{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.1909307875894988,
  "eval_steps": 10,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 1075.9285888671875,
      "epoch": 0.002386634844868735,
      "grad_norm": 0.25947600514173863,
      "kl": 0.0,
      "learning_rate": 7.692307692307692e-08,
      "loss": -0.0203,
      "reward": 1.1196562051773071,
      "reward_std": 0.6183667778968811,
      "rewards/": 4.645899772644043,
      "rewards/math_compute_score": 0.2380952388048172,
      "step": 1
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 827.4761962890625,
      "epoch": 0.00477326968973747,
      "grad_norm": 0.34697911338033505,
      "kl": 0.0,
      "learning_rate": 1.5384615384615385e-07,
      "loss": 0.1454,
      "reward": 1.1186012029647827,
      "reward_std": 0.9155174493789673,
      "rewards/": 4.069196701049805,
      "rewards/math_compute_score": 0.380952388048172,
      "step": 2
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 720.5714721679688,
      "epoch": 0.007159904534606206,
      "grad_norm": 0.32782371953159684,
      "kl": 1.9550323486328125e-05,
      "learning_rate": 2.3076923076923078e-07,
      "loss": 0.1558,
      "reward": 0.7518136501312256,
      "reward_std": 0.8834309577941895,
      "rewards/": 3.473353862762451,
      "rewards/math_compute_score": 0.0714285746216774,
      "step": 3
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1059.40478515625,
      "epoch": 0.00954653937947494,
      "grad_norm": 0.2765682789441865,
      "kl": 2.300739288330078e-05,
      "learning_rate": 3.076923076923077e-07,
      "loss": 0.2423,
      "reward": 0.8247302770614624,
      "reward_std": 0.8945653438568115,
      "rewards/": 3.456984758377075,
      "rewards/math_compute_score": 0.1666666716337204,
      "step": 4
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 960.7380981445312,
      "epoch": 0.011933174224343675,
      "grad_norm": 0.33895220525238917,
      "kl": 4.792213439941406e-05,
      "learning_rate": 3.8461538461538463e-07,
      "loss": 0.1198,
      "reward": 0.4260934293270111,
      "reward_std": 1.0363324880599976,
      "rewards/": 2.797133684158325,
      "rewards/math_compute_score": -0.1666666716337204,
      "step": 5
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1122.1190185546875,
      "epoch": 0.014319809069212411,
      "grad_norm": 0.627130910533,
      "kl": 1.6808509826660156e-05,
      "learning_rate": 4.6153846153846156e-07,
      "loss": 0.1927,
      "reward": 0.5224749445915222,
      "reward_std": 1.1910170316696167,
      "rewards/": 2.231422185897827,
      "rewards/math_compute_score": 0.095238097012043,
      "step": 6
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1001.7857055664062,
      "epoch": 0.016706443914081145,
      "grad_norm": 0.403975286781731,
      "kl": 4.9114227294921875e-05,
      "learning_rate": 5.384615384615384e-07,
      "loss": 0.0551,
      "reward": 0.82244473695755,
      "reward_std": 0.849882960319519,
      "rewards/": 4.207461357116699,
      "rewards/math_compute_score": -0.02380952425301075,
      "step": 7
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1074.1190185546875,
      "epoch": 0.01909307875894988,
      "grad_norm": 0.2627348199882925,
      "kl": 4.673004150390625e-05,
      "learning_rate": 6.153846153846154e-07,
      "loss": 0.1452,
      "reward": 0.9726702570915222,
      "reward_std": 0.9153575301170349,
      "rewards/": 4.291922569274902,
      "rewards/math_compute_score": 0.1428571492433548,
      "step": 8
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 696.3095092773438,
      "epoch": 0.021479713603818614,
      "grad_norm": 0.33742159571899366,
      "kl": 2.8848648071289062e-05,
      "learning_rate": 6.923076923076922e-07,
      "loss": 0.0265,
      "reward": 0.965959906578064,
      "reward_std": 0.6284574270248413,
      "rewards/": 3.115513563156128,
      "rewards/math_compute_score": 0.4285714328289032,
      "step": 9
    },
    {
      "epoch": 0.02386634844868735,
      "grad_norm": 0.35280633916428544,
      "learning_rate": 7.692307692307693e-07,
      "loss": 0.0653,
      "step": 10
    },
    {
      "epoch": 0.02386634844868735,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 1023.7500305175781,
      "eval_kl": 3.4868717193603516e-05,
      "eval_loss": 0.198669895529747,
      "eval_reward": 0.9118510186672211,
      "eval_reward_std": 0.8892157077789307,
      "eval_rewards/": 3.9640169739723206,
      "eval_rewards/math_compute_score": 0.1488095261156559,
      "eval_runtime": 82.3895,
      "eval_samples_per_second": 0.255,
      "eval_steps_per_second": 0.012,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 938.1309509277344,
      "epoch": 0.026252983293556086,
      "grad_norm": 0.30608376776798635,
      "kl": 4.845857620239258e-05,
      "learning_rate": 8.461538461538461e-07,
      "loss": 0.1056,
      "reward": 0.5514880195260048,
      "reward_std": 0.9235334098339081,
      "rewards/": 2.6145827770233154,
      "rewards/math_compute_score": 0.0357142835855484,
      "step": 11
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1085.9761962890625,
      "epoch": 0.028639618138424822,
      "grad_norm": 0.2661507893474065,
      "kl": 4.2438507080078125e-05,
      "learning_rate": 9.230769230769231e-07,
      "loss": 0.2348,
      "reward": 0.9119373559951782,
      "reward_std": 0.7280297875404358,
      "rewards/": 3.988258123397827,
      "rewards/math_compute_score": 0.1428571492433548,
      "step": 12
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1281.1190185546875,
      "epoch": 0.031026252983293555,
      "grad_norm": 0.26920466593905773,
      "kl": 3.2901763916015625e-05,
      "learning_rate": 1e-06,
      "loss": 0.1775,
      "reward": 0.6415899395942688,
      "reward_std": 0.9059087038040161,
      "rewards/": 3.493664026260376,
      "rewards/math_compute_score": -0.0714285746216774,
      "step": 13
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1072.7857666015625,
      "epoch": 0.03341288782816229,
      "grad_norm": 0.31831495486107775,
      "kl": 4.267692565917969e-05,
      "learning_rate": 9.99985031250522e-07,
      "loss": 0.3527,
      "reward": 0.7675060033798218,
      "reward_std": 1.03658926486969,
      "rewards/": 3.170863628387451,
      "rewards/math_compute_score": 0.1666666716337204,
      "step": 14
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1117.40478515625,
      "epoch": 0.03579952267303103,
      "grad_norm": 0.2987798632353295,
      "kl": 4.220008850097656e-05,
      "learning_rate": 9.999401258983425e-07,
      "loss": 0.1567,
      "reward": 0.8110119104385376,
      "reward_std": 0.8140324950218201,
      "rewards/": 3.293154716491699,
      "rewards/math_compute_score": 0.190476194024086,
      "step": 15
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1116.9285888671875,
      "epoch": 0.03818615751789976,
      "grad_norm": 0.2575689090925313,
      "kl": 4.506111145019531e-05,
      "learning_rate": 9.998652866321687e-07,
      "loss": 0.1441,
      "reward": 0.39009490609169006,
      "reward_std": 0.8612415194511414,
      "rewards/": 2.5219030380249023,
      "rewards/math_compute_score": -0.1428571492433548,
      "step": 16
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 905.047607421875,
      "epoch": 0.0405727923627685,
      "grad_norm": 0.3749791786039388,
      "kl": 6.341934204101562e-05,
      "learning_rate": 9.997605179330017e-07,
      "loss": 0.2651,
      "reward": 0.7404214143753052,
      "reward_std": 0.6791912317276001,
      "rewards/": 3.035440444946289,
      "rewards/math_compute_score": 0.1666666716337204,
      "step": 17
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 824.1666870117188,
      "epoch": 0.04295942720763723,
      "grad_norm": 0.32853762246024193,
      "kl": 2.7060508728027344e-05,
      "learning_rate": 9.996258260738674e-07,
      "loss": 0.3643,
      "reward": 0.9851795434951782,
      "reward_std": 0.7530649304389954,
      "rewards/": 3.592564344406128,
      "rewards/math_compute_score": 0.3333333432674408,
      "step": 18
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1304.6190185546875,
      "epoch": 0.045346062052505964,
      "grad_norm": 0.21651829402479825,
      "kl": 5.269050598144531e-05,
      "learning_rate": 9.994612191194405e-07,
      "loss": 0.0839,
      "reward": 0.5019060373306274,
      "reward_std": 1.0315256118774414,
      "rewards/": 2.7952444553375244,
      "rewards/math_compute_score": -0.0714285746216774,
      "step": 19
    },
    {
      "epoch": 0.0477326968973747,
      "grad_norm": 0.26887161838116597,
      "learning_rate": 9.992667069255618e-07,
      "loss": 0.1969,
      "step": 20
    },
    {
      "epoch": 0.0477326968973747,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 1136.1309814453125,
      "eval_kl": 5.7637691497802734e-05,
      "eval_loss": 0.25490644574165344,
      "eval_reward": 0.8720096871256828,
      "eval_reward_std": 0.9841141104698181,
      "eval_rewards/": 3.83623868227005,
      "eval_rewards/math_compute_score": 0.13095238618552685,
      "eval_runtime": 82.545,
      "eval_samples_per_second": 0.254,
      "eval_steps_per_second": 0.012,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 981.9642639160156,
      "epoch": 0.050119331742243436,
      "grad_norm": 0.34449148369838256,
      "kl": 0.0001131296157836914,
      "learning_rate": 9.990423011386488e-07,
      "loss": 0.2118,
      "reward": 0.9287342429161072,
      "reward_std": 0.7616195380687714,
      "rewards/": 3.977004289627075,
      "rewards/math_compute_score": 0.16666667070239782,
      "step": 21
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 877.3095703125,
      "epoch": 0.05250596658711217,
      "grad_norm": 0.40398884988921724,
      "kl": 0.000102996826171875,
      "learning_rate": 9.987880151949975e-07,
      "loss": 0.1459,
      "reward": 1.012574553489685,
      "reward_std": 0.7148696780204773,
      "rewards/": 4.300967216491699,
      "rewards/math_compute_score": 0.190476194024086,
      "step": 22
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1151.3333740234375,
      "epoch": 0.05489260143198091,
      "grad_norm": 0.3204894614801331,
      "kl": 0.00017547607421875,
      "learning_rate": 9.985038643199778e-07,
      "loss": 0.2659,
      "reward": 0.4764881134033203,
      "reward_std": 0.9507336616516113,
      "rewards/": 2.763392925262451,
      "rewards/math_compute_score": -0.095238097012043,
      "step": 23
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1214.59521484375,
      "epoch": 0.057279236276849645,
      "grad_norm": 0.2915535040877492,
      "kl": 9.870529174804688e-05,
      "learning_rate": 9.981898655271234e-07,
      "loss": 0.2398,
      "reward": 0.5920386910438538,
      "reward_std": 0.9118747115135193,
      "rewards/": 3.5316221714019775,
      "rewards/math_compute_score": -0.1428571492433548,
      "step": 24
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1111.761962890625,
      "epoch": 0.059665871121718374,
      "grad_norm": 0.3966341374575089,
      "kl": 0.00014209747314453125,
      "learning_rate": 9.978460376171112e-07,
      "loss": 0.074,
      "reward": 0.5987061262130737,
      "reward_std": 0.8957276344299316,
      "rewards/": 2.6125779151916504,
      "rewards/math_compute_score": 0.095238097012043,
      "step": 25
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 937.5238037109375,
      "epoch": 0.06205250596658711,
      "grad_norm": 0.24086304315962492,
      "kl": 9.107589721679688e-05,
      "learning_rate": 9.974724011766361e-07,
      "loss": 0.1505,
      "reward": 1.1114305257797241,
      "reward_std": 0.7139089107513428,
      "rewards/": 3.8428664207458496,
      "rewards/math_compute_score": 0.4285714328289032,
      "step": 26
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 806.90478515625,
      "epoch": 0.06443914081145585,
      "grad_norm": 0.36587422053686375,
      "kl": 0.00021839141845703125,
      "learning_rate": 9.970689785771798e-07,
      "loss": 0.335,
      "reward": 1.220070719718933,
      "reward_std": 0.8377094268798828,
      "rewards/": 4.862258434295654,
      "rewards/math_compute_score": 0.3095238208770752,
      "step": 27
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 840.7857055664062,
      "epoch": 0.06682577565632458,
      "grad_norm": 1.1836785865916923,
      "kl": 0.00023555755615234375,
      "learning_rate": 9.96635793973669e-07,
      "loss": 0.3326,
      "reward": 1.1378720998764038,
      "reward_std": 0.8822279572486877,
      "rewards/": 4.356026649475098,
      "rewards/math_compute_score": 0.3333333432674408,
      "step": 28
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 894.5952758789062,
      "epoch": 0.06921241050119331,
      "grad_norm": 0.2976095645322702,
      "kl": 0.0002040863037109375,
      "learning_rate": 9.961728733030316e-07,
      "loss": 0.1709,
      "reward": 0.7952183485031128,
      "reward_std": 0.8282038569450378,
      "rewards/": 3.0237107276916504,
      "rewards/math_compute_score": 0.2380952388048172,
      "step": 29
    },
    {
      "epoch": 0.07159904534606205,
      "grad_norm": 0.31922105827783287,
      "learning_rate": 9.956802442826415e-07,
      "loss": 0.1756,
      "step": 30
    },
    {
      "epoch": 0.07159904534606205,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 1086.5595245361328,
      "eval_kl": 0.00028777122497558594,
      "eval_loss": 0.11829498410224915,
      "eval_reward": 0.9524402394890785,
      "eval_reward_std": 0.8144673109054565,
      "eval_rewards/": 3.9764870405197144,
      "eval_rewards/math_compute_score": 0.1964285746216774,
      "eval_runtime": 83.0591,
      "eval_samples_per_second": 0.253,
      "eval_steps_per_second": 0.012,
      "step": 30
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 918.5238037109375,
      "epoch": 0.07398568019093078,
      "grad_norm": 0.3688063540106728,
      "kl": 0.00035762786865234375,
      "learning_rate": 9.951579364086603e-07,
      "loss": 0.3767,
      "reward": 0.8680548071861267,
      "reward_std": 0.8301426470279694,
      "rewards/": 3.2926549911499023,
      "rewards/math_compute_score": 0.2619047649204731,
      "step": 31
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1212.3809814453125,
      "epoch": 0.07637231503579953,
      "grad_norm": 0.2490900765535506,
      "kl": 0.0002994537353515625,
      "learning_rate": 9.946059809542706e-07,
      "loss": 0.1623,
      "reward": 0.7525901794433594,
      "reward_std": 0.8568804860115051,
      "rewards/": 3.477236747741699,
      "rewards/math_compute_score": 0.0714285746216774,
      "step": 32
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 725.40478515625,
      "epoch": 0.07875894988066826,
      "grad_norm": 0.39900214613105317,
      "kl": 0.000522613525390625,
      "learning_rate": 9.940244109678041e-07,
      "loss": 0.0071,
      "reward": 1.2588740587234497,
      "reward_std": 0.6066949367523193,
      "rewards/": 5.0562744140625,
      "rewards/math_compute_score": 0.3095238208770752,
      "step": 33
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 698.9761962890625,
      "epoch": 0.081145584725537,
      "grad_norm": 0.3908939527833547,
      "kl": 0.000396728515625,
      "learning_rate": 9.93413261270763e-07,
      "loss": 0.268,
      "reward": 1.276125431060791,
      "reward_std": 0.6631749868392944,
      "rewards/": 3.8091983795166016,
      "rewards/math_compute_score": 0.6428571343421936,
      "step": 34
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 908.0238037109375,
      "epoch": 0.08353221957040573,
      "grad_norm": 0.3377115570250511,
      "kl": 0.000377655029296875,
      "learning_rate": 9.927725684557339e-07,
      "loss": 0.1234,
      "reward": 1.2422432899475098,
      "reward_std": 0.6443389654159546,
      "rewards/": 5.163597583770752,
      "rewards/math_compute_score": 0.261904776096344,
      "step": 35
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1112.8095703125,
      "epoch": 0.08591885441527446,
      "grad_norm": 0.23706974723352947,
      "kl": 0.000385284423828125,
      "learning_rate": 9.921023708841973e-07,
      "loss": -0.1353,
      "reward": 0.601748526096344,
      "reward_std": 0.5709776282310486,
      "rewards/": 3.294456958770752,
      "rewards/math_compute_score": -0.0714285746216774,
      "step": 36
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 867.7619018554688,
      "epoch": 0.0883054892601432,
      "grad_norm": 0.3434544369292518,
      "kl": 0.00067138671875,
      "learning_rate": 9.914027086842322e-07,
      "loss": 0.0705,
      "reward": 1.0254464149475098,
      "reward_std": 1.1037424802780151,
      "rewards/": 3.5081846714019775,
      "rewards/math_compute_score": 0.4047619104385376,
      "step": 37
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 792.3095092773438,
      "epoch": 0.09069212410501193,
      "grad_norm": 0.3791096432652416,
      "kl": 0.000576019287109375,
      "learning_rate": 9.906736237481108e-07,
      "loss": 0.1116,
      "reward": 1.2115607261657715,
      "reward_std": 0.8592520356178284,
      "rewards/": 4.153041362762451,
      "rewards/math_compute_score": 0.4761904776096344,
      "step": 38
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 772.5714721679688,
      "epoch": 0.09307875894988067,
      "grad_norm": 0.3544290113443038,
      "kl": 0.000850677490234375,
      "learning_rate": 9.899151597297922e-07,
      "loss": 0.3191,
      "reward": 1.139400839805603,
      "reward_std": 0.6882633566856384,
      "rewards/": 3.9827184677124023,
      "rewards/math_compute_score": 0.4285714328289032,
      "step": 39
    },
    {
      "epoch": 0.0954653937947494,
      "grad_norm": 0.42043440113349795,
      "learning_rate": 9.891273620423082e-07,
      "loss": 0.1347,
      "step": 40
    },
    {
      "epoch": 0.0954653937947494,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 1038.15478515625,
      "eval_kl": 0.0018901824951171875,
      "eval_loss": 0.12366585433483124,
      "eval_reward": 1.030681535601616,
      "eval_reward_std": 0.8943771868944168,
      "eval_rewards/": 4.439121901988983,
      "eval_rewards/math_compute_score": 0.17857143329456449,
      "eval_runtime": 82.5285,
      "eval_samples_per_second": 0.254,
      "eval_steps_per_second": 0.012,
      "step": 40
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 908.3333435058594,
      "epoch": 0.09785202863961814,
      "grad_norm": 0.25194007930707885,
      "kl": 0.0010833740234375,
      "learning_rate": 9.883102778550434e-07,
      "loss": 0.2062,
      "reward": 1.3001837134361267,
      "reward_std": 0.7053222358226776,
      "rewards/": 4.977108955383301,
      "rewards/math_compute_score": 0.380952388048172,
      "step": 41
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 900.2619018554688,
      "epoch": 0.10023866348448687,
      "grad_norm": 0.32787405490701843,
      "kl": 0.0010986328125,
      "learning_rate": 9.874639560909118e-07,
      "loss": 0.0606,
      "reward": 1.039421796798706,
      "reward_std": 0.8191918730735779,
      "rewards/": 3.197108745574951,
      "rewards/math_compute_score": 0.5,
      "step": 42
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 917.90478515625,
      "epoch": 0.1026252983293556,
      "grad_norm": 0.5163718395183062,
      "kl": 0.00128936767578125,
      "learning_rate": 9.865884474234275e-07,
      "loss": 0.2818,
      "reward": 1.222414493560791,
      "reward_std": 0.7349604368209839,
      "rewards/": 4.493024826049805,
      "rewards/math_compute_score": 0.4047619104385376,
      "step": 43
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 690.2857055664062,
      "epoch": 0.10501193317422435,
      "grad_norm": 0.3207693941377101,
      "kl": 0.0008087158203125,
      "learning_rate": 9.856838042736696e-07,
      "loss": 0.1025,
      "reward": 1.2743303775787354,
      "reward_std": 0.7797837257385254,
      "rewards/": 4.46688985824585,
      "rewards/math_compute_score": 0.4761904776096344,
      "step": 44
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 676.0714111328125,
      "epoch": 0.10739856801909307,
      "grad_norm": 0.28894735083744694,
      "kl": 0.0009918212890625,
      "learning_rate": 9.847500808071456e-07,
      "loss": 0.1353,
      "reward": 1.2259044647216797,
      "reward_std": 0.7126508355140686,
      "rewards/": 3.558094024658203,
      "rewards/math_compute_score": 0.6428571343421936,
      "step": 45
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 775.40478515625,
      "epoch": 0.10978520286396182,
      "grad_norm": 0.27186613080235555,
      "kl": 0.000629425048828125,
      "learning_rate": 9.837873329305457e-07,
      "loss": 0.1791,
      "reward": 1.423451542854309,
      "reward_std": 0.6115496754646301,
      "rewards/": 4.545828819274902,
      "rewards/math_compute_score": 0.6428571343421936,
      "step": 46
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1084.90478515625,
      "epoch": 0.11217183770883055,
      "grad_norm": 0.2663709639869927,
      "kl": 0.001007080078125,
      "learning_rate": 9.82795618288397e-07,
      "loss": 0.1393,
      "reward": 0.9681082367897034,
      "reward_std": 0.7282014489173889,
      "rewards/": 4.173874855041504,
      "rewards/math_compute_score": 0.1666666716337204,
      "step": 47
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 844.6428833007812,
      "epoch": 0.11455847255369929,
      "grad_norm": 0.30223981125620725,
      "kl": 0.00147247314453125,
      "learning_rate": 9.817749962596114e-07,
      "loss": 0.18,
      "reward": 1.0321242809295654,
      "reward_std": 0.8400474190711975,
      "rewards/": 4.208240509033203,
      "rewards/math_compute_score": 0.2380952388048172,
      "step": 48
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 560.5238037109375,
      "epoch": 0.11694510739856802,
      "grad_norm": 0.4242899512098847,
      "kl": 0.00177764892578125,
      "learning_rate": 9.807255279539312e-07,
      "loss": 0.1896,
      "reward": 1.313750982284546,
      "reward_std": 0.46607664227485657,
      "rewards/": 4.283040523529053,
      "rewards/math_compute_score": 0.5714285969734192,
      "step": 49
    },
    {
      "epoch": 0.11933174224343675,
      "grad_norm": 0.348729238348003,
      "learning_rate": 9.796472762082685e-07,
      "loss": 0.2283,
      "step": 50
    },
    {
      "epoch": 0.11933174224343675,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 933.5654907226562,
      "eval_kl": 0.0038604736328125,
      "eval_loss": 0.1660795509815216,
      "eval_reward": 1.0914957970380783,
      "eval_reward_std": 0.7910807579755783,
      "eval_rewards/": 4.457478940486908,
      "eval_rewards/math_compute_score": 0.25000000558793545,
      "eval_runtime": 81.0284,
      "eval_samples_per_second": 0.259,
      "eval_steps_per_second": 0.012,
      "step": 50
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 785.5595397949219,
      "epoch": 0.12171837708830549,
      "grad_norm": 0.27596087991963414,
      "kl": 0.001678466796875,
      "learning_rate": 9.785403055829448e-07,
      "loss": -0.001,
      "reward": 1.2578682899475098,
      "reward_std": 0.7686284184455872,
      "rewards/": 4.384579658508301,
      "rewards/math_compute_score": 0.4761904776096344,
      "step": 51
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 916.0714721679688,
      "epoch": 0.12410501193317422,
      "grad_norm": 0.2741628390919074,
      "kl": 0.00164794921875,
      "learning_rate": 9.77404682357824e-07,
      "loss": 0.1083,
      "reward": 1.1715024709701538,
      "reward_std": 0.6183907985687256,
      "rewards/": 4.333702564239502,
      "rewards/math_compute_score": 0.380952388048172,
      "step": 52
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 819.5714721679688,
      "epoch": 0.12649164677804295,
      "grad_norm": 0.35526991630751986,
      "kl": 0.0020904541015625,
      "learning_rate": 9.762404745283437e-07,
      "loss": 0.2932,
      "reward": 1.2613282203674316,
      "reward_std": 0.5622918605804443,
      "rewards/": 4.497117042541504,
      "rewards/math_compute_score": 0.4523809552192688,
      "step": 53
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 878.7857055664062,
      "epoch": 0.1288782816229117,
      "grad_norm": 0.5981296635875077,
      "kl": 0.00194549560546875,
      "learning_rate": 9.75047751801446e-07,
      "loss": 0.2455,
      "reward": 1.058624029159546,
      "reward_std": 0.8494831323623657,
      "rewards/": 4.2455010414123535,
      "rewards/math_compute_score": 0.261904776096344,
      "step": 54
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 753.7380981445312,
      "epoch": 0.13126491646778043,
      "grad_norm": 0.2280494532860685,
      "kl": 0.000865936279296875,
      "learning_rate": 9.738265855914012e-07,
      "loss": -0.0038,
      "reward": 1.4897369146347046,
      "reward_std": 0.6451947689056396,
      "rewards/": 4.972493648529053,
      "rewards/math_compute_score": 0.6190476417541504,
      "step": 55
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 732.0,
      "epoch": 0.13365155131264916,
      "grad_norm": 0.3162020489120827,
      "kl": 0.0020599365234375,
      "learning_rate": 9.725770490155338e-07,
      "loss": 0.2325,
      "reward": 1.3167736530303955,
      "reward_std": 0.5749982595443726,
      "rewards/": 4.393392086029053,
      "rewards/math_compute_score": 0.5476190447807312,
      "step": 56
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 780.0238037109375,
      "epoch": 0.1360381861575179,
      "grad_norm": 0.5756772020485215,
      "kl": 0.001495361328125,
      "learning_rate": 9.712992168898435e-07,
      "loss": 0.1519,
      "reward": 1.3120629787445068,
      "reward_std": 0.6641415953636169,
      "rewards/": 5.226981163024902,
      "rewards/math_compute_score": 0.3333333432674408,
      "step": 57
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 556.5,
      "epoch": 0.13842482100238662,
      "grad_norm": 0.4844983583796039,
      "kl": 0.0034942626953125,
      "learning_rate": 9.699931657245263e-07,
      "loss": -0.0877,
      "reward": 1.2563197612762451,
      "reward_std": 0.745265543460846,
      "rewards/": 4.376836776733398,
      "rewards/math_compute_score": 0.4761904776096344,
      "step": 58
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 635.2380981445312,
      "epoch": 0.14081145584725538,
      "grad_norm": 0.4758034742756555,
      "kl": 0.0026397705078125,
      "learning_rate": 9.686589737193928e-07,
      "loss": 0.053,
      "reward": 1.0557314157485962,
      "reward_std": 1.0078890323638916,
      "rewards/": 3.754848003387451,
      "rewards/math_compute_score": 0.380952388048172,
      "step": 59
    },
    {
      "epoch": 0.1431980906921241,
      "grad_norm": 0.3072937144155207,
      "learning_rate": 9.67296720759187e-07,
      "loss": 0.0474,
      "step": 60
    },
    {
      "epoch": 0.1431980906921241,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 837.0000152587891,
      "eval_kl": 0.013916015625,
      "eval_loss": 0.11870573461055756,
      "eval_reward": 1.0395183265209198,
      "eval_reward_std": 0.7884587794542313,
      "eval_rewards/": 4.197591245174408,
      "eval_rewards/math_compute_score": 0.2500000111758709,
      "eval_runtime": 79.2917,
      "eval_samples_per_second": 0.265,
      "eval_steps_per_second": 0.013,
      "step": 60
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 674.9404907226562,
      "epoch": 0.14558472553699284,
      "grad_norm": 0.4086529116673979,
      "kl": 0.00350189208984375,
      "learning_rate": 9.659064884088016e-07,
      "loss": 0.1354,
      "reward": 1.1227608919143677,
      "reward_std": 0.6351146399974823,
      "rewards/": 4.042375802993774,
      "rewards/math_compute_score": 0.3928571492433548,
      "step": 61
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 730.952392578125,
      "epoch": 0.14797136038186157,
      "grad_norm": 0.33492026311818307,
      "kl": 0.0020599365234375,
      "learning_rate": 9.644883599083957e-07,
      "loss": -0.0413,
      "reward": 1.2523438930511475,
      "reward_std": 0.5973415374755859,
      "rewards/": 5.023623466491699,
      "rewards/math_compute_score": 0.3095238208770752,
      "step": 62
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 591.1190795898438,
      "epoch": 0.15035799522673032,
      "grad_norm": 0.30746445342498774,
      "kl": 0.00110626220703125,
      "learning_rate": 9.630424201684103e-07,
      "loss": -0.024,
      "reward": 1.3001348972320557,
      "reward_std": 0.5841225385665894,
      "rewards/": 4.4054365158081055,
      "rewards/math_compute_score": 0.523809552192688,
      "step": 63
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 739.0714721679688,
      "epoch": 0.15274463007159905,
      "grad_norm": 0.33486655671684584,
      "kl": 0.0015411376953125,
      "learning_rate": 9.615687557644848e-07,
      "loss": 0.1172,
      "reward": 1.4695312976837158,
      "reward_std": 0.6374803185462952,
      "rewards/": 4.871465682983398,
      "rewards/math_compute_score": 0.6190476417541504,
      "step": 64
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 820.4285888671875,
      "epoch": 0.15513126491646778,
      "grad_norm": 0.7636558012245415,
      "kl": 0.0037994384765625,
      "learning_rate": 9.600674549322716e-07,
      "loss": 0.2518,
      "reward": 1.0687873363494873,
      "reward_std": 0.8015285730361938,
      "rewards/": 3.534412384033203,
      "rewards/math_compute_score": 0.4523809552192688,
      "step": 65
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 991.5238037109375,
      "epoch": 0.1575178997613365,
      "grad_norm": 0.24283471181996055,
      "kl": 0.00176239013671875,
      "learning_rate": 9.585386075621552e-07,
      "loss": 0.1276,
      "reward": 1.4505395889282227,
      "reward_std": 0.3715428113937378,
      "rewards/": 5.157459259033203,
      "rewards/math_compute_score": 0.523809552192688,
      "step": 66
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 666.2857055664062,
      "epoch": 0.15990453460620524,
      "grad_norm": 0.32163982732892316,
      "kl": 0.003021240234375,
      "learning_rate": 9.569823051938689e-07,
      "loss": 0.3253,
      "reward": 1.4773437976837158,
      "reward_std": 0.6230893731117249,
      "rewards/": 4.339099884033203,
      "rewards/math_compute_score": 0.761904776096344,
      "step": 67
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 603.547607421875,
      "epoch": 0.162291169451074,
      "grad_norm": 0.5401807083101557,
      "kl": 0.00982666015625,
      "learning_rate": 9.553986410110134e-07,
      "loss": 0.4025,
      "reward": 0.9949312210083008,
      "reward_std": 0.4883081316947937,
      "rewards/": 3.641322612762451,
      "rewards/math_compute_score": 0.3333333432674408,
      "step": 68
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 857.6190795898438,
      "epoch": 0.16467780429594273,
      "grad_norm": 0.2558130935281063,
      "kl": 0.00057220458984375,
      "learning_rate": 9.537877098354784e-07,
      "loss": -0.0058,
      "reward": 1.5268882513046265,
      "reward_std": 0.5736187696456909,
      "rewards/": 5.34872579574585,
      "rewards/math_compute_score": 0.5714285969734192,
      "step": 69
    },
    {
      "epoch": 0.16706443914081145,
      "grad_norm": 0.27521723291274197,
      "learning_rate": 9.52149608121765e-07,
      "loss": 0.0567,
      "step": 70
    },
    {
      "epoch": 0.16706443914081145,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 882.9702453613281,
      "eval_kl": 0.03376007080078125,
      "eval_loss": 0.26883435249328613,
      "eval_reward": 1.151841551065445,
      "eval_reward_std": 0.8478920459747314,
      "eval_rewards/": 4.473493456840515,
      "eval_rewards/math_compute_score": 0.3214285783469677,
      "eval_runtime": 78.4604,
      "eval_samples_per_second": 0.268,
      "eval_steps_per_second": 0.013,
      "step": 70
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 712.1309509277344,
      "epoch": 0.16945107398568018,
      "grad_norm": 0.286488510896718,
      "kl": 0.00177001953125,
      "learning_rate": 9.504844339512094e-07,
      "loss": 0.2365,
      "reward": 1.5305490493774414,
      "reward_std": 0.5794989466667175,
      "rewards/": 5.12893533706665,
      "rewards/math_compute_score": 0.6309524029493332,
      "step": 71
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 807.7857055664062,
      "epoch": 0.1718377088305489,
      "grad_norm": 0.3110498028785089,
      "kl": 0.0013275146484375,
      "learning_rate": 9.487922870261121e-07,
      "loss": -0.0117,
      "reward": 1.6091006994247437,
      "reward_std": 0.5738370418548584,
      "rewards/": 5.188360214233398,
      "rewards/math_compute_score": 0.7142857313156128,
      "step": 72
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 771.6666870117188,
      "epoch": 0.17422434367541767,
      "grad_norm": 0.28968592215813294,
      "kl": 0.0019989013671875,
      "learning_rate": 9.470732686637664e-07,
      "loss": 0.0653,
      "reward": 1.406017541885376,
      "reward_std": 0.7840931415557861,
      "rewards/": 5.220563888549805,
      "rewards/math_compute_score": 0.4523809552192688,
      "step": 73
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 776.0238037109375,
      "epoch": 0.1766109785202864,
      "grad_norm": 0.31133583857342856,
      "kl": 0.0069580078125,
      "learning_rate": 9.45327481790393e-07,
      "loss": -0.0325,
      "reward": 1.483282208442688,
      "reward_std": 0.7019442915916443,
      "rewards/": 5.321172714233398,
      "rewards/math_compute_score": 0.523809552192688,
      "step": 74
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 764.5,
      "epoch": 0.17899761336515513,
      "grad_norm": 0.36739420438573467,
      "kl": 0.004913330078125,
      "learning_rate": 9.435550309349776e-07,
      "loss": -0.025,
      "reward": 1.2999628782272339,
      "reward_std": 0.7012795805931091,
      "rewards/": 5.26171875,
      "rewards/math_compute_score": 0.3095238208770752,
      "step": 75
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 666.9761962890625,
      "epoch": 0.18138424821002386,
      "grad_norm": 0.33428186789934067,
      "kl": 0.0022125244140625,
      "learning_rate": 9.417560222230114e-07,
      "loss": 0.0816,
      "reward": 1.252938985824585,
      "reward_std": 0.6224194765090942,
      "rewards/": 4.07421875,
      "rewards/math_compute_score": 0.5476190447807312,
      "step": 76
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 703.4285888671875,
      "epoch": 0.18377088305489261,
      "grad_norm": 0.3021119658752094,
      "kl": 0.0036773681640625,
      "learning_rate": 9.399305633701372e-07,
      "loss": 0.0133,
      "reward": 1.5353423357009888,
      "reward_std": 0.5898163318634033,
      "rewards/": 5.390996932983398,
      "rewards/math_compute_score": 0.5714285969734192,
      "step": 77
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 743.4761962890625,
      "epoch": 0.18615751789976134,
      "grad_norm": 0.4512916551501506,
      "kl": 0.003326416015625,
      "learning_rate": 9.380787636757e-07,
      "loss": 0.3175,
      "reward": 1.4912109375,
      "reward_std": 0.6637881994247437,
      "rewards/": 5.360816478729248,
      "rewards/math_compute_score": 0.523809552192688,
      "step": 78
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 760.452392578125,
      "epoch": 0.18854415274463007,
      "grad_norm": 0.3498194905087342,
      "kl": 0.00142669677734375,
      "learning_rate": 9.362007340162028e-07,
      "loss": 0.0411,
      "reward": 1.5506510734558105,
      "reward_std": 0.3708002269268036,
      "rewards/": 5.277064800262451,
      "rewards/math_compute_score": 0.6190476417541504,
      "step": 79
    },
    {
      "epoch": 0.1909307875894988,
      "grad_norm": 0.23488028047184029,
      "learning_rate": 9.342965868386673e-07,
      "loss": -0.0472,
      "step": 80
    },
    {
      "epoch": 0.1909307875894988,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 915.2738342285156,
      "eval_kl": 0.02166748046875,
      "eval_loss": 0.17802654206752777,
      "eval_reward": 1.2084915190935135,
      "eval_reward_std": 0.7365763932466507,
      "eval_rewards/": 4.542457342147827,
      "eval_rewards/math_compute_score": 0.3750000111758709,
      "eval_runtime": 78.6127,
      "eval_samples_per_second": 0.267,
      "eval_steps_per_second": 0.013,
      "step": 80
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 419,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 40,
  "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": 14,
  "trial_name": null,
  "trial_params": null
}