{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9981333333333333,
  "eval_steps": 100,
  "global_step": 394,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 579.8596701371042,
      "epoch": 0.002533333333333333,
      "grad_norm": 0.2689627707004547,
      "kl": -9.097551044664885e-08,
      "learning_rate": 7.500000000000001e-08,
      "loss": 0.0054,
      "reward": 0.6754386079938788,
      "reward_std": 0.29693952202796936,
      "rewards/accuracy_reward": 0.6754386079938788,
      "rewards/format_reward": 0.0,
      "step": 1
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 623.1096651177658,
      "epoch": 0.012666666666666666,
      "grad_norm": 0.34043097496032715,
      "kl": 0.00013973210987291838,
      "learning_rate": 3.75e-07,
      "loss": -0.0101,
      "reward": 0.6052631694627436,
      "reward_std": 0.3655273945708024,
      "rewards/accuracy_reward": 0.6052631694627436,
      "rewards/format_reward": 0.0,
      "step": 5
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 574.2947516190378,
      "epoch": 0.025333333333333333,
      "grad_norm": 0.16866758465766907,
      "kl": 0.0002021940130936472,
      "learning_rate": 7.5e-07,
      "loss": -0.0133,
      "reward": 0.6228070329678687,
      "reward_std": 0.3261075675487518,
      "rewards/accuracy_reward": 0.6228070329678687,
      "rewards/format_reward": 0.0,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 593.8473849647924,
      "epoch": 0.038,
      "grad_norm": 0.13506826758384705,
      "kl": 0.00030391090794613484,
      "learning_rate": 1.125e-06,
      "loss": -0.0171,
      "reward": 0.6192982615608918,
      "reward_std": 0.3493922035945089,
      "rewards/accuracy_reward": 0.6192982615608918,
      "rewards/format_reward": 0.0,
      "step": 15
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 614.5087890625,
      "epoch": 0.050666666666666665,
      "grad_norm": 0.37117841839790344,
      "kl": 0.0014222195273951481,
      "learning_rate": 1.5e-06,
      "loss": -0.0263,
      "reward": 0.6087719443597291,
      "reward_std": 0.35981847079176654,
      "rewards/accuracy_reward": 0.6087719443597291,
      "rewards/format_reward": 0.0,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 610.6105422170539,
      "epoch": 0.06333333333333334,
      "grad_norm": 0.6702502369880676,
      "kl": 0.008063848395096627,
      "learning_rate": 1.875e-06,
      "loss": -0.0323,
      "reward": 0.6000000163128502,
      "reward_std": 0.3776336585220538,
      "rewards/accuracy_reward": 0.6000000163128502,
      "rewards/format_reward": 0.0,
      "step": 25
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 581.708786171361,
      "epoch": 0.076,
      "grad_norm": 0.13385862112045288,
      "kl": 0.08222049913908305,
      "learning_rate": 2.25e-06,
      "loss": -0.003,
      "reward": 0.5719298386260083,
      "reward_std": 0.3216610023849889,
      "rewards/accuracy_reward": 0.5719298386260083,
      "rewards/format_reward": 0.0,
      "step": 30
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 580.1789640727796,
      "epoch": 0.08866666666666667,
      "grad_norm": 0.1613887995481491,
      "kl": 0.4072471217105263,
      "learning_rate": 2.6250000000000003e-06,
      "loss": -0.0034,
      "reward": 0.5684210676895944,
      "reward_std": 0.3606692671775818,
      "rewards/accuracy_reward": 0.5684210676895944,
      "rewards/format_reward": 0.0,
      "step": 35
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 573.603520443565,
      "epoch": 0.10133333333333333,
      "grad_norm": 4.28112268447876,
      "kl": 0.28351083052785775,
      "learning_rate": 3e-06,
      "loss": -0.0148,
      "reward": 0.6192982620314548,
      "reward_std": 0.34681932079164607,
      "rewards/accuracy_reward": 0.6192982620314548,
      "rewards/format_reward": 0.0,
      "step": 40
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 597.3666823537726,
      "epoch": 0.114,
      "grad_norm": 0.09016856551170349,
      "kl": 0.1709989447342722,
      "learning_rate": 2.998523534736735e-06,
      "loss": -0.0274,
      "reward": 0.6052631729527523,
      "reward_std": 0.37054079206366286,
      "rewards/accuracy_reward": 0.6052631729527523,
      "rewards/format_reward": 0.0,
      "step": 45
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 588.608786171361,
      "epoch": 0.12666666666666668,
      "grad_norm": 0.2084578275680542,
      "kl": 0.2026009107890882,
      "learning_rate": 2.994097045546504e-06,
      "loss": -0.0133,
      "reward": 0.5824561528469386,
      "reward_std": 0.3644034429600364,
      "rewards/accuracy_reward": 0.5824561528469386,
      "rewards/format_reward": 0.0,
      "step": 50
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 592.9579096743935,
      "epoch": 0.13933333333333334,
      "grad_norm": 1.332377552986145,
      "kl": 0.253752979479338,
      "learning_rate": 2.986729246506011e-06,
      "loss": -0.0199,
      "reward": 0.6315789619558736,
      "reward_std": 0.34024513646175986,
      "rewards/accuracy_reward": 0.6315789619558736,
      "rewards/format_reward": 0.0,
      "step": 55
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 580.366682032535,
      "epoch": 0.152,
      "grad_norm": 0.056244488805532455,
      "kl": 17.54536361694336,
      "learning_rate": 2.976434642014389e-06,
      "loss": 0.8101,
      "reward": 0.6280701911763141,
      "reward_std": 0.3611795199544806,
      "rewards/accuracy_reward": 0.6280701911763141,
      "rewards/format_reward": 0.0,
      "step": 60
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 602.4929979826275,
      "epoch": 0.16466666666666666,
      "grad_norm": 2.9207770824432373,
      "kl": 0.25210153680098685,
      "learning_rate": 2.9632334982395456e-06,
      "loss": 0.0016,
      "reward": 0.6473684381497534,
      "reward_std": 0.3406482906718003,
      "rewards/accuracy_reward": 0.6473684381497534,
      "rewards/format_reward": 0.0,
      "step": 65
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 550.657911119963,
      "epoch": 0.17733333333333334,
      "grad_norm": 2.1469576358795166,
      "kl": 1.1785144203587583,
      "learning_rate": 2.947151803221774e-06,
      "loss": 0.0246,
      "reward": 0.5982456306093618,
      "reward_std": 0.3625405713131553,
      "rewards/accuracy_reward": 0.5982456306093618,
      "rewards/format_reward": 0.0,
      "step": 70
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 630.1982622648541,
      "epoch": 0.19,
      "grad_norm": 0.437906414270401,
      "kl": 0.2049952456825658,
      "learning_rate": 2.928221215713164e-06,
      "loss": -0.0136,
      "reward": 0.6368421231445514,
      "reward_std": 0.3507869362831116,
      "rewards/accuracy_reward": 0.6368421231445514,
      "rewards/format_reward": 0.0,
      "step": 75
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 554.66141614412,
      "epoch": 0.20266666666666666,
      "grad_norm": 0.20336699485778809,
      "kl": 0.840125154194079,
      "learning_rate": 2.906479002853542e-06,
      "loss": -0.0331,
      "reward": 0.6403508934535478,
      "reward_std": 0.33991540262573644,
      "rewards/accuracy_reward": 0.6403508934535478,
      "rewards/format_reward": 0.0,
      "step": 80
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 571.3140498111122,
      "epoch": 0.21533333333333332,
      "grad_norm": 0.566848874092102,
      "kl": 0.42301218133223684,
      "learning_rate": 2.8819679668056195e-06,
      "loss": -0.0154,
      "reward": 0.6754386137974889,
      "reward_std": 0.3516690385969062,
      "rewards/accuracy_reward": 0.6754386137974889,
      "rewards/format_reward": 0.0,
      "step": 85
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 576.1280825966282,
      "epoch": 0.228,
      "grad_norm": 5.783878326416016,
      "kl": 0.3097421746504934,
      "learning_rate": 2.8547363604937856e-06,
      "loss": 0.0077,
      "reward": 0.612280716237269,
      "reward_std": 0.355850856241427,
      "rewards/accuracy_reward": 0.612280716237269,
      "rewards/format_reward": 0.0,
      "step": 90
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 615.8421213250411,
      "epoch": 0.24066666666666667,
      "grad_norm": 0.29518190026283264,
      "kl": 0.10967632092927632,
      "learning_rate": 2.824837792612416e-06,
      "loss": -0.0353,
      "reward": 0.6122807160804146,
      "reward_std": 0.3359477858794363,
      "rewards/accuracy_reward": 0.6122807160804146,
      "rewards/format_reward": 0.0,
      "step": 95
    },
    {
      "epoch": 0.25333333333333335,
      "grad_norm": 0.09735328704118729,
      "learning_rate": 2.792331122090709e-06,
      "loss": -0.0192,
      "step": 100
    },
    {
      "epoch": 0.25333333333333335,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 581.6460822784423,
      "eval_kl": 0.4462999359130859,
      "eval_loss": -0.020424701273441315,
      "eval_reward": 0.4882333454877138,
      "eval_reward_std": 0.3652301513493061,
      "eval_rewards/accuracy_reward": 0.4882333454877138,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 95085.0702,
      "eval_samples_per_second": 0.053,
      "eval_steps_per_second": 0.009,
      "step": 100
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 560.9368566412675,
      "epoch": 0.266,
      "grad_norm": 13.26487922668457,
      "kl": 0.22525361713610198,
      "learning_rate": 2.7572803422217976e-06,
      "loss": -0.0193,
      "reward": 0.6263158046885541,
      "reward_std": 0.3673114654264952,
      "rewards/accuracy_reward": 0.6263158046885541,
      "rewards/format_reward": 0.0,
      "step": 105
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 572.0772078664679,
      "epoch": 0.2786666666666667,
      "grad_norm": 5.774922847747803,
      "kl": 1.1548378392269736,
      "learning_rate": 2.71975445468425e-06,
      "loss": -0.0186,
      "reward": 0.6631579132456529,
      "reward_std": 0.3699968554471668,
      "rewards/accuracy_reward": 0.6631579132456529,
      "rewards/format_reward": 0.0,
      "step": 110
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 617.7017713044819,
      "epoch": 0.29133333333333333,
      "grad_norm": 0.3488950729370117,
      "kl": 0.4739940442537007,
      "learning_rate": 2.679827333703964e-06,
      "loss": 0.0054,
      "reward": 0.5789473808125446,
      "reward_std": 0.359339523942847,
      "rewards/accuracy_reward": 0.5789473808125446,
      "rewards/format_reward": 0.0,
      "step": 115
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 584.6894883005242,
      "epoch": 0.304,
      "grad_norm": 2.7635438442230225,
      "kl": 0.8890348735608553,
      "learning_rate": 2.637577580623858e-06,
      "loss": -0.0151,
      "reward": 0.5947368579475503,
      "reward_std": 0.34380959868431094,
      "rewards/accuracy_reward": 0.5947368579475503,
      "rewards/format_reward": 0.0,
      "step": 120
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 615.0772094726562,
      "epoch": 0.31666666666666665,
      "grad_norm": 1.1497012376785278,
      "kl": 0.36164293791118424,
      "learning_rate": 2.593088369167671e-06,
      "loss": -0.0181,
      "reward": 0.6192982593649312,
      "reward_std": 0.3435111723448101,
      "rewards/accuracy_reward": 0.6192982593649312,
      "rewards/format_reward": 0.0,
      "step": 125
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 576.5666825143915,
      "epoch": 0.3293333333333333,
      "grad_norm": 6.164768695831299,
      "kl": 0.5363926937705592,
      "learning_rate": 2.5464472817024772e-06,
      "loss": 0.018,
      "reward": 0.6684210694149921,
      "reward_std": 0.368413172583831,
      "rewards/accuracy_reward": 0.6684210694149921,
      "rewards/format_reward": 0.0,
      "step": 130
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 566.4719445479543,
      "epoch": 0.342,
      "grad_norm": 2.0130836963653564,
      "kl": 1.256256103515625,
      "learning_rate": 2.497746136822254e-06,
      "loss": 0.0444,
      "reward": 0.642105276176804,
      "reward_std": 0.33451331728383116,
      "rewards/accuracy_reward": 0.642105276176804,
      "rewards/format_reward": 0.0,
      "step": 135
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 615.5193124871505,
      "epoch": 0.3546666666666667,
      "grad_norm": 2.369405746459961,
      "kl": 0.7050068102384869,
      "learning_rate": 2.4470808085919304e-06,
      "loss": -0.0144,
      "reward": 0.6421052804118709,
      "reward_std": 0.3574345387910542,
      "rewards/accuracy_reward": 0.6421052804118709,
      "rewards/format_reward": 0.0,
      "step": 140
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 576.3912459524055,
      "epoch": 0.36733333333333335,
      "grad_norm": 4.594287872314453,
      "kl": 2.437645841899671,
      "learning_rate": 2.3945510378077523e-06,
      "loss": 0.0796,
      "reward": 0.6842105448246002,
      "reward_std": 0.350753252757223,
      "rewards/accuracy_reward": 0.6842105448246002,
      "rewards/format_reward": 0.0,
      "step": 145
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 596.2193150570519,
      "epoch": 0.38,
      "grad_norm": 2.5179603099823,
      "kl": 1.6508487099095395,
      "learning_rate": 2.340260235645519e-06,
      "loss": 0.0431,
      "reward": 0.663157911363401,
      "reward_std": 0.3722002707029644,
      "rewards/accuracy_reward": 0.663157911363401,
      "rewards/format_reward": 0.0,
      "step": 150
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 579.5965046129728,
      "epoch": 0.39266666666666666,
      "grad_norm": 76.81166076660156,
      "kl": 4.629216886821546,
      "learning_rate": 2.2843152800832416e-06,
      "loss": 0.094,
      "reward": 0.5807017698099739,
      "reward_std": 0.37164790002923265,
      "rewards/accuracy_reward": 0.5807017698099739,
      "rewards/format_reward": 0.0,
      "step": 155
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 604.2456311677631,
      "epoch": 0.4053333333333333,
      "grad_norm": 2.316026210784912,
      "kl": 5.967475971422697,
      "learning_rate": 2.2268263054989753e-06,
      "loss": 0.242,
      "reward": 0.5666666804175627,
      "reward_std": 0.3491695720898478,
      "rewards/accuracy_reward": 0.5666666804175627,
      "rewards/format_reward": 0.0,
      "step": 160
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 585.1210690146999,
      "epoch": 0.418,
      "grad_norm": 1.3882242441177368,
      "kl": 1.542066072162829,
      "learning_rate": 2.167906485858047e-06,
      "loss": 0.031,
      "reward": 0.6631579112065465,
      "reward_std": 0.32709676968423945,
      "rewards/accuracy_reward": 0.6631579112065465,
      "rewards/format_reward": 0.0,
      "step": 165
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 609.7789647152549,
      "epoch": 0.43066666666666664,
      "grad_norm": 23.364316940307617,
      "kl": 3.0389545641447366,
      "learning_rate": 2.1076718119164804e-06,
      "loss": 0.0985,
      "reward": 0.5368421203211734,
      "reward_std": 0.36661528725373116,
      "rewards/accuracy_reward": 0.5368421203211734,
      "rewards/format_reward": 0.0,
      "step": 170
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 589.2105436626233,
      "epoch": 0.44333333333333336,
      "grad_norm": 2.261157274246216,
      "kl": 3.034598581414474,
      "learning_rate": 2.0462408628792335e-06,
      "loss": 0.0464,
      "reward": 0.5877193132513447,
      "reward_std": 0.32118205584977805,
      "rewards/accuracy_reward": 0.5877193132513447,
      "rewards/format_reward": 0.0,
      "step": 175
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 583.0210664447985,
      "epoch": 0.456,
      "grad_norm": 1.490301251411438,
      "kl": 4.092078279194079,
      "learning_rate": 1.9837345729627633e-06,
      "loss": 0.0794,
      "reward": 0.6087719440460205,
      "reward_std": 0.34601063853815983,
      "rewards/accuracy_reward": 0.6087719440460205,
      "rewards/format_reward": 0.0,
      "step": 180
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 591.0193131296259,
      "epoch": 0.4686666666666667,
      "grad_norm": 3.377087354660034,
      "kl": 39.17373753597862,
      "learning_rate": 1.9202759933214665e-06,
      "loss": 1.5296,
      "reward": 0.5684210672190315,
      "reward_std": 0.3722002681932951,
      "rewards/accuracy_reward": 0.5684210672190315,
      "rewards/format_reward": 0.0,
      "step": 185
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 628.3315945273952,
      "epoch": 0.48133333333333334,
      "grad_norm": 15.796894073486328,
      "kl": 8.018802682976974,
      "learning_rate": 1.8559900498066726e-06,
      "loss": 0.2629,
      "reward": 0.5877193149767423,
      "reward_std": 0.3454245896715867,
      "rewards/accuracy_reward": 0.5877193149767423,
      "rewards/format_reward": 0.0,
      "step": 190
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 616.9403683311061,
      "epoch": 0.494,
      "grad_norm": 0.7116127014160156,
      "kl": 11.00859888980263,
      "learning_rate": 1.7910032970350677e-06,
      "loss": 0.5039,
      "reward": 0.614035103195592,
      "reward_std": 0.3618389898224881,
      "rewards/accuracy_reward": 0.614035103195592,
      "rewards/format_reward": 0.0,
      "step": 195
    },
    {
      "epoch": 0.5066666666666667,
      "grad_norm": 0.7055822014808655,
      "learning_rate": 1.7254436692507058e-06,
      "loss": 0.2295,
      "step": 200
    },
    {
      "epoch": 0.5066666666666667,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 581.6460156143188,
      "eval_kl": 44.771087231445314,
      "eval_loss": 1.8106327056884766,
      "eval_reward": 0.4892333455443382,
      "eval_reward_std": 0.362811917424202,
      "eval_rewards/accuracy_reward": 0.4891666788816452,
      "eval_rewards/format_reward": 6.666666865348816e-05,
      "eval_runtime": 95474.2074,
      "eval_samples_per_second": 0.052,
      "eval_steps_per_second": 0.009,
      "step": 200
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 624.9587893837377,
      "epoch": 0.5193333333333333,
      "grad_norm": 0.2618827223777771,
      "kl": 4.135612246864721,
      "learning_rate": 1.6594402284710481e-06,
      "loss": 0.0336,
      "reward": 0.564035101783903,
      "reward_std": 0.3579249236144518,
      "rewards/accuracy_reward": 0.564035101783903,
      "rewards/format_reward": 0.0,
      "step": 205
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 621.8368573640522,
      "epoch": 0.532,
      "grad_norm": 1.038638710975647,
      "kl": 0.783648681640625,
      "learning_rate": 1.593122910412851e-06,
      "loss": 0.0144,
      "reward": 0.6175438767985294,
      "reward_std": 0.35596638729697777,
      "rewards/accuracy_reward": 0.6175438767985294,
      "rewards/format_reward": 0.0,
      "step": 210
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 612.468435990183,
      "epoch": 0.5446666666666666,
      "grad_norm": 15.33467960357666,
      "kl": 5.278036338404605,
      "learning_rate": 1.5266222686980693e-06,
      "loss": 0.2523,
      "reward": 0.5964912433373301,
      "reward_std": 0.344139332834043,
      "rewards/accuracy_reward": 0.5964912433373301,
      "rewards/format_reward": 0.0,
      "step": 215
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 579.1210691753187,
      "epoch": 0.5573333333333333,
      "grad_norm": 14.52474308013916,
      "kl": 0.7972836143092106,
      "learning_rate": 1.460069217843338e-06,
      "loss": -0.0145,
      "reward": 0.5754386097192764,
      "reward_std": 0.3417132879558362,
      "rewards/accuracy_reward": 0.5754386097192764,
      "rewards/format_reward": 0.0,
      "step": 220
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 567.8386103579872,
      "epoch": 0.57,
      "grad_norm": 4.826749324798584,
      "kl": 1.95933837890625,
      "learning_rate": 1.3935947755389924e-06,
      "loss": -0.0023,
      "reward": 0.6614035228365346,
      "reward_std": 0.33425700288069876,
      "rewards/accuracy_reward": 0.6614035228365346,
      "rewards/format_reward": 0.0,
      "step": 225
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 609.7438760857833,
      "epoch": 0.5826666666666667,
      "grad_norm": 2.4607062339782715,
      "kl": 2.425996800472862,
      "learning_rate": 1.3273298047249756e-06,
      "loss": 0.0442,
      "reward": 0.542105278059056,
      "reward_std": 0.3630532004331288,
      "rewards/accuracy_reward": 0.542105278059056,
      "rewards/format_reward": 0.0,
      "step": 230
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 633.8508912739001,
      "epoch": 0.5953333333333334,
      "grad_norm": 1.241820216178894,
      "kl": 4.220809454666941,
      "learning_rate": 1.2614047559713923e-06,
      "loss": 0.1149,
      "reward": 0.5491228218141355,
      "reward_std": 0.3291172884012524,
      "rewards/accuracy_reward": 0.5491228218141355,
      "rewards/format_reward": 0.0,
      "step": 235
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 607.8807173879524,
      "epoch": 0.608,
      "grad_norm": 4.545963287353516,
      "kl": 1.1027640894839639,
      "learning_rate": 1.1959494106708598e-06,
      "loss": 0.0236,
      "reward": 0.6000000144305982,
      "reward_std": 0.36793422448007684,
      "rewards/accuracy_reward": 0.6000000144305982,
      "rewards/format_reward": 0.0,
      "step": 240
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 598.5158052143298,
      "epoch": 0.6206666666666667,
      "grad_norm": 3.975374221801758,
      "kl": 17.801213314658717,
      "learning_rate": 1.1310926255482204e-06,
      "loss": 0.6585,
      "reward": 0.6385965044561185,
      "reward_std": 0.33627751940175105,
      "rewards/accuracy_reward": 0.6385965044561185,
      "rewards/format_reward": 0.0,
      "step": 245
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 573.9087850470291,
      "epoch": 0.6333333333333333,
      "grad_norm": 1.1734281778335571,
      "kl": 2.3962697882401316,
      "learning_rate": 1.0669620789905688e-06,
      "loss": 0.027,
      "reward": 0.5877193123102188,
      "reward_std": 0.38500809261673374,
      "rewards/accuracy_reward": 0.5877193123102188,
      "rewards/format_reward": 0.0,
      "step": 250
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 606.8912443462171,
      "epoch": 0.646,
      "grad_norm": 2.513296365737915,
      "kl": 2.2693404348273027,
      "learning_rate": 1.0036840196969795e-06,
      "loss": 0.0564,
      "reward": 0.6245614178870854,
      "reward_std": 0.32143837056661906,
      "rewards/accuracy_reward": 0.6245614178870854,
      "rewards/format_reward": 0.0,
      "step": 255
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 567.249137316252,
      "epoch": 0.6586666666666666,
      "grad_norm": 1.0174099206924438,
      "kl": 3.5286929481907894,
      "learning_rate": 9.413830181427508e-07,
      "loss": 0.0882,
      "reward": 0.6456140494660327,
      "reward_std": 0.3340006878501491,
      "rewards/accuracy_reward": 0.6456140494660327,
      "rewards/format_reward": 0.0,
      "step": 260
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 610.2052803441098,
      "epoch": 0.6713333333333333,
      "grad_norm": 1.029402256011963,
      "kl": 2.918760922080592,
      "learning_rate": 8.801817213474331e-07,
      "loss": 0.0485,
      "reward": 0.5929824714597903,
      "reward_std": 0.34714905211800023,
      "rewards/accuracy_reward": 0.5929824714597903,
      "rewards/format_reward": 0.0,
      "step": 265
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 638.8596645957546,
      "epoch": 0.684,
      "grad_norm": 1.5382988452911377,
      "kl": 2.563796193976151,
      "learning_rate": 8.202006114294044e-07,
      "loss": 0.0392,
      "reward": 0.5807017699668282,
      "reward_std": 0.36489082386619165,
      "rewards/accuracy_reward": 0.5807017699668282,
      "rewards/format_reward": 0.0,
      "step": 270
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 608.0421232524671,
      "epoch": 0.6966666666666667,
      "grad_norm": 16.033626556396484,
      "kl": 1.5960372121710527,
      "learning_rate": 7.615577684223272e-07,
      "loss": 0.0303,
      "reward": 0.5894736991116875,
      "reward_std": 0.3749873553451739,
      "rewards/accuracy_reward": 0.5894736991116875,
      "rewards/format_reward": 0.0,
      "step": 275
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 600.415805053711,
      "epoch": 0.7093333333333334,
      "grad_norm": 2.8626461029052734,
      "kl": 1.9747047825863486,
      "learning_rate": 7.043686378203864e-07,
      "loss": 0.0147,
      "reward": 0.6368421217328624,
      "reward_std": 0.37260342334446156,
      "rewards/accuracy_reward": 0.6368421217328624,
      "rewards/format_reward": 0.0,
      "step": 280
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 616.5772094726562,
      "epoch": 0.722,
      "grad_norm": 1.4815430641174316,
      "kl": 2.7695633737664473,
      "learning_rate": 6.487458033099425e-07,
      "loss": 0.0412,
      "reward": 0.5385965032012839,
      "reward_std": 0.35034166982299403,
      "rewards/accuracy_reward": 0.5385965032012839,
      "rewards/format_reward": 0.0,
      "step": 285
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 645.3333529823705,
      "epoch": 0.7346666666666667,
      "grad_norm": 0.6981754302978516,
      "kl": 2.3487503854851974,
      "learning_rate": 5.947987651349942e-07,
      "loss": 0.0472,
      "reward": 0.566666682927232,
      "reward_std": 0.3772305058805566,
      "rewards/accuracy_reward": 0.566666682927232,
      "rewards/format_reward": 0.0,
      "step": 290
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 615.3684377569901,
      "epoch": 0.7473333333333333,
      "grad_norm": 2.056654214859009,
      "kl": 2.483375308388158,
      "learning_rate": 5.426337245327703e-07,
      "loss": 0.0787,
      "reward": 0.5491228204024465,
      "reward_std": 0.3359477874479796,
      "rewards/accuracy_reward": 0.5491228204024465,
      "rewards/format_reward": 0.0,
      "step": 295
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.973532497882843,
      "learning_rate": 4.923533746638108e-07,
      "loss": 0.0014,
      "step": 300
    },
    {
      "epoch": 0.76,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 579.6517157089233,
      "eval_kl": 2.4090683868408203,
      "eval_loss": 0.046611957252025604,
      "eval_reward": 0.49193334555327894,
      "eval_reward_std": 0.35773685903549196,
      "eval_rewards/accuracy_reward": 0.49193334555327894,
      "eval_rewards/format_reward": 0.0,
      "eval_runtime": 94754.8266,
      "eval_samples_per_second": 0.053,
      "eval_steps_per_second": 0.009,
      "step": 300
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 594.8903664036801,
      "epoch": 0.7726666666666666,
      "grad_norm": 0.28998905420303345,
      "kl": 1.3069589715254935,
      "learning_rate": 4.440566984481256e-07,
      "loss": 0.0171,
      "reward": 0.6210526459311184,
      "reward_std": 0.34825076147129663,
      "rewards/accuracy_reward": 0.6210526459311184,
      "rewards/format_reward": 0.0,
      "step": 305
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 618.7473841616982,
      "epoch": 0.7853333333333333,
      "grad_norm": 0.46500492095947266,
      "kl": 1.2823197214226973,
      "learning_rate": 3.978387737053994e-07,
      "loss": 0.0136,
      "reward": 0.5666666805744172,
      "reward_std": 0.32507625222206116,
      "rewards/accuracy_reward": 0.5666666805744172,
      "rewards/format_reward": 0.0,
      "step": 310
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 618.5772081877055,
      "epoch": 0.798,
      "grad_norm": 0.40139040350914,
      "kl": 1.015325927734375,
      "learning_rate": 3.5379058598286167e-07,
      "loss": -0.0158,
      "reward": 0.5438596634488356,
      "reward_std": 0.3694023759741532,
      "rewards/accuracy_reward": 0.5438596634488356,
      "rewards/format_reward": 0.0,
      "step": 315
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 594.957911119963,
      "epoch": 0.8106666666666666,
      "grad_norm": 1.3908910751342773,
      "kl": 1.2546104029605263,
      "learning_rate": 3.119988494392894e-07,
      "loss": 0.0289,
      "reward": 0.6508772078313325,
      "reward_std": 0.3327888513866224,
      "rewards/accuracy_reward": 0.6508772078313325,
      "rewards/format_reward": 0.0,
      "step": 320
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 582.7333484850432,
      "epoch": 0.8233333333333334,
      "grad_norm": 4.380345344543457,
      "kl": 418.79378180252877,
      "learning_rate": 2.725458361377465e-07,
      "loss": 15.2043,
      "reward": 0.608771941379497,
      "reward_std": 0.34012960415137444,
      "rewards/accuracy_reward": 0.608771941379497,
      "rewards/format_reward": 0.0,
      "step": 325
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 634.2175588507401,
      "epoch": 0.836,
      "grad_norm": 6.396597862243652,
      "kl": 1.1365401418585526,
      "learning_rate": 2.3550921408312737e-07,
      "loss": 0.0132,
      "reward": 0.5859649261361675,
      "reward_std": 0.34545827006038865,
      "rewards/accuracy_reward": 0.5859649261361675,
      "rewards/format_reward": 0.0,
      "step": 330
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 614.2140493292558,
      "epoch": 0.8486666666666667,
      "grad_norm": 0.9369886517524719,
      "kl": 1.5584103232935855,
      "learning_rate": 2.0096189432334195e-07,
      "loss": 0.0201,
      "reward": 0.6035087874061182,
      "reward_std": 0.3459035368342149,
      "rewards/accuracy_reward": 0.6035087874061182,
      "rewards/format_reward": 0.0,
      "step": 335
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 619.8245766087582,
      "epoch": 0.8613333333333333,
      "grad_norm": 0.42530328035354614,
      "kl": 1.4711069207442433,
      "learning_rate": 1.6897188741514286e-07,
      "loss": 0.0519,
      "reward": 0.575438610503548,
      "reward_std": 0.3146836676095661,
      "rewards/accuracy_reward": 0.575438610503548,
      "rewards/format_reward": 0.0,
      "step": 340
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 577.3737000314812,
      "epoch": 0.874,
      "grad_norm": 0.3861980140209198,
      "kl": 1.7285689504523025,
      "learning_rate": 1.396021695371582e-07,
      "loss": -0.0088,
      "reward": 0.6385965082206224,
      "reward_std": 0.37675155840421976,
      "rewards/accuracy_reward": 0.6385965082206224,
      "rewards/format_reward": 0.0,
      "step": 345
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 629.012297941509,
      "epoch": 0.8866666666666667,
      "grad_norm": 4.221518039703369,
      "kl": 1.9362998560855262,
      "learning_rate": 1.1291055851370623e-07,
      "loss": 0.056,
      "reward": 0.582456154886045,
      "reward_std": 0.38160365223884585,
      "rewards/accuracy_reward": 0.582456154886045,
      "rewards/format_reward": 0.0,
      "step": 350
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 589.6280856483861,
      "epoch": 0.8993333333333333,
      "grad_norm": 2.3517720699310303,
      "kl": 1.8255171926398026,
      "learning_rate": 8.894959999345015e-08,
      "loss": 0.0361,
      "reward": 0.5754386076801702,
      "reward_std": 0.34303222267251265,
      "rewards/accuracy_reward": 0.5754386076801702,
      "rewards/format_reward": 0.0,
      "step": 355
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 629.0754549528423,
      "epoch": 0.912,
      "grad_norm": 1.3438383340835571,
      "kl": 1.5210680509868422,
      "learning_rate": 6.776646400696212e-08,
      "loss": 0.0234,
      "reward": 0.5789473809693989,
      "reward_std": 0.3563298034040551,
      "rewards/accuracy_reward": 0.5789473809693989,
      "rewards/format_reward": 0.0,
      "step": 360
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 585.6263320119757,
      "epoch": 0.9246666666666666,
      "grad_norm": 1.6730248928070068,
      "kl": 3.8371864720394737,
      "learning_rate": 4.940285210684375e-08,
      "loss": 0.075,
      "reward": 0.6596491382310264,
      "reward_std": 0.3641158220015074,
      "rewards/accuracy_reward": 0.6596491382310264,
      "rewards/format_reward": 0.0,
      "step": 365
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 634.8017714651007,
      "epoch": 0.9373333333333334,
      "grad_norm": 0.6605441570281982,
      "kl": 1.9925395764802631,
      "learning_rate": 3.389491527319999e-08,
      "loss": 0.0375,
      "reward": 0.6578947547234987,
      "reward_std": 0.41116404282419305,
      "rewards/accuracy_reward": 0.6578947547234987,
      "rewards/format_reward": 0.0,
      "step": 370
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 600.2000179893092,
      "epoch": 0.95,
      "grad_norm": 1.5404088497161865,
      "kl": 1.4656384919819079,
      "learning_rate": 2.127318274608381e-08,
      "loss": 0.0059,
      "reward": 0.5789473817536706,
      "reward_std": 0.350234569060175,
      "rewards/accuracy_reward": 0.5789473817536706,
      "rewards/format_reward": 0.0,
      "step": 375
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 573.8017702604595,
      "epoch": 0.9626666666666667,
      "grad_norm": 2.143493175506592,
      "kl": 1.3233176783511513,
      "learning_rate": 1.1562501925013125e-08,
      "loss": 0.0301,
      "reward": 0.6631579116771096,
      "reward_std": 0.3379683045964492,
      "rewards/accuracy_reward": 0.6631579116771096,
      "rewards/format_reward": 0.0,
      "step": 380
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 616.8175617418791,
      "epoch": 0.9753333333333334,
      "grad_norm": 0.5864923596382141,
      "kl": 1.3502071982935855,
      "learning_rate": 4.781989453874814e-09,
      "loss": 0.0052,
      "reward": 0.5807017691825566,
      "reward_std": 0.3575416382990385,
      "rewards/accuracy_reward": 0.5807017691825566,
      "rewards/format_reward": 0.0,
      "step": 385
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 612.1122984233656,
      "epoch": 0.988,
      "grad_norm": 0.3266775906085968,
      "kl": 1.6176237407483554,
      "learning_rate": 9.44993587509657e-10,
      "loss": 0.0364,
      "reward": 0.5947368560652984,
      "reward_std": 0.300396885683662,
      "rewards/accuracy_reward": 0.5947368560652984,
      "rewards/format_reward": 0.0,
      "step": 390
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 612.8508915148283,
      "epoch": 0.9981333333333333,
      "kl": 1.1464434171977795,
      "reward": 0.603070187529451,
      "reward_std": 0.27440278663447026,
      "rewards/accuracy_reward": 0.603070187529451,
      "rewards/format_reward": 0.0,
      "step": 394,
      "total_flos": 0.0,
      "train_loss": 0.2665646580783036,
      "train_runtime": 484901.6937,
      "train_samples_per_second": 0.015,
      "train_steps_per_second": 0.001
    }
  ],
  "logging_steps": 5,
  "max_steps": 394,
  "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": 2,
  "trial_name": null,
  "trial_params": null
}