{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8,
  "eval_steps": 100,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.5915625745430588,
      "epoch": 0.016,
      "grad_norm": 5.71875,
      "learning_rate": 1.8e-07,
      "loss": 0.8466,
      "mean_token_accuracy": 0.8017937403172255,
      "num_tokens": 340696.0,
      "step": 10
    },
    {
      "entropy": 0.5926362803205848,
      "epoch": 0.032,
      "grad_norm": 5.59375,
      "learning_rate": 3.8e-07,
      "loss": 0.8487,
      "mean_token_accuracy": 0.8023959375917912,
      "num_tokens": 675930.0,
      "step": 20
    },
    {
      "entropy": 0.591606929153204,
      "epoch": 0.048,
      "grad_norm": 4.96875,
      "learning_rate": 5.800000000000001e-07,
      "loss": 0.8302,
      "mean_token_accuracy": 0.8044086921960115,
      "num_tokens": 1007956.0,
      "step": 30
    },
    {
      "entropy": 0.6204556789249182,
      "epoch": 0.064,
      "grad_norm": 4.9375,
      "learning_rate": 7.8e-07,
      "loss": 0.8457,
      "mean_token_accuracy": 0.7998181059956551,
      "num_tokens": 1340560.0,
      "step": 40
    },
    {
      "entropy": 0.6393004363402724,
      "epoch": 0.08,
      "grad_norm": 4.3125,
      "learning_rate": 9.800000000000001e-07,
      "loss": 0.8343,
      "mean_token_accuracy": 0.8025688324123621,
      "num_tokens": 1679432.0,
      "step": 50
    },
    {
      "entropy": 0.6593567343428731,
      "epoch": 0.096,
      "grad_norm": 2.5625,
      "learning_rate": 1.1800000000000001e-06,
      "loss": 0.8046,
      "mean_token_accuracy": 0.8078926544636488,
      "num_tokens": 2021570.0,
      "step": 60
    },
    {
      "entropy": 0.7051705561578274,
      "epoch": 0.112,
      "grad_norm": 2.8125,
      "learning_rate": 1.3800000000000001e-06,
      "loss": 0.8246,
      "mean_token_accuracy": 0.8029463469982148,
      "num_tokens": 2359297.0,
      "step": 70
    },
    {
      "entropy": 0.754444376192987,
      "epoch": 0.128,
      "grad_norm": 2.4375,
      "learning_rate": 1.5800000000000001e-06,
      "loss": 0.8435,
      "mean_token_accuracy": 0.79781529083848,
      "num_tokens": 2683168.0,
      "step": 80
    },
    {
      "entropy": 0.7250820267945528,
      "epoch": 0.144,
      "grad_norm": 1.90625,
      "learning_rate": 1.7800000000000001e-06,
      "loss": 0.8195,
      "mean_token_accuracy": 0.8042717296630144,
      "num_tokens": 3019426.0,
      "step": 90
    },
    {
      "entropy": 0.7252413807436824,
      "epoch": 0.16,
      "grad_norm": 2.109375,
      "learning_rate": 1.98e-06,
      "loss": 0.8097,
      "mean_token_accuracy": 0.8047077178955078,
      "num_tokens": 3354390.0,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_chemistry_entropy": 0.7109885907173157,
      "eval_chemistry_loss": 0.804060161113739,
      "eval_chemistry_mean_token_accuracy": 0.8113441648483276,
      "eval_chemistry_num_tokens": 3354390.0,
      "eval_chemistry_runtime": 34.8657,
      "eval_chemistry_samples_per_second": 14.341,
      "eval_chemistry_steps_per_second": 3.585,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_physics_entropy": 0.7321163377761841,
      "eval_physics_loss": 0.8287621140480042,
      "eval_physics_mean_token_accuracy": 0.8033607420921326,
      "eval_physics_num_tokens": 3354390.0,
      "eval_physics_runtime": 41.3727,
      "eval_physics_samples_per_second": 12.085,
      "eval_physics_steps_per_second": 3.021,
      "step": 100
    },
    {
      "entropy": 0.7267085557803512,
      "epoch": 0.176,
      "grad_norm": 2.234375,
      "learning_rate": 2.1800000000000003e-06,
      "loss": 0.8058,
      "mean_token_accuracy": 0.8033798363059759,
      "num_tokens": 3684374.0,
      "step": 110
    },
    {
      "entropy": 0.7212623069062829,
      "epoch": 0.192,
      "grad_norm": 1.828125,
      "learning_rate": 2.38e-06,
      "loss": 0.7989,
      "mean_token_accuracy": 0.8043727930635214,
      "num_tokens": 4023915.0,
      "step": 120
    },
    {
      "entropy": 0.7022871913388371,
      "epoch": 0.208,
      "grad_norm": 2.1875,
      "learning_rate": 2.5800000000000003e-06,
      "loss": 0.7758,
      "mean_token_accuracy": 0.8102051325142383,
      "num_tokens": 4352045.0,
      "step": 130
    },
    {
      "entropy": 0.6965231081470847,
      "epoch": 0.224,
      "grad_norm": 1.8203125,
      "learning_rate": 2.7800000000000005e-06,
      "loss": 0.7726,
      "mean_token_accuracy": 0.8089764468371868,
      "num_tokens": 4681371.0,
      "step": 140
    },
    {
      "entropy": 0.6956626677885651,
      "epoch": 0.24,
      "grad_norm": 1.6640625,
      "learning_rate": 2.9800000000000003e-06,
      "loss": 0.7608,
      "mean_token_accuracy": 0.8104452796280384,
      "num_tokens": 5021784.0,
      "step": 150
    },
    {
      "entropy": 0.708347842283547,
      "epoch": 0.256,
      "grad_norm": 1.9453125,
      "learning_rate": 3.1800000000000005e-06,
      "loss": 0.7636,
      "mean_token_accuracy": 0.809263763204217,
      "num_tokens": 5367308.0,
      "step": 160
    },
    {
      "entropy": 0.7138571232557297,
      "epoch": 0.272,
      "grad_norm": 2.03125,
      "learning_rate": 3.3800000000000007e-06,
      "loss": 0.7721,
      "mean_token_accuracy": 0.8050564888864755,
      "num_tokens": 5699101.0,
      "step": 170
    },
    {
      "entropy": 0.7222439678385854,
      "epoch": 0.288,
      "grad_norm": 1.7109375,
      "learning_rate": 3.58e-06,
      "loss": 0.7766,
      "mean_token_accuracy": 0.8063390862196684,
      "num_tokens": 6022837.0,
      "step": 180
    },
    {
      "entropy": 0.7117253359407186,
      "epoch": 0.304,
      "grad_norm": 2.265625,
      "learning_rate": 3.7800000000000002e-06,
      "loss": 0.7535,
      "mean_token_accuracy": 0.809041840955615,
      "num_tokens": 6343826.0,
      "step": 190
    },
    {
      "entropy": 0.7200546801090241,
      "epoch": 0.32,
      "grad_norm": 1.7265625,
      "learning_rate": 3.980000000000001e-06,
      "loss": 0.7524,
      "mean_token_accuracy": 0.8073045868426562,
      "num_tokens": 6670164.0,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_chemistry_entropy": 0.711975371837616,
      "eval_chemistry_loss": 0.7488377690315247,
      "eval_chemistry_mean_token_accuracy": 0.8110983457565307,
      "eval_chemistry_num_tokens": 6670164.0,
      "eval_chemistry_runtime": 34.8588,
      "eval_chemistry_samples_per_second": 14.344,
      "eval_chemistry_steps_per_second": 3.586,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_physics_entropy": 0.7178810501098633,
      "eval_physics_loss": 0.7585415244102478,
      "eval_physics_mean_token_accuracy": 0.8081966361999512,
      "eval_physics_num_tokens": 6670164.0,
      "eval_physics_runtime": 41.7087,
      "eval_physics_samples_per_second": 11.988,
      "eval_physics_steps_per_second": 2.997,
      "step": 200
    },
    {
      "entropy": 0.7111945543438196,
      "epoch": 0.336,
      "grad_norm": 1.578125,
      "learning_rate": 4.18e-06,
      "loss": 0.7313,
      "mean_token_accuracy": 0.8092666126787662,
      "num_tokens": 7003404.0,
      "step": 210
    },
    {
      "entropy": 0.6962820990011096,
      "epoch": 0.352,
      "grad_norm": 1.8828125,
      "learning_rate": 4.38e-06,
      "loss": 0.712,
      "mean_token_accuracy": 0.8133333619683981,
      "num_tokens": 7342593.0,
      "step": 220
    },
    {
      "entropy": 0.6933605512604117,
      "epoch": 0.368,
      "grad_norm": 1.6484375,
      "learning_rate": 4.58e-06,
      "loss": 0.6986,
      "mean_token_accuracy": 0.8163042381405831,
      "num_tokens": 7675569.0,
      "step": 230
    },
    {
      "entropy": 0.7312716225162148,
      "epoch": 0.384,
      "grad_norm": 1.9296875,
      "learning_rate": 4.78e-06,
      "loss": 0.7319,
      "mean_token_accuracy": 0.8038436233997345,
      "num_tokens": 8012218.0,
      "step": 240
    },
    {
      "entropy": 0.6996616214513779,
      "epoch": 0.4,
      "grad_norm": 1.796875,
      "learning_rate": 4.980000000000001e-06,
      "loss": 0.6917,
      "mean_token_accuracy": 0.8134940605610609,
      "num_tokens": 8350120.0,
      "step": 250
    },
    {
      "entropy": 0.6945684082806111,
      "epoch": 0.416,
      "grad_norm": 1.6953125,
      "learning_rate": 5.18e-06,
      "loss": 0.678,
      "mean_token_accuracy": 0.8151624854654074,
      "num_tokens": 8688773.0,
      "step": 260
    },
    {
      "entropy": 0.7035746525973081,
      "epoch": 0.432,
      "grad_norm": 1.8359375,
      "learning_rate": 5.380000000000001e-06,
      "loss": 0.6888,
      "mean_token_accuracy": 0.8118865180760622,
      "num_tokens": 9022094.0,
      "step": 270
    },
    {
      "entropy": 0.6850203935056924,
      "epoch": 0.448,
      "grad_norm": 1.6640625,
      "learning_rate": 5.580000000000001e-06,
      "loss": 0.6727,
      "mean_token_accuracy": 0.8186012674123049,
      "num_tokens": 9360608.0,
      "step": 280
    },
    {
      "entropy": 0.6806540634483099,
      "epoch": 0.464,
      "grad_norm": 1.875,
      "learning_rate": 5.78e-06,
      "loss": 0.6592,
      "mean_token_accuracy": 0.8200464811176061,
      "num_tokens": 9688896.0,
      "step": 290
    },
    {
      "entropy": 0.6852828785777092,
      "epoch": 0.48,
      "grad_norm": 1.828125,
      "learning_rate": 5.98e-06,
      "loss": 0.6672,
      "mean_token_accuracy": 0.8180346965789795,
      "num_tokens": 10025864.0,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_chemistry_entropy": 0.7274510788917542,
      "eval_chemistry_loss": 0.7009354829788208,
      "eval_chemistry_mean_token_accuracy": 0.8133535690307617,
      "eval_chemistry_num_tokens": 10025864.0,
      "eval_chemistry_runtime": 34.811,
      "eval_chemistry_samples_per_second": 14.363,
      "eval_chemistry_steps_per_second": 3.591,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_physics_entropy": 0.7100022671222687,
      "eval_physics_loss": 0.7013525366783142,
      "eval_physics_mean_token_accuracy": 0.813646514415741,
      "eval_physics_num_tokens": 10025864.0,
      "eval_physics_runtime": 41.4407,
      "eval_physics_samples_per_second": 12.065,
      "eval_physics_steps_per_second": 3.016,
      "step": 300
    },
    {
      "entropy": 0.7103753052651882,
      "epoch": 0.496,
      "grad_norm": 1.6171875,
      "learning_rate": 6.18e-06,
      "loss": 0.6847,
      "mean_token_accuracy": 0.812765546515584,
      "num_tokens": 10362573.0,
      "step": 310
    },
    {
      "entropy": 0.6982777336612344,
      "epoch": 0.512,
      "grad_norm": 1.8125,
      "learning_rate": 6.380000000000001e-06,
      "loss": 0.6716,
      "mean_token_accuracy": 0.8169939771294594,
      "num_tokens": 10694440.0,
      "step": 320
    },
    {
      "entropy": 0.708159969188273,
      "epoch": 0.528,
      "grad_norm": 1.7421875,
      "learning_rate": 6.5800000000000005e-06,
      "loss": 0.693,
      "mean_token_accuracy": 0.8125004164874554,
      "num_tokens": 11022357.0,
      "step": 330
    },
    {
      "entropy": 0.699538960121572,
      "epoch": 0.544,
      "grad_norm": 1.828125,
      "learning_rate": 6.780000000000001e-06,
      "loss": 0.6711,
      "mean_token_accuracy": 0.8153361644595861,
      "num_tokens": 11345911.0,
      "step": 340
    },
    {
      "entropy": 0.6910958087071777,
      "epoch": 0.56,
      "grad_norm": 1.5078125,
      "learning_rate": 6.98e-06,
      "loss": 0.6659,
      "mean_token_accuracy": 0.8177970543503761,
      "num_tokens": 11687225.0,
      "step": 350
    },
    {
      "entropy": 0.6805725138634443,
      "epoch": 0.576,
      "grad_norm": 1.6484375,
      "learning_rate": 7.180000000000001e-06,
      "loss": 0.6565,
      "mean_token_accuracy": 0.8207161612808704,
      "num_tokens": 12015720.0,
      "step": 360
    },
    {
      "entropy": 0.6865102933719754,
      "epoch": 0.592,
      "grad_norm": 1.84375,
      "learning_rate": 7.3800000000000005e-06,
      "loss": 0.6638,
      "mean_token_accuracy": 0.8175321508198976,
      "num_tokens": 12343393.0,
      "step": 370
    },
    {
      "entropy": 0.6841516578570008,
      "epoch": 0.608,
      "grad_norm": 1.6328125,
      "learning_rate": 7.58e-06,
      "loss": 0.6582,
      "mean_token_accuracy": 0.8192304316908121,
      "num_tokens": 12671474.0,
      "step": 380
    },
    {
      "entropy": 0.7005494212731719,
      "epoch": 0.624,
      "grad_norm": 1.6015625,
      "learning_rate": 7.78e-06,
      "loss": 0.6674,
      "mean_token_accuracy": 0.8156571265310049,
      "num_tokens": 12996077.0,
      "step": 390
    },
    {
      "entropy": 0.6937378702685237,
      "epoch": 0.64,
      "grad_norm": 1.6171875,
      "learning_rate": 7.980000000000002e-06,
      "loss": 0.6701,
      "mean_token_accuracy": 0.8157013915479183,
      "num_tokens": 13325291.0,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_chemistry_entropy": 0.7147737612724304,
      "eval_chemistry_loss": 0.692807137966156,
      "eval_chemistry_mean_token_accuracy": 0.8137030324935913,
      "eval_chemistry_num_tokens": 13325291.0,
      "eval_chemistry_runtime": 34.8141,
      "eval_chemistry_samples_per_second": 14.362,
      "eval_chemistry_steps_per_second": 3.591,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_physics_entropy": 0.684398841381073,
      "eval_physics_loss": 0.6776748895645142,
      "eval_physics_mean_token_accuracy": 0.8169920797348023,
      "eval_physics_num_tokens": 13325291.0,
      "eval_physics_runtime": 41.4558,
      "eval_physics_samples_per_second": 12.061,
      "eval_physics_steps_per_second": 3.015,
      "step": 400
    },
    {
      "entropy": 0.6792764134705067,
      "epoch": 0.656,
      "grad_norm": 1.625,
      "learning_rate": 8.18e-06,
      "loss": 0.6538,
      "mean_token_accuracy": 0.8188512332737445,
      "num_tokens": 13671434.0,
      "step": 410
    },
    {
      "entropy": 0.6741176716983318,
      "epoch": 0.672,
      "grad_norm": 1.703125,
      "learning_rate": 8.380000000000001e-06,
      "loss": 0.6479,
      "mean_token_accuracy": 0.8206427529454231,
      "num_tokens": 13993668.0,
      "step": 420
    },
    {
      "entropy": 0.6748962674289942,
      "epoch": 0.688,
      "grad_norm": 1.5625,
      "learning_rate": 8.580000000000001e-06,
      "loss": 0.6489,
      "mean_token_accuracy": 0.821071095019579,
      "num_tokens": 14334907.0,
      "step": 430
    },
    {
      "entropy": 0.6882562624290586,
      "epoch": 0.704,
      "grad_norm": 1.765625,
      "learning_rate": 8.78e-06,
      "loss": 0.6631,
      "mean_token_accuracy": 0.8166982632130384,
      "num_tokens": 14670639.0,
      "step": 440
    },
    {
      "entropy": 0.6710160901769996,
      "epoch": 0.72,
      "grad_norm": 1.828125,
      "learning_rate": 8.98e-06,
      "loss": 0.6451,
      "mean_token_accuracy": 0.8207047082483768,
      "num_tokens": 14987995.0,
      "step": 450
    },
    {
      "entropy": 0.6616904048249126,
      "epoch": 0.736,
      "grad_norm": 1.65625,
      "learning_rate": 9.180000000000002e-06,
      "loss": 0.6374,
      "mean_token_accuracy": 0.8214268710464239,
      "num_tokens": 15324659.0,
      "step": 460
    },
    {
      "entropy": 0.6631792033091187,
      "epoch": 0.752,
      "grad_norm": 1.6171875,
      "learning_rate": 9.38e-06,
      "loss": 0.6379,
      "mean_token_accuracy": 0.8226557578891516,
      "num_tokens": 15657342.0,
      "step": 470
    },
    {
      "entropy": 0.6816092301160097,
      "epoch": 0.768,
      "grad_norm": 1.640625,
      "learning_rate": 9.58e-06,
      "loss": 0.6582,
      "mean_token_accuracy": 0.8162681456655264,
      "num_tokens": 15988993.0,
      "step": 480
    },
    {
      "entropy": 0.6584033869206906,
      "epoch": 0.784,
      "grad_norm": 1.6484375,
      "learning_rate": 9.780000000000001e-06,
      "loss": 0.6307,
      "mean_token_accuracy": 0.8244805630296469,
      "num_tokens": 16333857.0,
      "step": 490
    },
    {
      "entropy": 0.6675900464877487,
      "epoch": 0.8,
      "grad_norm": 1.625,
      "learning_rate": 9.980000000000001e-06,
      "loss": 0.6444,
      "mean_token_accuracy": 0.8208535581827163,
      "num_tokens": 16664690.0,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_chemistry_entropy": 0.7186227707862854,
      "eval_chemistry_loss": 0.6891804337501526,
      "eval_chemistry_mean_token_accuracy": 0.812873167514801,
      "eval_chemistry_num_tokens": 16664690.0,
      "eval_chemistry_runtime": 34.7308,
      "eval_chemistry_samples_per_second": 14.396,
      "eval_chemistry_steps_per_second": 3.599,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_physics_entropy": 0.6725036988258362,
      "eval_physics_loss": 0.6597840785980225,
      "eval_physics_mean_token_accuracy": 0.8193980932235718,
      "eval_physics_num_tokens": 16664690.0,
      "eval_physics_runtime": 41.0989,
      "eval_physics_samples_per_second": 12.166,
      "eval_physics_steps_per_second": 3.041,
      "step": 500
    }
  ],
  "logging_steps": 10,
  "max_steps": 10000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 16,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.238007550370562e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}