low_mean": 0.001195083423226606,
      "clip_ratio/low_min": 6.795987519581104e-05,
      "clip_ratio/region_mean": 0.0023622034932486713,
      "epoch": 9.438483965014576,
      "grad_norm": 0.18078528344631195,
      "learning_rate": 1e-06,
      "loss": -0.0147,
      "step": 623
    },
    {
      "clip_ratio/high_max": 0.0026416391556267627,
      "clip_ratio/high_mean": 0.0011515861806401517,
      "clip_ratio/low_mean": 0.0011183592196175596,
      "clip_ratio/low_min": 3.810975613305345e-05,
      "clip_ratio/region_mean": 0.002269945412990637,
      "epoch": 9.447813411078718,
      "grad_norm": 0.12711338698863983,
      "learning_rate": 1e-06,
      "loss": -0.0082,
      "step": 624
    },
    {
      "clip_ratio/high_max": 0.0026272208997397684,
      "clip_ratio/high_mean": 0.0011124699121864978,
      "clip_ratio/low_mean": 0.0011897631957253907,
      "clip_ratio/low_min": 1.7796128304325975e-05,
      "clip_ratio/region_mean": 0.002302233115187846,
      "epoch": 9.457142857142857,
      "grad_norm": 0.11111123859882355,
      "learning_rate": 1e-06,
      "loss": 0.0005,
      "step": 625
    },
    {
      "clip_ratio/high_max": 0.0028883457998745143,
      "clip_ratio/high_mean": 0.0011402333620935678,
      "clip_ratio/low_mean": 0.001043206502799876,
      "clip_ratio/low_min": 5.6443625908286776e-05,
      "clip_ratio/region_mean": 0.002183439864893444,
      "epoch": 9.466472303206997,
      "grad_norm": 0.12038857489824295,
      "learning_rate": 1e-06,
      "loss": -0.0258,
      "step": 626
    },
    {
      "clip_ratio/high_max": 0.002864792331820354,
      "clip_ratio/high_mean": 0.0011926564111490734,
      "clip_ratio/low_mean": 0.0011926289735129103,
      "clip_ratio/low_min": 0.0001204743512062123,
      "clip_ratio/region_mean": 0.0023852853701100685,
      "epoch": 9.475801749271136,
      "grad_norm": 0.14227232336997986,
      "learning_rate": 1e-06,
      "loss": 0.0279,
      "step": 627
    },
    {
      "clip_ratio/high_max": 0.002558366431912873,
      "clip_ratio/high_mean": 0.0010591216378088575,
      "clip_ratio/low_mean": 0.0011159516870975494,
      "clip_ratio/low_min": 9.450146899325773e-05,
      "clip_ratio/region_mean": 0.002175073343096301,
      "epoch": 9.485131195335278,
      "grad_norm": 0.11924881488084793,
      "learning_rate": 1e-06,
      "loss": 0.011,
      "step": 628
    },
    {
      "clip_ratio/high_max": 0.0024306028644787148,
      "clip_ratio/high_mean": 0.0010062136061605997,
      "clip_ratio/low_mean": 0.0014047953554836567,
      "clip_ratio/low_min": 0.00024120151101669762,
      "clip_ratio/region_mean": 0.002411008972558193,
      "epoch": 9.494460641399417,
      "grad_norm": 0.11524834483861923,
      "learning_rate": 1e-06,
      "loss": 0.0282,
      "step": 629
    },
    {
      "clip_ratio/high_max": 0.0027105009721708484,
      "clip_ratio/high_mean": 0.0012046547562931664,
      "clip_ratio/low_mean": 0.001090580968593713,
      "clip_ratio/low_min": 6.118566216173349e-05,
      "clip_ratio/region_mean": 0.002295235695783049,
      "epoch": 9.503790087463557,
      "grad_norm": 4.149825096130371,
      "learning_rate": 1e-06,
      "loss": -0.0233,
      "step": 630
    },
    {
      "clip_ratio/high_max": 0.0022502470965264365,
      "clip_ratio/high_mean": 0.001068422930984525,
      "clip_ratio/low_mean": 0.0013431821171252523,
      "clip_ratio/low_min": 0.0001410236109222751,
      "clip_ratio/region_mean": 0.0024116049899021164,
      "epoch": 9.513119533527696,
      "grad_norm": 0.12179077416658401,
      "learning_rate": 1e-06,
      "loss": 0.0296,
      "step": 631
    },
    {
      "clip_ratio/high_max": 0.002633572967170039,
      "clip_ratio/high_mean": 0.0012700480001512915,
      "clip_ratio/low_mean": 0.00114913525612792,
      "clip_ratio/low_min": 0.00012557354421005584,
      "clip_ratio/region_mean": 0.0024191832198994234,
      "epoch": 9.522448979591836,
      "grad_norm": 0.12621724605560303,
      "learning_rate": 1e-06,
      "loss": -0.0546,
      "step": 632
    },
    {
      "clip_ratio/high_max": 0.0033828141604317352,
      "clip_ratio/high_mean": 0.0013177760229154956,
      "clip_ratio/low_mean": 0.001145079781053937,
      "clip_ratio/low_min": 0.0001697611123745446,
      "clip_ratio/region_mean": 0.0024628558530821465,
      "epoch": 9.531778425655977,
      "grad_norm": 0.12737131118774414,
      "learning_rate": 1e-06,
      "loss": -0.0104,
      "step": 633
    },
    {
      "clip_ratio/high_max": 0.002435559654259123,
      "clip_ratio/high_mean": 0.0010674743780327844,
      "clip_ratio/low_mean": 0.0010726855052780593,
      "clip_ratio/low_min": 6.857414427940967e-05,
      "clip_ratio/region_mean": 0.0021401599078672007,
      "epoch": 9.541107871720117,
      "grad_norm": 0.11658084392547607,
      "learning_rate": 1e-06,
      "loss": 0.0054,
      "step": 634
    },
    {
      "clip_ratio/high_max": 0.0029381924541667104,
      "clip_ratio/high_mean": 0.0012834291337640025,
      "clip_ratio/low_mean": 0.0012484374419727828,
      "clip_ratio/low_min": 2.2325415557133965e-05,
      "clip_ratio/region_mean": 0.002531866564822849,
      "epoch": 9.550437317784256,
      "grad_norm": 0.12238677591085434,
      "learning_rate": 1e-06,
      "loss": -0.0065,
      "step": 635
    },
    {
      "clip_ratio/high_max": 0.0028432906219677534,
      "clip_ratio/high_mean": 0.0012824474415538134,
      "clip_ratio/low_mean": 0.001128096177126281,
      "clip_ratio/low_min": 7.687859124416718e-05,
      "clip_ratio/region_mean": 0.002410543638688978,
      "epoch": 9.559766763848396,
      "grad_norm": 0.11918799579143524,
      "learning_rate": 1e-06,
      "loss": -0.0446,
      "step": 636
    },
    {
      "clip_ratio/high_max": 0.002640255057485774,
      "clip_ratio/high_mean": 0.0011537502614373807,
      "clip_ratio/low_mean": 0.0015121560973057058,
      "clip_ratio/low_min": 7.015793380560353e-05,
      "clip_ratio/region_mean": 0.0026659063005354255,
      "epoch": 9.569096209912537,
      "grad_norm": 0.18935620784759521,
      "learning_rate": 1e-06,
      "loss": 0.018,
      "step": 637
    },
    {
      "clip_ratio/high_max": 0.0026502571563469246,
      "clip_ratio/high_mean": 0.0011769416414608713,
      "clip_ratio/low_mean": 0.0012465130421333015,
      "clip_ratio/low_min": 3.7534337025135756e-05,
      "clip_ratio/region_mean": 0.0024234546726802364,
      "epoch": 9.578425655976677,
      "grad_norm": 0.12795966863632202,
      "learning_rate": 1e-06,
      "loss": -0.0058,
      "step": 638
    },
    {
      "clip_ratio/high_max": 0.002897474951168988,
      "clip_ratio/high_mean": 0.0012982441949134227,
      "clip_ratio/low_mean": 0.001345323344139615,
      "clip_ratio/low_min": 0.0001741444539220538,
      "clip_ratio/region_mean": 0.0026435674881213345,
      "epoch": 9.587755102040816,
      "grad_norm": 0.12297380715608597,
      "learning_rate": 1e-06,
      "loss": -0.019,
      "step": 639
    },
    {
      "clip_ratio/high_max": 0.002759557122772094,
      "clip_ratio/high_mean": 0.001197715235321084,
      "clip_ratio/low_mean": 0.0012491223606048152,
      "clip_ratio/low_min": 4.050881034345366e-05,
      "clip_ratio/region_mean": 0.0024468376068398356,
      "epoch": 9.597084548104956,
      "grad_norm": 0.11560143530368805,
      "learning_rate": 1e-06,
      "loss": 0.0147,
      "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": 0.036411830357142905,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4087.0,
      "completions/mean_length": 689.0717163085938,
      "completions/mean_terminated_length": 560.3316040039062,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 10.00932944606414,
      "grad_norm": 0.1392713338136673,
      "learning_rate": 1e-06,
      "loss": -0.032,
      "num_tokens": 408630472.0,
      "reward": 0.6119384765625,
      "reward_std": 0.180939719080925,
      "rewards/simpleverify_reward/mean": 0.6119384765625,
      "rewards/simpleverify_reward/std": 0.48731300234794617,
      "step": 641
    },
    {
      "clip_ratio/high_max": 0.002047834153927397,
      "clip_ratio/high_mean": 0.0008650950385344913,
      "clip_ratio/low_mean": 0.00047096276421143557,
      "clip_ratio/low_min": 5.592924571828917e-05,
      "clip_ratio/region_mean": 0.001336057797743706,
      "epoch": 10.018658892128279,
      "grad_norm": 0.13607341051101685,
      "learning_rate": 1e-06,
      "loss": -0.002,
      "step": 642
    },
    {
      "clip_ratio/high_max": 0.0022278865799307823,
      "clip_ratio/high_mean": 0.0009834486918407492,
      "clip_ratio/low_mean": 0.0004665450869651977,
      "clip_ratio/low_min": 9.999999747378752e-06,
      "clip_ratio/region_mean": 0.001449993760616053,
      "epoch": 10.02798833819242,
      "grad_norm": 0.14427782595157623,
      "learning_rate": 1e-06,
      "loss": -0.0355,
      "step": 643
    },
    {
      "clip_ratio/high_max": 0.00204466410286841,
      "clip_ratio/high_mean": 0.0008348910723725567,
      "clip_ratio/low_mean": 0.0006804180611652555,
      "clip_ratio/low_min": 1.4253135304898024e-05,
      "clip_ratio/region_mean": 0.001515309137175791,
      "epoch": 10.03731778425656,
      "grad_norm": 0.137557253241539,
      "learning_rate": 1e-06,
      "loss": 0.0036,
      "step": 644
    },
    {
      "clip_ratio/high_max": 0.0022067810787120834,
      "clip_ratio/high_mean": 0.0009102083386096638,
      "clip_ratio/low_mean": 0.0006301733064901782,
      "clip_ratio/low_min": 1.498441633884795e-05,
      "clip_ratio/region_mean": 0.0015403816432808526,
      "epoch": 10.0466472303207,
      "grad_norm": 0.12389273196458817,
      "learning_rate": 1e-06,
      "loss": -0.0166,
      "step": 645
    },
    {
      "clip_ratio/high_max": 0.002310533782292623,
      "clip_ratio/high_mean": 0.0009628772040741751,
      "clip_ratio/low_mean": 0.0007191718659669277,
      "clip_ratio/low_min": 0.0001007345927064307,
      "clip_ratio/region_mean": 0.0016820490891404916,
      "epoch": 10.055976676384839,
      "grad_norm": 0.122986800968647,
      "learning_rate": 1e-06,
      "loss": -0.0057,
      "step": 646
    },
    {
      "clip_ratio/high_max": 0.0022726235183654353,
      "clip_ratio/high_mean": 0.0009261680079362122,
      "clip_ratio/low_mean": 0.0007612035115016624,
      "clip_ratio/low_min": 8.911988334148191e-05,
      "clip_ratio/region_mean": 0.0016873714994289912,
      "epoch": 10.06530612244898,
      "grad_norm": 0.12592777609825134,
      "learning_rate": 1e-06,
      "loss": 0.004,
      "step": 647
    },
    {
      "clip_ratio/high_max": 0.002151245153072523,
      "clip_ratio/high_mean": 0.0009111481977015501,
      "clip_ratio/low_mean": 0.0009201193934131879,
      "clip_ratio/low_min": 9.33001347220852e-05,
      "clip_ratio/region_mean": 0.0018312676111236215,
      "epoch": 10.07463556851312,
      "grad_norm": 0.20106640458106995,
      "learning_rate": 1e-06,
      "loss": 0.0087,
      "step": 648
    },
    {
      "clip_ratio/high_max": 0.0021997812364134006,
      "clip_ratio/high_mean": 0.0009732377184263896,
      "clip_ratio/low_mean": 0.000970530974882422,
      "clip_ratio/low_min": 3.126488627458457e-05,
      "clip_ratio/region_mean": 0.0019437686787568964,
      "epoch": 10.08396501457726,
      "grad_norm": 0.11654148250818253,
      "learning_rate": 1e-06,
      "loss": 0.0082,
      "step": 649
    },
    {
      "clip_ratio/high_max": 0.002358433324843645,
      "clip_ratio/high_mean": 0.000947766635363223,
      "clip_ratio/low_mean": 0.0008258764564743615,
      "clip_ratio/low_min": 2.7778428375313524e-05,
      "clip_ratio/region_mean": 0.0017736431182129309,
      "epoch": 10.093294460641399,
      "grad_norm": 0.12819696962833405,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 650
    },
    {
      "clip_ratio/high_max": 0.0024035481619648635,
      "clip_ratio/high_mean": 0.0010478258845978417,
      "clip_ratio/low_mean": 0.0008721863960090559,
      "clip_ratio/low_min": 9.82086748990696e-06,
      "clip_ratio/region_mean": 0.001920012255141046,
      "epoch": 10.102623906705539,
      "grad_norm": 0.1123422384262085,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "step": 651
    },
    {
      "clip_ratio/high_max": 0.0023037685532472096,
      "clip_ratio/high_mean": 0.0010170627847401192,
      "clip_ratio/low_mean": 0.0008383774784306297,
      "clip_ratio/low_min": 3.5487936656863894e-05,
      "clip_ratio/region_mean": 0.001855440241342876,
      "epoch": 10.11195335276968,
      "grad_norm": 0.1292382776737213,
      "learning_rate": 1e-06,
      "loss": 0.0204,
      "step": 652
    },
    {
      "clip_ratio/high_max": 0.002392032190982718,
      "clip_ratio/high_mean": 0.0010367856884840876,
      "clip_ratio/low_mean": 0.0009508761158940615,
      "clip_ratio/low_min": 7.856137472117553e-05,
      "clip_ratio/region_mean": 0.0019876617734553292,
      "epoch": 10.12128279883382,
      "grad_norm": 0.1392720639705658,
      "learning_rate": 1e-06,
      "loss": -0.0116,
      "step": 653
    },
    {
      "clip_ratio/high_max": 0.0025919851104845293,
      "clip_ratio/high_mean": 0.001032770062010968,
      "clip_ratio/low_mean": 0.0008818659116514027,
      "clip_ratio/low_min": 1.6864543795236386e-05,
      "clip_ratio/region_mean": 0.0019146359481965192,
      "epoch": 10.130612244897959,
      "grad_norm": 0.16294026374816895,
      "learning_rate": 1e-06,
      "loss": -0.0109,
      "step": 654
    },
    {
      "clip_ratio/high_max": 0.0024329038860742003,
      "clip_ratio/high_mean": 0.0009516924801573623,
      "clip_ratio/low_mean": 0.000850626505780383,
      "clip_ratio/low_min": 6.167898027342744e-05,
      "clip_ratio/region_mean": 0.0018023189550149255,
      "epoch": 10.139941690962099,
      "grad_norm": 0.12228190153837204,
      "learning_rate": 1e-06,
      "loss": 0.0106,
      "step": 655
    },
    {
      "clip_ratio/high_max": 0.002595309360913234,
      "clip_ratio/high_mean": 0.0010521063704800326,
      "clip_ratio/low_mean": 0.0010832557891262695,
      "clip_ratio/low_min": 2.2494152290164493e-05,
      "clip_ratio/region_mean": 0.002135362185072154,
      "epoch": 10.14927113702624,
      "grad_norm": 0.1252080351114273,
      "learning_rate": 1e-06,
      "loss": 0.0118,
      "step": 656
    },
    {
      "clip_ratio/high_max": 0.0027463661135698203,
      "clip_ratio/high_mean": 0.0009227972986991517,
      "clip_ratio/low_mean": 0.0010185221362917218,
      "clip_ratio/low_min": 4.5262389903655276e-05,
      "clip_ratio/region_mean": 0.0019413193840591703,
      "epoch": 10.15860058309038,
      "grad_norm": 0.13458426296710968,
      "learning_rate": 1e-06,
      "loss": 0.0381,
      "step": 657
    },
    {
      "clip_ratio/high_max": 0.002248806515126489,
      "clip_ratio/high_mean": 0.001009656858514063,
      "clip_ratio/low_mean": 0.0012476445772335865,
      "clip_ratio/low_min": 5.814436281070812e-05,
      "clip_ratio/region_mean": 0.002257301428471692,
      "epoch": 10.167930029154519,
      "grad_norm": 0.12581020593643188,
      "learning_rate": 1e-06,
      "loss": 0.0137,
      "step": 658
    },
    {
      "clip_ratio/high_max": 0.0023237678869918454,
      "clip_ratio/high_mean": 0.0009907921230478678,
      "clip_ratio/low_mean": 0.0010463258113304619,
      "clip_ratio/low_min": 0.0001229994868481299,
      "clip_ratio/region_mean": 0.0020371178834466264,
      "epoch": 10.177259475218658,
      "grad_norm": 0.11078030616044998,
      "learning_rate": 1e-06,
      "loss": -0.0194,
      "step": 659
    },
    {
      "clip_ratio/high_max": 0.002240027672087308,
      "clip_ratio/high_mean": 0.0010262998184771277,
      "clip_ratio/low_mean": 0.0009901681496558012,
      "clip_ratio/low_min": 2.6652451197151095e-05,
      "clip_ratio/region_mean": 0.0020164679444860667,
      "epoch": 10.186588921282798,
      "grad_norm": 0.12092120945453644,
      "learning_rate": 1e-06,
      "loss": -0.0162,
      "step": 660
    },
    {
      "clip_ratio/high_max": 0.002720078598940745,
      "clip_ratio/high_mean": 0.0011530970223248005,
      "clip_ratio/low_mean": 0.0009657608370616799,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002118857912137173,
      "epoch": 10.19591836734694,
      "grad_norm": 0.11478060483932495,
      "learning_rate": 1e-06,
      "loss": -0.0625,
      "step": 661
    },
    {
      "clip_ratio/high_max": 0.002589636256743688,
      "clip_ratio/high_mean": 0.0010263687290716916,
      "clip_ratio/low_mean": 0.0010298481556674233,
      "clip_ratio/low_min": 4.5873623093939386e-05,
      "clip_ratio/region_mean": 0.0020562169156619348,
      "epoch": 10.205247813411079,
      "grad_norm": 0.12798132002353668,
      "learning_rate": 1e-06,
      "loss": -0.0017,
      "step": 662
    },
    {
      "clip_ratio/high_max": 0.0025797197959036566,
      "clip_ratio/high_mean": 0.0010226304239040473,
      "clip_ratio/low_mean": 0.0010953042183245998,
      "clip_ratio/low_min": 8.636050006316509e-05,
      "clip_ratio/region_mean": 0.0021179346294957213,
      "epoch": 10.214577259475218,
      "grad_norm": 0.13217544555664062,
      "learning_rate": 1e-06,
      "loss": 0.0054,
      "step": 663
    },
    {
      "clip_ratio/high_max": 0.002498672816727776,
      "clip_ratio/high_mean": 0.0010616590225254185,
      "clip_ratio/low_mean": 0.0011730836195056327,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002234742627479136,
      "epoch": 10.223906705539358,
      "grad_norm": 0.11584997922182083,
      "learning_rate": 1e-06,
      "loss": -0.0079,
      "step": 664
    },
    {
      "clip_ratio/high_max": 0.0022701531743223313,
      "clip_ratio/high_mean": 0.0009171811143460218,
      "clip_ratio/low_mean": 0.0010016705309681129,
      "clip_ratio/low_min": 4.604869445756776e-05,
      "clip_ratio/region_mean": 0.0019188516162103042,
      "epoch": 10.2332361516035,
      "grad_norm": 0.12133453041315079,
      "learning_rate": 1e-06,
      "loss": 0.0436,
      "step": 665
    },
    {
      "clip_ratio/high_max": 0.002220187958300812,
      "clip_ratio/high_mean": 0.0009100147872231901,
      "clip_ratio/low_mean": 0.001300268151680939,
      "clip_ratio/low_min": 2.7959726139670238e-05,
      "clip_ratio/region_mean": 0.002210282960732002,
      "epoch": 10.242565597667639,
      "grad_norm": 0.11900375783443451,
      "learning_rate": 1e-06,
      "loss": -0.0053,
      "step": 666
    },
    {
      "clip_ratio/high_max": 0.002403224028967088,
      "clip_ratio/high_mean": 0.0010102004207510618,
      "clip_ratio/low_mean": 0.0012406728637870401,
      "clip_ratio/low_min": 9.695904373074882e-05,
      "clip_ratio/region_mean": 0.002250873265438713,
      "epoch": 10.251895043731778,
      "grad_norm": 0.1301468163728714,
      "learning_rate": 1e-06,
      "loss": -0.0038,
      "step": 667
    },
    {
      "clip_ratio/high_max": 0.0027375788195058703,
      "clip_ratio/high_mean": 0.0009631810626160586,
      "clip_ratio/low_mean": 0.0011323421203996986,
      "clip_ratio/low_min": 5.215851251705317e-05,
      "clip_ratio/region_mean": 0.002095523159368895,
      "epoch": 10.261224489795918,
      "grad_norm": 0.11487710475921631,
      "learning_rate": 1e-06,
      "loss": 0.0007,
      "step": 668
    },
    {
      "clip_ratio/high_max": 0.00240037602634402,
      "clip_ratio/high_mean": 0.0011246110479987692,
      "clip_ratio/low_mean": 0.0010027175630966667,
      "clip_ratio/low_min": 1.436451384506654e-05,
      "clip_ratio/region_mean": 0.0021273286183713935,
      "epoch": 10.270553935860057,
      "grad_norm": 0.12883462011814117,
      "learning_rate": 1e-06,
      "loss": -0.0357,
      "step": 669
    },
    {
      "clip_ratio/high_max": 0.0024460777422063984,
      "clip_ratio/high_mean": 0.0009644778874644544,
      "clip_ratio/low_mean": 0.0010287969125784002,
      "clip_ratio/low_min": 6.640352057729615e-05,
      "clip_ratio/region_mean": 0.0019932748000428546,
      "epoch": 10.279883381924199,
      "grad_norm": 0.11545933037996292,
      "learning_rate": 1e-06,
      "loss": 0.0129,
      "step": 670
    },
    {
      "clip_ratio/high_max": 0.002622639702167362,
      "clip_ratio/high_mean": 0.0011205963382963091,
      "clip_ratio/low_mean": 0.0010524799108679872,
      "clip_ratio/low_min": 0.0001541666715638712,
      "clip_ratio/region_mean": 0.0021730762673541903,
      "epoch": 10.289212827988338,
      "grad_norm": 0.1184234619140625,
      "learning_rate": 1e-06,
      "loss": -0.0069,
      "step": 671
    },
    {
      "clip_ratio/high_max": 0.0026612194487825036,
      "clip_ratio/high_mean": 0.0011800450593000278,
      "clip_ratio/low_mean": 0.001131166485720314,
      "clip_ratio/low_min": 6.303721056610812e-05,
      "clip_ratio/region_mean": 0.002311211545020342,
      "epoch": 10.298542274052478,
      "grad_norm": 0.1345868855714798,
      "learning_rate": 1e-06,
      "loss": -0.0174,
      "step": 672
    },
    {
      "clip_ratio/high_max": 0.002358785131946206,
      "clip_ratio/high_mean": 0.0011492310877656564,
      "clip_ratio/low_mean": 0.001042951074850862,
      "clip_ratio/low_min": 0.00013271452189655975,
      "clip_ratio/region_mean": 0.002192182233557105,
      "epoch": 10.307871720116617,
      "grad_norm": 0.12077160179615021,
      "learning_rate": 1e-06,
      "loss": 0.0164,
      "step": 673
    },
    {
      "clip_ratio/high_max": 0.002329846676730085,
      "clip_ratio/high_mean": 0.0010293162886227947,
      "clip_ratio/low_mean": 0.000970665880231536,
      "clip_ratio/low_min": 2.9770025321340654e-05,
      "clip_ratio/region_mean": 0.0019999821670353413,
      "epoch": 10.317201166180759,
      "grad_norm": 0.11701883375644684,
      "learning_rate": 1e-06,
      "loss": 0.0141,
      "step": 674
    },
    {
      "clip_ratio/high_max": 0.0023262734903255478,
      "clip_ratio/high_mean": 0.0009797258080652682,
      "clip_ratio/low_mean": 0.0009553791642247234,
      "clip_ratio/low_min": 1.7269963791477494e-05,
      "clip_ratio/region_mean": 0.001935104992298875,
      "epoch": 10.326530612244898,
      "grad_norm": 0.12469476461410522,
      "learning_rate": 1e-06,
      "loss": 0.0126,
      "step": 675
    },
    {
      "clip_ratio/high_max": 0.002977002179250121,
      "clip_ratio/high_mean": 0.0011426664823375177,
      "clip_ratio/low_mean": 0.0009623113328416366,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002104977822455112,
      "epoch": 10.335860058309038,
      "grad_norm": 0.11522013694047928,
      "learning_rate": 1e-06,
      "loss": -0.0093,
      "step": 676
    },
    {
      "clip_ratio/high_max": 0.0025358507846249267,
      "clip_ratio/high_mean": 0.0011223089568375144,
      "clip_ratio/low_mean": 0.0012691402262134943,
      "clip_ratio/low_min": 7.694023042859044e-05,
      "clip_ratio/region_mean": 0.0023914492048788816,
      "epoch": 10.345189504373177,
      "grad_norm": 0.1352730393409729,
      "learning_rate": 1e-06,
      "loss": 0.0167,
      "step": 677
    },
    {
      "clip_ratio/high_max": 0.002420452430669684,
      "clip_ratio/high_mean": 0.0011734755935322028,
      "clip_ratio/low_mean": 0.0009007682801893679,
      "clip_ratio/low_min": 3.0252333999669645e-05,
      "clip_ratio/region_mean": 0.0020742438355227932,
      "epoch": 10.354518950437317,
      "grad_norm": 0.12467822432518005,
      "learning_rate": 1e-06,
      "loss": -0.0279,
      "step": 678
    },
    {
      "clip_ratio/high_max": 0.002822324739099713,
      "clip_ratio/high_mean": 0.0010816074500326067,
      "clip_ratio/low_mean": 0.0009819369370234199,
      "clip_ratio/low_min": 5.467043229145929e-05,
      "clip_ratio/region_mean": 0.0020635444088838995,
      "epoch": 10.363848396501458,
      "grad_norm": 0.11344575136899948,
      "learning_rate": 1e-06,
      "loss": -0.0348,
      "step": 679
    },
    {
      "clip_ratio/high_max": 0.002812238846672699,
      "clip_ratio/high_mean": 0.001204537635203451,
      "clip_ratio/low_mean": 0.0010692879404814448,
      "clip_ratio/low_min": 4.6952697630331386e-05,
      "clip_ratio/region_mean": 0.002273825608426705,
      "epoch": 10.373177842565598,
      "grad_norm": 0.13063548505306244,
      "learning_rate": 1e-06,
      "loss": -0.0351,
      "step": 680
    },
    {
      "clip_ratio/high_max": 0.003112280297500547,
      "clip_ratio/high_mean": 0.0011077583130827406,
      "clip_ratio/low_mean": 0.001213093175465474,
      "clip_ratio/low_min": 2.037822014244739e-05,
      "clip_ratio/region_mean": 0.0023208514976431616,
      "epoch": 10.382507288629737,
      "grad_norm": 0.12051547318696976,
      "learning_rate": 1e-06,
      "loss": 0.0171,
      "step": 681
    },
    {
      "clip_ratio/high_max": 0.002610737799841445,
      "clip_ratio/high_mean": 0.0011157767839904409,
      "clip_ratio/low_mean": 0.0013045312698523048,
      "clip_ratio/low_min": 5.999003042234108e-05,
      "clip_ratio/region_mean": 0.0024203080392908305,
      "epoch": 10.391836734693877,
      "grad_norm": 0.13831959664821625,
      "learning_rate": 1e-06,
      "loss": 0.0116,
      "step": 682
    },
    {
      "clip_ratio/high_max": 0.0028272876006667502,
      "clip_ratio/high_mean": 0.0011715241562342271,
      "clip_ratio/low_mean": 0.0010931353972409852,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022646595316473395,
      "epoch": 10.401166180758018,
      "grad_norm": 0.1293380856513977,
      "learning_rate": 1e-06,
      "loss": -0.0352,
      "step": 683
    },
    {
      "clip_ratio/high_max": 0.00275197607697919,
      "clip_ratio/high_mean": 0.0011955971167481039,
      "clip_ratio/low_mean": 0.001124811311456142,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023204084354802035,
      "epoch": 10.410495626822158,
      "grad_norm": 0.11363451182842255,
      "learning_rate": 1e-06,
      "loss": -0.0146,
      "step": 684
    },
    {
      "clip_ratio/high_max": 0.0025618452855269425,
      "clip_ratio/high_mean": 0.001123327234381577,
      "clip_ratio/low_mean": 0.001032562580803642,
      "clip_ratio/low_min": 1.3138532267475966e-05,
      "clip_ratio/region_mean": 0.0021558897860813886,
      "epoch": 10.419825072886297,
      "grad_norm": 0.11953888088464737,
      "learning_rate": 1e-06,
      "loss": -0.0254,
      "step": 685
    },
    {
      "clip_ratio/high_max": 0.00277746456413297,
      "clip_ratio/high_mean": 0.0010868894460145384,
      "clip_ratio/low_mean": 0.0009991054794227239,
      "clip_ratio/low_min": 7.77307304815622e-05,
      "clip_ratio/region_mean": 0.002085994950903114,
      "epoch": 10.429154518950437,
      "grad_norm": 0.12367735058069229,
      "learning_rate": 1e-06,
      "loss": 0.0017,
      "step": 686
    },
    {
      "clip_ratio/high_max": 0.0025198418443324044,
      "clip_ratio/high_mean": 0.0011185725234099664,
      "clip_ratio/low_mean": 0.0011533467932167696,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022719192929798737,
      "epoch": 10.438483965014576,
      "grad_norm": 0.1284836083650589,
      "learning_rate": 1e-06,
      "loss": -0.0126,
      "step": 687
    },
    {
      "clip_ratio/high_max": 0.0029426096443785354,
      "clip_ratio/high_mean": 0.0013686332567885984,
      "clip_ratio/low_mean": 0.0010915840393863618,
      "clip_ratio/low_min": 1.524762137705693e-05,
      "clip_ratio/region_mean": 0.0024602173070888966,
      "epoch": 10.447813411078718,
      "grad_norm": 0.11925153434276581,
      "learning_rate": 1e-06,
      "loss": -0.0365,
      "step": 688
    },
    {
      "clip_ratio/high_max": 0.0026962149859173223,
      "clip_ratio/high_mean": 0.001117982745199697,
      "clip_ratio/low_mean": 0.001111999030399602,
      "clip_ratio/low_min": 5.7968771216110326e-05,
      "clip_ratio/region_mean": 0.0022299817901512142,
      "epoch": 10.457142857142857,
      "grad_norm": 0.12644325196743011,
      "learning_rate": 1e-06,
      "loss": 0.0007,
      "step": 689
    },
    {
      "clip_ratio/high_max": 0.00284798932261765,
      "clip_ratio/high_mean": 0.0011122077685286058,
      "clip_ratio/low_mean": 0.0013346712476050016,
      "clip_ratio/low_min": 0.0001674002287472831,
      "clip_ratio/region_mean": 0.002446878919727169,
      "epoch": 10.466472303206997,
      "grad_norm": 0.15829947590827942,
      "learning_rate": 1e-06,
      "loss": 0.0149,
      "step": 690
    },
    {
      "clip_ratio/high_max": 0.0028364363170112483,
      "clip_ratio/high_mean": 0.0011795817226811778,
      "clip_ratio/low_mean": 0.0011272262781858444,
      "clip_ratio/low_min": 5.4322379583027214e-05,
      "clip_ratio/region_mean": 0.0023068079390213825,
      "epoch": 10.475801749271136,
      "grad_norm": 0.12783053517341614,
      "learning_rate": 1e-06,
      "loss": -0.0039,
      "step": 691
    },
    {
      "clip_ratio/high_max": 0.002701826801057905,
      "clip_ratio/high_mean": 0.0011491954064695165,
      "clip_ratio/low_mean": 0.0010792308130476158,
      "clip_ratio/low_min": 6.33924973953981e-05,
      "clip_ratio/region_mean": 0.002228426186775323,
      "epoch": 10.485131195335278,
      "grad_norm": 0.13505765795707703,
      "learning_rate": 1e-06,
      "loss": 0.0011,
      "step": 692
    },
    {
      "clip_ratio/high_max": 0.002801874201395549,
      "clip_ratio/high_mean": 0.0011267380687058903,
      "clip_ratio/low_mean": 0.0012592140174092492,
      "clip_ratio/low_min": 6.472423046943732e-05,
      "clip_ratio/region_mean": 0.0023859520806581713,
      "epoch": 10.494460641399417,
      "grad_norm": 0.1198091208934784,
      "learning_rate": 1e-06,
      "loss": -0.006,
      "step": 693
    },
    {
      "clip_ratio/high_max": 0.00272663126088446,
      "clip_ratio/high_mean": 0.0011502458728500642,
      "clip_ratio/low_mean": 0.0011762642261601286,
      "clip_ratio/low_min": 5.099959162180312e-05,
      "clip_ratio/region_mean": 0.0023265101553988643,
      "epoch": 10.503790087463557,
      "grad_norm": 0.1183510273694992,
      "learning_rate": 1e-06,
      "loss": -0.0282,
      "step": 694
    },
    {
      "clip_ratio/high_max": 0.00226325225958135,
      "clip_ratio/high_mean": 0.0010219620453426614,
      "clip_ratio/low_mean": 0.0011328844866511645,
      "clip_ratio/low_min": 7.217723759822547e-05,
      "clip_ratio/region_mean": 0.0021548465665546246,
      "epoch": 10.513119533527696,
      "grad_norm": 0.11545974761247635,
      "learning_rate": 1e-06,
      "loss": 0.0347,
      "step": 695
    },
    {
      "clip_ratio/high_max": 0.0026324745631427504,
      "clip_ratio/high_mean": 0.0010099662504217122,
      "clip_ratio/low_mean": 0.0012412101968948264,
      "clip_ratio/low_min": 7.962570816744119e-05,
      "clip_ratio/region_mean": 0.002251176440040581,
      "epoch": 10.522448979591836,
      "grad_norm": 0.12273969501256943,
      "learning_rate": 1e-06,
      "loss": 0.007,
      "step": 696
    },
    {
      "clip_ratio/high_max": 0.002909318311139941,
      "clip_ratio/high_mean": 0.0013815920901834033,
      "clip_ratio/low_mean": 0.001027212639201025,
      "clip_ratio/low_min": 1.5206812349788379e-05,
      "clip_ratio/region_mean": 0.0024088047721306793,
      "epoch": 10.531778425655977,
      "grad_norm": 0.13470061123371124,
      "learning_rate": 1e-06,
      "loss": -0.0138,
      "step": 697
    },
    {
      "clip_ratio/high_max": 0.0028308058535913005,
      "clip_ratio/high_mean": 0.0011884839223057497,
      "clip_ratio/low_mean": 0.0012319489069341216,
      "clip_ratio/low_min": 4.3097268644487485e-05,
      "clip_ratio/region_mean": 0.002420432814687956,
      "epoch": 10.541107871720117,
      "grad_norm": 0.12553003430366516,
      "learning_rate": 1e-06,
      "loss": -0.0059,
      "step": 698
    },
    {
      "clip_ratio/high_max": 0.0024274217066704296,
      "clip_ratio/high_mean": 0.001128029794926988,
      "clip_ratio/low_mean": 0.001073542824087781,
      "clip_ratio/low_min": 7.914487468951847e-05,
      "clip_ratio/region_mean": 0.002201572591729928,
      "epoch": 10.550437317784256,
      "grad_norm": 0.13157522678375244,
      "learning_rate": 1e-06,
      "loss": -0.0077,
      "step": 699
    },
    {
      "clip_ratio/high_max": 0.0026281162499799393,
      "clip_ratio/high_mean": 0.001287640701775672,
      "clip_ratio/low_mean": 0.0010792171415232588,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023668578287470154,
      "epoch": 10.559766763848396,
      "grad_norm": 0.12107131630182266,
      "learning_rate": 1e-06,
      "loss": -0.0182,
      "step": 700
    },
    {
      "clip_ratio/high_max": 0.002921500949014444,
      "clip_ratio/high_mean": 0.0013067157597106416,
      "clip_ratio/low_mean": 0.0010316732441424392,
      "clip_ratio/low_min": 2.7345685339241754e-05,
      "clip_ratio/region_mean": 0.002338389022042975,
      "epoch": 10.569096209912537,
      "grad_norm": 0.11855398863554001,
      "learning_rate": 1e-06,
      "loss": -0.0406,
      "step": 701
    },
    {
      "clip_ratio/high_max": 0.0028459496825234964,
      "clip_ratio/high_mean": 0.0013375794660532847,
      "clip_ratio/low_mean": 0.0009570107904437464,
      "clip_ratio/low_min": 2.015478821704164e-05,
      "clip_ratio/region_mean": 0.0022945903037907556,
      "epoch": 10.578425655976677,
      "grad_norm": 0.11987543851137161,
      "learning_rate": 1e-06,
      "loss": -0.0312,
      "step": 702
    },
    {
      "clip_ratio/high_max": 0.0028598568751476705,
      "clip_ratio/high_mean": 0.0013222416673670523,
      "clip_ratio/low_mean": 0.0010981945051753428,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024204362198361196,
      "epoch": 10.587755102040816,
      "grad_norm": 0.12368352711200714,
      "learning_rate": 1e-06,
      "loss": -0.0507,
      "step": 703
    },
    {
      "clip_ratio/high_max": 0.0027670677736750804,
      "clip_ratio/high_mean": 0.0010696100234781625,
      "clip_ratio/low_mean": 0.0010607376643747557,
      "clip_ratio/low_min": 1.981610694201663e-05,
      "clip_ratio/region_mean": 0.002130347696947865,
      "epoch": 10.597084548104956,
      "grad_norm": 0.13114728033542633,
      "learning_rate": 1e-06,
      "loss": 0.0038,
      "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": 0.0411202566964286,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4067.0,
      "completions/mean_length": 704.5460815429688,
      "completions/mean_terminated_length": 559.1080322265625,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 11.00932944606414,
      "grad_norm": 0.12951792776584625,
      "learning_rate": 1e-06,
      "loss": 0.0076,
      "num_tokens": 445078717.0,
      "reward": 0.6148332953453064,
      "reward_std": 0.181431844830513,
      "rewards/simpleverify_reward/mean": 0.6148332953453064,
      "rewards/simpleverify_reward/std": 0.4866389334201813,
      "step": 705
    },
    {
      "clip_ratio/high_max": 0.0024895566384657286,
      "clip_ratio/high_mean": 0.001000732179818442,
      "clip_ratio/low_mean": 0.0005967781635263236,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015975103597156703,
      "epoch": 11.018658892128279,
      "grad_norm": 0.1346265971660614,
      "learning_rate": 1e-06,
      "loss": 0.0223,
      "step": 706
    },
    {
      "clip_ratio/high_max": 0.0019017110389540903,
      "clip_ratio/high_mean": 0.0008059542378759943,
      "clip_ratio/low_mean": 0.0006481111213361146,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014540653573931195,
      "epoch": 11.02798833819242,
      "grad_norm": 0.12726321816444397,
      "learning_rate": 1e-06,
      "loss": -0.0029,
      "step": 707
    },
    {
      "clip_ratio/high_max": 0.001742305763400509,
      "clip_ratio/high_mean": 0.0007656710840819869,
      "clip_ratio/low_mean": 0.0006633641851294669,
      "clip_ratio/low_min": 2.565629347373033e-05,
      "clip_ratio/region_mean": 0.001429035281034885,
      "epoch": 11.03731778425656,
      "grad_norm": 0.12016823142766953,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "step": 708
    },
    {
      "clip_ratio/high_max": 0.0022843821134301834,
      "clip_ratio/high_mean": 0.0009043302670761477,
      "clip_ratio/low_mean": 0.0005947845074842917,
      "clip_ratio/low_min": 3.191217911080457e-05,
      "clip_ratio/region_mean": 0.00149911477274145,
      "epoch": 11.0466472303207,
      "grad_norm": 0.11984176933765411,
      "learning_rate": 1e-06,
      "loss": -0.0322,
      "step": 709
    },
    {
      "clip_ratio/high_max": 0.0019807598291663453,
      "clip_ratio/high_mean": 0.0008598407857789425,
      "clip_ratio/low_mean": 0.0007699524157942506,
      "clip_ratio/low_min": 4.4958812395634595e-05,
      "clip_ratio/region_mean": 0.0016297931688313838,
      "epoch": 11.055976676384839,
      "grad_norm": 0.1348883956670761,
      "learning_rate": 1e-06,
      "loss": -0.0167,
      "step": 710
    },
    {
      "clip_ratio/high_max": 0.0025368011883983854,
      "clip_ratio/high_mean": 0.0010764137769001536,
      "clip_ratio/low_mean": 0.0006921420699654846,
      "clip_ratio/low_min": 6.029195265000453e-05,
      "clip_ratio/region_mean": 0.0017685558341327123,
      "epoch": 11.06530612244898,
      "grad_norm": 0.12955395877361298,
      "learning_rate": 1e-06,
      "loss": -0.0284,
      "step": 711
    },
    {
      "clip_ratio/high_max": 0.0019479884358588606,
      "clip_ratio/high_mean": 0.0008761910139583051,
      "clip_ratio/low_mean": 0.0008445257117273286,
      "clip_ratio/low_min": 4.405466461321339e-05,
      "clip_ratio/region_mean": 0.0017207166674779728,
      "epoch": 11.07463556851312,
      "grad_norm": 0.131577730178833,
      "learning_rate": 1e-06,
      "loss": -0.01,
      "step": 712
    },
    {
      "clip_ratio/high_max": 0.0024894445232348517,
      "clip_ratio/high_mean": 0.0009285551295761252,
      "clip_ratio/low_mean": 0.0008968574893515324,
      "clip_ratio/low_min": 3.550074870872777e-05,
      "clip_ratio/region_mean": 0.0018254126116517,
      "epoch": 11.08396501457726,
      "grad_norm": 0.1403147280216217,
      "learning_rate": 1e-06,
      "loss": 0.0326,
      "step": 713
    },
    {
      "clip_ratio/high_max": 0.0021842913993168622,
      "clip_ratio/high_mean": 0.000965073184488574,
      "clip_ratio/low_mean": 0.0007783819610267528,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017434551264159381,
      "epoch": 11.093294460641399,
      "grad_norm": 0.12377360463142395,
      "learning_rate": 1e-06,
      "loss": -0.0221,
      "step": 714
    },
    {
      "clip_ratio/high_max": 0.0026069017985719256,
      "clip_ratio/high_mean": 0.0010752888483693823,
      "clip_ratio/low_mean": 0.0007933147626317805,
      "clip_ratio/low_min": 4.6561715862480924e-05,
      "clip_ratio/region_mean": 0.001868603598268237,
      "epoch": 11.102623906705539,
      "grad_norm": 0.1331249624490738,
      "learning_rate": 1e-06,
      "loss": -0.0632,
      "step": 715
    },
    {
      "clip_ratio/high_max": 0.0023348920076387003,
      "clip_ratio/high_mean": 0.0009418703884875868,
      "clip_ratio/low_mean": 0.0009070927280845353,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018489631547708996,
      "epoch": 11.11195335276968,
      "grad_norm": 0.12545324862003326,
      "learning_rate": 1e-06,
      "loss": 0.013,
      "step": 716
    },
    {
      "clip_ratio/high_max": 0.0022802559251431376,
      "clip_ratio/high_mean": 0.0009894158720271662,
      "clip_ratio/low_mean": 0.0008970978014986031,
      "clip_ratio/low_min": 1.7720441974233836e-05,
      "clip_ratio/region_mean": 0.0018865136444219388,
      "epoch": 11.12128279883382,
      "grad_norm": 0.14978772401809692,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 717
    },
    {
      "clip_ratio/high_max": 0.0026689244841691107,
      "clip_ratio/high_mean": 0.0011391809239285067,
      "clip_ratio/low_mean": 0.0009504259451205144,
      "clip_ratio/low_min": 2.439982108626282e-05,
      "clip_ratio/region_mean": 0.0020896068745059893,
      "epoch": 11.130612244897959,
      "grad_norm": 0.14861105382442474,
      "learning_rate": 1e-06,
      "loss": 0.0112,
      "step": 718
    },
    {
      "clip_ratio/high_max": 0.00223810021270765,
      "clip_ratio/high_mean": 0.000930754715227522,
      "clip_ratio/low_mean": 0.001068526689778082,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001999281368625816,
      "epoch": 11.139941690962099,
      "grad_norm": 0.12481657415628433,
      "learning_rate": 1e-06,
      "loss": -0.0018,
      "step": 719
    },
    {
      "clip_ratio/high_max": 0.002785318487440236,
      "clip_ratio/high_mean": 0.001166293030109955,
      "clip_ratio/low_mean": 0.0010012327038566582,
      "clip_ratio/low_min": 2.1742911485489458e-05,
      "clip_ratio/region_mean": 0.0021675257085007615,
      "epoch": 11.14927113702624,
      "grad_norm": 0.11951804906129837,
      "learning_rate": 1e-06,
      "loss": -0.0406,
      "step": 720
    },
    {
      "clip_ratio/high_max": 0.0023329493415076286,
      "clip_ratio/high_mean": 0.0009614210248400923,
      "clip_ratio/low_mean": 0.0009302967100666137,
      "clip_ratio/low_min": 7.445224946422968e-05,
      "clip_ratio/region_mean": 0.0018917177185358014,
      "epoch": 11.15860058309038,
      "grad_norm": 0.12068294733762741,
      "learning_rate": 1e-06,
      "loss": -0.0064,
      "step": 721
    },
    {
      "clip_ratio/high_max": 0.0021991022731526755,
      "clip_ratio/high_mean": 0.0009441447273275116,
      "clip_ratio/low_mean": 0.0009330413085990585,
      "clip_ratio/low_min": 3.205401571904076e-05,
      "clip_ratio/region_mean": 0.0018771860777633265,
      "epoch": 11.167930029154519,
      "grad_norm": 0.11660280078649521,
      "learning_rate": 1e-06,
      "loss": 0.0035,
      "step": 722
    },
    {
      "clip_ratio/high_max": 0.00244561888393946,
      "clip_ratio/high_mean": 0.0010596034298941959,
      "clip_ratio/low_mean": 0.0009064220203072182,
      "clip_ratio/low_min": 6.806187957408838e-05,
      "clip_ratio/region_mean": 0.0019660254183690995,
      "epoch": 11.177259475218658,
      "grad_norm": 0.11755045503377914,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 723
    },
    {
      "clip_ratio/high_max": 0.002438738629280124,
      "clip_ratio/high_mean": 0.0010361317436036188,
      "clip_ratio/low_mean": 0.000995213542410056,
      "clip_ratio/low_min": 9.091229549085256e-05,
      "clip_ratio/region_mean": 0.0020313452841946855,
      "epoch": 11.186588921282798,
      "grad_norm": 0.12721142172813416,
      "learning_rate": 1e-06,
      "loss": 0.0133,
      "step": 724
    },
    {
      "clip_ratio/high_max": 0.002381140257057268,
      "clip_ratio/high_mean": 0.0009218243103532586,
      "clip_ratio/low_mean": 0.0008442512171313865,
      "clip_ratio/low_min": 3.742515036719851e-05,
      "clip_ratio/region_mean": 0.0017660755038377829,
      "epoch": 11.19591836734694,
      "grad_norm": 0.12327289581298828,
      "learning_rate": 1e-06,
      "loss": 0.0034,
      "step": 725
    },
    {
      "clip_ratio/high_max": 0.002510574799089227,
      "clip_ratio/high_mean": 0.001081999600501149,
      "clip_ratio/low_mean": 0.000987239349342417,
      "clip_ratio/low_min": 8.373669697903097e-05,
      "clip_ratio/region_mean": 0.002069238958938513,
      "epoch": 11.205247813411079,
      "grad_norm": 0.13151709735393524,
      "learning_rate": 1e-06,
      "loss": 0.0173,
      "step": 726
    },
    {
      "clip_ratio/high_max": 0.0023962165432749316,
      "clip_ratio/high_mean": 0.001015934754832415,
      "clip_ratio/low_mean": 0.0011323489979986334,
      "clip_ratio/low_min": 3.128910975647159e-05,
      "clip_ratio/region_mean": 0.0021482837400981225,
      "epoch": 11.214577259475218,
      "grad_norm": 0.11238902062177658,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 727
    },
    {
      "clip_ratio/high_max": 0.0026177162799285725,
      "clip_ratio/high_mean": 0.0009803035973163787,
      "clip_ratio/low_mean": 0.0009010215417220024,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018813251663232222,
      "epoch": 11.223906705539358,
      "grad_norm": 0.11881382763385773,
      "learning_rate": 1e-06,
      "loss": -0.0198,
      "step": 728
    },
    {
      "clip_ratio/high_max": 0.0021843644572072662,
      "clip_ratio/high_mean": 0.0010725448009907268,
      "clip_ratio/low_mean": 0.0010054527301690541,
      "clip_ratio/low_min": 2.4708440832910128e-05,
      "clip_ratio/region_mean": 0.002077997567539569,
      "epoch": 11.2332361516035,
      "grad_norm": 0.17188583314418793,
      "learning_rate": 1e-06,
      "loss": 0.0185,
      "step": 729
    },
    {
      "clip_ratio/high_max": 0.002614750243083108,
      "clip_ratio/high_mean": 0.0012701251234830124,
      "clip_ratio/low_mean": 0.0010940887150354683,
      "clip_ratio/low_min": 6.12630810792325e-05,
      "clip_ratio/region_mean": 0.0023642137603019364,
      "epoch": 11.242565597667639,
      "grad_norm": 0.12362480908632278,
      "learning_rate": 1e-06,
      "loss": -0.0523,
      "step": 730
    },
    {
      "clip_ratio/high_max": 0.0028457361986511387,
      "clip_ratio/high_mean": 0.0011353273948770948,
      "clip_ratio/low_mean": 0.0010526663954806281,
      "clip_ratio/low_min": 0.00013636371477332432,
      "clip_ratio/region_mean": 0.002187993814004585,
      "epoch": 11.251895043731778,
      "grad_norm": 0.12405002117156982,
      "learning_rate": 1e-06,
      "loss": 0.0146,
      "step": 731
    },
    {
      "clip_ratio/high_max": 0.00239518877788214,
      "clip_ratio/high_mean": 0.0009262924450013088,
      "clip_ratio/low_mean": 0.0010145465892037464,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019408389998716302,
      "epoch": 11.261224489795918,
      "grad_norm": 0.12108957022428513,
      "learning_rate": 1e-06,
      "loss": 0.0316,
      "step": 732
    },
    {
      "clip_ratio/high_max": 0.002161001590138767,
      "clip_ratio/high_mean": 0.0009232370721292682,
      "clip_ratio/low_mean": 0.0011434762782300822,
      "clip_ratio/low_min": 9.173826310870936e-05,
      "clip_ratio/region_mean": 0.0020667133649112657,
      "epoch": 11.270553935860057,
      "grad_norm": 0.13761521875858307,
      "learning_rate": 1e-06,
      "loss": 0.0257,
      "step": 733
    },
    {
      "clip_ratio/high_max": 0.002515211104764603,
      "clip_ratio/high_mean": 0.0009791414495339268,
      "clip_ratio/low_mean": 0.0011099152488895925,
      "clip_ratio/low_min": 5.0089560318156146e-05,
      "clip_ratio/region_mean": 0.002089056739350781,
      "epoch": 11.279883381924199,
      "grad_norm": 0.1308261752128601,
      "learning_rate": 1e-06,
      "loss": 0.0514,
      "step": 734
    },
    {
      "clip_ratio/high_max": 0.002347918962186668,
      "clip_ratio/high_mean": 0.0009449677945667645,
      "clip_ratio/low_mean": 0.0009651322561694542,
      "clip_ratio/low_min": 7.460100096068345e-05,
      "clip_ratio/region_mean": 0.0019101000289083458,
      "epoch": 11.289212827988338,
      "grad_norm": 0.11759794503450394,
      "learning_rate": 1e-06,
      "loss": 0.0155,
      "step": 735
    },
    {
      "clip_ratio/high_max": 0.002399704262643354,
      "clip_ratio/high_mean": 0.0009696418019302655,
      "clip_ratio/low_mean": 0.0010621196779538877,
      "clip_ratio/low_min": 5.5973654525587335e-05,
      "clip_ratio/region_mean": 0.002031761468970217,
      "epoch": 11.298542274052478,
      "grad_norm": 0.12267669290304184,
      "learning_rate": 1e-06,
      "loss": 0.0107,
      "step": 736
    },
    {
      "clip_ratio/high_max": 0.0023604551824973896,
      "clip_ratio/high_mean": 0.0011393387085263385,
      "clip_ratio/low_mean": 0.0009235406751031405,
      "clip_ratio/low_min": 2.9429937967506703e-05,
      "clip_ratio/region_mean": 0.0020628794154617935,
      "epoch": 11.307871720116617,
      "grad_norm": 0.11160534620285034,
      "learning_rate": 1e-06,
      "loss": -0.0345,
      "step": 737
    },
    {
      "clip_ratio/high_max": 0.002713168109039543,
      "clip_ratio/high_mean": 0.0010418410347483587,
      "clip_ratio/low_mean": 0.0011020058441317815,
      "clip_ratio/low_min": 0.00011018855548172724,
      "clip_ratio/region_mean": 0.0021438468756969087,
      "epoch": 11.317201166180759,
      "grad_norm": 0.12543310225009918,
      "learning_rate": 1e-06,
      "loss": 0.0164,
      "step": 738
    },
    {
      "clip_ratio/high_max": 0.0028527663671411574,
      "clip_ratio/high_mean": 0.0011213679899810813,
      "clip_ratio/low_mean": 0.0009325786959379911,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020539467004709877,
      "epoch": 11.326530612244898,
      "grad_norm": 0.15083754062652588,
      "learning_rate": 1e-06,
      "loss": -0.0146,
      "step": 739
    },
    {
      "clip_ratio/high_max": 0.0030303685562103055,
      "clip_ratio/high_mean": 0.0011599207973631565,
      "clip_ratio/low_mean": 0.0013503695226972923,
      "clip_ratio/low_min": 6.762324028386502e-05,
      "clip_ratio/region_mean": 0.002510290330974385,
      "epoch": 11.335860058309038,
      "grad_norm": 0.24084800481796265,
      "learning_rate": 1e-06,
      "loss": -0.0045,
      "step": 740
    },
    {
      "clip_ratio/high_max": 0.0020095816762477625,
      "clip_ratio/high_mean": 0.0009474514572502812,
      "clip_ratio/low_mean": 0.0012158709614595864,
      "clip_ratio/low_min": 2.384585968684405e-05,
      "clip_ratio/region_mean": 0.0021633224096149206,
      "epoch": 11.345189504373177,
      "grad_norm": 0.1264019012451172,
      "learning_rate": 1e-06,
      "loss": 0.054,
      "step": 741
    },
    {
      "clip_ratio/high_max": 0.0025180551238008775,
      "clip_ratio/high_mean": 0.001131332595832646,
      "clip_ratio/low_mean": 0.0009596096006134758,
      "clip_ratio/low_min": 3.4610492548381444e-05,
      "clip_ratio/region_mean": 0.0020909421727992594,
      "epoch": 11.354518950437317,
      "grad_norm": 0.12203217297792435,
      "learning_rate": 1e-06,
      "loss": -0.0248,
      "step": 742
    },
    {
      "clip_ratio/high_max": 0.0030798617590335198,
      "clip_ratio/high_mean": 0.0013094159221509472,
      "clip_ratio/low_mean": 0.000991718527075136,
      "clip_ratio/low_min": 3.9871658373158425e-05,
      "clip_ratio/region_mean": 0.0023011344601400197,
      "epoch": 11.363848396501458,
      "grad_norm": 0.13002605736255646,
      "learning_rate": 1e-06,
      "loss": -0.018,
      "step": 743
    },
    {
      "clip_ratio/high_max": 0.0027146751308464445,
      "clip_ratio/high_mean": 0.0012330370445852168,
      "clip_ratio/low_mean": 0.0009880520810838789,
      "clip_ratio/low_min": 0.00010868485333048739,
      "clip_ratio/region_mean": 0.002221089096565265,
      "epoch": 11.373177842565598,
      "grad_norm": 0.13661731779575348,
      "learning_rate": 1e-06,
      "loss": -0.0582,
      "step": 744
    },
    {
      "clip_ratio/high_max": 0.002614112469018437,
      "clip_ratio/high_mean": 0.0012876811924797948,
      "clip_ratio/low_mean": 0.0010484185404493473,
      "clip_ratio/low_min": 8.594152677687816e-05,
      "clip_ratio/region_mean": 0.0023360997074632905,
      "epoch": 11.382507288629737,
      "grad_norm": 0.13901357352733612,
      "learning_rate": 1e-06,
      "loss": -0.0137,
      "step": 745
    },
    {
      "clip_ratio/high_max": 0.002895220648497343,
      "clip_ratio/high_mean": 0.0012374354955682065,
      "clip_ratio/low_mean": 0.0009863231771305436,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002223758645413909,
      "epoch": 11.391836734693877,
      "grad_norm": 0.12362924218177795,
      "learning_rate": 1e-06,
      "loss": -0.0252,
      "step": 746
    },
    {
      "clip_ratio/high_max": 0.0026101976764039136,
      "clip_ratio/high_mean": 0.001134576006734278,
      "clip_ratio/low_mean": 0.0011314359962852905,
      "clip_ratio/low_min": 5.206631522014504e-05,
      "clip_ratio/region_mean": 0.0022660120666841976,
      "epoch": 11.401166180758018,
      "grad_norm": 0.12327460199594498,
      "learning_rate": 1e-06,
      "loss": -0.0087,
      "step": 747
    },
    {
      "clip_ratio/high_max": 0.002580092885182239,
      "clip_ratio/high_mean": 0.001262037254491588,
      "clip_ratio/low_mean": 0.0010065585374832153,
      "clip_ratio/low_min": 0.00011536663078004494,
      "clip_ratio/region_mean": 0.00226859580288874,
      "epoch": 11.410495626822158,
      "grad_norm": 0.12078581750392914,
      "learning_rate": 1e-06,
      "loss": -0.0297,
      "step": 748
    },
    {
      "clip_ratio/high_max": 0.003057281966903247,
      "clip_ratio/high_mean": 0.0012085692305845441,
      "clip_ratio/low_mean": 0.0011419986767577939,
      "clip_ratio/low_min": 6.670238690276165e-05,
      "clip_ratio/region_mean": 0.002350567912799306,
      "epoch": 11.419825072886297,
      "grad_norm": 0.12613429129123688,
      "learning_rate": 1e-06,
      "loss": -0.0415,
      "step": 749
    },
    {
      "clip_ratio/high_max": 0.00254219192720484,
      "clip_ratio/high_mean": 0.0011747742610168643,
      "clip_ratio/low_mean": 0.001075635615052306,
      "clip_ratio/low_min": 1.7841850421973504e-05,
      "clip_ratio/region_mean": 0.0022504099033540115,
      "epoch": 11.429154518950437,
      "grad_norm": 0.12662017345428467,
      "learning_rate": 1e-06,
      "loss": 0.0049,
      "step": 750
    },
    {
      "clip_ratio/high_max": 0.002502478026144672,
      "clip_ratio/high_mean": 0.0010446332362334942,
      "clip_ratio/low_mean": 0.0012332304595474852,
      "clip_ratio/low_min": 8.38929736346472e-05,
      "clip_ratio/region_mean": 0.0022778636775910854,
      "epoch": 11.438483965014576,
      "grad_norm": 0.13240495324134827,
      "learning_rate": 1e-06,
      "loss": -0.0294,
      "step": 751
    },
    {
      "clip_ratio/high_max": 0.0027811097825178877,
      "clip_ratio/high_mean": 0.001113544170948444,
      "clip_ratio/low_mean": 0.0011844903619930847,
      "clip_ratio/low_min": 6.30276099400362e-05,
      "clip_ratio/region_mean": 0.0022980345675023273,
      "epoch": 11.447813411078718,
      "grad_norm": 0.1276206225156784,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 752
    },
    {
      "clip_ratio/high_max": 0.0030890490452293307,
      "clip_ratio/high_mean": 0.0011506654882396106,
      "clip_ratio/low_mean": 0.0012208038824610412,
      "clip_ratio/low_min": 4.82346149510704e-05,
      "clip_ratio/region_mean": 0.0023714693379588425,
      "epoch": 11.457142857142857,
      "grad_norm": 0.13413143157958984,
      "learning_rate": 1e-06,
      "loss": -0.0112,
      "step": 753
    },
    {
      "clip_ratio/high_max": 0.0031963121437001973,
      "clip_ratio/high_mean": 0.0013220494729466736,
      "clip_ratio/low_mean": 0.0012697162783297244,
      "clip_ratio/low_min": 3.198566992068663e-05,
      "clip_ratio/region_mean": 0.002591765791294165,
      "epoch": 11.466472303206997,
      "grad_norm": 0.13057121634483337,
      "learning_rate": 1e-06,
      "loss": -0.0176,
      "step": 754
    },
    {
      "clip_ratio/high_max": 0.00294013433449436,
      "clip_ratio/high_mean": 0.0012466764637792949,
      "clip_ratio/low_mean": 0.0014524683028867003,
      "clip_ratio/low_min": 0.00019288149997009896,
      "clip_ratio/region_mean": 0.002699144803045783,
      "epoch": 11.475801749271136,
      "grad_norm": 0.11764442175626755,
      "learning_rate": 1e-06,
      "loss": -0.0065,
      "step": 755
    },
    {
      "clip_ratio/high_max": 0.002724162877711933,
      "clip_ratio/high_mean": 0.0011036147552658804,
      "clip_ratio/low_mean": 0.0011416962333896663,
      "clip_ratio/low_min": 6.9522475314443e-05,
      "clip_ratio/region_mean": 0.00224531094863778,
      "epoch": 11.485131195335278,
      "grad_norm": 0.12831537425518036,
      "learning_rate": 1e-06,
      "loss": 0.0018,
      "step": 756
    },
    {
      "clip_ratio/high_max": 0.002910467497713398,
      "clip_ratio/high_mean": 0.0012290866106923204,
      "clip_ratio/low_mean": 0.0012091222270100843,
      "clip_ratio/low_min": 5.258953842712799e-05,
      "clip_ratio/region_mean": 0.002438208866806235,
      "epoch": 11.494460641399417,
      "grad_norm": 0.1343003511428833,
      "learning_rate": 1e-06,
      "loss": -0.0358,
      "step": 757
    },
    {
      "clip_ratio/high_max": 0.0025345139220007695,
      "clip_ratio/high_mean": 0.0011368698706064606,
      "clip_ratio/low_mean": 0.0011714310985553311,
      "clip_ratio/low_min": 4.6779983676970005e-05,
      "clip_ratio/region_mean": 0.002308300980075728,
      "epoch": 11.503790087463557,
      "grad_norm": 0.12139391899108887,
      "learning_rate": 1e-06,
      "loss": 0.0057,
      "step": 758
    },
    {
      "clip_ratio/high_max": 0.002611616342619527,
      "clip_ratio/high_mean": 0.0011538451690285,
      "clip_ratio/low_mean": 0.0011436322620284045,
      "clip_ratio/low_min": 0.00012903800961794332,
      "clip_ratio/region_mean": 0.002297477491083555,
      "epoch": 11.513119533527696,
      "grad_norm": 0.1262354552745819,
      "learning_rate": 1e-06,
      "loss": -0.0227,
      "step": 759
    },
    {
      "clip_ratio/high_max": 0.002452380213071592,
      "clip_ratio/high_mean": 0.0010706639877753332,
      "clip_ratio/low_mean": 0.001350758289845544,
      "clip_ratio/low_min": 0.00016404510461143218,
      "clip_ratio/region_mean": 0.0024214223230956122,
      "epoch": 11.522448979591836,
      "grad_norm": 0.12216098606586456,
      "learning_rate": 1e-06,
      "loss": 0.0463,
      "step": 760
    },
    {
      "clip_ratio/high_max": 0.0032536294929741416,
      "clip_ratio/high_mean": 0.0014053815648367163,
      "clip_ratio/low_mean": 0.0011871640526806004,
      "clip_ratio/low_min": 8.931851607485441e-05,
      "clip_ratio/region_mean": 0.002592545555671677,
      "epoch": 11.531778425655977,
      "grad_norm": 0.11559303849935532,
      "learning_rate": 1e-06,
      "loss": -0.0374,
      "step": 761
    },
    {
      "clip_ratio/high_max": 0.0032493172475369647,
      "clip_ratio/high_mean": 0.0013286821958899964,
      "clip_ratio/low_mean": 0.0008384985521843191,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021671807262464426,
      "epoch": 11.541107871720117,
      "grad_norm": 0.11857152730226517,
      "learning_rate": 1e-06,
      "loss": -0.0626,
      "step": 762
    },
    {
      "clip_ratio/high_max": 0.0028308008913882077,
      "clip_ratio/high_mean": 0.0011588010929699522,
      "clip_ratio/low_mean": 0.0011810539381258423,
      "clip_ratio/low_min": 7.284887396963313e-05,
      "clip_ratio/region_mean": 0.002339855011086911,
      "epoch": 11.550437317784256,
      "grad_norm": 0.11263173073530197,
      "learning_rate": 1e-06,
      "loss": -0.0117,
      "step": 763
    },
    {
      "clip_ratio/high_max": 0.00292706972322776,
      "clip_ratio/high_mean": 0.0011468915627119713,
      "clip_ratio/low_mean": 0.0013457503573590657,
      "clip_ratio/low_min": 1.806358341127634e-05,
      "clip_ratio/region_mean": 0.002492641913704574,
      "epoch": 11.559766763848396,
      "grad_norm": 0.1806187480688095,
      "learning_rate": 1e-06,
      "loss": -0.0201,
      "step": 764
    },
    {
      "clip_ratio/high_max": 0.002420545235509053,
      "clip_ratio/high_mean": 0.0010321046538592782,
      "clip_ratio/low_mean": 0.0014250486055971123,
      "clip_ratio/low_min": 0.00011616915980994236,
      "clip_ratio/region_mean": 0.002457153248542454,
      "epoch": 11.569096209912537,
      "grad_norm": 0.11586566269397736,
      "learning_rate": 1e-06,
      "loss": -0.0038,
      "step": 765
    },
    {
      "clip_ratio/high_max": 0.002730551153945271,
      "clip_ratio/high_mean": 0.0011794552083301824,
      "clip_ratio/low_mean": 0.0012456532458600122,
      "clip_ratio/low_min": 8.259169226221275e-05,
      "clip_ratio/region_mean": 0.0024251084178104065,
      "epoch": 11.578425655976677,
      "grad_norm": 0.1318320780992508,
      "learning_rate": 1e-06,
      "loss": -0.0186,
      "step": 766
    },
    {
      "clip_ratio/high_max": 0.003323022261611186,
      "clip_ratio/high_mean": 0.0013961967397335684,
      "clip_ratio/low_mean": 0.0010770241315185558,
      "clip_ratio/low_min": 0.00016406031681981403,
      "clip_ratio/region_mean": 0.002473220869433135,
      "epoch": 11.587755102040816,
      "grad_norm": 0.12317539751529694,
      "learning_rate": 1e-06,
      "loss": -0.0453,
      "step": 767
    },
    {
      "clip_ratio/high_max": 0.002865194182959385,
      "clip_ratio/high_mean": 0.001273908117582323,
      "clip_ratio/low_mean": 0.0011619734432315454,
      "clip_ratio/low_min": 4.2286872485419735e-05,
      "clip_ratio/region_mean": 0.002435881513520144,
      "epoch": 11.597084548104956,
      "grad_norm": 0.13387911021709442,
      "learning_rate": 1e-06,
      "loss": -0.0092,
      "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": 0.042881556919642905,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4094.0,
      "completions/mean_length": 711.5376586914062,
      "completions/mean_terminated_length": 559.904296875,
      "completions/min_length": 3.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 12.00932944606414,
      "grad_norm": 0.12344630062580109,
      "learning_rate": 1e-06,
      "loss": 0.0313,
      "num_tokens": 481489698.0,
      "reward": 0.6261858344078064,
      "reward_std": 0.1730087548494339,
      "rewards/simpleverify_reward/mean": 0.6261858344078064,
      "rewards/simpleverify_reward/std": 0.48381945490837097,
      "step": 769
    },
    {
      "clip_ratio/high_max": 0.001982039520953549,
      "clip_ratio/high_mean": 0.000854600011734874,
      "clip_ratio/low_mean": 0.0005202167030802229,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0013748167039011605,
      "epoch": 12.018658892128279,
      "grad_norm": 0.12510351836681366,
      "learning_rate": 1e-06,
      "loss": -0.0493,
      "step": 770
    },
    {
      "clip_ratio/high_max": 0.0018740668492682744,
      "clip_ratio/high_mean": 0.0007808510454196949,
      "clip_ratio/low_mean": 0.0004294324535294436,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0012102835316909477,
      "epoch": 12.02798833819242,
      "grad_norm": 0.1354844719171524,
      "learning_rate": 1e-06,
      "loss": -0.0253,
      "step": 771
    },
    {
      "clip_ratio/high_max": 0.0019721556527656503,
      "clip_ratio/high_mean": 0.0007854262312321225,
      "clip_ratio/low_mean": 0.0006107273266025004,
      "clip_ratio/low_min": 2.8629018743231427e-05,
      "clip_ratio/region_mean": 0.0013961535551061388,
      "epoch": 12.03731778425656,
      "grad_norm": 0.12820011377334595,
      "learning_rate": 1e-06,
      "loss": 0.0082,
      "step": 772
    },
    {
      "clip_ratio/high_max": 0.001995891892875079,
      "clip_ratio/high_mean": 0.0008941908308770508,
      "clip_ratio/low_mean": 0.0005449218706417014,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014391127042472363,
      "epoch": 12.0466472303207,
      "grad_norm": 0.14113283157348633,
      "learning_rate": 1e-06,
      "loss": -0.0236,
      "step": 773
    },
    {
      "clip_ratio/high_max": 0.00212800425288151,
      "clip_ratio/high_mean": 0.0008312418367495411,
      "clip_ratio/low_mean": 0.0005162234765521134,
      "clip_ratio/low_min": 1.27577059174655e-05,
      "clip_ratio/region_mean": 0.0013474652987497393,
      "epoch": 12.055976676384839,
      "grad_norm": 0.12578293681144714,
      "learning_rate": 1e-06,
      "loss": -0.0031,
      "step": 774
    },
    {
      "clip_ratio/high_max": 0.00207770339329727,
      "clip_ratio/high_mean": 0.0008650214276713086,
      "clip_ratio/low_mean": 0.0008076959747995716,
      "clip_ratio/low_min": 1.3898154065827839e-05,
      "clip_ratio/region_mean": 0.0016727174224797636,
      "epoch": 12.06530612244898,
      "grad_norm": 0.13031761348247528,
      "learning_rate": 1e-06,
      "loss": 0.0144,
      "step": 775
    },
    {
      "clip_ratio/high_max": 0.002384588609857019,
      "clip_ratio/high_mean": 0.0010925433452939615,
      "clip_ratio/low_mean": 0.0008189423151634401,
      "clip_ratio/low_min": 4.086302578798495e-05,
      "clip_ratio/region_mean": 0.0019114856913802214,
      "epoch": 12.07463556851312,
      "grad_norm": 0.1263548582792282,
      "learning_rate": 1e-06,
      "loss": -0.0248,
      "step": 776
    },
    {
      "clip_ratio/high_max": 0.002328166832739953,
      "clip_ratio/high_mean": 0.0009084943048947025,
      "clip_ratio/low_mean": 0.0007481899956474081,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001656684293266153,
      "epoch": 12.08396501457726,
      "grad_norm": 0.12414831668138504,
      "learning_rate": 1e-06,
      "loss": -0.0062,
      "step": 777
    },
    {
      "clip_ratio/high_max": 0.00216972162888851,
      "clip_ratio/high_mean": 0.0009422271541552618,
      "clip_ratio/low_mean": 0.0008486050264764344,
      "clip_ratio/low_min": 4.4227801481611095e-05,
      "clip_ratio/region_mean": 0.0017908321678987704,
      "epoch": 12.093294460641399,
      "grad_norm": 0.1291738897562027,
      "learning_rate": 1e-06,
      "loss": -0.0171,
      "step": 778
    },
    {
      "clip_ratio/high_max": 0.002544457332987804,
      "clip_ratio/high_mean": 0.0011278465208306443,
      "clip_ratio/low_mean": 0.0008483778992740554,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001976224419195205,
      "epoch": 12.102623906705539,
      "grad_norm": 0.12321454286575317,
      "learning_rate": 1e-06,
      "loss": -0.0424,
      "step": 779
    },
    {
      "clip_ratio/high_max": 0.0022426467839977704,
      "clip_ratio/high_mean": 0.0009335520517197438,
      "clip_ratio/low_mean": 0.0007896645665823598,
      "clip_ratio/low_min": 4.5110986320651136e-05,
      "clip_ratio/region_mean": 0.0017232166501344182,
      "epoch": 12.11195335276968,
      "grad_norm": 0.11238522082567215,
      "learning_rate": 1e-06,
      "loss": -0.0324,
      "step": 780
    },
    {
      "clip_ratio/high_max": 0.0020828996202908456,
      "clip_ratio/high_mean": 0.0009151240883511491,
      "clip_ratio/low_mean": 0.0008842037786962464,
      "clip_ratio/low_min": 2.4680372916918714e-05,
      "clip_ratio/region_mean": 0.0017993278815993108,
      "epoch": 12.12128279883382,
      "grad_norm": 0.13164466619491577,
      "learning_rate": 1e-06,
      "loss": -0.0101,
      "step": 781
    },
    {
      "clip_ratio/high_max": 0.0017927900298673194,
      "clip_ratio/high_mean": 0.0008759616248426028,
      "clip_ratio/low_mean": 0.0010220490330539178,
      "clip_ratio/low_min": 5.633168257190846e-05,
      "clip_ratio/region_mean": 0.001898010632430669,
      "epoch": 12.130612244897959,
      "grad_norm": 0.13335801661014557,
      "learning_rate": 1e-06,
      "loss": -0.0222,
      "step": 782
    },
    {
      "clip_ratio/high_max": 0.0024858457618393004,
      "clip_ratio/high_mean": 0.0009664298386269365,
      "clip_ratio/low_mean": 0.0010804744051711168,
      "clip_ratio/low_min": 6.860592839075252e-05,
      "clip_ratio/region_mean": 0.002046904221060686,
      "epoch": 12.139941690962099,
      "grad_norm": 0.12098106741905212,
      "learning_rate": 1e-06,
      "loss": -0.0088,
      "step": 783
    },
    {
      "clip_ratio/high_max": 0.002409208995231893,
      "clip_ratio/high_mean": 0.0010208608000539243,
      "clip_ratio/low_mean": 0.0008848585193845793,
      "clip_ratio/low_min": 0.00013395625683187973,
      "clip_ratio/region_mean": 0.001905719327623956,
      "epoch": 12.14927113702624,
      "grad_norm": 0.12097387760877609,
      "learning_rate": 1e-06,
      "loss": -0.0151,
      "step": 784
    },
    {
      "clip_ratio/high_max": 0.0022664294665446505,
      "clip_ratio/high_mean": 0.0010413722302473616,
      "clip_ratio/low_mean": 0.0008924802932597231,
      "clip_ratio/low_min": 4.85528735225671e-05,
      "clip_ratio/region_mean": 0.0019338524944032542,
      "epoch": 12.15860058309038,
      "grad_norm": 0.11832785606384277,
      "learning_rate": 1e-06,
      "loss": -0.0347,
      "step": 785
    },
    {
      "clip_ratio/high_max": 0.0019698569376487285,
      "clip_ratio/high_mean": 0.000833251640870003,
      "clip_ratio/low_mean": 0.0009684636133897584,
      "clip_ratio/low_min": 3.361736889928579e-05,
      "clip_ratio/region_mean": 0.0018017152360698674,
      "epoch": 12.167930029154519,
      "grad_norm": 0.11735925078392029,
      "learning_rate": 1e-06,
      "loss": 0.0627,
      "step": 786
    },
    {
      "clip_ratio/high_max": 0.0023884439797257073,
      "clip_ratio/high_mean": 0.0010025802948803175,
      "clip_ratio/low_mean": 0.0010688419788493775,
      "clip_ratio/low_min": 8.836338929540943e-05,
      "clip_ratio/region_mean": 0.0020714222409878857,
      "epoch": 12.177259475218658,
      "grad_norm": 0.12652647495269775,
      "learning_rate": 1e-06,
      "loss": 0.0046,
      "step": 787
    },
    {
      "clip_ratio/high_max": 0.0021524110779864714,
      "clip_ratio/high_mean": 0.0008977284724096535,
      "clip_ratio/low_mean": 0.0009293914608861087,
      "clip_ratio/low_min": 3.6704934245790355e-05,
      "clip_ratio/region_mean": 0.0018271198860020377,
      "epoch": 12.186588921282798,
      "grad_norm": 0.1132396012544632,
      "learning_rate": 1e-06,
      "loss": 0.0223,
      "step": 788
    },
    {
      "clip_ratio/high_max": 0.0023530725593445823,
      "clip_ratio/high_mean": 0.0009946757309080567,
      "clip_ratio/low_mean": 0.0010905117524089292,
      "clip_ratio/low_min": 2.194672833866207e-05,
      "clip_ratio/region_mean": 0.002085187479679007,
      "epoch": 12.19591836734694,
      "grad_norm": 0.12985463440418243,
      "learning_rate": 1e-06,
      "loss": -0.0084,
      "step": 789
    },
    {
      "clip_ratio/high_max": 0.00261299560952466,
      "clip_ratio/high_mean": 0.0011101696763944346,
      "clip_ratio/low_mean": 0.0008699340041857795,
      "clip_ratio/low_min": 3.603508594096638e-05,
      "clip_ratio/region_mean": 0.0019801036396529526,
      "epoch": 12.205247813411079,
      "grad_norm": 0.1184607520699501,
      "learning_rate": 1e-06,
      "loss": -0.0285,
      "step": 790
    },
    {
      "clip_ratio/high_max": 0.002717594805289991,
      "clip_ratio/high_mean": 0.0011332717731420416,
      "clip_ratio/low_mean": 0.0010408923462819075,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021741640885011293,
      "epoch": 12.214577259475218,
      "grad_norm": 0.13032302260398865,
      "learning_rate": 1e-06,
      "loss": 0.0097,
      "step": 791
    },
    {
      "clip_ratio/high_max": 0.002858444830053486,
      "clip_ratio/high_mean": 0.0011575926982914098,
      "clip_ratio/low_mean": 0.0009146167885774048,
      "clip_ratio/low_min": 9.365544701722683e-05,
      "clip_ratio/region_mean": 0.002072209506877698,
      "epoch": 12.223906705539358,
      "grad_norm": 0.12457794696092606,
      "learning_rate": 1e-06,
      "loss": -0.0125,
      "step": 792
    },
    {
      "clip_ratio/high_max": 0.002127605323039461,
      "clip_ratio/high_mean": 0.0010039906974270707,
      "clip_ratio/low_mean": 0.0009453637940168846,
      "clip_ratio/low_min": 2.7352298275218345e-05,
      "clip_ratio/region_mean": 0.001949354467797093,
      "epoch": 12.2332361516035,
      "grad_norm": 0.11974716931581497,
      "learning_rate": 1e-06,
      "loss": -0.0246,
      "step": 793
    },
    {
      "clip_ratio/high_max": 0.0025850981910480186,
      "clip_ratio/high_mean": 0.0009408178066223627,
      "clip_ratio/low_mean": 0.0011635788177954964,
      "clip_ratio/low_min": 8.449673623545095e-05,
      "clip_ratio/region_mean": 0.0021043966480647214,
      "epoch": 12.242565597667639,
      "grad_norm": 0.1367463916540146,
      "learning_rate": 1e-06,
      "loss": 0.0294,
      "step": 794
    },
    {
      "clip_ratio/high_max": 0.0022464876601588912,
      "clip_ratio/high_mean": 0.0009718564178911038,
      "clip_ratio/low_mean": 0.0010182424794038525,
      "clip_ratio/low_min": 7.940767136460636e-05,
      "clip_ratio/region_mean": 0.0019900988991139457,
      "epoch": 12.251895043731778,
      "grad_norm": 0.1325341910123825,
      "learning_rate": 1e-06,
      "loss": -0.0359,
      "step": 795
    },
    {
      "clip_ratio/high_max": 0.002219196583610028,
      "clip_ratio/high_mean": 0.000921394939723541,
      "clip_ratio/low_mean": 0.001008641684165923,
      "clip_ratio/low_min": 3.4312379284529015e-05,
      "clip_ratio/region_mean": 0.0019300366402603686,
      "epoch": 12.261224489795918,
      "grad_norm": 0.11588510870933533,
      "learning_rate": 1e-06,
      "loss": 0.0114,
      "step": 796
    },
    {
      "clip_ratio/high_max": 0.0027079485953436233,
      "clip_ratio/high_mean": 0.0010720413556555286,
      "clip_ratio/low_mean": 0.000967370633588871,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002039411996520357,
      "epoch": 12.270553935860057,
      "grad_norm": 0.13519005477428436,
      "learning_rate": 1e-06,
      "loss": 0.0123,
      "step": 797
    },
    {
      "clip_ratio/high_max": 0.0024604934587841853,
      "clip_ratio/high_mean": 0.0010963392996927723,
      "clip_ratio/low_mean": 0.0008216893675125903,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019180286617483944,
      "epoch": 12.279883381924199,
      "grad_norm": 0.11680714040994644,
      "learning_rate": 1e-06,
      "loss": -0.0728,
      "step": 798
    },
    {
      "clip_ratio/high_max": 0.0023618988307134714,
      "clip_ratio/high_mean": 0.0009946611808118178,
      "clip_ratio/low_mean": 0.0012194803930469789,
      "clip_ratio/low_min": 8.814891043584794e-05,
      "clip_ratio/region_mean": 0.002214141597505659,
      "epoch": 12.289212827988338,
      "grad_norm": 0.13376952707767487,
      "learning_rate": 1e-06,
      "loss": 0.0463,
      "step": 799
    },
    {
      "clip_ratio/high_max": 0.0025391012877662433,
      "clip_ratio/high_mean": 0.001017661686091742,
      "clip_ratio/low_mean": 0.0010309930039511528,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020486547145992517,
      "epoch": 12.298542274052478,
      "grad_norm": 0.12554509937763214,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 800
    },
    {
      "clip_ratio/high_max": 0.002699669246794656,
      "clip_ratio/high_mean": 0.0011564240412553772,
      "clip_ratio/low_mean": 0.0009551375896990066,
      "clip_ratio/low_min": 4.852309211855754e-05,
      "clip_ratio/region_mean": 0.0021115616254974157,
      "epoch": 12.307871720116617,
      "grad_norm": 0.10621404647827148,
      "learning_rate": 1e-06,
      "loss": -0.0342,
      "step": 801
    },
    {
      "clip_ratio/high_max": 0.002637111676449422,
      "clip_ratio/high_mean": 0.0010181827656197129,
      "clip_ratio/low_mean": 0.0010148733017558698,
      "clip_ratio/low_min": 0.00013490813853422878,
      "clip_ratio/region_mean": 0.0020330560801085085,
      "epoch": 12.317201166180759,
      "grad_norm": 0.11737849563360214,
      "learning_rate": 1e-06,
      "loss": 0.0047,
      "step": 802
    },
    {
      "clip_ratio/high_max": 0.0028218860461493023,
      "clip_ratio/high_mean": 0.001119993376050843,
      "clip_ratio/low_mean": 0.000954437826294452,
      "clip_ratio/low_min": 3.095094598393189e-05,
      "clip_ratio/region_mean": 0.0020744312641909346,
      "epoch": 12.326530612244898,
      "grad_norm": 0.132485494017601,
      "learning_rate": 1e-06,
      "loss": -0.0059,
      "step": 803
    },
    {
      "clip_ratio/high_max": 0.002450728105031885,
      "clip_ratio/high_mean": 0.0009607499578123679,
      "clip_ratio/low_mean": 0.0008987103092295001,
      "clip_ratio/low_min": 1.2525049896794371e-05,
      "clip_ratio/region_mean": 0.0018594602515804581,
      "epoch": 12.335860058309038,
      "grad_norm": 0.11985781043767929,
      "learning_rate": 1e-06,
      "loss": -0.0088,
      "step": 804
    },
    {
      "clip_ratio/high_max": 0.0026656500267563388,
      "clip_ratio/high_mean": 0.001185635082947556,
      "clip_ratio/low_mean": 0.0009007929838844575,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020864279867964797,
      "epoch": 12.345189504373177,
      "grad_norm": 0.11155782639980316,
      "learning_rate": 1e-06,
      "loss": -0.0509,
      "step": 805
    },
    {
      "clip_ratio/high_max": 0.00200622856937116,
      "clip_ratio/high_mean": 0.000978238083916949,
      "clip_ratio/low_mean": 0.0009624249269108986,
      "clip_ratio/low_min": 8.195697682822356e-05,
      "clip_ratio/region_mean": 0.001940663052664604,
      "epoch": 12.354518950437317,
      "grad_norm": 0.13074827194213867,
      "learning_rate": 1e-06,
      "loss": 0.027,
      "step": 806
    },
    {
      "clip_ratio/high_max": 0.0024923087330535054,
      "clip_ratio/high_mean": 0.0010706364992074668,
      "clip_ratio/low_mean": 0.0009327660391136305,
      "clip_ratio/low_min": 1.4737090168637224e-05,
      "clip_ratio/region_mean": 0.002003402536502108,
      "epoch": 12.363848396501458,
      "grad_norm": 0.1202230229973793,
      "learning_rate": 1e-06,
      "loss": -0.0184,
      "step": 807
    },
    {
      "clip_ratio/high_max": 0.0026743922499008477,
      "clip_ratio/high_mean": 0.0011361265987943625,
      "clip_ratio/low_mean": 0.0008625164191471413,
      "clip_ratio/low_min": 1.5992834960343316e-05,
      "clip_ratio/region_mean": 0.001998643001570599,
      "epoch": 12.373177842565598,
      "grad_norm": 0.12350955605506897,
      "learning_rate": 1e-06,
      "loss": -0.0033,
      "step": 808
    },
    {
      "clip_ratio/high_max": 0.0022155088154249825,
      "clip_ratio/high_mean": 0.000983483800155227,
      "clip_ratio/low_mean": 0.0010891106339840917,
      "clip_ratio/low_min": 0.00011132718009321252,
      "clip_ratio/region_mean": 0.0020725944777950644,
      "epoch": 12.382507288629737,
      "grad_norm": 0.12260499596595764,
      "learning_rate": 1e-06,
      "loss": 0.0162,
      "step": 809
    },
    {
      "clip_ratio/high_max": 0.00277569120953558,
      "clip_ratio/high_mean": 0.0012253020504431333,
      "clip_ratio/low_mean": 0.0010829724778886884,
      "clip_ratio/low_min": 6.794331875425996e-05,
      "clip_ratio/region_mean": 0.002308274539245758,
      "epoch": 12.391836734693877,
      "grad_norm": 0.12466733157634735,
      "learning_rate": 1e-06,
      "loss": -0.0023,
      "step": 810
    },
    {
      "clip_ratio/high_max": 0.0023973426868906245,
      "clip_ratio/high_mean": 0.0010406812180008274,
      "clip_ratio/low_mean": 0.0011516396698425524,
      "clip_ratio/low_min": 0.00010277959972881945,
      "clip_ratio/region_mean": 0.0021923209351371042,
      "epoch": 12.401166180758018,
      "grad_norm": 0.12773151695728302,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "step": 811
    },
    {
      "clip_ratio/high_max": 0.0024643362339702435,
      "clip_ratio/high_mean": 0.0009923261059157085,
      "clip_ratio/low_mean": 0.0009225446028722217,
      "clip_ratio/low_min": 1.9476472516544163e-05,
      "clip_ratio/region_mean": 0.001914870703330962,
      "epoch": 12.410495626822158,
      "grad_norm": 0.12679500877857208,
      "learning_rate": 1e-06,
      "loss": 0.0239,
      "step": 812
    },
    {
      "clip_ratio/high_max": 0.0028649861851590686,
      "clip_ratio/high_mean": 0.0012896785337943584,
      "clip_ratio/low_mean": 0.0008719073084648699,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021615857767756097,
      "epoch": 12.419825072886297,
      "grad_norm": 0.12192143499851227,
      "learning_rate": 1e-06,
      "loss": -0.0683,
      "step": 813
    },
    {
      "clip_ratio/high_max": 0.0022357447451213375,
      "clip_ratio/high_mean": 0.001055348438967485,
      "clip_ratio/low_mean": 0.0010795788039104082,
      "clip_ratio/low_min": 3.757138620130718e-05,
      "clip_ratio/region_mean": 0.002134927264705766,
      "epoch": 12.429154518950437,
      "grad_norm": 0.11535592377185822,
      "learning_rate": 1e-06,
      "loss": -0.0201,
      "step": 814
    },
    {
      "clip_ratio/high_max": 0.00261023678467609,
      "clip_ratio/high_mean": 0.0010642307079251623,
      "clip_ratio/low_mean": 0.0012588719800987747,
      "clip_ratio/low_min": 0.0001318553877354134,
      "clip_ratio/region_mean": 0.0023231026498251595,
      "epoch": 12.438483965014576,
      "grad_norm": 0.1488509625196457,
      "learning_rate": 1e-06,
      "loss": 0.0388,
      "step": 815
    },
    {
      "clip_ratio/high_max": 0.002711935027036816,
      "clip_ratio/high_mean": 0.0011548065886017866,
      "clip_ratio/low_mean": 0.0011424495532992296,
      "clip_ratio/low_min": 0.00012122636508138385,
      "clip_ratio/region_mean": 0.0022972561346250586,
      "epoch": 12.447813411078718,
      "grad_norm": 0.11600733548402786,
      "learning_rate": 1e-06,
      "loss": -0.0042,
      "step": 816
    },
    {
      "clip_ratio/high_max": 0.002823381742928177,
      "clip_ratio/high_mean": 0.001177904719952494,
      "clip_ratio/low_mean": 0.0010691349270928185,
      "clip_ratio/low_min": 7.76533206590102e-05,
      "clip_ratio/region_mean": 0.0022470396434073336,
      "epoch": 12.457142857142857,
      "grad_norm": 0.126500204205513,
      "learning_rate": 1e-06,
      "loss": -0.0411,
      "step": 817
    },
    {
      "clip_ratio/high_max": 0.002562906942330301,
      "clip_ratio/high_mean": 0.0011726165867003147,
      "clip_ratio/low_mean": 0.0009087096295843367,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020813262672163546,
      "epoch": 12.466472303206997,
      "grad_norm": 0.11181089282035828,
      "learning_rate": 1e-06,
      "loss": -0.0637,
      "step": 818
    },
    {
      "clip_ratio/high_max": 0.002906414258177392,
      "clip_ratio/high_mean": 0.0010576440254226327,
      "clip_ratio/low_mean": 0.0009552777355565922,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002012921766436193,
      "epoch": 12.475801749271136,
      "grad_norm": 0.11564455181360245,
      "learning_rate": 1e-06,
      "loss": 0.006,
      "step": 819
    },
    {
      "clip_ratio/high_max": 0.002191215658967849,
      "clip_ratio/high_mean": 0.0010204338850599015,
      "clip_ratio/low_mean": 0.0012578281384776346,
      "clip_ratio/low_min": 7.845249638194218e-05,
      "clip_ratio/region_mean": 0.0022782620217185467,
      "epoch": 12.485131195335278,
      "grad_norm": 0.1232963427901268,
      "learning_rate": 1e-06,
      "loss": 0.0096,
      "step": 820
    },
    {
      "clip_ratio/high_max": 0.0027991028400720097,
      "clip_ratio/high_mean": 0.001012583195915795,
      "clip_ratio/low_mean": 0.0009260970527975587,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019386802523513325,
      "epoch": 12.494460641399417,
      "grad_norm": 0.1163288801908493,
      "learning_rate": 1e-06,
      "loss": -0.0082,
      "step": 821
    },
    {
      "clip_ratio/high_max": 0.0027422422790550627,
      "clip_ratio/high_mean": 0.001195595152239548,
      "clip_ratio/low_mean": 0.0009774352020031074,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002173030356061645,
      "epoch": 12.503790087463557,
      "grad_norm": 0.12730132043361664,
      "learning_rate": 1e-06,
      "loss": -0.0323,
      "step": 822
    },
    {
      "clip_ratio/high_max": 0.0025074906006921083,
      "clip_ratio/high_mean": 0.0011130772436445113,
      "clip_ratio/low_mean": 0.0010403667711216258,
      "clip_ratio/low_min": 0.0001829902485042112,
      "clip_ratio/region_mean": 0.0021534439001698047,
      "epoch": 12.513119533527696,
      "grad_norm": 0.1222425103187561,
      "learning_rate": 1e-06,
      "loss": -0.0127,
      "step": 823
    },
    {
      "clip_ratio/high_max": 0.0022122438240330666,
      "clip_ratio/high_mean": 0.0009195006714435294,
      "clip_ratio/low_mean": 0.0010709663856687257,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001990467091673054,
      "epoch": 12.522448979591836,
      "grad_norm": 0.12297448515892029,
      "learning_rate": 1e-06,
      "loss": 0.025,
      "step": 824
    },
    {
      "clip_ratio/high_max": 0.00236048595252214,
      "clip_ratio/high_mean": 0.0010837599802471232,
      "clip_ratio/low_mean": 0.0010024465736933053,
      "clip_ratio/low_min": 6.323777051875368e-05,
      "clip_ratio/region_mean": 0.0020862065357505344,
      "epoch": 12.531778425655977,
      "grad_norm": 0.1306031346321106,
      "learning_rate": 1e-06,
      "loss": -0.0088,
      "step": 825
    },
    {
      "clip_ratio/high_max": 0.002812127182551194,
      "clip_ratio/high_mean": 0.0011292535746179055,
      "clip_ratio/low_mean": 0.0010131762010132661,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021424297883640975,
      "epoch": 12.541107871720117,
      "grad_norm": 0.11995650827884674,
      "learning_rate": 1e-06,
      "loss": -0.0123,
      "step": 826
    },
    {
      "clip_ratio/high_max": 0.0029836383764632046,
      "clip_ratio/high_mean": 0.001176291967567522,
      "clip_ratio/low_mean": 0.0012050632140017115,
      "clip_ratio/low_min": 5.442845485958969e-05,
      "clip_ratio/region_mean": 0.0023813552033971064,
      "epoch": 12.550437317784256,
      "grad_norm": 0.12899866700172424,
      "learning_rate": 1e-06,
      "loss": -0.0237,
      "step": 827
    },
    {
      "clip_ratio/high_max": 0.0027070367214037105,
      "clip_ratio/high_mean": 0.0011146233809995465,
      "clip_ratio/low_mean": 0.0011652918146864977,
      "clip_ratio/low_min": 8.443019032711163e-05,
      "clip_ratio/region_mean": 0.0022799151556682773,
      "epoch": 12.559766763848396,
      "grad_norm": 0.1434909552335739,
      "learning_rate": 1e-06,
      "loss": -0.0385,
      "step": 828
    },
    {
      "clip_ratio/high_max": 0.0032792178681120276,
      "clip_ratio/high_mean": 0.001310497567828861,
      "clip_ratio/low_mean": 0.001049492742822622,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00235999035066925,
      "epoch": 12.569096209912537,
      "grad_norm": 0.1292286515235901,
      "learning_rate": 1e-06,
      "loss": -0.0159,
      "step": 829
    },
    {
      "clip_ratio/high_max": 0.0026349084728281014,
      "clip_ratio/high_mean": 0.0012092284778191242,
      "clip_ratio/low_mean": 0.001106671250454383,
      "clip_ratio/low_min": 4.5870066969655454e-05,
      "clip_ratio/region_mean": 0.0023158997209975496,
      "epoch": 12.578425655976677,
      "grad_norm": 0.1191285103559494,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 830
    },
    {
      "clip_ratio/high_max": 0.0027554127882467583,
      "clip_ratio/high_mean": 0.001035985453199828,
      "clip_ratio/low_mean": 0.0011049505956179928,
      "clip_ratio/low_min": 6.122842205513734e-05,
      "clip_ratio/region_mean": 0.002140936026989948,
      "epoch": 12.587755102040816,
      "grad_norm": 0.11864615976810455,
      "learning_rate": 1e-06,
      "loss": -0.0224,
      "step": 831
    },
    {
      "clip_ratio/high_max": 0.0022250700785662048,
      "clip_ratio/high_mean": 0.001027084690576885,
      "clip_ratio/low_mean": 0.0010864675286939018,
      "clip_ratio/low_min": 0.00010635959915816784,
      "clip_ratio/region_mean": 0.0021135521747055463,
      "epoch": 12.597084548104956,
      "grad_norm": 0.12421823292970657,
      "learning_rate": 1e-06,
      "loss": -0.0137,
      "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": 0.0462123325892857,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4076.0,
      "completions/mean_length": 715.6293334960938,
      "completions/mean_terminated_length": 551.8456420898438,
      "completions/min_length": 5.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 13.00932944606414,
      "grad_norm": 0.13414518535137177,
      "learning_rate": 1e-06,
      "loss": -0.0081,
      "num_tokens": 517420488.0,
      "reward": 0.632516086101532,
      "reward_std": 0.17390279471874237,
      "rewards/simpleverify_reward/mean": 0.6325160264968872,
      "rewards/simpleverify_reward/std": 0.4821240305900574,
      "step": 833
    },
    {
      "clip_ratio/high_max": 0.0019830406636174303,
      "clip_ratio/high_mean": 0.000829216519377951,
      "clip_ratio/low_mean": 0.000586779029617901,
      "clip_ratio/low_min": 4.197316684440011e-05,
      "clip_ratio/region_mean": 0.001415995524439495,
      "epoch": 13.018658892128279,
      "grad_norm": 0.12629295885562897,
      "learning_rate": 1e-06,
      "loss": 0.0185,
      "step": 834
    },
    {
      "clip_ratio/high_max": 0.002519905785447918,
      "clip_ratio/high_mean": 0.0009686950106697623,
      "clip_ratio/low_mean": 0.00043712686965591274,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014058219057915267,
      "epoch": 13.02798833819242,
      "grad_norm": 0.1319560408592224,
      "learning_rate": 1e-06,
      "loss": -0.0305,
      "step": 835
    },
    {
      "clip_ratio/high_max": 0.0020775691082235426,
      "clip_ratio/high_mean": 0.0008421860475209542,
      "clip_ratio/low_mean": 0.00046514722089341376,
      "clip_ratio/low_min": 1.4977234968682751e-05,
      "clip_ratio/region_mean": 0.001307333259319421,
      "epoch": 13.03731778425656,
      "grad_norm": 0.13775719702243805,
      "learning_rate": 1e-06,
      "loss": -0.0125,
      "step": 836
    },
    {
      "clip_ratio/high_max": 0.0022130011893750634,
      "clip_ratio/high_mean": 0.0009837952857196797,
      "clip_ratio/low_mean": 0.0005082932439108845,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001492088515078649,
      "epoch": 13.0466472303207,
      "grad_norm": 0.11993347108364105,
      "learning_rate": 1e-06,
      "loss": -0.0384,
      "step": 837
    },
    {
      "clip_ratio/high_max": 0.0018777957957354374,
      "clip_ratio/high_mean": 0.0007917759703559568,
      "clip_ratio/low_mean": 0.0006243171310416074,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014160930950311013,
      "epoch": 13.055976676384839,
      "grad_norm": 0.12756216526031494,
      "learning_rate": 1e-06,
      "loss": 0.0073,
      "step": 838
    },
    {
      "clip_ratio/high_max": 0.002551495796069503,
      "clip_ratio/high_mean": 0.001109146967792185,
      "clip_ratio/low_mean": 0.0005956410623184638,
      "clip_ratio/low_min": 1.6399895685026422e-05,
      "clip_ratio/region_mean": 0.0017047880028258078,
      "epoch": 13.06530612244898,
      "grad_norm": 0.12596020102500916,
      "learning_rate": 1e-06,
      "loss": -0.0747,
      "step": 839
    },
    {
      "clip_ratio/high_max": 0.0025459724383836146,
      "clip_ratio/high_mean": 0.000972008281678427,
      "clip_ratio/low_mean": 0.0006664229513262399,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016384312220907304,
      "epoch": 13.07463556851312,
      "grad_norm": 0.12022960931062698,
      "learning_rate": 1e-06,
      "loss": 0.0079,
      "step": 840
    },
    {
      "clip_ratio/high_max": 0.002262051966681611,
      "clip_ratio/high_mean": 0.001028937302180566,
      "clip_ratio/low_mean": 0.00063529398994433,
      "clip_ratio/low_min": 2.2340045688906685e-05,
      "clip_ratio/region_mean": 0.0016642313203192316,
      "epoch": 13.08396501457726,
      "grad_norm": 0.12536147236824036,
      "learning_rate": 1e-06,
      "loss": -0.0695,
      "step": 841
    },
    {
      "clip_ratio/high_max": 0.0019748207269003615,
      "clip_ratio/high_mean": 0.0008613280915596988,
      "clip_ratio/low_mean": 0.0006797303367420682,
      "clip_ratio/low_min": 1.2124150998715777e-05,
      "clip_ratio/region_mean": 0.0015410584746859968,
      "epoch": 13.093294460641399,
      "grad_norm": 0.13648957014083862,
      "learning_rate": 1e-06,
      "loss": -0.005,
      "step": 842
    },
    {
      "clip_ratio/high_max": 0.0021365920729294885,
      "clip_ratio/high_mean": 0.0008052554057940142,
      "clip_ratio/low_mean": 0.00081346208571631,
      "clip_ratio/low_min": 3.0599756428273395e-05,
      "clip_ratio/region_mean": 0.0016187174915103242,
      "epoch": 13.102623906705539,
      "grad_norm": 0.10979168862104416,
      "learning_rate": 1e-06,
      "loss": 0.0112,
      "step": 843
    },
    {
      "clip_ratio/high_max": 0.00238527319743298,
      "clip_ratio/high_mean": 0.000985304759524297,
      "clip_ratio/low_mean": 0.0008162400827131933,
      "clip_ratio/low_min": 5.832644546899246e-05,
      "clip_ratio/region_mean": 0.0018015448222286068,
      "epoch": 13.11195335276968,
      "grad_norm": 0.13775122165679932,
      "learning_rate": 1e-06,
      "loss": -0.016,
      "step": 844
    },
    {
      "clip_ratio/high_max": 0.0024285401159431785,
      "clip_ratio/high_mean": 0.0010776535491459072,
      "clip_ratio/low_mean": 0.0009538254817016423,
      "clip_ratio/low_min": 3.557199670467526e-05,
      "clip_ratio/region_mean": 0.0020314791036071256,
      "epoch": 13.12128279883382,
      "grad_norm": 0.1199856624007225,
      "learning_rate": 1e-06,
      "loss": -0.0403,
      "step": 845
    },
    {
      "clip_ratio/high_max": 0.0021228468031040393,
      "clip_ratio/high_mean": 0.0009413827538082842,
      "clip_ratio/low_mean": 0.0008255953034677077,
      "clip_ratio/low_min": 3.4370357752777636e-05,
      "clip_ratio/region_mean": 0.0017669780645519495,
      "epoch": 13.130612244897959,
      "grad_norm": 0.12401924282312393,
      "learning_rate": 1e-06,
      "loss": -0.0354,
      "step": 846
    },
    {
      "clip_ratio/high_max": 0.0023706748979748227,
      "clip_ratio/high_mean": 0.0009638159781388822,
      "clip_ratio/low_mean": 0.0009370964025947615,
      "clip_ratio/low_min": 1.938585592142772e-05,
      "clip_ratio/region_mean": 0.001900912364362739,
      "epoch": 13.139941690962099,
      "grad_norm": 0.11542689800262451,
      "learning_rate": 1e-06,
      "loss": -0.0238,
      "step": 847
    },
    {
      "clip_ratio/high_max": 0.0022999602297204547,
      "clip_ratio/high_mean": 0.0009680519342509797,
      "clip_ratio/low_mean": 0.000854471565617132,
      "clip_ratio/low_min": 2.4925224352045916e-05,
      "clip_ratio/region_mean": 0.001822523510782048,
      "epoch": 13.14927113702624,
      "grad_norm": 0.11744103580713272,
      "learning_rate": 1e-06,
      "loss": -0.0344,
      "step": 848
    },
    {
      "clip_ratio/high_max": 0.0021957652061246336,
      "clip_ratio/high_mean": 0.0010111743104062043,
      "clip_ratio/low_mean": 0.0008652507840452017,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001876425085356459,
      "epoch": 13.15860058309038,
      "grad_norm": 0.15531907975673676,
      "learning_rate": 1e-06,
      "loss": -0.0228,
      "step": 849
    },
    {
      "clip_ratio/high_max": 0.002185033998102881,
      "clip_ratio/high_mean": 0.0009567482593411114,
      "clip_ratio/low_mean": 0.0009446445146750193,
      "clip_ratio/low_min": 1.256533960258821e-05,
      "clip_ratio/region_mean": 0.001901392766740173,
      "epoch": 13.167930029154519,
      "grad_norm": 0.1306806057691574,
      "learning_rate": 1e-06,
      "loss": 0.0044,
      "step": 850
    },
    {
      "clip_ratio/high_max": 0.002366046766837826,
      "clip_ratio/high_mean": 0.0009903839863909525,
      "clip_ratio/low_mean": 0.0010551582599873655,
      "clip_ratio/low_min": 4.884605550614651e-05,
      "clip_ratio/region_mean": 0.002045542249106802,
      "epoch": 13.177259475218658,
      "grad_norm": 0.12941259145736694,
      "learning_rate": 1e-06,
      "loss": 0.0096,
      "step": 851
    },
    {
      "clip_ratio/high_max": 0.0019359269681444857,
      "clip_ratio/high_mean": 0.0008711113023309736,
      "clip_ratio/low_mean": 0.0010929880018011318,
      "clip_ratio/low_min": 9.626067276258254e-05,
      "clip_ratio/region_mean": 0.0019640991995402146,
      "epoch": 13.186588921282798,
      "grad_norm": 0.12935452163219452,
      "learning_rate": 1e-06,
      "loss": 0.0089,
      "step": 852
    },
    {
      "clip_ratio/high_max": 0.002536605370551115,
      "clip_ratio/high_mean": 0.0009768370000529103,
      "clip_ratio/low_mean": 0.0009711972415971104,
      "clip_ratio/low_min": 4.192036249150988e-05,
      "clip_ratio/region_mean": 0.0019480342816677876,
      "epoch": 13.19591836734694,
      "grad_norm": 0.12276334315538406,
      "learning_rate": 1e-06,
      "loss": -0.0215,
      "step": 853
    },
    {
      "clip_ratio/high_max": 0.002183154851081781,
      "clip_ratio/high_mean": 0.000942244058023789,
      "clip_ratio/low_mean": 0.0011278330639470369,
      "clip_ratio/low_min": 6.183659661473939e-05,
      "clip_ratio/region_mean": 0.0020700771565316245,
      "epoch": 13.205247813411079,
      "grad_norm": 0.12107101082801819,
      "learning_rate": 1e-06,
      "loss": 0.0097,
      "step": 854
    },
    {
      "clip_ratio/high_max": 0.002543337039242033,
      "clip_ratio/high_mean": 0.0009860966620180989,
      "clip_ratio/low_mean": 0.0009956289177353028,
      "clip_ratio/low_min": 2.192982537962962e-05,
      "clip_ratio/region_mean": 0.001981725567020476,
      "epoch": 13.214577259475218,
      "grad_norm": 0.12721332907676697,
      "learning_rate": 1e-06,
      "loss": -0.0085,
      "step": 855
    },
    {
      "clip_ratio/high_max": 0.0023810036436771043,
      "clip_ratio/high_mean": 0.0010373615623393562,
      "clip_ratio/low_mean": 0.00102116559719434,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002058527199551463,
      "epoch": 13.223906705539358,
      "grad_norm": 0.12266834080219269,
      "learning_rate": 1e-06,
      "loss": -0.0433,
      "step": 856
    },
    {
      "clip_ratio/high_max": 0.002492389270628337,
      "clip_ratio/high_mean": 0.0009353009791084332,
      "clip_ratio/low_mean": 0.0009484468100708909,
      "clip_ratio/low_min": 2.4457052859361283e-05,
      "clip_ratio/region_mean": 0.0018837477909983136,
      "epoch": 13.2332361516035,
      "grad_norm": 0.14596796035766602,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 857
    },
    {
      "clip_ratio/high_max": 0.002661496699147392,
      "clip_ratio/high_mean": 0.0012286690507608,
      "clip_ratio/low_mean": 0.0008065060428634752,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020351750499685295,
      "epoch": 13.242565597667639,
      "grad_norm": 0.12908431887626648,
      "learning_rate": 1e-06,
      "loss": -0.0752,
      "step": 858
    },
    {
      "clip_ratio/high_max": 0.0025324390226160176,
      "clip_ratio/high_mean": 0.0010505361897230614,
      "clip_ratio/low_mean": 0.001047112589731114,
      "clip_ratio/low_min": 7.459139942511683e-05,
      "clip_ratio/region_mean": 0.002097648808558006,
      "epoch": 13.251895043731778,
      "grad_norm": 0.12802553176879883,
      "learning_rate": 1e-06,
      "loss": -0.0043,
      "step": 859
    },
    {
      "clip_ratio/high_max": 0.0023204233293654397,
      "clip_ratio/high_mean": 0.001041752941091545,
      "clip_ratio/low_mean": 0.0009563090334268054,
      "clip_ratio/low_min": 5.52955661987653e-05,
      "clip_ratio/region_mean": 0.001998061918129679,
      "epoch": 13.261224489795918,
      "grad_norm": 0.13391587138175964,
      "learning_rate": 1e-06,
      "loss": -0.0129,
      "step": 860
    },
    {
      "clip_ratio/high_max": 0.002237422355392482,
      "clip_ratio/high_mean": 0.0009927220271492843,
      "clip_ratio/low_mean": 0.001034128505125409,
      "clip_ratio/low_min": 1.8121194443665445e-05,
      "clip_ratio/region_mean": 0.0020268505613785237,
      "epoch": 13.270553935860057,
      "grad_norm": 0.12003553658723831,
      "learning_rate": 1e-06,
      "loss": -0.0107,
      "step": 861
    },
    {
      "clip_ratio/high_max": 0.0026718492299551144,
      "clip_ratio/high_mean": 0.0011057402061851462,
      "clip_ratio/low_mean": 0.000998317602352472,
      "clip_ratio/low_min": 2.183024844271131e-05,
      "clip_ratio/region_mean": 0.0021040578503743745,
      "epoch": 13.279883381924199,
      "grad_norm": 0.11906471103429794,
      "learning_rate": 1e-06,
      "loss": -0.0058,
      "step": 862
    },
    {
      "clip_ratio/high_max": 0.002436835922708269,
      "clip_ratio/high_mean": 0.0011377226146578323,
      "clip_ratio/low_mean": 0.0011466565392765915,
      "clip_ratio/low_min": 6.766712613170967e-05,
      "clip_ratio/region_mean": 0.0022843791739433073,
      "epoch": 13.289212827988338,
      "grad_norm": 0.13598018884658813,
      "learning_rate": 1e-06,
      "loss": -0.032,
      "step": 863
    },
    {
      "clip_ratio/high_max": 0.00240250728893443,
      "clip_ratio/high_mean": 0.0009240321760444203,
      "clip_ratio/low_mean": 0.0013703959302802105,
      "clip_ratio/low_min": 5.9777506976388395e-05,
      "clip_ratio/region_mean": 0.00229442810814362,
      "epoch": 13.298542274052478,
      "grad_norm": 0.12569187581539154,
      "learning_rate": 1e-06,
      "loss": 0.0154,
      "step": 864
    },
    {
      "clip_ratio/high_max": 0.0024366669458686374,
      "clip_ratio/high_mean": 0.0011648315248748986,
      "clip_ratio/low_mean": 0.001137299001129577,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023021304805297405,
      "epoch": 13.307871720116617,
      "grad_norm": 0.12890926003456116,
      "learning_rate": 1e-06,
      "loss": -0.0111,
      "step": 865
    },
    {
      "clip_ratio/high_max": 0.0026204201349173672,
      "clip_ratio/high_mean": 0.0010493731751921587,
      "clip_ratio/low_mean": 0.0010249173756164964,
      "clip_ratio/low_min": 7.662937423447147e-05,
      "clip_ratio/region_mean": 0.0020742905762745067,
      "epoch": 13.317201166180759,
      "grad_norm": 0.12786240875720978,
      "learning_rate": 1e-06,
      "loss": -0.0215,
      "step": 866
    },
    {
      "clip_ratio/high_max": 0.0025590010554878972,
      "clip_ratio/high_mean": 0.0010956065634672996,
      "clip_ratio/low_mean": 0.0010608637512632413,
      "clip_ratio/low_min": 2.8189984732307494e-05,
      "clip_ratio/region_mean": 0.0021564703274634667,
      "epoch": 13.326530612244898,
      "grad_norm": 0.130562424659729,
      "learning_rate": 1e-06,
      "loss": -0.0255,
      "step": 867
    },
    {
      "clip_ratio/high_max": 0.0033513108792249113,
      "clip_ratio/high_mean": 0.001242486308910884,
      "clip_ratio/low_mean": 0.001135861690272577,
      "clip_ratio/low_min": 9.898459393298253e-05,
      "clip_ratio/region_mean": 0.002378347904596012,
      "epoch": 13.335860058309038,
      "grad_norm": 0.14444439113140106,
      "learning_rate": 1e-06,
      "loss": 0.0276,
      "step": 868
    },
    {
      "clip_ratio/high_max": 0.003126221774436999,
      "clip_ratio/high_mean": 0.001074210802471498,
      "clip_ratio/low_mean": 0.0009838177684287075,
      "clip_ratio/low_min": 4.917968726658728e-05,
      "clip_ratio/region_mean": 0.002058028578176163,
      "epoch": 13.345189504373177,
      "grad_norm": 0.12873801589012146,
      "learning_rate": 1e-06,
      "loss": -0.026,
      "step": 869
    },
    {
      "clip_ratio/high_max": 0.0026007405613199808,
      "clip_ratio/high_mean": 0.0009901737230393337,
      "clip_ratio/low_mean": 0.0011166982949362136,
      "clip_ratio/low_min": 4.587155854096636e-05,
      "clip_ratio/region_mean": 0.0021068719725008123,
      "epoch": 13.354518950437317,
      "grad_norm": 0.12822173535823822,
      "learning_rate": 1e-06,
      "loss": 0.0314,
      "step": 870
    },
    {
      "clip_ratio/high_max": 0.0028998194757150486,
      "clip_ratio/high_mean": 0.001148977673437912,
      "clip_ratio/low_mean": 0.0009931444365065545,
      "clip_ratio/low_min": 3.6829504097113386e-05,
      "clip_ratio/region_mean": 0.0021421221099444665,
      "epoch": 13.363848396501458,
      "grad_norm": 0.12017261981964111,
      "learning_rate": 1e-06,
      "loss": -0.0332,
      "step": 871
    },
    {
      "clip_ratio/high_max": 0.002861479646526277,
      "clip_ratio/high_mean": 0.0011401223673601635,
      "clip_ratio/low_mean": 0.001102754142266349,
      "clip_ratio/low_min": 4.90837701363489e-05,
      "clip_ratio/region_mean": 0.002242876493255608,
      "epoch": 13.373177842565598,
      "grad_norm": 0.12584428489208221,
      "learning_rate": 1e-06,
      "loss": 0.0115,
      "step": 872
    },
    {
      "clip_ratio/high_max": 0.002275574619488907,
      "clip_ratio/high_mean": 0.0010041766422546061,
      "clip_ratio/low_mean": 0.0010493665486137616,
      "clip_ratio/low_min": 4.434720904100686e-05,
      "clip_ratio/region_mean": 0.002053543126748991,
      "epoch": 13.382507288629737,
      "grad_norm": 0.12351217120885849,
      "learning_rate": 1e-06,
      "loss": -0.0225,
      "step": 873
    },
    {
      "clip_ratio/high_max": 0.0027673511940520257,
      "clip_ratio/high_mean": 0.0013027097847952973,
      "clip_ratio/low_mean": 0.0009989740756282117,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002301683896803297,
      "epoch": 13.391836734693877,
      "grad_norm": 0.1255517154932022,
      "learning_rate": 1e-06,
      "loss": -0.0638,
      "step": 874
    },
    {
      "clip_ratio/high_max": 0.002537508604291361,
      "clip_ratio/high_mean": 0.0011440946836955845,
      "clip_ratio/low_mean": 0.0010802688193507493,
      "clip_ratio/low_min": 1.6747051631682552e-05,
      "clip_ratio/region_mean": 0.0022243634884944186,
      "epoch": 13.401166180758018,
      "grad_norm": 0.13426025211811066,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "step": 875
    },
    {
      "clip_ratio/high_max": 0.0023940327009768225,
      "clip_ratio/high_mean": 0.0009865128122328315,
      "clip_ratio/low_mean": 0.001143872446846217,
      "clip_ratio/low_min": 0.00011511968023114605,
      "clip_ratio/region_mean": 0.002130385211785324,
      "epoch": 13.410495626822158,
      "grad_norm": 0.13342955708503723,
      "learning_rate": 1e-06,
      "loss": 0.0557,
      "step": 876
    },
    {
      "clip_ratio/high_max": 0.0021587336450465955,
      "clip_ratio/high_mean": 0.0009314640647062333,
      "clip_ratio/low_mean": 0.0009676771696831565,
      "clip_ratio/low_min": 3.904512686858652e-05,
      "clip_ratio/region_mean": 0.0018991412507602945,
      "epoch": 13.419825072886297,
      "grad_norm": 169.31161499023438,
      "learning_rate": 1e-06,
      "loss": -0.0246,
      "step": 877
    },
    {
      "clip_ratio/high_max": 0.0026572888673399575,
      "clip_ratio/high_mean": 0.0011478262495074887,
      "clip_ratio/low_mean": 0.0009573229472152889,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021051491421530955,
      "epoch": 13.429154518950437,
      "grad_norm": 0.12571832537651062,
      "learning_rate": 1e-06,
      "loss": -0.0261,
      "step": 878
    },
    {
      "clip_ratio/high_max": 0.0026892770220001694,
      "clip_ratio/high_mean": 0.0011772154048230732,
      "clip_ratio/low_mean": 0.00111311878390552,
      "clip_ratio/low_min": 0.00010643437963153701,
      "clip_ratio/region_mean": 0.0022903342178324237,
      "epoch": 13.438483965014576,
      "grad_norm": 0.12426993250846863,
      "learning_rate": 1e-06,
      "loss": -0.0212,
      "step": 879
    },
    {
      "clip_ratio/high_max": 0.0029829922132194042,
      "clip_ratio/high_mean": 0.0013295427525008563,
      "clip_ratio/low_mean": 0.0010098710590682458,
      "clip_ratio/low_min": 6.025064431014471e-05,
      "clip_ratio/region_mean": 0.0023394138042931445,
      "epoch": 13.447813411078718,
      "grad_norm": 0.13399332761764526,
      "learning_rate": 1e-06,
      "loss": -0.0399,
      "step": 880
    },
    {
      "clip_ratio/high_max": 0.0030106031917966902,
      "clip_ratio/high_mean": 0.0011709192658599932,
      "clip_ratio/low_mean": 0.0010956884179904591,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022666076765744947,
      "epoch": 13.457142857142857,
      "grad_norm": 0.11155635118484497,
      "learning_rate": 1e-06,
      "loss": -0.0429,
      "step": 881
    },
    {
      "clip_ratio/high_max": 0.002703793994442094,
      "clip_ratio/high_mean": 0.0013057460164418444,
      "clip_ratio/low_mean": 0.0011353579229762545,
      "clip_ratio/low_min": 2.9592803912237287e-05,
      "clip_ratio/region_mean": 0.0024411039194092155,
      "epoch": 13.466472303206997,
      "grad_norm": 0.19284729659557343,
      "learning_rate": 1e-06,
      "loss": -0.0232,
      "step": 882
    },
    {
      "clip_ratio/high_max": 0.0027425165462773293,
      "clip_ratio/high_mean": 0.001239390367118176,
      "clip_ratio/low_mean": 0.0011183500464539975,
      "clip_ratio/low_min": 2.1777002984890714e-05,
      "clip_ratio/region_mean": 0.002357740420848131,
      "epoch": 13.475801749271136,
      "grad_norm": 0.12508141994476318,
      "learning_rate": 1e-06,
      "loss": -0.0114,
      "step": 883
    },
    {
      "clip_ratio/high_max": 0.0029979851315147243,
      "clip_ratio/high_mean": 0.001300359141168883,
      "clip_ratio/low_mean": 0.0013067465224594343,
      "clip_ratio/low_min": 8.133424853440374e-05,
      "clip_ratio/region_mean": 0.002607105576316826,
      "epoch": 13.485131195335278,
      "grad_norm": 0.1267627328634262,
      "learning_rate": 1e-06,
      "loss": -0.0156,
      "step": 884
    },
    {
      "clip_ratio/high_max": 0.003407486656215042,
      "clip_ratio/high_mean": 0.0013521736182156019,
      "clip_ratio/low_mean": 0.0011613209499046206,
      "clip_ratio/low_min": 4.535382322501391e-05,
      "clip_ratio/region_mean": 0.0025134946263278835,
      "epoch": 13.494460641399417,
      "grad_norm": 0.12223848700523376,
      "learning_rate": 1e-06,
      "loss": -0.0419,
      "step": 885
    },
    {
      "clip_ratio/high_max": 0.0033824653874034993,
      "clip_ratio/high_mean": 0.0014186567859724164,
      "clip_ratio/low_mean": 0.0012371748562145513,
      "clip_ratio/low_min": 1.256533960258821e-05,
      "clip_ratio/region_mean": 0.0026558317040326074,
      "epoch": 13.503790087463557,
      "grad_norm": 0.13529939949512482,
      "learning_rate": 1e-06,
      "loss": -0.0143,
      "step": 886
    },
    {
      "clip_ratio/high_max": 0.0028677445152425207,
      "clip_ratio/high_mean": 0.001251653406143305,
      "clip_ratio/low_mean": 0.00124818472431798,
      "clip_ratio/low_min": 9.836300978349755e-05,
      "clip_ratio/region_mean": 0.0024998381777550094,
      "epoch": 13.513119533527696,
      "grad_norm": 0.12827137112617493,
      "learning_rate": 1e-06,
      "loss": -0.0363,
      "step": 887
    },
    {
      "clip_ratio/high_max": 0.002635695884237066,
      "clip_ratio/high_mean": 0.0011737718996300828,
      "clip_ratio/low_mean": 0.0012570931066875346,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024308649881277233,
      "epoch": 13.522448979591836,
      "grad_norm": 0.1189485639333725,
      "learning_rate": 1e-06,
      "loss": -0.0117,
      "step": 888
    },
    {
      "clip_ratio/high_max": 0.0027452050198917277,
      "clip_ratio/high_mean": 0.00115804493361793,
      "clip_ratio/low_mean": 0.0012230700776854064,
      "clip_ratio/low_min": 5.9580554079730064e-05,
      "clip_ratio/region_mean": 0.0023811150022083893,
      "epoch": 13.531778425655977,
      "grad_norm": 0.12266857922077179,
      "learning_rate": 1e-06,
      "loss": 0.0057,
      "step": 889
    },
    {
      "clip_ratio/high_max": 0.0028218238439876586,
      "clip_ratio/high_mean": 0.0012595596017490607,
      "clip_ratio/low_mean": 0.0015888533744146116,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002848412965249736,
      "epoch": 13.541107871720117,
      "grad_norm": 0.22287705540657043,
      "learning_rate": 1e-06,
      "loss": 0.0243,
      "step": 890
    },
    {
      "clip_ratio/high_max": 0.00284343878593063,
      "clip_ratio/high_mean": 0.0011793407102231868,
      "clip_ratio/low_mean": 0.0012273779684619512,
      "clip_ratio/low_min": 5.764590241597034e-05,
      "clip_ratio/region_mean": 0.0024067187041509897,
      "epoch": 13.550437317784256,
      "grad_norm": 0.13702525198459625,
      "learning_rate": 1e-06,
      "loss": -0.0187,
      "step": 891
    },
    {
      "clip_ratio/high_max": 0.0022639693816017825,
      "clip_ratio/high_mean": 0.0010550611314101843,
      "clip_ratio/low_mean": 0.0012934463629790116,
      "clip_ratio/low_min": 0.00016416675498476252,
      "clip_ratio/region_mean": 0.002348507525312016,
      "epoch": 13.559766763848396,
      "grad_norm": 0.13432511687278748,
      "learning_rate": 1e-06,
      "loss": -0.005,
      "step": 892
    },
    {
      "clip_ratio/high_max": 0.0024704822135390714,
      "clip_ratio/high_mean": 0.0011700084432959557,
      "clip_ratio/low_mean": 0.0015362148824351607,
      "clip_ratio/low_min": 2.7103209504275583e-05,
      "clip_ratio/region_mean": 0.002706223356653936,
      "epoch": 13.569096209912537,
      "grad_norm": 0.1321430504322052,
      "learning_rate": 1e-06,
      "loss": 0.0151,
      "step": 893
    },
    {
      "clip_ratio/high_max": 0.0029745934589300305,
      "clip_ratio/high_mean": 0.0011174059800396208,
      "clip_ratio/low_mean": 0.0013322115100891097,
      "clip_ratio/low_min": 2.7630416298052296e-05,
      "clip_ratio/region_mean": 0.0024496175174135715,
      "epoch": 13.578425655976677,
      "grad_norm": 0.12417371571063995,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 894
    },
    {
      "clip_ratio/high_max": 0.0031292916391976178,
      "clip_ratio/high_mean": 0.0013402972799667623,
      "clip_ratio/low_mean": 0.0013383482728386298,
      "clip_ratio/low_min": 4.3599582568276674e-05,
      "clip_ratio/region_mean": 0.002678645570995286,
      "epoch": 13.587755102040816,
      "grad_norm": 0.14334237575531006,
      "learning_rate": 1e-06,
      "loss": -0.0505,
      "step": 895
    },
    {
      "clip_ratio/high_max": 0.00333062898425851,
      "clip_ratio/high_mean": 0.001381813061016146,
      "clip_ratio/low_mean": 0.0014490344256046228,
      "clip_ratio/low_min": 0.0001234303508681478,
      "clip_ratio/region_mean": 0.0028308474211371504,
      "epoch": 13.597084548104956,
      "grad_norm": 0.13221228122711182,
      "learning_rate": 1e-06,
      "loss": -0.0132,
      "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": 0.0567103794642857,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4095.0,
      "completions/mean_length": 755.4510498046875,
      "completions/mean_terminated_length": 554.6179809570312,
      "completions/min_length": 74.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 14.00932944606414,
      "grad_norm": 0.14969795942306519,
      "learning_rate": 1e-06,
      "loss": -0.0266,
      "num_tokens": 553113682.0,
      "reward": 0.6208147406578064,
      "reward_std": 0.17625229060649872,
      "rewards/simpleverify_reward/mean": 0.6208147406578064,
      "rewards/simpleverify_reward/std": 0.4851885437965393,
      "step": 897
    },
    {
      "clip_ratio/high_max": 0.0021321147651178762,
      "clip_ratio/high_mean": 0.0009024411319842329,
      "clip_ratio/low_mean": 0.0004685198714469152,
      "clip_ratio/low_min": 1.8529499357100576e-05,
      "clip_ratio/region_mean": 0.0013709610102523584,
      "epoch": 14.018658892128279,
      "grad_norm": 0.13002531230449677,
      "learning_rate": 1e-06,
      "loss": -0.0539,
      "step": 898
    },
    {
      "clip_ratio/high_max": 0.0022830904053989798,
      "clip_ratio/high_mean": 0.0009456026236875914,
      "clip_ratio/low_mean": 0.0005751479429818573,
      "clip_ratio/low_min": 2.8988868507440202e-05,
      "clip_ratio/region_mean": 0.001520750480267452,
      "epoch": 14.02798833819242,
      "grad_norm": 0.13815423846244812,
      "learning_rate": 1e-06,
      "loss": -0.0121,
      "step": 899
    },
    {
      "clip_ratio/high_max": 0.0021398777134891134,
      "clip_ratio/high_mean": 0.0009047834009834332,
      "clip_ratio/low_mean": 0.000516072619575425,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014208560387487523,
      "epoch": 14.03731778425656,
      "grad_norm": 0.13017743825912476,
      "learning_rate": 1e-06,
      "loss": -0.0709,
      "step": 900
    },
    {
      "clip_ratio/high_max": 0.001989206466532778,
      "clip_ratio/high_mean": 0.0008325697745021898,
      "clip_ratio/low_mean": 0.0005534265819733264,
      "clip_ratio/low_min": 5.677307581208879e-05,
      "clip_ratio/region_mean": 0.0013859963473805692,
      "epoch": 14.0466472303207,
      "grad_norm": 0.12986066937446594,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 901
    },
    {
      "clip_ratio/high_max": 0.0024154558777809143,
      "clip_ratio/high_mean": 0.000894890799827408,
      "clip_ratio/low_mean": 0.0005908537896175403,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014857446039968636,
      "epoch": 14.055976676384839,
      "grad_norm": 0.14485520124435425,
      "learning_rate": 1e-06,
      "loss": -0.012,
      "step": 902
    },
    {
      "clip_ratio/high_max": 0.002370783913647756,
      "clip_ratio/high_mean": 0.0009642664736020379,
      "clip_ratio/low_mean": 0.0007347641367232427,
      "clip_ratio/low_min": 4.8053605496534146e-05,
      "clip_ratio/region_mean": 0.0016990305957733653,
      "epoch": 14.06530612244898,
      "grad_norm": 0.14210668206214905,
      "learning_rate": 1e-06,
      "loss": -0.0205,
      "step": 903
    },
    {
      "clip_ratio/high_max": 0.0022944658921915106,
      "clip_ratio/high_mean": 0.0009390147315571085,
      "clip_ratio/low_mean": 0.0006571683288711938,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015961830504238605,
      "epoch": 14.07463556851312,
      "grad_norm": 0.12780068814754486,
      "learning_rate": 1e-06,
      "loss": -0.0047,
      "step": 904
    },
    {
      "clip_ratio/high_max": 0.0020741868356708437,
      "clip_ratio/high_mean": 0.0009202171459037345,
      "clip_ratio/low_mean": 0.0006916583797647036,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016118755338538904,
      "epoch": 14.08396501457726,
      "grad_norm": 0.13621775805950165,
      "learning_rate": 1e-06,
      "loss": -0.0249,
      "step": 905
    },
    {
      "clip_ratio/high_max": 0.002060031853034161,
      "clip_ratio/high_mean": 0.0008378832617381704,
      "clip_ratio/low_mean": 0.0006564872364833718,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014943705027690157,
      "epoch": 14.093294460641399,
      "grad_norm": 0.12371636927127838,
      "learning_rate": 1e-06,
      "loss": 0.0006,
      "step": 906
    },
    {
      "clip_ratio/high_max": 0.0022291329005383886,
      "clip_ratio/high_mean": 0.000812384401797317,
      "clip_ratio/low_mean": 0.000691331773850834,
      "clip_ratio/low_min": 2.2045855075703003e-05,
      "clip_ratio/region_mean": 0.0015037161865620874,
      "epoch": 14.102623906705539,
      "grad_norm": 0.11786044389009476,
      "learning_rate": 1e-06,
      "loss": -0.0108,
      "step": 907
    },
    {
      "clip_ratio/high_max": 0.0022796319535700604,
      "clip_ratio/high_mean": 0.0009864562343864236,
      "clip_ratio/low_mean": 0.0008921745311454288,
      "clip_ratio/low_min": 2.2316601643979084e-05,
      "clip_ratio/region_mean": 0.0018786307737173047,
      "epoch": 14.11195335276968,
      "grad_norm": 0.14417092502117157,
      "learning_rate": 1e-06,
      "loss": -0.0218,
      "step": 908
    },
    {
      "clip_ratio/high_max": 0.002153070417989511,
      "clip_ratio/high_mean": 0.0009623824680602411,
      "clip_ratio/low_mean": 0.0008726408050279133,
      "clip_ratio/low_min": 6.024687354511116e-05,
      "clip_ratio/region_mean": 0.001835023271269165,
      "epoch": 14.12128279883382,
      "grad_norm": 0.12148705869913101,
      "learning_rate": 1e-06,
      "loss": 0.0134,
      "step": 909
    },
    {
      "clip_ratio/high_max": 0.001995118276681751,
      "clip_ratio/high_mean": 0.0008352095665031811,
      "clip_ratio/low_mean": 0.0008332733141287463,
      "clip_ratio/low_min": 4.991334481019294e-05,
      "clip_ratio/region_mean": 0.0016684829024598002,
      "epoch": 14.130612244897959,
      "grad_norm": 0.12048781663179398,
      "learning_rate": 1e-06,
      "loss": -0.0127,
      "step": 910
    },
    {
      "clip_ratio/high_max": 0.0025849801531876437,
      "clip_ratio/high_mean": 0.0011180993606103584,
      "clip_ratio/low_mean": 0.0007752993515168782,
      "clip_ratio/low_min": 3.476073470665142e-05,
      "clip_ratio/region_mean": 0.0018933986648335122,
      "epoch": 14.139941690962099,
      "grad_norm": 0.12970127165317535,
      "learning_rate": 1e-06,
      "loss": -0.0652,
      "step": 911
    },
    {
      "clip_ratio/high_max": 0.002135657730832463,
      "clip_ratio/high_mean": 0.0008769598407525336,
      "clip_ratio/low_mean": 0.0009745557708811248,
      "clip_ratio/low_min": 6.452128764067311e-05,
      "clip_ratio/region_mean": 0.0018515156116336584,
      "epoch": 14.14927113702624,
      "grad_norm": 0.11937867105007172,
      "learning_rate": 1e-06,
      "loss": 0.0028,
      "step": 912
    },
    {
      "clip_ratio/high_max": 0.001953250155565911,
      "clip_ratio/high_mean": 0.0008245341159636155,
      "clip_ratio/low_mean": 0.0010529869605306885,
      "clip_ratio/low_min": 5.208333459449932e-05,
      "clip_ratio/region_mean": 0.0018775210191961378,
      "epoch": 14.15860058309038,
      "grad_norm": 0.11515150964260101,
      "learning_rate": 1e-06,
      "loss": 0.0128,
      "step": 913
    },
    {
      "clip_ratio/high_max": 0.002455089539580513,
      "clip_ratio/high_mean": 0.0010868431236303877,
      "clip_ratio/low_mean": 0.0009100359884541831,
      "clip_ratio/low_min": 8.621878805570304e-06,
      "clip_ratio/region_mean": 0.001996879102080129,
      "epoch": 14.167930029154519,
      "grad_norm": 0.12120584398508072,
      "learning_rate": 1e-06,
      "loss": -0.0255,
      "step": 914
    },
    {
      "clip_ratio/high_max": 0.0023854992068663705,
      "clip_ratio/high_mean": 0.0010666750276868697,
      "clip_ratio/low_mean": 0.0009492768058407819,
      "clip_ratio/low_min": 3.2316442229785025e-05,
      "clip_ratio/region_mean": 0.00201595183898462,
      "epoch": 14.177259475218658,
      "grad_norm": 0.13539746403694153,
      "learning_rate": 1e-06,
      "loss": -0.0224,
      "step": 915
    },
    {
      "clip_ratio/high_max": 0.0022336070323945023,
      "clip_ratio/high_mean": 0.000979129787083366,
      "clip_ratio/low_mean": 0.0010042173016699962,
      "clip_ratio/low_min": 4.6387925976887345e-05,
      "clip_ratio/region_mean": 0.0019833470360026695,
      "epoch": 14.186588921282798,
      "grad_norm": 0.11462336033582687,
      "learning_rate": 1e-06,
      "loss": -0.0209,
      "step": 916
    },
    {
      "clip_ratio/high_max": 0.002345109787711408,
      "clip_ratio/high_mean": 0.0009678079386503669,
      "clip_ratio/low_mean": 0.0011101311465608887,
      "clip_ratio/low_min": 0.00016861287076608278,
      "clip_ratio/region_mean": 0.002077939127048012,
      "epoch": 14.19591836734694,
      "grad_norm": 0.13211119174957275,
      "learning_rate": 1e-06,
      "loss": 0.0308,
      "step": 917
    },
    {
      "clip_ratio/high_max": 0.0022739557971362956,
      "clip_ratio/high_mean": 0.0009184006657960708,
      "clip_ratio/low_mean": 0.000950915860812529,
      "clip_ratio/low_min": 3.9835543248045724e-05,
      "clip_ratio/region_mean": 0.0018693165038712323,
      "epoch": 14.205247813411079,
      "grad_norm": 0.13406263291835785,
      "learning_rate": 1e-06,
      "loss": 0.0035,
      "step": 918
    },
    {
      "clip_ratio/high_max": 0.002644911073730327,
      "clip_ratio/high_mean": 0.0010202525809290819,
      "clip_ratio/low_mean": 0.001061573395418236,
      "clip_ratio/low_min": 4.7071792323549744e-05,
      "clip_ratio/region_mean": 0.002081826009089127,
      "epoch": 14.214577259475218,
      "grad_norm": 0.1266704946756363,
      "learning_rate": 1e-06,
      "loss": -0.0083,
      "step": 919
    },
    {
      "clip_ratio/high_max": 0.0023664222608204,
      "clip_ratio/high_mean": 0.001073367559001781,
      "clip_ratio/low_mean": 0.0011838728933071252,
      "clip_ratio/low_min": 0.00015215515668387525,
      "clip_ratio/region_mean": 0.0022572404996026307,
      "epoch": 14.223906705539358,
      "grad_norm": 0.19711726903915405,
      "learning_rate": 1e-06,
      "loss": -0.0265,
      "step": 920
    },
    {
      "clip_ratio/high_max": 0.0026139828405575827,
      "clip_ratio/high_mean": 0.0011277917292318307,
      "clip_ratio/low_mean": 0.0009537962541799061,
      "clip_ratio/low_min": 3.863409256155137e-05,
      "clip_ratio/region_mean": 0.002081587983411737,
      "epoch": 14.2332361516035,
      "grad_norm": 0.1462983787059784,
      "learning_rate": 1e-06,
      "loss": -0.0053,
      "step": 921
    },
    {
      "clip_ratio/high_max": 0.0030707967816852033,
      "clip_ratio/high_mean": 0.0013405666541075334,
      "clip_ratio/low_mean": 0.0007491079650208121,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002089674620947335,
      "epoch": 14.242565597667639,
      "grad_norm": 0.14318053424358368,
      "learning_rate": 1e-06,
      "loss": -0.0822,
      "step": 922
    },
    {
      "clip_ratio/high_max": 0.002247482319944538,
      "clip_ratio/high_mean": 0.0010654392408468993,
      "clip_ratio/low_mean": 0.0009853786941675935,
      "clip_ratio/low_min": 6.670921356999315e-05,
      "clip_ratio/region_mean": 0.002050817958661355,
      "epoch": 14.251895043731778,
      "grad_norm": 0.12203782051801682,
      "learning_rate": 1e-06,
      "loss": -0.032,
      "step": 923
    },
    {
      "clip_ratio/high_max": 0.002335211720492225,
      "clip_ratio/high_mean": 0.001002703960693907,
      "clip_ratio/low_mean": 0.0009630866152292583,
      "clip_ratio/low_min": 3.251755333621986e-05,
      "clip_ratio/region_mean": 0.001965790521353483,
      "epoch": 14.261224489795918,
      "grad_norm": 0.1335146427154541,
      "learning_rate": 1e-06,
      "loss": -0.0164,
      "step": 924
    },
    {
      "clip_ratio/high_max": 0.002520498019293882,
      "clip_ratio/high_mean": 0.0010447496115375543,
      "clip_ratio/low_mean": 0.001091637819627067,
      "clip_ratio/low_min": 8.63748691699584e-05,
      "clip_ratio/region_mean": 0.002136387411155738,
      "epoch": 14.270553935860057,
      "grad_norm": 0.13873517513275146,
      "learning_rate": 1e-06,
      "loss": -0.0168,
      "step": 925
    },
    {
      "clip_ratio/high_max": 0.002843659560312517,
      "clip_ratio/high_mean": 0.0011851942908833735,
      "clip_ratio/low_mean": 0.0010693122567317914,
      "clip_ratio/low_min": 3.4054914067382924e-05,
      "clip_ratio/region_mean": 0.002254506529425271,
      "epoch": 14.279883381924199,
      "grad_norm": 0.13015536963939667,
      "learning_rate": 1e-06,
      "loss": 0.0139,
      "step": 926
    },
    {
      "clip_ratio/high_max": 0.0022768777307646815,
      "clip_ratio/high_mean": 0.0010951667118206387,
      "clip_ratio/low_mean": 0.0010495628375792876,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021447295221150853,
      "epoch": 14.289212827988338,
      "grad_norm": 0.126378133893013,
      "learning_rate": 1e-06,
      "loss": -0.0296,
      "step": 927
    },
    {
      "clip_ratio/high_max": 0.0024771498938207515,
      "clip_ratio/high_mean": 0.0010968905589834321,
      "clip_ratio/low_mean": 0.000995513319139718,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020924038835801184,
      "epoch": 14.298542274052478,
      "grad_norm": 0.12050725519657135,
      "learning_rate": 1e-06,
      "loss": -0.0201,
      "step": 928
    },
    {
      "clip_ratio/high_max": 0.0022828545479569584,
      "clip_ratio/high_mean": 0.0010056908104161266,
      "clip_ratio/low_mean": 0.0010011418344220147,
      "clip_ratio/low_min": 4.3749618271249346e-05,
      "clip_ratio/region_mean": 0.0020068326630280353,
      "epoch": 14.307871720116617,
      "grad_norm": 0.13950660824775696,
      "learning_rate": 1e-06,
      "loss": -0.0087,
      "step": 929
    },
    {
      "clip_ratio/high_max": 0.002354379110329319,
      "clip_ratio/high_mean": 0.0009210168664139928,
      "clip_ratio/low_mean": 0.001131519387854496,
      "clip_ratio/low_min": 5.436201990960399e-05,
      "clip_ratio/region_mean": 0.002052536314295139,
      "epoch": 14.317201166180759,
      "grad_norm": 0.12580522894859314,
      "learning_rate": 1e-06,
      "loss": 0.0346,
      "step": 930
    },
    {
      "clip_ratio/high_max": 0.00236603675148217,
      "clip_ratio/high_mean": 0.0009832563082454726,
      "clip_ratio/low_mean": 0.0010800404124893248,
      "clip_ratio/low_min": 6.691541420877911e-05,
      "clip_ratio/region_mean": 0.00206329671345884,
      "epoch": 14.326530612244898,
      "grad_norm": 0.13978849351406097,
      "learning_rate": 1e-06,
      "loss": 0.0174,
      "step": 931
    },
    {
      "clip_ratio/high_max": 0.0024994046543724835,
      "clip_ratio/high_mean": 0.0010441215763421496,
      "clip_ratio/low_mean": 0.0010637633786245715,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021078849022160284,
      "epoch": 14.335860058309038,
      "grad_norm": 0.12466134876012802,
      "learning_rate": 1e-06,
      "loss": -0.0146,
      "step": 932
    },
    {
      "clip_ratio/high_max": 0.0025941191852325574,
      "clip_ratio/high_mean": 0.0010674317891243845,
      "clip_ratio/low_mean": 0.0012813621251552831,
      "clip_ratio/low_min": 0.00015002742475189734,
      "clip_ratio/region_mean": 0.002348794019781053,
      "epoch": 14.345189504373177,
      "grad_norm": 0.13128824532032013,
      "learning_rate": 1e-06,
      "loss": -0.0697,
      "step": 933
    },
    {
      "clip_ratio/high_max": 0.002672317343240138,
      "clip_ratio/high_mean": 0.0010759210090327542,
      "clip_ratio/low_mean": 0.0009295324853155762,
      "clip_ratio/low_min": 6.443041093007196e-05,
      "clip_ratio/region_mean": 0.002005453541642055,
      "epoch": 14.354518950437317,
      "grad_norm": 0.12702219188213348,
      "learning_rate": 1e-06,
      "loss": -0.0073,
      "step": 934
    },
    {
      "clip_ratio/high_max": 0.0025854838095256127,
      "clip_ratio/high_mean": 0.0011564231426746119,
      "clip_ratio/low_mean": 0.000901870291272644,
      "clip_ratio/low_min": 6.69623950670939e-05,
      "clip_ratio/region_mean": 0.00205829345213715,
      "epoch": 14.363848396501458,
      "grad_norm": 0.10431494563817978,
      "learning_rate": 1e-06,
      "loss": -0.0282,
      "step": 935
    },
    {
      "clip_ratio/high_max": 0.0031442917825188488,
      "clip_ratio/high_mean": 0.0013375504713621922,
      "clip_ratio/low_mean": 0.0010702581530495081,
      "clip_ratio/low_min": 1.628452264412772e-05,
      "clip_ratio/region_mean": 0.002407808569842018,
      "epoch": 14.373177842565598,
      "grad_norm": 0.1322004199028015,
      "learning_rate": 1e-06,
      "loss": -0.0572,
      "step": 936
    },
    {
      "clip_ratio/high_max": 0.0028658176597673446,
      "clip_ratio/high_mean": 0.001189300099213142,
      "clip_ratio/low_mean": 0.0010691876595956273,
      "clip_ratio/low_min": 9.669288556324318e-05,
      "clip_ratio/region_mean": 0.0022584877515328117,
      "epoch": 14.382507288629737,
      "grad_norm": 0.13608138263225555,
      "learning_rate": 1e-06,
      "loss": -0.0288,
      "step": 937
    },
    {
      "clip_ratio/high_max": 0.0023503157062805258,
      "clip_ratio/high_mean": 0.0009750783065101132,
      "clip_ratio/low_mean": 0.001167000493296655,
      "clip_ratio/low_min": 8.580875601182925e-05,
      "clip_ratio/region_mean": 0.002142078836186556,
      "epoch": 14.391836734693877,
      "grad_norm": 0.12495539337396622,
      "learning_rate": 1e-06,
      "loss": 0.0133,
      "step": 938
    },
    {
      "clip_ratio/high_max": 0.0022415608109440655,
      "clip_ratio/high_mean": 0.0009491406817687675,
      "clip_ratio/low_mean": 0.0011565360055101337,
      "clip_ratio/low_min": 5.4371883379644714e-05,
      "clip_ratio/region_mean": 0.0021056766563560814,
      "epoch": 14.401166180758018,
      "grad_norm": 0.1135687455534935,
      "learning_rate": 1e-06,
      "loss": 0.0051,
      "step": 939
    },
    {
      "clip_ratio/high_max": 0.002703091755392961,
      "clip_ratio/high_mean": 0.0010301889396941988,
      "clip_ratio/low_mean": 0.0012471113259380218,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002277300249261316,
      "epoch": 14.410495626822158,
      "grad_norm": 0.12958236038684845,
      "learning_rate": 1e-06,
      "loss": 0.0285,
      "step": 940
    },
    {
      "clip_ratio/high_max": 0.0025375607510795817,
      "clip_ratio/high_mean": 0.001088365101168165,
      "clip_ratio/low_mean": 0.0009787153721845243,
      "clip_ratio/low_min": 3.8747675716876984e-05,
      "clip_ratio/region_mean": 0.0020670805097324774,
      "epoch": 14.419825072886297,
      "grad_norm": 9.442768096923828,
      "learning_rate": 1e-06,
      "loss": 0.0111,
      "step": 941
    },
    {
      "clip_ratio/high_max": 0.0030651435008621775,
      "clip_ratio/high_mean": 0.0013379396332311444,
      "clip_ratio/low_mean": 0.0011468818338471465,
      "clip_ratio/low_min": 2.466942896717228e-05,
      "clip_ratio/region_mean": 0.0024848214889061637,
      "epoch": 14.429154518950437,
      "grad_norm": 0.13983182609081268,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 942
    },
    {
      "clip_ratio/high_max": 0.0029531323016271926,
      "clip_ratio/high_mean": 0.0011889290435647126,
      "clip_ratio/low_mean": 0.0010633777219482,
      "clip_ratio/low_min": 6.785496589145623e-05,
      "clip_ratio/region_mean": 0.002252306730952114,
      "epoch": 14.438483965014576,
      "grad_norm": 0.14745375514030457,
      "learning_rate": 1e-06,
      "loss": -0.0342,
      "step": 943
    },
    {
      "clip_ratio/high_max": 0.0028511406635516323,
      "clip_ratio/high_mean": 0.0011925216422241647,
      "clip_ratio/low_mean": 0.0010012209040723974,
      "clip_ratio/low_min": 7.85820320743369e-05,
      "clip_ratio/region_mean": 0.0021937425626674667,
      "epoch": 14.447813411078718,
      "grad_norm": 0.12538109719753265,
      "learning_rate": 1e-06,
      "loss": -0.0538,
      "step": 944
    },
    {
      "clip_ratio/high_max": 0.002897155274695251,
      "clip_ratio/high_mean": 0.0012587637884280412,
      "clip_ratio/low_mean": 0.0012891206752101425,
      "clip_ratio/low_min": 0.00025955435557989404,
      "clip_ratio/region_mean": 0.00254788450547494,
      "epoch": 14.457142857142857,
      "grad_norm": 0.13619297742843628,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 945
    },
    {
      "clip_ratio/high_max": 0.0027963031388935633,
      "clip_ratio/high_mean": 0.0012036244552291464,
      "clip_ratio/low_mean": 0.0012208813459437806,
      "clip_ratio/low_min": 1.8565275240689516e-05,
      "clip_ratio/region_mean": 0.002424505779345054,
      "epoch": 14.466472303206997,
      "grad_norm": 0.12614497542381287,
      "learning_rate": 1e-06,
      "loss": -0.0217,
      "step": 946
    },
    {
      "clip_ratio/high_max": 0.0031505838414886966,
      "clip_ratio/high_mean": 0.0012830314590246417,
      "clip_ratio/low_mean": 0.0013210248434916139,
      "clip_ratio/low_min": 0.00010843661584658548,
      "clip_ratio/region_mean": 0.00260405625886051,
      "epoch": 14.475801749271136,
      "grad_norm": 0.11964280903339386,
      "learning_rate": 1e-06,
      "loss": -0.0396,
      "step": 947
    },
    {
      "clip_ratio/high_max": 0.0026417123517603613,
      "clip_ratio/high_mean": 0.0011750264529837295,
      "clip_ratio/low_mean": 0.0011886841311934404,
      "clip_ratio/low_min": 1.3363267498789355e-05,
      "clip_ratio/region_mean": 0.0023637106132810004,
      "epoch": 14.485131195335278,
      "grad_norm": 0.11716482788324356,
      "learning_rate": 1e-06,
      "loss": -0.0256,
      "step": 948
    },
    {
      "clip_ratio/high_max": 0.003004811624123249,
      "clip_ratio/high_mean": 0.001232846869243076,
      "clip_ratio/low_mean": 0.0012912543406855548,
      "clip_ratio/low_min": 2.8191248929942958e-05,
      "clip_ratio/region_mean": 0.002524101219023578,
      "epoch": 14.494460641399417,
      "grad_norm": 0.11609780788421631,
      "learning_rate": 1e-06,
      "loss": -0.0191,
      "step": 949
    },
    {
      "clip_ratio/high_max": 0.0024833397910697386,
      "clip_ratio/high_mean": 0.0010518038980080746,
      "clip_ratio/low_mean": 0.0014151472623780137,
      "clip_ratio/low_min": 0.00021751447457063477,
      "clip_ratio/region_mean": 0.0024669511258252896,
      "epoch": 14.503790087463557,
      "grad_norm": 0.14569492638111115,
      "learning_rate": 1e-06,
      "loss": 0.0324,
      "step": 950
    },
    {
      "clip_ratio/high_max": 0.003082641000219155,
      "clip_ratio/high_mean": 0.0012449124187696725,
      "clip_ratio/low_mean": 0.0013270887320686597,
      "clip_ratio/low_min": 0.00018013046610576566,
      "clip_ratio/region_mean": 0.002572001176304184,
      "epoch": 14.513119533527696,
      "grad_norm": 0.13831104338169098,
      "learning_rate": 1e-06,
      "loss": -0.0115,
      "step": 951
    },
    {
      "clip_ratio/high_max": 0.002846763440174982,
      "clip_ratio/high_mean": 0.0013010627662879415,
      "clip_ratio/low_mean": 0.0013778409720544005,
      "clip_ratio/low_min": 0.00023421824880642816,
      "clip_ratio/region_mean": 0.002678903802006971,
      "epoch": 14.522448979591836,
      "grad_norm": 0.12651324272155762,
      "learning_rate": 1e-06,
      "loss": -0.0168,
      "step": 952
    },
    {
      "clip_ratio/high_max": 0.0027706861146725714,
      "clip_ratio/high_mean": 0.0012503973157436121,
      "clip_ratio/low_mean": 0.001177241669211071,
      "clip_ratio/low_min": 1.3941556971985847e-05,
      "clip_ratio/region_mean": 0.0024276388867292553,
      "epoch": 14.531778425655977,
      "grad_norm": 0.11958511173725128,
      "learning_rate": 1e-06,
      "loss": -0.0616,
      "step": 953
    },
    {
      "clip_ratio/high_max": 0.0027408288879087195,
      "clip_ratio/high_mean": 0.0011930600921914447,
      "clip_ratio/low_mean": 0.0012990522118343506,
      "clip_ratio/low_min": 0.0001541691890452057,
      "clip_ratio/region_mean": 0.002492112311301753,
      "epoch": 14.541107871720117,
      "grad_norm": 0.13056538999080658,
      "learning_rate": 1e-06,
      "loss": 0.0033,
      "step": 954
    },
    {
      "clip_ratio/high_max": 0.0033028795442078263,
      "clip_ratio/high_mean": 0.0012961380343767814,
      "clip_ratio/low_mean": 0.0013840430256095715,
      "clip_ratio/low_min": 3.1748090805194806e-05,
      "clip_ratio/region_mean": 0.0026801810672623105,
      "epoch": 14.550437317784256,
      "grad_norm": 0.1939610093832016,
      "learning_rate": 1e-06,
      "loss": -0.0077,
      "step": 955
    },
    {
      "clip_ratio/high_max": 0.0032667394043528475,
      "clip_ratio/high_mean": 0.0013424872740870342,
      "clip_ratio/low_mean": 0.0012122152184019797,
      "clip_ratio/low_min": 5.3465810196939856e-05,
      "clip_ratio/region_mean": 0.002554702492489014,
      "epoch": 14.559766763848396,
      "grad_norm": 0.13248389959335327,
      "learning_rate": 1e-06,
      "loss": -0.047,
      "step": 956
    },
    {
      "clip_ratio/high_max": 0.0028414863991201855,
      "clip_ratio/high_mean": 0.0013341539415705483,
      "clip_ratio/low_mean": 0.0012563679883896839,
      "clip_ratio/low_min": 0.00010780977936519776,
      "clip_ratio/region_mean": 0.0025905219445121475,
      "epoch": 14.569096209912537,
      "grad_norm": 0.11197350174188614,
      "learning_rate": 1e-06,
      "loss": -0.0341,
      "step": 957
    },
    {
      "clip_ratio/high_max": 0.0026658251917979214,
      "clip_ratio/high_mean": 0.0011300213773210999,
      "clip_ratio/low_mean": 0.001313895558268996,
      "clip_ratio/low_min": 8.55326934470213e-05,
      "clip_ratio/region_mean": 0.0024439169574179687,
      "epoch": 14.578425655976677,
      "grad_norm": 0.1402246505022049,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 958
    },
    {
      "clip_ratio/high_max": 0.002564258931670338,
      "clip_ratio/high_mean": 0.0010224614070466487,
      "clip_ratio/low_mean": 0.0013112426786392462,
      "clip_ratio/low_min": 3.2391810236731544e-05,
      "clip_ratio/region_mean": 0.0023337041202466935,
      "epoch": 14.587755102040816,
      "grad_norm": 0.125115767121315,
      "learning_rate": 1e-06,
      "loss": 0.0198,
      "step": 959
    },
    {
      "clip_ratio/high_max": 0.003075114087550901,
      "clip_ratio/high_mean": 0.0013087887800793396,
      "clip_ratio/low_mean": 0.001271657447432517,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002580446242063772,
      "epoch": 14.597084548104956,
      "grad_norm": 0.1357250213623047,
      "learning_rate": 1e-06,
      "loss": -0.0224,
      "step": 960
    },
    {
      "epoch": 14.597084548104956,
      "step": 960,
      "total_flos": 0.0,
      "train_loss": -0.0033660151399809973,
      "train_runtime": 58553.53,
      "train_samples_per_second": 24.484,
      "train_steps_per_second": 0.027
    }
  ],
  "logging_steps": 1,
  "max_steps": 1600,
  "num_input_tokens_seen": 553113682,
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}