{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.49588954060483026,
  "eval_steps": 500,
  "global_step": 1000,
  "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": 0.0535714285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14410.0,
      "completions/mean_length": 1952.3037109375,
      "completions/mean_terminated_length": 1193.3773193359375,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0004958895406048302,
      "grad_norm": 1.7557120323181152,
      "learning_rate": 0.0,
      "loss": 0.0,
      "num_tokens": 942024.0,
      "reward": 0.299107164144516,
      "reward_std": 0.3110422194004059,
      "rewards/accuracy_reward/mean": 0.2991071343421936,
      "rewards/accuracy_reward/std": 0.45837873220443726,
      "step": 1
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0009917790812096604,
      "grad_norm": 1.7560921907424927,
      "learning_rate": 1e-08,
      "loss": 0.0,
      "step": 2
    },
    {
      "clip_ratio/high_max": 0.011167433188802534,
      "clip_ratio/high_mean": 0.0024931100216463165,
      "clip_ratio/low_mean": 0.00226416202940527,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004757272006827407,
      "epoch": 0.0014876686218144909,
      "grad_norm": 1.9170538187026978,
      "learning_rate": 2e-08,
      "loss": 0.0005,
      "step": 3
    },
    {
      "clip_ratio/high_max": 0.010536649010646215,
      "clip_ratio/high_mean": 0.0023931153812100092,
      "clip_ratio/low_mean": 0.0026474937169780333,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005040609064963064,
      "epoch": 0.001983558162419321,
      "grad_norm": 1.7436535358428955,
      "learning_rate": 3e-08,
      "loss": 0.0006,
      "step": 4
    },
    {
      "clip_ratio/high_max": 0.010404693607597437,
      "clip_ratio/high_mean": 0.0023042125537813263,
      "clip_ratio/low_mean": 0.0027232365831650895,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0050274491140953614,
      "epoch": 0.0024794477030241516,
      "grad_norm": 1.773264765739441,
      "learning_rate": 4e-08,
      "loss": 0.0006,
      "step": 5
    },
    {
      "clip_ratio/high_max": 0.011299804794816737,
      "clip_ratio/high_mean": 0.002474455901960937,
      "clip_ratio/low_mean": 0.002619469375076733,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005093925261462573,
      "epoch": 0.0029753372436289818,
      "grad_norm": 1.589799165725708,
      "learning_rate": 5e-08,
      "loss": 0.0005,
      "step": 6
    },
    {
      "clip_ratio/high_max": 0.009685181392342201,
      "clip_ratio/high_mean": 0.0022252411467889033,
      "clip_ratio/low_mean": 0.002265281937980035,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004490523052481876,
      "epoch": 0.003471226784233812,
      "grad_norm": 1.8007147312164307,
      "learning_rate": 6e-08,
      "loss": 0.0007,
      "step": 7
    },
    {
      "clip_ratio/high_max": 0.010020916586654494,
      "clip_ratio/high_mean": 0.0022770342420699308,
      "clip_ratio/low_mean": 0.002512929635201999,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004789963872099179,
      "epoch": 0.003967116324838642,
      "grad_norm": 1.7602940797805786,
      "learning_rate": 7e-08,
      "loss": 0.0005,
      "step": 8
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0825892857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13135.0,
      "completions/mean_length": 2278.692138671875,
      "completions/mean_terminated_length": 1101.055908203125,
      "completions/min_length": 6.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.004463005865443473,
      "grad_norm": 1.5323379039764404,
      "learning_rate": 8e-08,
      "loss": -0.0,
      "num_tokens": 2031942.0,
      "reward": 0.2901785969734192,
      "reward_std": 0.3322319984436035,
      "rewards/accuracy_reward/mean": 0.2901785671710968,
      "rewards/accuracy_reward/std": 0.4543520212173462,
      "step": 9
    },
    {
      "clip_ratio/high_max": 0.009046431141541689,
      "clip_ratio/high_mean": 0.00213518183522865,
      "clip_ratio/low_mean": 0.002453601546221762,
      "clip_ratio/low_min": 3.051757857974735e-06,
      "clip_ratio/region_mean": 0.004588783358485671,
      "epoch": 0.004958895406048303,
      "grad_norm": 1.6876938343048096,
      "learning_rate": 9e-08,
      "loss": 0.0007,
      "step": 10
    },
    {
      "clip_ratio/high_max": 0.009184943159198156,
      "clip_ratio/high_mean": 0.002262502872895311,
      "clip_ratio/low_mean": 0.00289479756156652,
      "clip_ratio/low_min": 3.051757857974735e-06,
      "clip_ratio/region_mean": 0.005157300442078849,
      "epoch": 0.005454784946653133,
      "grad_norm": 1.8953957557678223,
      "learning_rate": 1e-07,
      "loss": 0.0009,
      "step": 11
    },
    {
      "clip_ratio/high_max": 0.0087673398120387,
      "clip_ratio/high_mean": 0.002209051125419137,
      "clip_ratio/low_mean": 0.002892916354994668,
      "clip_ratio/low_min": 1.0172526572205243e-06,
      "clip_ratio/region_mean": 0.0051019674701819895,
      "epoch": 0.0059506744872579636,
      "grad_norm": 1.6153126955032349,
      "learning_rate": 1.0999999999999999e-07,
      "loss": 0.0006,
      "step": 12
    },
    {
      "clip_ratio/high_max": 0.008912842044082936,
      "clip_ratio/high_mean": 0.002234245628756071,
      "clip_ratio/low_mean": 0.002632593334055855,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004866838977250154,
      "epoch": 0.006446564027862794,
      "grad_norm": 2.0074634552001953,
      "learning_rate": 1.2e-07,
      "loss": 0.0009,
      "step": 13
    },
    {
      "clip_ratio/high_max": 0.009310546602137038,
      "clip_ratio/high_mean": 0.0022803476740591577,
      "clip_ratio/low_mean": 0.0023988454333903064,
      "clip_ratio/low_min": 5.086263172415784e-06,
      "clip_ratio/region_mean": 0.00467919313814491,
      "epoch": 0.006942453568467624,
      "grad_norm": 1.883607268333435,
      "learning_rate": 1.3e-07,
      "loss": 0.0007,
      "step": 14
    },
    {
      "clip_ratio/high_max": 0.009018985438160598,
      "clip_ratio/high_mean": 0.002211305206628822,
      "clip_ratio/low_mean": 0.003138838485483575,
      "clip_ratio/low_min": 1.0172526572205243e-06,
      "clip_ratio/region_mean": 0.005350143705072696,
      "epoch": 0.007438343109072454,
      "grad_norm": 1.8985334634780884,
      "learning_rate": 1.4e-07,
      "loss": 0.0008,
      "step": 15
    },
    {
      "clip_ratio/high_max": 0.009544631961944106,
      "clip_ratio/high_mean": 0.002238498590031668,
      "clip_ratio/low_mean": 0.0028333873611359195,
      "clip_ratio/low_min": 5.086263172415784e-06,
      "clip_ratio/region_mean": 0.0050718859492917545,
      "epoch": 0.007934232649677284,
      "grad_norm": 1.4888081550598145,
      "learning_rate": 1.5e-07,
      "loss": 0.0004,
      "step": 16
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0535714285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 11928.0,
      "completions/mean_length": 1747.243408203125,
      "completions/mean_terminated_length": 976.7098999023438,
      "completions/min_length": 13.0,
      "completions/min_terminated_length": 13.0,
      "epoch": 0.008430122190282115,
      "grad_norm": 0.28484830260276794,
      "learning_rate": 1.6e-07,
      "loss": -0.0,
      "num_tokens": 2882947.0,
      "reward": 0.314732164144516,
      "reward_std": 0.2812138497829437,
      "rewards/accuracy_reward/mean": 0.3147321343421936,
      "rewards/accuracy_reward/std": 0.4649282991886139,
      "step": 17
    },
    {
      "clip_ratio/high_max": 0.009855566173428087,
      "clip_ratio/high_mean": 0.002131032052830051,
      "clip_ratio/low_mean": 0.0021255130386634846,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004256545126736455,
      "epoch": 0.008926011730886946,
      "grad_norm": 0.3203405439853668,
      "learning_rate": 1.7000000000000001e-07,
      "loss": 0.0007,
      "step": 18
    },
    {
      "clip_ratio/high_max": 0.00998267206341552,
      "clip_ratio/high_mean": 0.0022234960342757404,
      "clip_ratio/low_mean": 0.002105405340444122,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0043289013528919895,
      "epoch": 0.009421901271491775,
      "grad_norm": 0.2954414188861847,
      "learning_rate": 1.8e-07,
      "loss": 0.0006,
      "step": 19
    },
    {
      "clip_ratio/high_max": 0.009869431212791824,
      "clip_ratio/high_mean": 0.0022057523510738974,
      "clip_ratio/low_mean": 0.00197554391866106,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00418129630452313,
      "epoch": 0.009917790812096606,
      "grad_norm": 0.299679696559906,
      "learning_rate": 1.8999999999999998e-07,
      "loss": 0.0006,
      "step": 20
    },
    {
      "clip_ratio/high_max": 0.009408266269019805,
      "clip_ratio/high_mean": 0.002097320965049221,
      "clip_ratio/low_mean": 0.002057437336475232,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004154758266849967,
      "epoch": 0.010413680352701436,
      "grad_norm": 0.2985828220844269,
      "learning_rate": 2e-07,
      "loss": 0.0006,
      "step": 21
    },
    {
      "clip_ratio/high_max": 0.009758115858858218,
      "clip_ratio/high_mean": 0.002193691598222358,
      "clip_ratio/low_mean": 0.002050312759422468,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004244004328029405,
      "epoch": 0.010909569893306267,
      "grad_norm": 0.2506781816482544,
      "learning_rate": 2.0999999999999997e-07,
      "loss": 0.0005,
      "step": 22
    },
    {
      "clip_ratio/high_max": 0.010414585964099388,
      "clip_ratio/high_mean": 0.0022704533669184457,
      "clip_ratio/low_mean": 0.0023323346608776774,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004602788030751981,
      "epoch": 0.011405459433911096,
      "grad_norm": 0.24824994802474976,
      "learning_rate": 2.1999999999999998e-07,
      "loss": 0.0005,
      "step": 23
    },
    {
      "clip_ratio/high_max": 0.01003274679533206,
      "clip_ratio/high_mean": 0.0022769658285142214,
      "clip_ratio/low_mean": 0.0021918380275280924,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0044688038315143785,
      "epoch": 0.011901348974515927,
      "grad_norm": 0.25023654103279114,
      "learning_rate": 2.3e-07,
      "loss": 0.0006,
      "step": 24
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0558035714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15110.0,
      "completions/mean_length": 1881.7032470703125,
      "completions/mean_terminated_length": 1085.113525390625,
      "completions/min_length": 130.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.012397238515120756,
      "grad_norm": 0.24290600419044495,
      "learning_rate": 2.4e-07,
      "loss": 0.0,
      "num_tokens": 3800838.0,
      "reward": 0.2745535969734192,
      "reward_std": 0.28127801418304443,
      "rewards/accuracy_reward/mean": 0.2745535671710968,
      "rewards/accuracy_reward/std": 0.4467879831790924,
      "step": 25
    },
    {
      "clip_ratio/high_max": 0.00800984944726224,
      "clip_ratio/high_mean": 0.0016682330235653353,
      "clip_ratio/low_mean": 0.0022027858335604833,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038710188537152135,
      "epoch": 0.012893128055725588,
      "grad_norm": 0.23155416548252106,
      "learning_rate": 2.5e-07,
      "loss": 0.0004,
      "step": 26
    },
    {
      "clip_ratio/high_max": 0.008114075211778982,
      "clip_ratio/high_mean": 0.0016869619898898236,
      "clip_ratio/low_mean": 0.0023123729831695528,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003999334932132115,
      "epoch": 0.013389017596330417,
      "grad_norm": 0.23805807530879974,
      "learning_rate": 2.6e-07,
      "loss": 0.0005,
      "step": 27
    },
    {
      "clip_ratio/high_max": 0.008217864167818334,
      "clip_ratio/high_mean": 0.0017665997902440722,
      "clip_ratio/low_mean": 0.0023187257736481115,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004085325558207842,
      "epoch": 0.013884907136935248,
      "grad_norm": 0.2324390709400177,
      "learning_rate": 2.7e-07,
      "loss": 0.0004,
      "step": 28
    },
    {
      "clip_ratio/high_max": 0.007910606589575764,
      "clip_ratio/high_mean": 0.001672662552437032,
      "clip_ratio/low_mean": 0.002422388981358381,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004095051536751271,
      "epoch": 0.014380796677540077,
      "grad_norm": 0.23118717968463898,
      "learning_rate": 2.8e-07,
      "loss": 0.0004,
      "step": 29
    },
    {
      "clip_ratio/high_max": 0.007963672807818512,
      "clip_ratio/high_mean": 0.0017139911774393113,
      "clip_ratio/low_mean": 0.0023007951597264764,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0040147863246602356,
      "epoch": 0.014876686218144908,
      "grad_norm": 0.23215720057487488,
      "learning_rate": 2.9e-07,
      "loss": 0.0004,
      "step": 30
    },
    {
      "clip_ratio/high_max": 0.007794904100592248,
      "clip_ratio/high_mean": 0.0016359256542273215,
      "clip_ratio/low_mean": 0.002221869435516055,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003857795079056814,
      "epoch": 0.015372575758749738,
      "grad_norm": 0.22901228070259094,
      "learning_rate": 3e-07,
      "loss": 0.0005,
      "step": 31
    },
    {
      "clip_ratio/high_max": 0.008335780330526177,
      "clip_ratio/high_mean": 0.0017504490574538067,
      "clip_ratio/low_mean": 0.0023030940897115215,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004053543140344118,
      "epoch": 0.015868465299354567,
      "grad_norm": 0.2314842790365219,
      "learning_rate": 3.1e-07,
      "loss": 0.0004,
      "step": 32
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.044642857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15033.0,
      "completions/mean_length": 1669.60498046875,
      "completions/mean_terminated_length": 1029.8668212890625,
      "completions/min_length": 10.0,
      "completions/min_terminated_length": 10.0,
      "epoch": 0.0163643548399594,
      "grad_norm": 0.541083574295044,
      "learning_rate": 3.2e-07,
      "loss": 0.0,
      "num_tokens": 4620285.0,
      "reward": 0.2790178656578064,
      "reward_std": 0.2512976825237274,
      "rewards/accuracy_reward/mean": 0.2790178656578064,
      "rewards/accuracy_reward/std": 0.449017733335495,
      "step": 33
    },
    {
      "clip_ratio/high_max": 0.010374229990702588,
      "clip_ratio/high_mean": 0.002174396147211155,
      "clip_ratio/low_mean": 0.002206031362959493,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004380427468277048,
      "epoch": 0.01686024438056423,
      "grad_norm": 0.32749730348587036,
      "learning_rate": 3.3e-07,
      "loss": 0.0002,
      "step": 34
    },
    {
      "clip_ratio/high_max": 0.00912506164786464,
      "clip_ratio/high_mean": 0.0019906305128643,
      "clip_ratio/low_mean": 0.001996911773460397,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003987542348113493,
      "epoch": 0.01735613392116906,
      "grad_norm": 0.3368091285228729,
      "learning_rate": 3.4000000000000003e-07,
      "loss": 0.0003,
      "step": 35
    },
    {
      "clip_ratio/high_max": 0.010064000460261013,
      "clip_ratio/high_mean": 0.002157824822461407,
      "clip_ratio/low_mean": 0.0025021279375891936,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004659952754082042,
      "epoch": 0.01785202346177389,
      "grad_norm": 0.36362946033477783,
      "learning_rate": 3.5e-07,
      "loss": 0.0001,
      "step": 36
    },
    {
      "clip_ratio/high_max": 0.009934016852639616,
      "clip_ratio/high_mean": 0.0021664341393261566,
      "clip_ratio/low_mean": 0.0023220588718686486,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004488492988457438,
      "epoch": 0.01834791300237872,
      "grad_norm": 0.3851204216480255,
      "learning_rate": 3.6e-07,
      "loss": 0.0002,
      "step": 37
    },
    {
      "clip_ratio/high_max": 0.010322031561372569,
      "clip_ratio/high_mean": 0.0022267843100962637,
      "clip_ratio/low_mean": 0.0024607914140801768,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004687575705247582,
      "epoch": 0.01884380254298355,
      "grad_norm": 0.24734635651111603,
      "learning_rate": 3.7e-07,
      "loss": 0.0,
      "step": 38
    },
    {
      "clip_ratio/high_max": 0.008923900830268394,
      "clip_ratio/high_mean": 0.0020342444249763503,
      "clip_ratio/low_mean": 0.002789703589371584,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004823948027478764,
      "epoch": 0.01933969208358838,
      "grad_norm": 0.2821483314037323,
      "learning_rate": 3.7999999999999996e-07,
      "loss": 0.0001,
      "step": 39
    },
    {
      "clip_ratio/high_max": 0.009451986486965325,
      "clip_ratio/high_mean": 0.002121220696153614,
      "clip_ratio/low_mean": 0.0030893643970557605,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005210585091845132,
      "epoch": 0.019835581624193212,
      "grad_norm": 0.24948416650295258,
      "learning_rate": 3.8999999999999997e-07,
      "loss": 0.0,
      "step": 40
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.033482142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15283.0,
      "completions/mean_length": 1671.024658203125,
      "completions/mean_terminated_length": 1196.8106689453125,
      "completions/min_length": 89.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.02033147116479804,
      "grad_norm": 0.22222869098186493,
      "learning_rate": 4e-07,
      "loss": -0.0,
      "num_tokens": 5442344.0,
      "reward": 0.2745535969734192,
      "reward_std": 0.2836872935295105,
      "rewards/accuracy_reward/mean": 0.2745535671710968,
      "rewards/accuracy_reward/std": 0.4467879831790924,
      "step": 41
    },
    {
      "clip_ratio/high_max": 0.009384013937960844,
      "clip_ratio/high_mean": 0.0019746449793274223,
      "clip_ratio/low_mean": 0.0023150522670221108,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004289697264539427,
      "epoch": 0.02082736070540287,
      "grad_norm": 0.20703142881393433,
      "learning_rate": 4.0999999999999994e-07,
      "loss": 0.0005,
      "step": 42
    },
    {
      "clip_ratio/high_max": 0.008630768632428953,
      "clip_ratio/high_mean": 0.0018106386814906728,
      "clip_ratio/low_mean": 0.0024419532423962664,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004252591888871393,
      "epoch": 0.021323250246007702,
      "grad_norm": 0.20188823342323303,
      "learning_rate": 4.1999999999999995e-07,
      "loss": 0.0004,
      "step": 43
    },
    {
      "clip_ratio/high_max": 0.008670552611874882,
      "clip_ratio/high_mean": 0.0018617365494719706,
      "clip_ratio/low_mean": 0.002528024277808072,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004389760828416911,
      "epoch": 0.021819139786612533,
      "grad_norm": 0.19356396794319153,
      "learning_rate": 4.2999999999999996e-07,
      "loss": 0.0004,
      "step": 44
    },
    {
      "clip_ratio/high_max": 0.00822552266981802,
      "clip_ratio/high_mean": 0.0017788405284591136,
      "clip_ratio/low_mean": 0.0026053979340758815,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004384238494822057,
      "epoch": 0.02231502932721736,
      "grad_norm": 0.1998298317193985,
      "learning_rate": 4.3999999999999997e-07,
      "loss": 0.0004,
      "step": 45
    },
    {
      "clip_ratio/high_max": 0.00948522142789443,
      "clip_ratio/high_mean": 0.0020446483040359453,
      "clip_ratio/low_mean": 0.00256819612786785,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004612844422808848,
      "epoch": 0.022810918867822192,
      "grad_norm": 0.19322603940963745,
      "learning_rate": 4.5e-07,
      "loss": 0.0003,
      "step": 46
    },
    {
      "clip_ratio/high_max": 0.00879126082691073,
      "clip_ratio/high_mean": 0.0018611239693200332,
      "clip_ratio/low_mean": 0.002711578200432996,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00457270219703787,
      "epoch": 0.023306808408427023,
      "grad_norm": 0.20106711983680725,
      "learning_rate": 4.6e-07,
      "loss": 0.0003,
      "step": 47
    },
    {
      "clip_ratio/high_max": 0.0086006731071393,
      "clip_ratio/high_mean": 0.001867603264599893,
      "clip_ratio/low_mean": 0.0028253323657736473,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0046929356449254556,
      "epoch": 0.023802697949031854,
      "grad_norm": 0.19773784279823303,
      "learning_rate": 4.6999999999999995e-07,
      "loss": 0.0003,
      "step": 48
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0424107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13924.0,
      "completions/mean_length": 1820.1563720703125,
      "completions/mean_terminated_length": 1220.489501953125,
      "completions/min_length": 102.0,
      "completions/min_terminated_length": 102.0,
      "epoch": 0.024298587489636682,
      "grad_norm": 0.20372866094112396,
      "learning_rate": 4.8e-07,
      "loss": -0.0,
      "num_tokens": 6326318.0,
      "reward": 0.2477678656578064,
      "reward_std": 0.2468036413192749,
      "rewards/accuracy_reward/mean": 0.2477678507566452,
      "rewards/accuracy_reward/std": 0.4321989119052887,
      "step": 49
    },
    {
      "clip_ratio/high_max": 0.008343736210008501,
      "clip_ratio/high_mean": 0.0016661981994730013,
      "clip_ratio/low_mean": 0.0019082582328167064,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003574456429305428,
      "epoch": 0.024794477030241513,
      "grad_norm": 0.19582587480545044,
      "learning_rate": 4.9e-07,
      "loss": 0.0004,
      "step": 50
    },
    {
      "clip_ratio/high_max": 0.008456764116999693,
      "clip_ratio/high_mean": 0.0016215470689076028,
      "clip_ratio/low_mean": 0.0018815803529719233,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003503127405565465,
      "epoch": 0.025290366570846344,
      "grad_norm": 0.1932041198015213,
      "learning_rate": 5e-07,
      "loss": 0.0004,
      "step": 51
    },
    {
      "clip_ratio/high_max": 0.009115832919633249,
      "clip_ratio/high_mean": 0.0017474604944709426,
      "clip_ratio/low_mean": 0.0019994199282677982,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003746880423022958,
      "epoch": 0.025786256111451175,
      "grad_norm": 0.19513899087905884,
      "learning_rate": 5.1e-07,
      "loss": 0.0003,
      "step": 52
    },
    {
      "clip_ratio/high_max": 0.009240591065463377,
      "clip_ratio/high_mean": 0.0018747792073554592,
      "clip_ratio/low_mean": 0.0019603708745989934,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038351501016222755,
      "epoch": 0.026282145652056006,
      "grad_norm": 0.18583644926548004,
      "learning_rate": 5.2e-07,
      "loss": 0.0003,
      "step": 53
    },
    {
      "clip_ratio/high_max": 0.009182901441818103,
      "clip_ratio/high_mean": 0.0017495169577159686,
      "clip_ratio/low_mean": 0.0020165997476624398,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037661166807083646,
      "epoch": 0.026778035192660834,
      "grad_norm": 0.18464510142803192,
      "learning_rate": 5.3e-07,
      "loss": 0.0003,
      "step": 54
    },
    {
      "clip_ratio/high_max": 0.008938012793805683,
      "clip_ratio/high_mean": 0.0017761641279321339,
      "clip_ratio/low_mean": 0.0019855888298252466,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037617529487761203,
      "epoch": 0.027273924733265665,
      "grad_norm": 0.1786801517009735,
      "learning_rate": 5.4e-07,
      "loss": 0.0003,
      "step": 55
    },
    {
      "clip_ratio/high_max": 0.009981536750274245,
      "clip_ratio/high_mean": 0.001974870063804701,
      "clip_ratio/low_mean": 0.0022026667857630855,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004177536858151143,
      "epoch": 0.027769814273870496,
      "grad_norm": 0.18090926110744476,
      "learning_rate": 5.5e-07,
      "loss": 0.0001,
      "step": 56
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13259.0,
      "completions/mean_length": 2491.4599609375,
      "completions/mean_terminated_length": 1367.0242919921875,
      "completions/min_length": 13.0,
      "completions/min_terminated_length": 13.0,
      "epoch": 0.028265703814475327,
      "grad_norm": 0.5446521043777466,
      "learning_rate": 5.6e-07,
      "loss": -0.0,
      "num_tokens": 7514132.0,
      "reward": 0.2901785969734192,
      "reward_std": 0.23979702591896057,
      "rewards/accuracy_reward/mean": 0.2901785671710968,
      "rewards/accuracy_reward/std": 0.4543520212173462,
      "step": 57
    },
    {
      "clip_ratio/high_max": 0.006588367810763884,
      "clip_ratio/high_mean": 0.0015784529946358816,
      "clip_ratio/low_mean": 0.0016161439206712203,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003194596910816472,
      "epoch": 0.028761593355080155,
      "grad_norm": 0.36485645174980164,
      "learning_rate": 5.699999999999999e-07,
      "loss": 0.0002,
      "step": 58
    },
    {
      "clip_ratio/high_max": 0.006623814862905419,
      "clip_ratio/high_mean": 0.001615854187093646,
      "clip_ratio/low_mean": 0.0018882707726248782,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035041250133645008,
      "epoch": 0.029257482895684986,
      "grad_norm": 0.26758676767349243,
      "learning_rate": 5.8e-07,
      "loss": 0.0002,
      "step": 59
    },
    {
      "clip_ratio/high_max": 0.006900099528138526,
      "clip_ratio/high_mean": 0.0017312117774963554,
      "clip_ratio/low_mean": 0.0019010523388658385,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036322641498429675,
      "epoch": 0.029753372436289817,
      "grad_norm": 0.2793244421482086,
      "learning_rate": 5.9e-07,
      "loss": 0.0001,
      "step": 60
    },
    {
      "clip_ratio/high_max": 0.007645924908501911,
      "clip_ratio/high_mean": 0.0017695804444883834,
      "clip_ratio/low_mean": 0.0020805536757251275,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003850134075406686,
      "epoch": 0.030249261976894648,
      "grad_norm": 0.252706378698349,
      "learning_rate": 6e-07,
      "loss": 0.0001,
      "step": 61
    },
    {
      "clip_ratio/high_max": 0.007272356910107192,
      "clip_ratio/high_mean": 0.00177738125603355,
      "clip_ratio/low_mean": 0.0024392984201142554,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042166796837648235,
      "epoch": 0.030745151517499476,
      "grad_norm": 0.19972658157348633,
      "learning_rate": 6.1e-07,
      "loss": -0.0,
      "step": 62
    },
    {
      "clip_ratio/high_max": 0.007467888461178518,
      "clip_ratio/high_mean": 0.001845906908329198,
      "clip_ratio/low_mean": 0.0025745886549088937,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004420495571991978,
      "epoch": 0.031241041058104307,
      "grad_norm": 0.24313460290431976,
      "learning_rate": 6.2e-07,
      "loss": -0.0002,
      "step": 63
    },
    {
      "clip_ratio/high_max": 0.007929844014142873,
      "clip_ratio/high_mean": 0.0018546227306615037,
      "clip_ratio/low_mean": 0.0027119049950385943,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004566527655185837,
      "epoch": 0.031736930598709134,
      "grad_norm": 0.19414299726486206,
      "learning_rate": 6.3e-07,
      "loss": -0.0003,
      "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": 0.046875,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14543.0,
      "completions/mean_length": 1793.696533203125,
      "completions/mean_terminated_length": 1126.5010986328125,
      "completions/min_length": 12.0,
      "completions/min_terminated_length": 12.0,
      "epoch": 0.032232820139313965,
      "grad_norm": 0.2702247202396393,
      "learning_rate": 6.4e-07,
      "loss": -0.0,
      "num_tokens": 8385660.0,
      "reward": 0.3504464328289032,
      "reward_std": 0.30217117071151733,
      "rewards/accuracy_reward/mean": 0.3504464328289032,
      "rewards/accuracy_reward/std": 0.47764310240745544,
      "step": 65
    },
    {
      "clip_ratio/high_max": 0.010357285143982153,
      "clip_ratio/high_mean": 0.0026252009647578234,
      "clip_ratio/low_mean": 0.002045263200685099,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004670464151786291,
      "epoch": 0.0327287096799188,
      "grad_norm": 0.23609422147274017,
      "learning_rate": 6.5e-07,
      "loss": 0.0003,
      "step": 66
    },
    {
      "clip_ratio/high_max": 0.010334131058698404,
      "clip_ratio/high_mean": 0.0024866511241725675,
      "clip_ratio/low_mean": 0.002378508413585223,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0048651595097908285,
      "epoch": 0.03322459922052363,
      "grad_norm": 0.23260565102100372,
      "learning_rate": 6.6e-07,
      "loss": 0.0003,
      "step": 67
    },
    {
      "clip_ratio/high_max": 0.010065025235235225,
      "clip_ratio/high_mean": 0.0025154042568829027,
      "clip_ratio/low_mean": 0.001975412642437391,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004490816849283874,
      "epoch": 0.03372048876112846,
      "grad_norm": 0.2346385419368744,
      "learning_rate": 6.7e-07,
      "loss": 0.0002,
      "step": 68
    },
    {
      "clip_ratio/high_max": 0.010829139186171233,
      "clip_ratio/high_mean": 0.0026294352874174365,
      "clip_ratio/low_mean": 0.0028190506814240734,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00544848591198388,
      "epoch": 0.03421637830173329,
      "grad_norm": 0.21574367582798004,
      "learning_rate": 6.800000000000001e-07,
      "loss": 0.0001,
      "step": 69
    },
    {
      "clip_ratio/high_max": 0.011391910975362407,
      "clip_ratio/high_mean": 0.0028609727419279807,
      "clip_ratio/low_mean": 0.0028425585494460393,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005703531327526434,
      "epoch": 0.03471226784233812,
      "grad_norm": 0.2143310010433197,
      "learning_rate": 6.9e-07,
      "loss": -0.0,
      "step": 70
    },
    {
      "clip_ratio/high_max": 0.011299408986815251,
      "clip_ratio/high_mean": 0.002873901948078128,
      "clip_ratio/low_mean": 0.003413775222043114,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006287677148065995,
      "epoch": 0.03520815738294295,
      "grad_norm": 0.20147114992141724,
      "learning_rate": 7e-07,
      "loss": -0.0002,
      "step": 71
    },
    {
      "clip_ratio/high_max": 0.012002673640381545,
      "clip_ratio/high_mean": 0.002957386867819878,
      "clip_ratio/low_mean": 0.0038766998823263066,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006834086738308542,
      "epoch": 0.03570404692354778,
      "grad_norm": 0.18365052342414856,
      "learning_rate": 7.1e-07,
      "loss": -0.0003,
      "step": 72
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.024553571428571397,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15010.0,
      "completions/mean_length": 1508.341552734375,
      "completions/mean_terminated_length": 1159.6727294921875,
      "completions/min_length": 12.0,
      "completions/min_terminated_length": 12.0,
      "epoch": 0.03619993646415261,
      "grad_norm": 0.20101158320903778,
      "learning_rate": 7.2e-07,
      "loss": -0.0,
      "num_tokens": 9134725.0,
      "reward": 0.361607164144516,
      "reward_std": 0.2796297073364258,
      "rewards/accuracy_reward/mean": 0.3616071343421936,
      "rewards/accuracy_reward/std": 0.48100295662879944,
      "step": 73
    },
    {
      "clip_ratio/high_max": 0.005742017186094017,
      "clip_ratio/high_mean": 0.0012924441111863416,
      "clip_ratio/low_mean": 0.0015960580457203832,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002888502118366887,
      "epoch": 0.03669582600475744,
      "grad_norm": 0.1593097746372223,
      "learning_rate": 7.3e-07,
      "loss": 0.0003,
      "step": 74
    },
    {
      "clip_ratio/high_max": 0.005765729106315121,
      "clip_ratio/high_mean": 0.0012644981788980658,
      "clip_ratio/low_mean": 0.0018857609875908565,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031502591655225842,
      "epoch": 0.03719171554536227,
      "grad_norm": 0.15791821479797363,
      "learning_rate": 7.4e-07,
      "loss": 0.0002,
      "step": 75
    },
    {
      "clip_ratio/high_max": 0.006536488142046437,
      "clip_ratio/high_mean": 0.0014311991665181267,
      "clip_ratio/low_mean": 0.0019422289910266954,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033734281541910605,
      "epoch": 0.0376876050859671,
      "grad_norm": 0.15379948914051056,
      "learning_rate": 7.5e-07,
      "loss": 0.0002,
      "step": 76
    },
    {
      "clip_ratio/high_max": 0.006879762720927829,
      "clip_ratio/high_mean": 0.0014608523933929973,
      "clip_ratio/low_mean": 0.0020882450667727426,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035490974432832445,
      "epoch": 0.03818349462657193,
      "grad_norm": 0.15018825232982635,
      "learning_rate": 7.599999999999999e-07,
      "loss": 0.0001,
      "step": 77
    },
    {
      "clip_ratio/high_max": 0.007366736174844846,
      "clip_ratio/high_mean": 0.0015952294488670304,
      "clip_ratio/low_mean": 0.0021738074013342157,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037690368735638913,
      "epoch": 0.03867938416717676,
      "grad_norm": 0.15790612995624542,
      "learning_rate": 7.699999999999999e-07,
      "loss": 0.0001,
      "step": 78
    },
    {
      "clip_ratio/high_max": 0.0076883125184394885,
      "clip_ratio/high_mean": 0.0016483859399158973,
      "clip_ratio/low_mean": 0.0022395542385993394,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038879401627127663,
      "epoch": 0.039175273707781594,
      "grad_norm": 0.14808905124664307,
      "learning_rate": 7.799999999999999e-07,
      "loss": -0.0001,
      "step": 79
    },
    {
      "clip_ratio/high_max": 0.00864371846546419,
      "clip_ratio/high_mean": 0.001838633846546145,
      "clip_ratio/low_mean": 0.0025456230890767984,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00438425692027522,
      "epoch": 0.039671163248386425,
      "grad_norm": 0.14132992923259735,
      "learning_rate": 7.9e-07,
      "loss": -0.0002,
      "step": 80
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.017857142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14512.0,
      "completions/mean_length": 1377.93310546875,
      "completions/mean_terminated_length": 1123.713623046875,
      "completions/min_length": 172.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.04016705278899125,
      "grad_norm": 0.20323733985424042,
      "learning_rate": 8e-07,
      "loss": 0.0,
      "num_tokens": 9819255.0,
      "reward": 0.3906250298023224,
      "reward_std": 0.2382306456565857,
      "rewards/accuracy_reward/mean": 0.390625,
      "rewards/accuracy_reward/std": 0.48843589425086975,
      "step": 81
    },
    {
      "clip_ratio/high_max": 0.008460259119601687,
      "clip_ratio/high_mean": 0.0019007410573976813,
      "clip_ratio/low_mean": 0.001299077861801834,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031998189124351484,
      "epoch": 0.04066294232959608,
      "grad_norm": 0.18832682073116302,
      "learning_rate": 8.1e-07,
      "loss": 0.0004,
      "step": 82
    },
    {
      "clip_ratio/high_max": 0.008343295463419054,
      "clip_ratio/high_mean": 0.0018727999513430404,
      "clip_ratio/low_mean": 0.0013857753385195792,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003258575263316743,
      "epoch": 0.04115883187020091,
      "grad_norm": 0.18185274302959442,
      "learning_rate": 8.199999999999999e-07,
      "loss": 0.0003,
      "step": 83
    },
    {
      "clip_ratio/high_max": 0.008886882356819115,
      "clip_ratio/high_mean": 0.002058483564724156,
      "clip_ratio/low_mean": 0.0015208641333970263,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003579347706363478,
      "epoch": 0.04165472141080574,
      "grad_norm": 0.17563284933567047,
      "learning_rate": 8.299999999999999e-07,
      "loss": 0.0002,
      "step": 84
    },
    {
      "clip_ratio/high_max": 0.009757792206073646,
      "clip_ratio/high_mean": 0.0021702875856135506,
      "clip_ratio/low_mean": 0.001593417201718239,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003763704718039662,
      "epoch": 0.04215061095141057,
      "grad_norm": 0.17286504805088043,
      "learning_rate": 8.399999999999999e-07,
      "loss": 0.0001,
      "step": 85
    },
    {
      "clip_ratio/high_max": 0.010141662336536683,
      "clip_ratio/high_mean": 0.002254493506370636,
      "clip_ratio/low_mean": 0.0018539100331054215,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004108403507416369,
      "epoch": 0.042646500492015404,
      "grad_norm": 0.17081013321876526,
      "learning_rate": 8.499999999999999e-07,
      "loss": -0.0,
      "step": 86
    },
    {
      "clip_ratio/high_max": 0.011031183723389404,
      "clip_ratio/high_mean": 0.0024779251598374685,
      "clip_ratio/low_mean": 0.0020636038942143387,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0045415290815071785,
      "epoch": 0.043142390032620236,
      "grad_norm": 0.1589944213628769,
      "learning_rate": 8.599999999999999e-07,
      "loss": -0.0002,
      "step": 87
    },
    {
      "clip_ratio/high_max": 0.011667547358229058,
      "clip_ratio/high_mean": 0.002645916350047628,
      "clip_ratio/low_mean": 0.002288584555401485,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0049345009410899365,
      "epoch": 0.04363827957322507,
      "grad_norm": 0.14980703592300415,
      "learning_rate": 8.699999999999999e-07,
      "loss": -0.0005,
      "step": 88
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0424107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15187.0,
      "completions/mean_length": 1854.4866943359375,
      "completions/mean_terminated_length": 1256.34033203125,
      "completions/min_length": 219.0,
      "completions/min_terminated_length": 219.0,
      "epoch": 0.0441341691138299,
      "grad_norm": 0.16792313754558563,
      "learning_rate": 8.799999999999999e-07,
      "loss": -0.0,
      "num_tokens": 10720529.0,
      "reward": 0.3214285969734192,
      "reward_std": 0.22161763906478882,
      "rewards/accuracy_reward/mean": 0.3214285671710968,
      "rewards/accuracy_reward/std": 0.4675469994544983,
      "step": 89
    },
    {
      "clip_ratio/high_max": 0.006549437445755757,
      "clip_ratio/high_mean": 0.0013885769656099,
      "clip_ratio/low_mean": 0.0009579541519428858,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023465311405743705,
      "epoch": 0.04463005865443472,
      "grad_norm": 0.15483498573303223,
      "learning_rate": 8.9e-07,
      "loss": 0.0002,
      "step": 90
    },
    {
      "clip_ratio/high_max": 0.006471253005656763,
      "clip_ratio/high_mean": 0.0013857557669325615,
      "clip_ratio/low_mean": 0.0010186496834307945,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024044054498517653,
      "epoch": 0.04512594819503955,
      "grad_norm": 0.15085436403751373,
      "learning_rate": 9e-07,
      "loss": 0.0001,
      "step": 91
    },
    {
      "clip_ratio/high_max": 0.006693593444651924,
      "clip_ratio/high_mean": 0.0014197043431067868,
      "clip_ratio/low_mean": 0.0011303372639872578,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025500416259092162,
      "epoch": 0.045621837735644384,
      "grad_norm": 0.148464173078537,
      "learning_rate": 9.1e-07,
      "loss": 0.0001,
      "step": 92
    },
    {
      "clip_ratio/high_max": 0.007533042684372049,
      "clip_ratio/high_mean": 0.0015869697517700843,
      "clip_ratio/low_mean": 0.0012826252967101937,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002869595006814052,
      "epoch": 0.046117727276249215,
      "grad_norm": 0.14271365106105804,
      "learning_rate": 9.2e-07,
      "loss": -0.0002,
      "step": 93
    },
    {
      "clip_ratio/high_max": 0.008238533709118201,
      "clip_ratio/high_mean": 0.0017262667431623413,
      "clip_ratio/low_mean": 0.0014477107552011148,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003173977509504766,
      "epoch": 0.046613616816854046,
      "grad_norm": 0.13681292533874512,
      "learning_rate": 9.3e-07,
      "loss": -0.0003,
      "step": 94
    },
    {
      "clip_ratio/high_max": 0.007675043532799464,
      "clip_ratio/high_mean": 0.0016133102124058496,
      "clip_ratio/low_mean": 0.0015988743328989585,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032121845488291,
      "epoch": 0.04710950635745888,
      "grad_norm": 0.13631173968315125,
      "learning_rate": 9.399999999999999e-07,
      "loss": -0.0003,
      "step": 95
    },
    {
      "clip_ratio/high_max": 0.008664268322718272,
      "clip_ratio/high_mean": 0.0018614575385527132,
      "clip_ratio/low_mean": 0.0016848377807718862,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003546295287378598,
      "epoch": 0.04760539589806371,
      "grad_norm": 0.1315957009792328,
      "learning_rate": 9.499999999999999e-07,
      "loss": -0.0005,
      "step": 96
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14992.0,
      "completions/mean_length": 1861.5469970703125,
      "completions/mean_terminated_length": 1426.1129150390625,
      "completions/min_length": 243.0,
      "completions/min_terminated_length": 243.0,
      "epoch": 0.04810128543866854,
      "grad_norm": 0.14852479100227356,
      "learning_rate": 9.6e-07,
      "loss": 0.0,
      "num_tokens": 11637726.0,
      "reward": 0.2388392984867096,
      "reward_std": 0.2324451208114624,
      "rewards/accuracy_reward/mean": 0.2388392835855484,
      "rewards/accuracy_reward/std": 0.4268510043621063,
      "step": 97
    },
    {
      "clip_ratio/high_max": 0.00659341320169915,
      "clip_ratio/high_mean": 0.0012911487785913778,
      "clip_ratio/low_mean": 0.0013692264427618284,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002660375233972445,
      "epoch": 0.048597174979273364,
      "grad_norm": 0.1425253450870514,
      "learning_rate": 9.7e-07,
      "loss": 0.0003,
      "step": 98
    },
    {
      "clip_ratio/high_max": 0.006399107505785651,
      "clip_ratio/high_mean": 0.0013406837740603805,
      "clip_ratio/low_mean": 0.001649623943478673,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029903077165727154,
      "epoch": 0.049093064519878195,
      "grad_norm": 0.14061778783798218,
      "learning_rate": 9.8e-07,
      "loss": 0.0001,
      "step": 99
    },
    {
      "clip_ratio/high_max": 0.007037838104224647,
      "clip_ratio/high_mean": 0.0014072924964239064,
      "clip_ratio/low_mean": 0.0016709519721871402,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030782444373471662,
      "epoch": 0.049588954060483026,
      "grad_norm": 0.13562434911727905,
      "learning_rate": 9.9e-07,
      "loss": 0.0001,
      "step": 100
    },
    {
      "clip_ratio/high_max": 0.0073998546049551805,
      "clip_ratio/high_mean": 0.0015221267690321838,
      "clip_ratio/low_mean": 0.0018762041920012962,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033983309695031494,
      "epoch": 0.05008484360108786,
      "grad_norm": 0.1329057663679123,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 101
    },
    {
      "clip_ratio/high_max": 0.008212845928937895,
      "clip_ratio/high_mean": 0.0016316274118253205,
      "clip_ratio/low_mean": 0.0020412229162616313,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003672850321891019,
      "epoch": 0.05058073314169269,
      "grad_norm": 0.12286388128995895,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 102
    },
    {
      "clip_ratio/high_max": 0.009515765505057061,
      "clip_ratio/high_mean": 0.001857984800608392,
      "clip_ratio/low_mean": 0.002240926984654834,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004098911816981854,
      "epoch": 0.05107662268229752,
      "grad_norm": 0.1234845519065857,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 103
    },
    {
      "clip_ratio/high_max": 0.009760147015185794,
      "clip_ratio/high_mean": 0.0019198810427951685,
      "clip_ratio/low_mean": 0.0024176864558853595,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004337567490438232,
      "epoch": 0.05157251222290235,
      "grad_norm": 0.11556278169155121,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 104
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0267857142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13964.0,
      "completions/mean_length": 1508.493408203125,
      "completions/mean_terminated_length": 1127.2591552734375,
      "completions/min_length": 222.0,
      "completions/min_terminated_length": 222.0,
      "epoch": 0.05206840176350718,
      "grad_norm": 0.1436101496219635,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 12384163.0,
      "reward": 0.3883928656578064,
      "reward_std": 0.2050851285457611,
      "rewards/accuracy_reward/mean": 0.3883928656578064,
      "rewards/accuracy_reward/std": 0.4879295527935028,
      "step": 105
    },
    {
      "clip_ratio/high_max": 0.005477254227116646,
      "clip_ratio/high_mean": 0.0012020476776797295,
      "clip_ratio/low_mean": 0.0012692021035149992,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00247124977772728,
      "epoch": 0.05256429130411201,
      "grad_norm": 0.13298627734184265,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 106
    },
    {
      "clip_ratio/high_max": 0.005473626290040556,
      "clip_ratio/high_mean": 0.0012191067766025299,
      "clip_ratio/low_mean": 0.0012302214293811176,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024493282269304473,
      "epoch": 0.05306018084471684,
      "grad_norm": 0.1409875750541687,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 107
    },
    {
      "clip_ratio/high_max": 0.005747606481236289,
      "clip_ratio/high_mean": 0.0012774351969255804,
      "clip_ratio/low_mean": 0.0014263192100258948,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027037544161885307,
      "epoch": 0.05355607038532167,
      "grad_norm": 0.13383662700653076,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 108
    },
    {
      "clip_ratio/high_max": 0.006393430630851071,
      "clip_ratio/high_mean": 0.0014089425960719382,
      "clip_ratio/low_mean": 0.0015831052301109594,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002992047789462049,
      "epoch": 0.0540519599259265,
      "grad_norm": 0.12524321675300598,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 109
    },
    {
      "clip_ratio/high_max": 0.0070557514645770425,
      "clip_ratio/high_mean": 0.0015346493315746557,
      "clip_ratio/low_mean": 0.001789445642401688,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033240949796891073,
      "epoch": 0.05454784946653133,
      "grad_norm": 0.12360424548387527,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 110
    },
    {
      "clip_ratio/high_max": 0.007985540268236946,
      "clip_ratio/high_mean": 0.0017138075637603833,
      "clip_ratio/low_mean": 0.0019596908058190365,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036734983325459325,
      "epoch": 0.05504373900713616,
      "grad_norm": 0.11155232787132263,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 111
    },
    {
      "clip_ratio/high_max": 0.008731966813684267,
      "clip_ratio/high_mean": 0.0018613774278719575,
      "clip_ratio/low_mean": 0.002224065162096167,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004085442571067688,
      "epoch": 0.05553962854774099,
      "grad_norm": 0.11124052107334137,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 112
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.044642857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13707.0,
      "completions/mean_length": 1783.935302734375,
      "completions/mean_terminated_length": 1149.5396728515625,
      "completions/min_length": 217.0,
      "completions/min_terminated_length": 217.0,
      "epoch": 0.05603551808834582,
      "grad_norm": 0.17350871860980988,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 13253462.0,
      "reward": 0.4196428656578064,
      "reward_std": 0.24122920632362366,
      "rewards/accuracy_reward/mean": 0.4196428656578064,
      "rewards/accuracy_reward/std": 0.4940522015094757,
      "step": 113
    },
    {
      "clip_ratio/high_max": 0.007698455454374198,
      "clip_ratio/high_mean": 0.0018029885293344705,
      "clip_ratio/low_mean": 0.0009671371690842534,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027701257138232904,
      "epoch": 0.056531407628950654,
      "grad_norm": 0.15774789452552795,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 114
    },
    {
      "clip_ratio/high_max": 0.008064122514042538,
      "clip_ratio/high_mean": 0.0018469209796876385,
      "clip_ratio/low_mean": 0.000852889451266492,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026998104365247855,
      "epoch": 0.05702729716955548,
      "grad_norm": 0.1533394753932953,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 115
    },
    {
      "clip_ratio/high_max": 0.00866946319820272,
      "clip_ratio/high_mean": 0.001963068180657501,
      "clip_ratio/low_mean": 0.0011198187007721572,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030828869107608625,
      "epoch": 0.05752318671016031,
      "grad_norm": 0.15033955872058868,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 116
    },
    {
      "clip_ratio/high_max": 0.010176094001508318,
      "clip_ratio/high_mean": 0.002274178921197745,
      "clip_ratio/low_mean": 0.0012626684958831902,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035368473559174163,
      "epoch": 0.05801907625076514,
      "grad_norm": 0.1445356160402298,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 117
    },
    {
      "clip_ratio/high_max": 0.011022347651305608,
      "clip_ratio/high_mean": 0.0025182671333823237,
      "clip_ratio/low_mean": 0.0015351211966958545,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0040533883498028445,
      "epoch": 0.05851496579136997,
      "grad_norm": 0.13612000644207,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 118
    },
    {
      "clip_ratio/high_max": 0.012174287498055492,
      "clip_ratio/high_mean": 0.0026913874078218214,
      "clip_ratio/low_mean": 0.001783059762431094,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004474447126312953,
      "epoch": 0.0590108553319748,
      "grad_norm": 0.12885288894176483,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 119
    },
    {
      "clip_ratio/high_max": 0.01353116701648105,
      "clip_ratio/high_mean": 0.0030173748127708677,
      "clip_ratio/low_mean": 0.0019601648395450866,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004977539636001893,
      "epoch": 0.059506744872579634,
      "grad_norm": 0.1223459541797638,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 120
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0424107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15066.0,
      "completions/mean_length": 1878.6295166015625,
      "completions/mean_terminated_length": 1281.552490234375,
      "completions/min_length": 164.0,
      "completions/min_terminated_length": 164.0,
      "epoch": 0.060002634413184465,
      "grad_norm": 0.1658322960138321,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 14161776.0,
      "reward": 0.3571428656578064,
      "reward_std": 0.24198371171951294,
      "rewards/accuracy_reward/mean": 0.3571428656578064,
      "rewards/accuracy_reward/std": 0.47969308495521545,
      "step": 121
    },
    {
      "clip_ratio/high_max": 0.006607086127587536,
      "clip_ratio/high_mean": 0.0013458100768275472,
      "clip_ratio/low_mean": 0.0012133605299595729,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025591706134946435,
      "epoch": 0.060498523953789296,
      "grad_norm": 0.1524488925933838,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 122
    },
    {
      "clip_ratio/high_max": 0.007686171215027571,
      "clip_ratio/high_mean": 0.0015321463959026005,
      "clip_ratio/low_mean": 0.0013129176180086688,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028450640202208888,
      "epoch": 0.06099441349439413,
      "grad_norm": 0.14416955411434174,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 123
    },
    {
      "clip_ratio/high_max": 0.008009471813238633,
      "clip_ratio/high_mean": 0.001639964398691518,
      "clip_ratio/low_mean": 0.0014685078907064053,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031084722722880542,
      "epoch": 0.06149030303499895,
      "grad_norm": 0.13974629342556,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 124
    },
    {
      "clip_ratio/high_max": 0.008719969417143147,
      "clip_ratio/high_mean": 0.0017941452920240408,
      "clip_ratio/low_mean": 0.0018138034279786552,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003607948734497768,
      "epoch": 0.06198619257560378,
      "grad_norm": 0.13295143842697144,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 125
    },
    {
      "clip_ratio/high_max": 0.009935141144524096,
      "clip_ratio/high_mean": 0.002051086260507873,
      "clip_ratio/low_mean": 0.001885160079950765,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003936246327612025,
      "epoch": 0.06248208211620861,
      "grad_norm": 0.12530085444450378,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 126
    },
    {
      "clip_ratio/high_max": 0.011307463153571007,
      "clip_ratio/high_mean": 0.002394097575916021,
      "clip_ratio/low_mean": 0.002258254509456492,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004652352088669431,
      "epoch": 0.06297797165681344,
      "grad_norm": 0.11875636130571365,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 127
    },
    {
      "clip_ratio/high_max": 0.012316636053583352,
      "clip_ratio/high_mean": 0.002579095651071839,
      "clip_ratio/low_mean": 0.002509830475560193,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00508892612742784,
      "epoch": 0.06347386119741827,
      "grad_norm": 0.1188141480088234,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "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": 0.0267857142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14042.0,
      "completions/mean_length": 1418.743408203125,
      "completions/mean_terminated_length": 1035.0389404296875,
      "completions/min_length": 231.0,
      "completions/min_terminated_length": 231.0,
      "epoch": 0.0639697507380231,
      "grad_norm": 0.16798920929431915,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 14875797.0,
      "reward": 0.3325892984867096,
      "reward_std": 0.2064494639635086,
      "rewards/accuracy_reward/mean": 0.3325892984867096,
      "rewards/accuracy_reward/std": 0.47166749835014343,
      "step": 129
    },
    {
      "clip_ratio/high_max": 0.006105694508732995,
      "clip_ratio/high_mean": 0.0012181344720829657,
      "clip_ratio/low_mean": 0.0014820045673786808,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027001390708392137,
      "epoch": 0.06446564027862793,
      "grad_norm": 0.15334412455558777,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 130
    },
    {
      "clip_ratio/high_max": 0.006340444029774517,
      "clip_ratio/high_mean": 0.0012607102917172597,
      "clip_ratio/low_mean": 0.001671608752985776,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029323190487957618,
      "epoch": 0.06496152981923277,
      "grad_norm": 0.1515379101037979,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 131
    },
    {
      "clip_ratio/high_max": 0.007701121658101329,
      "clip_ratio/high_mean": 0.001501674886640103,
      "clip_ratio/low_mean": 0.0019017575521047547,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034034324335152633,
      "epoch": 0.0654574193598376,
      "grad_norm": 0.14188846945762634,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 132
    },
    {
      "clip_ratio/high_max": 0.008396955785428872,
      "clip_ratio/high_mean": 0.0016355410136839055,
      "clip_ratio/low_mean": 0.00200198702191301,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003637528016042779,
      "epoch": 0.06595330890044243,
      "grad_norm": 0.13578101992607117,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 133
    },
    {
      "clip_ratio/high_max": 0.009826390638409066,
      "clip_ratio/high_mean": 0.0018557082714778517,
      "clip_ratio/low_mean": 0.0024351661677428638,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004290874408070522,
      "epoch": 0.06644919844104726,
      "grad_norm": 0.13408499956130981,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 134
    },
    {
      "clip_ratio/high_max": 0.010988741800247226,
      "clip_ratio/high_mean": 0.0020738973378229275,
      "clip_ratio/low_mean": 0.002854833371429777,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004928730726533104,
      "epoch": 0.0669450879816521,
      "grad_norm": 0.11510859429836273,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 135
    },
    {
      "clip_ratio/high_max": 0.012867531868323567,
      "clip_ratio/high_mean": 0.0024248734043794684,
      "clip_ratio/low_mean": 0.003065843535523527,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005490716930580675,
      "epoch": 0.06744097752225692,
      "grad_norm": 0.10885921865701675,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 136
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0379464285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15184.0,
      "completions/mean_length": 1902.83935546875,
      "completions/mean_terminated_length": 1372.04638671875,
      "completions/min_length": 222.0,
      "completions/min_terminated_length": 222.0,
      "epoch": 0.06793686706286174,
      "grad_norm": 0.1765299290418625,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 15802165.0,
      "reward": 0.3482142984867096,
      "reward_std": 0.28685468435287476,
      "rewards/accuracy_reward/mean": 0.3482142984867096,
      "rewards/accuracy_reward/std": 0.476936936378479,
      "step": 137
    },
    {
      "clip_ratio/high_max": 0.00656084850015759,
      "clip_ratio/high_mean": 0.0016516139573923283,
      "clip_ratio/low_mean": 0.001317441662223473,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029690556084460695,
      "epoch": 0.06843275660346658,
      "grad_norm": 0.16446584463119507,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 138
    },
    {
      "clip_ratio/high_max": 0.007360659572441364,
      "clip_ratio/high_mean": 0.0017940284168389553,
      "clip_ratio/low_mean": 0.0014389272652692853,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032329556606782717,
      "epoch": 0.0689286461440714,
      "grad_norm": 0.15533778071403503,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 139
    },
    {
      "clip_ratio/high_max": 0.008355895321074058,
      "clip_ratio/high_mean": 0.001987478873843429,
      "clip_ratio/low_mean": 0.0016593675947547126,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036468464872996265,
      "epoch": 0.06942453568467624,
      "grad_norm": 0.15210993587970734,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 140
    },
    {
      "clip_ratio/high_max": 0.009706687178550055,
      "clip_ratio/high_mean": 0.0022564270698239852,
      "clip_ratio/low_mean": 0.0020667512713998804,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004323178345657652,
      "epoch": 0.06992042522528107,
      "grad_norm": 0.13742290437221527,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 141
    },
    {
      "clip_ratio/high_max": 0.010770832406706177,
      "clip_ratio/high_mean": 0.0025574721685188706,
      "clip_ratio/low_mean": 0.0024319721286474305,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004989444223610917,
      "epoch": 0.0704163147658859,
      "grad_norm": 0.13200795650482178,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 142
    },
    {
      "clip_ratio/high_max": 0.012138543283072067,
      "clip_ratio/high_mean": 0.002921526517638995,
      "clip_ratio/low_mean": 0.002733672767135431,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005655199312968762,
      "epoch": 0.07091220430649073,
      "grad_norm": 0.12562298774719238,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 143
    },
    {
      "clip_ratio/high_max": 0.014214933667972218,
      "clip_ratio/high_mean": 0.0032494584306732577,
      "clip_ratio/low_mean": 0.0031378071059862123,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006387265531884623,
      "epoch": 0.07140809384709557,
      "grad_norm": 0.1178639829158783,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 144
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.029017857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13183.0,
      "completions/mean_length": 1794.200927734375,
      "completions/mean_terminated_length": 1388.7862548828125,
      "completions/min_length": 270.0,
      "completions/min_terminated_length": 270.0,
      "epoch": 0.07190398338770039,
      "grad_norm": 0.12785916030406952,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 16673487.0,
      "reward": 0.345982164144516,
      "reward_std": 0.2312304973602295,
      "rewards/accuracy_reward/mean": 0.3459821343421936,
      "rewards/accuracy_reward/std": 0.47621920704841614,
      "step": 145
    },
    {
      "clip_ratio/high_max": 0.005667658603670134,
      "clip_ratio/high_mean": 0.0011716152689587034,
      "clip_ratio/low_mean": 0.0010792287866934203,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022508440445108135,
      "epoch": 0.07239987292830521,
      "grad_norm": 0.11652223765850067,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 146
    },
    {
      "clip_ratio/high_max": 0.005903063020014088,
      "clip_ratio/high_mean": 0.0012611674510480952,
      "clip_ratio/low_mean": 0.0011948179520686608,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024559854019798877,
      "epoch": 0.07289576246891005,
      "grad_norm": 0.11023937910795212,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 147
    },
    {
      "clip_ratio/high_max": 0.006627768269027001,
      "clip_ratio/high_mean": 0.0013412817672815436,
      "clip_ratio/low_mean": 0.0013894805957761491,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002730762352257443,
      "epoch": 0.07339165200951488,
      "grad_norm": 0.1087583526968956,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 148
    },
    {
      "clip_ratio/high_max": 0.006822516583270044,
      "clip_ratio/high_mean": 0.001435831715753011,
      "clip_ratio/low_mean": 0.0015483070070558824,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002984138702686323,
      "epoch": 0.07388754155011971,
      "grad_norm": 0.10296270996332169,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 149
    },
    {
      "clip_ratio/high_max": 0.007530759674409637,
      "clip_ratio/high_mean": 0.0016300205136303703,
      "clip_ratio/low_mean": 0.0017121012821235126,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033421218295188737,
      "epoch": 0.07438343109072454,
      "grad_norm": 0.09627677500247955,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 150
    },
    {
      "clip_ratio/high_max": 0.008544436637748731,
      "clip_ratio/high_mean": 0.0018113672149411286,
      "clip_ratio/low_mean": 0.002010485835057807,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038218530891072078,
      "epoch": 0.07487932063132938,
      "grad_norm": 0.09216029942035675,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 151
    },
    {
      "clip_ratio/high_max": 0.009978570858947933,
      "clip_ratio/high_mean": 0.0020761053995101975,
      "clip_ratio/low_mean": 0.002330005711087324,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004406111059779505,
      "epoch": 0.0753752101719342,
      "grad_norm": 0.09015058726072311,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 152
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0379464285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15090.0,
      "completions/mean_length": 1819.0469970703125,
      "completions/mean_terminated_length": 1284.9488525390625,
      "completions/min_length": 331.0,
      "completions/min_terminated_length": 331.0,
      "epoch": 0.07587109971253903,
      "grad_norm": 0.14622217416763306,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 17557212.0,
      "reward": 0.4352678656578064,
      "reward_std": 0.18900910019874573,
      "rewards/accuracy_reward/mean": 0.4352678656578064,
      "rewards/accuracy_reward/std": 0.49634629487991333,
      "step": 153
    },
    {
      "clip_ratio/high_max": 0.005659708353050519,
      "clip_ratio/high_mean": 0.0012030532968765328,
      "clip_ratio/low_mean": 0.0007908830376948117,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019939363342018623,
      "epoch": 0.07636698925314386,
      "grad_norm": 0.18139085173606873,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 154
    },
    {
      "clip_ratio/high_max": 0.006661090445049922,
      "clip_ratio/high_mean": 0.001401245363013004,
      "clip_ratio/low_mean": 0.0009376902754354433,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002338935664795372,
      "epoch": 0.07686287879374869,
      "grad_norm": 0.11290682107210159,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 155
    },
    {
      "clip_ratio/high_max": 0.006873811591503909,
      "clip_ratio/high_mean": 0.0014643671934209124,
      "clip_ratio/low_mean": 0.001134077395050781,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025984445952076385,
      "epoch": 0.07735876833435353,
      "grad_norm": 0.1077532097697258,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 156
    },
    {
      "clip_ratio/high_max": 0.008123619914840674,
      "clip_ratio/high_mean": 0.0016902349866541044,
      "clip_ratio/low_mean": 0.0012848384563994841,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002975073412500251,
      "epoch": 0.07785465787495835,
      "grad_norm": 0.10621393471956253,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 157
    },
    {
      "clip_ratio/high_max": 0.008777697756158886,
      "clip_ratio/high_mean": 0.0018088821177570935,
      "clip_ratio/low_mean": 0.0013984661144661459,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003207348259365972,
      "epoch": 0.07835054741556319,
      "grad_norm": 0.0961860865354538,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 158
    },
    {
      "clip_ratio/high_max": 0.009814135270971747,
      "clip_ratio/high_mean": 0.0020628888781857313,
      "clip_ratio/low_mean": 0.0016472077375340177,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037100965769241157,
      "epoch": 0.07884643695616801,
      "grad_norm": 0.09149108082056046,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 159
    },
    {
      "clip_ratio/high_max": 0.010815903877301025,
      "clip_ratio/high_mean": 0.0022960680737469374,
      "clip_ratio/low_mean": 0.0020123186797320614,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004308386724687807,
      "epoch": 0.07934232649677285,
      "grad_norm": 0.08270243555307388,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 160
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.033482142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15121.0,
      "completions/mean_length": 1806.5224609375,
      "completions/mean_terminated_length": 1337.0023193359375,
      "completions/min_length": 258.0,
      "completions/min_terminated_length": 258.0,
      "epoch": 0.07983821603737767,
      "grad_norm": 0.15443798899650574,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 18436686.0,
      "reward": 0.283482164144516,
      "reward_std": 0.21396119892597198,
      "rewards/accuracy_reward/mean": 0.2834821343421936,
      "rewards/accuracy_reward/std": 0.4511922299861908,
      "step": 161
    },
    {
      "clip_ratio/high_max": 0.005153531521500554,
      "clip_ratio/high_mean": 0.0010412932369945338,
      "clip_ratio/low_mean": 0.001401589744091325,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002442882968352933,
      "epoch": 0.0803341055779825,
      "grad_norm": 0.13129809498786926,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 162
    },
    {
      "clip_ratio/high_max": 0.005997318068693858,
      "clip_ratio/high_mean": 0.0012127539116590924,
      "clip_ratio/low_mean": 0.001655648659379949,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028684025674010627,
      "epoch": 0.08082999511858734,
      "grad_norm": 0.11930796504020691,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 163
    },
    {
      "clip_ratio/high_max": 0.006631928679780685,
      "clip_ratio/high_mean": 0.0013359524418774527,
      "clip_ratio/low_mean": 0.0018817483114617062,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032177007979043992,
      "epoch": 0.08132588465919216,
      "grad_norm": 0.12321930378675461,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 164
    },
    {
      "clip_ratio/high_max": 0.007985461063071853,
      "clip_ratio/high_mean": 0.0015533415803474782,
      "clip_ratio/low_mean": 0.002046354326466826,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003599695894990873,
      "epoch": 0.081821774199797,
      "grad_norm": 0.12315398454666138,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 165
    },
    {
      "clip_ratio/high_max": 0.00882659627313842,
      "clip_ratio/high_mean": 0.0016988151924124395,
      "clip_ratio/low_mean": 0.0022770738921735756,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039758891025485354,
      "epoch": 0.08231766374040182,
      "grad_norm": 0.10716372728347778,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 166
    },
    {
      "clip_ratio/high_max": 0.01047008516616188,
      "clip_ratio/high_mean": 0.002028975052780879,
      "clip_ratio/low_mean": 0.0025866571995720733,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004615632269633352,
      "epoch": 0.08281355328100666,
      "grad_norm": 0.09270600229501724,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 167
    },
    {
      "clip_ratio/high_max": 0.012080547690857202,
      "clip_ratio/high_mean": 0.0022486634725282784,
      "clip_ratio/low_mean": 0.002896460116517119,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005145123561305809,
      "epoch": 0.08330944282161148,
      "grad_norm": 0.09748215228319168,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 168
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15226.0,
      "completions/mean_length": 2089.96875,
      "completions/mean_terminated_length": 1437.34423828125,
      "completions/min_length": 228.0,
      "completions/min_terminated_length": 228.0,
      "epoch": 0.08380533236221632,
      "grad_norm": 0.14761880040168762,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 19439696.0,
      "reward": 0.3325892984867096,
      "reward_std": 0.26370424032211304,
      "rewards/accuracy_reward/mean": 0.3325892984867096,
      "rewards/accuracy_reward/std": 0.47166749835014343,
      "step": 169
    },
    {
      "clip_ratio/high_max": 0.005630264877254376,
      "clip_ratio/high_mean": 0.0011459417504511293,
      "clip_ratio/low_mean": 0.001338898014410006,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002484839749740786,
      "epoch": 0.08430122190282115,
      "grad_norm": 0.1363615095615387,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 170
    },
    {
      "clip_ratio/high_max": 0.006470865751907695,
      "clip_ratio/high_mean": 0.001335740195145263,
      "clip_ratio/low_mean": 0.0016047543593344926,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00294049454623746,
      "epoch": 0.08479711144342597,
      "grad_norm": 0.13255184888839722,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 171
    },
    {
      "clip_ratio/high_max": 0.006610017265302304,
      "clip_ratio/high_mean": 0.0013500437295306256,
      "clip_ratio/low_mean": 0.0017538818549382995,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031039255918585695,
      "epoch": 0.08529300098403081,
      "grad_norm": 0.132757306098938,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 172
    },
    {
      "clip_ratio/high_max": 0.008400868724493193,
      "clip_ratio/high_mean": 0.0017124871517353313,
      "clip_ratio/low_mean": 0.0019818435508227594,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036943306868124637,
      "epoch": 0.08578889052463563,
      "grad_norm": 0.12050355970859528,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 173
    },
    {
      "clip_ratio/high_max": 0.009358600467749056,
      "clip_ratio/high_mean": 0.0018814588897839712,
      "clip_ratio/low_mean": 0.0024512652025805437,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004332724087362294,
      "epoch": 0.08628478006524047,
      "grad_norm": 0.1064835637807846,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 174
    },
    {
      "clip_ratio/high_max": 0.010181480456594727,
      "clip_ratio/high_mean": 0.0021284094955262844,
      "clip_ratio/low_mean": 0.0028606591108655266,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004989068635040894,
      "epoch": 0.0867806696058453,
      "grad_norm": 0.10880880802869797,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 175
    },
    {
      "clip_ratio/high_max": 0.011471953344880603,
      "clip_ratio/high_mean": 0.00243211564225021,
      "clip_ratio/low_mean": 0.0032352062512472912,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005667321849614382,
      "epoch": 0.08727655914645013,
      "grad_norm": 0.09929163753986359,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 176
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.033482142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14784.0,
      "completions/mean_length": 1697.0535888671875,
      "completions/mean_terminated_length": 1223.7413330078125,
      "completions/min_length": 184.0,
      "completions/min_terminated_length": 184.0,
      "epoch": 0.08777244868705496,
      "grad_norm": 0.15766136348247528,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 20270936.0,
      "reward": 0.4196428656578064,
      "reward_std": 0.21695247292518616,
      "rewards/accuracy_reward/mean": 0.4196428656578064,
      "rewards/accuracy_reward/std": 0.4940522015094757,
      "step": 177
    },
    {
      "clip_ratio/high_max": 0.00575263189512043,
      "clip_ratio/high_mean": 0.0012434693487648474,
      "clip_ratio/low_mean": 0.0009674090466376128,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002210878395999316,
      "epoch": 0.0882683382276598,
      "grad_norm": 0.14676572382450104,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 178
    },
    {
      "clip_ratio/high_max": 0.006333830430776288,
      "clip_ratio/high_mean": 0.001378407736410736,
      "clip_ratio/low_mean": 0.0010793070854901998,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002457714815136569,
      "epoch": 0.08876422776826462,
      "grad_norm": 0.13755740225315094,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 179
    },
    {
      "clip_ratio/high_max": 0.0071948116292333,
      "clip_ratio/high_mean": 0.0015607275802267395,
      "clip_ratio/low_mean": 0.0013942379896434431,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029549655678238196,
      "epoch": 0.08926011730886944,
      "grad_norm": 0.12935592234134674,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 180
    },
    {
      "clip_ratio/high_max": 0.008605137229096727,
      "clip_ratio/high_mean": 0.0018285875705714716,
      "clip_ratio/low_mean": 0.001623936208119403,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003452523767919047,
      "epoch": 0.08975600684947428,
      "grad_norm": 0.11166780441999435,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 181
    },
    {
      "clip_ratio/high_max": 0.010663332741387421,
      "clip_ratio/high_mean": 0.0022009146337040875,
      "clip_ratio/low_mean": 0.002063667196694041,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004264581843017368,
      "epoch": 0.0902518963900791,
      "grad_norm": 0.1055390015244484,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 182
    },
    {
      "clip_ratio/high_max": 0.011422514293371933,
      "clip_ratio/high_mean": 0.002359501293994981,
      "clip_ratio/low_mean": 0.0024237451528961174,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004783246448823775,
      "epoch": 0.09074778593068394,
      "grad_norm": 0.09475918114185333,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 183
    },
    {
      "clip_ratio/high_max": 0.01339945546715171,
      "clip_ratio/high_mean": 0.0027737269019780797,
      "clip_ratio/low_mean": 0.0027750119719485156,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005548738890865934,
      "epoch": 0.09124367547128877,
      "grad_norm": 0.09180509299039841,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 184
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0424107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14065.0,
      "completions/mean_length": 1991.997802734375,
      "completions/mean_terminated_length": 1399.9417724609375,
      "completions/min_length": 261.0,
      "completions/min_terminated_length": 261.0,
      "epoch": 0.0917395650118936,
      "grad_norm": 0.15154431760311127,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 21231175.0,
      "reward": 0.4196428656578064,
      "reward_std": 0.26498445868492126,
      "rewards/accuracy_reward/mean": 0.4196428656578064,
      "rewards/accuracy_reward/std": 0.4940522015094757,
      "step": 185
    },
    {
      "clip_ratio/high_max": 0.006591190207473119,
      "clip_ratio/high_mean": 0.00147313250249681,
      "clip_ratio/low_mean": 0.001192055610601983,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002665188089849835,
      "epoch": 0.09223545455249843,
      "grad_norm": 0.13247278332710266,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 186
    },
    {
      "clip_ratio/high_max": 0.006785153800592525,
      "clip_ratio/high_mean": 0.0015203524817479774,
      "clip_ratio/low_mean": 0.0014185176568588531,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029388701395873795,
      "epoch": 0.09273134409310325,
      "grad_norm": 0.13051877915859222,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 187
    },
    {
      "clip_ratio/high_max": 0.007997232298293966,
      "clip_ratio/high_mean": 0.0017625569546453335,
      "clip_ratio/low_mean": 0.0016067795538958762,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033693364912323887,
      "epoch": 0.09322723363370809,
      "grad_norm": 0.12517009675502777,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 188
    },
    {
      "clip_ratio/high_max": 0.0091287037521397,
      "clip_ratio/high_mean": 0.002053121275821468,
      "clip_ratio/low_mean": 0.0018726336868581939,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003925754996089381,
      "epoch": 0.09372312317431292,
      "grad_norm": 0.11099936813116074,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 189
    },
    {
      "clip_ratio/high_max": 0.010884680828894489,
      "clip_ratio/high_mean": 0.0024277166403408046,
      "clip_ratio/low_mean": 0.002229206206749268,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004656922905269312,
      "epoch": 0.09421901271491775,
      "grad_norm": 0.1030183732509613,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 190
    },
    {
      "clip_ratio/high_max": 0.012303421422984684,
      "clip_ratio/high_mean": 0.0027457725659587595,
      "clip_ratio/low_mean": 0.002810156991529311,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005555929565161932,
      "epoch": 0.09471490225552258,
      "grad_norm": 0.09813059866428375,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 191
    },
    {
      "clip_ratio/high_max": 0.014568056551070185,
      "clip_ratio/high_mean": 0.0032373372550864588,
      "clip_ratio/low_mean": 0.0032348169780220815,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006472154265793506,
      "epoch": 0.09521079179612742,
      "grad_norm": 0.09101764857769012,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "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": 0.060267857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15227.0,
      "completions/mean_length": 2662.696533203125,
      "completions/mean_terminated_length": 1848.380126953125,
      "completions/min_length": 266.0,
      "completions/min_terminated_length": 266.0,
      "epoch": 0.09570668133673224,
      "grad_norm": 0.13822278380393982,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 22495663.0,
      "reward": 0.283482164144516,
      "reward_std": 0.22665229439735413,
      "rewards/accuracy_reward/mean": 0.2834821343421936,
      "rewards/accuracy_reward/std": 0.4511922299861908,
      "step": 193
    },
    {
      "clip_ratio/high_max": 0.004764056209751288,
      "clip_ratio/high_mean": 0.0009360919843857118,
      "clip_ratio/low_mean": 0.0012738735700850157,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002209965570841632,
      "epoch": 0.09620257087733708,
      "grad_norm": 0.12064450979232788,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 194
    },
    {
      "clip_ratio/high_max": 0.005409542414781754,
      "clip_ratio/high_mean": 0.00105767305876725,
      "clip_ratio/low_mean": 0.0014943578379984501,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025520308981867856,
      "epoch": 0.0966984604179419,
      "grad_norm": 0.11369688808917999,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 195
    },
    {
      "clip_ratio/high_max": 0.006081661211283063,
      "clip_ratio/high_mean": 0.0011815820580522995,
      "clip_ratio/low_mean": 0.0017916806218636339,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029732626926488592,
      "epoch": 0.09719434995854673,
      "grad_norm": 0.10454533994197845,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 196
    },
    {
      "clip_ratio/high_max": 0.007153306702093687,
      "clip_ratio/high_mean": 0.0013967678351036739,
      "clip_ratio/low_mean": 0.001976817574359302,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033735853985490394,
      "epoch": 0.09769023949915157,
      "grad_norm": 0.10156543552875519,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 197
    },
    {
      "clip_ratio/high_max": 0.008660263092679088,
      "clip_ratio/high_mean": 0.001650839888043265,
      "clip_ratio/low_mean": 0.002408119866117886,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0040589597265352495,
      "epoch": 0.09818612903975639,
      "grad_norm": 0.09431879967451096,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 198
    },
    {
      "clip_ratio/high_max": 0.009787996641534846,
      "clip_ratio/high_mean": 0.0018413710522509064,
      "clip_ratio/low_mean": 0.0027344752028284347,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004575846260195249,
      "epoch": 0.09868201858036123,
      "grad_norm": 0.08861619979143143,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 199
    },
    {
      "clip_ratio/high_max": 0.011136546323541552,
      "clip_ratio/high_mean": 0.0021025219698458386,
      "clip_ratio/low_mean": 0.0034286200001361067,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005531141980100074,
      "epoch": 0.09917790812096605,
      "grad_norm": 0.08138272911310196,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 200
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.049107142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15264.0,
      "completions/mean_length": 2304.071533203125,
      "completions/mean_terminated_length": 1629.8216552734375,
      "completions/min_length": 172.0,
      "completions/min_terminated_length": 172.0,
      "epoch": 0.09967379766157089,
      "grad_norm": 0.1615176647901535,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 23596775.0,
      "reward": 0.4174107313156128,
      "reward_std": 0.24799828231334686,
      "rewards/accuracy_reward/mean": 0.4174107015132904,
      "rewards/accuracy_reward/std": 0.4936830997467041,
      "step": 201
    },
    {
      "clip_ratio/high_max": 0.00633759233278397,
      "clip_ratio/high_mean": 0.001436032823903588,
      "clip_ratio/low_mean": 0.0011868806570873858,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002622913483719458,
      "epoch": 0.10016968720217571,
      "grad_norm": 0.13745781779289246,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 202
    },
    {
      "clip_ratio/high_max": 0.008035448312512017,
      "clip_ratio/high_mean": 0.0017341734160254418,
      "clip_ratio/low_mean": 0.001380968984676656,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003115142387287051,
      "epoch": 0.10066557674278055,
      "grad_norm": 0.13050714135169983,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 203
    },
    {
      "clip_ratio/high_max": 0.008869542751199333,
      "clip_ratio/high_mean": 0.0019407469362704433,
      "clip_ratio/low_mean": 0.0017499529437827732,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036906998948325054,
      "epoch": 0.10116146628338538,
      "grad_norm": 0.12193664908409119,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 204
    },
    {
      "clip_ratio/high_max": 0.01065450461828732,
      "clip_ratio/high_mean": 0.00234058579098928,
      "clip_ratio/low_mean": 0.0020613652645806724,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004401951035106322,
      "epoch": 0.1016573558239902,
      "grad_norm": 0.11155785620212555,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 205
    },
    {
      "clip_ratio/high_max": 0.011966985537583241,
      "clip_ratio/high_mean": 0.002659810371369531,
      "clip_ratio/low_mean": 0.0025068984941754024,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005166708857359481,
      "epoch": 0.10215324536459504,
      "grad_norm": 0.10905234515666962,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 206
    },
    {
      "clip_ratio/high_max": 0.013504636248399038,
      "clip_ratio/high_mean": 0.0030254417856667715,
      "clip_ratio/low_mean": 0.002962806735013146,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00598824858025182,
      "epoch": 0.10264913490519986,
      "grad_norm": 0.10773346573114395,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 207
    },
    {
      "clip_ratio/high_max": 0.015613143368682358,
      "clip_ratio/high_mean": 0.0034018258711512317,
      "clip_ratio/low_mean": 0.0033973643548961263,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00679919026515563,
      "epoch": 0.1031450244458047,
      "grad_norm": 0.09303463995456696,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 208
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.060267857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15317.0,
      "completions/mean_length": 2619.65185546875,
      "completions/mean_terminated_length": 1802.574951171875,
      "completions/min_length": 154.0,
      "completions/min_terminated_length": 154.0,
      "epoch": 0.10364091398640952,
      "grad_norm": 0.19247539341449738,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 24836795.0,
      "reward": 0.4196428656578064,
      "reward_std": 0.23641349375247955,
      "rewards/accuracy_reward/mean": 0.4196428656578064,
      "rewards/accuracy_reward/std": 0.4940522015094757,
      "step": 209
    },
    {
      "clip_ratio/high_max": 0.007083041306032101,
      "clip_ratio/high_mean": 0.001665385589831203,
      "clip_ratio/low_mean": 0.0008557677440990119,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025211533347828663,
      "epoch": 0.10413680352701436,
      "grad_norm": 0.16681571304798126,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 210
    },
    {
      "clip_ratio/high_max": 0.007800049163051881,
      "clip_ratio/high_mean": 0.0018590454365039477,
      "clip_ratio/low_mean": 0.0011530675674862323,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030121130012048525,
      "epoch": 0.10463269306761919,
      "grad_norm": 0.1375265121459961,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 211
    },
    {
      "clip_ratio/high_max": 0.008686238548762049,
      "clip_ratio/high_mean": 0.0020515664893991925,
      "clip_ratio/low_mean": 0.0013481640011434592,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033997304699369124,
      "epoch": 0.10512858260822402,
      "grad_norm": 0.12510205805301666,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 212
    },
    {
      "clip_ratio/high_max": 0.009739783001350588,
      "clip_ratio/high_mean": 0.0023990972099454666,
      "clip_ratio/low_mean": 0.0017181543503284047,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004117251572097302,
      "epoch": 0.10562447214882885,
      "grad_norm": 0.11066936701536179,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 213
    },
    {
      "clip_ratio/high_max": 0.011630541735939914,
      "clip_ratio/high_mean": 0.002824748571129021,
      "clip_ratio/low_mean": 0.002076585936549691,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004901334498754295,
      "epoch": 0.10612036168943367,
      "grad_norm": 0.10947967320680618,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 214
    },
    {
      "clip_ratio/high_max": 0.013071252491499763,
      "clip_ratio/high_mean": 0.0031906635904306313,
      "clip_ratio/low_mean": 0.0024207685697206216,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00561143216509663,
      "epoch": 0.10661625123003851,
      "grad_norm": 0.10359339416027069,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 215
    },
    {
      "clip_ratio/high_max": 0.014569470284186536,
      "clip_ratio/high_mean": 0.003592358571040677,
      "clip_ratio/low_mean": 0.0028184628851022353,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006410821497411234,
      "epoch": 0.10711214077064334,
      "grad_norm": 0.09484237432479858,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 216
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13722.0,
      "completions/mean_length": 1805.7366943359375,
      "completions/mean_terminated_length": 1368.5023193359375,
      "completions/min_length": 253.0,
      "completions/min_terminated_length": 253.0,
      "epoch": 0.10760803031124817,
      "grad_norm": 0.17061293125152588,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 25714629.0,
      "reward": 0.4665178656578064,
      "reward_std": 0.26587599515914917,
      "rewards/accuracy_reward/mean": 0.4665178656578064,
      "rewards/accuracy_reward/std": 0.4994353950023651,
      "step": 217
    },
    {
      "clip_ratio/high_max": 0.007145964798837667,
      "clip_ratio/high_mean": 0.0016563646606755356,
      "clip_ratio/low_mean": 0.0013217935525204894,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029781582143186824,
      "epoch": 0.108103919851853,
      "grad_norm": 0.15041764080524445,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 218
    },
    {
      "clip_ratio/high_max": 0.007984811742062448,
      "clip_ratio/high_mean": 0.0018248687426876131,
      "clip_ratio/low_mean": 0.0014801340780792316,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003305002824163239,
      "epoch": 0.10859980939245784,
      "grad_norm": 0.14258238673210144,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 219
    },
    {
      "clip_ratio/high_max": 0.00902629775555397,
      "clip_ratio/high_mean": 0.0020472174894621276,
      "clip_ratio/low_mean": 0.0018085287131555106,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003855746221233858,
      "epoch": 0.10909569893306266,
      "grad_norm": 0.13006500899791718,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 220
    },
    {
      "clip_ratio/high_max": 0.010611409627017565,
      "clip_ratio/high_mean": 0.002458063800986565,
      "clip_ratio/low_mean": 0.002287096044511827,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004745159879348648,
      "epoch": 0.10959158847366748,
      "grad_norm": 0.1230592131614685,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 221
    },
    {
      "clip_ratio/high_max": 0.012497835068643326,
      "clip_ratio/high_mean": 0.0028276503267079534,
      "clip_ratio/low_mean": 0.0026617833075590625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005489433690854639,
      "epoch": 0.11008747801427232,
      "grad_norm": 0.10992691665887833,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 222
    },
    {
      "clip_ratio/high_max": 0.015048458997625858,
      "clip_ratio/high_mean": 0.003359110537530796,
      "clip_ratio/low_mean": 0.003259662740248359,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006618773289119417,
      "epoch": 0.11058336755487715,
      "grad_norm": 0.10510656982660294,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 223
    },
    {
      "clip_ratio/high_max": 0.016616192733636126,
      "clip_ratio/high_mean": 0.003702498448546976,
      "clip_ratio/low_mean": 0.003826247535727134,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007528745993113262,
      "epoch": 0.11107925709548198,
      "grad_norm": 0.09582949429750443,
      "learning_rate": 1e-06,
      "loss": -0.0017,
      "step": 224
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0580357142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15064.0,
      "completions/mean_length": 2377.473388671875,
      "completions/mean_terminated_length": 1577.6019287109375,
      "completions/min_length": 249.0,
      "completions/min_terminated_length": 249.0,
      "epoch": 0.11157514663608681,
      "grad_norm": 0.14727763831615448,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 26848633.0,
      "reward": 0.4263392984867096,
      "reward_std": 0.1816357970237732,
      "rewards/accuracy_reward/mean": 0.4263392984867096,
      "rewards/accuracy_reward/std": 0.49509719014167786,
      "step": 225
    },
    {
      "clip_ratio/high_max": 0.005575898368988419,
      "clip_ratio/high_mean": 0.0011847607534036797,
      "clip_ratio/low_mean": 0.0007495374159418589,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019342981500187761,
      "epoch": 0.11207103617669165,
      "grad_norm": 0.13605214655399323,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 226
    },
    {
      "clip_ratio/high_max": 0.006962522710637131,
      "clip_ratio/high_mean": 0.0014650205378075043,
      "clip_ratio/low_mean": 0.0009693835000916806,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002434404029855841,
      "epoch": 0.11256692571729647,
      "grad_norm": 0.11827627569437027,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 227
    },
    {
      "clip_ratio/high_max": 0.008523965123458765,
      "clip_ratio/high_mean": 0.0017352990580548067,
      "clip_ratio/low_mean": 0.0013286902202196416,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030639892488864007,
      "epoch": 0.11306281525790131,
      "grad_norm": 0.10426940023899078,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 228
    },
    {
      "clip_ratio/high_max": 0.009379505434480961,
      "clip_ratio/high_mean": 0.0018777082577798865,
      "clip_ratio/low_mean": 0.001659053401994015,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003536761659717058,
      "epoch": 0.11355870479850613,
      "grad_norm": 0.09800033271312714,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 229
    },
    {
      "clip_ratio/high_max": 0.010569351419690065,
      "clip_ratio/high_mean": 0.0021632709003824857,
      "clip_ratio/low_mean": 0.0018823143674637777,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004045585295926912,
      "epoch": 0.11405459433911096,
      "grad_norm": 0.09491190314292908,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 230
    },
    {
      "clip_ratio/high_max": 0.011878590703418013,
      "clip_ratio/high_mean": 0.0024436759858872392,
      "clip_ratio/low_mean": 0.0022841379864928513,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004727813969140016,
      "epoch": 0.1145504838797158,
      "grad_norm": 0.08002258092164993,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 231
    },
    {
      "clip_ratio/high_max": 0.013594871452369262,
      "clip_ratio/high_mean": 0.002824678500473965,
      "clip_ratio/low_mean": 0.0025918860286111567,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0054165645779136185,
      "epoch": 0.11504637342032062,
      "grad_norm": 0.0758328065276146,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 232
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0401785714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14540.0,
      "completions/mean_length": 2183.19873046875,
      "completions/mean_terminated_length": 1631.611572265625,
      "completions/min_length": 368.0,
      "completions/min_terminated_length": 368.0,
      "epoch": 0.11554226296092546,
      "grad_norm": 0.14348861575126648,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 27892362.0,
      "reward": 0.4151785969734192,
      "reward_std": 0.2311626821756363,
      "rewards/accuracy_reward/mean": 0.4151785671710968,
      "rewards/accuracy_reward/std": 0.49330368638038635,
      "step": 233
    },
    {
      "clip_ratio/high_max": 0.004572688825646765,
      "clip_ratio/high_mean": 0.0009770175120138447,
      "clip_ratio/low_mean": 0.0011343072013403344,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021113247285597936,
      "epoch": 0.11603815250153028,
      "grad_norm": 0.12673529982566833,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 234
    },
    {
      "clip_ratio/high_max": 0.0051074605053145206,
      "clip_ratio/high_mean": 0.0010825664833191695,
      "clip_ratio/low_mean": 0.0012781789196480986,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023607454586738186,
      "epoch": 0.11653404204213512,
      "grad_norm": 0.12059280276298523,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 235
    },
    {
      "clip_ratio/high_max": 0.0063248513943108264,
      "clip_ratio/high_mean": 0.0013654494150614482,
      "clip_ratio/low_mean": 0.001622831374618272,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029882808330796706,
      "epoch": 0.11702993158273994,
      "grad_norm": 0.10548672080039978,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 236
    },
    {
      "clip_ratio/high_max": 0.007766624581563519,
      "clip_ratio/high_mean": 0.0016655040046771319,
      "clip_ratio/low_mean": 0.0019972942585297915,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036627982813683957,
      "epoch": 0.11752582112334478,
      "grad_norm": 0.09665427356958389,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 237
    },
    {
      "clip_ratio/high_max": 0.009213452774929465,
      "clip_ratio/high_mean": 0.001968944126019778,
      "clip_ratio/low_mean": 0.0023616348956068123,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004330579014606428,
      "epoch": 0.1180217106639496,
      "grad_norm": 0.09623239189386368,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 238
    },
    {
      "clip_ratio/high_max": 0.010335083772588405,
      "clip_ratio/high_mean": 0.0022035732531549,
      "clip_ratio/low_mean": 0.002697875138778727,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004901448340717707,
      "epoch": 0.11851760020455443,
      "grad_norm": 0.08508098870515823,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 239
    },
    {
      "clip_ratio/high_max": 0.011828866006908356,
      "clip_ratio/high_mean": 0.002530613350245403,
      "clip_ratio/low_mean": 0.0030383812668901555,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005568994655845927,
      "epoch": 0.11901348974515927,
      "grad_norm": 0.07687724381685257,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 240
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0647321428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14666.0,
      "completions/mean_length": 2732.21435546875,
      "completions/mean_terminated_length": 1858.21484375,
      "completions/min_length": 224.0,
      "completions/min_terminated_length": 224.0,
      "epoch": 0.11950937928576409,
      "grad_norm": 0.1774415820837021,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 29187722.0,
      "reward": 0.3660714328289032,
      "reward_std": 0.28954288363456726,
      "rewards/accuracy_reward/mean": 0.3660714328289032,
      "rewards/accuracy_reward/std": 0.482267826795578,
      "step": 241
    },
    {
      "clip_ratio/high_max": 0.006568017310200958,
      "clip_ratio/high_mean": 0.0015881573935985216,
      "clip_ratio/low_mean": 0.0014422406773064722,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003030398062946915,
      "epoch": 0.12000526882636893,
      "grad_norm": 0.1478789895772934,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 242
    },
    {
      "clip_ratio/high_max": 0.007990589881956112,
      "clip_ratio/high_mean": 0.0018287813832102984,
      "clip_ratio/low_mean": 0.0018121727773632301,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003640954165803123,
      "epoch": 0.12050115836697375,
      "grad_norm": 0.14355163276195526,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 243
    },
    {
      "clip_ratio/high_max": 0.009448767112189671,
      "clip_ratio/high_mean": 0.0022383080922736553,
      "clip_ratio/low_mean": 0.00216435350125721,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004402661648327921,
      "epoch": 0.12099704790757859,
      "grad_norm": 0.1309833526611328,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 244
    },
    {
      "clip_ratio/high_max": 0.011090917421824997,
      "clip_ratio/high_mean": 0.0025824567819654476,
      "clip_ratio/low_mean": 0.002770778725107448,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005353235530492384,
      "epoch": 0.12149293744818342,
      "grad_norm": 0.11801817268133163,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 245
    },
    {
      "clip_ratio/high_max": 0.012930098298966186,
      "clip_ratio/high_mean": 0.0030115190556898597,
      "clip_ratio/low_mean": 0.003448758773629379,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006460277739279263,
      "epoch": 0.12198882698878825,
      "grad_norm": 0.1052909791469574,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 246
    },
    {
      "clip_ratio/high_max": 0.01438806414080318,
      "clip_ratio/high_mean": 0.0034677208814173355,
      "clip_ratio/low_mean": 0.004076660541159072,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00754438139119884,
      "epoch": 0.12248471652939308,
      "grad_norm": 0.09513720870018005,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 247
    },
    {
      "clip_ratio/high_max": 0.016224967323068995,
      "clip_ratio/high_mean": 0.0038587773251492763,
      "clip_ratio/low_mean": 0.004803074750498126,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.008661852046316199,
      "epoch": 0.1229806060699979,
      "grad_norm": 0.11291947960853577,
      "learning_rate": 1e-06,
      "loss": -0.0017,
      "step": 248
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15011.0,
      "completions/mean_length": 2719.716552734375,
      "completions/mean_terminated_length": 1877.031005859375,
      "completions/min_length": 322.0,
      "completions/min_terminated_length": 322.0,
      "epoch": 0.12347649561060274,
      "grad_norm": 0.15207715332508087,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 30474147.0,
      "reward": 0.3973214328289032,
      "reward_std": 0.23912444710731506,
      "rewards/accuracy_reward/mean": 0.3973214328289032,
      "rewards/accuracy_reward/std": 0.48989057540893555,
      "step": 249
    },
    {
      "clip_ratio/high_max": 0.005877129322470864,
      "clip_ratio/high_mean": 0.001184506365461857,
      "clip_ratio/low_mean": 0.0011675611808072972,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002352067525862367,
      "epoch": 0.12397238515120756,
      "grad_norm": 0.13811887800693512,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 250
    },
    {
      "clip_ratio/high_max": 0.00637557283971546,
      "clip_ratio/high_mean": 0.0012599111664712837,
      "clip_ratio/low_mean": 0.0013449626113697377,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026048737818200607,
      "epoch": 0.1244682746918124,
      "grad_norm": 0.11866877228021622,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 251
    },
    {
      "clip_ratio/high_max": 0.008030319178942591,
      "clip_ratio/high_mean": 0.001583187975029432,
      "clip_ratio/low_mean": 0.0018804646097123623,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034636525942914886,
      "epoch": 0.12496416423241723,
      "grad_norm": 0.10853729397058487,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 252
    },
    {
      "clip_ratio/high_max": 0.009370527717692312,
      "clip_ratio/high_mean": 0.0018263227671013738,
      "clip_ratio/low_mean": 0.0023786711109323733,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042049938874697546,
      "epoch": 0.12546005377302205,
      "grad_norm": 0.09561814367771149,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 253
    },
    {
      "clip_ratio/high_max": 0.011312071877910057,
      "clip_ratio/high_mean": 0.0021609061409435526,
      "clip_ratio/low_mean": 0.0028486353451171453,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005009541506296955,
      "epoch": 0.1259559433136269,
      "grad_norm": 0.09443897753953934,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 254
    },
    {
      "clip_ratio/high_max": 0.013235370259280899,
      "clip_ratio/high_mean": 0.002532104238980537,
      "clip_ratio/low_mean": 0.0033094211335082946,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005841525347932475,
      "epoch": 0.12645183285423173,
      "grad_norm": 0.08161179721355438,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 255
    },
    {
      "clip_ratio/high_max": 0.014080045182709,
      "clip_ratio/high_mean": 0.00272944167818423,
      "clip_ratio/low_mean": 0.0037545617026353284,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006484003446530551,
      "epoch": 0.12694772239483654,
      "grad_norm": 0.07649612426757812,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "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": 0.0580357142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14235.0,
      "completions/mean_length": 2712.3662109375,
      "completions/mean_terminated_length": 1933.1280517578125,
      "completions/min_length": 312.0,
      "completions/min_terminated_length": 312.0,
      "epoch": 0.12744361193544138,
      "grad_norm": 0.15578414499759674,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 31758015.0,
      "reward": 0.4330357313156128,
      "reward_std": 0.26039132475852966,
      "rewards/accuracy_reward/mean": 0.4330357015132904,
      "rewards/accuracy_reward/std": 0.4960494041442871,
      "step": 257
    },
    {
      "clip_ratio/high_max": 0.006448145077229128,
      "clip_ratio/high_mean": 0.0013809574429615168,
      "clip_ratio/low_mean": 0.0012016006953103897,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025825581160461297,
      "epoch": 0.1279395014760462,
      "grad_norm": 0.14260150492191315,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 258
    },
    {
      "clip_ratio/high_max": 0.007452258385455934,
      "clip_ratio/high_mean": 0.001590196805409505,
      "clip_ratio/low_mean": 0.0015059718758152485,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030961687280068872,
      "epoch": 0.12843539101665105,
      "grad_norm": 0.1301211416721344,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 259
    },
    {
      "clip_ratio/high_max": 0.008949931881943485,
      "clip_ratio/high_mean": 0.0019565254533517873,
      "clip_ratio/low_mean": 0.0019264169530401887,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038829424574942095,
      "epoch": 0.12893128055725586,
      "grad_norm": 0.1416904479265213,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 260
    },
    {
      "clip_ratio/high_max": 0.010612619855237426,
      "clip_ratio/high_mean": 0.0023335655573646363,
      "clip_ratio/low_mean": 0.002526175813102327,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004859741377003957,
      "epoch": 0.1294271700978607,
      "grad_norm": 0.1125275120139122,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 261
    },
    {
      "clip_ratio/high_max": 0.012662957960856147,
      "clip_ratio/high_mean": 0.002749044522715849,
      "clip_ratio/low_mean": 0.0030092436793722754,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005758288187280414,
      "epoch": 0.12992305963846554,
      "grad_norm": 0.10432489961385727,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 262
    },
    {
      "clip_ratio/high_max": 0.014124593635642668,
      "clip_ratio/high_mean": 0.0030784265718466486,
      "clip_ratio/low_mean": 0.0034766162610821993,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006555042829859303,
      "epoch": 0.13041894917907038,
      "grad_norm": 0.09645858407020569,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 263
    },
    {
      "clip_ratio/high_max": 0.01616412650764687,
      "clip_ratio/high_mean": 0.0035450603327262797,
      "clip_ratio/low_mean": 0.004020521693121282,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007565581970993662,
      "epoch": 0.1309148387196752,
      "grad_norm": 0.08041694015264511,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 264
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0870535714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15357.0,
      "completions/mean_length": 3549.055908203125,
      "completions/mean_terminated_length": 2422.828857421875,
      "completions/min_length": 321.0,
      "completions/min_terminated_length": 321.0,
      "epoch": 0.13141072826028002,
      "grad_norm": 0.1444324254989624,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 33423984.0,
      "reward": 0.3906250298023224,
      "reward_std": 0.2482999712228775,
      "rewards/accuracy_reward/mean": 0.390625,
      "rewards/accuracy_reward/std": 0.48843589425086975,
      "step": 265
    },
    {
      "clip_ratio/high_max": 0.006377484636686859,
      "clip_ratio/high_mean": 0.0015167708909302746,
      "clip_ratio/low_mean": 0.0008180464966471845,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002334817379960441,
      "epoch": 0.13190661780088486,
      "grad_norm": 0.12835930287837982,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 266
    },
    {
      "clip_ratio/high_max": 0.006809843729570275,
      "clip_ratio/high_mean": 0.0017315928630523558,
      "clip_ratio/low_mean": 0.0010637097186645406,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002795302594677196,
      "epoch": 0.13240250734148967,
      "grad_norm": 0.12246652692556381,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 267
    },
    {
      "clip_ratio/high_max": 0.008456480498352903,
      "clip_ratio/high_mean": 0.0021246022977265966,
      "clip_ratio/low_mean": 0.0013383543373493012,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003462956673502049,
      "epoch": 0.1328983968820945,
      "grad_norm": 0.09665338695049286,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 268
    },
    {
      "clip_ratio/high_max": 0.009697665147541557,
      "clip_ratio/high_mean": 0.002419994297042649,
      "clip_ratio/low_mean": 0.0016239907669159948,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004043985070893541,
      "epoch": 0.13339428642269935,
      "grad_norm": 0.09391030669212341,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 269
    },
    {
      "clip_ratio/high_max": 0.010523427179577993,
      "clip_ratio/high_mean": 0.00264090272185058,
      "clip_ratio/low_mean": 0.002030432144238148,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0046713348529010545,
      "epoch": 0.1338901759633042,
      "grad_norm": 0.08494468778371811,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 270
    },
    {
      "clip_ratio/high_max": 0.012809070940420497,
      "clip_ratio/high_mean": 0.0031614465933671454,
      "clip_ratio/low_mean": 0.002396356242343245,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00555780277136364,
      "epoch": 0.134386065503909,
      "grad_norm": 0.08098793774843216,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 271
    },
    {
      "clip_ratio/high_max": 0.014248309813410742,
      "clip_ratio/high_mean": 0.003558117498869251,
      "clip_ratio/low_mean": 0.0028042558117249428,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006362373289448442,
      "epoch": 0.13488195504451383,
      "grad_norm": 0.07798866927623749,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 272
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0959821428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15091.0,
      "completions/mean_length": 3675.169921875,
      "completions/mean_terminated_length": 2434.55810546875,
      "completions/min_length": 254.0,
      "completions/min_terminated_length": 254.0,
      "epoch": 0.13537784458511867,
      "grad_norm": 0.14326226711273193,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 35138492.0,
      "reward": 0.3995535969734192,
      "reward_std": 0.2318316251039505,
      "rewards/accuracy_reward/mean": 0.3995535671710968,
      "rewards/accuracy_reward/std": 0.49035418033599854,
      "step": 273
    },
    {
      "clip_ratio/high_max": 0.0054711799830329255,
      "clip_ratio/high_mean": 0.0012710295173974373,
      "clip_ratio/low_mean": 0.0007048643107054886,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001975893830262976,
      "epoch": 0.13587373412572348,
      "grad_norm": 0.12484049797058105,
      "learning_rate": 1e-06,
      "loss": 0.0002,
      "step": 274
    },
    {
      "clip_ratio/high_max": 0.006297263613305404,
      "clip_ratio/high_mean": 0.0014965298494189483,
      "clip_ratio/low_mean": 0.0009108908226949097,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002407420685131001,
      "epoch": 0.13636962366632832,
      "grad_norm": 0.1116931140422821,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 275
    },
    {
      "clip_ratio/high_max": 0.0078036238646745915,
      "clip_ratio/high_mean": 0.0017888989184484672,
      "clip_ratio/low_mean": 0.0013059310222161002,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003094829928386389,
      "epoch": 0.13686551320693316,
      "grad_norm": 0.1091800406575203,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 276
    },
    {
      "clip_ratio/high_max": 0.009140143371041631,
      "clip_ratio/high_mean": 0.00220867155076121,
      "clip_ratio/low_mean": 0.0015409892049547125,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037496607442335517,
      "epoch": 0.137361402747538,
      "grad_norm": 0.09608090668916702,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 277
    },
    {
      "clip_ratio/high_max": 0.010150866028197925,
      "clip_ratio/high_mean": 0.002446641004326011,
      "clip_ratio/low_mean": 0.0019205742944450321,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004367215359025067,
      "epoch": 0.1378572922881428,
      "grad_norm": 0.4167588949203491,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 278
    },
    {
      "clip_ratio/high_max": 0.011765411079977639,
      "clip_ratio/high_mean": 0.0028528592347356607,
      "clip_ratio/low_mean": 0.0022660386358666074,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00511889790504938,
      "epoch": 0.13835318182874765,
      "grad_norm": 0.07375650107860565,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 279
    },
    {
      "clip_ratio/high_max": 0.01325578243995551,
      "clip_ratio/high_mean": 0.003156949260755937,
      "clip_ratio/low_mean": 0.0025393590119051623,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00569630817108191,
      "epoch": 0.13884907136935248,
      "grad_norm": 0.0728866383433342,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 280
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15106.0,
      "completions/mean_length": 3504.727783203125,
      "completions/mean_terminated_length": 2468.830078125,
      "completions/min_length": 227.0,
      "completions/min_terminated_length": 227.0,
      "epoch": 0.1393449609099573,
      "grad_norm": 0.16383588314056396,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 36782490.0,
      "reward": 0.3950892984867096,
      "reward_std": 0.2452845275402069,
      "rewards/accuracy_reward/mean": 0.3950892984867096,
      "rewards/accuracy_reward/std": 0.4894163906574249,
      "step": 281
    },
    {
      "clip_ratio/high_max": 0.005577909329076647,
      "clip_ratio/high_mean": 0.0012697071304046403,
      "clip_ratio/low_mean": 0.0008283566485260963,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020980637880256836,
      "epoch": 0.13984085045056213,
      "grad_norm": 0.1292608678340912,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 282
    },
    {
      "clip_ratio/high_max": 0.006710757059408934,
      "clip_ratio/high_mean": 0.0015532555710251472,
      "clip_ratio/low_mean": 0.0010216572088665998,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00257491279057831,
      "epoch": 0.14033673999116697,
      "grad_norm": 0.10077671706676483,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 283
    },
    {
      "clip_ratio/high_max": 0.007945709905470721,
      "clip_ratio/high_mean": 0.0018301656887160789,
      "clip_ratio/low_mean": 0.0012243162148308784,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030544818741873314,
      "epoch": 0.1408326295317718,
      "grad_norm": 0.09652832895517349,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 284
    },
    {
      "clip_ratio/high_max": 0.009132172137469752,
      "clip_ratio/high_mean": 0.0021120579090165847,
      "clip_ratio/low_mean": 0.0016151200732110738,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037271779597176646,
      "epoch": 0.14132851907237662,
      "grad_norm": 0.08959442377090454,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 285
    },
    {
      "clip_ratio/high_max": 0.011052867230318952,
      "clip_ratio/high_mean": 0.002560072075084463,
      "clip_ratio/low_mean": 0.001991191458500907,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004551263509711134,
      "epoch": 0.14182440861298146,
      "grad_norm": 0.07796226441860199,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 286
    },
    {
      "clip_ratio/high_max": 0.012516249458712991,
      "clip_ratio/high_mean": 0.002875388613119867,
      "clip_ratio/low_mean": 0.0024152862818027643,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005290674904699699,
      "epoch": 0.1423202981535863,
      "grad_norm": 0.07379800081253052,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 287
    },
    {
      "clip_ratio/high_max": 0.01377072733521345,
      "clip_ratio/high_mean": 0.003171671638483531,
      "clip_ratio/low_mean": 0.0027846265870721254,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005956298236014845,
      "epoch": 0.14281618769419113,
      "grad_norm": 0.07033225893974304,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 288
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14651.0,
      "completions/mean_length": 3600.24560546875,
      "completions/mean_terminated_length": 2603.65625,
      "completions/min_length": 327.0,
      "completions/min_terminated_length": 327.0,
      "epoch": 0.14331207723479594,
      "grad_norm": 0.15649835765361786,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 38464944.0,
      "reward": 0.4196428656578064,
      "reward_std": 0.24183818697929382,
      "rewards/accuracy_reward/mean": 0.4196428656578064,
      "rewards/accuracy_reward/std": 0.4940522015094757,
      "step": 289
    },
    {
      "clip_ratio/high_max": 0.006670758248219499,
      "clip_ratio/high_mean": 0.0015356099396512946,
      "clip_ratio/low_mean": 0.001023724948851168,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025593349046175717,
      "epoch": 0.14380796677540078,
      "grad_norm": 0.1319458782672882,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 290
    },
    {
      "clip_ratio/high_max": 0.008044726004300173,
      "clip_ratio/high_mean": 0.0018136099884031864,
      "clip_ratio/low_mean": 0.0011866120343313469,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030002220355527243,
      "epoch": 0.14430385631600562,
      "grad_norm": 0.11506669968366623,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 291
    },
    {
      "clip_ratio/high_max": 0.010243642598652514,
      "clip_ratio/high_mean": 0.002305913888676514,
      "clip_ratio/low_mean": 0.0016369934044746515,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039429072703569545,
      "epoch": 0.14479974585661043,
      "grad_norm": 0.11019513756036758,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 292
    },
    {
      "clip_ratio/high_max": 0.0118645579823351,
      "clip_ratio/high_mean": 0.0026657014693682868,
      "clip_ratio/low_mean": 0.002017713775842367,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004683415236286237,
      "epoch": 0.14529563539721527,
      "grad_norm": 0.09220142662525177,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 293
    },
    {
      "clip_ratio/high_max": 0.013652935504069319,
      "clip_ratio/high_mean": 0.003042028268964714,
      "clip_ratio/low_mean": 0.002399561575231246,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005441589899419341,
      "epoch": 0.1457915249378201,
      "grad_norm": 0.10637906193733215,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 294
    },
    {
      "clip_ratio/high_max": 0.015800845212652348,
      "clip_ratio/high_mean": 0.003453988366345584,
      "clip_ratio/low_mean": 0.0028340612796569076,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006288049729846534,
      "epoch": 0.14628741447842494,
      "grad_norm": 0.0765334889292717,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 295
    },
    {
      "clip_ratio/high_max": 0.016966881419648416,
      "clip_ratio/high_mean": 0.0037756673523290374,
      "clip_ratio/low_mean": 0.0032014116019354333,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00697707899598754,
      "epoch": 0.14678330401902975,
      "grad_norm": 0.0709555596113205,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 296
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15030.0,
      "completions/mean_length": 3135.47998046875,
      "completions/mean_terminated_length": 2320.511962890625,
      "completions/min_length": 238.0,
      "completions/min_terminated_length": 238.0,
      "epoch": 0.1472791935596346,
      "grad_norm": 0.126410111784935,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 39936183.0,
      "reward": 0.4486607313156128,
      "reward_std": 0.17818525433540344,
      "rewards/accuracy_reward/mean": 0.4486607015132904,
      "rewards/accuracy_reward/std": 0.49791330099105835,
      "step": 297
    },
    {
      "clip_ratio/high_max": 0.004919985483866185,
      "clip_ratio/high_mean": 0.0010503727121431439,
      "clip_ratio/low_mean": 0.0007280871345187734,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017784598438765897,
      "epoch": 0.14777508310023943,
      "grad_norm": 0.11244285851716995,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 298
    },
    {
      "clip_ratio/high_max": 0.0059285995848767925,
      "clip_ratio/high_mean": 0.00126362220134979,
      "clip_ratio/low_mean": 0.0008842683183445388,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021478905324272546,
      "epoch": 0.14827097264084424,
      "grad_norm": 0.09911437332630157,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 299
    },
    {
      "clip_ratio/high_max": 0.00677051429374842,
      "clip_ratio/high_mean": 0.0014324709279662784,
      "clip_ratio/low_mean": 0.0011858902103938362,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026183611147132524,
      "epoch": 0.14876686218144908,
      "grad_norm": 0.08921737968921661,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 300
    },
    {
      "clip_ratio/high_max": 0.008020192650292302,
      "clip_ratio/high_mean": 0.0016660246799347078,
      "clip_ratio/low_mean": 0.0014759700008539767,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031419946930100195,
      "epoch": 0.14926275172205392,
      "grad_norm": 0.0762285366654396,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 301
    },
    {
      "clip_ratio/high_max": 0.009361766165966401,
      "clip_ratio/high_mean": 0.001954654929249955,
      "clip_ratio/low_mean": 0.0018178664822130486,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037725213632597843,
      "epoch": 0.14975864126265875,
      "grad_norm": 0.06784319132566452,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 302
    },
    {
      "clip_ratio/high_max": 0.010397596626717132,
      "clip_ratio/high_mean": 0.002210390476648172,
      "clip_ratio/low_mean": 0.0021183535672548715,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004328744069368895,
      "epoch": 0.15025453080326356,
      "grad_norm": 0.06508739292621613,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 303
    },
    {
      "clip_ratio/high_max": 0.011540717176103499,
      "clip_ratio/high_mean": 0.002472909445714322,
      "clip_ratio/low_mean": 0.00238316861077692,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0048560781283413235,
      "epoch": 0.1507504203438684,
      "grad_norm": 0.06117117404937744,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 304
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15265.0,
      "completions/mean_length": 4016.81494140625,
      "completions/mean_terminated_length": 3055.528076171875,
      "completions/min_length": 267.0,
      "completions/min_terminated_length": 267.0,
      "epoch": 0.15124630988447324,
      "grad_norm": 0.14686031639575958,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 41808628.0,
      "reward": 0.4508928656578064,
      "reward_std": 0.24506837129592896,
      "rewards/accuracy_reward/mean": 0.4508928656578064,
      "rewards/accuracy_reward/std": 0.4981389045715332,
      "step": 305
    },
    {
      "clip_ratio/high_max": 0.005665526401571697,
      "clip_ratio/high_mean": 0.0012230748084220977,
      "clip_ratio/low_mean": 0.0009819235197312537,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002204998346769571,
      "epoch": 0.15174219942507805,
      "grad_norm": 0.12485535442829132,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 306
    },
    {
      "clip_ratio/high_max": 0.007684251009777654,
      "clip_ratio/high_mean": 0.0016462774947285652,
      "clip_ratio/low_mean": 0.0013315824984374558,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002977860019200307,
      "epoch": 0.1522380889656829,
      "grad_norm": 0.10339293628931046,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 307
    },
    {
      "clip_ratio/high_max": 0.008854790528857848,
      "clip_ratio/high_mean": 0.0018770759475046361,
      "clip_ratio/low_mean": 0.0015806339431492233,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034577099272610212,
      "epoch": 0.15273397850628773,
      "grad_norm": 0.5912436246871948,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 308
    },
    {
      "clip_ratio/high_max": 0.010691582319850568,
      "clip_ratio/high_mean": 0.0022500033414871723,
      "clip_ratio/low_mean": 0.0019586168934893067,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042086202220161795,
      "epoch": 0.15322986804689256,
      "grad_norm": 0.09112657606601715,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 309
    },
    {
      "clip_ratio/high_max": 0.011613657337875338,
      "clip_ratio/high_mean": 0.002451784629101894,
      "clip_ratio/low_mean": 0.002257186602037109,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004708971198851941,
      "epoch": 0.15372575758749737,
      "grad_norm": 0.08381519466638565,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 310
    },
    {
      "clip_ratio/high_max": 0.013281400191772263,
      "clip_ratio/high_mean": 0.002822937051405461,
      "clip_ratio/low_mean": 0.002566428224668016,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005389365282098879,
      "epoch": 0.1542216471281022,
      "grad_norm": 0.07305855304002762,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 311
    },
    {
      "clip_ratio/high_max": 0.014578504480596166,
      "clip_ratio/high_mean": 0.0031123171888793877,
      "clip_ratio/low_mean": 0.0029421416027162195,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006054458836842969,
      "epoch": 0.15471753666870705,
      "grad_norm": 0.07195312529802322,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 312
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0915178571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14615.0,
      "completions/mean_length": 3559.4443359375,
      "completions/mean_terminated_length": 2370.6904296875,
      "completions/min_length": 318.0,
      "completions/min_terminated_length": 318.0,
      "epoch": 0.1552134262093119,
      "grad_norm": 0.14236600697040558,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 43469587.0,
      "reward": 0.5178571939468384,
      "reward_std": 0.22153069078922272,
      "rewards/accuracy_reward/mean": 0.5178571343421936,
      "rewards/accuracy_reward/std": 0.5002396106719971,
      "step": 313
    },
    {
      "clip_ratio/high_max": 0.0052734395721927285,
      "clip_ratio/high_mean": 0.0012246713927197561,
      "clip_ratio/low_mean": 0.0007349880690981081,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001959659439535244,
      "epoch": 0.1557093157499167,
      "grad_norm": 0.12461022287607193,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 314
    },
    {
      "clip_ratio/high_max": 0.006739895517966943,
      "clip_ratio/high_mean": 0.0015525012586294906,
      "clip_ratio/low_mean": 0.0010561524733248007,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026086537120590947,
      "epoch": 0.15620520529052154,
      "grad_norm": 0.1089615449309349,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 315
    },
    {
      "clip_ratio/high_max": 0.008273348739749053,
      "clip_ratio/high_mean": 0.0018939112628686416,
      "clip_ratio/low_mean": 0.0013733057392073533,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003267217023676494,
      "epoch": 0.15670109483112638,
      "grad_norm": 0.09143701195716858,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 316
    },
    {
      "clip_ratio/high_max": 0.009840944505413063,
      "clip_ratio/high_mean": 0.00220846956335663,
      "clip_ratio/low_mean": 0.0017706967923913908,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039791663116375275,
      "epoch": 0.15719698437173119,
      "grad_norm": 0.09338968247175217,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 317
    },
    {
      "clip_ratio/high_max": 0.011040043940738542,
      "clip_ratio/high_mean": 0.002485523017185187,
      "clip_ratio/low_mean": 0.002088604591335752,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004574127614432655,
      "epoch": 0.15769287391233602,
      "grad_norm": 0.07617510855197906,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 318
    },
    {
      "clip_ratio/high_max": 0.01219565378050902,
      "clip_ratio/high_mean": 0.0027402433775023383,
      "clip_ratio/low_mean": 0.0025042901061169687,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005244533466793655,
      "epoch": 0.15818876345294086,
      "grad_norm": 0.07351313531398773,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 319
    },
    {
      "clip_ratio/high_max": 0.013358933061681455,
      "clip_ratio/high_mean": 0.003005422647220257,
      "clip_ratio/low_mean": 0.0028229399076735717,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005828362555803324,
      "epoch": 0.1586846529935457,
      "grad_norm": 0.06698370724916458,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "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": 0.0558035714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14887.0,
      "completions/mean_length": 3161.024658203125,
      "completions/mean_terminated_length": 2440.044921875,
      "completions/min_length": 396.0,
      "completions/min_terminated_length": 396.0,
      "epoch": 0.1591805425341505,
      "grad_norm": 0.16317249834537506,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 44953902.0,
      "reward": 0.535714328289032,
      "reward_std": 0.2504725456237793,
      "rewards/accuracy_reward/mean": 0.5357142686843872,
      "rewards/accuracy_reward/std": 0.4992803931236267,
      "step": 321
    },
    {
      "clip_ratio/high_max": 0.005669975893852097,
      "clip_ratio/high_mean": 0.0013859297536100712,
      "clip_ratio/low_mean": 0.0011049316219953198,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024908613585239436,
      "epoch": 0.15967643207475535,
      "grad_norm": 0.1327161192893982,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 322
    },
    {
      "clip_ratio/high_max": 0.0067322146242077,
      "clip_ratio/high_mean": 0.0016162155334313866,
      "clip_ratio/low_mean": 0.0014021679949109966,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030183835009438553,
      "epoch": 0.16017232161536019,
      "grad_norm": 0.13251926004886627,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 323
    },
    {
      "clip_ratio/high_max": 0.00824541402835166,
      "clip_ratio/high_mean": 0.0019280501389857818,
      "clip_ratio/low_mean": 0.0017043607021776097,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003632410838577016,
      "epoch": 0.160668211155965,
      "grad_norm": 0.09692376106977463,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 324
    },
    {
      "clip_ratio/high_max": 0.010072230952573591,
      "clip_ratio/high_mean": 0.0023690963873832516,
      "clip_ratio/low_mean": 0.0022011057887993957,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004570202209265517,
      "epoch": 0.16116410069656983,
      "grad_norm": 0.08947577327489853,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 325
    },
    {
      "clip_ratio/high_max": 0.011420661287047551,
      "clip_ratio/high_mean": 0.002686522116391643,
      "clip_ratio/low_mean": 0.0026162677017964597,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005302789791471696,
      "epoch": 0.16165999023717467,
      "grad_norm": 0.07811744511127472,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 326
    },
    {
      "clip_ratio/high_max": 0.013255334481073078,
      "clip_ratio/high_mean": 0.00311091079538528,
      "clip_ratio/low_mean": 0.0030284696321132287,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006139380352351509,
      "epoch": 0.1621558797777795,
      "grad_norm": 0.0818217396736145,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 327
    },
    {
      "clip_ratio/high_max": 0.01410065642267,
      "clip_ratio/high_mean": 0.0033825923665062874,
      "clip_ratio/low_mean": 0.0035110677847569605,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006893660026662474,
      "epoch": 0.16265176931838432,
      "grad_norm": 0.07265164703130722,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 328
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1049107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14448.0,
      "completions/mean_length": 4628.83740234375,
      "completions/mean_terminated_length": 3371.06982421875,
      "completions/min_length": 367.0,
      "completions/min_terminated_length": 367.0,
      "epoch": 0.16314765885898916,
      "grad_norm": 0.1280871033668518,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 47104437.0,
      "reward": 0.3392857313156128,
      "reward_std": 0.2127424031496048,
      "rewards/accuracy_reward/mean": 0.3392857015132904,
      "rewards/accuracy_reward/std": 0.47399619221687317,
      "step": 329
    },
    {
      "clip_ratio/high_max": 0.004253544518178387,
      "clip_ratio/high_mean": 0.0009996277299819667,
      "clip_ratio/low_mean": 0.000983549214595314,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019831769325833193,
      "epoch": 0.163643548399594,
      "grad_norm": 0.10383424907922745,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 330
    },
    {
      "clip_ratio/high_max": 0.00472206168979028,
      "clip_ratio/high_mean": 0.0010792290476047128,
      "clip_ratio/low_mean": 0.00110601806932209,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021852471395504836,
      "epoch": 0.1641394379401988,
      "grad_norm": 0.09656023234128952,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 331
    },
    {
      "clip_ratio/high_max": 0.006087784698138421,
      "clip_ratio/high_mean": 0.0013562359051775275,
      "clip_ratio/low_mean": 0.001397305124896775,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002753540991193404,
      "epoch": 0.16463532748080364,
      "grad_norm": 0.08476672321557999,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 332
    },
    {
      "clip_ratio/high_max": 0.007023445453341992,
      "clip_ratio/high_mean": 0.0015841302149510739,
      "clip_ratio/low_mean": 0.0017769369617326447,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033610671570158956,
      "epoch": 0.16513121702140848,
      "grad_norm": 0.07479941099882126,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 333
    },
    {
      "clip_ratio/high_max": 0.008126462882501073,
      "clip_ratio/high_mean": 0.0018623068481247174,
      "clip_ratio/low_mean": 0.0020735001619414106,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003935807016205217,
      "epoch": 0.16562710656201332,
      "grad_norm": 0.06919775903224945,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 334
    },
    {
      "clip_ratio/high_max": 0.009035618346388219,
      "clip_ratio/high_mean": 0.002056833901178834,
      "clip_ratio/low_mean": 0.0024367831405243123,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004493617046364307,
      "epoch": 0.16612299610261813,
      "grad_norm": 0.06692374497652054,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 335
    },
    {
      "clip_ratio/high_max": 0.009889918997942004,
      "clip_ratio/high_mean": 0.002253761231258977,
      "clip_ratio/low_mean": 0.0027739887094639926,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005027749901955758,
      "epoch": 0.16661888564322297,
      "grad_norm": 0.060890376567840576,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 336
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0758928571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14613.0,
      "completions/mean_length": 3668.727783203125,
      "completions/mean_terminated_length": 2708.574951171875,
      "completions/min_length": 375.0,
      "completions/min_terminated_length": 375.0,
      "epoch": 0.1671147751838278,
      "grad_norm": 0.14432071149349213,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 48818611.0,
      "reward": 0.3638392984867096,
      "reward_std": 0.22927908599376678,
      "rewards/accuracy_reward/mean": 0.3638392984867096,
      "rewards/accuracy_reward/std": 0.48164102435112,
      "step": 337
    },
    {
      "clip_ratio/high_max": 0.004436411308688548,
      "clip_ratio/high_mean": 0.0009715282241131717,
      "clip_ratio/low_mean": 0.0010442937921197881,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020158220095254364,
      "epoch": 0.16761066472443265,
      "grad_norm": 0.12347402423620224,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 338
    },
    {
      "clip_ratio/high_max": 0.005357418150197191,
      "clip_ratio/high_mean": 0.0011695574368673078,
      "clip_ratio/low_mean": 0.0014092954229454335,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002578852871010895,
      "epoch": 0.16810655426503746,
      "grad_norm": 0.10098458081483841,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 339
    },
    {
      "clip_ratio/high_max": 0.006655354384747625,
      "clip_ratio/high_mean": 0.001455609288029791,
      "clip_ratio/low_mean": 0.001811169613574748,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032667789000697667,
      "epoch": 0.1686024438056423,
      "grad_norm": 0.08813236653804779,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 340
    },
    {
      "clip_ratio/high_max": 0.008330293101153075,
      "clip_ratio/high_mean": 0.0018024635961637614,
      "clip_ratio/low_mean": 0.002225381239441049,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004027844843221828,
      "epoch": 0.16909833334624713,
      "grad_norm": 0.08137432485818863,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 341
    },
    {
      "clip_ratio/high_max": 0.009471288857639593,
      "clip_ratio/high_mean": 0.0020878490967675134,
      "clip_ratio/low_mean": 0.0026628493685620924,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004750698453790392,
      "epoch": 0.16959422288685194,
      "grad_norm": 0.07489914447069168,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 342
    },
    {
      "clip_ratio/high_max": 0.010788274219066807,
      "clip_ratio/high_mean": 0.0023690672529710355,
      "clip_ratio/low_mean": 0.003141679227610439,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005510746454092441,
      "epoch": 0.17009011242745678,
      "grad_norm": 0.06540001928806305,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 343
    },
    {
      "clip_ratio/high_max": 0.012050565813296998,
      "clip_ratio/high_mean": 0.0026216989398903934,
      "clip_ratio/low_mean": 0.003475501957382221,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006097200923250057,
      "epoch": 0.17058600196806162,
      "grad_norm": 0.06230587139725685,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 344
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.044642857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15176.0,
      "completions/mean_length": 3489.281494140625,
      "completions/mean_terminated_length": 2934.57470703125,
      "completions/min_length": 344.0,
      "completions/min_terminated_length": 344.0,
      "epoch": 0.17108189150866646,
      "grad_norm": 0.1722555309534073,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 50455553.0,
      "reward": 0.4508928656578064,
      "reward_std": 0.2740766406059265,
      "rewards/accuracy_reward/mean": 0.4508928656578064,
      "rewards/accuracy_reward/std": 0.4981389045715332,
      "step": 345
    },
    {
      "clip_ratio/high_max": 0.0060495146208268125,
      "clip_ratio/high_mean": 0.0013770283715075493,
      "clip_ratio/low_mean": 0.0011496989625356946,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025267273294957704,
      "epoch": 0.17157778104927127,
      "grad_norm": 0.13526687026023865,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 346
    },
    {
      "clip_ratio/high_max": 0.008160603454598458,
      "clip_ratio/high_mean": 0.0018218857335341454,
      "clip_ratio/low_mean": 0.001582853292859454,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034047390245177667,
      "epoch": 0.1720736705898761,
      "grad_norm": 0.11619925498962402,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 347
    },
    {
      "clip_ratio/high_max": 0.009790590298507595,
      "clip_ratio/high_mean": 0.002214399411514023,
      "clip_ratio/low_mean": 0.0020851400375363482,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004299539426938281,
      "epoch": 0.17256956013048094,
      "grad_norm": 0.09626959264278412,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 348
    },
    {
      "clip_ratio/high_max": 0.01197797365966835,
      "clip_ratio/high_mean": 0.0026312323616366484,
      "clip_ratio/low_mean": 0.002604069463131964,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005235301806351345,
      "epoch": 0.17306544967108575,
      "grad_norm": 0.08880406618118286,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 349
    },
    {
      "clip_ratio/high_max": 0.014279339098720811,
      "clip_ratio/high_mean": 0.003142576940717845,
      "clip_ratio/low_mean": 0.0032373373638847625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00637991435087315,
      "epoch": 0.1735613392116906,
      "grad_norm": 0.07943679392337799,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 350
    },
    {
      "clip_ratio/high_max": 0.015249884592776652,
      "clip_ratio/high_mean": 0.003419980058879446,
      "clip_ratio/low_mean": 0.0037371691203134105,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007157149215345271,
      "epoch": 0.17405722875229543,
      "grad_norm": 0.0738634467124939,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 351
    },
    {
      "clip_ratio/high_max": 0.017058583871403243,
      "clip_ratio/high_mean": 0.0037940883012197446,
      "clip_ratio/low_mean": 0.004149678828184733,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007943767197502893,
      "epoch": 0.17455311829290027,
      "grad_norm": 0.06597808748483658,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 352
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.044642857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14706.0,
      "completions/mean_length": 3154.7255859375,
      "completions/mean_terminated_length": 2584.385498046875,
      "completions/min_length": 318.0,
      "completions/min_terminated_length": 318.0,
      "epoch": 0.17504900783350508,
      "grad_norm": 0.1501922905445099,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 51935670.0,
      "reward": 0.5334821939468384,
      "reward_std": 0.2505558729171753,
      "rewards/accuracy_reward/mean": 0.5334821343421936,
      "rewards/accuracy_reward/std": 0.4994353950023651,
      "step": 353
    },
    {
      "clip_ratio/high_max": 0.005317979524988914,
      "clip_ratio/high_mean": 0.0013697741842406685,
      "clip_ratio/low_mean": 0.000866140882635591,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00223591505391596,
      "epoch": 0.17554489737410992,
      "grad_norm": 0.1301083117723465,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 354
    },
    {
      "clip_ratio/high_max": 0.00661379159646458,
      "clip_ratio/high_mean": 0.0017282572280237218,
      "clip_ratio/low_mean": 0.0011766521466824997,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002904909370954556,
      "epoch": 0.17604078691471475,
      "grad_norm": 0.10958127677440643,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 355
    },
    {
      "clip_ratio/high_max": 0.008834865922835888,
      "clip_ratio/high_mean": 0.0022207542297110194,
      "clip_ratio/low_mean": 0.0015417926647387503,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003762546903999464,
      "epoch": 0.1765366764553196,
      "grad_norm": 0.09234565496444702,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 356
    },
    {
      "clip_ratio/high_max": 0.010535629418882309,
      "clip_ratio/high_mean": 0.0026377597610007797,
      "clip_ratio/low_mean": 0.0020314283099196473,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004669188037951244,
      "epoch": 0.1770325659959244,
      "grad_norm": 0.08738448470830917,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 357
    },
    {
      "clip_ratio/high_max": 0.012396201273077168,
      "clip_ratio/high_mean": 0.0030429866669692274,
      "clip_ratio/low_mean": 0.0024582817211467045,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00550126838243159,
      "epoch": 0.17752845553652924,
      "grad_norm": 0.07476630061864853,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 358
    },
    {
      "clip_ratio/high_max": 0.01393779604404699,
      "clip_ratio/high_mean": 0.0033599079242776497,
      "clip_ratio/low_mean": 0.002914163818331872,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00627407178581052,
      "epoch": 0.17802434507713408,
      "grad_norm": 0.0695321336388588,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 359
    },
    {
      "clip_ratio/high_max": 0.015608344849169953,
      "clip_ratio/high_mean": 0.0038145430494296306,
      "clip_ratio/low_mean": 0.0033133341883058165,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0071278771702054655,
      "epoch": 0.1785202346177389,
      "grad_norm": 0.06272797286510468,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 360
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0401785714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15276.0,
      "completions/mean_length": 3055.15185546875,
      "completions/mean_terminated_length": 2540.065185546875,
      "completions/min_length": 376.0,
      "completions/min_terminated_length": 376.0,
      "epoch": 0.17901612415834373,
      "grad_norm": 0.18425078690052032,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 53372906.0,
      "reward": 0.5491071939468384,
      "reward_std": 0.2974432706832886,
      "rewards/accuracy_reward/mean": 0.5491071343421936,
      "rewards/accuracy_reward/std": 0.4981389045715332,
      "step": 361
    },
    {
      "clip_ratio/high_max": 0.006046652260010887,
      "clip_ratio/high_mean": 0.0016411840840078185,
      "clip_ratio/low_mean": 0.00127512164840482,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029163057151890825,
      "epoch": 0.17951201369894856,
      "grad_norm": 0.14748848974704742,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 362
    },
    {
      "clip_ratio/high_max": 0.007747321349597769,
      "clip_ratio/high_mean": 0.0020953717250904447,
      "clip_ratio/low_mean": 0.0015991368852610321,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003694508609441982,
      "epoch": 0.1800079032395534,
      "grad_norm": 0.12851925194263458,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 363
    },
    {
      "clip_ratio/high_max": 0.009306051852036035,
      "clip_ratio/high_mean": 0.0024550539706069685,
      "clip_ratio/low_mean": 0.0021360377879773296,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004591091728798347,
      "epoch": 0.1805037927801582,
      "grad_norm": 0.10891084372997284,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 364
    },
    {
      "clip_ratio/high_max": 0.011281812523520784,
      "clip_ratio/high_mean": 0.0029100043088874372,
      "clip_ratio/low_mean": 0.002554156179940037,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005464160449264455,
      "epoch": 0.18099968232076305,
      "grad_norm": 0.09985747933387756,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 365
    },
    {
      "clip_ratio/high_max": 0.012931635399581864,
      "clip_ratio/high_mean": 0.0033578694278730836,
      "clip_ratio/low_mean": 0.003112799251994147,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006470668722613482,
      "epoch": 0.1814955718613679,
      "grad_norm": 0.09169958531856537,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 366
    },
    {
      "clip_ratio/high_max": 0.014877521123707993,
      "clip_ratio/high_mean": 0.003866886290779803,
      "clip_ratio/low_mean": 0.003586682850254874,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0074535690910124686,
      "epoch": 0.1819914614019727,
      "grad_norm": 0.08017182350158691,
      "learning_rate": 1e-06,
      "loss": -0.0016,
      "step": 367
    },
    {
      "clip_ratio/high_max": 0.016146414342074422,
      "clip_ratio/high_mean": 0.004222355472393247,
      "clip_ratio/low_mean": 0.004162560772783763,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.008384916287468513,
      "epoch": 0.18248735094257754,
      "grad_norm": 0.08676199615001678,
      "learning_rate": 1e-06,
      "loss": -0.0018,
      "step": 368
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0513392857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15054.0,
      "completions/mean_length": 3504.950927734375,
      "completions/mean_terminated_length": 2863.38330078125,
      "completions/min_length": 570.0,
      "completions/min_terminated_length": 570.0,
      "epoch": 0.18298324048318237,
      "grad_norm": 0.14170844852924347,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 55011356.0,
      "reward": 0.4955357313156128,
      "reward_std": 0.24882563948631287,
      "rewards/accuracy_reward/mean": 0.4955357015132904,
      "rewards/accuracy_reward/std": 0.5005390048027039,
      "step": 369
    },
    {
      "clip_ratio/high_max": 0.004646013008823502,
      "clip_ratio/high_mean": 0.0011542072161319084,
      "clip_ratio/low_mean": 0.0009821104121101598,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002136317638814944,
      "epoch": 0.1834791300237872,
      "grad_norm": 0.12357435375452042,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 370
    },
    {
      "clip_ratio/high_max": 0.005793183836431126,
      "clip_ratio/high_mean": 0.0014281390322139487,
      "clip_ratio/low_mean": 0.0012530143555977702,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026811533944055554,
      "epoch": 0.18397501956439202,
      "grad_norm": 0.17298002541065216,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 371
    },
    {
      "clip_ratio/high_max": 0.007317651172343176,
      "clip_ratio/high_mean": 0.0018188284943789768,
      "clip_ratio/low_mean": 0.0017797864190924884,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035986149314339855,
      "epoch": 0.18447090910499686,
      "grad_norm": 0.09287194162607193,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 372
    },
    {
      "clip_ratio/high_max": 0.009019258250191342,
      "clip_ratio/high_mean": 0.002195928734181507,
      "clip_ratio/low_mean": 0.0021876547682495584,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004383583502203692,
      "epoch": 0.1849667986456017,
      "grad_norm": 0.08684719353914261,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 373
    },
    {
      "clip_ratio/high_max": 0.010302761103957891,
      "clip_ratio/high_mean": 0.0024505330152351235,
      "clip_ratio/low_mean": 0.0026107504588708252,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005061283496615943,
      "epoch": 0.1854626881862065,
      "grad_norm": 0.0791984349489212,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 374
    },
    {
      "clip_ratio/high_max": 0.011507758368679788,
      "clip_ratio/high_mean": 0.0027939639148826245,
      "clip_ratio/low_mean": 0.003000340549760949,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005794304506707704,
      "epoch": 0.18595857772681135,
      "grad_norm": 0.07033291459083557,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 375
    },
    {
      "clip_ratio/high_max": 0.01263812727847835,
      "clip_ratio/high_mean": 0.0030569692489734734,
      "clip_ratio/low_mean": 0.003410516794929208,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006467486069595907,
      "epoch": 0.18645446726741619,
      "grad_norm": 0.06631667166948318,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 376
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0647321428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15353.0,
      "completions/mean_length": 3779.498046875,
      "completions/mean_terminated_length": 2977.9833984375,
      "completions/min_length": 414.0,
      "completions/min_terminated_length": 414.0,
      "epoch": 0.18695035680802102,
      "grad_norm": 0.13441690802574158,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 56772003.0,
      "reward": 0.4821428656578064,
      "reward_std": 0.2302519828081131,
      "rewards/accuracy_reward/mean": 0.4821428656578064,
      "rewards/accuracy_reward/std": 0.5002396702766418,
      "step": 377
    },
    {
      "clip_ratio/high_max": 0.004977090897227754,
      "clip_ratio/high_mean": 0.0012332339406384563,
      "clip_ratio/low_mean": 0.0007161793876093725,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019494133312036865,
      "epoch": 0.18744624634862583,
      "grad_norm": 0.1057291179895401,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 378
    },
    {
      "clip_ratio/high_max": 0.006201162080287759,
      "clip_ratio/high_mean": 0.001516254264743111,
      "clip_ratio/low_mean": 0.0008884114172360569,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024046656758400786,
      "epoch": 0.18794213588923067,
      "grad_norm": 0.09323406964540482,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 379
    },
    {
      "clip_ratio/high_max": 0.007826728320651455,
      "clip_ratio/high_mean": 0.0018629988117027096,
      "clip_ratio/low_mean": 0.0011133523096304998,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029763511101918994,
      "epoch": 0.1884380254298355,
      "grad_norm": 0.08136916905641556,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 380
    },
    {
      "clip_ratio/high_max": 0.009307103529863525,
      "clip_ratio/high_mean": 0.0022823879912721168,
      "clip_ratio/low_mean": 0.0014416575834275136,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037240455549181206,
      "epoch": 0.18893391497044035,
      "grad_norm": 0.07331608235836029,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 381
    },
    {
      "clip_ratio/high_max": 0.010672244439774659,
      "clip_ratio/high_mean": 0.0025343601400891203,
      "clip_ratio/low_mean": 0.0017274036731578235,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004261763785507355,
      "epoch": 0.18942980451104516,
      "grad_norm": 0.0666651651263237,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 382
    },
    {
      "clip_ratio/high_max": 0.011982558644376695,
      "clip_ratio/high_mean": 0.0028554157520375156,
      "clip_ratio/low_mean": 0.0020352498634110816,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0048906656047620345,
      "epoch": 0.18992569405165,
      "grad_norm": 0.07091226428747177,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 383
    },
    {
      "clip_ratio/high_max": 0.013237950053735403,
      "clip_ratio/high_mean": 0.0031526219031547953,
      "clip_ratio/low_mean": 0.0022910519151082553,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0054436737791547785,
      "epoch": 0.19042158359225483,
      "grad_norm": 0.053962334990501404,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "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": 0.0848214285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14967.0,
      "completions/mean_length": 4375.04052734375,
      "completions/mean_terminated_length": 3356.921875,
      "completions/min_length": 511.0,
      "completions/min_terminated_length": 511.0,
      "epoch": 0.19091747313285964,
      "grad_norm": 0.13026458024978638,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 58805069.0,
      "reward": 0.4330357313156128,
      "reward_std": 0.2380809187889099,
      "rewards/accuracy_reward/mean": 0.4330357015132904,
      "rewards/accuracy_reward/std": 0.4960494041442871,
      "step": 385
    },
    {
      "clip_ratio/high_max": 0.004268600256182253,
      "clip_ratio/high_mean": 0.0009145703995727672,
      "clip_ratio/low_mean": 0.000767990947394992,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016825613421360686,
      "epoch": 0.19141336267346448,
      "grad_norm": 0.1117580384016037,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 386
    },
    {
      "clip_ratio/high_max": 0.005697618353224243,
      "clip_ratio/high_mean": 0.0011996960826081704,
      "clip_ratio/low_mean": 0.0009788860950266098,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002178582138640195,
      "epoch": 0.19190925221406932,
      "grad_norm": 0.0984165295958519,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 387
    },
    {
      "clip_ratio/high_max": 0.007137903632610687,
      "clip_ratio/high_mean": 0.0014684743421184976,
      "clip_ratio/low_mean": 0.0013097594086275421,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027782337579083105,
      "epoch": 0.19240514175467416,
      "grad_norm": 0.08528845012187958,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 388
    },
    {
      "clip_ratio/high_max": 0.008633767696665018,
      "clip_ratio/high_mean": 0.0017320934218787443,
      "clip_ratio/low_mean": 0.0017317767628810543,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034638701608855627,
      "epoch": 0.19290103129527897,
      "grad_norm": 0.0739128366112709,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 389
    },
    {
      "clip_ratio/high_max": 0.010156362168345368,
      "clip_ratio/high_mean": 0.002060353543583915,
      "clip_ratio/low_mean": 0.002068007927164217,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004128361472794495,
      "epoch": 0.1933969208358838,
      "grad_norm": 0.06466557085514069,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 390
    },
    {
      "clip_ratio/high_max": 0.01151603932157741,
      "clip_ratio/high_mean": 0.002328390736010988,
      "clip_ratio/low_mean": 0.0023950066126872116,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004723397389170714,
      "epoch": 0.19389281037648864,
      "grad_norm": 0.059904273599386215,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 391
    },
    {
      "clip_ratio/high_max": 0.01279359708132688,
      "clip_ratio/high_mean": 0.002583393165423331,
      "clip_ratio/low_mean": 0.002762120979014071,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005345514173313859,
      "epoch": 0.19438869991709345,
      "grad_norm": 0.0562138557434082,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 392
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.046875,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15127.0,
      "completions/mean_length": 3185.915283203125,
      "completions/mean_terminated_length": 2587.189697265625,
      "completions/min_length": 274.0,
      "completions/min_terminated_length": 274.0,
      "epoch": 0.1948845894576983,
      "grad_norm": 0.16532553732395172,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 60305751.0,
      "reward": 0.5089285969734192,
      "reward_std": 0.25016582012176514,
      "rewards/accuracy_reward/mean": 0.5089285969734192,
      "rewards/accuracy_reward/std": 0.5004791617393494,
      "step": 393
    },
    {
      "clip_ratio/high_max": 0.005360226119591971,
      "clip_ratio/high_mean": 0.0013038629186894468,
      "clip_ratio/low_mean": 0.0007254103820741875,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020292733097448945,
      "epoch": 0.19538047899830313,
      "grad_norm": 0.13716700673103333,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 394
    },
    {
      "clip_ratio/high_max": 0.007002672478847671,
      "clip_ratio/high_mean": 0.00167720664967419,
      "clip_ratio/low_mean": 0.001221183803409076,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002898390480368107,
      "epoch": 0.19587636853890797,
      "grad_norm": 0.12155735492706299,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 395
    },
    {
      "clip_ratio/high_max": 0.008641017891932279,
      "clip_ratio/high_mean": 0.0020856833634752547,
      "clip_ratio/low_mean": 0.001683833964762016,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037695173423344386,
      "epoch": 0.19637225807951278,
      "grad_norm": 0.09781133383512497,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 396
    },
    {
      "clip_ratio/high_max": 0.009979311282222625,
      "clip_ratio/high_mean": 0.0024628912437947292,
      "clip_ratio/low_mean": 0.002090424461357543,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0045533156971941935,
      "epoch": 0.19686814762011762,
      "grad_norm": 0.08708653599023819,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 397
    },
    {
      "clip_ratio/high_max": 0.012204926024423912,
      "clip_ratio/high_mean": 0.0029274142475514964,
      "clip_ratio/low_mean": 0.002460379155309056,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005387793362388038,
      "epoch": 0.19736403716072246,
      "grad_norm": 0.07621461898088455,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 398
    },
    {
      "clip_ratio/high_max": 0.013181974052713485,
      "clip_ratio/high_mean": 0.003254235976783093,
      "clip_ratio/low_mean": 0.0029458544115641416,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006200090396305313,
      "epoch": 0.19785992670132727,
      "grad_norm": 0.07164672762155533,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 399
    },
    {
      "clip_ratio/high_max": 0.014604697727918392,
      "clip_ratio/high_mean": 0.0035447830596240237,
      "clip_ratio/low_mean": 0.003225161468662918,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006769944573534303,
      "epoch": 0.1983558162419321,
      "grad_norm": 0.06585543602705002,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 400
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14653.0,
      "completions/mean_length": 3728.2412109375,
      "completions/mean_terminated_length": 2952.79052734375,
      "completions/min_length": 562.0,
      "completions/min_terminated_length": 562.0,
      "epoch": 0.19885170578253694,
      "grad_norm": 0.13796038925647736,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 62044491.0,
      "reward": 0.5066964626312256,
      "reward_std": 0.25001609325408936,
      "rewards/accuracy_reward/mean": 0.5066964030265808,
      "rewards/accuracy_reward/std": 0.5005140900611877,
      "step": 401
    },
    {
      "clip_ratio/high_max": 0.00463702038541669,
      "clip_ratio/high_mean": 0.00099152610175679,
      "clip_ratio/low_mean": 0.0009452444538737836,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019367705444892636,
      "epoch": 0.19934759532314178,
      "grad_norm": 0.12020285427570343,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 402
    },
    {
      "clip_ratio/high_max": 0.00649249876732938,
      "clip_ratio/high_mean": 0.0013471761320715814,
      "clip_ratio/low_mean": 0.0012487838040442512,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025959599333873484,
      "epoch": 0.1998434848637466,
      "grad_norm": 0.10032843798398972,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 403
    },
    {
      "clip_ratio/high_max": 0.00864829788042698,
      "clip_ratio/high_mean": 0.0017352121232079298,
      "clip_ratio/low_mean": 0.0016113981703256286,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033466103031969396,
      "epoch": 0.20033937440435143,
      "grad_norm": 0.08476389944553375,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 404
    },
    {
      "clip_ratio/high_max": 0.009833205538598122,
      "clip_ratio/high_mean": 0.0020075914018207186,
      "clip_ratio/low_mean": 0.0019956688868205674,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004003260319223045,
      "epoch": 0.20083526394495627,
      "grad_norm": 0.08041660487651825,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 405
    },
    {
      "clip_ratio/high_max": 0.011801565386122093,
      "clip_ratio/high_mean": 0.002356615364874415,
      "clip_ratio/low_mean": 0.0024547713621814182,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004811386723304167,
      "epoch": 0.2013311534855611,
      "grad_norm": 0.0694650262594223,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 406
    },
    {
      "clip_ratio/high_max": 0.012914798258861993,
      "clip_ratio/high_mean": 0.0025973978978299783,
      "clip_ratio/low_mean": 0.0028262677133170655,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005423665599664673,
      "epoch": 0.20182704302616591,
      "grad_norm": 0.06558109819889069,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 407
    },
    {
      "clip_ratio/high_max": 0.014285735614976147,
      "clip_ratio/high_mean": 0.002863555713702226,
      "clip_ratio/low_mean": 0.0032563239362843888,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006119879682955798,
      "epoch": 0.20232293256677075,
      "grad_norm": 0.060023434460163116,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 408
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0714285714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15230.0,
      "completions/mean_length": 3544.946533203125,
      "completions/mean_terminated_length": 2636.09619140625,
      "completions/min_length": 411.0,
      "completions/min_terminated_length": 411.0,
      "epoch": 0.2028188221073756,
      "grad_norm": 0.1762905716896057,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 63701763.0,
      "reward": 0.515625,
      "reward_std": 0.22988249361515045,
      "rewards/accuracy_reward/mean": 0.515625,
      "rewards/accuracy_reward/std": 0.500314474105835,
      "step": 409
    },
    {
      "clip_ratio/high_max": 0.00507850404119381,
      "clip_ratio/high_mean": 0.0012195634874387906,
      "clip_ratio/low_mean": 0.000979278982526921,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021988424741721246,
      "epoch": 0.2033147116479804,
      "grad_norm": 0.13215990364551544,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 410
    },
    {
      "clip_ratio/high_max": 0.006495412533695344,
      "clip_ratio/high_mean": 0.0015583202412017272,
      "clip_ratio/low_mean": 0.0013488135703028092,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029071338155972626,
      "epoch": 0.20381060118858524,
      "grad_norm": 0.11273898929357529,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 411
    },
    {
      "clip_ratio/high_max": 0.007805386638210621,
      "clip_ratio/high_mean": 0.001850381606686824,
      "clip_ratio/low_mean": 0.0018108607453086734,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036612423538144867,
      "epoch": 0.20430649072919008,
      "grad_norm": 0.1519012600183487,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 412
    },
    {
      "clip_ratio/high_max": 0.00950299604483007,
      "clip_ratio/high_mean": 0.0023122798540953227,
      "clip_ratio/low_mean": 0.0023357019879313157,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0046479818622628954,
      "epoch": 0.20480238026979491,
      "grad_norm": 0.08081427216529846,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 413
    },
    {
      "clip_ratio/high_max": 0.011020542538972222,
      "clip_ratio/high_mean": 0.0026820088851309265,
      "clip_ratio/low_mean": 0.0027386177935113665,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005420626618843016,
      "epoch": 0.20529826981039972,
      "grad_norm": 0.07367555052042007,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 414
    },
    {
      "clip_ratio/high_max": 0.012409536428094725,
      "clip_ratio/high_mean": 0.0030116679765797016,
      "clip_ratio/low_mean": 0.003178644896934202,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006190312961280142,
      "epoch": 0.20579415935100456,
      "grad_norm": 0.06465550512075424,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 415
    },
    {
      "clip_ratio/high_max": 0.013538740251533454,
      "clip_ratio/high_mean": 0.003335505437462416,
      "clip_ratio/low_mean": 0.0036186909744628792,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006954196391689038,
      "epoch": 0.2062900488916094,
      "grad_norm": 0.06344138830900192,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 416
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14853.0,
      "completions/mean_length": 4035.12060546875,
      "completions/mean_terminated_length": 3045.56787109375,
      "completions/min_length": 425.0,
      "completions/min_terminated_length": 425.0,
      "epoch": 0.2067859384322142,
      "grad_norm": 0.11404595524072647,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 65578777.0,
      "reward": 0.4598214626312256,
      "reward_std": 0.15128032863140106,
      "rewards/accuracy_reward/mean": 0.4598214328289032,
      "rewards/accuracy_reward/std": 0.49894019961357117,
      "step": 417
    },
    {
      "clip_ratio/high_max": 0.004530197585154383,
      "clip_ratio/high_mean": 0.0008743663287305026,
      "clip_ratio/low_mean": 0.0006304480768903886,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015048144011871045,
      "epoch": 0.20728182797281905,
      "grad_norm": 0.0932532250881195,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 418
    },
    {
      "clip_ratio/high_max": 0.005475403346281382,
      "clip_ratio/high_mean": 0.0010449164574310998,
      "clip_ratio/low_mean": 0.0007658048328949008,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018107212913491821,
      "epoch": 0.2077777175134239,
      "grad_norm": 0.08132358640432358,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 419
    },
    {
      "clip_ratio/high_max": 0.00731919482768717,
      "clip_ratio/high_mean": 0.0013916684119976708,
      "clip_ratio/low_mean": 0.0010524058699274974,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024440742577098717,
      "epoch": 0.20827360705402873,
      "grad_norm": 0.07456531375646591,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 420
    },
    {
      "clip_ratio/high_max": 0.008545182336092694,
      "clip_ratio/high_mean": 0.0015942435825309076,
      "clip_ratio/low_mean": 0.0012943349813667737,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028885785707188916,
      "epoch": 0.20876949659463354,
      "grad_norm": 0.06341797858476639,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 421
    },
    {
      "clip_ratio/high_max": 0.009764025286131073,
      "clip_ratio/high_mean": 0.0018743361288215965,
      "clip_ratio/low_mean": 0.0015438072066444875,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034181433566118358,
      "epoch": 0.20926538613523837,
      "grad_norm": 0.056401945650577545,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 422
    },
    {
      "clip_ratio/high_max": 0.010967018995870603,
      "clip_ratio/high_mean": 0.002094032144668745,
      "clip_ratio/low_mean": 0.0017524398772366112,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003846471981432842,
      "epoch": 0.2097612756758432,
      "grad_norm": 0.04792223498225212,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 423
    },
    {
      "clip_ratio/high_max": 0.011574416552321054,
      "clip_ratio/high_mean": 0.0021984398208587663,
      "clip_ratio/low_mean": 0.0019251812771017285,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004123621075450501,
      "epoch": 0.21025716521644805,
      "grad_norm": 0.045221805572509766,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 424
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.060267857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14792.0,
      "completions/mean_length": 3275.79931640625,
      "completions/mean_terminated_length": 2500.802978515625,
      "completions/min_length": 537.0,
      "completions/min_terminated_length": 537.0,
      "epoch": 0.21075305475705286,
      "grad_norm": 0.15538586676120758,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 67108879.0,
      "reward": 0.5491071939468384,
      "reward_std": 0.2613605558872223,
      "rewards/accuracy_reward/mean": 0.5491071343421936,
      "rewards/accuracy_reward/std": 0.4981389045715332,
      "step": 425
    },
    {
      "clip_ratio/high_max": 0.005528582954866579,
      "clip_ratio/high_mean": 0.0012774916239095546,
      "clip_ratio/low_mean": 0.0009220308199928695,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021995224592501472,
      "epoch": 0.2112489442976577,
      "grad_norm": 0.12544164061546326,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 426
    },
    {
      "clip_ratio/high_max": 0.006558242201208486,
      "clip_ratio/high_mean": 0.0015360877242756033,
      "clip_ratio/low_mean": 0.00123604797590815,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027721357118934975,
      "epoch": 0.21174483383826254,
      "grad_norm": 0.11000988632440567,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 427
    },
    {
      "clip_ratio/high_max": 0.008548070384676976,
      "clip_ratio/high_mean": 0.0019885412186795293,
      "clip_ratio/low_mean": 0.0016795839324004191,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036681251363006595,
      "epoch": 0.21224072337886735,
      "grad_norm": 0.09506767243146896,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 428
    },
    {
      "clip_ratio/high_max": 0.009887067637464497,
      "clip_ratio/high_mean": 0.0023098716490039806,
      "clip_ratio/low_mean": 0.00216175341665803,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004471625045880501,
      "epoch": 0.21273661291947218,
      "grad_norm": 0.08440399169921875,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 429
    },
    {
      "clip_ratio/high_max": 0.011388537071979954,
      "clip_ratio/high_mean": 0.0026806539776771388,
      "clip_ratio/low_mean": 0.0026239127562348585,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005304566736413108,
      "epoch": 0.21323250246007702,
      "grad_norm": 0.08359929174184799,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 430
    },
    {
      "clip_ratio/high_max": 0.01291180717635143,
      "clip_ratio/high_mean": 0.0030492418668472965,
      "clip_ratio/low_mean": 0.003071555567885298,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006120797446783399,
      "epoch": 0.21372839200068186,
      "grad_norm": 0.06780005246400833,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 431
    },
    {
      "clip_ratio/high_max": 0.014198555574694183,
      "clip_ratio/high_mean": 0.0033321578785034944,
      "clip_ratio/low_mean": 0.0034812341866654606,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006813392078584002,
      "epoch": 0.21422428154128667,
      "grad_norm": 0.065037801861763,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 432
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1026785714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15357.0,
      "completions/mean_length": 4444.9375,
      "completions/mean_terminated_length": 3195.9501953125,
      "completions/min_length": 377.0,
      "completions/min_terminated_length": 377.0,
      "epoch": 0.2147201710818915,
      "grad_norm": 0.10776380449533463,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 69175651.0,
      "reward": 0.5178571939468384,
      "reward_std": 0.19163449108600616,
      "rewards/accuracy_reward/mean": 0.5178571343421936,
      "rewards/accuracy_reward/std": 0.5002396702766418,
      "step": 433
    },
    {
      "clip_ratio/high_max": 0.004068339492732775,
      "clip_ratio/high_mean": 0.0007457235252559258,
      "clip_ratio/low_mean": 0.0007869228821846264,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015326464101690362,
      "epoch": 0.21521606062249635,
      "grad_norm": 0.09072855114936829,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 434
    },
    {
      "clip_ratio/high_max": 0.005139882778166793,
      "clip_ratio/high_mean": 0.0009147170544565597,
      "clip_ratio/low_mean": 0.0009278307386466622,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001842547792421101,
      "epoch": 0.21571195016310116,
      "grad_norm": 0.07887856662273407,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 435
    },
    {
      "clip_ratio/high_max": 0.006670363502053078,
      "clip_ratio/high_mean": 0.0011766672732846928,
      "clip_ratio/low_mean": 0.0011994558032597524,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023761230695527047,
      "epoch": 0.216207839703706,
      "grad_norm": 0.06860994547605515,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 436
    },
    {
      "clip_ratio/high_max": 0.007684277516091242,
      "clip_ratio/high_mean": 0.0013737276246956753,
      "clip_ratio/low_mean": 0.0013881475628068074,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002761875187388796,
      "epoch": 0.21670372924431083,
      "grad_norm": 0.059473808854818344,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 437
    },
    {
      "clip_ratio/high_max": 0.009012819795316318,
      "clip_ratio/high_mean": 0.0016145716654136777,
      "clip_ratio/low_mean": 0.00167717211377294,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003291743781119294,
      "epoch": 0.21719961878491567,
      "grad_norm": 0.09302248805761337,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 438
    },
    {
      "clip_ratio/high_max": 0.010294995288859354,
      "clip_ratio/high_mean": 0.0018571001455711666,
      "clip_ratio/low_mean": 0.002011342912624059,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003868443044666492,
      "epoch": 0.21769550832552048,
      "grad_norm": 0.050270430743694305,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 439
    },
    {
      "clip_ratio/high_max": 0.011548595233762171,
      "clip_ratio/high_mean": 0.0020531676016162237,
      "clip_ratio/low_mean": 0.0023031504964592386,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004356318109330459,
      "epoch": 0.21819139786612532,
      "grad_norm": 0.04656864330172539,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 440
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0982142857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15288.0,
      "completions/mean_length": 4221.40869140625,
      "completions/mean_terminated_length": 3008.29443359375,
      "completions/min_length": 537.0,
      "completions/min_terminated_length": 537.0,
      "epoch": 0.21868728740673016,
      "grad_norm": 0.14369437098503113,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 71140386.0,
      "reward": 0.5066964626312256,
      "reward_std": 0.21613015234470367,
      "rewards/accuracy_reward/mean": 0.5066964030265808,
      "rewards/accuracy_reward/std": 0.5005140900611877,
      "step": 441
    },
    {
      "clip_ratio/high_max": 0.004118800066862605,
      "clip_ratio/high_mean": 0.0009871734329180981,
      "clip_ratio/low_mean": 0.000729863827700683,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001717037258686105,
      "epoch": 0.21918317694733497,
      "grad_norm": 0.1144103929400444,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 442
    },
    {
      "clip_ratio/high_max": 0.005682986735337181,
      "clip_ratio/high_mean": 0.0012496387805640552,
      "clip_ratio/low_mean": 0.0009627295587506524,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022123683393147076,
      "epoch": 0.2196790664879398,
      "grad_norm": 0.0952599048614502,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 443
    },
    {
      "clip_ratio/high_max": 0.006952893105335534,
      "clip_ratio/high_mean": 0.0015528247774909687,
      "clip_ratio/low_mean": 0.0013255377998575568,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028783625809865043,
      "epoch": 0.22017495602854464,
      "grad_norm": 0.07809757441282272,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 444
    },
    {
      "clip_ratio/high_max": 0.008518959541106597,
      "clip_ratio/high_mean": 0.0018925583713098604,
      "clip_ratio/low_mean": 0.001695341246886528,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035878996518476924,
      "epoch": 0.22067084556914948,
      "grad_norm": 0.06632252782583237,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 445
    },
    {
      "clip_ratio/high_max": 0.010118398280610563,
      "clip_ratio/high_mean": 0.0022347670151248167,
      "clip_ratio/low_mean": 0.0020371524772144767,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004271919448456174,
      "epoch": 0.2211667351097543,
      "grad_norm": 0.05824476480484009,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 446
    },
    {
      "clip_ratio/high_max": 0.011211685490707168,
      "clip_ratio/high_mean": 0.0024787128022580873,
      "clip_ratio/low_mean": 0.002338887537462142,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004817600366095576,
      "epoch": 0.22166262465035913,
      "grad_norm": 0.053080227226018906,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 447
    },
    {
      "clip_ratio/high_max": 0.01235234280466102,
      "clip_ratio/high_mean": 0.002750737295627914,
      "clip_ratio/low_mean": 0.0026462924677161936,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005397029778578144,
      "epoch": 0.22215851419096397,
      "grad_norm": 0.05126945674419403,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "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": 0.0647321428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15335.0,
      "completions/mean_length": 3647.03369140625,
      "completions/mean_terminated_length": 2836.350830078125,
      "completions/min_length": 530.0,
      "completions/min_terminated_length": 530.0,
      "epoch": 0.2226544037315688,
      "grad_norm": 0.1219181939959526,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 72842969.0,
      "reward": 0.5334821939468384,
      "reward_std": 0.1950194537639618,
      "rewards/accuracy_reward/mean": 0.5334821343421936,
      "rewards/accuracy_reward/std": 0.4994353950023651,
      "step": 449
    },
    {
      "clip_ratio/high_max": 0.005156044970135554,
      "clip_ratio/high_mean": 0.0010260329257789635,
      "clip_ratio/low_mean": 0.0007200726063274487,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017461055331295938,
      "epoch": 0.22315029327217362,
      "grad_norm": 0.10195298492908478,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 450
    },
    {
      "clip_ratio/high_max": 0.006365708635712508,
      "clip_ratio/high_mean": 0.0012830129611529628,
      "clip_ratio/low_mean": 0.0009369284711056025,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002219941433850181,
      "epoch": 0.22364618281277845,
      "grad_norm": 0.08095495402812958,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 451
    },
    {
      "clip_ratio/high_max": 0.007970699644829438,
      "clip_ratio/high_mean": 0.001584316279036102,
      "clip_ratio/low_mean": 0.0011697411616751197,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002754057427864609,
      "epoch": 0.2241420723533833,
      "grad_norm": 0.07171594351530075,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 452
    },
    {
      "clip_ratio/high_max": 0.009396939036378171,
      "clip_ratio/high_mean": 0.0019178523473328823,
      "clip_ratio/low_mean": 0.001444859197590631,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003362711559930176,
      "epoch": 0.2246379618939881,
      "grad_norm": 0.06140248477458954,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 453
    },
    {
      "clip_ratio/high_max": 0.010762107465779991,
      "clip_ratio/high_mean": 0.0021850370740139624,
      "clip_ratio/low_mean": 0.0017448815159468722,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003929918590984016,
      "epoch": 0.22513385143459294,
      "grad_norm": 0.0587463304400444,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 454
    },
    {
      "clip_ratio/high_max": 0.011696712510456564,
      "clip_ratio/high_mean": 0.0023931990904202394,
      "clip_ratio/low_mean": 0.001982751139053107,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004375950185931288,
      "epoch": 0.22562974097519778,
      "grad_norm": 0.046890877187252045,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 455
    },
    {
      "clip_ratio/high_max": 0.01314992189509212,
      "clip_ratio/high_mean": 0.00267372543203237,
      "clip_ratio/low_mean": 0.0022860417832362145,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004959767251420999,
      "epoch": 0.22612563051580262,
      "grad_norm": 0.0466814748942852,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 456
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15248.0,
      "completions/mean_length": 3660.899658203125,
      "completions/mean_terminated_length": 2638.648193359375,
      "completions/min_length": 493.0,
      "completions/min_terminated_length": 493.0,
      "epoch": 0.22662152005640743,
      "grad_norm": 0.1524345576763153,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 74552060.0,
      "reward": 0.535714328289032,
      "reward_std": 0.20358654856681824,
      "rewards/accuracy_reward/mean": 0.5357142686843872,
      "rewards/accuracy_reward/std": 0.4992803931236267,
      "step": 457
    },
    {
      "clip_ratio/high_max": 0.004597266279233736,
      "clip_ratio/high_mean": 0.0010232576426005835,
      "clip_ratio/low_mean": 0.0007856594349959778,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018089170943085264,
      "epoch": 0.22711740959701227,
      "grad_norm": 0.12642796337604523,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 458
    },
    {
      "clip_ratio/high_max": 0.006217878295501578,
      "clip_ratio/high_mean": 0.0013452283064907533,
      "clip_ratio/low_mean": 0.0011555453926348491,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025007736855968687,
      "epoch": 0.2276132991376171,
      "grad_norm": 0.109556645154953,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 459
    },
    {
      "clip_ratio/high_max": 0.008312020672747167,
      "clip_ratio/high_mean": 0.0018165637270612933,
      "clip_ratio/low_mean": 0.0016601147005985695,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003476678408560474,
      "epoch": 0.2281091886782219,
      "grad_norm": 0.08621096611022949,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 460
    },
    {
      "clip_ratio/high_max": 0.009690301696537063,
      "clip_ratio/high_mean": 0.002107658736349549,
      "clip_ratio/low_mean": 0.002084870053067789,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004192528798739659,
      "epoch": 0.22860507821882675,
      "grad_norm": 0.07575183361768723,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 461
    },
    {
      "clip_ratio/high_max": 0.01126297584414715,
      "clip_ratio/high_mean": 0.002516196469514398,
      "clip_ratio/low_mean": 0.002540492926073057,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005056689302818995,
      "epoch": 0.2291009677594316,
      "grad_norm": 0.07227878272533417,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 462
    },
    {
      "clip_ratio/high_max": 0.012443004889064468,
      "clip_ratio/high_mean": 0.002815927000483498,
      "clip_ratio/low_mean": 0.0029309915473731962,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005746918522163469,
      "epoch": 0.22959685730003643,
      "grad_norm": 0.06670142710208893,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 463
    },
    {
      "clip_ratio/high_max": 0.013598087738500908,
      "clip_ratio/high_mean": 0.003080658131693781,
      "clip_ratio/low_mean": 0.0032570954804214125,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006337753649404476,
      "epoch": 0.23009274684064124,
      "grad_norm": 0.061638686805963516,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 464
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1004464285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15221.0,
      "completions/mean_length": 4371.3125,
      "completions/mean_terminated_length": 3144.287841796875,
      "completions/min_length": 617.0,
      "completions/min_terminated_length": 617.0,
      "epoch": 0.23058863638124608,
      "grad_norm": 0.12249234318733215,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 76578464.0,
      "reward": 0.5022321939468384,
      "reward_std": 0.2633311152458191,
      "rewards/accuracy_reward/mean": 0.5022321343421936,
      "rewards/accuracy_reward/std": 0.5005539655685425,
      "step": 465
    },
    {
      "clip_ratio/high_max": 0.003436656520534598,
      "clip_ratio/high_mean": 0.0008248511497868094,
      "clip_ratio/low_mean": 0.0006392696876105219,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014641208249486226,
      "epoch": 0.23108452592185091,
      "grad_norm": 0.1011197492480278,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 466
    },
    {
      "clip_ratio/high_max": 0.004296709959817235,
      "clip_ratio/high_mean": 0.0010076274168113741,
      "clip_ratio/low_mean": 0.0007941094604007048,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018017368888649798,
      "epoch": 0.23158041546245572,
      "grad_norm": 0.09222371876239777,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 467
    },
    {
      "clip_ratio/high_max": 0.005345331259377417,
      "clip_ratio/high_mean": 0.0012185637947368377,
      "clip_ratio/low_mean": 0.001112318552287661,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023308823328989092,
      "epoch": 0.23207630500306056,
      "grad_norm": 0.08248686045408249,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 468
    },
    {
      "clip_ratio/high_max": 0.006682626499241451,
      "clip_ratio/high_mean": 0.001534512062335125,
      "clip_ratio/low_mean": 0.0014052871130729727,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002939799165687873,
      "epoch": 0.2325721945436654,
      "grad_norm": 0.07556495815515518,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 469
    },
    {
      "clip_ratio/high_max": 0.007987351764313644,
      "clip_ratio/high_mean": 0.001804000394713512,
      "clip_ratio/low_mean": 0.001693685533439293,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034976859351445455,
      "epoch": 0.23306808408427024,
      "grad_norm": 0.06272947788238525,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 470
    },
    {
      "clip_ratio/high_max": 0.008962880077888258,
      "clip_ratio/high_mean": 0.002051796031082631,
      "clip_ratio/low_mean": 0.0019696422285733206,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004021438280688017,
      "epoch": 0.23356397362487505,
      "grad_norm": 0.05313817039132118,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 471
    },
    {
      "clip_ratio/high_max": 0.009758148065884598,
      "clip_ratio/high_mean": 0.0022374426289388794,
      "clip_ratio/low_mean": 0.002268601463242703,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0045060440606903285,
      "epoch": 0.2340598631654799,
      "grad_norm": 0.04962389916181564,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 472
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14952.0,
      "completions/mean_length": 3822.67431640625,
      "completions/mean_terminated_length": 2844.934814453125,
      "completions/min_length": 460.0,
      "completions/min_terminated_length": 460.0,
      "epoch": 0.23455575270608472,
      "grad_norm": 0.15953747928142548,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 78363478.0,
      "reward": 0.4910714626312256,
      "reward_std": 0.25730663537979126,
      "rewards/accuracy_reward/mean": 0.4910714328289032,
      "rewards/accuracy_reward/std": 0.5004791617393494,
      "step": 473
    },
    {
      "clip_ratio/high_max": 0.004754895212499832,
      "clip_ratio/high_mean": 0.0010136481109839224,
      "clip_ratio/low_mean": 0.0009114771769986874,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001925125278376072,
      "epoch": 0.23505164224668956,
      "grad_norm": 0.13997046649456024,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 474
    },
    {
      "clip_ratio/high_max": 0.006309939044513158,
      "clip_ratio/high_mean": 0.0013360954228573974,
      "clip_ratio/low_mean": 0.0013086503613521927,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026447457921676687,
      "epoch": 0.23554753178729437,
      "grad_norm": 0.10134938359260559,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 475
    },
    {
      "clip_ratio/high_max": 0.0074433176650927635,
      "clip_ratio/high_mean": 0.0015674490045967104,
      "clip_ratio/low_mean": 0.0018143007044955084,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003381749726486305,
      "epoch": 0.2360434213278992,
      "grad_norm": 0.0961502268910408,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 476
    },
    {
      "clip_ratio/high_max": 0.009113674499531044,
      "clip_ratio/high_mean": 0.0019357294611381803,
      "clip_ratio/low_mean": 0.0022902121864945,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004225941636832431,
      "epoch": 0.23653931086850405,
      "grad_norm": 0.07644295692443848,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 477
    },
    {
      "clip_ratio/high_max": 0.010566608267254196,
      "clip_ratio/high_mean": 0.002212565500258279,
      "clip_ratio/low_mean": 0.002761938217986426,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004974503714038292,
      "epoch": 0.23703520040910886,
      "grad_norm": 0.06998546421527863,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 478
    },
    {
      "clip_ratio/high_max": 0.011664411178571754,
      "clip_ratio/high_mean": 0.002485709392658464,
      "clip_ratio/low_mean": 0.00313910661270711,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005624816011732037,
      "epoch": 0.2375310899497137,
      "grad_norm": 0.05778935179114342,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 479
    },
    {
      "clip_ratio/high_max": 0.01253418052874622,
      "clip_ratio/high_mean": 0.002712228658992899,
      "clip_ratio/low_mean": 0.0035245080216554925,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006236736657228903,
      "epoch": 0.23802697949031854,
      "grad_norm": 0.055169835686683655,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 480
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0915178571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15049.0,
      "completions/mean_length": 3985.013671875,
      "completions/mean_terminated_length": 2839.130126953125,
      "completions/min_length": 473.0,
      "completions/min_terminated_length": 473.0,
      "epoch": 0.23852286903092337,
      "grad_norm": 0.16006724536418915,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 80220500.0,
      "reward": 0.439732164144516,
      "reward_std": 0.23070621490478516,
      "rewards/accuracy_reward/mean": 0.4397321343421936,
      "rewards/accuracy_reward/std": 0.49690937995910645,
      "step": 481
    },
    {
      "clip_ratio/high_max": 0.004883393130512559,
      "clip_ratio/high_mean": 0.0010071629571939411,
      "clip_ratio/low_mean": 0.0010930865490763608,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002100249526392872,
      "epoch": 0.23901875857152818,
      "grad_norm": 0.1354125291109085,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 482
    },
    {
      "clip_ratio/high_max": 0.006939107506696018,
      "clip_ratio/high_mean": 0.0013538359303311154,
      "clip_ratio/low_mean": 0.001489013804643946,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028428497616914683,
      "epoch": 0.23951464811213302,
      "grad_norm": 0.10562701523303986,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 483
    },
    {
      "clip_ratio/high_max": 0.008818808913929388,
      "clip_ratio/high_mean": 0.0017517810642857512,
      "clip_ratio/low_mean": 0.002028367436537337,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037801485141244484,
      "epoch": 0.24001053765273786,
      "grad_norm": 0.09243745356798172,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 484
    },
    {
      "clip_ratio/high_max": 0.010838732803676976,
      "clip_ratio/high_mean": 0.002088199913941935,
      "clip_ratio/low_mean": 0.0025605361394127613,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004648736056878988,
      "epoch": 0.24050642719334267,
      "grad_norm": 0.08058091253042221,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 485
    },
    {
      "clip_ratio/high_max": 0.0125474270516861,
      "clip_ratio/high_mean": 0.0024351198107979144,
      "clip_ratio/low_mean": 0.003094899724374045,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005530019586331036,
      "epoch": 0.2410023167339475,
      "grad_norm": 0.0689515769481659,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 486
    },
    {
      "clip_ratio/high_max": 0.014113937628280837,
      "clip_ratio/high_mean": 0.0027221355003348435,
      "clip_ratio/low_mean": 0.003561490661127209,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006283626141339482,
      "epoch": 0.24149820627455235,
      "grad_norm": 0.06269052624702454,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 487
    },
    {
      "clip_ratio/high_max": 0.015424532448378159,
      "clip_ratio/high_mean": 0.0029952066010991985,
      "clip_ratio/low_mean": 0.004070159170851184,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0070653659086019616,
      "epoch": 0.24199409581515718,
      "grad_norm": 0.0609232634305954,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 488
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1049107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14840.0,
      "completions/mean_length": 3724.4130859375,
      "completions/mean_terminated_length": 2360.640869140625,
      "completions/min_length": 460.0,
      "completions/min_terminated_length": 460.0,
      "epoch": 0.242489985355762,
      "grad_norm": 0.16524240374565125,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 81965725.0,
      "reward": 0.6049107313156128,
      "reward_std": 0.1893800050020218,
      "rewards/accuracy_reward/mean": 0.6049107313156128,
      "rewards/accuracy_reward/std": 0.4894163906574249,
      "step": 489
    },
    {
      "clip_ratio/high_max": 0.004454094676475506,
      "clip_ratio/high_mean": 0.0010590586639409594,
      "clip_ratio/low_mean": 0.0006380485616546139,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016971072225260286,
      "epoch": 0.24298587489636683,
      "grad_norm": 0.10527802258729935,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 490
    },
    {
      "clip_ratio/high_max": 0.00599363890796667,
      "clip_ratio/high_mean": 0.0014016432630796771,
      "clip_ratio/low_mean": 0.0009028448275785195,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00230448811282713,
      "epoch": 0.24348176443697167,
      "grad_norm": 0.08470255881547928,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 491
    },
    {
      "clip_ratio/high_max": 0.0074015626014443114,
      "clip_ratio/high_mean": 0.001762600811161974,
      "clip_ratio/low_mean": 0.0012003649637790659,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029629657968825995,
      "epoch": 0.2439776539775765,
      "grad_norm": 0.07405374944210052,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 492
    },
    {
      "clip_ratio/high_max": 0.009289774549870344,
      "clip_ratio/high_mean": 0.0021911536414336297,
      "clip_ratio/low_mean": 0.0014661368295492139,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036572904891727376,
      "epoch": 0.24447354351818132,
      "grad_norm": 0.06254327297210693,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 493
    },
    {
      "clip_ratio/high_max": 0.010575361777227954,
      "clip_ratio/high_mean": 0.0024851206869698217,
      "clip_ratio/low_mean": 0.0017582630841843638,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004243383769335196,
      "epoch": 0.24496943305878616,
      "grad_norm": 0.05737098306417465,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 494
    },
    {
      "clip_ratio/high_max": 0.01176126077734807,
      "clip_ratio/high_mean": 0.002798127993401067,
      "clip_ratio/low_mean": 0.001975622172039948,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0047737501968185825,
      "epoch": 0.245465322599391,
      "grad_norm": 0.05171648785471916,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 495
    },
    {
      "clip_ratio/high_max": 0.01276100261202373,
      "clip_ratio/high_mean": 0.003066246964635866,
      "clip_ratio/low_mean": 0.0022294173041927934,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005295664287018553,
      "epoch": 0.2459612121399958,
      "grad_norm": 0.04542279615998268,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 496
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13909.0,
      "completions/mean_length": 3801.801513671875,
      "completions/mean_terminated_length": 2822.293212890625,
      "completions/min_length": 469.0,
      "completions/min_terminated_length": 469.0,
      "epoch": 0.24645710168060064,
      "grad_norm": 0.14687815308570862,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 83733740.0,
      "reward": 0.609375,
      "reward_std": 0.1914825141429901,
      "rewards/accuracy_reward/mean": 0.609375,
      "rewards/accuracy_reward/std": 0.48843589425086975,
      "step": 497
    },
    {
      "clip_ratio/high_max": 0.004469586749110022,
      "clip_ratio/high_mean": 0.0009459532691380446,
      "clip_ratio/low_mean": 0.0005306131718043616,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001476566453675332,
      "epoch": 0.24695299122120548,
      "grad_norm": 0.10114692896604538,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 498
    },
    {
      "clip_ratio/high_max": 0.005757172179073677,
      "clip_ratio/high_mean": 0.0012318215592586057,
      "clip_ratio/low_mean": 0.0007434702134219151,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001975291785583977,
      "epoch": 0.24744888076181032,
      "grad_norm": 0.08469536900520325,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 499
    },
    {
      "clip_ratio/high_max": 0.006943546783077181,
      "clip_ratio/high_mean": 0.0014651960277660692,
      "clip_ratio/low_mean": 0.0009939497286950427,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024591457572569198,
      "epoch": 0.24794477030241513,
      "grad_norm": 0.12979383766651154,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 500
    },
    {
      "clip_ratio/high_max": 0.00875586569236475,
      "clip_ratio/high_mean": 0.0018325992805330316,
      "clip_ratio/low_mean": 0.0012687966731164124,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003101395942394447,
      "epoch": 0.24844065984301997,
      "grad_norm": 0.06260770559310913,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 501
    },
    {
      "clip_ratio/high_max": 0.010057194656837964,
      "clip_ratio/high_mean": 0.0021137679041203228,
      "clip_ratio/low_mean": 0.001455870558515926,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003569638469343772,
      "epoch": 0.2489365493836248,
      "grad_norm": 0.057483796030282974,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 502
    },
    {
      "clip_ratio/high_max": 0.011593471972446423,
      "clip_ratio/high_mean": 0.002428657067412132,
      "clip_ratio/low_mean": 0.0016682338501823324,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00409689093430643,
      "epoch": 0.24943243892422962,
      "grad_norm": 0.05195270851254463,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 503
    },
    {
      "clip_ratio/high_max": 0.012819161674997304,
      "clip_ratio/high_mean": 0.002656804804246349,
      "clip_ratio/low_mean": 0.0019081759203345428,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0045649807834706735,
      "epoch": 0.24992832846483445,
      "grad_norm": 0.04540488123893738,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 504
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09375,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14862.0,
      "completions/mean_length": 4307.88427734375,
      "completions/mean_terminated_length": 3164.5615234375,
      "completions/min_length": 473.0,
      "completions/min_terminated_length": 473.0,
      "epoch": 0.2504242180054393,
      "grad_norm": 0.14783190190792084,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 85735232.0,
      "reward": 0.5,
      "reward_std": 0.26940926909446716,
      "rewards/accuracy_reward/mean": 0.5,
      "rewards/accuracy_reward/std": 0.5005589723587036,
      "step": 505
    },
    {
      "clip_ratio/high_max": 0.004857275832364394,
      "clip_ratio/high_mean": 0.0012053648904384318,
      "clip_ratio/low_mean": 0.0007309673021893559,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019363321930541133,
      "epoch": 0.2509201075460441,
      "grad_norm": 0.10955046862363815,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 506
    },
    {
      "clip_ratio/high_max": 0.00623985481888667,
      "clip_ratio/high_mean": 0.0015178006510723208,
      "clip_ratio/low_mean": 0.000986136513489555,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002503937178516935,
      "epoch": 0.25141599708664897,
      "grad_norm": 0.09693163633346558,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 507
    },
    {
      "clip_ratio/high_max": 0.007957742896905984,
      "clip_ratio/high_mean": 0.001926314870843271,
      "clip_ratio/low_mean": 0.0014055229276266346,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003331837783662195,
      "epoch": 0.2519118866272538,
      "grad_norm": 0.08484119921922684,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 508
    },
    {
      "clip_ratio/high_max": 0.009566208427713718,
      "clip_ratio/high_mean": 0.0023176357260581426,
      "clip_ratio/low_mean": 0.0017180117996815625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0040356475410590065,
      "epoch": 0.2524077761678586,
      "grad_norm": 0.06823431700468063,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 509
    },
    {
      "clip_ratio/high_max": 0.011003718243955518,
      "clip_ratio/high_mean": 0.0026380443889593153,
      "clip_ratio/low_mean": 0.0020399380214541907,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004677982416978921,
      "epoch": 0.25290366570846345,
      "grad_norm": 0.06305266916751862,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 510
    },
    {
      "clip_ratio/high_max": 0.012029972935124533,
      "clip_ratio/high_mean": 0.002894026073818168,
      "clip_ratio/low_mean": 0.0024165432991196667,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005310569365974516,
      "epoch": 0.25339955524906826,
      "grad_norm": 0.05719885602593422,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 511
    },
    {
      "clip_ratio/high_max": 0.012978574110093177,
      "clip_ratio/high_mean": 0.0031267703100184008,
      "clip_ratio/low_mean": 0.002650991426833116,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0057777617239480605,
      "epoch": 0.2538954447896731,
      "grad_norm": 0.0630265399813652,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "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": 0.1116071428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15110.0,
      "completions/mean_length": 4265.41552734375,
      "completions/mean_terminated_length": 2871.623046875,
      "completions/min_length": 452.0,
      "completions/min_terminated_length": 452.0,
      "epoch": 0.25439133433027794,
      "grad_norm": 0.1493770182132721,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 87714810.0,
      "reward": 0.5223214626312256,
      "reward_std": 0.23439790308475494,
      "rewards/accuracy_reward/mean": 0.5223214030265808,
      "rewards/accuracy_reward/std": 0.5000599026679993,
      "step": 513
    },
    {
      "clip_ratio/high_max": 0.004341995226241124,
      "clip_ratio/high_mean": 0.0010596634576813813,
      "clip_ratio/low_mean": 0.000644405551042837,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001704069006791542,
      "epoch": 0.25488722387088275,
      "grad_norm": 0.1180933266878128,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 514
    },
    {
      "clip_ratio/high_max": 0.005806659196423425,
      "clip_ratio/high_mean": 0.0014568993010470876,
      "clip_ratio/low_mean": 0.0009864479036991725,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024433472290184,
      "epoch": 0.2553831134114876,
      "grad_norm": 0.09350404888391495,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 515
    },
    {
      "clip_ratio/high_max": 0.007218969943096454,
      "clip_ratio/high_mean": 0.0018336030454975116,
      "clip_ratio/low_mean": 0.0012733776645177386,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003106980711891083,
      "epoch": 0.2558790029520924,
      "grad_norm": 0.07985076308250427,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 516
    },
    {
      "clip_ratio/high_max": 0.008595991488618893,
      "clip_ratio/high_mean": 0.002189655959682568,
      "clip_ratio/low_mean": 0.0015888866840896299,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037785426629852736,
      "epoch": 0.25637489249269724,
      "grad_norm": 0.06508482247591019,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 517
    },
    {
      "clip_ratio/high_max": 0.009652947266658884,
      "clip_ratio/high_mean": 0.002442342116864893,
      "clip_ratio/low_mean": 0.0018642950129788005,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004306637130866875,
      "epoch": 0.2568707820333021,
      "grad_norm": 0.058969005942344666,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 518
    },
    {
      "clip_ratio/high_max": 0.010935900991171366,
      "clip_ratio/high_mean": 0.002794830161747086,
      "clip_ratio/low_mean": 0.0021454166045487,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004940246737533016,
      "epoch": 0.2573666715739069,
      "grad_norm": 0.051776688545942307,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 519
    },
    {
      "clip_ratio/high_max": 0.011798734787589638,
      "clip_ratio/high_mean": 0.003000844556481752,
      "clip_ratio/low_mean": 0.0024073543743838854,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005408198998338776,
      "epoch": 0.2578625611145117,
      "grad_norm": 0.05219506844878197,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 520
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0669642857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14794.0,
      "completions/mean_length": 3793.770263671875,
      "completions/mean_terminated_length": 2963.65771484375,
      "completions/min_length": 559.0,
      "completions/min_terminated_length": 559.0,
      "epoch": 0.2583584506551166,
      "grad_norm": 0.14981748163700104,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 89484067.0,
      "reward": 0.4732142984867096,
      "reward_std": 0.22146931290626526,
      "rewards/accuracy_reward/mean": 0.4732142984867096,
      "rewards/accuracy_reward/std": 0.4998401701450348,
      "step": 521
    },
    {
      "clip_ratio/high_max": 0.003950096517655766,
      "clip_ratio/high_mean": 0.0008219761076588838,
      "clip_ratio/low_mean": 0.0008047419958927549,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016267181131297548,
      "epoch": 0.2588543401957214,
      "grad_norm": 0.11224875599145889,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 522
    },
    {
      "clip_ratio/high_max": 0.00542423916158441,
      "clip_ratio/high_mean": 0.0011176586149304057,
      "clip_ratio/low_mean": 0.001159872121263561,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002277530696574104,
      "epoch": 0.2593502297363262,
      "grad_norm": 0.09458872675895691,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 523
    },
    {
      "clip_ratio/high_max": 0.006852434566098964,
      "clip_ratio/high_mean": 0.0014227463848328625,
      "clip_ratio/low_mean": 0.0015482412732694684,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002970987656226498,
      "epoch": 0.2598461192769311,
      "grad_norm": 0.07645193487405777,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 524
    },
    {
      "clip_ratio/high_max": 0.008351729393325513,
      "clip_ratio/high_mean": 0.0017420928070350783,
      "clip_ratio/low_mean": 0.001966794158192897,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037088869721628726,
      "epoch": 0.2603420088175359,
      "grad_norm": 0.07278440147638321,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 525
    },
    {
      "clip_ratio/high_max": 0.009593880487955175,
      "clip_ratio/high_mean": 0.001989909953863389,
      "clip_ratio/low_mean": 0.0023905666793098135,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004380476612823259,
      "epoch": 0.26083789835814075,
      "grad_norm": 0.05901899188756943,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 526
    },
    {
      "clip_ratio/high_max": 0.010696817254938651,
      "clip_ratio/high_mean": 0.0022178482513481867,
      "clip_ratio/low_mean": 0.0028444488904142418,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005062297199401655,
      "epoch": 0.26133378789874556,
      "grad_norm": 0.05158602446317673,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 527
    },
    {
      "clip_ratio/high_max": 0.011636638842901448,
      "clip_ratio/high_mean": 0.002407494481303729,
      "clip_ratio/low_mean": 0.003206009229529627,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00561350371572189,
      "epoch": 0.2618296774393504,
      "grad_norm": 0.04619830474257469,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 528
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0870535714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15152.0,
      "completions/mean_length": 4176.67431640625,
      "completions/mean_terminated_length": 3110.29345703125,
      "completions/min_length": 374.0,
      "completions/min_terminated_length": 374.0,
      "epoch": 0.26232556697995524,
      "grad_norm": 0.15256592631340027,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 91426585.0,
      "reward": 0.5178571939468384,
      "reward_std": 0.2662454843521118,
      "rewards/accuracy_reward/mean": 0.5178571343421936,
      "rewards/accuracy_reward/std": 0.5002396702766418,
      "step": 529
    },
    {
      "clip_ratio/high_max": 0.005451788043956185,
      "clip_ratio/high_mean": 0.0012903545705285069,
      "clip_ratio/low_mean": 0.000820325915583453,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021106804879309493,
      "epoch": 0.26282145652056005,
      "grad_norm": 0.12578822672367096,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 530
    },
    {
      "clip_ratio/high_max": 0.006781797324947547,
      "clip_ratio/high_mean": 0.00156719309620712,
      "clip_ratio/low_mean": 0.001269116413766369,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002836309494341549,
      "epoch": 0.26331734606116486,
      "grad_norm": 0.10137467831373215,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 531
    },
    {
      "clip_ratio/high_max": 0.008884640359610785,
      "clip_ratio/high_mean": 0.002021414271894173,
      "clip_ratio/low_mean": 0.0017886791737282692,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003810093453466834,
      "epoch": 0.2638132356017697,
      "grad_norm": 0.09026478230953217,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 532
    },
    {
      "clip_ratio/high_max": 0.010739824225311168,
      "clip_ratio/high_mean": 0.002431884740417445,
      "clip_ratio/low_mean": 0.0023689530759156696,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004800837798029534,
      "epoch": 0.26430912514237453,
      "grad_norm": 0.07889021188020706,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 533
    },
    {
      "clip_ratio/high_max": 0.012166408352641156,
      "clip_ratio/high_mean": 0.002757816926077794,
      "clip_ratio/low_mean": 0.0027618607117574356,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00551967763567518,
      "epoch": 0.26480501468297934,
      "grad_norm": 0.07322157174348831,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 534
    },
    {
      "clip_ratio/high_max": 0.01386645071397652,
      "clip_ratio/high_mean": 0.0031303205014410196,
      "clip_ratio/low_mean": 0.00310369223689122,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006234012756976881,
      "epoch": 0.2653009042235842,
      "grad_norm": 0.06080270931124687,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 535
    },
    {
      "clip_ratio/high_max": 0.015400494270579657,
      "clip_ratio/high_mean": 0.0035024892099499993,
      "clip_ratio/low_mean": 0.0035584435134978776,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007060932653985219,
      "epoch": 0.265796793764189,
      "grad_norm": 0.05285323038697243,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 536
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.049107142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14679.0,
      "completions/mean_length": 3438.6005859375,
      "completions/mean_terminated_length": 2822.94140625,
      "completions/min_length": 491.0,
      "completions/min_terminated_length": 491.0,
      "epoch": 0.26629268330479383,
      "grad_norm": 0.15915709733963013,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 93040014.0,
      "reward": 0.5446428656578064,
      "reward_std": 0.1892252415418625,
      "rewards/accuracy_reward/mean": 0.5446428656578064,
      "rewards/accuracy_reward/std": 0.49855971336364746,
      "step": 537
    },
    {
      "clip_ratio/high_max": 0.005110996489747777,
      "clip_ratio/high_mean": 0.001032715491419367,
      "clip_ratio/low_mean": 0.0006480617177544445,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016807772135507548,
      "epoch": 0.2667885728453987,
      "grad_norm": 0.11555469036102295,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 538
    },
    {
      "clip_ratio/high_max": 0.0064345237660745624,
      "clip_ratio/high_mean": 0.001332392581844033,
      "clip_ratio/low_mean": 0.0009694408434484103,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023018334113658057,
      "epoch": 0.2672844623860035,
      "grad_norm": 0.08347176015377045,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 539
    },
    {
      "clip_ratio/high_max": 0.00801094108101097,
      "clip_ratio/high_mean": 0.001611358860600376,
      "clip_ratio/low_mean": 0.0013173928526271084,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002928751717263367,
      "epoch": 0.2677803519266084,
      "grad_norm": 0.07132827490568161,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 540
    },
    {
      "clip_ratio/high_max": 0.009617876341508236,
      "clip_ratio/high_mean": 0.0019888620922756672,
      "clip_ratio/low_mean": 0.0016889007833924552,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036777628370145976,
      "epoch": 0.2682762414672132,
      "grad_norm": 0.05626622214913368,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 541
    },
    {
      "clip_ratio/high_max": 0.010640593025527778,
      "clip_ratio/high_mean": 0.0022156155482662143,
      "clip_ratio/low_mean": 0.0020090217013830625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004224637252264074,
      "epoch": 0.268772131007818,
      "grad_norm": 0.051098696887493134,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 542
    },
    {
      "clip_ratio/high_max": 0.01202059087336238,
      "clip_ratio/high_mean": 0.002464114607164447,
      "clip_ratio/low_mean": 0.002227342093988227,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00469145672286686,
      "epoch": 0.26926802054842286,
      "grad_norm": 0.045788396149873734,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 543
    },
    {
      "clip_ratio/high_max": 0.013114083063555881,
      "clip_ratio/high_mean": 0.002723263468396908,
      "clip_ratio/low_mean": 0.0025023638532957193,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005225627333857119,
      "epoch": 0.26976391008902767,
      "grad_norm": 0.045516978949308395,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 544
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0825892857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15353.0,
      "completions/mean_length": 3937.6943359375,
      "completions/mean_terminated_length": 2909.40869140625,
      "completions/min_length": 562.0,
      "completions/min_terminated_length": 562.0,
      "epoch": 0.2702597996296325,
      "grad_norm": 0.14626286923885345,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 94875861.0,
      "reward": 0.520089328289032,
      "reward_std": 0.24108001589775085,
      "rewards/accuracy_reward/mean": 0.5200892686843872,
      "rewards/accuracy_reward/std": 0.5001547932624817,
      "step": 545
    },
    {
      "clip_ratio/high_max": 0.005130915144036408,
      "clip_ratio/high_mean": 0.0011077191613821924,
      "clip_ratio/low_mean": 0.0007650917850696715,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018728109362200485,
      "epoch": 0.27075568917023735,
      "grad_norm": 0.245662122964859,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 546
    },
    {
      "clip_ratio/high_max": 0.006534100697535905,
      "clip_ratio/high_mean": 0.0014063537238939716,
      "clip_ratio/low_mean": 0.0009894766068327954,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023958303327162866,
      "epoch": 0.27125157871084216,
      "grad_norm": 0.6688205003738403,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 547
    },
    {
      "clip_ratio/high_max": 0.00817560136965767,
      "clip_ratio/high_mean": 0.0017431567882795207,
      "clip_ratio/low_mean": 0.0013993774288110217,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003142534213111503,
      "epoch": 0.27174746825144697,
      "grad_norm": 0.4002957344055176,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 548
    },
    {
      "clip_ratio/high_max": 0.009589249086275231,
      "clip_ratio/high_mean": 0.002096917062999637,
      "clip_ratio/low_mean": 0.001770351504546852,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038672685386700323,
      "epoch": 0.27224335779205183,
      "grad_norm": 0.06766475737094879,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 549
    },
    {
      "clip_ratio/high_max": 0.011577226187000633,
      "clip_ratio/high_mean": 0.0024669525389526825,
      "clip_ratio/low_mean": 0.0021057325209312694,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004572685071252636,
      "epoch": 0.27273924733265664,
      "grad_norm": 0.06359925121068954,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 550
    },
    {
      "clip_ratio/high_max": 0.01272953881561989,
      "clip_ratio/high_mean": 0.002728021335485664,
      "clip_ratio/low_mean": 0.0024506925526566192,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005178713941859314,
      "epoch": 0.2732351368732615,
      "grad_norm": 0.0575854629278183,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 551
    },
    {
      "clip_ratio/high_max": 0.013577941779658431,
      "clip_ratio/high_mean": 0.002954891333274645,
      "clip_ratio/low_mean": 0.002717137849003848,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005672029164998094,
      "epoch": 0.2737310264138663,
      "grad_norm": 0.04984666779637337,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 552
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0736607142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15327.0,
      "completions/mean_length": 3697.294921875,
      "completions/mean_terminated_length": 2769.89892578125,
      "completions/min_length": 495.0,
      "completions/min_terminated_length": 495.0,
      "epoch": 0.27422691595447113,
      "grad_norm": 0.19841612875461578,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 96595129.0,
      "reward": 0.4977678656578064,
      "reward_std": 0.2410772293806076,
      "rewards/accuracy_reward/mean": 0.4977678656578064,
      "rewards/accuracy_reward/std": 0.5005539655685425,
      "step": 553
    },
    {
      "clip_ratio/high_max": 0.006290726214501774,
      "clip_ratio/high_mean": 0.001576083955114882,
      "clip_ratio/low_mean": 0.000889548478994584,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024656324098941695,
      "epoch": 0.274722805495076,
      "grad_norm": 0.13250431418418884,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 554
    },
    {
      "clip_ratio/high_max": 0.008153628967193072,
      "clip_ratio/high_mean": 0.0020548494574086362,
      "clip_ratio/low_mean": 0.0013598500127045554,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034146994939874276,
      "epoch": 0.2752186950356808,
      "grad_norm": 0.11211083084344864,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 555
    },
    {
      "clip_ratio/high_max": 0.010422139803267783,
      "clip_ratio/high_mean": 0.002600651453349201,
      "clip_ratio/low_mean": 0.0018149783393255348,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004415629751292727,
      "epoch": 0.2757145845762856,
      "grad_norm": 0.0891462191939354,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 556
    },
    {
      "clip_ratio/high_max": 0.01253717457439052,
      "clip_ratio/high_mean": 0.00311663784304983,
      "clip_ratio/low_mean": 0.002398705940322543,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005515343778824899,
      "epoch": 0.2762104741168905,
      "grad_norm": 0.07669910788536072,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 557
    },
    {
      "clip_ratio/high_max": 0.014368110631039599,
      "clip_ratio/high_mean": 0.003623614041316614,
      "clip_ratio/low_mean": 0.0027884484343303484,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006412062457457068,
      "epoch": 0.2767063636574953,
      "grad_norm": 0.06892888993024826,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 558
    },
    {
      "clip_ratio/high_max": 0.015634667357517174,
      "clip_ratio/high_mean": 0.003953255948090373,
      "clip_ratio/low_mean": 0.003248407738738024,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007201663641353662,
      "epoch": 0.2772022531981001,
      "grad_norm": 0.06336460262537003,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 559
    },
    {
      "clip_ratio/high_max": 0.017049210178811336,
      "clip_ratio/high_mean": 0.004275318497093394,
      "clip_ratio/low_mean": 0.0036044309049430012,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00787974942522851,
      "epoch": 0.27769814273870497,
      "grad_norm": 0.05687689036130905,
      "learning_rate": 1e-06,
      "loss": -0.0016,
      "step": 560
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1049107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15075.0,
      "completions/mean_length": 4495.201171875,
      "completions/mean_terminated_length": 3221.770751953125,
      "completions/min_length": 515.0,
      "completions/min_terminated_length": 515.0,
      "epoch": 0.2781940322793098,
      "grad_norm": 0.1525355875492096,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 98680083.0,
      "reward": 0.5558035969734192,
      "reward_std": 0.223112553358078,
      "rewards/accuracy_reward/mean": 0.5558035969734192,
      "rewards/accuracy_reward/std": 0.4974316656589508,
      "step": 561
    },
    {
      "clip_ratio/high_max": 0.0036886919515382033,
      "clip_ratio/high_mean": 0.0008406199037835904,
      "clip_ratio/low_mean": 0.0006845760966029957,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015251960126079211,
      "epoch": 0.2786899218199146,
      "grad_norm": 0.12775088846683502,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 562
    },
    {
      "clip_ratio/high_max": 0.005117497098581225,
      "clip_ratio/high_mean": 0.0011939007490582298,
      "clip_ratio/low_mean": 0.0009567914115677922,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002150692163468193,
      "epoch": 0.27918581136051945,
      "grad_norm": 0.09355885535478592,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 563
    },
    {
      "clip_ratio/high_max": 0.006869574857773841,
      "clip_ratio/high_mean": 0.0015603783218693934,
      "clip_ratio/low_mean": 0.0014314146412175432,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029917929564362566,
      "epoch": 0.27968170090112426,
      "grad_norm": 0.09069592505693436,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 564
    },
    {
      "clip_ratio/high_max": 0.008425398458712152,
      "clip_ratio/high_mean": 0.0018967918483667745,
      "clip_ratio/low_mean": 0.001750999145087917,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036477910102803435,
      "epoch": 0.28017759044172913,
      "grad_norm": 0.0653475970029831,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 565
    },
    {
      "clip_ratio/high_max": 0.009519128880128847,
      "clip_ratio/high_mean": 0.002136293264129563,
      "clip_ratio/low_mean": 0.002058310664722285,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004194603933683538,
      "epoch": 0.28067347998233394,
      "grad_norm": 0.05737563595175743,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 566
    },
    {
      "clip_ratio/high_max": 0.010124762093255413,
      "clip_ratio/high_mean": 0.002297909740718751,
      "clip_ratio/low_mean": 0.002375402758502787,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004673312496379367,
      "epoch": 0.28116936952293875,
      "grad_norm": 0.053140509873628616,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 567
    },
    {
      "clip_ratio/high_max": 0.011225109283259371,
      "clip_ratio/high_mean": 0.0025384121568094997,
      "clip_ratio/low_mean": 0.00266051222013175,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005198924401156546,
      "epoch": 0.2816652590635436,
      "grad_norm": 0.05255644768476486,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 568
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0959821428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15135.0,
      "completions/mean_length": 4795.93310546875,
      "completions/mean_terminated_length": 3674.316162109375,
      "completions/min_length": 498.0,
      "completions/min_terminated_length": 498.0,
      "epoch": 0.2821611486041484,
      "grad_norm": 0.13389532268047333,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 100899093.0,
      "reward": 0.3995535969734192,
      "reward_std": 0.2389032542705536,
      "rewards/accuracy_reward/mean": 0.3995535671710968,
      "rewards/accuracy_reward/std": 0.49035418033599854,
      "step": 569
    },
    {
      "clip_ratio/high_max": 0.0037824977762284107,
      "clip_ratio/high_mean": 0.0007790930239934823,
      "clip_ratio/low_mean": 0.0007490004282999507,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015280934612746933,
      "epoch": 0.28265703814475324,
      "grad_norm": 0.13652947545051575,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 570
    },
    {
      "clip_ratio/high_max": 0.004969561905454611,
      "clip_ratio/high_mean": 0.00100237539140835,
      "clip_ratio/low_mean": 0.001079085924800438,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002081461332466006,
      "epoch": 0.2831529276853581,
      "grad_norm": 0.0895213782787323,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 571
    },
    {
      "clip_ratio/high_max": 0.006074626860936405,
      "clip_ratio/high_mean": 0.0012474961654334038,
      "clip_ratio/low_mean": 0.0014743387788485052,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027218349249551466,
      "epoch": 0.2836488172259629,
      "grad_norm": 0.07592698931694031,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 572
    },
    {
      "clip_ratio/high_max": 0.0073286318583996035,
      "clip_ratio/high_mean": 0.0015331430786318379,
      "clip_ratio/low_mean": 0.0018647538957452525,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033978969723307273,
      "epoch": 0.2841447067665677,
      "grad_norm": 0.06460409611463547,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 573
    },
    {
      "clip_ratio/high_max": 0.00845350976305781,
      "clip_ratio/high_mean": 0.0017534584571876621,
      "clip_ratio/low_mean": 0.0022283287579512034,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003981787226734923,
      "epoch": 0.2846405963071726,
      "grad_norm": 0.05999760329723358,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 574
    },
    {
      "clip_ratio/high_max": 0.009157386382867116,
      "clip_ratio/high_mean": 0.0019280269266346295,
      "clip_ratio/low_mean": 0.00256424967926705,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0044922765833916856,
      "epoch": 0.2851364858477774,
      "grad_norm": 0.05561729148030281,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 575
    },
    {
      "clip_ratio/high_max": 0.009848261463048402,
      "clip_ratio/high_mean": 0.002100782626257569,
      "clip_ratio/low_mean": 0.0028886563814012334,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004989439021073849,
      "epoch": 0.28563237538838226,
      "grad_norm": 0.04907912015914917,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "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": 0.0870535714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15117.0,
      "completions/mean_length": 3669.779296875,
      "completions/mean_terminated_length": 2555.063720703125,
      "completions/min_length": 531.0,
      "completions/min_terminated_length": 531.0,
      "epoch": 0.2861282649289871,
      "grad_norm": 0.17374281585216522,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 102608394.0,
      "reward": 0.4508928656578064,
      "reward_std": 0.1934325248003006,
      "rewards/accuracy_reward/mean": 0.4508928656578064,
      "rewards/accuracy_reward/std": 0.4981389045715332,
      "step": 577
    },
    {
      "clip_ratio/high_max": 0.004003773494332563,
      "clip_ratio/high_mean": 0.0008081698720161512,
      "clip_ratio/low_mean": 0.0009022203499284842,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017103902266626392,
      "epoch": 0.2866241544695919,
      "grad_norm": 0.12304027378559113,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 578
    },
    {
      "clip_ratio/high_max": 0.005771454147179611,
      "clip_ratio/high_mean": 0.0011506475971145846,
      "clip_ratio/low_mean": 0.00125187324425724,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002402520831537913,
      "epoch": 0.28712004401019675,
      "grad_norm": 0.16680306196212769,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 579
    },
    {
      "clip_ratio/high_max": 0.007371023966697976,
      "clip_ratio/high_mean": 0.0014505190911222599,
      "clip_ratio/low_mean": 0.0016646456234639118,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031151647456226783,
      "epoch": 0.28761593355080156,
      "grad_norm": 0.0920347273349762,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 580
    },
    {
      "clip_ratio/high_max": 0.008702493498276453,
      "clip_ratio/high_mean": 0.001717685885978426,
      "clip_ratio/low_mean": 0.0021492682591315315,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038669540847422468,
      "epoch": 0.28811182309140637,
      "grad_norm": 0.06943738460540771,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 581
    },
    {
      "clip_ratio/high_max": 0.009983888136048336,
      "clip_ratio/high_mean": 0.001986518015655747,
      "clip_ratio/low_mean": 0.002533799090372213,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004520317088008596,
      "epoch": 0.28860771263201124,
      "grad_norm": 0.0586155503988266,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 582
    },
    {
      "clip_ratio/high_max": 0.010875726933591068,
      "clip_ratio/high_mean": 0.0021476342362802825,
      "clip_ratio/low_mean": 0.0028724006864422336,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005020034952849528,
      "epoch": 0.28910360217261605,
      "grad_norm": 0.06827613711357117,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 583
    },
    {
      "clip_ratio/high_max": 0.011618484913924476,
      "clip_ratio/high_mean": 0.002325613173070451,
      "clip_ratio/low_mean": 0.003239650884211187,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005565264093547739,
      "epoch": 0.28959949171322086,
      "grad_norm": 0.054129231721162796,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 584
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0758928571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15247.0,
      "completions/mean_length": 4681.75927734375,
      "completions/mean_terminated_length": 3804.802001953125,
      "completions/min_length": 213.0,
      "completions/min_terminated_length": 213.0,
      "epoch": 0.2900953812538257,
      "grad_norm": 0.1667572259902954,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 104778334.0,
      "reward": 0.4843750298023224,
      "reward_std": 0.25077059864997864,
      "rewards/accuracy_reward/mean": 0.484375,
      "rewards/accuracy_reward/std": 0.500314474105835,
      "step": 585
    },
    {
      "clip_ratio/high_max": 0.0038728937488485826,
      "clip_ratio/high_mean": 0.0010025462825069553,
      "clip_ratio/low_mean": 0.0007995512098659674,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018020974773662601,
      "epoch": 0.29059127079443053,
      "grad_norm": 0.10752402245998383,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 586
    },
    {
      "clip_ratio/high_max": 0.005430833001810242,
      "clip_ratio/high_mean": 0.0013609175216515723,
      "clip_ratio/low_mean": 0.0011495534297409904,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002510470946845089,
      "epoch": 0.29108716033503534,
      "grad_norm": 0.08760404586791992,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 587
    },
    {
      "clip_ratio/high_max": 0.006453846530348528,
      "clip_ratio/high_mean": 0.0016514218032170902,
      "clip_ratio/low_mean": 0.001535644241812406,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031870660254753602,
      "epoch": 0.2915830498756402,
      "grad_norm": 0.07918339967727661,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 588
    },
    {
      "clip_ratio/high_max": 0.007999007277248893,
      "clip_ratio/high_mean": 0.0020415372532625042,
      "clip_ratio/low_mean": 0.0019082985759268922,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003949835830781012,
      "epoch": 0.292078939416245,
      "grad_norm": 0.06914311647415161,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 589
    },
    {
      "clip_ratio/high_max": 0.008792559325229377,
      "clip_ratio/high_mean": 0.0022888867988513084,
      "clip_ratio/low_mean": 0.0022367472126916255,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004525634055426053,
      "epoch": 0.2925748289568499,
      "grad_norm": 0.06592115014791489,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 590
    },
    {
      "clip_ratio/high_max": 0.009940911837475142,
      "clip_ratio/high_mean": 0.002596848179564404,
      "clip_ratio/low_mean": 0.0025330274204407033,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005129875562488451,
      "epoch": 0.2930707184974547,
      "grad_norm": 0.06255247443914413,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 591
    },
    {
      "clip_ratio/high_max": 0.01083745317737339,
      "clip_ratio/high_mean": 0.0027935304988204734,
      "clip_ratio/low_mean": 0.0027890540950465947,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0055825846279731195,
      "epoch": 0.2935666080380595,
      "grad_norm": 0.06578902900218964,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 592
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1116071428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15293.0,
      "completions/mean_length": 4385.79931640625,
      "completions/mean_terminated_length": 3007.130615234375,
      "completions/min_length": 320.0,
      "completions/min_terminated_length": 320.0,
      "epoch": 0.2940624975786644,
      "grad_norm": 0.14899121224880219,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 106815292.0,
      "reward": 0.5,
      "reward_std": 0.19419486820697784,
      "rewards/accuracy_reward/mean": 0.5,
      "rewards/accuracy_reward/std": 0.5005589723587036,
      "step": 593
    },
    {
      "clip_ratio/high_max": 0.004082401872437913,
      "clip_ratio/high_mean": 0.0008325461335516593,
      "clip_ratio/low_mean": 0.0006944156103827481,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015269617465492047,
      "epoch": 0.2945583871192692,
      "grad_norm": 0.10063159465789795,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 594
    },
    {
      "clip_ratio/high_max": 0.005890189135243418,
      "clip_ratio/high_mean": 0.0011666333330140333,
      "clip_ratio/low_mean": 0.0009624248076534059,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021290581321409263,
      "epoch": 0.295054276659874,
      "grad_norm": 0.07893205434083939,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 595
    },
    {
      "clip_ratio/high_max": 0.007213999293526285,
      "clip_ratio/high_mean": 0.0014417907775623462,
      "clip_ratio/low_mean": 0.0012655240036565374,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027073147680312104,
      "epoch": 0.29555016620047886,
      "grad_norm": 0.06929469108581543,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 596
    },
    {
      "clip_ratio/high_max": 0.008527637182851322,
      "clip_ratio/high_mean": 0.001710883275791275,
      "clip_ratio/low_mean": 0.0015690011156266337,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032798843235468667,
      "epoch": 0.29604605574108367,
      "grad_norm": 0.06370684504508972,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 597
    },
    {
      "clip_ratio/high_max": 0.00950574791568215,
      "clip_ratio/high_mean": 0.001905556669953512,
      "clip_ratio/low_mean": 0.0018612431013025343,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003766799772165541,
      "epoch": 0.2965419452816885,
      "grad_norm": 0.05125043913722038,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 598
    },
    {
      "clip_ratio/high_max": 0.010529582261369796,
      "clip_ratio/high_mean": 0.0021096975988257327,
      "clip_ratio/low_mean": 0.002179938594053965,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004289636225621507,
      "epoch": 0.29703783482229335,
      "grad_norm": 0.04632648453116417,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 599
    },
    {
      "clip_ratio/high_max": 0.011229304280277574,
      "clip_ratio/high_mean": 0.0022540790732819005,
      "clip_ratio/low_mean": 0.00246637234681657,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004720451425782812,
      "epoch": 0.29753372436289816,
      "grad_norm": 0.04573115333914757,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 600
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15204.0,
      "completions/mean_length": 4455.015625,
      "completions/mean_terminated_length": 3502.15283203125,
      "completions/min_length": 485.0,
      "completions/min_terminated_length": 485.0,
      "epoch": 0.298029613903503,
      "grad_norm": 0.17833608388900757,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 108885507.0,
      "reward": 0.4151785969734192,
      "reward_std": 0.27685460448265076,
      "rewards/accuracy_reward/mean": 0.4151785671710968,
      "rewards/accuracy_reward/std": 0.49330368638038635,
      "step": 601
    },
    {
      "clip_ratio/high_max": 0.005415241066657472,
      "clip_ratio/high_mean": 0.001277280823160254,
      "clip_ratio/low_mean": 0.0009477466774683307,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022250274982980045,
      "epoch": 0.29852550344410783,
      "grad_norm": 0.1590942144393921,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 602
    },
    {
      "clip_ratio/high_max": 0.007241536055516917,
      "clip_ratio/high_mean": 0.0016546535061934264,
      "clip_ratio/low_mean": 0.001256232715746819,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029108862281645997,
      "epoch": 0.29902139298471264,
      "grad_norm": 0.13695557415485382,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 603
    },
    {
      "clip_ratio/high_max": 0.009280456513806712,
      "clip_ratio/high_mean": 0.00210263060307625,
      "clip_ratio/low_mean": 0.001686525468187483,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037891561005380936,
      "epoch": 0.2995172825253175,
      "grad_norm": 0.09165150672197342,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 604
    },
    {
      "clip_ratio/high_max": 0.011084699795901543,
      "clip_ratio/high_mean": 0.0024927679564825667,
      "clip_ratio/low_mean": 0.002155688947652834,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004648456872928364,
      "epoch": 0.3000131720659223,
      "grad_norm": 0.0776260495185852,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 605
    },
    {
      "clip_ratio/high_max": 0.012653891626541736,
      "clip_ratio/high_mean": 0.0028188158312332234,
      "clip_ratio/low_mean": 0.002552683995929783,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005371499855755246,
      "epoch": 0.30050906160652713,
      "grad_norm": 0.06523507833480835,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 606
    },
    {
      "clip_ratio/high_max": 0.013735487522353651,
      "clip_ratio/high_mean": 0.0031192714432108914,
      "clip_ratio/low_mean": 0.0029321490338247713,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006051420457879431,
      "epoch": 0.301004951147132,
      "grad_norm": 0.06066508963704109,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 607
    },
    {
      "clip_ratio/high_max": 0.014927839034498902,
      "clip_ratio/high_mean": 0.0033864890210679732,
      "clip_ratio/low_mean": 0.003316161968598408,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006702650989609538,
      "epoch": 0.3015008406877368,
      "grad_norm": 0.05669444054365158,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 608
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.060267857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14983.0,
      "completions/mean_length": 3856.013671875,
      "completions/mean_terminated_length": 3118.228271484375,
      "completions/min_length": 412.0,
      "completions/min_terminated_length": 412.0,
      "epoch": 0.3019967302283416,
      "grad_norm": 0.156690776348114,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 110686825.0,
      "reward": 0.5446428656578064,
      "reward_std": 0.2310115396976471,
      "rewards/accuracy_reward/mean": 0.5446428656578064,
      "rewards/accuracy_reward/std": 0.49855974316596985,
      "step": 609
    },
    {
      "clip_ratio/high_max": 0.004245824191457359,
      "clip_ratio/high_mean": 0.001062123846850227,
      "clip_ratio/low_mean": 0.0006697886588540314,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017319125081485254,
      "epoch": 0.3024926197689465,
      "grad_norm": 0.1271880716085434,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 610
    },
    {
      "clip_ratio/high_max": 0.005934277691267198,
      "clip_ratio/high_mean": 0.0014074543704509779,
      "clip_ratio/low_mean": 0.0009791500798428387,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002386604482126131,
      "epoch": 0.3029885093095513,
      "grad_norm": 0.08787835389375687,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 611
    },
    {
      "clip_ratio/high_max": 0.007365619625488762,
      "clip_ratio/high_mean": 0.0017741551605467976,
      "clip_ratio/low_mean": 0.0013281415763799487,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031022967405078816,
      "epoch": 0.3034843988501561,
      "grad_norm": 0.07231655716896057,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 612
    },
    {
      "clip_ratio/high_max": 0.00850761752553808,
      "clip_ratio/high_mean": 0.002032486745974893,
      "clip_ratio/low_mean": 0.001649316901534803,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036818036178374314,
      "epoch": 0.30398028839076097,
      "grad_norm": 0.06141653656959534,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 613
    },
    {
      "clip_ratio/high_max": 0.009516204900137382,
      "clip_ratio/high_mean": 0.002286244220158551,
      "clip_ratio/low_mean": 0.001969130333236535,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004255374598869821,
      "epoch": 0.3044761779313658,
      "grad_norm": 0.05853695422410965,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 614
    },
    {
      "clip_ratio/high_max": 0.010381666685134405,
      "clip_ratio/high_mean": 0.0025411636797798565,
      "clip_ratio/low_mean": 0.002284652714706681,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004825816367883817,
      "epoch": 0.30497206747197064,
      "grad_norm": 0.05119042843580246,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 615
    },
    {
      "clip_ratio/high_max": 0.011320198264002101,
      "clip_ratio/high_mean": 0.002740189698215545,
      "clip_ratio/low_mean": 0.0025509444083127164,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005291134082654025,
      "epoch": 0.30546795701257545,
      "grad_norm": 0.04755592718720436,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 616
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0959821428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15149.0,
      "completions/mean_length": 4564.076171875,
      "completions/mean_terminated_length": 3417.842041015625,
      "completions/min_length": 573.0,
      "completions/min_terminated_length": 573.0,
      "epoch": 0.30596384655318026,
      "grad_norm": 0.14335563778877258,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 112804915.0,
      "reward": 0.392857164144516,
      "reward_std": 0.18968310952186584,
      "rewards/accuracy_reward/mean": 0.3928571343421936,
      "rewards/accuracy_reward/std": 0.4889315068721771,
      "step": 617
    },
    {
      "clip_ratio/high_max": 0.0033785533742047846,
      "clip_ratio/high_mean": 0.0007266363650160201,
      "clip_ratio/low_mean": 0.0006474020408404613,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0013740384098355207,
      "epoch": 0.30645973609378513,
      "grad_norm": 0.10379937291145325,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 618
    },
    {
      "clip_ratio/high_max": 0.004461435099074151,
      "clip_ratio/high_mean": 0.0009600561329534685,
      "clip_ratio/low_mean": 0.0009439702855615906,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001904026422039351,
      "epoch": 0.30695562563438994,
      "grad_norm": 0.10271034389734268,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 619
    },
    {
      "clip_ratio/high_max": 0.0056230138943647034,
      "clip_ratio/high_mean": 0.0012069351096215541,
      "clip_ratio/low_mean": 0.0013038613615208305,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025107964592052667,
      "epoch": 0.30745151517499475,
      "grad_norm": 0.06846829503774643,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 620
    },
    {
      "clip_ratio/high_max": 0.006631481250224169,
      "clip_ratio/high_mean": 0.0014464789564954117,
      "clip_ratio/low_mean": 0.001607147286279087,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030536262228224587,
      "epoch": 0.3079474047155996,
      "grad_norm": 0.058289941400289536,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 621
    },
    {
      "clip_ratio/high_max": 0.007384371845546411,
      "clip_ratio/high_mean": 0.0015692003507865593,
      "clip_ratio/low_mean": 0.001925962082850674,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034951624515997537,
      "epoch": 0.3084432942562044,
      "grad_norm": 0.04890524595975876,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 622
    },
    {
      "clip_ratio/high_max": 0.00792730626926641,
      "clip_ratio/high_mean": 0.0017100272143579787,
      "clip_ratio/low_mean": 0.0022301485709022018,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003940175827665371,
      "epoch": 0.30893918379680924,
      "grad_norm": 0.045743901282548904,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 623
    },
    {
      "clip_ratio/high_max": 0.00866753539594356,
      "clip_ratio/high_mean": 0.0018654739869816694,
      "clip_ratio/low_mean": 0.0024567309459371245,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004322204963045806,
      "epoch": 0.3094350733374141,
      "grad_norm": 0.04297325760126114,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 624
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1049107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15355.0,
      "completions/mean_length": 4631.466796875,
      "completions/mean_terminated_length": 3374.007568359375,
      "completions/min_length": 370.0,
      "completions/min_terminated_length": 370.0,
      "epoch": 0.3099309628780189,
      "grad_norm": 0.18953722715377808,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 114946996.0,
      "reward": 0.5245535969734192,
      "reward_std": 0.2647455036640167,
      "rewards/accuracy_reward/mean": 0.5245535969734192,
      "rewards/accuracy_reward/std": 0.49995502829551697,
      "step": 625
    },
    {
      "clip_ratio/high_max": 0.004906961250526365,
      "clip_ratio/high_mean": 0.001130426292888842,
      "clip_ratio/low_mean": 0.0007594474197532008,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001889873708023515,
      "epoch": 0.3104268524186238,
      "grad_norm": 0.17972366511821747,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 626
    },
    {
      "clip_ratio/high_max": 0.00685769809751946,
      "clip_ratio/high_mean": 0.0015536923032186678,
      "clip_ratio/low_mean": 0.0011452578297621585,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026989501552634465,
      "epoch": 0.3109227419592286,
      "grad_norm": 0.10291782021522522,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 627
    },
    {
      "clip_ratio/high_max": 0.008073494826476235,
      "clip_ratio/high_mean": 0.001843287178473929,
      "clip_ratio/low_mean": 0.0015613591168346375,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034046463306367514,
      "epoch": 0.3114186314998334,
      "grad_norm": 0.08868975937366486,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 628
    },
    {
      "clip_ratio/high_max": 0.009544137055854662,
      "clip_ratio/high_mean": 0.002136229963525693,
      "clip_ratio/low_mean": 0.0018298736531221493,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003966103646234842,
      "epoch": 0.31191452104043826,
      "grad_norm": 0.5552672147750854,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 629
    },
    {
      "clip_ratio/high_max": 0.010268668989738217,
      "clip_ratio/high_mean": 0.0023730824382255378,
      "clip_ratio/low_mean": 0.0021823669024740866,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0045554493299277965,
      "epoch": 0.3124104105810431,
      "grad_norm": 0.06833735853433609,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 630
    },
    {
      "clip_ratio/high_max": 0.01165786784622469,
      "clip_ratio/high_mean": 0.002645252219735994,
      "clip_ratio/low_mean": 0.0025720886938529475,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005217340918534319,
      "epoch": 0.3129063001216479,
      "grad_norm": 0.06418905407190323,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 631
    },
    {
      "clip_ratio/high_max": 0.012505956228778814,
      "clip_ratio/high_mean": 0.0028435836479729915,
      "clip_ratio/low_mean": 0.0028557344933801687,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0056993181351572275,
      "epoch": 0.31340218966225275,
      "grad_norm": 0.06464637815952301,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 632
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0892857142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15052.0,
      "completions/mean_length": 4880.359375,
      "completions/mean_terminated_length": 3852.94384765625,
      "completions/min_length": 552.0,
      "completions/min_terminated_length": 552.0,
      "epoch": 0.31389807920285756,
      "grad_norm": 0.1537923663854599,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 117202373.0,
      "reward": 0.5178571939468384,
      "reward_std": 0.24198651313781738,
      "rewards/accuracy_reward/mean": 0.5178571343421936,
      "rewards/accuracy_reward/std": 0.5002396702766418,
      "step": 633
    },
    {
      "clip_ratio/high_max": 0.004816387649952958,
      "clip_ratio/high_mean": 0.001062483715372764,
      "clip_ratio/low_mean": 0.0007005687679111361,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017630524635023903,
      "epoch": 0.31439396874346237,
      "grad_norm": 0.11314970999956131,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 634
    },
    {
      "clip_ratio/high_max": 0.006614141955651576,
      "clip_ratio/high_mean": 0.0014389287955509644,
      "clip_ratio/low_mean": 0.0009691579923583049,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002408086791547248,
      "epoch": 0.31488985828406724,
      "grad_norm": 0.09495263546705246,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 635
    },
    {
      "clip_ratio/high_max": 0.007992982697942352,
      "clip_ratio/high_mean": 0.0017181586893002532,
      "clip_ratio/low_mean": 0.0012310220539575312,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002949180733594403,
      "epoch": 0.31538574782467205,
      "grad_norm": 0.07521307468414307,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 636
    },
    {
      "clip_ratio/high_max": 0.009746427499521815,
      "clip_ratio/high_mean": 0.002081432294971819,
      "clip_ratio/low_mean": 0.0015452812034482122,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036267134983063443,
      "epoch": 0.31588163736527686,
      "grad_norm": 0.06367422640323639,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 637
    },
    {
      "clip_ratio/high_max": 0.010810432774633227,
      "clip_ratio/high_mean": 0.0023420776475404637,
      "clip_ratio/low_mean": 0.001882099954059413,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004224177595460787,
      "epoch": 0.3163775269058817,
      "grad_norm": 0.062004756182432175,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 638
    },
    {
      "clip_ratio/high_max": 0.011817869866717956,
      "clip_ratio/high_mean": 0.0025814310495206882,
      "clip_ratio/low_mean": 0.0021793705243453587,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004760801575685036,
      "epoch": 0.31687341644648653,
      "grad_norm": 0.052178479731082916,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 639
    },
    {
      "clip_ratio/high_max": 0.012560903360281372,
      "clip_ratio/high_mean": 0.0027464690688248083,
      "clip_ratio/low_mean": 0.0024398665168519074,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0051863355984096415,
      "epoch": 0.3173693059870914,
      "grad_norm": 0.04779940843582153,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "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": 0.0736607142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15090.0,
      "completions/mean_length": 4320.84375,
      "completions/mean_terminated_length": 3443.031494140625,
      "completions/min_length": 379.0,
      "completions/min_terminated_length": 379.0,
      "epoch": 0.3178651955276962,
      "grad_norm": 0.1865006983280182,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 119209215.0,
      "reward": 0.4910714626312256,
      "reward_std": 0.24138033390045166,
      "rewards/accuracy_reward/mean": 0.4910714328289032,
      "rewards/accuracy_reward/std": 0.5004791617393494,
      "step": 641
    },
    {
      "clip_ratio/high_max": 0.005367552598727343,
      "clip_ratio/high_mean": 0.0010444022711908474,
      "clip_ratio/low_mean": 0.0011051287431200763,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002149531010672945,
      "epoch": 0.318361085068301,
      "grad_norm": 0.12773331999778748,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 642
    },
    {
      "clip_ratio/high_max": 0.007694323810937931,
      "clip_ratio/high_mean": 0.0014566784385579012,
      "clip_ratio/low_mean": 0.001630577309470027,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030872557408656576,
      "epoch": 0.3188569746089059,
      "grad_norm": 0.10200220346450806,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 643
    },
    {
      "clip_ratio/high_max": 0.009773146874067606,
      "clip_ratio/high_mean": 0.0018137427384772309,
      "clip_ratio/low_mean": 0.0019822640921347556,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037960068566462724,
      "epoch": 0.3193528641495107,
      "grad_norm": 0.08319564163684845,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 644
    },
    {
      "clip_ratio/high_max": 0.011305692391033517,
      "clip_ratio/high_mean": 0.002115803849619624,
      "clip_ratio/low_mean": 0.0026292932018350257,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004745097040540713,
      "epoch": 0.3198487536901155,
      "grad_norm": 0.07582846283912659,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 645
    },
    {
      "clip_ratio/high_max": 0.012479840972446254,
      "clip_ratio/high_mean": 0.0023632868321783462,
      "clip_ratio/low_mean": 0.003090228006954021,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005453514850159991,
      "epoch": 0.32034464323072037,
      "grad_norm": 0.06651268154382706,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 646
    },
    {
      "clip_ratio/high_max": 0.013654399272127193,
      "clip_ratio/high_mean": 0.002599291323122088,
      "clip_ratio/low_mean": 0.003518950412171762,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006118241712101735,
      "epoch": 0.3208405327713252,
      "grad_norm": 0.07175970822572708,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 647
    },
    {
      "clip_ratio/high_max": 0.01428761533497891,
      "clip_ratio/high_mean": 0.002745986309264481,
      "clip_ratio/low_mean": 0.003922377906974361,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006668364216238842,
      "epoch": 0.32133642231193,
      "grad_norm": 0.05304485186934471,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 648
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0892857142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15297.0,
      "completions/mean_length": 4325.28125,
      "completions/mean_terminated_length": 3243.4462890625,
      "completions/min_length": 645.0,
      "completions/min_terminated_length": 645.0,
      "epoch": 0.32183231185253486,
      "grad_norm": 0.14293630421161652,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 121218085.0,
      "reward": 0.4977678656578064,
      "reward_std": 0.22154980897903442,
      "rewards/accuracy_reward/mean": 0.4977678656578064,
      "rewards/accuracy_reward/std": 0.5005539655685425,
      "step": 649
    },
    {
      "clip_ratio/high_max": 0.003911158053597319,
      "clip_ratio/high_mean": 0.0008615881711193651,
      "clip_ratio/low_mean": 0.0008563279681652602,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001717916119559959,
      "epoch": 0.32232820139313967,
      "grad_norm": 0.11448021233081818,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 650
    },
    {
      "clip_ratio/high_max": 0.004919158442135085,
      "clip_ratio/high_mean": 0.00106774552295974,
      "clip_ratio/low_mean": 0.0011288335213066603,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021965790529066,
      "epoch": 0.32282409093374453,
      "grad_norm": 0.09107816964387894,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 651
    },
    {
      "clip_ratio/high_max": 0.0064950756313919555,
      "clip_ratio/high_mean": 0.001421145861968398,
      "clip_ratio/low_mean": 0.0015349040766068356,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002956049906970293,
      "epoch": 0.32331998047434934,
      "grad_norm": 0.08218993991613388,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 652
    },
    {
      "clip_ratio/high_max": 0.007623853296536254,
      "clip_ratio/high_mean": 0.001668284397055686,
      "clip_ratio/low_mean": 0.0019592262215155642,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036275106358516496,
      "epoch": 0.32381587001495415,
      "grad_norm": 0.07819053530693054,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 653
    },
    {
      "clip_ratio/high_max": 0.008684996089868946,
      "clip_ratio/high_mean": 0.0018844681853806833,
      "clip_ratio/low_mean": 0.0023437434167590254,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004228211611916777,
      "epoch": 0.324311759555559,
      "grad_norm": 0.05633116513490677,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 654
    },
    {
      "clip_ratio/high_max": 0.009515623969491571,
      "clip_ratio/high_mean": 0.0020915050922667433,
      "clip_ratio/low_mean": 0.002684546324189796,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004776051397129777,
      "epoch": 0.32480764909616383,
      "grad_norm": 0.06837611645460129,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 655
    },
    {
      "clip_ratio/high_max": 0.0101389060546353,
      "clip_ratio/high_mean": 0.0022341593253258907,
      "clip_ratio/low_mean": 0.003004758176757605,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0052389174852578435,
      "epoch": 0.32530353863676864,
      "grad_norm": 0.04658592864871025,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 656
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09375,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14966.0,
      "completions/mean_length": 4112.607421875,
      "completions/mean_terminated_length": 2949.083740234375,
      "completions/min_length": 606.0,
      "completions/min_terminated_length": 606.0,
      "epoch": 0.3257994281773735,
      "grad_norm": 0.20744387805461884,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 123131397.0,
      "reward": 0.5245535969734192,
      "reward_std": 0.2948128283023834,
      "rewards/accuracy_reward/mean": 0.5245535969734192,
      "rewards/accuracy_reward/std": 0.49995502829551697,
      "step": 657
    },
    {
      "clip_ratio/high_max": 0.005882623359866557,
      "clip_ratio/high_mean": 0.0015073840539798766,
      "clip_ratio/low_mean": 0.000934926640923095,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024423107015536516,
      "epoch": 0.3262953177179783,
      "grad_norm": 0.1469687819480896,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 658
    },
    {
      "clip_ratio/high_max": 0.007806036732290522,
      "clip_ratio/high_mean": 0.002013251710025088,
      "clip_ratio/low_mean": 0.0014727271401540065,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003485978829303349,
      "epoch": 0.3267912072585831,
      "grad_norm": 0.11442055553197861,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 659
    },
    {
      "clip_ratio/high_max": 0.009536647308777901,
      "clip_ratio/high_mean": 0.0024569748098883792,
      "clip_ratio/low_mean": 0.0019471458934106067,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004404120710205461,
      "epoch": 0.327287096799188,
      "grad_norm": 0.0960107371211052,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 660
    },
    {
      "clip_ratio/high_max": 0.011166463070367172,
      "clip_ratio/high_mean": 0.00286776181053483,
      "clip_ratio/low_mean": 0.0024131748990612323,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0052809367025474785,
      "epoch": 0.3277829863397928,
      "grad_norm": 0.07922998070716858,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 661
    },
    {
      "clip_ratio/high_max": 0.012924326360007399,
      "clip_ratio/high_mean": 0.0033150166761970468,
      "clip_ratio/low_mean": 0.002772903235722879,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006087919919991691,
      "epoch": 0.3282788758803976,
      "grad_norm": 0.06695511192083359,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 662
    },
    {
      "clip_ratio/high_max": 0.013919287603130215,
      "clip_ratio/high_mean": 0.003615616054730708,
      "clip_ratio/low_mean": 0.0031426191325181208,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006758235158486059,
      "epoch": 0.3287747654210025,
      "grad_norm": 0.05969513580203056,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 663
    },
    {
      "clip_ratio/high_max": 0.015076924935783609,
      "clip_ratio/high_mean": 0.003908458569412687,
      "clip_ratio/low_mean": 0.0035126001365597403,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007421058660838753,
      "epoch": 0.3292706549616073,
      "grad_norm": 0.057852957397699356,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 664
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14579.0,
      "completions/mean_length": 3973.035888671875,
      "completions/mean_terminated_length": 3213.90478515625,
      "completions/min_length": 454.0,
      "completions/min_terminated_length": 454.0,
      "epoch": 0.32976654450221216,
      "grad_norm": 0.19494716823101044,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 124994133.0,
      "reward": 0.5334821939468384,
      "reward_std": 0.2903907895088196,
      "rewards/accuracy_reward/mean": 0.5334821343421936,
      "rewards/accuracy_reward/std": 0.4994353950023651,
      "step": 665
    },
    {
      "clip_ratio/high_max": 0.005248925303931173,
      "clip_ratio/high_mean": 0.0012740671838855633,
      "clip_ratio/low_mean": 0.0010801344801052437,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023542016688224976,
      "epoch": 0.33026243404281697,
      "grad_norm": 0.1259927898645401,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 666
    },
    {
      "clip_ratio/high_max": 0.006659511008365371,
      "clip_ratio/high_mean": 0.001598823303311292,
      "clip_ratio/low_mean": 0.0015225941579046776,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031214175223794882,
      "epoch": 0.3307583235834218,
      "grad_norm": 0.10057657957077026,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 667
    },
    {
      "clip_ratio/high_max": 0.008493310651829233,
      "clip_ratio/high_mean": 0.0020621863964152,
      "clip_ratio/low_mean": 0.0019434514621252674,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0040056378857116215,
      "epoch": 0.33125421312402664,
      "grad_norm": 0.09602950513362885,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 668
    },
    {
      "clip_ratio/high_max": 0.009739288745549857,
      "clip_ratio/high_mean": 0.0023905545999696187,
      "clip_ratio/low_mean": 0.0023841782758609043,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004774732895384659,
      "epoch": 0.33175010266463145,
      "grad_norm": 0.07162962108850479,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 669
    },
    {
      "clip_ratio/high_max": 0.010758462951343972,
      "clip_ratio/high_mean": 0.0026367997115812614,
      "clip_ratio/low_mean": 0.00285176277679966,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005488562488608295,
      "epoch": 0.33224599220523626,
      "grad_norm": 0.06820335984230042,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 670
    },
    {
      "clip_ratio/high_max": 0.011485564053145936,
      "clip_ratio/high_mean": 0.0028628899162868038,
      "clip_ratio/low_mean": 0.0031987442037006986,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0060616341233981075,
      "epoch": 0.33274188174584113,
      "grad_norm": 0.06063077226281166,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 671
    },
    {
      "clip_ratio/high_max": 0.01225969100050861,
      "clip_ratio/high_mean": 0.003064539798288024,
      "clip_ratio/low_mean": 0.00354637284999626,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006610912667383673,
      "epoch": 0.33323777128644594,
      "grad_norm": 0.0562279038131237,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 672
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0758928571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14443.0,
      "completions/mean_length": 3750.218994140625,
      "completions/mean_terminated_length": 2796.75830078125,
      "completions/min_length": 356.0,
      "completions/min_terminated_length": 356.0,
      "epoch": 0.33373366082705075,
      "grad_norm": 0.1675998866558075,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 126737431.0,
      "reward": 0.5892857313156128,
      "reward_std": 0.22769217193126678,
      "rewards/accuracy_reward/mean": 0.5892857313156128,
      "rewards/accuracy_reward/std": 0.4925134479999542,
      "step": 673
    },
    {
      "clip_ratio/high_max": 0.0042725367293314775,
      "clip_ratio/high_mean": 0.0009355773307788695,
      "clip_ratio/low_mean": 0.0006702608467605842,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016058381756920426,
      "epoch": 0.3342295503676556,
      "grad_norm": 0.1200026348233223,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 674
    },
    {
      "clip_ratio/high_max": 0.0059017878184022265,
      "clip_ratio/high_mean": 0.0012421780643308011,
      "clip_ratio/low_mean": 0.0010276608777530782,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002269838935262669,
      "epoch": 0.3347254399082604,
      "grad_norm": 0.0875861719250679,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 675
    },
    {
      "clip_ratio/high_max": 0.00735165249898273,
      "clip_ratio/high_mean": 0.001555234531451788,
      "clip_ratio/low_mean": 0.0014237943394164176,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029790288672302268,
      "epoch": 0.3352213294488653,
      "grad_norm": 0.07213940471410751,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 676
    },
    {
      "clip_ratio/high_max": 0.008849543908581836,
      "clip_ratio/high_mean": 0.0018436296631989535,
      "clip_ratio/low_mean": 0.0017912037639007394,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003634833417891059,
      "epoch": 0.3357172189894701,
      "grad_norm": 0.06718581914901733,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 677
    },
    {
      "clip_ratio/high_max": 0.009405151817190927,
      "clip_ratio/high_mean": 0.001969405224372167,
      "clip_ratio/low_mean": 0.002075841003943424,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004045246200803376,
      "epoch": 0.3362131085300749,
      "grad_norm": 0.06045062467455864,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 678
    },
    {
      "clip_ratio/high_max": 0.010586051277641673,
      "clip_ratio/high_mean": 0.002229157984857011,
      "clip_ratio/low_mean": 0.0023584234135114457,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004587581370287808,
      "epoch": 0.3367089980706798,
      "grad_norm": 0.05438857525587082,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 679
    },
    {
      "clip_ratio/high_max": 0.010991409752023173,
      "clip_ratio/high_mean": 0.002353765540647146,
      "clip_ratio/low_mean": 0.002636528151242601,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004990293686205405,
      "epoch": 0.3372048876112846,
      "grad_norm": 0.049491193145513535,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 680
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15332.0,
      "completions/mean_length": 4357.92431640625,
      "completions/mean_terminated_length": 3396.57763671875,
      "completions/min_length": 617.0,
      "completions/min_terminated_length": 617.0,
      "epoch": 0.3377007771518894,
      "grad_norm": 0.14952437579631805,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 128757677.0,
      "reward": 0.5,
      "reward_std": 0.22846665978431702,
      "rewards/accuracy_reward/mean": 0.5,
      "rewards/accuracy_reward/std": 0.5005589723587036,
      "step": 681
    },
    {
      "clip_ratio/high_max": 0.004651974599255482,
      "clip_ratio/high_mean": 0.0010732276491580706,
      "clip_ratio/low_mean": 0.0007689850019687583,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018422126554469287,
      "epoch": 0.33819666669249426,
      "grad_norm": 18.12948989868164,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "step": 682
    },
    {
      "clip_ratio/high_max": 0.00534556259844976,
      "clip_ratio/high_mean": 0.0012541411056190555,
      "clip_ratio/low_mean": 0.0009993745723022585,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002253515684969898,
      "epoch": 0.3386925562330991,
      "grad_norm": 0.11014117300510406,
      "learning_rate": 1e-06,
      "loss": 0.0031,
      "step": 683
    },
    {
      "clip_ratio/high_max": 0.006483786105491163,
      "clip_ratio/high_mean": 0.0014960457067445532,
      "clip_ratio/low_mean": 0.0012830486284656217,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027790943249783595,
      "epoch": 0.3391884457737039,
      "grad_norm": 0.09230881929397583,
      "learning_rate": 1e-06,
      "loss": 0.0029,
      "step": 684
    },
    {
      "clip_ratio/high_max": 0.007635143692823476,
      "clip_ratio/high_mean": 0.0017589562389730418,
      "clip_ratio/low_mean": 0.001582868769446577,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033418249995520455,
      "epoch": 0.33968433531430875,
      "grad_norm": 0.07602237164974213,
      "learning_rate": 1e-06,
      "loss": 0.0027,
      "step": 685
    },
    {
      "clip_ratio/high_max": 0.008678274256453733,
      "clip_ratio/high_mean": 0.002010476978512088,
      "clip_ratio/low_mean": 0.0019329948156610044,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00394347179280885,
      "epoch": 0.34018022485491356,
      "grad_norm": 0.0713312178850174,
      "learning_rate": 1e-06,
      "loss": 0.0025,
      "step": 686
    },
    {
      "clip_ratio/high_max": 0.009660058554800344,
      "clip_ratio/high_mean": 0.0022611537267493986,
      "clip_ratio/low_mean": 0.0022447817896136257,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0045059354861223255,
      "epoch": 0.3406761143955184,
      "grad_norm": 0.06009386107325554,
      "learning_rate": 1e-06,
      "loss": 0.0024,
      "step": 687
    },
    {
      "clip_ratio/high_max": 0.010444182113133138,
      "clip_ratio/high_mean": 0.0024341394994280563,
      "clip_ratio/low_mean": 0.0025834569482867664,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005017596444304218,
      "epoch": 0.34117200393612324,
      "grad_norm": 0.05558209866285324,
      "learning_rate": 1e-06,
      "loss": 0.0023,
      "step": 688
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0758928571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14989.0,
      "completions/mean_length": 4068.34619140625,
      "completions/mean_terminated_length": 3141.011962890625,
      "completions/min_length": 340.0,
      "completions/min_terminated_length": 340.0,
      "epoch": 0.34166789347672805,
      "grad_norm": 0.19211696088314056,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 130646472.0,
      "reward": 0.4799107313156128,
      "reward_std": 0.18111374974250793,
      "rewards/accuracy_reward/mean": 0.4799107015132904,
      "rewards/accuracy_reward/std": 0.5001547932624817,
      "step": 689
    },
    {
      "clip_ratio/high_max": 0.004094927198821097,
      "clip_ratio/high_mean": 0.0008146479827928488,
      "clip_ratio/low_mean": 0.0007372933937972448,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015519413943820837,
      "epoch": 0.3421637830173329,
      "grad_norm": 0.17597569525241852,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 690
    },
    {
      "clip_ratio/high_max": 0.005588866401012638,
      "clip_ratio/high_mean": 0.0011150000091220136,
      "clip_ratio/low_mean": 0.0010515504198451708,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021665504357315513,
      "epoch": 0.3426596725579377,
      "grad_norm": 0.09394050389528275,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 691
    },
    {
      "clip_ratio/high_max": 0.006899865031300578,
      "clip_ratio/high_mean": 0.0013960876783585263,
      "clip_ratio/low_mean": 0.0013912776338429467,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027873653284586908,
      "epoch": 0.34315556209854253,
      "grad_norm": 0.06896509230136871,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 692
    },
    {
      "clip_ratio/high_max": 0.008063922037763405,
      "clip_ratio/high_mean": 0.0016326139442526255,
      "clip_ratio/low_mean": 0.0017616868389040974,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003394300773379655,
      "epoch": 0.3436514516391474,
      "grad_norm": 0.06277701258659363,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 693
    },
    {
      "clip_ratio/high_max": 0.009032347043103073,
      "clip_ratio/high_mean": 0.0018587334618587192,
      "clip_ratio/low_mean": 0.0020551393000687312,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00391387276158639,
      "epoch": 0.3441473411797522,
      "grad_norm": 0.05819186195731163,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 694
    },
    {
      "clip_ratio/high_max": 0.009996168768338975,
      "clip_ratio/high_mean": 0.002049741413884476,
      "clip_ratio/low_mean": 0.0023516853734690812,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004401426779850226,
      "epoch": 0.344643230720357,
      "grad_norm": 0.04919612035155296,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 695
    },
    {
      "clip_ratio/high_max": 0.010728351495345123,
      "clip_ratio/high_mean": 0.002211906750289927,
      "clip_ratio/low_mean": 0.002657439114727822,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0048693458611523965,
      "epoch": 0.3451391202609619,
      "grad_norm": 0.06544066220521927,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 696
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15323.0,
      "completions/mean_length": 3997.46435546875,
      "completions/mean_terminated_length": 3034.53759765625,
      "completions/min_length": 437.0,
      "completions/min_terminated_length": 437.0,
      "epoch": 0.3456350098015667,
      "grad_norm": 0.205281063914299,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 132506080.0,
      "reward": 0.4888392984867096,
      "reward_std": 0.24174758791923523,
      "rewards/accuracy_reward/mean": 0.4888392984867096,
      "rewards/accuracy_reward/std": 0.5004342794418335,
      "step": 697
    },
    {
      "clip_ratio/high_max": 0.005233985608356306,
      "clip_ratio/high_mean": 0.0011656030683298013,
      "clip_ratio/low_mean": 0.000864940742417275,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020305437819843064,
      "epoch": 0.3461308993421715,
      "grad_norm": 0.28875499963760376,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 698
    },
    {
      "clip_ratio/high_max": 0.007223539478218299,
      "clip_ratio/high_mean": 0.0015718694867246086,
      "clip_ratio/low_mean": 0.0012807776516865488,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002852647147847165,
      "epoch": 0.34662678888277637,
      "grad_norm": 0.09991331398487091,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 699
    },
    {
      "clip_ratio/high_max": 0.008691559620274347,
      "clip_ratio/high_mean": 0.0019232296440350183,
      "clip_ratio/low_mean": 0.0016729679921354546,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035961976386715833,
      "epoch": 0.3471226784233812,
      "grad_norm": 0.08459225296974182,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 700
    },
    {
      "clip_ratio/high_max": 0.00988080024944793,
      "clip_ratio/high_mean": 0.002234746963949874,
      "clip_ratio/low_mean": 0.002079171721561579,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004313918727802957,
      "epoch": 0.34761856796398605,
      "grad_norm": 0.07736601680517197,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 701
    },
    {
      "clip_ratio/high_max": 0.011188218562892871,
      "clip_ratio/high_mean": 0.00251442307171601,
      "clip_ratio/low_mean": 0.002426991577976878,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004941414603308658,
      "epoch": 0.34811445750459086,
      "grad_norm": 0.06576650589704514,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 702
    },
    {
      "clip_ratio/high_max": 0.012093732304492733,
      "clip_ratio/high_mean": 0.0027313944979141525,
      "clip_ratio/low_mean": 0.002801314383304998,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00553270885211532,
      "epoch": 0.34861034704519567,
      "grad_norm": 0.05892506614327431,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 703
    },
    {
      "clip_ratio/high_max": 0.012887891210993985,
      "clip_ratio/high_mean": 0.002912992472829501,
      "clip_ratio/low_mean": 0.0031254340940449765,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006038426533450547,
      "epoch": 0.34910623658580053,
      "grad_norm": 0.0522688552737236,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "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": 0.0892857142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15286.0,
      "completions/mean_length": 4582.390625,
      "completions/mean_terminated_length": 3525.762451171875,
      "completions/min_length": 572.0,
      "completions/min_terminated_length": 572.0,
      "epoch": 0.34960212612640534,
      "grad_norm": 0.1413990557193756,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 134633383.0,
      "reward": 0.4642857313156128,
      "reward_std": 0.21079377830028534,
      "rewards/accuracy_reward/mean": 0.4642857015132904,
      "rewards/accuracy_reward/std": 0.4992803633213043,
      "step": 705
    },
    {
      "clip_ratio/high_max": 0.0035929256519011687,
      "clip_ratio/high_mean": 0.0007802902096045727,
      "clip_ratio/low_mean": 0.0007279429333379994,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015082331283338135,
      "epoch": 0.35009801566701015,
      "grad_norm": 0.09104672074317932,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 706
    },
    {
      "clip_ratio/high_max": 0.004383167328342097,
      "clip_ratio/high_mean": 0.0009400261810696975,
      "clip_ratio/low_mean": 0.0009433393903464093,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018833655572052521,
      "epoch": 0.350593905207615,
      "grad_norm": 0.07467487454414368,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 707
    },
    {
      "clip_ratio/high_max": 0.005502809599420289,
      "clip_ratio/high_mean": 0.0011873411729084182,
      "clip_ratio/low_mean": 0.0012850855528654392,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002472426706844999,
      "epoch": 0.35108979474821983,
      "grad_norm": 0.06616140902042389,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 708
    },
    {
      "clip_ratio/high_max": 0.0063785670663492056,
      "clip_ratio/high_mean": 0.001361531843940611,
      "clip_ratio/low_mean": 0.0015226474901055553,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002884179332795611,
      "epoch": 0.35158568428882464,
      "grad_norm": 0.0565788596868515,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 709
    },
    {
      "clip_ratio/high_max": 0.007250101216413896,
      "clip_ratio/high_mean": 0.0015367774262813327,
      "clip_ratio/low_mean": 0.001798473327994543,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003335250735290174,
      "epoch": 0.3520815738294295,
      "grad_norm": 0.05350051447749138,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 710
    },
    {
      "clip_ratio/high_max": 0.00792662753519835,
      "clip_ratio/high_mean": 0.0016884340038814116,
      "clip_ratio/low_mean": 0.0020567957543562443,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037452297619893216,
      "epoch": 0.3525774633700343,
      "grad_norm": 0.05233151465654373,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 711
    },
    {
      "clip_ratio/high_max": 0.008530078164767474,
      "clip_ratio/high_mean": 0.0018357960002504115,
      "clip_ratio/low_mean": 0.002218278471332269,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0040540745030739345,
      "epoch": 0.3530733529106392,
      "grad_norm": 0.040067799389362335,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 712
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0580357142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13628.0,
      "completions/mean_length": 3604.4912109375,
      "completions/mean_terminated_length": 2880.218017578125,
      "completions/min_length": 459.0,
      "completions/min_terminated_length": 459.0,
      "epoch": 0.353569242451244,
      "grad_norm": 0.1639689952135086,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 136316763.0,
      "reward": 0.6227678656578064,
      "reward_std": 0.21936315298080444,
      "rewards/accuracy_reward/mean": 0.6227678656578064,
      "rewards/accuracy_reward/std": 0.4852356016635895,
      "step": 713
    },
    {
      "clip_ratio/high_max": 0.0035107492130919127,
      "clip_ratio/high_mean": 0.000782873902380743,
      "clip_ratio/low_mean": 0.0006413061908574491,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014241801045500324,
      "epoch": 0.3540651319918488,
      "grad_norm": 0.14966607093811035,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 714
    },
    {
      "clip_ratio/high_max": 0.004894256499028415,
      "clip_ratio/high_mean": 0.0010943594593300077,
      "clip_ratio/low_mean": 0.0010012835838892897,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020956430334990728,
      "epoch": 0.35456102153245367,
      "grad_norm": 0.08262798190116882,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 715
    },
    {
      "clip_ratio/high_max": 0.005676330898495507,
      "clip_ratio/high_mean": 0.0012465750178307644,
      "clip_ratio/low_mean": 0.0013203582187770735,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025669332353572827,
      "epoch": 0.3550569110730585,
      "grad_norm": 0.06557364016771317,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 716
    },
    {
      "clip_ratio/high_max": 0.006658929811237613,
      "clip_ratio/high_mean": 0.001470733035603189,
      "clip_ratio/low_mean": 0.0015929685414448613,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030637015970569337,
      "epoch": 0.3555528006136633,
      "grad_norm": 0.06385992467403412,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 717
    },
    {
      "clip_ratio/high_max": 0.007011306624917779,
      "clip_ratio/high_mean": 0.0016106526613839378,
      "clip_ratio/low_mean": 0.001887109018070987,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003497761677863309,
      "epoch": 0.35604869015426815,
      "grad_norm": 0.05100933834910393,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 718
    },
    {
      "clip_ratio/high_max": 0.007671306728298077,
      "clip_ratio/high_mean": 0.0017525139710414805,
      "clip_ratio/low_mean": 0.002140672758173423,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038931867147766752,
      "epoch": 0.35654457969487297,
      "grad_norm": 0.04617922008037567,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 719
    },
    {
      "clip_ratio/high_max": 0.007940828254504595,
      "clip_ratio/high_mean": 0.0018220564197690692,
      "clip_ratio/low_mean": 0.002339029566314821,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004161085971645662,
      "epoch": 0.3570404692354778,
      "grad_norm": 0.045129794627428055,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 720
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0758928571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14701.0,
      "completions/mean_length": 3849.72119140625,
      "completions/mean_terminated_length": 2904.432373046875,
      "completions/min_length": 537.0,
      "completions/min_terminated_length": 537.0,
      "epoch": 0.35753635877608264,
      "grad_norm": 0.24126586318016052,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 138107166.0,
      "reward": 0.5848214626312256,
      "reward_std": 0.29038938879966736,
      "rewards/accuracy_reward/mean": 0.5848214030265808,
      "rewards/accuracy_reward/std": 0.49330365657806396,
      "step": 721
    },
    {
      "clip_ratio/high_max": 0.006038854968210217,
      "clip_ratio/high_mean": 0.0016797125023231274,
      "clip_ratio/low_mean": 0.0010568451984909188,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027365576825104654,
      "epoch": 0.35803224831668745,
      "grad_norm": 0.16425463557243347,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 722
    },
    {
      "clip_ratio/high_max": 0.007900510301624308,
      "clip_ratio/high_mean": 0.0021979054745315807,
      "clip_ratio/low_mean": 0.001599072969156623,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037969784352753777,
      "epoch": 0.35852813785729226,
      "grad_norm": 0.10588867962360382,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 723
    },
    {
      "clip_ratio/high_max": 0.009415073131094687,
      "clip_ratio/high_mean": 0.002617165477204253,
      "clip_ratio/low_mean": 0.002081082518770927,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004698247968008218,
      "epoch": 0.3590240273978971,
      "grad_norm": 0.09237731993198395,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 724
    },
    {
      "clip_ratio/high_max": 0.011255758025072282,
      "clip_ratio/high_mean": 0.0031455795938200026,
      "clip_ratio/low_mean": 0.0026357749602539116,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005781354631835711,
      "epoch": 0.35951991693850194,
      "grad_norm": 0.08081110566854477,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 725
    },
    {
      "clip_ratio/high_max": 0.012172036258562002,
      "clip_ratio/high_mean": 0.003387996964647755,
      "clip_ratio/low_mean": 0.003038867100713105,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006426864116292563,
      "epoch": 0.3600158064791068,
      "grad_norm": 0.07359214127063751,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 726
    },
    {
      "clip_ratio/high_max": 0.01386896589610842,
      "clip_ratio/high_mean": 0.003799766264819482,
      "clip_ratio/low_mean": 0.0034574555015751685,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0072572217368360725,
      "epoch": 0.3605116960197116,
      "grad_norm": 0.06517906486988068,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 727
    },
    {
      "clip_ratio/high_max": 0.014918616561772069,
      "clip_ratio/high_mean": 0.004066121494361141,
      "clip_ratio/low_mean": 0.0037813946514688723,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007847516062611248,
      "epoch": 0.3610075855603164,
      "grad_norm": 0.058299802243709564,
      "learning_rate": 1e-06,
      "loss": -0.0015,
      "step": 728
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1339285714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15313.0,
      "completions/mean_length": 5117.25927734375,
      "completions/mean_terminated_length": 3533.329833984375,
      "completions/min_length": 536.0,
      "completions/min_terminated_length": 536.0,
      "epoch": 0.3615034751009213,
      "grad_norm": 0.1800694465637207,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 140467586.0,
      "reward": 0.4933035969734192,
      "reward_std": 0.26025426387786865,
      "rewards/accuracy_reward/mean": 0.4933035671710968,
      "rewards/accuracy_reward/std": 0.5005140900611877,
      "step": 729
    },
    {
      "clip_ratio/high_max": 0.005063509799583699,
      "clip_ratio/high_mean": 0.0011563595734287446,
      "clip_ratio/low_mean": 0.0009027509175894011,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002059110497157235,
      "epoch": 0.3619993646415261,
      "grad_norm": 0.11735457926988602,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 730
    },
    {
      "clip_ratio/high_max": 0.006603261741474853,
      "clip_ratio/high_mean": 0.0014815102637157906,
      "clip_ratio/low_mean": 0.001347767783386189,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028292780134506756,
      "epoch": 0.3624952541821309,
      "grad_norm": 0.1024734303355217,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 731
    },
    {
      "clip_ratio/high_max": 0.007727427560894284,
      "clip_ratio/high_mean": 0.0017435502061289299,
      "clip_ratio/low_mean": 0.0017655043685635974,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035090546080027707,
      "epoch": 0.3629911437227358,
      "grad_norm": 1.2042393684387207,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 732
    },
    {
      "clip_ratio/high_max": 0.00915769841412839,
      "clip_ratio/high_mean": 0.0020627574419904704,
      "clip_ratio/low_mean": 0.0021574123047685134,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004220169750624336,
      "epoch": 0.3634870332633406,
      "grad_norm": 0.07030929625034332,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 733
    },
    {
      "clip_ratio/high_max": 0.010513519377127523,
      "clip_ratio/high_mean": 0.00236704171038582,
      "clip_ratio/low_mean": 0.0024957704238204315,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0048628120857756585,
      "epoch": 0.3639829228039454,
      "grad_norm": 0.06355348229408264,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 734
    },
    {
      "clip_ratio/high_max": 0.011395861492928816,
      "clip_ratio/high_mean": 0.002585216267107171,
      "clip_ratio/low_mean": 0.002843554588935149,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0054287708226183895,
      "epoch": 0.36447881234455026,
      "grad_norm": 0.057497426867485046,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 735
    },
    {
      "clip_ratio/high_max": 0.012405151366692735,
      "clip_ratio/high_mean": 0.0027835200930894644,
      "clip_ratio/low_mean": 0.0031284330891594436,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005911953214308596,
      "epoch": 0.3649747018851551,
      "grad_norm": 0.05146970599889755,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 736
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1272321428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15065.0,
      "completions/mean_length": 4624.4443359375,
      "completions/mean_terminated_length": 3059.414306640625,
      "completions/min_length": 581.0,
      "completions/min_terminated_length": 581.0,
      "epoch": 0.36547059142575994,
      "grad_norm": 0.17509746551513672,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 142617825.0,
      "reward": 0.4732142984867096,
      "reward_std": 0.1755598485469818,
      "rewards/accuracy_reward/mean": 0.4732142984867096,
      "rewards/accuracy_reward/std": 0.4998401701450348,
      "step": 737
    },
    {
      "clip_ratio/high_max": 0.003881735667164321,
      "clip_ratio/high_mean": 0.0008420111819305021,
      "clip_ratio/low_mean": 0.0007080350880244168,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015500462500597223,
      "epoch": 0.36596648096636475,
      "grad_norm": 0.09267657995223999,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 738
    },
    {
      "clip_ratio/high_max": 0.00489343130357156,
      "clip_ratio/high_mean": 0.0010633170755909305,
      "clip_ratio/low_mean": 0.000944006388976959,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002007323467637434,
      "epoch": 0.36646237050696956,
      "grad_norm": 0.0758083164691925,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 739
    },
    {
      "clip_ratio/high_max": 0.006175110742333345,
      "clip_ratio/high_mean": 0.0013266882938296476,
      "clip_ratio/low_mean": 0.001227234997600135,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025539233155313923,
      "epoch": 0.3669582600475744,
      "grad_norm": 0.06819026172161102,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 740
    },
    {
      "clip_ratio/high_max": 0.007203283399576321,
      "clip_ratio/high_mean": 0.0015562441526526527,
      "clip_ratio/low_mean": 0.001514632809971772,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030708769180591844,
      "epoch": 0.36745414958817924,
      "grad_norm": 0.05707360431551933,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 741
    },
    {
      "clip_ratio/high_max": 0.008184492497093743,
      "clip_ratio/high_mean": 0.0017737706025400257,
      "clip_ratio/low_mean": 0.0017912580319716653,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035650286399686593,
      "epoch": 0.36795003912878405,
      "grad_norm": 0.04604189097881317,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 742
    },
    {
      "clip_ratio/high_max": 0.00884456143467105,
      "clip_ratio/high_mean": 0.0019499604472912324,
      "clip_ratio/low_mean": 0.0020329535318523995,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003982914011885441,
      "epoch": 0.3684459286693889,
      "grad_norm": 0.04337774217128754,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 743
    },
    {
      "clip_ratio/high_max": 0.009354653953778325,
      "clip_ratio/high_mean": 0.002065843910258991,
      "clip_ratio/low_mean": 0.0022533986389134952,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00431924255190097,
      "epoch": 0.3689418182099937,
      "grad_norm": 0.04147830978035927,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 744
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0758928571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15131.0,
      "completions/mean_length": 3783.107421875,
      "completions/mean_terminated_length": 2832.347900390625,
      "completions/min_length": 397.0,
      "completions/min_terminated_length": 397.0,
      "epoch": 0.36943770775059853,
      "grad_norm": 0.21698197722434998,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 144380225.0,
      "reward": 0.5758928656578064,
      "reward_std": 0.2440071403980255,
      "rewards/accuracy_reward/mean": 0.5758928656578064,
      "rewards/accuracy_reward/std": 0.4947591722011566,
      "step": 745
    },
    {
      "clip_ratio/high_max": 0.0051123158118571155,
      "clip_ratio/high_mean": 0.0011199256996405893,
      "clip_ratio/low_mean": 0.0008487796847305162,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001968705368199153,
      "epoch": 0.3699335972912034,
      "grad_norm": 0.13389962911605835,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 746
    },
    {
      "clip_ratio/high_max": 0.006733585618349025,
      "clip_ratio/high_mean": 0.0014698513618895959,
      "clip_ratio/low_mean": 0.0012925573198856455,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002762408666058036,
      "epoch": 0.3704294868318082,
      "grad_norm": 0.10370028764009476,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 747
    },
    {
      "clip_ratio/high_max": 0.00808738555861055,
      "clip_ratio/high_mean": 0.0017974667248381593,
      "clip_ratio/low_mean": 0.0017781918925550144,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003575658577574359,
      "epoch": 0.370925376372413,
      "grad_norm": 0.08613869547843933,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 748
    },
    {
      "clip_ratio/high_max": 0.00966828596938285,
      "clip_ratio/high_mean": 0.002155485084585962,
      "clip_ratio/low_mean": 0.0021663054745317822,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004321790524045355,
      "epoch": 0.3714212659130179,
      "grad_norm": 0.07188727706670761,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 749
    },
    {
      "clip_ratio/high_max": 0.01069090174860321,
      "clip_ratio/high_mean": 0.0023723129738755233,
      "clip_ratio/low_mean": 0.0025385538036175603,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004910866806312697,
      "epoch": 0.3719171554536227,
      "grad_norm": 0.06314442306756973,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 750
    },
    {
      "clip_ratio/high_max": 0.011408634825784247,
      "clip_ratio/high_mean": 0.002558401084115758,
      "clip_ratio/low_mean": 0.002900100001681949,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005458501083012379,
      "epoch": 0.37241304499422756,
      "grad_norm": 0.05636752396821976,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 751
    },
    {
      "clip_ratio/high_max": 0.012058366872224724,
      "clip_ratio/high_mean": 0.002734327488269628,
      "clip_ratio/low_mean": 0.0031338840404941948,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0058682114804469165,
      "epoch": 0.37290893453483237,
      "grad_norm": 0.05504003167152405,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 752
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0513392857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15216.0,
      "completions/mean_length": 3631.04248046875,
      "completions/mean_terminated_length": 2996.298583984375,
      "completions/min_length": 318.0,
      "completions/min_terminated_length": 318.0,
      "epoch": 0.3734048240754372,
      "grad_norm": 0.18282631039619446,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 146074348.0,
      "reward": 0.5401785969734192,
      "reward_std": 0.22838754951953888,
      "rewards/accuracy_reward/mean": 0.5401785969734192,
      "rewards/accuracy_reward/std": 0.49894022941589355,
      "step": 753
    },
    {
      "clip_ratio/high_max": 0.003958556084398879,
      "clip_ratio/high_mean": 0.0009063436496035138,
      "clip_ratio/low_mean": 0.0009436215464120323,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018499651973797882,
      "epoch": 0.37390071361604205,
      "grad_norm": 0.1281830221414566,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 754
    },
    {
      "clip_ratio/high_max": 0.004978527833372937,
      "clip_ratio/high_mean": 0.001123980573083827,
      "clip_ratio/low_mean": 0.001355288925651621,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002479269490322622,
      "epoch": 0.37439660315664686,
      "grad_norm": 0.08778057992458344,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 755
    },
    {
      "clip_ratio/high_max": 0.005916157642786857,
      "clip_ratio/high_mean": 0.00132075011060806,
      "clip_ratio/low_mean": 0.001853975237963823,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031747253342473414,
      "epoch": 0.37489249269725167,
      "grad_norm": 0.07385500520467758,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 756
    },
    {
      "clip_ratio/high_max": 0.006917779624927789,
      "clip_ratio/high_mean": 0.0015545581127298647,
      "clip_ratio/low_mean": 0.002228531602440853,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037830897181265755,
      "epoch": 0.37538838223785653,
      "grad_norm": 0.05921715870499611,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 757
    },
    {
      "clip_ratio/high_max": 0.007724054590653395,
      "clip_ratio/high_mean": 0.001741541027513449,
      "clip_ratio/low_mean": 0.002608080403661006,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0043496214548213175,
      "epoch": 0.37588427177846134,
      "grad_norm": 0.05509370565414429,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 758
    },
    {
      "clip_ratio/high_max": 0.00831384387856815,
      "clip_ratio/high_mean": 0.001880316576716723,
      "clip_ratio/low_mean": 0.0029665939423466625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0048469105204276275,
      "epoch": 0.37638016131906615,
      "grad_norm": 0.05065090209245682,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 759
    },
    {
      "clip_ratio/high_max": 0.008878013017238118,
      "clip_ratio/high_mean": 0.0020276232094147417,
      "clip_ratio/low_mean": 0.0033116694589807594,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005339292698408826,
      "epoch": 0.376876050859671,
      "grad_norm": 0.0456579253077507,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 760
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0915178571428571,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15024.0,
      "completions/mean_length": 4560.38427734375,
      "completions/mean_terminated_length": 3472.4619140625,
      "completions/min_length": 468.0,
      "completions/min_terminated_length": 468.0,
      "epoch": 0.37737194040027583,
      "grad_norm": 0.16578903794288635,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 148192056.0,
      "reward": 0.504464328289032,
      "reward_std": 0.1820894330739975,
      "rewards/accuracy_reward/mean": 0.5044642686843872,
      "rewards/accuracy_reward/std": 0.5005390048027039,
      "step": 761
    },
    {
      "clip_ratio/high_max": 0.004302716147776664,
      "clip_ratio/high_mean": 0.0008141291724541588,
      "clip_ratio/low_mean": 0.0005720789283145677,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0013862080999729187,
      "epoch": 0.3778678299408807,
      "grad_norm": 0.14039528369903564,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 762
    },
    {
      "clip_ratio/high_max": 0.005761069243362726,
      "clip_ratio/high_mean": 0.0010999400933542347,
      "clip_ratio/low_mean": 0.000856861219062921,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001956801316282508,
      "epoch": 0.3783637194814855,
      "grad_norm": 0.08250514417886734,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 763
    },
    {
      "clip_ratio/high_max": 0.007105201127615146,
      "clip_ratio/high_mean": 0.0013333722849893093,
      "clip_ratio/low_mean": 0.001099359016109247,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024327313049070654,
      "epoch": 0.3788596090220903,
      "grad_norm": 0.0633288025856018,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 764
    },
    {
      "clip_ratio/high_max": 0.008230254616137245,
      "clip_ratio/high_mean": 0.0015412922939503915,
      "clip_ratio/low_mean": 0.0013635825553706127,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029048748501736554,
      "epoch": 0.3793554985626952,
      "grad_norm": 0.05535858869552612,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 765
    },
    {
      "clip_ratio/high_max": 0.009624274001907906,
      "clip_ratio/high_mean": 0.0017784395431590383,
      "clip_ratio/low_mean": 0.0016259778329867913,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034044173908114317,
      "epoch": 0.3798513881033,
      "grad_norm": 0.048737697303295135,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 766
    },
    {
      "clip_ratio/high_max": 0.010487879814718326,
      "clip_ratio/high_mean": 0.0019509984194883145,
      "clip_ratio/low_mean": 0.0018020023784401928,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037530007803070475,
      "epoch": 0.3803472776439048,
      "grad_norm": 0.045172613114118576,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 767
    },
    {
      "clip_ratio/high_max": 0.011173837487149285,
      "clip_ratio/high_mean": 0.0020874804183677043,
      "clip_ratio/low_mean": 0.001984358137065101,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004071838557138108,
      "epoch": 0.38084316718450967,
      "grad_norm": 0.03987935557961464,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "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": 0.1049107142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15271.0,
      "completions/mean_length": 4589.998046875,
      "completions/mean_terminated_length": 3327.678466796875,
      "completions/min_length": 490.0,
      "completions/min_terminated_length": 490.0,
      "epoch": 0.3813390567251145,
      "grad_norm": 0.1376647800207138,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 150322455.0,
      "reward": 0.4754464626312256,
      "reward_std": 0.1656225174665451,
      "rewards/accuracy_reward/mean": 0.4754464328289032,
      "rewards/accuracy_reward/std": 0.4999549984931946,
      "step": 769
    },
    {
      "clip_ratio/high_max": 0.0037574724556179717,
      "clip_ratio/high_mean": 0.0007190204991047722,
      "clip_ratio/low_mean": 0.0005412974635987666,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0012603179498000827,
      "epoch": 0.3818349462657193,
      "grad_norm": 0.14862807095050812,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 770
    },
    {
      "clip_ratio/high_max": 0.0048632820980856195,
      "clip_ratio/high_mean": 0.0009414562573510921,
      "clip_ratio/low_mean": 0.0007428376446796392,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016842938857735135,
      "epoch": 0.38233083580632415,
      "grad_norm": 0.08464793115854263,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 771
    },
    {
      "clip_ratio/high_max": 0.00618448977911612,
      "clip_ratio/high_mean": 0.0011852780412482389,
      "clip_ratio/low_mean": 0.0010327842393280662,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022180622681844397,
      "epoch": 0.38282672534692896,
      "grad_norm": 0.07845152914524078,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 772
    },
    {
      "clip_ratio/high_max": 0.007602500652865274,
      "clip_ratio/high_mean": 0.0014381655400939053,
      "clip_ratio/low_mean": 0.00125723899031982,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026954045633829082,
      "epoch": 0.3833226148875338,
      "grad_norm": 0.08710558712482452,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 773
    },
    {
      "clip_ratio/high_max": 0.008667309759402997,
      "clip_ratio/high_mean": 0.0016340390666300664,
      "clip_ratio/low_mean": 0.001507260645439601,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031412997195729986,
      "epoch": 0.38381850442813864,
      "grad_norm": 0.04805275425314903,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 774
    },
    {
      "clip_ratio/high_max": 0.00924473946361104,
      "clip_ratio/high_mean": 0.0017656431846262421,
      "clip_ratio/low_mean": 0.001719630505931491,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034852736880566226,
      "epoch": 0.38431439396874345,
      "grad_norm": 0.04765886440873146,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 775
    },
    {
      "clip_ratio/high_max": 0.009933743214787683,
      "clip_ratio/high_mean": 0.0018863321506614739,
      "clip_ratio/low_mean": 0.0019245982705342612,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00381093040232372,
      "epoch": 0.3848102835093483,
      "grad_norm": 0.04532298073172569,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 776
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0982142857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14802.0,
      "completions/mean_length": 4100.4130859375,
      "completions/mean_terminated_length": 2874.121337890625,
      "completions/min_length": 440.0,
      "completions/min_terminated_length": 440.0,
      "epoch": 0.3853061730499531,
      "grad_norm": 0.18981106579303741,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 152227944.0,
      "reward": 0.598214328289032,
      "reward_std": 0.22979693114757538,
      "rewards/accuracy_reward/mean": 0.5982142686843872,
      "rewards/accuracy_reward/std": 0.49080711603164673,
      "step": 777
    },
    {
      "clip_ratio/high_max": 0.004942331532220123,
      "clip_ratio/high_mean": 0.0011510113713484316,
      "clip_ratio/low_mean": 0.0007085724992066389,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001859583851910429,
      "epoch": 0.38580206259055794,
      "grad_norm": 0.14411267638206482,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 778
    },
    {
      "clip_ratio/high_max": 0.0063993691892392235,
      "clip_ratio/high_mean": 0.0014450848877913813,
      "clip_ratio/low_mean": 0.0010442086878015289,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024892935653042514,
      "epoch": 0.3862979521311628,
      "grad_norm": 0.20455290377140045,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 779
    },
    {
      "clip_ratio/high_max": 0.007894121134086163,
      "clip_ratio/high_mean": 0.0017919136982982309,
      "clip_ratio/low_mean": 0.0013236304869224114,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031155441774899373,
      "epoch": 0.3867938416717676,
      "grad_norm": 0.07698752731084824,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 780
    },
    {
      "clip_ratio/high_max": 0.008792230781182298,
      "clip_ratio/high_mean": 0.0020193686852962855,
      "clip_ratio/low_mean": 0.0016230247367730044,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036423934070626274,
      "epoch": 0.3872897312123724,
      "grad_norm": 0.06770796328783035,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 781
    },
    {
      "clip_ratio/high_max": 0.010082515866088215,
      "clip_ratio/high_mean": 0.0022638745017502515,
      "clip_ratio/low_mean": 0.0018895457806138438,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00415342029191379,
      "epoch": 0.3877856207529773,
      "grad_norm": 0.0637950524687767,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 782
    },
    {
      "clip_ratio/high_max": 0.010908333028055495,
      "clip_ratio/high_mean": 0.002444524298880424,
      "clip_ratio/low_mean": 0.002103913367818677,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0045484376969398,
      "epoch": 0.3882815102935821,
      "grad_norm": 0.05213182047009468,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 783
    },
    {
      "clip_ratio/high_max": 0.011470392688352149,
      "clip_ratio/high_mean": 0.0025715547867548594,
      "clip_ratio/low_mean": 0.0022774994017709105,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00484905418852577,
      "epoch": 0.3887773998341869,
      "grad_norm": 0.051174651831388474,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 784
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.044642857142857095,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15244.0,
      "completions/mean_length": 3523.263671875,
      "completions/mean_terminated_length": 2970.144775390625,
      "completions/min_length": 67.0,
      "completions/min_terminated_length": 67.0,
      "epoch": 0.3892732893747918,
      "grad_norm": 0.23878303170204163,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 153871790.0,
      "reward": 0.546875,
      "reward_std": 0.25272703170776367,
      "rewards/accuracy_reward/mean": 0.546875,
      "rewards/accuracy_reward/std": 0.49835437536239624,
      "step": 785
    },
    {
      "clip_ratio/high_max": 0.005508710439244169,
      "clip_ratio/high_mean": 0.001361496690833519,
      "clip_ratio/low_mean": 0.0013116315433308046,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026731282368928078,
      "epoch": 0.3897691789153966,
      "grad_norm": 0.1314787119626999,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 786
    },
    {
      "clip_ratio/high_max": 0.007152205187594518,
      "clip_ratio/high_mean": 0.0017869562768737524,
      "clip_ratio/low_mean": 0.0018660959128737886,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003653052206573193,
      "epoch": 0.39026506845600145,
      "grad_norm": 0.10566339641809464,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 787
    },
    {
      "clip_ratio/high_max": 0.008377977483178256,
      "clip_ratio/high_mean": 0.0021046964604920504,
      "clip_ratio/low_mean": 0.0023423149891641515,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004447011398951872,
      "epoch": 0.39076095799660626,
      "grad_norm": 0.08378065377473831,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 788
    },
    {
      "clip_ratio/high_max": 0.009718321125546936,
      "clip_ratio/high_mean": 0.0024314700281138357,
      "clip_ratio/low_mean": 0.002779811531127052,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005211281592892192,
      "epoch": 0.39125684753721107,
      "grad_norm": 0.07672514021396637,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 789
    },
    {
      "clip_ratio/high_max": 0.01072242425652803,
      "clip_ratio/high_mean": 0.0027133506323480105,
      "clip_ratio/low_mean": 0.003238915203837678,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005952265790256206,
      "epoch": 0.39175273707781594,
      "grad_norm": 0.07393396645784378,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 790
    },
    {
      "clip_ratio/high_max": 0.011657618297249428,
      "clip_ratio/high_mean": 0.0029297955256879504,
      "clip_ratio/low_mean": 0.003651243512649671,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0065810390278784325,
      "epoch": 0.39224862661842075,
      "grad_norm": 0.07168807834386826,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 791
    },
    {
      "clip_ratio/high_max": 0.012554772007206338,
      "clip_ratio/high_mean": 0.0031802229193544918,
      "clip_ratio/low_mean": 0.004009354544450616,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007189577458120766,
      "epoch": 0.39274451615902556,
      "grad_norm": 0.056267473846673965,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 792
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15175.0,
      "completions/mean_length": 4379.89306640625,
      "completions/mean_terminated_length": 3420.466064453125,
      "completions/min_length": 487.0,
      "completions/min_terminated_length": 487.0,
      "epoch": 0.3932404056996304,
      "grad_norm": 0.16454346477985382,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 155905942.0,
      "reward": 0.4754464626312256,
      "reward_std": 0.17540507018566132,
      "rewards/accuracy_reward/mean": 0.4754464328289032,
      "rewards/accuracy_reward/std": 0.4999549984931946,
      "step": 793
    },
    {
      "clip_ratio/high_max": 0.004200187609967543,
      "clip_ratio/high_mean": 0.0007818047311047849,
      "clip_ratio/low_mean": 0.0006368023765048747,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0014186071016411006,
      "epoch": 0.39373629524023523,
      "grad_norm": 0.11581408977508545,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 794
    },
    {
      "clip_ratio/high_max": 0.005700293837435311,
      "clip_ratio/high_mean": 0.0010469455933161953,
      "clip_ratio/low_mean": 0.0009498104525960116,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019967560438090004,
      "epoch": 0.39423218478084004,
      "grad_norm": 0.07670275866985321,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 795
    },
    {
      "clip_ratio/high_max": 0.006668094769338495,
      "clip_ratio/high_mean": 0.0012168045434464148,
      "clip_ratio/low_mean": 0.0012261132654316498,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002442917801090516,
      "epoch": 0.3947280743214449,
      "grad_norm": 0.06462813913822174,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 796
    },
    {
      "clip_ratio/high_max": 0.0077836311247665435,
      "clip_ratio/high_mean": 0.0014173613553793984,
      "clip_ratio/low_mean": 0.0015388513540983695,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00295621273744473,
      "epoch": 0.3952239638620497,
      "grad_norm": 0.049734581261873245,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 797
    },
    {
      "clip_ratio/high_max": 0.008759313601331087,
      "clip_ratio/high_mean": 0.0015967399776855018,
      "clip_ratio/low_mean": 0.0017848460581717518,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003381586060640984,
      "epoch": 0.39571985340265453,
      "grad_norm": 0.04648972302675247,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 798
    },
    {
      "clip_ratio/high_max": 0.009270462724089157,
      "clip_ratio/high_mean": 0.0017004258315864718,
      "clip_ratio/low_mean": 0.00208305850776469,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037834843615200953,
      "epoch": 0.3962157429432594,
      "grad_norm": 0.040940798819065094,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 799
    },
    {
      "clip_ratio/high_max": 0.009811503205128247,
      "clip_ratio/high_mean": 0.0018084397888742387,
      "clip_ratio/low_mean": 0.002284551327193185,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004092991101060761,
      "epoch": 0.3967116324838642,
      "grad_norm": 0.04130401834845543,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 800
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13974.0,
      "completions/mean_length": 3733.95556640625,
      "completions/mean_terminated_length": 2748.697509765625,
      "completions/min_length": 436.0,
      "completions/min_terminated_length": 436.0,
      "epoch": 0.3972075220244691,
      "grad_norm": 0.15385322272777557,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 157651378.0,
      "reward": 0.613839328289032,
      "reward_std": 0.15421023964881897,
      "rewards/accuracy_reward/mean": 0.6138392686843872,
      "rewards/accuracy_reward/std": 0.4874124526977539,
      "step": 801
    },
    {
      "clip_ratio/high_max": 0.0036967343685319065,
      "clip_ratio/high_mean": 0.0007856237848500314,
      "clip_ratio/low_mean": 0.0005491959103665067,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001334819688963762,
      "epoch": 0.3977034115650739,
      "grad_norm": 0.10151074826717377,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 802
    },
    {
      "clip_ratio/high_max": 0.004847840189540875,
      "clip_ratio/high_mean": 0.0010148976900836715,
      "clip_ratio/low_mean": 0.0009026105603879842,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001917508257974987,
      "epoch": 0.3981993011056787,
      "grad_norm": 0.07983680814504623,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 803
    },
    {
      "clip_ratio/high_max": 0.0058168402829323895,
      "clip_ratio/high_mean": 0.00122696539710887,
      "clip_ratio/low_mean": 0.0011549387529612432,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002381904134836077,
      "epoch": 0.39869519064628356,
      "grad_norm": 0.0702308788895607,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 804
    },
    {
      "clip_ratio/high_max": 0.006595573102458729,
      "clip_ratio/high_mean": 0.0013913741979649785,
      "clip_ratio/low_mean": 0.001405118146294626,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002796492338347889,
      "epoch": 0.39919108018688837,
      "grad_norm": 0.05231950059533119,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 805
    },
    {
      "clip_ratio/high_max": 0.0073634641121316236,
      "clip_ratio/high_mean": 0.001547384783407324,
      "clip_ratio/low_mean": 0.0016585922123795172,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003205977012839867,
      "epoch": 0.3996869697274932,
      "grad_norm": 0.04529507830739021,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 806
    },
    {
      "clip_ratio/high_max": 0.008019508335564751,
      "clip_ratio/high_mean": 0.0016865106449586165,
      "clip_ratio/low_mean": 0.0018648320485681324,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035513426946636173,
      "epoch": 0.40018285926809805,
      "grad_norm": 0.0405164435505867,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 807
    },
    {
      "clip_ratio/high_max": 0.008556687189411605,
      "clip_ratio/high_mean": 0.001838619287354959,
      "clip_ratio/low_mean": 0.001991953137121527,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0038305724065139657,
      "epoch": 0.40067874880870286,
      "grad_norm": 0.042359888553619385,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 808
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0848214285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15336.0,
      "completions/mean_length": 4506.8955078125,
      "completions/mean_terminated_length": 3500.99755859375,
      "completions/min_length": 670.0,
      "completions/min_terminated_length": 670.0,
      "epoch": 0.40117463834930767,
      "grad_norm": 0.1649194359779358,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 159739499.0,
      "reward": 0.4799107313156128,
      "reward_std": 0.21703803539276123,
      "rewards/accuracy_reward/mean": 0.4799107015132904,
      "rewards/accuracy_reward/std": 0.5001547932624817,
      "step": 809
    },
    {
      "clip_ratio/high_max": 0.004432332112628501,
      "clip_ratio/high_mean": 0.0009638664278099895,
      "clip_ratio/low_mean": 0.0007449681968410005,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017088346189666481,
      "epoch": 0.40167052788991253,
      "grad_norm": 0.10734858363866806,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 810
    },
    {
      "clip_ratio/high_max": 0.00614311762728903,
      "clip_ratio/high_mean": 0.0012692774264451145,
      "clip_ratio/low_mean": 0.0010090055225191463,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022782829400966875,
      "epoch": 0.40216641743051734,
      "grad_norm": 0.0900370329618454,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 811
    },
    {
      "clip_ratio/high_max": 0.007147946005716221,
      "clip_ratio/high_mean": 0.00151513429455008,
      "clip_ratio/low_mean": 0.00133625161515738,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028513859451777535,
      "epoch": 0.4026623069711222,
      "grad_norm": 0.06731628626585007,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 812
    },
    {
      "clip_ratio/high_max": 0.008369758234039182,
      "clip_ratio/high_mean": 0.0017642542343310197,
      "clip_ratio/low_mean": 0.0016527357038285118,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003416989902689238,
      "epoch": 0.403158196511727,
      "grad_norm": 0.06017053499817848,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 813
    },
    {
      "clip_ratio/high_max": 0.009616394720069366,
      "clip_ratio/high_mean": 0.0020208561741128506,
      "clip_ratio/low_mean": 0.0019589388634813076,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003979795071245462,
      "epoch": 0.40365408605233183,
      "grad_norm": 0.05148792266845703,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 814
    },
    {
      "clip_ratio/high_max": 0.010670185558410594,
      "clip_ratio/high_mean": 0.002240627565242903,
      "clip_ratio/low_mean": 0.0022416968477045884,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004482324458876974,
      "epoch": 0.4041499755929367,
      "grad_norm": 0.04554063081741333,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 815
    },
    {
      "clip_ratio/high_max": 0.011550949326192494,
      "clip_ratio/high_mean": 0.0024109758851409424,
      "clip_ratio/low_mean": 0.0024723311357774946,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004883307044110552,
      "epoch": 0.4046458651335415,
      "grad_norm": 0.04057490825653076,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 816
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0803571428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15068.0,
      "completions/mean_length": 3950.982421875,
      "completions/mean_terminated_length": 2954.07763671875,
      "completions/min_length": 466.0,
      "completions/min_terminated_length": 466.0,
      "epoch": 0.4051417546741463,
      "grad_norm": 0.20390227437019348,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 161578339.0,
      "reward": 0.4933035969734192,
      "reward_std": 0.2189052850008011,
      "rewards/accuracy_reward/mean": 0.4933035671710968,
      "rewards/accuracy_reward/std": 0.5005140900611877,
      "step": 817
    },
    {
      "clip_ratio/high_max": 0.003898435250903276,
      "clip_ratio/high_mean": 0.0007994448356214434,
      "clip_ratio/low_mean": 0.0008928397599561322,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016922846102715994,
      "epoch": 0.4056376442147512,
      "grad_norm": 0.12252761423587799,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 818
    },
    {
      "clip_ratio/high_max": 0.0052922533136552374,
      "clip_ratio/high_mean": 0.0011036981439076499,
      "clip_ratio/low_mean": 0.0012702088759510843,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023739070209103375,
      "epoch": 0.406133533755356,
      "grad_norm": 0.28560563921928406,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 819
    },
    {
      "clip_ratio/high_max": 0.006460019017595187,
      "clip_ratio/high_mean": 0.0013471196241710004,
      "clip_ratio/low_mean": 0.0017101458864772212,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003057265502093287,
      "epoch": 0.4066294232959608,
      "grad_norm": 0.07237254083156586,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 820
    },
    {
      "clip_ratio/high_max": 0.007290399327757768,
      "clip_ratio/high_mean": 0.0015135667812273823,
      "clip_ratio/low_mean": 0.0020653432785593395,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003578910042278949,
      "epoch": 0.40712531283656567,
      "grad_norm": 0.06444113701581955,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 821
    },
    {
      "clip_ratio/high_max": 0.008001560599041113,
      "clip_ratio/high_mean": 0.00164790846523033,
      "clip_ratio/low_mean": 0.0023536817039939706,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004001590153848156,
      "epoch": 0.4076212023771705,
      "grad_norm": 0.05788607522845268,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 822
    },
    {
      "clip_ratio/high_max": 0.008638409827653959,
      "clip_ratio/high_mean": 0.0017905210330013688,
      "clip_ratio/low_mean": 0.0026916271696109106,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004482148217675785,
      "epoch": 0.40811709191777534,
      "grad_norm": 0.05619983747601509,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 823
    },
    {
      "clip_ratio/high_max": 0.008999527371088334,
      "clip_ratio/high_mean": 0.001870336324031996,
      "clip_ratio/low_mean": 0.0029791756493295907,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004849511986321886,
      "epoch": 0.40861298145838015,
      "grad_norm": 0.04471541941165924,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 824
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0558035714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15272.0,
      "completions/mean_length": 3856.578369140625,
      "completions/mean_terminated_length": 3176.706787109375,
      "completions/min_length": 433.0,
      "completions/min_terminated_length": 433.0,
      "epoch": 0.40910887099898496,
      "grad_norm": 0.1509619802236557,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 163385022.0,
      "reward": 0.5066964626312256,
      "reward_std": 0.14729197323322296,
      "rewards/accuracy_reward/mean": 0.5066964030265808,
      "rewards/accuracy_reward/std": 0.5005140900611877,
      "step": 825
    },
    {
      "clip_ratio/high_max": 0.002661901860847138,
      "clip_ratio/high_mean": 0.0005665436473236696,
      "clip_ratio/low_mean": 0.0006541446965684372,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0012206883411636227,
      "epoch": 0.40960476053958983,
      "grad_norm": 0.08393595367670059,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 826
    },
    {
      "clip_ratio/high_max": 0.0034747078534564935,
      "clip_ratio/high_mean": 0.0007285456545105262,
      "clip_ratio/low_mean": 0.0008404471620906406,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015689928195570246,
      "epoch": 0.41010065008019464,
      "grad_norm": 0.0712735652923584,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 827
    },
    {
      "clip_ratio/high_max": 0.0042114661910090945,
      "clip_ratio/high_mean": 0.0008972895254828472,
      "clip_ratio/low_mean": 0.0010370498819156637,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001934339407398511,
      "epoch": 0.41059653962079945,
      "grad_norm": 0.06373745203018188,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 828
    },
    {
      "clip_ratio/high_max": 0.0048986442325258395,
      "clip_ratio/high_mean": 0.001021727817487772,
      "clip_ratio/low_mean": 0.0012789023001005262,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023006301271379925,
      "epoch": 0.4110924291614043,
      "grad_norm": 0.05151547119021416,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 829
    },
    {
      "clip_ratio/high_max": 0.005287966348987538,
      "clip_ratio/high_mean": 0.001130954245581961,
      "clip_ratio/low_mean": 0.0014963608630296221,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026273151070199674,
      "epoch": 0.4115883187020091,
      "grad_norm": 0.044478558003902435,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 830
    },
    {
      "clip_ratio/high_max": 0.005698319037037436,
      "clip_ratio/high_mean": 0.0012020703206871985,
      "clip_ratio/low_mean": 0.0016987960975711758,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029008664278080687,
      "epoch": 0.41208420824261394,
      "grad_norm": 0.0490901917219162,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 831
    },
    {
      "clip_ratio/high_max": 0.005982632857921999,
      "clip_ratio/high_mean": 0.00127781717696962,
      "clip_ratio/low_mean": 0.0018763889929687139,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031542061806248967,
      "epoch": 0.4125800977832188,
      "grad_norm": 0.037380803376436234,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "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": 0.0982142857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14946.0,
      "completions/mean_length": 4400.3798828125,
      "completions/mean_terminated_length": 3206.75732421875,
      "completions/min_length": 422.0,
      "completions/min_terminated_length": 422.0,
      "epoch": 0.4130759873238236,
      "grad_norm": 0.21273459494113922,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 165428112.0,
      "reward": 0.5401785969734192,
      "reward_std": 0.2195100486278534,
      "rewards/accuracy_reward/mean": 0.5401785969734192,
      "rewards/accuracy_reward/std": 0.49894022941589355,
      "step": 833
    },
    {
      "clip_ratio/high_max": 0.004404968005474075,
      "clip_ratio/high_mean": 0.0009307112500209769,
      "clip_ratio/low_mean": 0.0009587565894548788,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018894678464675962,
      "epoch": 0.4135718768644284,
      "grad_norm": 0.12677262723445892,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 834
    },
    {
      "clip_ratio/high_max": 0.006143011078165728,
      "clip_ratio/high_mean": 0.0012531692696029495,
      "clip_ratio/low_mean": 0.0012853301812469908,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002538499445790876,
      "epoch": 0.4140677664050333,
      "grad_norm": 0.09401555359363556,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 835
    },
    {
      "clip_ratio/high_max": 0.007499228733649943,
      "clip_ratio/high_mean": 0.0015323940760936239,
      "clip_ratio/low_mean": 0.0017563950391945582,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0032887890865822555,
      "epoch": 0.4145636559456381,
      "grad_norm": 0.07729651033878326,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 836
    },
    {
      "clip_ratio/high_max": 0.008748344214836834,
      "clip_ratio/high_mean": 0.0017940437373908935,
      "clip_ratio/low_mean": 0.002094458077465333,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003888501818437362,
      "epoch": 0.41505954548624296,
      "grad_norm": 0.06593268364667892,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 837
    },
    {
      "clip_ratio/high_max": 0.00959191429137718,
      "clip_ratio/high_mean": 0.001981086073556071,
      "clip_ratio/low_mean": 0.0024553924218366774,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004436478484876716,
      "epoch": 0.4155554350268478,
      "grad_norm": 0.05693370848894119,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 838
    },
    {
      "clip_ratio/high_max": 0.010764876587927574,
      "clip_ratio/high_mean": 0.002181582480261568,
      "clip_ratio/low_mean": 0.0028295791079244736,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0050111615792047814,
      "epoch": 0.4160513245674526,
      "grad_norm": 0.04980418086051941,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 839
    },
    {
      "clip_ratio/high_max": 0.011335981253068894,
      "clip_ratio/high_mean": 0.0023098779647625634,
      "clip_ratio/low_mean": 0.003050530635960058,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0053604086988343624,
      "epoch": 0.41654721410805745,
      "grad_norm": 0.04576840251684189,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 840
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0669642857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 13961.0,
      "completions/mean_length": 3688.107421875,
      "completions/mean_terminated_length": 2850.411376953125,
      "completions/min_length": 460.0,
      "completions/min_terminated_length": 460.0,
      "epoch": 0.41704310364866226,
      "grad_norm": 0.22967024147510529,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 167152056.0,
      "reward": 0.5691964626312256,
      "reward_std": 0.2836872935295105,
      "rewards/accuracy_reward/mean": 0.5691964030265808,
      "rewards/accuracy_reward/std": 0.4957422614097595,
      "step": 841
    },
    {
      "clip_ratio/high_max": 0.005212460479015135,
      "clip_ratio/high_mean": 0.001353140593664648,
      "clip_ratio/low_mean": 0.0010302312473413622,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023833718541936832,
      "epoch": 0.41753899318926707,
      "grad_norm": 0.14577677845954895,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 842
    },
    {
      "clip_ratio/high_max": 0.006787248501495924,
      "clip_ratio/high_mean": 0.0017771379521036579,
      "clip_ratio/low_mean": 0.001537803496489687,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003314941442113195,
      "epoch": 0.41803488272987194,
      "grad_norm": 0.1087193712592125,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 843
    },
    {
      "clip_ratio/high_max": 0.008008223983779317,
      "clip_ratio/high_mean": 0.0021170908098611108,
      "clip_ratio/low_mean": 0.001982011945699469,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004099102837244573,
      "epoch": 0.41853077227047675,
      "grad_norm": 0.1311848759651184,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 844
    },
    {
      "clip_ratio/high_max": 0.009380891624459764,
      "clip_ratio/high_mean": 0.002452186231948872,
      "clip_ratio/low_mean": 0.0023321387902228707,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004784325005857681,
      "epoch": 0.41902666181108156,
      "grad_norm": 0.07558256387710571,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 845
    },
    {
      "clip_ratio/high_max": 0.010117848243680783,
      "clip_ratio/high_mean": 0.0026902477511612233,
      "clip_ratio/low_mean": 0.0027557192672702513,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005445967039122479,
      "epoch": 0.4195225513516864,
      "grad_norm": 0.06615594029426575,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 846
    },
    {
      "clip_ratio/high_max": 0.01059344207533286,
      "clip_ratio/high_mean": 0.0028440084042813396,
      "clip_ratio/low_mean": 0.0030778217430906807,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0059218301339569734,
      "epoch": 0.42001844089229123,
      "grad_norm": 0.06021525710821152,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 847
    },
    {
      "clip_ratio/high_max": 0.011401931384170894,
      "clip_ratio/high_mean": 0.00306475276920537,
      "clip_ratio/low_mean": 0.003341452070685591,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006406204852282826,
      "epoch": 0.4205143304328961,
      "grad_norm": 0.05921819806098938,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 848
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0982142857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15170.0,
      "completions/mean_length": 5043.779296875,
      "completions/mean_terminated_length": 3920.230224609375,
      "completions/min_length": 598.0,
      "completions/min_terminated_length": 598.0,
      "epoch": 0.4210102199735009,
      "grad_norm": 0.14894352853298187,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 169483749.0,
      "reward": 0.4665178656578064,
      "reward_std": 0.16336661577224731,
      "rewards/accuracy_reward/mean": 0.4665178656578064,
      "rewards/accuracy_reward/std": 0.4994353950023651,
      "step": 849
    },
    {
      "clip_ratio/high_max": 0.004671092497574136,
      "clip_ratio/high_mean": 0.0008467535088243494,
      "clip_ratio/low_mean": 0.00048688690526432765,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0013336404126675916,
      "epoch": 0.4215061095141057,
      "grad_norm": 0.10258540511131287,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 850
    },
    {
      "clip_ratio/high_max": 0.005818646863190224,
      "clip_ratio/high_mean": 0.001029825596901901,
      "clip_ratio/low_mean": 0.000733589352648778,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017634149514833553,
      "epoch": 0.4220019990547106,
      "grad_norm": 0.2581718862056732,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 851
    },
    {
      "clip_ratio/high_max": 0.00684992488095304,
      "clip_ratio/high_mean": 0.0012290920379882664,
      "clip_ratio/low_mean": 0.0009307069803980994,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002159799024639142,
      "epoch": 0.4224978885953154,
      "grad_norm": 0.067273810505867,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 852
    },
    {
      "clip_ratio/high_max": 0.007906045264462591,
      "clip_ratio/high_mean": 0.0014211724409278759,
      "clip_ratio/low_mean": 0.0011587451540435723,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025799175875818037,
      "epoch": 0.4229937781359202,
      "grad_norm": 0.051393914967775345,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 853
    },
    {
      "clip_ratio/high_max": 0.00873704714467749,
      "clip_ratio/high_mean": 0.0015741494567009795,
      "clip_ratio/low_mean": 0.0013502584693014796,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002924407928048822,
      "epoch": 0.4234896676765251,
      "grad_norm": 0.04500604420900345,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 854
    },
    {
      "clip_ratio/high_max": 0.009467062021940365,
      "clip_ratio/high_mean": 0.0017161920775379258,
      "clip_ratio/low_mean": 0.0015021007720861235,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00321829283916486,
      "epoch": 0.4239855572171299,
      "grad_norm": 0.03953379765152931,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 855
    },
    {
      "clip_ratio/high_max": 0.010269738844726817,
      "clip_ratio/high_mean": 0.0018523489029576012,
      "clip_ratio/low_mean": 0.0016459623325317807,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003498311253679276,
      "epoch": 0.4244814467577347,
      "grad_norm": 0.034719835966825485,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 856
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.049107142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15168.0,
      "completions/mean_length": 3770.7568359375,
      "completions/mean_terminated_length": 3172.251220703125,
      "completions/min_length": 527.0,
      "completions/min_terminated_length": 527.0,
      "epoch": 0.42497733629833956,
      "grad_norm": 0.17755459249019623,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 171244680.0,
      "reward": 0.5446428656578064,
      "reward_std": 0.19291327893733978,
      "rewards/accuracy_reward/mean": 0.5446428656578064,
      "rewards/accuracy_reward/std": 0.49855971336364746,
      "step": 857
    },
    {
      "clip_ratio/high_max": 0.004513471541031322,
      "clip_ratio/high_mean": 0.0009882671859031689,
      "clip_ratio/low_mean": 0.0008022959948448261,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017905631789290055,
      "epoch": 0.42547322583894437,
      "grad_norm": 1.3421016931533813,
      "learning_rate": 1e-06,
      "loss": 0.0004,
      "step": 858
    },
    {
      "clip_ratio/high_max": 0.005667302370966354,
      "clip_ratio/high_mean": 0.0012532697796814318,
      "clip_ratio/low_mean": 0.0010438031639523615,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002297072894180019,
      "epoch": 0.4259691153795492,
      "grad_norm": 0.08500746637582779,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 859
    },
    {
      "clip_ratio/high_max": 0.006632747353251034,
      "clip_ratio/high_mean": 0.001467432533218016,
      "clip_ratio/low_mean": 0.001365214777933943,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028326472847766127,
      "epoch": 0.42646500492015404,
      "grad_norm": 0.11359663307666779,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 860
    },
    {
      "clip_ratio/high_max": 0.0075080787792103365,
      "clip_ratio/high_mean": 0.0016965025138233614,
      "clip_ratio/low_mean": 0.0016699100265213929,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003366412541254249,
      "epoch": 0.42696089446075886,
      "grad_norm": 0.06130491942167282,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 861
    },
    {
      "clip_ratio/high_max": 0.008471315904898802,
      "clip_ratio/high_mean": 0.0019072211293860164,
      "clip_ratio/low_mean": 0.0019977205638497253,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039049417264322983,
      "epoch": 0.4274567840013637,
      "grad_norm": 0.052641384303569794,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 862
    },
    {
      "clip_ratio/high_max": 0.009059973368493957,
      "clip_ratio/high_mean": 0.0020228214059443417,
      "clip_ratio/low_mean": 0.002224932522267409,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004247753888193984,
      "epoch": 0.42795267354196853,
      "grad_norm": 0.05726415663957596,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 863
    },
    {
      "clip_ratio/high_max": 0.009807001395529369,
      "clip_ratio/high_mean": 0.0021851565893484803,
      "clip_ratio/low_mean": 0.002440617762886177,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004625774377927883,
      "epoch": 0.42844856308257334,
      "grad_norm": 0.047413744032382965,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 864
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0736607142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15315.0,
      "completions/mean_length": 4379.23681640625,
      "completions/mean_terminated_length": 3506.067626953125,
      "completions/min_length": 612.0,
      "completions/min_terminated_length": 612.0,
      "epoch": 0.4289444526231782,
      "grad_norm": 0.1972431242465973,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 173277106.0,
      "reward": 0.5424107313156128,
      "reward_std": 0.22139866650104523,
      "rewards/accuracy_reward/mean": 0.5424107313156128,
      "rewards/accuracy_reward/std": 0.49875500798225403,
      "step": 865
    },
    {
      "clip_ratio/high_max": 0.004492866939472151,
      "clip_ratio/high_mean": 0.0010453692445935303,
      "clip_ratio/low_mean": 0.0008878680746420287,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019332373244651535,
      "epoch": 0.429440342163783,
      "grad_norm": 0.11346317082643509,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 866
    },
    {
      "clip_ratio/high_max": 0.005918551993090659,
      "clip_ratio/high_mean": 0.0013525236342957214,
      "clip_ratio/low_mean": 0.0011942279481900187,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025467515556556464,
      "epoch": 0.42993623170438783,
      "grad_norm": 0.11527877300977707,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 867
    },
    {
      "clip_ratio/high_max": 0.007043459805572638,
      "clip_ratio/high_mean": 0.00159465502929379,
      "clip_ratio/low_mean": 0.001516526365548998,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003111181420081266,
      "epoch": 0.4304321212449927,
      "grad_norm": 0.0759681984782219,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 868
    },
    {
      "clip_ratio/high_max": 0.008146383584971773,
      "clip_ratio/high_mean": 0.0018108761760231573,
      "clip_ratio/low_mean": 0.001905580462562284,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037164566138017108,
      "epoch": 0.4309280107855975,
      "grad_norm": 0.06727921962738037,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 869
    },
    {
      "clip_ratio/high_max": 0.008713478953723097,
      "clip_ratio/high_mean": 0.0019564910471672192,
      "clip_ratio/low_mean": 0.0022087745244334656,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004165265633673698,
      "epoch": 0.4314239003262023,
      "grad_norm": 0.055219244211912155,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 870
    },
    {
      "clip_ratio/high_max": 0.0096415595326107,
      "clip_ratio/high_mean": 0.002180251499339647,
      "clip_ratio/low_mean": 0.0024700304543330276,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004650281923204602,
      "epoch": 0.4319197898668072,
      "grad_norm": 0.0504583939909935,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 871
    },
    {
      "clip_ratio/high_max": 0.010159111519897124,
      "clip_ratio/high_mean": 0.0023118088715818885,
      "clip_ratio/low_mean": 0.0027135505656588066,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005025359443607158,
      "epoch": 0.432415679407412,
      "grad_norm": 0.04914911091327667,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 872
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0736607142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15227.0,
      "completions/mean_length": 4170.29931640625,
      "completions/mean_terminated_length": 3280.515625,
      "completions/min_length": 592.0,
      "completions/min_terminated_length": 592.0,
      "epoch": 0.43291156894801686,
      "grad_norm": 0.1743612438440323,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 175213192.0,
      "reward": 0.6049107313156128,
      "reward_std": 0.21651656925678253,
      "rewards/accuracy_reward/mean": 0.6049107313156128,
      "rewards/accuracy_reward/std": 0.4894163906574249,
      "step": 873
    },
    {
      "clip_ratio/high_max": 0.00447799716857844,
      "clip_ratio/high_mean": 0.001072680483048316,
      "clip_ratio/low_mean": 0.0005487443096399147,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016214247903008072,
      "epoch": 0.43340745848862167,
      "grad_norm": 0.12835824489593506,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 874
    },
    {
      "clip_ratio/high_max": 0.005553831075303606,
      "clip_ratio/high_mean": 0.0013553968942687789,
      "clip_ratio/low_mean": 0.000775492685647805,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021308895773017866,
      "epoch": 0.4339033480292265,
      "grad_norm": 0.08954957127571106,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 875
    },
    {
      "clip_ratio/high_max": 0.0068191095160727855,
      "clip_ratio/high_mean": 0.0016692573517502751,
      "clip_ratio/low_mean": 0.000995148247511679,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026644056165423535,
      "epoch": 0.43439923756983134,
      "grad_norm": 0.06377153843641281,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 876
    },
    {
      "clip_ratio/high_max": 0.00803879860904999,
      "clip_ratio/high_mean": 0.0019475506592243619,
      "clip_ratio/low_mean": 0.0012351543491604389,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00318270505658802,
      "epoch": 0.43489512711043615,
      "grad_norm": 0.05333618074655533,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 877
    },
    {
      "clip_ratio/high_max": 0.008701858158019604,
      "clip_ratio/high_mean": 0.0021191736796026817,
      "clip_ratio/low_mean": 0.0014361713174366741,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035553450211409654,
      "epoch": 0.43539101665104096,
      "grad_norm": 0.054149337112903595,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 878
    },
    {
      "clip_ratio/high_max": 0.009431708313059062,
      "clip_ratio/high_mean": 0.0022889275551278843,
      "clip_ratio/low_mean": 0.0016073169335868442,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003896244493262202,
      "epoch": 0.43588690619164583,
      "grad_norm": 0.042574334889650345,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 879
    },
    {
      "clip_ratio/high_max": 0.009967978050553938,
      "clip_ratio/high_mean": 0.0024425542387689347,
      "clip_ratio/low_mean": 0.0017689151345621212,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004211469343317731,
      "epoch": 0.43638279573225064,
      "grad_norm": 0.041106753051280975,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 880
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0736607142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15095.0,
      "completions/mean_length": 4303.1005859375,
      "completions/mean_terminated_length": 3423.877197265625,
      "completions/min_length": 644.0,
      "completions/min_terminated_length": 644.0,
      "epoch": 0.43687868527285545,
      "grad_norm": 0.1538999080657959,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 177211757.0,
      "reward": 0.551339328289032,
      "reward_std": 0.1548805981874466,
      "rewards/accuracy_reward/mean": 0.5513392686843872,
      "rewards/accuracy_reward/std": 0.49791330099105835,
      "step": 881
    },
    {
      "clip_ratio/high_max": 0.0029548519087256864,
      "clip_ratio/high_mean": 0.0006555385368756106,
      "clip_ratio/low_mean": 0.0004470874266644387,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001102625963767423,
      "epoch": 0.4373745748134603,
      "grad_norm": 0.4993912875652313,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 882
    },
    {
      "clip_ratio/high_max": 0.0038017553652025526,
      "clip_ratio/high_mean": 0.0008304417092404037,
      "clip_ratio/low_mean": 0.0006630245982250926,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001493466322472159,
      "epoch": 0.4378704643540651,
      "grad_norm": 0.07182348519563675,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 883
    },
    {
      "clip_ratio/high_max": 0.004695879015343962,
      "clip_ratio/high_mean": 0.0010033970679614868,
      "clip_ratio/low_mean": 0.0008269851148270391,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001830382167327116,
      "epoch": 0.43836635389466994,
      "grad_norm": 0.059129808098077774,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 884
    },
    {
      "clip_ratio/high_max": 0.005353052696591476,
      "clip_ratio/high_mean": 0.0011644704577520315,
      "clip_ratio/low_mean": 0.0010229662591427768,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0021874367350847024,
      "epoch": 0.4388622434352748,
      "grad_norm": 0.06279369443655014,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 885
    },
    {
      "clip_ratio/high_max": 0.00574306249836809,
      "clip_ratio/high_mean": 0.0012647730009121005,
      "clip_ratio/low_mean": 0.0011708908643868199,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024356639016787085,
      "epoch": 0.4393581329758796,
      "grad_norm": 0.04626602679491043,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 886
    },
    {
      "clip_ratio/high_max": 0.006368720809405204,
      "clip_ratio/high_mean": 0.0013964144181954907,
      "clip_ratio/low_mean": 0.0013294558955294633,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027258702841663762,
      "epoch": 0.4398540225164845,
      "grad_norm": 0.04381536319851875,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 887
    },
    {
      "clip_ratio/high_max": 0.00669404203290469,
      "clip_ratio/high_mean": 0.0014635565303251497,
      "clip_ratio/low_mean": 0.001476926889608876,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029404834263004886,
      "epoch": 0.4403499120570893,
      "grad_norm": 0.037596434354782104,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 888
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0558035714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15137.0,
      "completions/mean_length": 3727.654296875,
      "completions/mean_terminated_length": 3040.1630859375,
      "completions/min_length": 338.0,
      "completions/min_terminated_length": 338.0,
      "epoch": 0.4408458015976941,
      "grad_norm": 0.23854117095470428,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 178943562.0,
      "reward": 0.4821428656578064,
      "reward_std": 0.23506547510623932,
      "rewards/accuracy_reward/mean": 0.4821428656578064,
      "rewards/accuracy_reward/std": 0.5002396702766418,
      "step": 889
    },
    {
      "clip_ratio/high_max": 0.005412945960415527,
      "clip_ratio/high_mean": 0.0012072762831394357,
      "clip_ratio/low_mean": 0.001121024686540295,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002328300943645445,
      "epoch": 0.44134169113829896,
      "grad_norm": 0.17134426534175873,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 890
    },
    {
      "clip_ratio/high_max": 0.0071489280089735985,
      "clip_ratio/high_mean": 0.001611888253592042,
      "clip_ratio/low_mean": 0.0017159932365302666,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003327881470795546,
      "epoch": 0.4418375806789038,
      "grad_norm": 0.10025232285261154,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 891
    },
    {
      "clip_ratio/high_max": 0.008778816161793657,
      "clip_ratio/high_mean": 0.0019385089476600115,
      "clip_ratio/low_mean": 0.0021646049721084637,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004103113958990434,
      "epoch": 0.4423334702195086,
      "grad_norm": 0.08078654110431671,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 892
    },
    {
      "clip_ratio/high_max": 0.009849452169873985,
      "clip_ratio/high_mean": 0.0021683621766896977,
      "clip_ratio/low_mean": 0.002630523415746211,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004798885607669945,
      "epoch": 0.44282935976011345,
      "grad_norm": 0.06788304448127747,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 893
    },
    {
      "clip_ratio/high_max": 0.01070223539863946,
      "clip_ratio/high_mean": 0.002370040362166037,
      "clip_ratio/low_mean": 0.0030023781255295034,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00537241846723191,
      "epoch": 0.44332524930071826,
      "grad_norm": 0.059710707515478134,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 894
    },
    {
      "clip_ratio/high_max": 0.011548483165825019,
      "clip_ratio/high_mean": 0.002534810678298527,
      "clip_ratio/low_mean": 0.0033975081214521197,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005932318865234265,
      "epoch": 0.44382113884132307,
      "grad_norm": 0.05507204681634903,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 895
    },
    {
      "clip_ratio/high_max": 0.01220234097490902,
      "clip_ratio/high_mean": 0.002698059700833255,
      "clip_ratio/low_mean": 0.0036820252021243505,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006380084929332952,
      "epoch": 0.44431702838192794,
      "grad_norm": 0.05132966488599777,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "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": 0.0669642857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15210.0,
      "completions/mean_length": 3665.69873046875,
      "completions/mean_terminated_length": 2826.39453125,
      "completions/min_length": 585.0,
      "completions/min_terminated_length": 585.0,
      "epoch": 0.44481291792253275,
      "grad_norm": 0.16573666036128998,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 180656387.0,
      "reward": 0.6004464626312256,
      "reward_std": 0.20996643602848053,
      "rewards/accuracy_reward/mean": 0.6004464030265808,
      "rewards/accuracy_reward/std": 0.49035418033599854,
      "step": 897
    },
    {
      "clip_ratio/high_max": 0.004208982229101821,
      "clip_ratio/high_mean": 0.0009348877520096721,
      "clip_ratio/low_mean": 0.0008629600062022291,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017978477612814459,
      "epoch": 0.4453088074631376,
      "grad_norm": 0.19805718958377838,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "step": 898
    },
    {
      "clip_ratio/high_max": 0.005367393914639251,
      "clip_ratio/high_mean": 0.0011916224093511119,
      "clip_ratio/low_mean": 0.0011720957708121205,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023637181830054033,
      "epoch": 0.4458046970037424,
      "grad_norm": 0.08317714184522629,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 899
    },
    {
      "clip_ratio/high_max": 0.006239509362785611,
      "clip_ratio/high_mean": 0.0013734843901147542,
      "clip_ratio/low_mean": 0.0015066302036075285,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028801145890611224,
      "epoch": 0.44630058654434723,
      "grad_norm": 0.06758680194616318,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 900
    },
    {
      "clip_ratio/high_max": 0.007257704342919169,
      "clip_ratio/high_mean": 0.0016271656584194716,
      "clip_ratio/low_mean": 0.0018885256956764351,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003515691363645601,
      "epoch": 0.4467964760849521,
      "grad_norm": 0.06702382862567902,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 901
    },
    {
      "clip_ratio/high_max": 0.00796907426774851,
      "clip_ratio/high_mean": 0.0018035017092188355,
      "clip_ratio/low_mean": 0.0022058087668028747,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004009310470792116,
      "epoch": 0.4472923656255569,
      "grad_norm": 0.05213538557291031,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 902
    },
    {
      "clip_ratio/high_max": 0.008544079002604121,
      "clip_ratio/high_mean": 0.0019407544291425438,
      "clip_ratio/low_mean": 0.0025310702069418767,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0044718246108459425,
      "epoch": 0.4477882551661617,
      "grad_norm": 0.04657350480556488,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 903
    },
    {
      "clip_ratio/high_max": 0.009110802468057955,
      "clip_ratio/high_mean": 0.002083765023144224,
      "clip_ratio/low_mean": 0.0028159801593119482,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004899745168586378,
      "epoch": 0.4482841447067666,
      "grad_norm": 0.04073105752468109,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 904
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09375,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15170.0,
      "completions/mean_length": 4715.8505859375,
      "completions/mean_terminated_length": 3614.7314453125,
      "completions/min_length": 538.0,
      "completions/min_terminated_length": 538.0,
      "epoch": 0.4487800342473714,
      "grad_norm": 0.19831746816635132,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 182847640.0,
      "reward": 0.4776785969734192,
      "reward_std": 0.18599504232406616,
      "rewards/accuracy_reward/mean": 0.4776785671710968,
      "rewards/accuracy_reward/std": 0.5000599026679993,
      "step": 905
    },
    {
      "clip_ratio/high_max": 0.004319153556934907,
      "clip_ratio/high_mean": 0.0009661904486506501,
      "clip_ratio/low_mean": 0.0008116711624666095,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017778616170289752,
      "epoch": 0.4492759237879762,
      "grad_norm": 0.109736368060112,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 906
    },
    {
      "clip_ratio/high_max": 0.005801673267342267,
      "clip_ratio/high_mean": 0.0012545149788820709,
      "clip_ratio/low_mean": 0.0010595819406375995,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0023140969237829268,
      "epoch": 0.44977181332858107,
      "grad_norm": 0.07861312478780746,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 907
    },
    {
      "clip_ratio/high_max": 0.006808901378462906,
      "clip_ratio/high_mean": 0.0014688444460944083,
      "clip_ratio/low_mean": 0.0013828114687157722,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0028516559286799747,
      "epoch": 0.4502677028691859,
      "grad_norm": 0.06872899830341339,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 908
    },
    {
      "clip_ratio/high_max": 0.007993284580152249,
      "clip_ratio/high_mean": 0.0017185957563015108,
      "clip_ratio/low_mean": 0.0017007173163392508,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034193131132269627,
      "epoch": 0.4507635924097907,
      "grad_norm": 0.05994388833642006,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 909
    },
    {
      "clip_ratio/high_max": 0.008578642658903846,
      "clip_ratio/high_mean": 0.0018734810842033767,
      "clip_ratio/low_mean": 0.001949239052237317,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003822720122116152,
      "epoch": 0.45125948195039556,
      "grad_norm": 0.056528009474277496,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 910
    },
    {
      "clip_ratio/high_max": 0.00952097821937059,
      "clip_ratio/high_mean": 0.002036040135408257,
      "clip_ratio/low_mean": 0.0022153981054771066,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00425143826760177,
      "epoch": 0.45175537149100037,
      "grad_norm": 0.0461927205324173,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 911
    },
    {
      "clip_ratio/high_max": 0.00981687495004735,
      "clip_ratio/high_mean": 0.0021304431597855,
      "clip_ratio/low_mean": 0.002441213439965395,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004571656605548924,
      "epoch": 0.45225126103160523,
      "grad_norm": 0.039836809039115906,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 912
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0848214285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14767.0,
      "completions/mean_length": 4007.62744140625,
      "completions/mean_terminated_length": 2955.4560546875,
      "completions/min_length": 403.0,
      "completions/min_terminated_length": 403.0,
      "epoch": 0.45274715057221004,
      "grad_norm": 0.3141884207725525,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 184711513.0,
      "reward": 0.5111607313156128,
      "reward_std": 0.281735897064209,
      "rewards/accuracy_reward/mean": 0.5111607313156128,
      "rewards/accuracy_reward/std": 0.5004342198371887,
      "step": 913
    },
    {
      "clip_ratio/high_max": 0.0071463833628513385,
      "clip_ratio/high_mean": 0.001745873612435389,
      "clip_ratio/low_mean": 0.001244047051841335,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002989920647451072,
      "epoch": 0.45324304011281485,
      "grad_norm": 0.1537674367427826,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 914
    },
    {
      "clip_ratio/high_max": 0.009056569746462628,
      "clip_ratio/high_mean": 0.0021949931983726856,
      "clip_ratio/low_mean": 0.001837043048567466,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004032036295029684,
      "epoch": 0.4537389296534197,
      "grad_norm": 0.10952853411436081,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 915
    },
    {
      "clip_ratio/high_max": 0.010832594089151826,
      "clip_ratio/high_mean": 0.0026356534744991222,
      "clip_ratio/low_mean": 0.0024501083878476493,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00508576190986787,
      "epoch": 0.45423481919402453,
      "grad_norm": 0.09650857746601105,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 916
    },
    {
      "clip_ratio/high_max": 0.01222961077655782,
      "clip_ratio/high_mean": 0.0030027621460249065,
      "clip_ratio/low_mean": 0.0029551090580071104,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005957871282589622,
      "epoch": 0.45473070873462934,
      "grad_norm": 0.07295569032430649,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 917
    },
    {
      "clip_ratio/high_max": 0.013463196442899061,
      "clip_ratio/high_mean": 0.003290801151706546,
      "clip_ratio/low_mean": 0.0033741922510444056,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006664993397862418,
      "epoch": 0.4552265982752342,
      "grad_norm": 0.07218813896179199,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 918
    },
    {
      "clip_ratio/high_max": 0.014497788488370134,
      "clip_ratio/high_mean": 0.003550546990481962,
      "clip_ratio/low_mean": 0.003820389551833614,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00737093658244703,
      "epoch": 0.455722487815839,
      "grad_norm": 0.08692403882741928,
      "learning_rate": 1e-06,
      "loss": -0.0013,
      "step": 919
    },
    {
      "clip_ratio/high_max": 0.01528158136352431,
      "clip_ratio/high_mean": 0.0037501025881283567,
      "clip_ratio/low_mean": 0.004114604819960732,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.007864707407861715,
      "epoch": 0.4562183773564438,
      "grad_norm": 0.062047701328992844,
      "learning_rate": 1e-06,
      "loss": -0.0014,
      "step": 920
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.049107142857142905,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15336.0,
      "completions/mean_length": 3879.9443359375,
      "completions/mean_terminated_length": 3287.077392578125,
      "completions/min_length": 458.0,
      "completions/min_terminated_length": 458.0,
      "epoch": 0.4567142668970487,
      "grad_norm": 0.2362489551305771,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 186522848.0,
      "reward": 0.5915178656578064,
      "reward_std": 0.24866807460784912,
      "rewards/accuracy_reward/mean": 0.5915178656578064,
      "rewards/accuracy_reward/std": 0.49210265278816223,
      "step": 921
    },
    {
      "clip_ratio/high_max": 0.0049713052776496625,
      "clip_ratio/high_mean": 0.0012365963957563508,
      "clip_ratio/low_mean": 0.0010205966716512194,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002257193092646048,
      "epoch": 0.4572101564376535,
      "grad_norm": 0.18838028609752655,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 922
    },
    {
      "clip_ratio/high_max": 0.006303152040345594,
      "clip_ratio/high_mean": 0.0015439111832620256,
      "clip_ratio/low_mean": 0.0014952513736261608,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030391625359698082,
      "epoch": 0.45770604597825837,
      "grad_norm": 0.17716169357299805,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 923
    },
    {
      "clip_ratio/high_max": 0.007224060071166605,
      "clip_ratio/high_mean": 0.001786226597005225,
      "clip_ratio/low_mean": 0.0019842049484850577,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003770431535485841,
      "epoch": 0.4582019355188632,
      "grad_norm": 16.375141143798828,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 924
    },
    {
      "clip_ratio/high_max": 0.008090468736554612,
      "clip_ratio/high_mean": 0.002030975299703641,
      "clip_ratio/low_mean": 0.002375751361796574,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004406726646266179,
      "epoch": 0.458697825059468,
      "grad_norm": 0.07446043193340302,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 925
    },
    {
      "clip_ratio/high_max": 0.008830307782773161,
      "clip_ratio/high_mean": 0.0022138807278224704,
      "clip_ratio/low_mean": 0.0027104571627205587,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004924337919874233,
      "epoch": 0.45919371460007286,
      "grad_norm": 0.8497096300125122,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 926
    },
    {
      "clip_ratio/high_max": 0.009307094809628325,
      "clip_ratio/high_mean": 0.0023483646300519467,
      "clip_ratio/low_mean": 0.003012928445059515,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005361293030546221,
      "epoch": 0.45968960414067767,
      "grad_norm": 0.0806020051240921,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 927
    },
    {
      "clip_ratio/high_max": 0.009601381461834535,
      "clip_ratio/high_mean": 0.002453034691825451,
      "clip_ratio/low_mean": 0.0032992097640089924,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005752244434916065,
      "epoch": 0.4601854936812825,
      "grad_norm": 0.07537529617547989,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 928
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1339285714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14977.0,
      "completions/mean_length": 5035.2548828125,
      "completions/mean_terminated_length": 3438.644287109375,
      "completions/min_length": 638.0,
      "completions/min_terminated_length": 638.0,
      "epoch": 0.46068138322188734,
      "grad_norm": 0.24095949530601501,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 188850610.0,
      "reward": 0.5022321939468384,
      "reward_std": 0.29616081714630127,
      "rewards/accuracy_reward/mean": 0.5022321343421936,
      "rewards/accuracy_reward/std": 0.5005539655685425,
      "step": 929
    },
    {
      "clip_ratio/high_max": 0.006254400994293974,
      "clip_ratio/high_mean": 0.001520936446468113,
      "clip_ratio/low_mean": 0.001044737047294575,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002565673483786668,
      "epoch": 0.46117727276249215,
      "grad_norm": 0.18769270181655884,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 930
    },
    {
      "clip_ratio/high_max": 0.008000390744200558,
      "clip_ratio/high_mean": 0.0019427187889959896,
      "clip_ratio/low_mean": 0.0016138144737425364,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035565332354963175,
      "epoch": 0.46167316230309696,
      "grad_norm": 0.12635208666324615,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 931
    },
    {
      "clip_ratio/high_max": 0.00973225771485886,
      "clip_ratio/high_mean": 0.0023619052294634457,
      "clip_ratio/low_mean": 0.0020454027642529127,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004407307990732079,
      "epoch": 0.46216905184370183,
      "grad_norm": 0.09880950301885605,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 932
    },
    {
      "clip_ratio/high_max": 0.011101106050773524,
      "clip_ratio/high_mean": 0.0026793386805366026,
      "clip_ratio/low_mean": 0.0024387267693839476,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005118065508213476,
      "epoch": 0.46266494138430664,
      "grad_norm": 0.07798777520656586,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 933
    },
    {
      "clip_ratio/high_max": 0.011990256469289307,
      "clip_ratio/high_mean": 0.002905200122313545,
      "clip_ratio/low_mean": 0.0027617699751374403,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00566697005979222,
      "epoch": 0.46316083092491145,
      "grad_norm": 0.07257165014743805,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 934
    },
    {
      "clip_ratio/high_max": 0.013043650320469169,
      "clip_ratio/high_mean": 0.0031661778411944397,
      "clip_ratio/low_mean": 0.003071156607205694,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006237334366232972,
      "epoch": 0.4636567204655163,
      "grad_norm": 0.06158596649765968,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 935
    },
    {
      "clip_ratio/high_max": 0.013658515417773742,
      "clip_ratio/high_mean": 0.003325282290461473,
      "clip_ratio/low_mean": 0.0033454511570596424,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0066707333626254695,
      "epoch": 0.4641526100061211,
      "grad_norm": 0.05491550639271736,
      "learning_rate": 1e-06,
      "loss": -0.0012,
      "step": 936
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0736607142857143,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14799.0,
      "completions/mean_length": 4385.951171875,
      "completions/mean_terminated_length": 3513.315673828125,
      "completions/min_length": 407.0,
      "completions/min_terminated_length": 407.0,
      "epoch": 0.464648499546726,
      "grad_norm": 0.21977640688419342,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 190878244.0,
      "reward": 0.5133928656578064,
      "reward_std": 0.22304978966712952,
      "rewards/accuracy_reward/mean": 0.5133928656578064,
      "rewards/accuracy_reward/std": 0.5003793835639954,
      "step": 937
    },
    {
      "clip_ratio/high_max": 0.004473419903661124,
      "clip_ratio/high_mean": 0.0010264618708788475,
      "clip_ratio/low_mean": 0.0009468012979141349,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0019732631799342926,
      "epoch": 0.4651443890873308,
      "grad_norm": 0.1380385160446167,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 938
    },
    {
      "clip_ratio/high_max": 0.005591695377006545,
      "clip_ratio/high_mean": 0.0012542120002763113,
      "clip_ratio/low_mean": 0.0013874645824216714,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002641676604980603,
      "epoch": 0.4656402786279356,
      "grad_norm": 0.09364288300275803,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 939
    },
    {
      "clip_ratio/high_max": 0.006716206091368804,
      "clip_ratio/high_mean": 0.0015324364396747114,
      "clip_ratio/low_mean": 0.0017871524214569945,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0033195888690897846,
      "epoch": 0.4661361681685405,
      "grad_norm": 0.07495006173849106,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 940
    },
    {
      "clip_ratio/high_max": 0.007673586063901894,
      "clip_ratio/high_mean": 0.0017548607072512823,
      "clip_ratio/low_mean": 0.0021412687769952754,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003896129508575541,
      "epoch": 0.4666320577091453,
      "grad_norm": 0.06458839029073715,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 941
    },
    {
      "clip_ratio/high_max": 0.008657420936287963,
      "clip_ratio/high_mean": 0.0019800236336777743,
      "clip_ratio/low_mean": 0.0024688691037226818,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004448892770597013,
      "epoch": 0.4671279472497501,
      "grad_norm": 0.05573840066790581,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 942
    },
    {
      "clip_ratio/high_max": 0.009318816741142655,
      "clip_ratio/high_mean": 0.002137825287945816,
      "clip_ratio/low_mean": 0.0028033778721692215,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00494120321036462,
      "epoch": 0.46762383679035496,
      "grad_norm": 0.04900655150413513,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 943
    },
    {
      "clip_ratio/high_max": 0.010027763393736677,
      "clip_ratio/high_mean": 0.002269439820338448,
      "clip_ratio/low_mean": 0.003145782072806469,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005415221927250968,
      "epoch": 0.4681197263309598,
      "grad_norm": 0.04570871964097023,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 944
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0379464285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15193.0,
      "completions/mean_length": 3566.73681640625,
      "completions/mean_terminated_length": 3101.572998046875,
      "completions/min_length": 465.0,
      "completions/min_terminated_length": 465.0,
      "epoch": 0.4686156158715646,
      "grad_norm": 0.15824024379253387,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 192546366.0,
      "reward": 0.6004464626312256,
      "reward_std": 0.20905573666095734,
      "rewards/accuracy_reward/mean": 0.6004464030265808,
      "rewards/accuracy_reward/std": 0.49035418033599854,
      "step": 945
    },
    {
      "clip_ratio/high_max": 0.0035929817368014483,
      "clip_ratio/high_mean": 0.0008802355860098032,
      "clip_ratio/low_mean": 0.0007507192659659268,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016309548532262852,
      "epoch": 0.46911150541216945,
      "grad_norm": 0.46077775955200195,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 946
    },
    {
      "clip_ratio/high_max": 0.004513349493208807,
      "clip_ratio/high_mean": 0.0011193642553735117,
      "clip_ratio/low_mean": 0.0010168550167009016,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002136219249223359,
      "epoch": 0.46960739495277426,
      "grad_norm": 0.08891453593969345,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 947
    },
    {
      "clip_ratio/high_max": 0.005494546108820941,
      "clip_ratio/high_mean": 0.0013495171060640132,
      "clip_ratio/low_mean": 0.0013894451445821687,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0027389622409828007,
      "epoch": 0.4701032844933791,
      "grad_norm": 0.08016993850469589,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 948
    },
    {
      "clip_ratio/high_max": 0.006404233847206342,
      "clip_ratio/high_mean": 0.0015461417199276184,
      "clip_ratio/low_mean": 0.0016596346388269012,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003205776356480783,
      "epoch": 0.47059917403398394,
      "grad_norm": 0.06641203165054321,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 949
    },
    {
      "clip_ratio/high_max": 0.007004444603808224,
      "clip_ratio/high_mean": 0.0016941427261372155,
      "clip_ratio/low_mean": 0.0019578683652525797,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036520110716082854,
      "epoch": 0.47109506357458875,
      "grad_norm": 0.059881553053855896,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 950
    },
    {
      "clip_ratio/high_max": 0.0076192917658772785,
      "clip_ratio/high_mean": 0.0018234715862490702,
      "clip_ratio/low_mean": 0.0021960884338341202,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004019560030428693,
      "epoch": 0.4715909531151936,
      "grad_norm": 0.04774344712495804,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 951
    },
    {
      "clip_ratio/high_max": 0.008129569059747155,
      "clip_ratio/high_mean": 0.0019743742129776365,
      "clip_ratio/low_mean": 0.002415553832747719,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00438992798262916,
      "epoch": 0.4720868426557984,
      "grad_norm": 0.04034626856446266,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 952
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0513392857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15285.0,
      "completions/mean_length": 4077.0224609375,
      "completions/mean_terminated_length": 3466.4140625,
      "completions/min_length": 566.0,
      "completions/min_terminated_length": 566.0,
      "epoch": 0.47258273219640323,
      "grad_norm": 0.23956650495529175,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 194441008.0,
      "reward": 0.4955357313156128,
      "reward_std": 0.21914197504520416,
      "rewards/accuracy_reward/mean": 0.4955357015132904,
      "rewards/accuracy_reward/std": 0.5005390048027039,
      "step": 953
    },
    {
      "clip_ratio/high_max": 0.005404060308137559,
      "clip_ratio/high_mean": 0.0012086783202676088,
      "clip_ratio/low_mean": 0.0009429367956954593,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002151615131424478,
      "epoch": 0.4730786217370081,
      "grad_norm": 56.76111602783203,
      "learning_rate": 1e-06,
      "loss": 0.001,
      "step": 954
    },
    {
      "clip_ratio/high_max": 0.006403176533240185,
      "clip_ratio/high_mean": 0.0014173422562180349,
      "clip_ratio/low_mean": 0.0011984945823826365,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0026158368713424807,
      "epoch": 0.4735745112776129,
      "grad_norm": 0.25936996936798096,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 955
    },
    {
      "clip_ratio/high_max": 0.007705922793320497,
      "clip_ratio/high_mean": 0.00168982774334836,
      "clip_ratio/low_mean": 0.0015181338529259847,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003207961678128868,
      "epoch": 0.4740704008182177,
      "grad_norm": 0.09200084209442139,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 956
    },
    {
      "clip_ratio/high_max": 0.008699882611836074,
      "clip_ratio/high_mean": 0.0019050549257144667,
      "clip_ratio/low_mean": 0.001851097405051405,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037561523462272817,
      "epoch": 0.4745662903588226,
      "grad_norm": 0.07645759731531143,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 957
    },
    {
      "clip_ratio/high_max": 0.00952673474967014,
      "clip_ratio/high_mean": 0.0020903444801660953,
      "clip_ratio/low_mean": 0.0021662490736957807,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004256593552952381,
      "epoch": 0.4750621798994274,
      "grad_norm": 0.07408525049686432,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 958
    },
    {
      "clip_ratio/high_max": 0.0104289430200879,
      "clip_ratio/high_mean": 0.002271083177447508,
      "clip_ratio/low_mean": 0.0024536312589589215,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0047247144127595675,
      "epoch": 0.47555806944003226,
      "grad_norm": 0.06456511467695236,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 959
    },
    {
      "clip_ratio/high_max": 0.010998502624715911,
      "clip_ratio/high_mean": 0.0024388226072460384,
      "clip_ratio/low_mean": 0.0026455608217474946,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005084383374423851,
      "epoch": 0.47605395898063707,
      "grad_norm": 0.050994161516427994,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 960
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0982142857142857,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15250.0,
      "completions/mean_length": 4287.3798828125,
      "completions/mean_terminated_length": 3081.450439453125,
      "completions/min_length": 396.0,
      "completions/min_terminated_length": 396.0,
      "epoch": 0.4765498485212419,
      "grad_norm": 0.2394161969423294,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 196428330.0,
      "reward": 0.5446428656578064,
      "reward_std": 0.255958616733551,
      "rewards/accuracy_reward/mean": 0.5446428656578064,
      "rewards/accuracy_reward/std": 0.49855974316596985,
      "step": 961
    },
    {
      "clip_ratio/high_max": 0.006203767014085315,
      "clip_ratio/high_mean": 0.0013834458570727293,
      "clip_ratio/low_mean": 0.0010231113841427941,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0024065572497420362,
      "epoch": 0.47704573806184675,
      "grad_norm": 0.13435421884059906,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 962
    },
    {
      "clip_ratio/high_max": 0.008212852131691761,
      "clip_ratio/high_mean": 0.001792485035593927,
      "clip_ratio/low_mean": 0.0014088776428025085,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003201362691470422,
      "epoch": 0.47754162760245156,
      "grad_norm": 0.13069799542427063,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 963
    },
    {
      "clip_ratio/high_max": 0.009649175168306101,
      "clip_ratio/high_mean": 0.0021352429930630024,
      "clip_ratio/low_mean": 0.0018377392576667262,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039729822710796725,
      "epoch": 0.47803751714305637,
      "grad_norm": 0.09508538246154785,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 964
    },
    {
      "clip_ratio/high_max": 0.011188115622644546,
      "clip_ratio/high_mean": 0.0024600257147540106,
      "clip_ratio/low_mean": 0.002178504743483245,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004638530477677705,
      "epoch": 0.47853340668366123,
      "grad_norm": 0.0715222880244255,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 965
    },
    {
      "clip_ratio/high_max": 0.012110836560168536,
      "clip_ratio/high_mean": 0.0026822303138942516,
      "clip_ratio/low_mean": 0.002524245532981695,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005206475852901349,
      "epoch": 0.47902929622426604,
      "grad_norm": 0.07339341193437576,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 966
    },
    {
      "clip_ratio/high_max": 0.01310012603062205,
      "clip_ratio/high_mean": 0.002876516379728855,
      "clip_ratio/low_mean": 0.0027841018696790343,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0056606183097756,
      "epoch": 0.47952518576487085,
      "grad_norm": 0.05588371679186821,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 967
    },
    {
      "clip_ratio/high_max": 0.013902594808314461,
      "clip_ratio/high_mean": 0.0030660125421491102,
      "clip_ratio/low_mean": 0.0030156195681456666,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006081632069253828,
      "epoch": 0.4800210753054757,
      "grad_norm": 0.05212387070059776,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 968
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1026785714285714,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14594.0,
      "completions/mean_length": 4816.79052734375,
      "completions/mean_terminated_length": 3610.353271484375,
      "completions/min_length": 454.0,
      "completions/min_terminated_length": 454.0,
      "epoch": 0.48051696484608053,
      "grad_norm": 0.22279512882232666,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 198654204.0,
      "reward": 0.4531250298023224,
      "reward_std": 0.1772059053182602,
      "rewards/accuracy_reward/mean": 0.453125,
      "rewards/accuracy_reward/std": 0.49835437536239624,
      "step": 969
    },
    {
      "clip_ratio/high_max": 0.003983345159213059,
      "clip_ratio/high_mean": 0.0008358076161130157,
      "clip_ratio/low_mean": 0.0006847468181376826,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0015205544277705485,
      "epoch": 0.48101285438668534,
      "grad_norm": 4.7309064865112305,
      "learning_rate": 1e-06,
      "loss": 0.0003,
      "step": 970
    },
    {
      "clip_ratio/high_max": 0.004706267443907564,
      "clip_ratio/high_mean": 0.0009712838182167616,
      "clip_ratio/low_mean": 0.0008480970259370224,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001819380836423079,
      "epoch": 0.4815087439272902,
      "grad_norm": 0.09329564869403839,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 971
    },
    {
      "clip_ratio/high_max": 0.005434567443444394,
      "clip_ratio/high_mean": 0.0011199068985661143,
      "clip_ratio/low_mean": 0.0011251180726503662,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002245024953936081,
      "epoch": 0.482004633467895,
      "grad_norm": 0.07613395154476166,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 972
    },
    {
      "clip_ratio/high_max": 0.006461654487793567,
      "clip_ratio/high_mean": 0.0013332754524526536,
      "clip_ratio/low_mean": 0.0013728393796554883,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002706114846660057,
      "epoch": 0.4825005230084999,
      "grad_norm": 0.06102833151817322,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 973
    },
    {
      "clip_ratio/high_max": 0.0071823209946160205,
      "clip_ratio/high_mean": 0.0014750139757779834,
      "clip_ratio/low_mean": 0.0016197691518300417,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0030947831232879253,
      "epoch": 0.4829964125491047,
      "grad_norm": 0.05729799345135689,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 974
    },
    {
      "clip_ratio/high_max": 0.008018009360966971,
      "clip_ratio/high_mean": 0.0016277886802527064,
      "clip_ratio/low_mean": 0.001850973185582916,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034787618228619976,
      "epoch": 0.4834923020897095,
      "grad_norm": 0.05040638893842697,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 975
    },
    {
      "clip_ratio/high_max": 0.00855858653812902,
      "clip_ratio/high_mean": 0.001759706388838822,
      "clip_ratio/low_mean": 0.002038968690612819,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0037986750867275987,
      "epoch": 0.48398819163031437,
      "grad_norm": 0.046409349888563156,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 976
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0848214285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14360.0,
      "completions/mean_length": 4480.5380859375,
      "completions/mean_terminated_length": 3472.197509765625,
      "completions/min_length": 536.0,
      "completions/min_terminated_length": 536.0,
      "epoch": 0.4844840811709192,
      "grad_norm": 0.19158810377120972,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 200729157.0,
      "reward": 0.5446428656578064,
      "reward_std": 0.22207047045230865,
      "rewards/accuracy_reward/mean": 0.5446428656578064,
      "rewards/accuracy_reward/std": 0.49855974316596985,
      "step": 977
    },
    {
      "clip_ratio/high_max": 0.004142770030739484,
      "clip_ratio/high_mean": 0.0008861720104960114,
      "clip_ratio/low_mean": 0.0008520653242385379,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0017382373357577308,
      "epoch": 0.484979970711524,
      "grad_norm": 702.9291381835938,
      "learning_rate": 1e-06,
      "loss": 0.0061,
      "step": 978
    },
    {
      "clip_ratio/high_max": 0.004906820682663238,
      "clip_ratio/high_mean": 0.001047498696607363,
      "clip_ratio/low_mean": 0.0010465709291338499,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0020940696267643943,
      "epoch": 0.48547586025212885,
      "grad_norm": 0.9474620819091797,
      "learning_rate": 1e-06,
      "loss": 0.0001,
      "step": 979
    },
    {
      "clip_ratio/high_max": 0.005767632625065744,
      "clip_ratio/high_mean": 0.001222141008838662,
      "clip_ratio/low_mean": 0.0012980412859633361,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0025201822909366456,
      "epoch": 0.48597174979273366,
      "grad_norm": 0.09678491204977036,
      "learning_rate": 1e-06,
      "loss": -0.0002,
      "step": 980
    },
    {
      "clip_ratio/high_max": 0.007003190214163624,
      "clip_ratio/high_mean": 0.0014631894564445247,
      "clip_ratio/low_mean": 0.0015581977938836644,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003021387249646068,
      "epoch": 0.4864676393333385,
      "grad_norm": 0.07004240155220032,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 981
    },
    {
      "clip_ratio/high_max": 0.007521770505263703,
      "clip_ratio/high_mean": 0.0015856638733566797,
      "clip_ratio/low_mean": 0.0018240433928440325,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0034097072857548483,
      "epoch": 0.48696352887394334,
      "grad_norm": 0.06467533856630325,
      "learning_rate": 1e-06,
      "loss": -0.0005,
      "step": 982
    },
    {
      "clip_ratio/high_max": 0.00843491334853752,
      "clip_ratio/high_mean": 0.0017683011437839014,
      "clip_ratio/low_mean": 0.00212395123207898,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00389225237267965,
      "epoch": 0.48745941841454815,
      "grad_norm": 0.06209000200033188,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 983
    },
    {
      "clip_ratio/high_max": 0.009039175612997497,
      "clip_ratio/high_mean": 0.0018966497364090174,
      "clip_ratio/low_mean": 0.0023470546298085537,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0042437043757672654,
      "epoch": 0.487955307955153,
      "grad_norm": 0.0639437735080719,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 984
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0647321428571429,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 14730.0,
      "completions/mean_length": 3961.58056640625,
      "completions/mean_terminated_length": 3172.66845703125,
      "completions/min_length": 453.0,
      "completions/min_terminated_length": 453.0,
      "epoch": 0.4884511974957578,
      "grad_norm": 0.2136445790529251,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "num_tokens": 202577649.0,
      "reward": 0.4888392984867096,
      "reward_std": 0.2239413559436798,
      "rewards/accuracy_reward/mean": 0.4888392984867096,
      "rewards/accuracy_reward/std": 0.5004342198371887,
      "step": 985
    },
    {
      "clip_ratio/high_max": 0.004949687121552415,
      "clip_ratio/high_mean": 0.001164187876156575,
      "clip_ratio/low_mean": 0.001048341464411351,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0022125293289718684,
      "epoch": 0.48894708703636264,
      "grad_norm": 0.17942523956298828,
      "learning_rate": 1e-06,
      "loss": -0.0,
      "step": 986
    },
    {
      "clip_ratio/high_max": 0.006682007106064702,
      "clip_ratio/high_mean": 0.0014980861367348552,
      "clip_ratio/low_mean": 0.001484615931076405,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0029827020662196446,
      "epoch": 0.4894429765769675,
      "grad_norm": 0.11465496569871902,
      "learning_rate": 1e-06,
      "loss": -0.0003,
      "step": 987
    },
    {
      "clip_ratio/high_max": 0.007605832379340427,
      "clip_ratio/high_mean": 0.0017463834776663134,
      "clip_ratio/low_mean": 0.0019090073287770792,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003655390813946724,
      "epoch": 0.4899388661175723,
      "grad_norm": 0.08714410662651062,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 988
    },
    {
      "clip_ratio/high_max": 0.008747771891648881,
      "clip_ratio/high_mean": 0.0020092417043997557,
      "clip_ratio/low_mean": 0.0022874863138895307,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004296728018744034,
      "epoch": 0.4904347556581771,
      "grad_norm": 0.08845504373311996,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 989
    },
    {
      "clip_ratio/high_max": 0.009539950453472557,
      "clip_ratio/high_mean": 0.002199360355916724,
      "clip_ratio/low_mean": 0.0027215005275138537,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004920860872516641,
      "epoch": 0.490930645198782,
      "grad_norm": 0.17693985998630524,
      "learning_rate": 1e-06,
      "loss": -0.0008,
      "step": 990
    },
    {
      "clip_ratio/high_max": 0.010303709063009592,
      "clip_ratio/high_mean": 0.0023675163297411927,
      "clip_ratio/low_mean": 0.003015921774931485,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005383438130593277,
      "epoch": 0.4914265347393868,
      "grad_norm": 0.07366528362035751,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 991
    },
    {
      "clip_ratio/high_max": 0.010779929973068647,
      "clip_ratio/high_mean": 0.002483052138813946,
      "clip_ratio/low_mean": 0.0033364632126904326,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005819515370603767,
      "epoch": 0.4919224242799916,
      "grad_norm": 0.056267958134412766,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 992
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0848214285714286,
      "completions/max_length": 15360.0,
      "completions/max_terminated_length": 15151.0,
      "completions/mean_length": 4264.83935546875,
      "completions/mean_terminated_length": 3236.50732421875,
      "completions/min_length": 276.0,
      "completions/min_terminated_length": 276.0,
      "epoch": 0.4924183138205965,
      "grad_norm": 0.2338014841079712,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "num_tokens": 204566425.0,
      "reward": 0.5334821939468384,
      "reward_std": 0.223872110247612,
      "rewards/accuracy_reward/mean": 0.5334821343421936,
      "rewards/accuracy_reward/std": 0.4994353950023651,
      "step": 993
    },
    {
      "clip_ratio/high_max": 0.0055666660409769975,
      "clip_ratio/high_mean": 0.0011433825491167227,
      "clip_ratio/low_mean": 0.0011573854603739164,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.002300768004715792,
      "epoch": 0.4929142033612013,
      "grad_norm": 0.13174571096897125,
      "learning_rate": 1e-06,
      "loss": -0.0001,
      "step": 994
    },
    {
      "clip_ratio/high_max": 0.006978628414799459,
      "clip_ratio/high_mean": 0.001435181322221979,
      "clip_ratio/low_mean": 0.00168515859627405,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0031203399112200714,
      "epoch": 0.4934100929018061,
      "grad_norm": 0.1124512329697609,
      "learning_rate": 1e-06,
      "loss": -0.0004,
      "step": 995
    },
    {
      "clip_ratio/high_max": 0.008461850480671274,
      "clip_ratio/high_mean": 0.0017345205374112993,
      "clip_ratio/low_mean": 0.0021670986088793143,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0039016191576592973,
      "epoch": 0.49390598244241096,
      "grad_norm": 0.08595485240221024,
      "learning_rate": 1e-06,
      "loss": -0.0006,
      "step": 996
    },
    {
      "clip_ratio/high_max": 0.00963830480031902,
      "clip_ratio/high_mean": 0.0019641476928882184,
      "clip_ratio/low_mean": 0.002592599352283287,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004556747011520201,
      "epoch": 0.4944018719830158,
      "grad_norm": 0.07002006471157074,
      "learning_rate": 1e-06,
      "loss": -0.0007,
      "step": 997
    },
    {
      "clip_ratio/high_max": 0.010583490195131162,
      "clip_ratio/high_mean": 0.002153337743038719,
      "clip_ratio/low_mean": 0.0031309684277402994,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005284306183966692,
      "epoch": 0.49489776152362064,
      "grad_norm": 0.062010493129491806,
      "learning_rate": 1e-06,
      "loss": -0.0009,
      "step": 998
    },
    {
      "clip_ratio/high_max": 0.011191946286999155,
      "clip_ratio/high_mean": 0.002291137352131045,
      "clip_ratio/low_mean": 0.003431323656513996,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005722460980905453,
      "epoch": 0.49539365106422545,
      "grad_norm": 0.06856238096952438,
      "learning_rate": 1e-06,
      "loss": -0.001,
      "step": 999
    },
    {
      "clip_ratio/high_max": 0.011989421927864896,
      "clip_ratio/high_mean": 0.0024586047952652734,
      "clip_ratio/low_mean": 0.0037321477170735307,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006190752455950133,
      "epoch": 0.49588954060483026,
      "grad_norm": 0.05438092723488808,
      "learning_rate": 1e-06,
      "loss": -0.0011,
      "step": 1000
    },
    {
      "epoch": 0.49588954060483026,
      "step": 1000,
      "total_flos": 0.0,
      "train_loss": -0.00041466000025666894,
      "train_runtime": 232889.004,
      "train_samples_per_second": 1.924,
      "train_steps_per_second": 0.004
    }
  ],
  "logging_steps": 1,
  "max_steps": 1000,
  "num_input_tokens_seen": 204566425,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}