{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.9952153110047846,
  "eval_steps": 500,
  "global_step": 104,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 885.7360992431641,
      "epoch": 0.019138755980861243,
      "grad_norm": 0.0571745065537259,
      "kl": 0.0,
      "learning_rate": 9.09090909090909e-08,
      "loss": 0.0224,
      "reward": 0.6623263955116272,
      "reward_std": 0.20746709778904915,
      "rewards/accuracy_reward": 0.2500000037252903,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4123263880610466,
      "step": 1
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 855.6076354980469,
      "epoch": 0.03827751196172249,
      "grad_norm": 0.06704915151383242,
      "kl": 0.0,
      "learning_rate": 1.818181818181818e-07,
      "loss": 0.0199,
      "reward": 0.6901041716337204,
      "reward_std": 0.22956417128443718,
      "rewards/accuracy_reward": 0.2465277761220932,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4435763880610466,
      "step": 2
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 904.9861297607422,
      "epoch": 0.05741626794258373,
      "grad_norm": 0.05238384111260958,
      "kl": 2.7120113372802734e-05,
      "learning_rate": 2.727272727272727e-07,
      "loss": 0.0109,
      "reward": 0.5494791641831398,
      "reward_std": 0.17554907500743866,
      "rewards/accuracy_reward": 0.17361111380159855,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.375868059694767,
      "step": 3
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 880.2083435058594,
      "epoch": 0.07655502392344497,
      "grad_norm": 0.05735410600314712,
      "kl": 2.5987625122070312e-05,
      "learning_rate": 3.636363636363636e-07,
      "loss": 0.0214,
      "reward": 0.6362847238779068,
      "reward_std": 0.22956416755914688,
      "rewards/accuracy_reward": 0.2256944440305233,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4105902835726738,
      "step": 4
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 867.8368225097656,
      "epoch": 0.09569377990430622,
      "grad_norm": 0.056581697743245854,
      "kl": 2.8789043426513672e-05,
      "learning_rate": 4.545454545454545e-07,
      "loss": 0.0117,
      "reward": 0.665798619389534,
      "reward_std": 0.1829147655516863,
      "rewards/accuracy_reward": 0.25,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4157986119389534,
      "step": 5
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 884.0381927490234,
      "epoch": 0.11483253588516747,
      "grad_norm": 0.060995509922658335,
      "kl": 2.4437904357910156e-05,
      "learning_rate": 5.454545454545454e-07,
      "loss": 0.0206,
      "reward": 0.659722238779068,
      "reward_std": 0.24306795001029968,
      "rewards/accuracy_reward": 0.243055559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4166666716337204,
      "step": 6
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 878.7083435058594,
      "epoch": 0.1339712918660287,
      "grad_norm": 0.06312478683496295,
      "kl": 2.7120113372802734e-05,
      "learning_rate": 6.363636363636363e-07,
      "loss": 0.0152,
      "reward": 0.6484375149011612,
      "reward_std": 0.224653709679842,
      "rewards/accuracy_reward": 0.243055559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4053819477558136,
      "step": 7
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 871.4930572509766,
      "epoch": 0.15311004784688995,
      "grad_norm": 0.06419727373029128,
      "kl": 2.6404857635498047e-05,
      "learning_rate": 7.272727272727272e-07,
      "loss": 0.0246,
      "reward": 0.6631944477558136,
      "reward_std": 0.2381574772298336,
      "rewards/accuracy_reward": 0.2395833283662796,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.423611119389534,
      "step": 8
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 883.0347290039062,
      "epoch": 0.1722488038277512,
      "grad_norm": 0.07232208605276093,
      "kl": 2.5197863578796387e-05,
      "learning_rate": 8.181818181818182e-07,
      "loss": 0.0217,
      "reward": 0.6892361044883728,
      "reward_std": 0.270075511187315,
      "rewards/accuracy_reward": 0.2847222164273262,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4045138880610466,
      "step": 9
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 869.8229217529297,
      "epoch": 0.19138755980861244,
      "grad_norm": 0.05684777755384174,
      "kl": 2.7745962142944336e-05,
      "learning_rate": 9.09090909090909e-07,
      "loss": 0.0053,
      "reward": 0.697048619389534,
      "reward_std": 0.2197432443499565,
      "rewards/accuracy_reward": 0.2916666641831398,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4053819477558136,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 850.7152862548828,
      "epoch": 0.21052631578947367,
      "grad_norm": 0.07180408830721081,
      "kl": 2.5004148483276367e-05,
      "learning_rate": 1e-06,
      "loss": 0.0299,
      "reward": 0.706597238779068,
      "reward_std": 0.28726211935281754,
      "rewards/accuracy_reward": 0.2673611156642437,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.439236119389534,
      "step": 11
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 868.28125,
      "epoch": 0.22966507177033493,
      "grad_norm": 0.0646871027868035,
      "kl": 2.7686357498168945e-05,
      "learning_rate": 9.99743270918373e-07,
      "loss": 0.0177,
      "reward": 0.7300347238779068,
      "reward_std": 0.24429556727409363,
      "rewards/accuracy_reward": 0.2881944440305233,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4418402761220932,
      "step": 12
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 864.7395935058594,
      "epoch": 0.24880382775119617,
      "grad_norm": 0.06556460276965595,
      "kl": 2.7000904083251953e-05,
      "learning_rate": 9.989733766060318e-07,
      "loss": 0.0163,
      "reward": 0.712673619389534,
      "reward_std": 0.20255662500858307,
      "rewards/accuracy_reward": 0.2777777910232544,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4348958283662796,
      "step": 13
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 879.0590209960938,
      "epoch": 0.2679425837320574,
      "grad_norm": 0.06296015481337248,
      "kl": 3.0308961868286133e-05,
      "learning_rate": 9.976911955263528e-07,
      "loss": 0.0101,
      "reward": 0.6510416716337204,
      "reward_std": 0.22588133066892624,
      "rewards/accuracy_reward": 0.2326388917863369,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4184027761220932,
      "step": 14
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 885.1006927490234,
      "epoch": 0.28708133971291866,
      "grad_norm": 0.06103661524458471,
      "kl": 3.299117088317871e-05,
      "learning_rate": 9.958981906712076e-07,
      "loss": 0.0203,
      "reward": 0.677083358168602,
      "reward_std": 0.272530734539032,
      "rewards/accuracy_reward": 0.2534722238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4236111119389534,
      "step": 15
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 874.2257080078125,
      "epoch": 0.3062200956937799,
      "grad_norm": 0.0764541954488641,
      "kl": 5.513429641723633e-05,
      "learning_rate": 9.935964078916617e-07,
      "loss": 0.0251,
      "reward": 0.681423619389534,
      "reward_std": 0.21483278647065163,
      "rewards/accuracy_reward": 0.2708333395421505,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4105902835726738,
      "step": 16
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 852.8472290039062,
      "epoch": 0.3253588516746411,
      "grad_norm": 0.0763344765825447,
      "kl": 0.00010287761688232422,
      "learning_rate": 9.907884735636224e-07,
      "loss": 0.0222,
      "reward": 0.7560763955116272,
      "reward_std": 0.2909449674189091,
      "rewards/accuracy_reward": 0.3263888880610466,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4296875074505806,
      "step": 17
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 873.6388854980469,
      "epoch": 0.3444976076555024,
      "grad_norm": 0.07015709920099024,
      "kl": 0.0001728534698486328,
      "learning_rate": 9.874775915910958e-07,
      "loss": 0.0186,
      "reward": 0.7039930671453476,
      "reward_std": 0.224653709679842,
      "rewards/accuracy_reward": 0.274305559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4296875,
      "step": 18
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 865.6250152587891,
      "epoch": 0.36363636363636365,
      "grad_norm": 0.06286838529602348,
      "kl": 0.0002143383026123047,
      "learning_rate": 9.836675397504734e-07,
      "loss": 0.0188,
      "reward": 0.724826380610466,
      "reward_std": 0.23938511312007904,
      "rewards/accuracy_reward": 0.3090277835726738,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.415798619389534,
      "step": 19
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 851.5555572509766,
      "epoch": 0.3827751196172249,
      "grad_norm": 0.0652308138114061,
      "kl": 0.00026702880859375,
      "learning_rate": 9.793626653800218e-07,
      "loss": 0.019,
      "reward": 0.7230902761220932,
      "reward_std": 0.24184033274650574,
      "rewards/accuracy_reward": 0.288194440305233,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4348958283662796,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 861.40625,
      "epoch": 0.4019138755980861,
      "grad_norm": 0.054173134249339484,
      "kl": 0.00030875205993652344,
      "learning_rate": 9.74567880419491e-07,
      "loss": 0.0092,
      "reward": 0.6545138955116272,
      "reward_std": 0.2086947076022625,
      "rewards/accuracy_reward": 0.2361111119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4184027835726738,
      "step": 21
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 839.1805725097656,
      "epoch": 0.42105263157894735,
      "grad_norm": 0.06484920729256737,
      "kl": 0.0003552436828613281,
      "learning_rate": 9.692886558055006e-07,
      "loss": 0.019,
      "reward": 0.7100694626569748,
      "reward_std": 0.23079179227352142,
      "rewards/accuracy_reward": 0.2881944440305233,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4218749925494194,
      "step": 22
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 889.3159790039062,
      "epoch": 0.44019138755980863,
      "grad_norm": 0.0673083530273057,
      "kl": 0.0004329681396484375,
      "learning_rate": 9.635310152291037e-07,
      "loss": 0.02,
      "reward": 0.6163194626569748,
      "reward_std": 0.20132901147007942,
      "rewards/accuracy_reward": 0.2222222201526165,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.3940972238779068,
      "step": 23
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 872.4722290039062,
      "epoch": 0.45933014354066987,
      "grad_norm": 0.05489657330998097,
      "kl": 0.0005602836608886719,
      "learning_rate": 9.573015282626445e-07,
      "loss": 0.0217,
      "reward": 0.6996527910232544,
      "reward_std": 0.22097085788846016,
      "rewards/accuracy_reward": 0.274305559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4253472313284874,
      "step": 24
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 864.3889007568359,
      "epoch": 0.4784688995215311,
      "grad_norm": 0.06394072732120833,
      "kl": 0.0007648468017578125,
      "learning_rate": 9.506073028637582e-07,
      "loss": 0.0165,
      "reward": 0.6866319477558136,
      "reward_std": 0.26148219406604767,
      "rewards/accuracy_reward": 0.274305559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4123263880610466,
      "step": 25
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 859.4375152587891,
      "epoch": 0.49760765550239233,
      "grad_norm": 0.06801792884391539,
      "kl": 0.0011510848999023438,
      "learning_rate": 9.434559772650618e-07,
      "loss": 0.0179,
      "reward": 0.7569444477558136,
      "reward_std": 0.2357022576034069,
      "rewards/accuracy_reward": 0.329861119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4270833358168602,
      "step": 26
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 882.1701354980469,
      "epoch": 0.5167464114832536,
      "grad_norm": 0.06594394459261921,
      "kl": 0.001209259033203125,
      "learning_rate": 9.358557112587928e-07,
      "loss": 0.0201,
      "reward": 0.6857638955116272,
      "reward_std": 0.2651650384068489,
      "rewards/accuracy_reward": 0.2777777761220932,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.407986119389534,
      "step": 27
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 848.2395935058594,
      "epoch": 0.5358851674641149,
      "grad_norm": 0.0697274048868991,
      "kl": 0.001598358154296875,
      "learning_rate": 9.278151768863375e-07,
      "loss": 0.029,
      "reward": 0.7039930671453476,
      "reward_std": 0.2835792750120163,
      "rewards/accuracy_reward": 0.2881944440305233,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4157986119389534,
      "step": 28
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 858.8889007568359,
      "epoch": 0.5550239234449761,
      "grad_norm": 0.06562680518580778,
      "kl": 0.0018520355224609375,
      "learning_rate": 9.193435485432744e-07,
      "loss": 0.0227,
      "reward": 0.7395833283662796,
      "reward_std": 0.2676202692091465,
      "rewards/accuracy_reward": 0.3159722313284874,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.423611119389534,
      "step": 29
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 809.0694580078125,
      "epoch": 0.5741626794258373,
      "grad_norm": 0.07438927904369452,
      "kl": 0.002056121826171875,
      "learning_rate": 9.104504925112231e-07,
      "loss": 0.0256,
      "reward": 0.8541666716337204,
      "reward_std": 0.2749859616160393,
      "rewards/accuracy_reward": 0.399305559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.454861119389534,
      "step": 30
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 871.6979217529297,
      "epoch": 0.5933014354066986,
      "grad_norm": 0.049564487099522436,
      "kl": 0.00186920166015625,
      "learning_rate": 9.011461559284413e-07,
      "loss": 0.0095,
      "reward": 0.7387152910232544,
      "reward_std": 0.15099675580859184,
      "rewards/accuracy_reward": 0.3402777835726738,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.3984375074505806,
      "step": 31
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 834.6250152587891,
      "epoch": 0.6124401913875598,
      "grad_norm": 0.06896614921110253,
      "kl": 0.0023975372314453125,
      "learning_rate": 8.914411552117559e-07,
      "loss": 0.0227,
      "reward": 0.7560763955116272,
      "reward_std": 0.26884788274765015,
      "rewards/accuracy_reward": 0.336805559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4192708358168602,
      "step": 32
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 845.7048645019531,
      "epoch": 0.631578947368421,
      "grad_norm": 0.07262913185512221,
      "kl": 0.00289154052734375,
      "learning_rate": 8.813465639430386e-07,
      "loss": 0.0293,
      "reward": 0.7838541716337204,
      "reward_std": 0.26393742114305496,
      "rewards/accuracy_reward": 0.329861119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4539930522441864,
      "step": 33
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 825.7361145019531,
      "epoch": 0.6507177033492823,
      "grad_norm": 0.06923687691149574,
      "kl": 0.002803802490234375,
      "learning_rate": 8.70873900234049e-07,
      "loss": 0.0156,
      "reward": 0.8272569626569748,
      "reward_std": 0.2590269446372986,
      "rewards/accuracy_reward": 0.392361119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4348958358168602,
      "step": 34
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 813.9930725097656,
      "epoch": 0.6698564593301436,
      "grad_norm": 0.0634139401535218,
      "kl": 0.002788543701171875,
      "learning_rate": 8.600351135840589e-07,
      "loss": 0.0179,
      "reward": 0.8159722238779068,
      "reward_std": 0.24306795746088028,
      "rewards/accuracy_reward": 0.3715277835726738,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4444444477558136,
      "step": 35
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 888.8159942626953,
      "epoch": 0.6889952153110048,
      "grad_norm": 0.06115251503429904,
      "kl": 0.00292205810546875,
      "learning_rate": 8.488425712452576e-07,
      "loss": 0.0181,
      "reward": 0.6588541716337204,
      "reward_std": 0.2221984826028347,
      "rewards/accuracy_reward": 0.2777777798473835,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.3810763880610466,
      "step": 36
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 837.6458435058594,
      "epoch": 0.7081339712918661,
      "grad_norm": 0.06174697280825877,
      "kl": 0.002880096435546875,
      "learning_rate": 8.373090441114925e-07,
      "loss": 0.0255,
      "reward": 0.7873263955116272,
      "reward_std": 0.2197432555258274,
      "rewards/accuracy_reward": 0.3541666716337204,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4331597238779068,
      "step": 37
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 855.8090362548828,
      "epoch": 0.7272727272727273,
      "grad_norm": 0.07806365108296032,
      "kl": 0.00281524658203125,
      "learning_rate": 8.254476921464483e-07,
      "loss": 0.0277,
      "reward": 0.7621527910232544,
      "reward_std": 0.24797841161489487,
      "rewards/accuracy_reward": 0.3298611119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4322916716337204,
      "step": 38
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 857.6840515136719,
      "epoch": 0.7464114832535885,
      "grad_norm": 0.07203148213149006,
      "kl": 0.003299713134765625,
      "learning_rate": 8.132720493678884e-07,
      "loss": 0.023,
      "reward": 0.7803819328546524,
      "reward_std": 0.24184033647179604,
      "rewards/accuracy_reward": 0.3402777835726738,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4401041716337204,
      "step": 39
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 838.0764007568359,
      "epoch": 0.7655502392344498,
      "grad_norm": 0.06474279173647825,
      "kl": 0.00308990478515625,
      "learning_rate": 8.007960084050963e-07,
      "loss": 0.0212,
      "reward": 0.8741319477558136,
      "reward_std": 0.23447464406490326,
      "rewards/accuracy_reward": 0.4340277761220932,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4401041716337204,
      "step": 40
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 859.78125,
      "epoch": 0.784688995215311,
      "grad_norm": 0.06299207628008677,
      "kl": 0.003017425537109375,
      "learning_rate": 7.88033804647133e-07,
      "loss": 0.0193,
      "reward": 0.7256944626569748,
      "reward_std": 0.24306795746088028,
      "rewards/accuracy_reward": 0.3159722238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4097222238779068,
      "step": 41
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 840.0625152587891,
      "epoch": 0.8038277511961722,
      "grad_norm": 0.07344429258616171,
      "kl": 0.003047943115234375,
      "learning_rate": 7.75e-07,
      "loss": 0.0251,
      "reward": 0.8220486044883728,
      "reward_std": 0.20992231741547585,
      "rewards/accuracy_reward": 0.392361119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4296875074505806,
      "step": 42
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 819.0382232666016,
      "epoch": 0.8229665071770335,
      "grad_norm": 0.061499395941431584,
      "kl": 0.003070831298828125,
      "learning_rate": 7.6170946627124e-07,
      "loss": 0.0223,
      "reward": 0.8151041716337204,
      "reward_std": 0.20992232859134674,
      "rewards/accuracy_reward": 0.3784722164273262,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4366319477558136,
      "step": 43
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 863.3264007568359,
      "epoch": 0.8421052631578947,
      "grad_norm": 0.06745163743583922,
      "kl": 0.0031585693359375,
      "learning_rate": 7.481773682009354e-07,
      "loss": 0.0196,
      "reward": 0.7387152910232544,
      "reward_std": 0.20501185581088066,
      "rewards/accuracy_reward": 0.3159722238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.422743059694767,
      "step": 44
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 830.9930725097656,
      "epoch": 0.861244019138756,
      "grad_norm": 0.07628628253732472,
      "kl": 0.0036468505859375,
      "learning_rate": 7.344191461584643e-07,
      "loss": 0.0256,
      "reward": 0.8289930671453476,
      "reward_std": 0.2492060288786888,
      "rewards/accuracy_reward": 0.3819444477558136,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4470486119389534,
      "step": 45
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 834.5590362548828,
      "epoch": 0.8803827751196173,
      "grad_norm": 0.0697101918978914,
      "kl": 0.004314422607421875,
      "learning_rate": 7.204504985247592e-07,
      "loss": 0.0163,
      "reward": 0.8498263955116272,
      "reward_std": 0.24184034019708633,
      "rewards/accuracy_reward": 0.40625,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4435763955116272,
      "step": 46
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 846.2569427490234,
      "epoch": 0.8995215311004785,
      "grad_norm": 0.06859248959883628,
      "kl": 0.003566741943359375,
      "learning_rate": 7.06287363780169e-07,
      "loss": 0.0134,
      "reward": 0.7534722238779068,
      "reward_std": 0.22097086533904076,
      "rewards/accuracy_reward": 0.3333333358168602,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4201388955116272,
      "step": 47
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 813.9062652587891,
      "epoch": 0.9186602870813397,
      "grad_norm": 0.07616625495450065,
      "kl": 0.00397491455078125,
      "learning_rate": 6.91945902318365e-07,
      "loss": 0.0284,
      "reward": 0.862847238779068,
      "reward_std": 0.2700754962861538,
      "rewards/accuracy_reward": 0.4131944477558136,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4496527761220932,
      "step": 48
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 843.1215209960938,
      "epoch": 0.937799043062201,
      "grad_norm": 0.06987264134329038,
      "kl": 0.003589630126953125,
      "learning_rate": 6.774424780070367e-07,
      "loss": 0.0194,
      "reward": 0.8289930671453476,
      "reward_std": 0.2786688208580017,
      "rewards/accuracy_reward": 0.4027777761220932,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4262152835726738,
      "step": 49
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 848.1909790039062,
      "epoch": 0.9569377990430622,
      "grad_norm": 0.07535010112426463,
      "kl": 0.00441741943359375,
      "learning_rate": 6.627936395164242e-07,
      "loss": 0.0225,
      "reward": 0.8229166716337204,
      "reward_std": 0.27744119614362717,
      "rewards/accuracy_reward": 0.3645833358168602,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4583333358168602,
      "step": 50
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 844.8402862548828,
      "epoch": 0.9760765550239234,
      "grad_norm": 0.07005632661144727,
      "kl": 0.004360198974609375,
      "learning_rate": 6.480161014369832e-07,
      "loss": 0.0217,
      "reward": 0.8645833283662796,
      "reward_std": 0.25043364986777306,
      "rewards/accuracy_reward": 0.4270833358168602,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4375,
      "step": 51
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 859.1458587646484,
      "epoch": 0.9952153110047847,
      "grad_norm": 0.07899639612660056,
      "kl": 0.004245758056640625,
      "learning_rate": 6.331267252077352e-07,
      "loss": 0.0249,
      "reward": 0.8576388955116272,
      "reward_std": 0.25043364241719246,
      "rewards/accuracy_reward": 0.423611119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4340277761220932,
      "step": 52
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 818.5764007568359,
      "epoch": 1.0191387559808613,
      "grad_norm": 0.06800368909135272,
      "kl": 0.0054168701171875,
      "learning_rate": 6.181424998770595e-07,
      "loss": 0.0099,
      "reward": 0.9149305522441864,
      "reward_std": 0.25534411147236824,
      "rewards/accuracy_reward": 0.444444440305233,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4704861044883728,
      "step": 53
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 825.7847290039062,
      "epoch": 1.0382775119617225,
      "grad_norm": 0.0687845544933541,
      "kl": 0.00484466552734375,
      "learning_rate": 6.030805227178808e-07,
      "loss": 0.0266,
      "reward": 0.8932291716337204,
      "reward_std": 0.27130311727523804,
      "rewards/accuracy_reward": 0.4236111119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.469618059694767,
      "step": 54
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 831.6875152587891,
      "epoch": 1.0574162679425838,
      "grad_norm": 0.06231563710215865,
      "kl": 0.0048065185546875,
      "learning_rate": 5.879579797193701e-07,
      "loss": 0.0051,
      "reward": 0.8385416716337204,
      "reward_std": 0.23324701189994812,
      "rewards/accuracy_reward": 0.3958333358168602,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4427083358168602,
      "step": 55
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 813.4722290039062,
      "epoch": 1.076555023923445,
      "grad_norm": 0.07317349080061789,
      "kl": 0.0054473876953125,
      "learning_rate": 5.727921259774208e-07,
      "loss": 0.0342,
      "reward": 0.8819444626569748,
      "reward_std": 0.25043365359306335,
      "rewards/accuracy_reward": 0.4236111268401146,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4583333432674408,
      "step": 56
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 802.3784942626953,
      "epoch": 1.0956937799043063,
      "grad_norm": 0.07553172071799115,
      "kl": 0.00548553466796875,
      "learning_rate": 5.57600266006269e-07,
      "loss": 0.0246,
      "reward": 0.9357638955116272,
      "reward_std": 0.2970830574631691,
      "rewards/accuracy_reward": 0.4583333432674408,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4774305671453476,
      "step": 57
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 809.1076354980469,
      "epoch": 1.1148325358851674,
      "grad_norm": 0.08337759523447522,
      "kl": 0.0051727294921875,
      "learning_rate": 5.423997339937311e-07,
      "loss": 0.0237,
      "reward": 0.8854167014360428,
      "reward_std": 0.3019935190677643,
      "rewards/accuracy_reward": 0.4097222238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4756944477558136,
      "step": 58
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 813.4965362548828,
      "epoch": 1.1339712918660287,
      "grad_norm": 0.08335533132880214,
      "kl": 0.006591796875,
      "learning_rate": 5.272078740225792e-07,
      "loss": 0.025,
      "reward": 0.8862847238779068,
      "reward_std": 0.26148218661546707,
      "rewards/accuracy_reward": 0.4062500149011612,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4800347238779068,
      "step": 59
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 799.8507232666016,
      "epoch": 1.1531100478468899,
      "grad_norm": 0.07203309192614081,
      "kl": 0.0055999755859375,
      "learning_rate": 5.1204202028063e-07,
      "loss": 0.0212,
      "reward": 0.9774305373430252,
      "reward_std": 0.2627098001539707,
      "rewards/accuracy_reward": 0.4930555745959282,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4843750149011612,
      "step": 60
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 791.4027862548828,
      "epoch": 1.1722488038277512,
      "grad_norm": 0.07531690542769269,
      "kl": 0.00594329833984375,
      "learning_rate": 4.969194772821193e-07,
      "loss": 0.0127,
      "reward": 0.981770858168602,
      "reward_std": 0.26884788647294044,
      "rewards/accuracy_reward": 0.486111119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4956597313284874,
      "step": 61
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 856.4513854980469,
      "epoch": 1.1913875598086126,
      "grad_norm": 0.062172711166539586,
      "kl": 0.00542449951171875,
      "learning_rate": 4.818575001229406e-07,
      "loss": 0.0132,
      "reward": 0.8255208432674408,
      "reward_std": 0.20255662500858307,
      "rewards/accuracy_reward": 0.3819444477558136,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4435763955116272,
      "step": 62
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 865.0312652587891,
      "epoch": 1.2105263157894737,
      "grad_norm": 0.056253844990592314,
      "kl": 0.00551605224609375,
      "learning_rate": 4.668732747922649e-07,
      "loss": 0.0133,
      "reward": 0.767361119389534,
      "reward_std": 0.19641854986548424,
      "rewards/accuracy_reward": 0.3472222313284874,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4201389029622078,
      "step": 63
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 828.03125,
      "epoch": 1.229665071770335,
      "grad_norm": 0.06432066728138643,
      "kl": 0.005279541015625,
      "learning_rate": 4.5198389856301677e-07,
      "loss": 0.0216,
      "reward": 0.8559027761220932,
      "reward_std": 0.23324701562523842,
      "rewards/accuracy_reward": 0.4027777835726738,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4531250074505806,
      "step": 64
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 827.1354217529297,
      "epoch": 1.2488038277511961,
      "grad_norm": 0.06953232592329361,
      "kl": 0.005645751953125,
      "learning_rate": 4.3720636048357573e-07,
      "loss": 0.0207,
      "reward": 0.8515625,
      "reward_std": 0.20992231741547585,
      "rewards/accuracy_reward": 0.399305559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4522569477558136,
      "step": 65
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 830.3020782470703,
      "epoch": 1.2679425837320575,
      "grad_norm": 0.06833410523166436,
      "kl": 0.00586700439453125,
      "learning_rate": 4.225575219929631e-07,
      "loss": 0.0255,
      "reward": 0.8984375149011612,
      "reward_std": 0.29340019077062607,
      "rewards/accuracy_reward": 0.416666679084301,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4817708283662796,
      "step": 66
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 837.3264007568359,
      "epoch": 1.2870813397129186,
      "grad_norm": 0.06984826024421593,
      "kl": 0.00540924072265625,
      "learning_rate": 4.0805409768163504e-07,
      "loss": 0.0133,
      "reward": 0.864583358168602,
      "reward_std": 0.28726211935281754,
      "rewards/accuracy_reward": 0.4027777910232544,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4618055522441864,
      "step": 67
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 813.1770782470703,
      "epoch": 1.30622009569378,
      "grad_norm": 0.07284090789531786,
      "kl": 0.00634765625,
      "learning_rate": 3.9371263621983097e-07,
      "loss": 0.0162,
      "reward": 0.8993055671453476,
      "reward_std": 0.27744119614362717,
      "rewards/accuracy_reward": 0.4270833358168602,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4722222313284874,
      "step": 68
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 816.5382080078125,
      "epoch": 1.325358851674641,
      "grad_norm": 0.07703787035943734,
      "kl": 0.0065155029296875,
      "learning_rate": 3.795495014752408e-07,
      "loss": 0.0256,
      "reward": 0.956597238779068,
      "reward_std": 0.3044487535953522,
      "rewards/accuracy_reward": 0.461805559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4947916641831398,
      "step": 69
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 803.5173645019531,
      "epoch": 1.3444976076555024,
      "grad_norm": 0.07147321350786168,
      "kl": 0.0061187744140625,
      "learning_rate": 3.655808538415357e-07,
      "loss": 0.026,
      "reward": 0.9496528059244156,
      "reward_std": 0.2700754851102829,
      "rewards/accuracy_reward": 0.4687500074505806,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4809027761220932,
      "step": 70
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 849.3402709960938,
      "epoch": 1.3636363636363638,
      "grad_norm": 0.062268739561404315,
      "kl": 0.00612640380859375,
      "learning_rate": 3.5182263179906457e-07,
      "loss": 0.0233,
      "reward": 0.8515624850988388,
      "reward_std": 0.23692986741662025,
      "rewards/accuracy_reward": 0.3993055671453476,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4522569477558136,
      "step": 71
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 827.8125,
      "epoch": 1.3827751196172249,
      "grad_norm": 0.07644202407601894,
      "kl": 0.00571441650390625,
      "learning_rate": 3.3829053372876006e-07,
      "loss": 0.029,
      "reward": 0.8593750149011612,
      "reward_std": 0.2700754962861538,
      "rewards/accuracy_reward": 0.3888888955116272,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4704861119389534,
      "step": 72
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 824.4514007568359,
      "epoch": 1.401913875598086,
      "grad_norm": 0.07665948925496174,
      "kl": 0.0063323974609375,
      "learning_rate": 3.250000000000001e-07,
      "loss": 0.025,
      "reward": 0.9288194477558136,
      "reward_std": 0.2995382845401764,
      "rewards/accuracy_reward": 0.4375,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4913194477558136,
      "step": 73
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 798.7048645019531,
      "epoch": 1.4210526315789473,
      "grad_norm": 0.06940993863431441,
      "kl": 0.0066375732421875,
      "learning_rate": 3.1196619535286707e-07,
      "loss": 0.0166,
      "reward": 0.9149305522441864,
      "reward_std": 0.21360516920685768,
      "rewards/accuracy_reward": 0.423611119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.491319440305233,
      "step": 74
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 828.0729217529297,
      "epoch": 1.4401913875598087,
      "grad_norm": 0.07296376411015054,
      "kl": 0.006256103515625,
      "learning_rate": 2.992039915949036e-07,
      "loss": 0.0193,
      "reward": 0.9097222238779068,
      "reward_std": 0.24552317708730698,
      "rewards/accuracy_reward": 0.430555559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4791666716337204,
      "step": 75
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 816.3090362548828,
      "epoch": 1.4593301435406698,
      "grad_norm": 0.07617903452488152,
      "kl": 0.0065765380859375,
      "learning_rate": 2.867279506321115e-07,
      "loss": 0.017,
      "reward": 0.8394097238779068,
      "reward_std": 0.2835792824625969,
      "rewards/accuracy_reward": 0.3715277835726738,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4678819477558136,
      "step": 76
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 841.1701507568359,
      "epoch": 1.4784688995215312,
      "grad_norm": 0.0675783715098431,
      "kl": 0.0067291259765625,
      "learning_rate": 2.7455230785355164e-07,
      "loss": 0.0299,
      "reward": 0.8255208432674408,
      "reward_std": 0.20010139793157578,
      "rewards/accuracy_reward": 0.3750000074505806,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4505208358168602,
      "step": 77
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 863.5416717529297,
      "epoch": 1.4976076555023923,
      "grad_norm": 0.06878244801121526,
      "kl": 0.00676727294921875,
      "learning_rate": 2.6269095588850744e-07,
      "loss": 0.0208,
      "reward": 0.8307291865348816,
      "reward_std": 0.26148219406604767,
      "rewards/accuracy_reward": 0.378472238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4522569552063942,
      "step": 78
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 827.4583435058594,
      "epoch": 1.5167464114832536,
      "grad_norm": 0.07251844117312602,
      "kl": 0.007293701171875,
      "learning_rate": 2.511574287547423e-07,
      "loss": 0.0209,
      "reward": 0.884548619389534,
      "reward_std": 0.281124047935009,
      "rewards/accuracy_reward": 0.4131944552063942,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4713541716337204,
      "step": 79
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 815.7916870117188,
      "epoch": 1.535885167464115,
      "grad_norm": 0.06598452519685939,
      "kl": 0.00701141357421875,
      "learning_rate": 2.39964886415941e-07,
      "loss": 0.0212,
      "reward": 0.862847238779068,
      "reward_std": 0.25043364614248276,
      "rewards/accuracy_reward": 0.399305559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4635416641831398,
      "step": 80
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 809.65625,
      "epoch": 1.555023923444976,
      "grad_norm": 0.07534072488970928,
      "kl": 0.0063323974609375,
      "learning_rate": 2.291260997659509e-07,
      "loss": 0.0184,
      "reward": 0.9218750149011612,
      "reward_std": 0.28480689227581024,
      "rewards/accuracy_reward": 0.4340277761220932,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4878472164273262,
      "step": 81
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 849.3958282470703,
      "epoch": 1.5741626794258372,
      "grad_norm": 0.06432192299979189,
      "kl": 0.0068206787109375,
      "learning_rate": 2.1865343605696128e-07,
      "loss": 0.0169,
      "reward": 0.8359375149011612,
      "reward_std": 0.21728801727294922,
      "rewards/accuracy_reward": 0.3645833358168602,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4713541641831398,
      "step": 82
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 839.3923492431641,
      "epoch": 1.5933014354066986,
      "grad_norm": 0.0750282166543299,
      "kl": 0.00699615478515625,
      "learning_rate": 2.085588447882441e-07,
      "loss": 0.0295,
      "reward": 0.8680555671453476,
      "reward_std": 0.30690397322177887,
      "rewards/accuracy_reward": 0.3923611119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.475694440305233,
      "step": 83
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 862.53125,
      "epoch": 1.61244019138756,
      "grad_norm": 0.0705730634901919,
      "kl": 0.00653839111328125,
      "learning_rate": 1.9885384407155872e-07,
      "loss": 0.0264,
      "reward": 0.8081597238779068,
      "reward_std": 0.2737583592534065,
      "rewards/accuracy_reward": 0.3611111119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4470486119389534,
      "step": 84
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 832.2951354980469,
      "epoch": 1.631578947368421,
      "grad_norm": 0.07373687083533499,
      "kl": 0.00716400146484375,
      "learning_rate": 1.89549507488777e-07,
      "loss": 0.018,
      "reward": 0.9184027761220932,
      "reward_std": 0.22833656147122383,
      "rewards/accuracy_reward": 0.4305555671453476,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4878472238779068,
      "step": 85
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 803.6944580078125,
      "epoch": 1.6507177033492821,
      "grad_norm": 0.06861588416484331,
      "kl": 0.007843017578125,
      "learning_rate": 1.8065645145672575e-07,
      "loss": 0.0224,
      "reward": 0.9375000149011612,
      "reward_std": 0.24552318453788757,
      "rewards/accuracy_reward": 0.4513888955116272,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4861111119389534,
      "step": 86
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 832.5069580078125,
      "epoch": 1.6698564593301435,
      "grad_norm": 0.06855680549432348,
      "kl": 0.007110595703125,
      "learning_rate": 1.7218482311366245e-07,
      "loss": 0.0183,
      "reward": 0.8836805522441864,
      "reward_std": 0.2602545693516731,
      "rewards/accuracy_reward": 0.4131944477558136,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.470486119389534,
      "step": 87
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 819.6875152587891,
      "epoch": 1.6889952153110048,
      "grad_norm": 0.06513799373796432,
      "kl": 0.00678253173828125,
      "learning_rate": 1.64144288741207e-07,
      "loss": 0.027,
      "reward": 0.9244791865348816,
      "reward_std": 0.2590269595384598,
      "rewards/accuracy_reward": 0.447916679084301,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4765625,
      "step": 88
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 807.8333435058594,
      "epoch": 1.7081339712918662,
      "grad_norm": 0.06838857628608602,
      "kl": 0.00734710693359375,
      "learning_rate": 1.5654402273493807e-07,
      "loss": 0.0204,
      "reward": 0.9175347238779068,
      "reward_std": 0.2516612708568573,
      "rewards/accuracy_reward": 0.4409722238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4765624925494194,
      "step": 89
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 830.8576507568359,
      "epoch": 1.7272727272727273,
      "grad_norm": 0.06707500592677755,
      "kl": 0.00732421875,
      "learning_rate": 1.4939269713624181e-07,
      "loss": 0.0171,
      "reward": 0.9131944626569748,
      "reward_std": 0.22833655402064323,
      "rewards/accuracy_reward": 0.4409722313284874,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4722222238779068,
      "step": 90
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 805.3263854980469,
      "epoch": 1.7464114832535884,
      "grad_norm": 0.08133209124657947,
      "kl": 0.007537841796875,
      "learning_rate": 1.4269847173735543e-07,
      "loss": 0.0284,
      "reward": 0.942708358168602,
      "reward_std": 0.27989643067121506,
      "rewards/accuracy_reward": 0.4409722238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.5017361044883728,
      "step": 91
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 833.0833282470703,
      "epoch": 1.7655502392344498,
      "grad_norm": 0.06350369442926171,
      "kl": 0.00672149658203125,
      "learning_rate": 1.3646898477089625e-07,
      "loss": 0.0197,
      "reward": 0.8750000149011612,
      "reward_std": 0.24306794628500938,
      "rewards/accuracy_reward": 0.4131944477558136,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.461805559694767,
      "step": 92
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 840.3715362548828,
      "epoch": 1.784688995215311,
      "grad_norm": 0.07310106942013637,
      "kl": 0.00698089599609375,
      "learning_rate": 1.307113441944992e-07,
      "loss": 0.0204,
      "reward": 0.8324652761220932,
      "reward_std": 0.26639265939593315,
      "rewards/accuracy_reward": 0.3750000074505806,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4574652835726738,
      "step": 93
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 798.3472290039062,
      "epoch": 1.8038277511961722,
      "grad_norm": 0.07470777769552865,
      "kl": 0.0069580078125,
      "learning_rate": 1.2543211958050894e-07,
      "loss": 0.0225,
      "reward": 0.9123263955116272,
      "reward_std": 0.26884787529706955,
      "rewards/accuracy_reward": 0.4409722238779068,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4713541641831398,
      "step": 94
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 809.6944580078125,
      "epoch": 1.8229665071770333,
      "grad_norm": 0.06571441072063877,
      "kl": 0.007171630859375,
      "learning_rate": 1.2063733461997803e-07,
      "loss": 0.0208,
      "reward": 0.9305555671453476,
      "reward_std": 0.21606039628386497,
      "rewards/accuracy_reward": 0.447916679084301,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4826388955116272,
      "step": 95
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 814.4757080078125,
      "epoch": 1.8421052631578947,
      "grad_norm": 0.07474184052328098,
      "kl": 0.0072021484375,
      "learning_rate": 1.1633246024952662e-07,
      "loss": 0.0291,
      "reward": 0.8836805522441864,
      "reward_std": 0.2602545730769634,
      "rewards/accuracy_reward": 0.40625,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.477430559694767,
      "step": 96
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 817.5972290039062,
      "epoch": 1.861244019138756,
      "grad_norm": 0.08343014010656753,
      "kl": 0.00762939453125,
      "learning_rate": 1.1252240840890425e-07,
      "loss": 0.0248,
      "reward": 0.9192708283662796,
      "reward_std": 0.27130311727523804,
      "rewards/accuracy_reward": 0.4375000149011612,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4817708358168602,
      "step": 97
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 812.0278015136719,
      "epoch": 1.8803827751196174,
      "grad_norm": 0.07184568208672217,
      "kl": 0.00701904296875,
      "learning_rate": 1.0921152643637749e-07,
      "loss": 0.0289,
      "reward": 0.90625,
      "reward_std": 0.2725307270884514,
      "rewards/accuracy_reward": 0.4305555522441864,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4756944477558136,
      "step": 98
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 780.3958435058594,
      "epoch": 1.8995215311004785,
      "grad_norm": 0.06820176876725473,
      "kl": 0.0081787109375,
      "learning_rate": 1.064035921083383e-07,
      "loss": 0.0215,
      "reward": 0.9201388955116272,
      "reward_std": 0.22588132694363594,
      "rewards/accuracy_reward": 0.430555559694767,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4895833432674408,
      "step": 99
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 812.1562652587891,
      "epoch": 1.9186602870813396,
      "grad_norm": 0.07452483871871766,
      "kl": 0.00782012939453125,
      "learning_rate": 1.0410180932879252e-07,
      "loss": 0.0291,
      "reward": 0.9079861044883728,
      "reward_std": 0.31181444972753525,
      "rewards/accuracy_reward": 0.4201388880610466,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4878472313284874,
      "step": 100
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 827.5972442626953,
      "epoch": 1.937799043062201,
      "grad_norm": 0.06920787242163999,
      "kl": 0.00714874267578125,
      "learning_rate": 1.0230880447364718e-07,
      "loss": 0.0181,
      "reward": 0.8758680671453476,
      "reward_std": 0.2762135788798332,
      "rewards/accuracy_reward": 0.4027777835726738,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4730902835726738,
      "step": 101
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 848.9201354980469,
      "epoch": 1.9569377990430623,
      "grad_norm": 0.06359574498753441,
      "kl": 0.00762939453125,
      "learning_rate": 1.0102662339396817e-07,
      "loss": 0.0237,
      "reward": 0.8289930522441864,
      "reward_std": 0.23938510566949844,
      "rewards/accuracy_reward": 0.3611111119389534,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4678819552063942,
      "step": 102
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 827.0173645019531,
      "epoch": 1.9760765550239234,
      "grad_norm": 0.06784856725869416,
      "kl": 0.00757598876953125,
      "learning_rate": 1.002567290816269e-07,
      "loss": 0.0233,
      "reward": 0.9071180671453476,
      "reward_std": 0.2565717250108719,
      "rewards/accuracy_reward": 0.4340277761220932,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4730902835726738,
      "step": 103
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 824.1666870117188,
      "epoch": 1.9952153110047846,
      "grad_norm": 0.07003424025971802,
      "kl": 0.00720977783203125,
      "learning_rate": 1e-07,
      "loss": 0.0232,
      "reward": 0.8446180522441864,
      "reward_std": 0.24920602142810822,
      "rewards/accuracy_reward": 0.3784722313284874,
      "rewards/format_reward": 0.0,
      "rewards/tag_count_reward": 0.4661458358168602,
      "step": 104
    },
    {
      "epoch": 1.9952153110047846,
      "step": 104,
      "total_flos": 0.0,
      "train_loss": 0.02088056951134394,
      "train_runtime": 4550.0212,
      "train_samples_per_second": 3.297,
      "train_steps_per_second": 0.023
    }
  ],
  "logging_steps": 1,
  "max_steps": 104,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 10,
  "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": 24,
  "trial_name": null,
  "trial_params": null
}