{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9965010496850945,
  "eval_steps": 100,
  "global_step": 89,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 719.0248603820801,
      "epoch": 0.01119664100769769,
      "grad_norm": 5.308396339416504,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7022594679147005,
      "reward_std": 0.11167733068577945,
      "rewards/accuracy_reward": 0.4343112222850323,
      "rewards/semantic_entropy_math_reward": 0.7022594679147005,
      "rewards/total_entropy_reward": 1.2448364309966564,
      "step": 1
    },
    {
      "completion_length": 697.6454010009766,
      "epoch": 0.02239328201539538,
      "grad_norm": 3.310378313064575,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6712827906012535,
      "reward_std": 0.1011071486864239,
      "rewards/accuracy_reward": 0.352040808647871,
      "rewards/semantic_entropy_math_reward": 0.6712827868759632,
      "rewards/total_entropy_reward": 1.305861696600914,
      "step": 2
    },
    {
      "completion_length": 701.3826370239258,
      "epoch": 0.03358992302309307,
      "grad_norm": 4.717713356018066,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7075437195599079,
      "reward_std": 0.07735519809648395,
      "rewards/accuracy_reward": 0.4126275437884033,
      "rewards/semantic_entropy_math_reward": 0.7075437270104885,
      "rewards/total_entropy_reward": 1.2551036067306995,
      "step": 3
    },
    {
      "completion_length": 692.5803489685059,
      "epoch": 0.04478656403079076,
      "grad_norm": 4.699649810791016,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6926020309329033,
      "reward_std": 0.09249642631039023,
      "rewards/accuracy_reward": 0.4196428433060646,
      "rewards/semantic_entropy_math_reward": 0.6926020495593548,
      "rewards/total_entropy_reward": 1.2794943004846573,
      "step": 4
    },
    {
      "completion_length": 643.8430976867676,
      "epoch": 0.05598320503848846,
      "grad_norm": 7.118838310241699,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7232142686843872,
      "reward_std": 0.09167159674689174,
      "rewards/accuracy_reward": 0.41326529532670975,
      "rewards/semantic_entropy_math_reward": 0.7232142575085163,
      "rewards/total_entropy_reward": 1.2130939476191998,
      "step": 5
    },
    {
      "completion_length": 711.6734580993652,
      "epoch": 0.06717984604618614,
      "grad_norm": 4.829713344573975,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6625364199280739,
      "reward_std": 0.1073409270029515,
      "rewards/accuracy_reward": 0.37436223588883877,
      "rewards/semantic_entropy_math_reward": 0.6625364273786545,
      "rewards/total_entropy_reward": 1.320713147521019,
      "step": 6
    },
    {
      "completion_length": 676.0541915893555,
      "epoch": 0.07837648705388384,
      "grad_norm": 4.374125003814697,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7006195187568665,
      "reward_std": 0.09755392652004957,
      "rewards/accuracy_reward": 0.4247448882088065,
      "rewards/semantic_entropy_math_reward": 0.7006195187568665,
      "rewards/total_entropy_reward": 1.2528423443436623,
      "step": 7
    },
    {
      "completion_length": 712.5401573181152,
      "epoch": 0.08957312806158152,
      "grad_norm": 6.558979034423828,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.633381923660636,
      "reward_std": 0.10544528882019222,
      "rewards/accuracy_reward": 0.3463010173290968,
      "rewards/semantic_entropy_math_reward": 0.6333819553256035,
      "rewards/total_entropy_reward": 1.3834920637309551,
      "step": 8
    },
    {
      "completion_length": 711.6345539093018,
      "epoch": 0.10076976906927922,
      "grad_norm": 7.621776103973389,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6800291426479816,
      "reward_std": 0.10248321201652288,
      "rewards/accuracy_reward": 0.39604590833187103,
      "rewards/semantic_entropy_math_reward": 0.6800291538238525,
      "rewards/total_entropy_reward": 1.2968212738633156,
      "step": 9
    },
    {
      "completion_length": 674.1779174804688,
      "epoch": 0.11196641007697691,
      "grad_norm": 7.523331165313721,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6800291538238525,
      "reward_std": 0.11024063220247626,
      "rewards/accuracy_reward": 0.411352033726871,
      "rewards/semantic_entropy_math_reward": 0.6800291500985622,
      "rewards/total_entropy_reward": 1.2904038280248642,
      "step": 10
    },
    {
      "completion_length": 741.4017696380615,
      "epoch": 0.1231630510846746,
      "grad_norm": 6.336475372314453,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6415816266089678,
      "reward_std": 0.10356245189905167,
      "rewards/accuracy_reward": 0.3858418306335807,
      "rewards/semantic_entropy_math_reward": 0.6415816303342581,
      "rewards/total_entropy_reward": 1.363376997411251,
      "step": 11
    },
    {
      "completion_length": 692.8437347412109,
      "epoch": 0.13435969209237228,
      "grad_norm": 7.185462474822998,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7002550885081291,
      "reward_std": 0.10777880600653589,
      "rewards/accuracy_reward": 0.43048468697816133,
      "rewards/semantic_entropy_math_reward": 0.7002551108598709,
      "rewards/total_entropy_reward": 1.2493381686508656,
      "step": 12
    },
    {
      "completion_length": 704.6638870239258,
      "epoch": 0.14555633310007,
      "grad_norm": 17.740726470947266,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.673833817243576,
      "reward_std": 0.09210015833377838,
      "rewards/accuracy_reward": 0.43558672815561295,
      "rewards/semantic_entropy_math_reward": 0.6738338358700275,
      "rewards/total_entropy_reward": 1.3185552880167961,
      "step": 13
    },
    {
      "completion_length": 679.9374847412109,
      "epoch": 0.15675297410776767,
      "grad_norm": 9.5872802734375,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7075437121093273,
      "reward_std": 0.09575178450904787,
      "rewards/accuracy_reward": 0.40624999068677425,
      "rewards/semantic_entropy_math_reward": 0.7075437419116497,
      "rewards/total_entropy_reward": 1.237540539354086,
      "step": 14
    },
    {
      "completion_length": 731.2404136657715,
      "epoch": 0.16794961511546536,
      "grad_norm": 4.687215805053711,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6375728864222765,
      "reward_std": 0.108485147356987,
      "rewards/accuracy_reward": 0.3762755021452904,
      "rewards/semantic_entropy_math_reward": 0.6375729013234377,
      "rewards/total_entropy_reward": 1.3838917911052704,
      "step": 15
    },
    {
      "completion_length": 700.0522842407227,
      "epoch": 0.17914625612316304,
      "grad_norm": 8.418819427490234,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6547011584043503,
      "reward_std": 0.10533164534717798,
      "rewards/accuracy_reward": 0.40624999161809683,
      "rewards/semantic_entropy_math_reward": 0.6547011733055115,
      "rewards/total_entropy_reward": 1.3433180004358292,
      "step": 16
    },
    {
      "completion_length": 684.034423828125,
      "epoch": 0.19034289713086075,
      "grad_norm": 7.032946586608887,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.664905235171318,
      "reward_std": 0.0974241562653333,
      "rewards/accuracy_reward": 0.3832908095791936,
      "rewards/semantic_entropy_math_reward": 0.6649052500724792,
      "rewards/total_entropy_reward": 1.3396263718605042,
      "step": 17
    },
    {
      "completion_length": 677.9177227020264,
      "epoch": 0.20153953813855843,
      "grad_norm": 15.37868595123291,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6975218504667282,
      "reward_std": 0.09150373586453497,
      "rewards/accuracy_reward": 0.45153060369193554,
      "rewards/semantic_entropy_math_reward": 0.69752187281847,
      "rewards/total_entropy_reward": 1.262941613793373,
      "step": 18
    },
    {
      "completion_length": 683.5771484375,
      "epoch": 0.21273617914625612,
      "grad_norm": 36.41266632080078,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7057215645909309,
      "reward_std": 0.09910683194175363,
      "rewards/accuracy_reward": 0.4368622377514839,
      "rewards/semantic_entropy_math_reward": 0.7057215794920921,
      "rewards/total_entropy_reward": 1.2459207847714424,
      "step": 19
    },
    {
      "completion_length": 688.8073768615723,
      "epoch": 0.22393282015395383,
      "grad_norm": 27.405065536499023,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7097303122282028,
      "reward_std": 0.09203298413194716,
      "rewards/accuracy_reward": 0.4598214225843549,
      "rewards/semantic_entropy_math_reward": 0.7097303308546543,
      "rewards/total_entropy_reward": 1.2417229264974594,
      "step": 20
    },
    {
      "completion_length": 736.3092956542969,
      "epoch": 0.2351294611616515,
      "grad_norm": 19.091941833496094,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6541544962674379,
      "reward_std": 0.0986680502537638,
      "rewards/accuracy_reward": 0.4228316266089678,
      "rewards/semantic_entropy_math_reward": 0.6541545186191797,
      "rewards/total_entropy_reward": 1.3527532257139683,
      "step": 21
    },
    {
      "completion_length": 709.4795722961426,
      "epoch": 0.2463261021693492,
      "grad_norm": 14.640216827392578,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6836734637618065,
      "reward_std": 0.10867427196353674,
      "rewards/accuracy_reward": 0.46301019564270973,
      "rewards/semantic_entropy_math_reward": 0.6836734749376774,
      "rewards/total_entropy_reward": 1.2770788073539734,
      "step": 22
    },
    {
      "completion_length": 681.6218032836914,
      "epoch": 0.2575227431770469,
      "grad_norm": 13.511483192443848,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6967929936945438,
      "reward_std": 0.09672949556261301,
      "rewards/accuracy_reward": 0.44579080305993557,
      "rewards/semantic_entropy_math_reward": 0.6967929899692535,
      "rewards/total_entropy_reward": 1.2772413976490498,
      "step": 23
    },
    {
      "completion_length": 705.6811141967773,
      "epoch": 0.26871938418474456,
      "grad_norm": 15.547083854675293,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6461370252072811,
      "reward_std": 0.1031385101377964,
      "rewards/accuracy_reward": 0.3985969265922904,
      "rewards/semantic_entropy_math_reward": 0.6461370065808296,
      "rewards/total_entropy_reward": 1.3622385039925575,
      "step": 24
    },
    {
      "completion_length": 731.4566116333008,
      "epoch": 0.27991602519244224,
      "grad_norm": 8.434446334838867,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6233600564301014,
      "reward_std": 0.08827707398450002,
      "rewards/accuracy_reward": 0.3494897885248065,
      "rewards/semantic_entropy_math_reward": 0.6233600713312626,
      "rewards/total_entropy_reward": 1.4261119738221169,
      "step": 25
    },
    {
      "completion_length": 685.5197486877441,
      "epoch": 0.29111266620014,
      "grad_norm": 12.586360931396484,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6765670366585255,
      "reward_std": 0.1049469755962491,
      "rewards/accuracy_reward": 0.412627543322742,
      "rewards/semantic_entropy_math_reward": 0.6765670329332352,
      "rewards/total_entropy_reward": 1.3074346259236336,
      "step": 26
    },
    {
      "completion_length": 711.061840057373,
      "epoch": 0.30230930720783766,
      "grad_norm": 9.39128303527832,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6521501429378986,
      "reward_std": 0.1099298931658268,
      "rewards/accuracy_reward": 0.40624999068677425,
      "rewards/semantic_entropy_math_reward": 0.6521501652896404,
      "rewards/total_entropy_reward": 1.3500806987285614,
      "step": 27
    },
    {
      "completion_length": 707.5612106323242,
      "epoch": 0.31350594821553535,
      "grad_norm": 8.279614448547363,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6692784316837788,
      "reward_std": 0.10041180578991771,
      "rewards/accuracy_reward": 0.36096938233822584,
      "rewards/semantic_entropy_math_reward": 0.6692784205079079,
      "rewards/total_entropy_reward": 1.3163355849683285,
      "step": 28
    },
    {
      "completion_length": 675.4157962799072,
      "epoch": 0.32470258922323303,
      "grad_norm": 13.983796119689941,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7091836705803871,
      "reward_std": 0.09643913782201707,
      "rewards/accuracy_reward": 0.478316318243742,
      "rewards/semantic_entropy_math_reward": 0.7091836743056774,
      "rewards/total_entropy_reward": 1.2364819720387459,
      "step": 29
    },
    {
      "completion_length": 723.9189872741699,
      "epoch": 0.3358992302309307,
      "grad_norm": 13.94961166381836,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.614431481808424,
      "reward_std": 0.12222768180072308,
      "rewards/accuracy_reward": 0.3533163210377097,
      "rewards/semantic_entropy_math_reward": 0.614431481808424,
      "rewards/total_entropy_reward": 1.4096959978342056,
      "step": 30
    },
    {
      "completion_length": 721.4604339599609,
      "epoch": 0.3470958712386284,
      "grad_norm": 7.932522773742676,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6603498533368111,
      "reward_std": 0.09537219302728772,
      "rewards/accuracy_reward": 0.3992346851155162,
      "rewards/semantic_entropy_math_reward": 0.6603498607873917,
      "rewards/total_entropy_reward": 1.3325160779058933,
      "step": 31
    },
    {
      "completion_length": 702.3788185119629,
      "epoch": 0.3582925122463261,
      "grad_norm": 9.096883773803711,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6829446069896221,
      "reward_std": 0.09760210802778602,
      "rewards/accuracy_reward": 0.40816325694322586,
      "rewards/semantic_entropy_math_reward": 0.682944618165493,
      "rewards/total_entropy_reward": 1.303523451089859,
      "step": 32
    },
    {
      "completion_length": 700.5924644470215,
      "epoch": 0.3694891532540238,
      "grad_norm": 3.891953229904175,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7055393569171429,
      "reward_std": 0.10629667551256716,
      "rewards/accuracy_reward": 0.45982141979038715,
      "rewards/semantic_entropy_math_reward": 0.7055393569171429,
      "rewards/total_entropy_reward": 1.2391385585069656,
      "step": 33
    },
    {
      "completion_length": 716.7027950286865,
      "epoch": 0.3806857942617215,
      "grad_norm": 12.71116828918457,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6698250789195299,
      "reward_std": 0.09906701685395092,
      "rewards/accuracy_reward": 0.40369897056370974,
      "rewards/semantic_entropy_math_reward": 0.6698250807821751,
      "rewards/total_entropy_reward": 1.3271235637366772,
      "step": 34
    },
    {
      "completion_length": 769.0414352416992,
      "epoch": 0.3918824352694192,
      "grad_norm": 9.855749130249023,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6481414064764977,
      "reward_std": 0.10145364608615637,
      "rewards/accuracy_reward": 0.3743622349575162,
      "rewards/semantic_entropy_math_reward": 0.6481414288282394,
      "rewards/total_entropy_reward": 1.3565044924616814,
      "step": 35
    },
    {
      "completion_length": 722.349479675293,
      "epoch": 0.40307907627711687,
      "grad_norm": 5.131447792053223,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6638119556009769,
      "reward_std": 0.09085186710581183,
      "rewards/accuracy_reward": 0.4170918297022581,
      "rewards/semantic_entropy_math_reward": 0.6638119481503963,
      "rewards/total_entropy_reward": 1.350273534655571,
      "step": 36
    },
    {
      "completion_length": 703.8832778930664,
      "epoch": 0.41427571728481455,
      "grad_norm": 7.057519435882568,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.690597664564848,
      "reward_std": 0.09433076065033674,
      "rewards/accuracy_reward": 0.43686223309487104,
      "rewards/semantic_entropy_math_reward": 0.6905976496636868,
      "rewards/total_entropy_reward": 1.2801040560007095,
      "step": 37
    },
    {
      "completion_length": 725.2799606323242,
      "epoch": 0.42547235829251223,
      "grad_norm": 7.460778713226318,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6669096164405346,
      "reward_std": 0.10552376857958734,
      "rewards/accuracy_reward": 0.41326529532670975,
      "rewards/semantic_entropy_math_reward": 0.6669096313416958,
      "rewards/total_entropy_reward": 1.3323625773191452,
      "step": 38
    },
    {
      "completion_length": 725.7671966552734,
      "epoch": 0.4366689993002099,
      "grad_norm": 7.124528408050537,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6472303047776222,
      "reward_std": 0.10429512546397746,
      "rewards/accuracy_reward": 0.3909438718110323,
      "rewards/semantic_entropy_math_reward": 0.6472303159534931,
      "rewards/total_entropy_reward": 1.3546394035220146,
      "step": 39
    },
    {
      "completion_length": 697.7423439025879,
      "epoch": 0.44786564030790765,
      "grad_norm": 7.065954208374023,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7193877585232258,
      "reward_std": 0.09428266366012394,
      "rewards/accuracy_reward": 0.4336734600365162,
      "rewards/semantic_entropy_math_reward": 0.7193877547979355,
      "rewards/total_entropy_reward": 1.2143822945654392,
      "step": 40
    },
    {
      "completion_length": 703.3673400878906,
      "epoch": 0.45906228131560534,
      "grad_norm": 6.764792442321777,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.706632649526,
      "reward_std": 0.0942279752343893,
      "rewards/accuracy_reward": 0.4649234591051936,
      "rewards/semantic_entropy_math_reward": 0.7066326681524515,
      "rewards/total_entropy_reward": 1.238373503088951,
      "step": 41
    },
    {
      "completion_length": 651.4183578491211,
      "epoch": 0.470258922323303,
      "grad_norm": 16.47541618347168,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6813046522438526,
      "reward_std": 0.10622056131251156,
      "rewards/accuracy_reward": 0.43239795230329037,
      "rewards/semantic_entropy_math_reward": 0.681304682046175,
      "rewards/total_entropy_reward": 1.284136950969696,
      "step": 42
    },
    {
      "completion_length": 743.2646636962891,
      "epoch": 0.4814555633310007,
      "grad_norm": 15.174323081970215,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6525145824998617,
      "reward_std": 0.10325121926143765,
      "rewards/accuracy_reward": 0.3947704015299678,
      "rewards/semantic_entropy_math_reward": 0.652514586225152,
      "rewards/total_entropy_reward": 1.358573641628027,
      "step": 43
    },
    {
      "completion_length": 725.5962905883789,
      "epoch": 0.4926522043386984,
      "grad_norm": 7.421966075897217,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6867711264640093,
      "reward_std": 0.09842351987026632,
      "rewards/accuracy_reward": 0.4323979504406452,
      "rewards/semantic_entropy_math_reward": 0.6867711190134287,
      "rewards/total_entropy_reward": 1.2839920222759247,
      "step": 44
    },
    {
      "completion_length": 698.1390113830566,
      "epoch": 0.5038488453463961,
      "grad_norm": 9.836834907531738,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6851311810314655,
      "reward_std": 0.10096711455844343,
      "rewards/accuracy_reward": 0.44196427427232265,
      "rewards/semantic_entropy_math_reward": 0.6851312182843685,
      "rewards/total_entropy_reward": 1.290800966322422,
      "step": 45
    },
    {
      "completion_length": 737.8379898071289,
      "epoch": 0.5150454863540938,
      "grad_norm": 13.060182571411133,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6867711432278156,
      "reward_std": 0.10890168650075793,
      "rewards/accuracy_reward": 0.403061218559742,
      "rewards/semantic_entropy_math_reward": 0.6867711097002029,
      "rewards/total_entropy_reward": 1.2737670093774796,
      "step": 46
    },
    {
      "completion_length": 720.0937309265137,
      "epoch": 0.5262421273617914,
      "grad_norm": 17.424467086791992,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6268221419304609,
      "reward_std": 0.11281977966427803,
      "rewards/accuracy_reward": 0.39668366219848394,
      "rewards/semantic_entropy_math_reward": 0.6268221419304609,
      "rewards/total_entropy_reward": 1.396425575017929,
      "step": 47
    },
    {
      "completion_length": 730.9495983123779,
      "epoch": 0.5374387683694891,
      "grad_norm": 7.613661766052246,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6333818975836039,
      "reward_std": 0.09289166564121842,
      "rewards/accuracy_reward": 0.38201529905200005,
      "rewards/semantic_entropy_math_reward": 0.6333819199353456,
      "rewards/total_entropy_reward": 1.3891723416745663,
      "step": 48
    },
    {
      "completion_length": 730.4712867736816,
      "epoch": 0.5486354093771868,
      "grad_norm": 11.16511344909668,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.628097664564848,
      "reward_std": 0.10562112857587636,
      "rewards/accuracy_reward": 0.3871173355728388,
      "rewards/semantic_entropy_math_reward": 0.6280976608395576,
      "rewards/total_entropy_reward": 1.3984056264162064,
      "step": 49
    },
    {
      "completion_length": 700.4126129150391,
      "epoch": 0.5598320503848845,
      "grad_norm": 7.883020401000977,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6543367393314838,
      "reward_std": 0.08848634106107056,
      "rewards/accuracy_reward": 0.43494897056370974,
      "rewards/semantic_entropy_math_reward": 0.6543367449194193,
      "rewards/total_entropy_reward": 1.368666134774685,
      "step": 50
    },
    {
      "completion_length": 726.035701751709,
      "epoch": 0.5710286913925823,
      "grad_norm": 16.26348114013672,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6665451787412167,
      "reward_std": 0.09329186473041773,
      "rewards/accuracy_reward": 0.40114795323461294,
      "rewards/semantic_entropy_math_reward": 0.6665451973676682,
      "rewards/total_entropy_reward": 1.3389556668698788,
      "step": 51
    },
    {
      "completion_length": 719.875617980957,
      "epoch": 0.58222533240028,
      "grad_norm": 10.366020202636719,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6474125292152166,
      "reward_std": 0.10870802192948759,
      "rewards/accuracy_reward": 0.3934948956593871,
      "rewards/semantic_entropy_math_reward": 0.6474125385284424,
      "rewards/total_entropy_reward": 1.3537504002451897,
      "step": 52
    },
    {
      "completion_length": 691.538890838623,
      "epoch": 0.5934219734079776,
      "grad_norm": 10.94919204711914,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6860422715544701,
      "reward_std": 0.0946744061075151,
      "rewards/accuracy_reward": 0.489158159121871,
      "rewards/semantic_entropy_math_reward": 0.6860422790050507,
      "rewards/total_entropy_reward": 1.3007621616125107,
      "step": 53
    },
    {
      "completion_length": 705.672176361084,
      "epoch": 0.6046186144156753,
      "grad_norm": 23.8171329498291,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6762026119977236,
      "reward_std": 0.09615424065850675,
      "rewards/accuracy_reward": 0.39668366592377424,
      "rewards/semantic_entropy_math_reward": 0.6762026362121105,
      "rewards/total_entropy_reward": 1.3121707029640675,
      "step": 54
    },
    {
      "completion_length": 677.1664428710938,
      "epoch": 0.615815255423373,
      "grad_norm": 9.031086921691895,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6896865852177143,
      "reward_std": 0.1015757208224386,
      "rewards/accuracy_reward": 0.4317601965740323,
      "rewards/semantic_entropy_math_reward": 0.6896865479648113,
      "rewards/total_entropy_reward": 1.287308655679226,
      "step": 55
    },
    {
      "completion_length": 713.0114631652832,
      "epoch": 0.6270118964310707,
      "grad_norm": 10.871789932250977,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6656341031193733,
      "reward_std": 0.11424232507124543,
      "rewards/accuracy_reward": 0.4221938643604517,
      "rewards/semantic_entropy_math_reward": 0.6656341068446636,
      "rewards/total_entropy_reward": 1.3229641020298004,
      "step": 56
    },
    {
      "completion_length": 677.5694923400879,
      "epoch": 0.6382085374387684,
      "grad_norm": 8.904267311096191,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6729227248579264,
      "reward_std": 0.10780132608488202,
      "rewards/accuracy_reward": 0.4381377436220646,
      "rewards/semantic_entropy_math_reward": 0.6729227565228939,
      "rewards/total_entropy_reward": 1.3122014850378036,
      "step": 57
    },
    {
      "completion_length": 687.1345500946045,
      "epoch": 0.6494051784464661,
      "grad_norm": 19.945981979370117,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6621720213443041,
      "reward_std": 0.10787450219504535,
      "rewards/accuracy_reward": 0.4183673388324678,
      "rewards/semantic_entropy_math_reward": 0.6621720027178526,
      "rewards/total_entropy_reward": 1.3265771567821503,
      "step": 58
    },
    {
      "completion_length": 673.6224384307861,
      "epoch": 0.6606018194541637,
      "grad_norm": 25.872291564941406,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6900510117411613,
      "reward_std": 0.09328114939853549,
      "rewards/accuracy_reward": 0.3998724389821291,
      "rewards/semantic_entropy_math_reward": 0.6900510005652905,
      "rewards/total_entropy_reward": 1.2932880148291588,
      "step": 59
    },
    {
      "completion_length": 712.2027912139893,
      "epoch": 0.6717984604618614,
      "grad_norm": 12.222248077392578,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6426749210804701,
      "reward_std": 0.09614149574190378,
      "rewards/accuracy_reward": 0.41007652156986296,
      "rewards/semantic_entropy_math_reward": 0.6426749173551798,
      "rewards/total_entropy_reward": 1.383428543806076,
      "step": 60
    },
    {
      "completion_length": 739.6472969055176,
      "epoch": 0.6829951014695591,
      "grad_norm": 9.738767623901367,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6273687966167927,
      "reward_std": 0.11152111599221826,
      "rewards/accuracy_reward": 0.34885203186422586,
      "rewards/semantic_entropy_math_reward": 0.6273688077926636,
      "rewards/total_entropy_reward": 1.3971327617764473,
      "step": 61
    },
    {
      "completion_length": 685.2659378051758,
      "epoch": 0.6941917424772568,
      "grad_norm": 20.52920913696289,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6596209742128849,
      "reward_std": 0.10549976932816207,
      "rewards/accuracy_reward": 0.43622447922825813,
      "rewards/semantic_entropy_math_reward": 0.6596209909766912,
      "rewards/total_entropy_reward": 1.3266954682767391,
      "step": 62
    },
    {
      "completion_length": 684.1128730773926,
      "epoch": 0.7053883834849545,
      "grad_norm": 13.253375053405762,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6740160267800093,
      "reward_std": 0.11020976235158741,
      "rewards/accuracy_reward": 0.4419642761349678,
      "rewards/semantic_entropy_math_reward": 0.6740160342305899,
      "rewards/total_entropy_reward": 1.295092262327671,
      "step": 63
    },
    {
      "completion_length": 702.8647766113281,
      "epoch": 0.7165850244926522,
      "grad_norm": 11.436138153076172,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6612609177827835,
      "reward_std": 0.0923373675905168,
      "rewards/accuracy_reward": 0.43239795230329037,
      "rewards/semantic_entropy_math_reward": 0.661260936409235,
      "rewards/total_entropy_reward": 1.3441792502999306,
      "step": 64
    },
    {
      "completion_length": 713.3341636657715,
      "epoch": 0.72778166550035,
      "grad_norm": 15.771661758422852,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6313775349408388,
      "reward_std": 0.1141419094055891,
      "rewards/accuracy_reward": 0.4139030510559678,
      "rewards/semantic_entropy_math_reward": 0.6313775237649679,
      "rewards/total_entropy_reward": 1.3911906033754349,
      "step": 65
    },
    {
      "completion_length": 691.0420761108398,
      "epoch": 0.7389783065080476,
      "grad_norm": 9.556193351745605,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6576165966689587,
      "reward_std": 0.10697318986058235,
      "rewards/accuracy_reward": 0.3386479513719678,
      "rewards/semantic_entropy_math_reward": 0.657616626471281,
      "rewards/total_entropy_reward": 1.3421761691570282,
      "step": 66
    },
    {
      "completion_length": 688.1179618835449,
      "epoch": 0.7501749475157453,
      "grad_norm": 18.490732192993164,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6712827831506729,
      "reward_std": 0.10525703080929816,
      "rewards/accuracy_reward": 0.445790808647871,
      "rewards/semantic_entropy_math_reward": 0.6712828055024147,
      "rewards/total_entropy_reward": 1.3154872134327888,
      "step": 67
    },
    {
      "completion_length": 647.132007598877,
      "epoch": 0.761371588523443,
      "grad_norm": 7.94115686416626,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6920553781092167,
      "reward_std": 0.09936971426941454,
      "rewards/accuracy_reward": 0.43176019564270973,
      "rewards/semantic_entropy_math_reward": 0.6920554004609585,
      "rewards/total_entropy_reward": 1.2725396156311035,
      "step": 68
    },
    {
      "completion_length": 699.4005012512207,
      "epoch": 0.7725682295311407,
      "grad_norm": 12.146166801452637,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6652696616947651,
      "reward_std": 0.11806112038902938,
      "rewards/accuracy_reward": 0.409438768401742,
      "rewards/semantic_entropy_math_reward": 0.6652696691453457,
      "rewards/total_entropy_reward": 1.3024558648467064,
      "step": 69
    },
    {
      "completion_length": 667.9285659790039,
      "epoch": 0.7837648705388384,
      "grad_norm": 15.953444480895996,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.679846927523613,
      "reward_std": 0.09083303948864341,
      "rewards/accuracy_reward": 0.4311224427074194,
      "rewards/semantic_entropy_math_reward": 0.6798469312489033,
      "rewards/total_entropy_reward": 1.3062404170632362,
      "step": 70
    },
    {
      "completion_length": 705.4400367736816,
      "epoch": 0.794961511546536,
      "grad_norm": 12.320276260375977,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6556122340261936,
      "reward_std": 0.09324299031868577,
      "rewards/accuracy_reward": 0.37946427892893553,
      "rewards/semantic_entropy_math_reward": 0.6556122545152903,
      "rewards/total_entropy_reward": 1.3581575378775597,
      "step": 71
    },
    {
      "completion_length": 654.1511325836182,
      "epoch": 0.8061581525542337,
      "grad_norm": 22.0692138671875,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6794824972748756,
      "reward_std": 0.10328507493250072,
      "rewards/accuracy_reward": 0.4209183603525162,
      "rewards/semantic_entropy_math_reward": 0.6794825121760368,
      "rewards/total_entropy_reward": 1.2847715727984905,
      "step": 72
    },
    {
      "completion_length": 690.308666229248,
      "epoch": 0.8173547935619314,
      "grad_norm": 10.22040843963623,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6780247762799263,
      "reward_std": 0.105113809928298,
      "rewards/accuracy_reward": 0.45918366499245167,
      "rewards/semantic_entropy_math_reward": 0.678024772554636,
      "rewards/total_entropy_reward": 1.2880272567272186,
      "step": 73
    },
    {
      "completion_length": 696.0803337097168,
      "epoch": 0.8285514345696291,
      "grad_norm": 20.899675369262695,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6516034882515669,
      "reward_std": 0.10480827221181244,
      "rewards/accuracy_reward": 0.3852040730416775,
      "rewards/semantic_entropy_math_reward": 0.6516034882515669,
      "rewards/total_entropy_reward": 1.3282338082790375,
      "step": 74
    },
    {
      "completion_length": 649.8654270172119,
      "epoch": 0.8397480755773268,
      "grad_norm": 9.469245910644531,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7246719934046268,
      "reward_std": 0.08447014342527837,
      "rewards/accuracy_reward": 0.433035708963871,
      "rewards/semantic_entropy_math_reward": 0.7246720045804977,
      "rewards/total_entropy_reward": 1.220706269145012,
      "step": 75
    },
    {
      "completion_length": 746.2142791748047,
      "epoch": 0.8509447165850245,
      "grad_norm": 10.530217170715332,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.639395035803318,
      "reward_std": 0.10488821647595614,
      "rewards/accuracy_reward": 0.40816325321793556,
      "rewards/semantic_entropy_math_reward": 0.6393950544297695,
      "rewards/total_entropy_reward": 1.3760143592953682,
      "step": 76
    },
    {
      "completion_length": 655.6836585998535,
      "epoch": 0.8621413575927221,
      "grad_norm": 14.1907958984375,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.7228498421609402,
      "reward_std": 0.09376341942697763,
      "rewards/accuracy_reward": 0.4674744727090001,
      "rewards/semantic_entropy_math_reward": 0.7228498421609402,
      "rewards/total_entropy_reward": 1.2132117934525013,
      "step": 77
    },
    {
      "completion_length": 672.8137626647949,
      "epoch": 0.8733379986004198,
      "grad_norm": 13.5209379196167,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.652332354336977,
      "reward_std": 0.09460025187581778,
      "rewards/accuracy_reward": 0.3858418334275484,
      "rewards/semantic_entropy_math_reward": 0.6523323617875576,
      "rewards/total_entropy_reward": 1.351367250084877,
      "step": 78
    },
    {
      "completion_length": 696.9342880249023,
      "epoch": 0.8845346396081175,
      "grad_norm": 17.302839279174805,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6514212656766176,
      "reward_std": 0.09626807877793908,
      "rewards/accuracy_reward": 0.4062499897554517,
      "rewards/semantic_entropy_math_reward": 0.651421282440424,
      "rewards/total_entropy_reward": 1.366665042936802,
      "step": 79
    },
    {
      "completion_length": 682.3073806762695,
      "epoch": 0.8957312806158153,
      "grad_norm": 9.640515327453613,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6767492536455393,
      "reward_std": 0.10075846454128623,
      "rewards/accuracy_reward": 0.42028060276061296,
      "rewards/semantic_entropy_math_reward": 0.6767492648214102,
      "rewards/total_entropy_reward": 1.3104709684848785,
      "step": 80
    },
    {
      "completion_length": 722.4725646972656,
      "epoch": 0.906927921623513,
      "grad_norm": 6.948408603668213,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6521501280367374,
      "reward_std": 0.08798706252127886,
      "rewards/accuracy_reward": 0.4534438718110323,
      "rewards/semantic_entropy_math_reward": 0.6521501541137695,
      "rewards/total_entropy_reward": 1.3641655445098877,
      "step": 81
    },
    {
      "completion_length": 686.7238388061523,
      "epoch": 0.9181245626312107,
      "grad_norm": 47.02455139160156,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6811224482953548,
      "reward_std": 0.11016466980800033,
      "rewards/accuracy_reward": 0.42283162754029036,
      "rewards/semantic_entropy_math_reward": 0.6811224408447742,
      "rewards/total_entropy_reward": 1.2916913330554962,
      "step": 82
    },
    {
      "completion_length": 709.4598007202148,
      "epoch": 0.9293212036389084,
      "grad_norm": 31.659669876098633,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6086005661636591,
      "reward_std": 0.10625140275806189,
      "rewards/accuracy_reward": 0.37882652692496777,
      "rewards/semantic_entropy_math_reward": 0.60860057733953,
      "rewards/total_entropy_reward": 1.4536337479948997,
      "step": 83
    },
    {
      "completion_length": 661.408784866333,
      "epoch": 0.940517844646606,
      "grad_norm": 8.08792781829834,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6858600489795208,
      "reward_std": 0.09285477036610246,
      "rewards/accuracy_reward": 0.4221938652917743,
      "rewards/semantic_entropy_math_reward": 0.6858600452542305,
      "rewards/total_entropy_reward": 1.283115666359663,
      "step": 84
    },
    {
      "completion_length": 657.9400405883789,
      "epoch": 0.9517144856543037,
      "grad_norm": 21.801326751708984,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6805757954716682,
      "reward_std": 0.09728616522625089,
      "rewards/accuracy_reward": 0.43749999441206455,
      "rewards/semantic_entropy_math_reward": 0.6805758327245712,
      "rewards/total_entropy_reward": 1.3078671097755432,
      "step": 85
    },
    {
      "completion_length": 674.7142734527588,
      "epoch": 0.9629111266620014,
      "grad_norm": 11.559253692626953,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6862244829535484,
      "reward_std": 0.10050268471240997,
      "rewards/accuracy_reward": 0.4304846851155162,
      "rewards/semantic_entropy_math_reward": 0.686224490404129,
      "rewards/total_entropy_reward": 1.2905268669128418,
      "step": 86
    },
    {
      "completion_length": 636.0388870239258,
      "epoch": 0.9741077676696991,
      "grad_norm": 11.289789199829102,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6986151486635208,
      "reward_std": 0.10155197884887457,
      "rewards/accuracy_reward": 0.43686223216354847,
      "rewards/semantic_entropy_math_reward": 0.6986151374876499,
      "rewards/total_entropy_reward": 1.2592380531132221,
      "step": 87
    },
    {
      "completion_length": 683.5822525024414,
      "epoch": 0.9853044086773968,
      "grad_norm": 10.526566505432129,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6585276797413826,
      "reward_std": 0.10897616110742092,
      "rewards/accuracy_reward": 0.4017857098951936,
      "rewards/semantic_entropy_math_reward": 0.6585277020931244,
      "rewards/total_entropy_reward": 1.324688896536827,
      "step": 88
    },
    {
      "completion_length": 683.3411960601807,
      "epoch": 0.9965010496850945,
      "grad_norm": 19.354259490966797,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "reward": 0.6687317825853825,
      "reward_std": 0.09195416304282844,
      "rewards/accuracy_reward": 0.380102027207613,
      "rewards/semantic_entropy_math_reward": 0.6687317807227373,
      "rewards/total_entropy_reward": 1.3135743215680122,
      "step": 89
    },
    {
      "epoch": 0.9965010496850945,
      "step": 89,
      "total_flos": 0.0,
      "train_loss": 2.8008765204773544e-08,
      "train_runtime": 57265.8095,
      "train_samples_per_second": 0.349,
      "train_steps_per_second": 0.002
    }
  ],
  "logging_steps": 1,
  "max_steps": 89,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 20,
  "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": 2,
  "trial_name": null,
  "trial_params": null
}