{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9987438399845395,
  "eval_steps": 100,
  "global_step": 646,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 890.5919452667237,
      "epoch": 0.015460430959512996,
      "grad_norm": 0.0048114829798690025,
      "kl": 0.0004779815673828125,
      "learning_rate": 3.0769230769230774e-06,
      "loss": 0.0,
      "reward": 0.2079081600648351,
      "reward_std": 0.18389849485829474,
      "rewards/accuracy_reward": 0.2079081600648351,
      "rewards/format_reward": 0.0,
      "step": 10
    },
    {
      "completion_length": 815.0035549163819,
      "epoch": 0.03092086191902599,
      "grad_norm": 0.0037700873930097754,
      "kl": 0.005376839637756347,
      "learning_rate": 6.153846153846155e-06,
      "loss": 0.0002,
      "reward": 0.3485969334375113,
      "reward_std": 0.20335620292462409,
      "rewards/accuracy_reward": 0.3485969334375113,
      "rewards/format_reward": 0.0,
      "step": 20
    },
    {
      "completion_length": 783.1976894378662,
      "epoch": 0.04638129287853899,
      "grad_norm": 0.003633377980625143,
      "kl": 0.008260059356689452,
      "learning_rate": 9.230769230769232e-06,
      "loss": 0.0003,
      "reward": 0.41823978765169156,
      "reward_std": 0.214983982546255,
      "rewards/accuracy_reward": 0.41823978765169156,
      "rewards/format_reward": 0.0,
      "step": 30
    },
    {
      "completion_length": 796.1503639221191,
      "epoch": 0.06184172383805198,
      "grad_norm": 0.004207792051316878,
      "kl": 0.013314247131347656,
      "learning_rate": 1.230769230769231e-05,
      "loss": 0.0005,
      "reward": 0.45267856353893876,
      "reward_std": 0.22209063512273133,
      "rewards/accuracy_reward": 0.45267856353893876,
      "rewards/format_reward": 0.0,
      "step": 40
    },
    {
      "completion_length": 707.4483289718628,
      "epoch": 0.07730215479756498,
      "grad_norm": 0.0039377851262065435,
      "kl": 0.023092269897460938,
      "learning_rate": 1.5384615384615387e-05,
      "loss": 0.0009,
      "reward": 0.5378826450556516,
      "reward_std": 0.2228247532621026,
      "rewards/accuracy_reward": 0.5378826450556516,
      "rewards/format_reward": 0.0,
      "step": 50
    },
    {
      "completion_length": 738.098327255249,
      "epoch": 0.09276258575707798,
      "grad_norm": 0.004922000110022058,
      "kl": 0.03480720520019531,
      "learning_rate": 1.8461538461538465e-05,
      "loss": 0.0014,
      "reward": 0.5255101950955577,
      "reward_std": 0.21384936766698956,
      "rewards/accuracy_reward": 0.5255101950955577,
      "rewards/format_reward": 0.0,
      "step": 60
    },
    {
      "completion_length": 714.9747295379639,
      "epoch": 0.10822301671659097,
      "grad_norm": 0.025813508862428863,
      "kl": 0.11458740234375,
      "learning_rate": 1.999634547413886e-05,
      "loss": 0.0046,
      "reward": 0.5369897866621614,
      "reward_std": 0.24942327085882426,
      "rewards/accuracy_reward": 0.5369897866621614,
      "rewards/format_reward": 0.0,
      "step": 70
    },
    {
      "completion_length": 729.6705226898193,
      "epoch": 0.12368344767610397,
      "grad_norm": 0.004934077059395142,
      "kl": 0.1241455078125,
      "learning_rate": 1.9967125291968495e-05,
      "loss": 0.005,
      "reward": 0.5182397849857807,
      "reward_std": 0.22350065293721855,
      "rewards/accuracy_reward": 0.5182397849857807,
      "rewards/format_reward": 0.0,
      "step": 80
    },
    {
      "completion_length": 752.231746673584,
      "epoch": 0.13914387863561697,
      "grad_norm": 0.004490175226607394,
      "kl": 0.1960784912109375,
      "learning_rate": 1.990877034074683e-05,
      "loss": 0.0078,
      "reward": 0.47576529716607185,
      "reward_std": 0.21953069823794066,
      "rewards/accuracy_reward": 0.47576529716607185,
      "rewards/format_reward": 0.0,
      "step": 90
    },
    {
      "completion_length": 735.3539382934571,
      "epoch": 0.15460430959512997,
      "grad_norm": 0.11798471666893984,
      "kl": 0.30410003662109375,
      "learning_rate": 1.9821451197042028e-05,
      "loss": 0.0122,
      "reward": 0.4843112153466791,
      "reward_std": 0.22822934831492603,
      "rewards/accuracy_reward": 0.4843112153466791,
      "rewards/format_reward": 0.0,
      "step": 100
    },
    {
      "epoch": 0.15460430959512997,
      "eval_completion_length": 649.0448852539063,
      "eval_kl": 0.27952473958333335,
      "eval_loss": 0.009933823719620705,
      "eval_reward": 0.45850338935852053,
      "eval_reward_std": 0.26114755471547446,
      "eval_rewards/accuracy_reward": 0.45850338935852053,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 88.3141,
      "eval_samples_per_second": 1.121,
      "eval_steps_per_second": 0.17,
      "step": 100
    },
    {
      "completion_length": 732.8517738342285,
      "epoch": 0.17006474055464296,
      "grad_norm": 0.02615303400330413,
      "kl": 1.07412109375,
      "learning_rate": 1.9705423102261324e-05,
      "loss": 0.043,
      "reward": 0.3991071363911033,
      "reward_std": 0.2730441292747855,
      "rewards/accuracy_reward": 0.3991071363911033,
      "rewards/format_reward": 0.0,
      "step": 110
    },
    {
      "completion_length": 792.610315322876,
      "epoch": 0.18552517151415596,
      "grad_norm": 0.0287118341101585,
      "kl": 0.542120361328125,
      "learning_rate": 1.956102521655831e-05,
      "loss": 0.0217,
      "reward": 0.3727040741709061,
      "reward_std": 0.26482684616930785,
      "rewards/accuracy_reward": 0.3727040741709061,
      "rewards/format_reward": 0.0,
      "step": 120
    },
    {
      "completion_length": 742.854447555542,
      "epoch": 0.20098560247366895,
      "grad_norm": 0.026940670011693296,
      "kl": 0.4710235595703125,
      "learning_rate": 1.9388679627438486e-05,
      "loss": 0.0188,
      "reward": 0.46033162334933875,
      "reward_std": 0.24005382088944316,
      "rewards/accuracy_reward": 0.46033162334933875,
      "rewards/format_reward": 0.0,
      "step": 130
    },
    {
      "completion_length": 747.0299587249756,
      "epoch": 0.21644603343318194,
      "grad_norm": 0.044131709279297326,
      "kl": 0.3888519287109375,
      "learning_rate": 1.9188890115960967e-05,
      "loss": 0.0156,
      "reward": 0.4734693782404065,
      "reward_std": 0.2484902088996023,
      "rewards/accuracy_reward": 0.4734693782404065,
      "rewards/format_reward": 0.0,
      "step": 140
    },
    {
      "completion_length": 719.1231998443603,
      "epoch": 0.23190646439269494,
      "grad_norm": 0.04685629239096149,
      "kl": 0.6114959716796875,
      "learning_rate": 1.8962240684142923e-05,
      "loss": 0.0245,
      "reward": 0.4772959094494581,
      "reward_std": 0.26747574456967416,
      "rewards/accuracy_reward": 0.4772959094494581,
      "rewards/format_reward": 0.0,
      "step": 150
    },
    {
      "completion_length": 752.0598068237305,
      "epoch": 0.24736689535220793,
      "grad_norm": 0.00786304561850755,
      "kl": 0.27758941650390623,
      "learning_rate": 1.8709393847871146e-05,
      "loss": 0.0111,
      "reward": 0.4918367238715291,
      "reward_std": 0.24101508525200188,
      "rewards/accuracy_reward": 0.4918367238715291,
      "rewards/format_reward": 0.0,
      "step": 160
    },
    {
      "completion_length": 713.9837879180908,
      "epoch": 0.26282732631172095,
      "grad_norm": 0.005284057529020878,
      "kl": 0.22295989990234374,
      "learning_rate": 1.8431088700310846e-05,
      "loss": 0.0089,
      "reward": 0.5107142773456872,
      "reward_std": 0.2484118543099612,
      "rewards/accuracy_reward": 0.5107142773456872,
      "rewards/format_reward": 0.0,
      "step": 170
    },
    {
      "completion_length": 721.1976871490479,
      "epoch": 0.27828775727123395,
      "grad_norm": 0.0059488370043247725,
      "kl": 0.27528076171875,
      "learning_rate": 1.8128138751472432e-05,
      "loss": 0.011,
      "reward": 0.4794642778113484,
      "reward_std": 0.28188897627405823,
      "rewards/accuracy_reward": 0.4794642778113484,
      "rewards/format_reward": 0.0,
      "step": 180
    },
    {
      "completion_length": 685.0909267425537,
      "epoch": 0.29374818823074694,
      "grad_norm": 0.024776439473654245,
      "kl": 0.3252960205078125,
      "learning_rate": 1.780142955025139e-05,
      "loss": 0.013,
      "reward": 0.49030611482448877,
      "reward_std": 0.29441971494816244,
      "rewards/accuracy_reward": 0.49030611482448877,
      "rewards/format_reward": 0.0,
      "step": 190
    },
    {
      "completion_length": 688.9476896286011,
      "epoch": 0.30920861919025994,
      "grad_norm": 0.018511152998280257,
      "kl": 0.407830810546875,
      "learning_rate": 1.745191609589231e-05,
      "loss": 0.0163,
      "reward": 0.4463010121136904,
      "reward_std": 0.29204082568176093,
      "rewards/accuracy_reward": 0.4463010121136904,
      "rewards/format_reward": 0.0,
      "step": 200
    },
    {
      "epoch": 0.30920861919025994,
      "eval_completion_length": 684.5197184244792,
      "eval_kl": 0.31907552083333335,
      "eval_loss": 0.013225244358181953,
      "eval_reward": 0.5034013529618581,
      "eval_reward_std": 0.26736749211947125,
      "eval_rewards/accuracy_reward": 0.5034013529618581,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 88.6591,
      "eval_samples_per_second": 1.117,
      "eval_steps_per_second": 0.169,
      "step": 200
    },
    {
      "completion_length": 688.203684425354,
      "epoch": 0.32466905014977293,
      "grad_norm": 0.04166474923994624,
      "kl": 0.4008056640625,
      "learning_rate": 1.7080620046443503e-05,
      "loss": 0.016,
      "reward": 0.45382652347907426,
      "reward_std": 0.27682951451279225,
      "rewards/accuracy_reward": 0.45382652347907426,
      "rewards/format_reward": 0.0,
      "step": 210
    },
    {
      "completion_length": 671.052919960022,
      "epoch": 0.3401294811092859,
      "grad_norm": 0.03352718681096581,
      "kl": 0.3749359130859375,
      "learning_rate": 1.6688626732362192e-05,
      "loss": 0.015,
      "reward": 0.4804846870712936,
      "reward_std": 0.2816257219295949,
      "rewards/accuracy_reward": 0.4804846870712936,
      "rewards/format_reward": 0.0,
      "step": 220
    },
    {
      "completion_length": 643.8374883651734,
      "epoch": 0.3555899120687989,
      "grad_norm": 0.020351629255596172,
      "kl": 0.41683349609375,
      "learning_rate": 1.6277081983999742e-05,
      "loss": 0.0167,
      "reward": 0.4418367271311581,
      "reward_std": 0.28084711018018427,
      "rewards/accuracy_reward": 0.4418367271311581,
      "rewards/format_reward": 0.0,
      "step": 230
    },
    {
      "completion_length": 676.2223068237305,
      "epoch": 0.3710503430283119,
      "grad_norm": 0.05256794474201724,
      "kl": 0.4444091796875,
      "learning_rate": 1.5847188782240473e-05,
      "loss": 0.0178,
      "reward": 0.4517857059370726,
      "reward_std": 0.29922230960801244,
      "rewards/accuracy_reward": 0.4517857059370726,
      "rewards/format_reward": 0.0,
      "step": 240
    },
    {
      "completion_length": 668.4691181182861,
      "epoch": 0.3865107739878249,
      "grad_norm": 0.03902146766301656,
      "kl": 0.5255126953125,
      "learning_rate": 1.5400203742084508e-05,
      "loss": 0.021,
      "reward": 0.4383928496390581,
      "reward_std": 0.2969975466839969,
      "rewards/accuracy_reward": 0.4383928496390581,
      "rewards/format_reward": 0.0,
      "step": 250
    },
    {
      "completion_length": 603.6825130462646,
      "epoch": 0.4019712049473379,
      "grad_norm": 0.02207794461900282,
      "kl": 0.4320556640625,
      "learning_rate": 1.4937433439453465e-05,
      "loss": 0.0173,
      "reward": 0.45331631754525004,
      "reward_std": 0.29558597495779393,
      "rewards/accuracy_reward": 0.45331631754525004,
      "rewards/format_reward": 0.0,
      "step": 260
    },
    {
      "completion_length": 640.7654218673706,
      "epoch": 0.4174316359068509,
      "grad_norm": 0.02435264101678905,
      "kl": 0.510498046875,
      "learning_rate": 1.4460230591956097e-05,
      "loss": 0.0204,
      "reward": 0.4378826460801065,
      "reward_std": 0.2932774598710239,
      "rewards/accuracy_reward": 0.4378826460801065,
      "rewards/format_reward": 0.0,
      "step": 270
    },
    {
      "completion_length": 645.673583984375,
      "epoch": 0.4328920668663639,
      "grad_norm": 0.08816250567178784,
      "kl": 0.658056640625,
      "learning_rate": 1.3969990104777712e-05,
      "loss": 0.0263,
      "reward": 0.42423468651250007,
      "reward_std": 0.2902136994060129,
      "rewards/accuracy_reward": 0.42423468651250007,
      "rewards/format_reward": 0.0,
      "step": 280
    },
    {
      "completion_length": 620.3529205322266,
      "epoch": 0.4483524978258769,
      "grad_norm": 0.03880404727089487,
      "kl": 0.92225341796875,
      "learning_rate": 1.3468144993251735e-05,
      "loss": 0.0369,
      "reward": 0.45433672657236457,
      "reward_std": 0.29660415309481325,
      "rewards/accuracy_reward": 0.45433672657236457,
      "rewards/format_reward": 0.0,
      "step": 290
    },
    {
      "completion_length": 662.0209051132202,
      "epoch": 0.4638129287853899,
      "grad_norm": 0.0580905052695571,
      "kl": 1.16871337890625,
      "learning_rate": 1.295616219403197e-05,
      "loss": 0.0468,
      "reward": 0.4096938706934452,
      "reward_std": 0.3028755730483681,
      "rewards/accuracy_reward": 0.4096938706934452,
      "rewards/format_reward": 0.0,
      "step": 300
    },
    {
      "epoch": 0.4638129287853899,
      "eval_completion_length": 604.8503275553386,
      "eval_kl": 1.15458984375,
      "eval_loss": 0.04879453405737877,
      "eval_reward": 0.4707482943932215,
      "eval_reward_std": 0.27726571063200633,
      "eval_rewards/accuracy_reward": 0.4707482943932215,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 85.9413,
      "eval_samples_per_second": 1.152,
      "eval_steps_per_second": 0.175,
      "step": 300
    },
    {
      "completion_length": 618.5189908981323,
      "epoch": 0.47927335974490287,
      "grad_norm": 0.04382094795523676,
      "kl": 0.9635986328125,
      "learning_rate": 1.2435538277109919e-05,
      "loss": 0.0385,
      "reward": 0.44183672638610005,
      "reward_std": 0.28991906996816397,
      "rewards/accuracy_reward": 0.44183672638610005,
      "rewards/format_reward": 0.0,
      "step": 310
    },
    {
      "completion_length": 576.9373609542847,
      "epoch": 0.49473379070441587,
      "grad_norm": 0.03238899986450347,
      "kl": 0.90596923828125,
      "learning_rate": 1.19077950712113e-05,
      "loss": 0.0362,
      "reward": 0.4463010119274259,
      "reward_std": 0.27278245403431356,
      "rewards/accuracy_reward": 0.4463010119274259,
      "rewards/format_reward": 0.0,
      "step": 320
    },
    {
      "completion_length": 535.9521591186524,
      "epoch": 0.5101942216639289,
      "grad_norm": 0.018431343598236234,
      "kl": 0.5312744140625,
      "learning_rate": 1.137447521535908e-05,
      "loss": 0.0213,
      "reward": 0.5080357047729194,
      "reward_std": 0.2620480744168162,
      "rewards/accuracy_reward": 0.5080357047729194,
      "rewards/format_reward": 0.0,
      "step": 330
    },
    {
      "completion_length": 646.3938673019409,
      "epoch": 0.5256546526234419,
      "grad_norm": 0.02317552951279719,
      "kl": 0.513262939453125,
      "learning_rate": 1.0837137649606241e-05,
      "loss": 0.0205,
      "reward": 0.44579080818220973,
      "reward_std": 0.29469514368101957,
      "rewards/accuracy_reward": 0.44579080818220973,
      "rewards/format_reward": 0.0,
      "step": 340
    },
    {
      "completion_length": 586.0761362075806,
      "epoch": 0.5411150835829549,
      "grad_norm": 0.016226017261457597,
      "kl": 0.3769073486328125,
      "learning_rate": 1.0297353058119209e-05,
      "loss": 0.0151,
      "reward": 0.4859693799167871,
      "reward_std": 0.2713043099734932,
      "rewards/accuracy_reward": 0.4859693799167871,
      "rewards/format_reward": 0.0,
      "step": 350
    },
    {
      "completion_length": 571.9863378524781,
      "epoch": 0.5565755145424679,
      "grad_norm": 0.0504216391704931,
      "kl": 0.599774169921875,
      "learning_rate": 9.756699277932196e-06,
      "loss": 0.024,
      "reward": 0.48443876539822667,
      "reward_std": 0.28054420156404375,
      "rewards/accuracy_reward": 0.48443876539822667,
      "rewards/format_reward": 0.0,
      "step": 360
    },
    {
      "completion_length": 563.5765197753906,
      "epoch": 0.5720359455019809,
      "grad_norm": 0.13990571845181696,
      "kl": 0.52132568359375,
      "learning_rate": 9.216756686793163e-06,
      "loss": 0.0209,
      "reward": 0.45790815523359923,
      "reward_std": 0.26586609268561007,
      "rewards/accuracy_reward": 0.45790815523359923,
      "rewards/format_reward": 0.0,
      "step": 370
    },
    {
      "completion_length": 575.829453086853,
      "epoch": 0.5874963764614939,
      "grad_norm": 0.013870992065908727,
      "kl": 0.330126953125,
      "learning_rate": 8.67910358358298e-06,
      "loss": 0.0132,
      "reward": 0.4734693790320307,
      "reward_std": 0.27416237886063755,
      "rewards/accuracy_reward": 0.4734693790320307,
      "rewards/format_reward": 0.0,
      "step": 380
    },
    {
      "completion_length": 568.1160606384277,
      "epoch": 0.6029568074210069,
      "grad_norm": 0.010163464330427247,
      "kl": 0.2426544189453125,
      "learning_rate": 8.145311574811325e-06,
      "loss": 0.0097,
      "reward": 0.4954081534408033,
      "reward_std": 0.26710083298385146,
      "rewards/accuracy_reward": 0.4954081534408033,
      "rewards/format_reward": 0.0,
      "step": 390
    },
    {
      "completion_length": 665.6186065673828,
      "epoch": 0.6184172383805199,
      "grad_norm": 0.013029334976500356,
      "kl": 0.321246337890625,
      "learning_rate": 7.616940980675004e-06,
      "loss": 0.0128,
      "reward": 0.4474489719606936,
      "reward_std": 0.27854115669615565,
      "rewards/accuracy_reward": 0.4474489719606936,
      "rewards/format_reward": 0.0,
      "step": 400
    },
    {
      "epoch": 0.6184172383805199,
      "eval_completion_length": 627.0203979492187,
      "eval_kl": 0.2732747395833333,
      "eval_loss": 0.011420400813221931,
      "eval_reward": 0.5047619010011355,
      "eval_reward_std": 0.2536137938499451,
      "eval_rewards/accuracy_reward": 0.5047619010011355,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 87.816,
      "eval_samples_per_second": 1.127,
      "eval_steps_per_second": 0.171,
      "step": 400
    },
    {
      "completion_length": 646.4711584091186,
      "epoch": 0.6338776693400329,
      "grad_norm": 0.023099135498393764,
      "kl": 0.343701171875,
      "learning_rate": 7.095536274107046e-06,
      "loss": 0.0137,
      "reward": 0.4645408088341355,
      "reward_std": 0.2891133207827806,
      "rewards/accuracy_reward": 0.4645408088341355,
      "rewards/format_reward": 0.0,
      "step": 410
    },
    {
      "completion_length": 587.3729467391968,
      "epoch": 0.6493381002995459,
      "grad_norm": 0.013467292425479267,
      "kl": 0.2858642578125,
      "learning_rate": 6.58262156614881e-06,
      "loss": 0.0114,
      "reward": 0.503188765514642,
      "reward_std": 0.2590713477227837,
      "rewards/accuracy_reward": 0.503188765514642,
      "rewards/format_reward": 0.0,
      "step": 420
    },
    {
      "completion_length": 594.2052188873291,
      "epoch": 0.6647985312590589,
      "grad_norm": 0.013766538000227693,
      "kl": 0.285015869140625,
      "learning_rate": 6.079696150841634e-06,
      "loss": 0.0114,
      "reward": 0.4913265212439001,
      "reward_std": 0.2684762907214463,
      "rewards/accuracy_reward": 0.4913265212439001,
      "rewards/format_reward": 0.0,
      "step": 430
    },
    {
      "completion_length": 631.7939949035645,
      "epoch": 0.6802589622185718,
      "grad_norm": 0.07419237005783043,
      "kl": 0.346197509765625,
      "learning_rate": 5.588230122660672e-06,
      "loss": 0.0138,
      "reward": 0.4860969296656549,
      "reward_std": 0.2773646651767194,
      "rewards/accuracy_reward": 0.4860969296656549,
      "rewards/format_reward": 0.0,
      "step": 440
    },
    {
      "completion_length": 586.2428462982177,
      "epoch": 0.6957193931780848,
      "grad_norm": 0.015937060852966797,
      "kl": 0.2478759765625,
      "learning_rate": 5.109660079301668e-06,
      "loss": 0.0099,
      "reward": 0.5220663199201226,
      "reward_std": 0.2621162030380219,
      "rewards/accuracy_reward": 0.5220663199201226,
      "rewards/format_reward": 0.0,
      "step": 450
    },
    {
      "completion_length": 641.8402923583984,
      "epoch": 0.7111798241375978,
      "grad_norm": 0.04445799644422213,
      "kl": 0.3908416748046875,
      "learning_rate": 4.64538492238166e-06,
      "loss": 0.0156,
      "reward": 0.4795918288640678,
      "reward_std": 0.28201754316687583,
      "rewards/accuracy_reward": 0.4795918288640678,
      "rewards/format_reward": 0.0,
      "step": 460
    },
    {
      "completion_length": 648.5068748474121,
      "epoch": 0.7266402550971108,
      "grad_norm": 0.018272003157409882,
      "kl": 0.44942626953125,
      "learning_rate": 4.196761768328599e-06,
      "loss": 0.018,
      "reward": 0.45650509353727103,
      "reward_std": 0.2825955556239933,
      "rewards/accuracy_reward": 0.45650509353727103,
      "rewards/format_reward": 0.0,
      "step": 470
    },
    {
      "completion_length": 632.1244766235352,
      "epoch": 0.7421006860566238,
      "grad_norm": 0.03550172747608362,
      "kl": 0.54378662109375,
      "learning_rate": 3.7651019814126656e-06,
      "loss": 0.0218,
      "reward": 0.4636479509063065,
      "reward_std": 0.2770009428262711,
      "rewards/accuracy_reward": 0.4636479509063065,
      "rewards/format_reward": 0.0,
      "step": 480
    },
    {
      "completion_length": 632.8059818267823,
      "epoch": 0.7575611170161368,
      "grad_norm": 0.9437401498024955,
      "kl": 1.1802734375,
      "learning_rate": 3.3516673405151546e-06,
      "loss": 0.0472,
      "reward": 0.45255101229995487,
      "reward_std": 0.2746642493642867,
      "rewards/accuracy_reward": 0.45255101229995487,
      "rewards/format_reward": 0.0,
      "step": 490
    },
    {
      "completion_length": 586.2049638748169,
      "epoch": 0.7730215479756498,
      "grad_norm": 0.11836217149566043,
      "kl": 0.545013427734375,
      "learning_rate": 2.957666350839663e-06,
      "loss": 0.0218,
      "reward": 0.48941325647756456,
      "reward_std": 0.263414288777858,
      "rewards/accuracy_reward": 0.48941325647756456,
      "rewards/format_reward": 0.0,
      "step": 500
    },
    {
      "epoch": 0.7730215479756498,
      "eval_completion_length": 551.6952229817708,
      "eval_kl": 0.28020833333333334,
      "eval_loss": 0.011639236472547054,
      "eval_reward": 0.5387754996617635,
      "eval_reward_std": 0.2534260580937068,
      "eval_rewards/accuracy_reward": 0.5387754996617635,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 85.6752,
      "eval_samples_per_second": 1.156,
      "eval_steps_per_second": 0.175,
      "step": 500
    },
    {
      "completion_length": 598.6526651382446,
      "epoch": 0.7884819789351628,
      "grad_norm": 0.014324651971585192,
      "kl": 0.284661865234375,
      "learning_rate": 2.5842507113469307e-06,
      "loss": 0.0114,
      "reward": 0.49579080580733714,
      "reward_std": 0.2849952794611454,
      "rewards/accuracy_reward": 0.49579080580733714,
      "rewards/format_reward": 0.0,
      "step": 510
    },
    {
      "completion_length": 602.975754737854,
      "epoch": 0.8039424098946758,
      "grad_norm": 0.013302847380980731,
      "kl": 0.287591552734375,
      "learning_rate": 2.2325119482391466e-06,
      "loss": 0.0115,
      "reward": 0.4869897892698646,
      "reward_std": 0.28300182512030003,
      "rewards/accuracy_reward": 0.4869897892698646,
      "rewards/format_reward": 0.0,
      "step": 520
    },
    {
      "completion_length": 617.0672046661377,
      "epoch": 0.8194028408541888,
      "grad_norm": 0.011506624405655665,
      "kl": 0.316864013671875,
      "learning_rate": 1.9034782243345074e-06,
      "loss": 0.0127,
      "reward": 0.49005101155489683,
      "reward_std": 0.283530889172107,
      "rewards/accuracy_reward": 0.49005101155489683,
      "rewards/format_reward": 0.0,
      "step": 530
    },
    {
      "completion_length": 608.665803527832,
      "epoch": 0.8348632718137018,
      "grad_norm": 0.016337273376864313,
      "kl": 0.557275390625,
      "learning_rate": 1.5981113336584041e-06,
      "loss": 0.0223,
      "reward": 0.49119897168129684,
      "reward_std": 0.28141105216927825,
      "rewards/accuracy_reward": 0.49119897168129684,
      "rewards/format_reward": 0.0,
      "step": 540
    },
    {
      "completion_length": 625.1159269332886,
      "epoch": 0.8503237027732148,
      "grad_norm": 0.057861981590189904,
      "kl": 0.260540771484375,
      "learning_rate": 1.3173038900362977e-06,
      "loss": 0.0104,
      "reward": 0.4499999931082129,
      "reward_std": 0.27894868138246237,
      "rewards/accuracy_reward": 0.4499999931082129,
      "rewards/format_reward": 0.0,
      "step": 550
    },
    {
      "completion_length": 632.5747304916382,
      "epoch": 0.8657841337327278,
      "grad_norm": 0.030466073357798586,
      "kl": 0.293267822265625,
      "learning_rate": 1.0618767179063416e-06,
      "loss": 0.0117,
      "reward": 0.4618622355163097,
      "reward_std": 0.2925746965222061,
      "rewards/accuracy_reward": 0.4618622355163097,
      "rewards/format_reward": 0.0,
      "step": 560
    },
    {
      "completion_length": 641.7084041595459,
      "epoch": 0.8812445646922408,
      "grad_norm": 0.017331692484703218,
      "kl": 0.316583251953125,
      "learning_rate": 8.325764529785851e-07,
      "loss": 0.0127,
      "reward": 0.4483418272808194,
      "reward_std": 0.28196476846933366,
      "rewards/accuracy_reward": 0.4483418272808194,
      "rewards/format_reward": 0.0,
      "step": 570
    },
    {
      "completion_length": 646.2618488311767,
      "epoch": 0.8967049956517538,
      "grad_norm": 0.009327065043797485,
      "kl": 0.303216552734375,
      "learning_rate": 6.300733597542086e-07,
      "loss": 0.0121,
      "reward": 0.44196427753195167,
      "reward_std": 0.28285656329244374,
      "rewards/accuracy_reward": 0.44196427753195167,
      "rewards/format_reward": 0.0,
      "step": 580
    },
    {
      "completion_length": 626.1570028305053,
      "epoch": 0.9121654266112668,
      "grad_norm": 0.010427073776751795,
      "kl": 0.31864013671875,
      "learning_rate": 4.549593722844492e-07,
      "loss": 0.0127,
      "reward": 0.46224488839507105,
      "reward_std": 0.28828581105917694,
      "rewards/accuracy_reward": 0.46224488839507105,
      "rewards/format_reward": 0.0,
      "step": 590
    },
    {
      "completion_length": 628.7367259979249,
      "epoch": 0.9276258575707798,
      "grad_norm": 0.01713962537975438,
      "kl": 0.3144775390625,
      "learning_rate": 3.0774636389618196e-07,
      "loss": 0.0126,
      "reward": 0.4502550953067839,
      "reward_std": 0.2758318264503032,
      "rewards/accuracy_reward": 0.4502550953067839,
      "rewards/format_reward": 0.0,
      "step": 600
    },
    {
      "epoch": 0.9276258575707798,
      "eval_completion_length": 607.1224405924479,
      "eval_kl": 0.2647786458333333,
      "eval_loss": 0.011027935892343521,
      "eval_reward": 0.4816326439380646,
      "eval_reward_std": 0.25309162139892577,
      "eval_rewards/accuracy_reward": 0.4816326439380646,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 87.0986,
      "eval_samples_per_second": 1.137,
      "eval_steps_per_second": 0.172,
      "step": 600
    },
    {
      "completion_length": 633.0672080993652,
      "epoch": 0.9430862885302927,
      "grad_norm": 0.022562920761009898,
      "kl": 0.302935791015625,
      "learning_rate": 1.8886465094192895e-07,
      "loss": 0.0121,
      "reward": 0.46020407443866135,
      "reward_std": 0.2828258784487844,
      "rewards/accuracy_reward": 0.46020407443866135,
      "rewards/format_reward": 0.0,
      "step": 610
    },
    {
      "completion_length": 626.6780488967895,
      "epoch": 0.9585467194898057,
      "grad_norm": 0.012428624166763307,
      "kl": 0.3028564453125,
      "learning_rate": 9.866173494794462e-08,
      "loss": 0.0121,
      "reward": 0.447959177242592,
      "reward_std": 0.27789597446098924,
      "rewards/accuracy_reward": 0.447959177242592,
      "rewards/format_reward": 0.0,
      "step": 620
    },
    {
      "completion_length": 630.8656784057617,
      "epoch": 0.9740071504493187,
      "grad_norm": 0.009372661260439265,
      "kl": 0.305780029296875,
      "learning_rate": 3.7401286837214224e-08,
      "loss": 0.0122,
      "reward": 0.4577806035755202,
      "reward_std": 0.27292990586720406,
      "rewards/accuracy_reward": 0.4577806035755202,
      "rewards/format_reward": 0.0,
      "step": 630
    },
    {
      "completion_length": 624.4659309387207,
      "epoch": 0.9894675814088317,
      "grad_norm": 0.012393498840602491,
      "kl": 0.301348876953125,
      "learning_rate": 5.262376196544239e-09,
      "loss": 0.0121,
      "reward": 0.4693877460435033,
      "reward_std": 0.27738194689154627,
      "rewards/accuracy_reward": 0.4693877460435033,
      "rewards/format_reward": 0.0,
      "step": 640
    },
    {
      "completion_length": 626.0862976710001,
      "epoch": 0.9987438399845395,
      "kl": 0.31524658203125,
      "reward": 0.46088434507449466,
      "reward_std": 0.28149245004169643,
      "rewards/accuracy_reward": 0.46088434507449466,
      "rewards/format_reward": 0.0,
      "step": 646,
      "total_flos": 0.0,
      "train_loss": 0.01602011715643949,
      "train_runtime": 67938.2282,
      "train_samples_per_second": 1.066,
      "train_steps_per_second": 0.01
    }
  ],
  "logging_steps": 10,
  "max_steps": 646,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}