{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.4,
  "eval_steps": 100,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.4866574574261904,
      "epoch": 0.004,
      "grad_norm": 616.0,
      "learning_rate": 1.8e-07,
      "loss": 33.2311,
      "mean_token_accuracy": 0.6604987328872085,
      "num_tokens": 51831.0,
      "step": 10
    },
    {
      "entropy": 0.4859781404957175,
      "epoch": 0.008,
      "grad_norm": 676.0,
      "learning_rate": 3.8e-07,
      "loss": 33.1415,
      "mean_token_accuracy": 0.6665693156421184,
      "num_tokens": 103172.0,
      "step": 20
    },
    {
      "entropy": 0.5029121663421392,
      "epoch": 0.012,
      "grad_norm": 824.0,
      "learning_rate": 5.800000000000001e-07,
      "loss": 34.7849,
      "mean_token_accuracy": 0.6541365889832378,
      "num_tokens": 147915.0,
      "step": 30
    },
    {
      "entropy": 0.5095210419967771,
      "epoch": 0.016,
      "grad_norm": 588.0,
      "learning_rate": 7.8e-07,
      "loss": 33.2642,
      "mean_token_accuracy": 0.6610883103683591,
      "num_tokens": 195524.0,
      "step": 40
    },
    {
      "entropy": 0.5093817379325628,
      "epoch": 0.02,
      "grad_norm": 544.0,
      "learning_rate": 9.800000000000001e-07,
      "loss": 32.1157,
      "mean_token_accuracy": 0.6604763394221663,
      "num_tokens": 243747.0,
      "step": 50
    },
    {
      "entropy": 0.5527876092121005,
      "epoch": 0.024,
      "grad_norm": 628.0,
      "learning_rate": 1.1800000000000001e-06,
      "loss": 31.9125,
      "mean_token_accuracy": 0.6578346157446504,
      "num_tokens": 296999.0,
      "step": 60
    },
    {
      "entropy": 0.5809885170310736,
      "epoch": 0.028,
      "grad_norm": 498.0,
      "learning_rate": 1.3800000000000001e-06,
      "loss": 32.3494,
      "mean_token_accuracy": 0.6499254325404763,
      "num_tokens": 342114.0,
      "step": 70
    },
    {
      "entropy": 0.6243463078513741,
      "epoch": 0.032,
      "grad_norm": 430.0,
      "learning_rate": 1.5800000000000001e-06,
      "loss": 29.4405,
      "mean_token_accuracy": 0.6568279277533293,
      "num_tokens": 390903.0,
      "step": 80
    },
    {
      "entropy": 0.7070688692852855,
      "epoch": 0.036,
      "grad_norm": 422.0,
      "learning_rate": 1.7800000000000001e-06,
      "loss": 27.9455,
      "mean_token_accuracy": 0.6594765370711684,
      "num_tokens": 440359.0,
      "step": 90
    },
    {
      "entropy": 0.7858394492417574,
      "epoch": 0.04,
      "grad_norm": 334.0,
      "learning_rate": 1.98e-06,
      "loss": 27.3289,
      "mean_token_accuracy": 0.6568881323561072,
      "num_tokens": 488481.0,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_biology_entropy": 0.8150426388680935,
      "eval_biology_loss": 1.601250410079956,
      "eval_biology_mean_token_accuracy": 0.6667410479187965,
      "eval_biology_num_tokens": 488481.0,
      "eval_biology_runtime": 34.4305,
      "eval_biology_samples_per_second": 14.522,
      "eval_biology_steps_per_second": 14.522,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_chemistry_entropy": 0.42218632949888707,
      "eval_chemistry_loss": 0.7587577700614929,
      "eval_chemistry_mean_token_accuracy": 0.8110456941127777,
      "eval_chemistry_num_tokens": 488481.0,
      "eval_chemistry_runtime": 39.1702,
      "eval_chemistry_samples_per_second": 12.765,
      "eval_chemistry_steps_per_second": 12.765,
      "step": 100
    },
    {
      "entropy": 0.8809404099360109,
      "epoch": 0.044,
      "grad_norm": 246.0,
      "learning_rate": 2.1800000000000003e-06,
      "loss": 25.34,
      "mean_token_accuracy": 0.6623364068567753,
      "num_tokens": 537471.0,
      "step": 110
    },
    {
      "entropy": 0.9853538589552044,
      "epoch": 0.048,
      "grad_norm": 228.0,
      "learning_rate": 2.38e-06,
      "loss": 23.768,
      "mean_token_accuracy": 0.6667981075122953,
      "num_tokens": 589868.0,
      "step": 120
    },
    {
      "entropy": 1.1170461051166058,
      "epoch": 0.052,
      "grad_norm": 171.0,
      "learning_rate": 2.5800000000000003e-06,
      "loss": 23.5554,
      "mean_token_accuracy": 0.6613909218460321,
      "num_tokens": 639512.0,
      "step": 130
    },
    {
      "entropy": 1.1676542518660427,
      "epoch": 0.056,
      "grad_norm": 151.0,
      "learning_rate": 2.7800000000000005e-06,
      "loss": 22.6768,
      "mean_token_accuracy": 0.6657767463475466,
      "num_tokens": 690754.0,
      "step": 140
    },
    {
      "entropy": 1.2725389588624239,
      "epoch": 0.06,
      "grad_norm": 132.0,
      "learning_rate": 2.9800000000000003e-06,
      "loss": 23.1722,
      "mean_token_accuracy": 0.6627637758851052,
      "num_tokens": 737075.0,
      "step": 150
    },
    {
      "entropy": 1.258068286627531,
      "epoch": 0.064,
      "grad_norm": 142.0,
      "learning_rate": 3.1800000000000005e-06,
      "loss": 21.9045,
      "mean_token_accuracy": 0.6725248593837023,
      "num_tokens": 791190.0,
      "step": 160
    },
    {
      "entropy": 1.306521786376834,
      "epoch": 0.068,
      "grad_norm": 112.0,
      "learning_rate": 3.3800000000000007e-06,
      "loss": 22.1444,
      "mean_token_accuracy": 0.6677160531282424,
      "num_tokens": 841344.0,
      "step": 170
    },
    {
      "entropy": 1.2990117866545916,
      "epoch": 0.072,
      "grad_norm": 141.0,
      "learning_rate": 3.58e-06,
      "loss": 22.1479,
      "mean_token_accuracy": 0.6692137856036424,
      "num_tokens": 890075.0,
      "step": 180
    },
    {
      "entropy": 1.2258288636803627,
      "epoch": 0.076,
      "grad_norm": 99.0,
      "learning_rate": 3.7800000000000002e-06,
      "loss": 20.8997,
      "mean_token_accuracy": 0.6792997987940907,
      "num_tokens": 940599.0,
      "step": 190
    },
    {
      "entropy": 1.2281328312121331,
      "epoch": 0.08,
      "grad_norm": 123.0,
      "learning_rate": 3.980000000000001e-06,
      "loss": 20.9624,
      "mean_token_accuracy": 0.6809809722006321,
      "num_tokens": 989860.0,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_biology_entropy": 1.2447693885564803,
      "eval_biology_loss": 1.31557035446167,
      "eval_biology_mean_token_accuracy": 0.6810677705407142,
      "eval_biology_num_tokens": 989860.0,
      "eval_biology_runtime": 34.2892,
      "eval_biology_samples_per_second": 14.582,
      "eval_biology_steps_per_second": 14.582,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_chemistry_entropy": 0.6061330872625112,
      "eval_chemistry_loss": 0.6521626710891724,
      "eval_chemistry_mean_token_accuracy": 0.8165175079107284,
      "eval_chemistry_num_tokens": 989860.0,
      "eval_chemistry_runtime": 39.0885,
      "eval_chemistry_samples_per_second": 12.791,
      "eval_chemistry_steps_per_second": 12.791,
      "step": 200
    },
    {
      "entropy": 1.2204063970595598,
      "epoch": 0.084,
      "grad_norm": 99.5,
      "learning_rate": 4.18e-06,
      "loss": 20.4735,
      "mean_token_accuracy": 0.688131432235241,
      "num_tokens": 1041199.0,
      "step": 210
    },
    {
      "entropy": 1.3349104586988687,
      "epoch": 0.088,
      "grad_norm": 111.5,
      "learning_rate": 4.38e-06,
      "loss": 22.5142,
      "mean_token_accuracy": 0.6627503650262951,
      "num_tokens": 1087099.0,
      "step": 220
    },
    {
      "entropy": 1.2557024333626032,
      "epoch": 0.092,
      "grad_norm": 96.5,
      "learning_rate": 4.58e-06,
      "loss": 20.9139,
      "mean_token_accuracy": 0.6778965264558792,
      "num_tokens": 1135090.0,
      "step": 230
    },
    {
      "entropy": 1.280203009955585,
      "epoch": 0.096,
      "grad_norm": 105.0,
      "learning_rate": 4.78e-06,
      "loss": 21.3579,
      "mean_token_accuracy": 0.6716440364718437,
      "num_tokens": 1181777.0,
      "step": 240
    },
    {
      "entropy": 1.276487385854125,
      "epoch": 0.1,
      "grad_norm": 100.5,
      "learning_rate": 4.980000000000001e-06,
      "loss": 20.8968,
      "mean_token_accuracy": 0.6772379480302334,
      "num_tokens": 1229242.0,
      "step": 250
    },
    {
      "entropy": 1.2585501682013274,
      "epoch": 0.104,
      "grad_norm": 85.0,
      "learning_rate": 5.18e-06,
      "loss": 20.7316,
      "mean_token_accuracy": 0.679420518502593,
      "num_tokens": 1283881.0,
      "step": 260
    },
    {
      "entropy": 1.248391578346491,
      "epoch": 0.108,
      "grad_norm": 91.0,
      "learning_rate": 5.380000000000001e-06,
      "loss": 20.4527,
      "mean_token_accuracy": 0.682051457837224,
      "num_tokens": 1332507.0,
      "step": 270
    },
    {
      "entropy": 1.2360890079289675,
      "epoch": 0.112,
      "grad_norm": 84.5,
      "learning_rate": 5.580000000000001e-06,
      "loss": 20.2916,
      "mean_token_accuracy": 0.6859552793204784,
      "num_tokens": 1385513.0,
      "step": 280
    },
    {
      "entropy": 1.2920903146266938,
      "epoch": 0.116,
      "grad_norm": 90.5,
      "learning_rate": 5.78e-06,
      "loss": 20.9268,
      "mean_token_accuracy": 0.673469840362668,
      "num_tokens": 1432251.0,
      "step": 290
    },
    {
      "entropy": 1.2169338904321194,
      "epoch": 0.12,
      "grad_norm": 87.0,
      "learning_rate": 5.98e-06,
      "loss": 19.6752,
      "mean_token_accuracy": 0.6875041898339986,
      "num_tokens": 1483667.0,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_biology_entropy": 1.2245318589806558,
      "eval_biology_loss": 1.2560091018676758,
      "eval_biology_mean_token_accuracy": 0.6886654210686683,
      "eval_biology_num_tokens": 1483667.0,
      "eval_biology_runtime": 34.2037,
      "eval_biology_samples_per_second": 14.618,
      "eval_biology_steps_per_second": 14.618,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_chemistry_entropy": 0.6325840967297554,
      "eval_chemistry_loss": 0.6409750580787659,
      "eval_chemistry_mean_token_accuracy": 0.8179453414678574,
      "eval_chemistry_num_tokens": 1483667.0,
      "eval_chemistry_runtime": 39.0169,
      "eval_chemistry_samples_per_second": 12.815,
      "eval_chemistry_steps_per_second": 12.815,
      "step": 300
    },
    {
      "entropy": 1.2429257260635496,
      "epoch": 0.124,
      "grad_norm": 79.5,
      "learning_rate": 6.18e-06,
      "loss": 20.6172,
      "mean_token_accuracy": 0.6801770402118563,
      "num_tokens": 1534286.0,
      "step": 310
    },
    {
      "entropy": 1.2787830816581844,
      "epoch": 0.128,
      "grad_norm": 84.5,
      "learning_rate": 6.380000000000001e-06,
      "loss": 20.5543,
      "mean_token_accuracy": 0.6807437939569354,
      "num_tokens": 1582368.0,
      "step": 320
    },
    {
      "entropy": 1.2588950738310813,
      "epoch": 0.132,
      "grad_norm": 89.0,
      "learning_rate": 6.5800000000000005e-06,
      "loss": 20.3303,
      "mean_token_accuracy": 0.6814694199711084,
      "num_tokens": 1630878.0,
      "step": 330
    },
    {
      "entropy": 1.2704457983374595,
      "epoch": 0.136,
      "grad_norm": 83.0,
      "learning_rate": 6.780000000000001e-06,
      "loss": 20.6418,
      "mean_token_accuracy": 0.6803831990808249,
      "num_tokens": 1677060.0,
      "step": 340
    },
    {
      "entropy": 1.2321893341839314,
      "epoch": 0.14,
      "grad_norm": 87.5,
      "learning_rate": 6.98e-06,
      "loss": 19.8202,
      "mean_token_accuracy": 0.6891451958566904,
      "num_tokens": 1725186.0,
      "step": 350
    },
    {
      "entropy": 1.2445756699889898,
      "epoch": 0.144,
      "grad_norm": 92.0,
      "learning_rate": 7.180000000000001e-06,
      "loss": 20.2527,
      "mean_token_accuracy": 0.6785595057532191,
      "num_tokens": 1773764.0,
      "step": 360
    },
    {
      "entropy": 1.246081338264048,
      "epoch": 0.148,
      "grad_norm": 82.0,
      "learning_rate": 7.3800000000000005e-06,
      "loss": 20.08,
      "mean_token_accuracy": 0.6857875902205706,
      "num_tokens": 1823706.0,
      "step": 370
    },
    {
      "entropy": 1.2665387976914644,
      "epoch": 0.152,
      "grad_norm": 95.0,
      "learning_rate": 7.58e-06,
      "loss": 20.6091,
      "mean_token_accuracy": 0.6813197182491422,
      "num_tokens": 1873167.0,
      "step": 380
    },
    {
      "entropy": 1.2548780847340821,
      "epoch": 0.156,
      "grad_norm": 85.0,
      "learning_rate": 7.78e-06,
      "loss": 20.406,
      "mean_token_accuracy": 0.6822501845657826,
      "num_tokens": 1921484.0,
      "step": 390
    },
    {
      "entropy": 1.2407029550522566,
      "epoch": 0.16,
      "grad_norm": 89.0,
      "learning_rate": 7.980000000000002e-06,
      "loss": 19.8013,
      "mean_token_accuracy": 0.6853796949610114,
      "num_tokens": 1970077.0,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_biology_entropy": 1.214753457427025,
      "eval_biology_loss": 1.2285655736923218,
      "eval_biology_mean_token_accuracy": 0.6923919150233269,
      "eval_biology_num_tokens": 1970077.0,
      "eval_biology_runtime": 34.3823,
      "eval_biology_samples_per_second": 14.542,
      "eval_biology_steps_per_second": 14.542,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_chemistry_entropy": 0.6530047204196453,
      "eval_chemistry_loss": 0.644757866859436,
      "eval_chemistry_mean_token_accuracy": 0.8174828389883041,
      "eval_chemistry_num_tokens": 1970077.0,
      "eval_chemistry_runtime": 39.0828,
      "eval_chemistry_samples_per_second": 12.793,
      "eval_chemistry_steps_per_second": 12.793,
      "step": 400
    },
    {
      "entropy": 1.2693706121295691,
      "epoch": 0.164,
      "grad_norm": 75.0,
      "learning_rate": 8.18e-06,
      "loss": 20.5934,
      "mean_token_accuracy": 0.6737196171656251,
      "num_tokens": 2019053.0,
      "step": 410
    },
    {
      "entropy": 1.2044812317937612,
      "epoch": 0.168,
      "grad_norm": 72.5,
      "learning_rate": 8.380000000000001e-06,
      "loss": 19.2447,
      "mean_token_accuracy": 0.6942890204489232,
      "num_tokens": 2066815.0,
      "step": 420
    },
    {
      "entropy": 1.219953553378582,
      "epoch": 0.172,
      "grad_norm": 76.5,
      "learning_rate": 8.580000000000001e-06,
      "loss": 20.0236,
      "mean_token_accuracy": 0.6850357944145798,
      "num_tokens": 2116402.0,
      "step": 430
    },
    {
      "entropy": 1.1823513347655534,
      "epoch": 0.176,
      "grad_norm": 74.0,
      "learning_rate": 8.78e-06,
      "loss": 19.1286,
      "mean_token_accuracy": 0.6948613569140434,
      "num_tokens": 2168354.0,
      "step": 440
    },
    {
      "entropy": 1.2351123623549938,
      "epoch": 0.18,
      "grad_norm": 75.0,
      "learning_rate": 8.98e-06,
      "loss": 19.847,
      "mean_token_accuracy": 0.6850900270044804,
      "num_tokens": 2219158.0,
      "step": 450
    },
    {
      "entropy": 1.2182675819844007,
      "epoch": 0.184,
      "grad_norm": 83.0,
      "learning_rate": 9.180000000000002e-06,
      "loss": 19.9054,
      "mean_token_accuracy": 0.6874489855021239,
      "num_tokens": 2270617.0,
      "step": 460
    },
    {
      "entropy": 1.2002981722354888,
      "epoch": 0.188,
      "grad_norm": 74.5,
      "learning_rate": 9.38e-06,
      "loss": 18.9485,
      "mean_token_accuracy": 0.6958082247525453,
      "num_tokens": 2322034.0,
      "step": 470
    },
    {
      "entropy": 1.2613376142457127,
      "epoch": 0.192,
      "grad_norm": 98.5,
      "learning_rate": 9.58e-06,
      "loss": 20.7735,
      "mean_token_accuracy": 0.6739435123279691,
      "num_tokens": 2365822.0,
      "step": 480
    },
    {
      "entropy": 1.205817231722176,
      "epoch": 0.196,
      "grad_norm": 89.5,
      "learning_rate": 9.780000000000001e-06,
      "loss": 19.1434,
      "mean_token_accuracy": 0.6935921791940928,
      "num_tokens": 2413853.0,
      "step": 490
    },
    {
      "entropy": 1.2515719644725323,
      "epoch": 0.2,
      "grad_norm": 90.5,
      "learning_rate": 9.980000000000001e-06,
      "loss": 20.4938,
      "mean_token_accuracy": 0.6777064654976129,
      "num_tokens": 2459876.0,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_biology_entropy": 1.1872086516022682,
      "eval_biology_loss": 1.2085970640182495,
      "eval_biology_mean_token_accuracy": 0.6949169620871544,
      "eval_biology_num_tokens": 2459876.0,
      "eval_biology_runtime": 34.4,
      "eval_biology_samples_per_second": 14.535,
      "eval_biology_steps_per_second": 14.535,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_chemistry_entropy": 0.6488757311999798,
      "eval_chemistry_loss": 0.6450773477554321,
      "eval_chemistry_mean_token_accuracy": 0.8178821201324463,
      "eval_chemistry_num_tokens": 2459876.0,
      "eval_chemistry_runtime": 39.2328,
      "eval_chemistry_samples_per_second": 12.744,
      "eval_chemistry_steps_per_second": 12.744,
      "step": 500
    },
    {
      "entropy": 1.2209952611476182,
      "epoch": 0.204,
      "grad_norm": 73.0,
      "learning_rate": 1.018e-05,
      "loss": 19.6626,
      "mean_token_accuracy": 0.6881283605471253,
      "num_tokens": 2509187.0,
      "step": 510
    },
    {
      "entropy": 1.2497611749917268,
      "epoch": 0.208,
      "grad_norm": 85.5,
      "learning_rate": 1.038e-05,
      "loss": 19.9518,
      "mean_token_accuracy": 0.6810324152931571,
      "num_tokens": 2558762.0,
      "step": 520
    },
    {
      "entropy": 1.1757056951522826,
      "epoch": 0.212,
      "grad_norm": 78.0,
      "learning_rate": 1.0580000000000002e-05,
      "loss": 19.3754,
      "mean_token_accuracy": 0.6935973027721047,
      "num_tokens": 2608616.0,
      "step": 530
    },
    {
      "entropy": 1.2561812430620194,
      "epoch": 0.216,
      "grad_norm": 83.0,
      "learning_rate": 1.0780000000000002e-05,
      "loss": 20.0513,
      "mean_token_accuracy": 0.6874226013198494,
      "num_tokens": 2654953.0,
      "step": 540
    },
    {
      "entropy": 1.2020427081733942,
      "epoch": 0.22,
      "grad_norm": 78.5,
      "learning_rate": 1.0980000000000002e-05,
      "loss": 19.5412,
      "mean_token_accuracy": 0.688438569009304,
      "num_tokens": 2704686.0,
      "step": 550
    },
    {
      "entropy": 1.205025901272893,
      "epoch": 0.224,
      "grad_norm": 86.5,
      "learning_rate": 1.1180000000000001e-05,
      "loss": 19.7851,
      "mean_token_accuracy": 0.6885192114859819,
      "num_tokens": 2755347.0,
      "step": 560
    },
    {
      "entropy": 1.2252822391688825,
      "epoch": 0.228,
      "grad_norm": 74.5,
      "learning_rate": 1.138e-05,
      "loss": 19.3164,
      "mean_token_accuracy": 0.6909766353666782,
      "num_tokens": 2804340.0,
      "step": 570
    },
    {
      "entropy": 1.1826427452266217,
      "epoch": 0.232,
      "grad_norm": 80.0,
      "learning_rate": 1.1580000000000001e-05,
      "loss": 19.3402,
      "mean_token_accuracy": 0.6917916279286146,
      "num_tokens": 2852298.0,
      "step": 580
    },
    {
      "entropy": 1.1695974227041006,
      "epoch": 0.236,
      "grad_norm": 89.0,
      "learning_rate": 1.178e-05,
      "loss": 19.0702,
      "mean_token_accuracy": 0.6989134319126606,
      "num_tokens": 2901227.0,
      "step": 590
    },
    {
      "entropy": 1.2379697296768426,
      "epoch": 0.24,
      "grad_norm": 92.5,
      "learning_rate": 1.198e-05,
      "loss": 19.6279,
      "mean_token_accuracy": 0.6916081016883254,
      "num_tokens": 2947346.0,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_biology_entropy": 1.1614825129508972,
      "eval_biology_loss": 1.197817087173462,
      "eval_biology_mean_token_accuracy": 0.6965834299921989,
      "eval_biology_num_tokens": 2947346.0,
      "eval_biology_runtime": 34.3297,
      "eval_biology_samples_per_second": 14.565,
      "eval_biology_steps_per_second": 14.565,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_chemistry_entropy": 0.6421602384448052,
      "eval_chemistry_loss": 0.6457399725914001,
      "eval_chemistry_mean_token_accuracy": 0.8177337770462036,
      "eval_chemistry_num_tokens": 2947346.0,
      "eval_chemistry_runtime": 39.0493,
      "eval_chemistry_samples_per_second": 12.804,
      "eval_chemistry_steps_per_second": 12.804,
      "step": 600
    },
    {
      "entropy": 1.2068689458072186,
      "epoch": 0.244,
      "grad_norm": 83.0,
      "learning_rate": 1.218e-05,
      "loss": 19.9725,
      "mean_token_accuracy": 0.6830457381904125,
      "num_tokens": 2996175.0,
      "step": 610
    },
    {
      "entropy": 1.2800738830119371,
      "epoch": 0.248,
      "grad_norm": 77.5,
      "learning_rate": 1.2380000000000002e-05,
      "loss": 20.1671,
      "mean_token_accuracy": 0.6834735475480557,
      "num_tokens": 3043128.0,
      "step": 620
    },
    {
      "entropy": 1.190559478290379,
      "epoch": 0.252,
      "grad_norm": 79.5,
      "learning_rate": 1.2580000000000002e-05,
      "loss": 19.3606,
      "mean_token_accuracy": 0.6899204226210713,
      "num_tokens": 3090827.0,
      "step": 630
    },
    {
      "entropy": 1.1722685299813747,
      "epoch": 0.256,
      "grad_norm": 81.5,
      "learning_rate": 1.2780000000000001e-05,
      "loss": 18.8652,
      "mean_token_accuracy": 0.6963184822350741,
      "num_tokens": 3139957.0,
      "step": 640
    },
    {
      "entropy": 1.2144603542983532,
      "epoch": 0.26,
      "grad_norm": 82.0,
      "learning_rate": 1.2980000000000001e-05,
      "loss": 19.7174,
      "mean_token_accuracy": 0.685292916931212,
      "num_tokens": 3183188.0,
      "step": 650
    },
    {
      "entropy": 1.2338291574269533,
      "epoch": 0.264,
      "grad_norm": 80.0,
      "learning_rate": 1.3180000000000001e-05,
      "loss": 19.7757,
      "mean_token_accuracy": 0.6882328724488616,
      "num_tokens": 3232159.0,
      "step": 660
    },
    {
      "entropy": 1.1896044220775366,
      "epoch": 0.268,
      "grad_norm": 79.5,
      "learning_rate": 1.3380000000000002e-05,
      "loss": 19.1149,
      "mean_token_accuracy": 0.6937362670898437,
      "num_tokens": 3283908.0,
      "step": 670
    },
    {
      "entropy": 1.1706442803144455,
      "epoch": 0.272,
      "grad_norm": 98.0,
      "learning_rate": 1.3580000000000002e-05,
      "loss": 19.1832,
      "mean_token_accuracy": 0.6947279252111912,
      "num_tokens": 3335951.0,
      "step": 680
    },
    {
      "entropy": 1.2103422859683632,
      "epoch": 0.276,
      "grad_norm": 88.0,
      "learning_rate": 1.378e-05,
      "loss": 19.5378,
      "mean_token_accuracy": 0.6888058576732874,
      "num_tokens": 3387755.0,
      "step": 690
    },
    {
      "entropy": 1.1900839667767287,
      "epoch": 0.28,
      "grad_norm": 90.0,
      "learning_rate": 1.398e-05,
      "loss": 19.3154,
      "mean_token_accuracy": 0.6907492563128471,
      "num_tokens": 3437213.0,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_biology_entropy": 1.1723266056776047,
      "eval_biology_loss": 1.1935571432113647,
      "eval_biology_mean_token_accuracy": 0.6974302505254746,
      "eval_biology_num_tokens": 3437213.0,
      "eval_biology_runtime": 34.4282,
      "eval_biology_samples_per_second": 14.523,
      "eval_biology_steps_per_second": 14.523,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_chemistry_entropy": 0.6567539082467556,
      "eval_chemistry_loss": 0.6476752758026123,
      "eval_chemistry_mean_token_accuracy": 0.8169361530542374,
      "eval_chemistry_num_tokens": 3437213.0,
      "eval_chemistry_runtime": 39.1281,
      "eval_chemistry_samples_per_second": 12.779,
      "eval_chemistry_steps_per_second": 12.779,
      "step": 700
    },
    {
      "entropy": 1.1601122960448265,
      "epoch": 0.284,
      "grad_norm": 70.0,
      "learning_rate": 1.418e-05,
      "loss": 18.4856,
      "mean_token_accuracy": 0.7002454694360495,
      "num_tokens": 3489088.0,
      "step": 710
    },
    {
      "entropy": 1.1836338724941016,
      "epoch": 0.288,
      "grad_norm": 74.0,
      "learning_rate": 1.4380000000000001e-05,
      "loss": 19.4318,
      "mean_token_accuracy": 0.6893038578331471,
      "num_tokens": 3539731.0,
      "step": 720
    },
    {
      "entropy": 1.215116088837385,
      "epoch": 0.292,
      "grad_norm": 76.5,
      "learning_rate": 1.4580000000000001e-05,
      "loss": 19.2441,
      "mean_token_accuracy": 0.6934854052960873,
      "num_tokens": 3586433.0,
      "step": 730
    },
    {
      "entropy": 1.2150515001267195,
      "epoch": 0.296,
      "grad_norm": 87.5,
      "learning_rate": 1.478e-05,
      "loss": 20.0166,
      "mean_token_accuracy": 0.6854536816477775,
      "num_tokens": 3633448.0,
      "step": 740
    },
    {
      "entropy": 1.1540303017944098,
      "epoch": 0.3,
      "grad_norm": 77.5,
      "learning_rate": 1.498e-05,
      "loss": 18.4485,
      "mean_token_accuracy": 0.705385484918952,
      "num_tokens": 3685080.0,
      "step": 750
    },
    {
      "entropy": 1.1504614014178514,
      "epoch": 0.304,
      "grad_norm": 79.0,
      "learning_rate": 1.5180000000000002e-05,
      "loss": 18.9317,
      "mean_token_accuracy": 0.6943564370274544,
      "num_tokens": 3733488.0,
      "step": 760
    },
    {
      "entropy": 1.227951274625957,
      "epoch": 0.308,
      "grad_norm": 81.0,
      "learning_rate": 1.5380000000000002e-05,
      "loss": 19.6245,
      "mean_token_accuracy": 0.69070417098701,
      "num_tokens": 3779708.0,
      "step": 770
    },
    {
      "entropy": 1.1914822444319726,
      "epoch": 0.312,
      "grad_norm": 91.5,
      "learning_rate": 1.5580000000000003e-05,
      "loss": 19.3053,
      "mean_token_accuracy": 0.691599304229021,
      "num_tokens": 3825212.0,
      "step": 780
    },
    {
      "entropy": 1.2374459221959113,
      "epoch": 0.316,
      "grad_norm": 90.5,
      "learning_rate": 1.578e-05,
      "loss": 20.1615,
      "mean_token_accuracy": 0.6793428828939796,
      "num_tokens": 3870912.0,
      "step": 790
    },
    {
      "entropy": 1.204637792892754,
      "epoch": 0.32,
      "grad_norm": 87.0,
      "learning_rate": 1.5980000000000003e-05,
      "loss": 19.1533,
      "mean_token_accuracy": 0.6921680744737386,
      "num_tokens": 3920545.0,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_biology_entropy": 1.1595962298512459,
      "eval_biology_loss": 1.1908730268478394,
      "eval_biology_mean_token_accuracy": 0.6982393438220024,
      "eval_biology_num_tokens": 3920545.0,
      "eval_biology_runtime": 34.3027,
      "eval_biology_samples_per_second": 14.576,
      "eval_biology_steps_per_second": 14.576,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_chemistry_entropy": 0.6581202555298805,
      "eval_chemistry_loss": 0.6543570756912231,
      "eval_chemistry_mean_token_accuracy": 0.8150564895868302,
      "eval_chemistry_num_tokens": 3920545.0,
      "eval_chemistry_runtime": 39.1221,
      "eval_chemistry_samples_per_second": 12.781,
      "eval_chemistry_steps_per_second": 12.781,
      "step": 800
    },
    {
      "entropy": 1.1894057728350163,
      "epoch": 0.324,
      "grad_norm": 82.0,
      "learning_rate": 1.618e-05,
      "loss": 19.3278,
      "mean_token_accuracy": 0.6910803083330392,
      "num_tokens": 3967499.0,
      "step": 810
    },
    {
      "entropy": 1.1689968170598148,
      "epoch": 0.328,
      "grad_norm": 75.5,
      "learning_rate": 1.638e-05,
      "loss": 18.6792,
      "mean_token_accuracy": 0.6996502809226512,
      "num_tokens": 4016673.0,
      "step": 820
    },
    {
      "entropy": 1.1961608117446303,
      "epoch": 0.332,
      "grad_norm": 88.5,
      "learning_rate": 1.658e-05,
      "loss": 19.6175,
      "mean_token_accuracy": 0.6912782397121191,
      "num_tokens": 4063859.0,
      "step": 830
    },
    {
      "entropy": 1.2022534670308231,
      "epoch": 0.336,
      "grad_norm": 75.5,
      "learning_rate": 1.6780000000000002e-05,
      "loss": 19.3189,
      "mean_token_accuracy": 0.6922560017555952,
      "num_tokens": 4114077.0,
      "step": 840
    },
    {
      "entropy": 1.2139138951897621,
      "epoch": 0.34,
      "grad_norm": 80.0,
      "learning_rate": 1.698e-05,
      "loss": 19.6407,
      "mean_token_accuracy": 0.6891672356054187,
      "num_tokens": 4160168.0,
      "step": 850
    },
    {
      "entropy": 1.1922080295160413,
      "epoch": 0.344,
      "grad_norm": 76.5,
      "learning_rate": 1.718e-05,
      "loss": 19.3527,
      "mean_token_accuracy": 0.6934164728969335,
      "num_tokens": 4210403.0,
      "step": 860
    },
    {
      "entropy": 1.177280587144196,
      "epoch": 0.348,
      "grad_norm": 76.0,
      "learning_rate": 1.7380000000000003e-05,
      "loss": 18.7411,
      "mean_token_accuracy": 0.6976979669183493,
      "num_tokens": 4256944.0,
      "step": 870
    },
    {
      "entropy": 1.2034492146223783,
      "epoch": 0.352,
      "grad_norm": 77.5,
      "learning_rate": 1.758e-05,
      "loss": 19.6262,
      "mean_token_accuracy": 0.6897061046212911,
      "num_tokens": 4306949.0,
      "step": 880
    },
    {
      "entropy": 1.2024015821516514,
      "epoch": 0.356,
      "grad_norm": 70.5,
      "learning_rate": 1.7780000000000003e-05,
      "loss": 19.4871,
      "mean_token_accuracy": 0.6887454513460398,
      "num_tokens": 4355070.0,
      "step": 890
    },
    {
      "entropy": 1.1949962507933378,
      "epoch": 0.36,
      "grad_norm": 65.0,
      "learning_rate": 1.798e-05,
      "loss": 19.252,
      "mean_token_accuracy": 0.6892035495489836,
      "num_tokens": 4406104.0,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_biology_entropy": 1.149014304101467,
      "eval_biology_loss": 1.191479206085205,
      "eval_biology_mean_token_accuracy": 0.6979464264512062,
      "eval_biology_num_tokens": 4406104.0,
      "eval_biology_runtime": 34.5339,
      "eval_biology_samples_per_second": 14.479,
      "eval_biology_steps_per_second": 14.479,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_chemistry_entropy": 0.6591544916331769,
      "eval_chemistry_loss": 0.6548386812210083,
      "eval_chemistry_mean_token_accuracy": 0.8150800436735153,
      "eval_chemistry_num_tokens": 4406104.0,
      "eval_chemistry_runtime": 39.1161,
      "eval_chemistry_samples_per_second": 12.782,
      "eval_chemistry_steps_per_second": 12.782,
      "step": 900
    },
    {
      "entropy": 1.210719463415444,
      "epoch": 0.364,
      "grad_norm": 77.5,
      "learning_rate": 1.8180000000000002e-05,
      "loss": 19.6229,
      "mean_token_accuracy": 0.6920540288090706,
      "num_tokens": 4455015.0,
      "step": 910
    },
    {
      "entropy": 1.1651359049603343,
      "epoch": 0.368,
      "grad_norm": 85.0,
      "learning_rate": 1.8380000000000004e-05,
      "loss": 18.965,
      "mean_token_accuracy": 0.6945374647155405,
      "num_tokens": 4504001.0,
      "step": 920
    },
    {
      "entropy": 1.2309420805424451,
      "epoch": 0.372,
      "grad_norm": 79.0,
      "learning_rate": 1.858e-05,
      "loss": 19.4645,
      "mean_token_accuracy": 0.6926089435815811,
      "num_tokens": 4550269.0,
      "step": 930
    },
    {
      "entropy": 1.1543804241344333,
      "epoch": 0.376,
      "grad_norm": 83.5,
      "learning_rate": 1.878e-05,
      "loss": 18.9949,
      "mean_token_accuracy": 0.695966462418437,
      "num_tokens": 4597559.0,
      "step": 940
    },
    {
      "entropy": 1.1860109277069568,
      "epoch": 0.38,
      "grad_norm": 88.0,
      "learning_rate": 1.898e-05,
      "loss": 19.0122,
      "mean_token_accuracy": 0.6945275101810694,
      "num_tokens": 4646903.0,
      "step": 950
    },
    {
      "entropy": 1.1854952551424502,
      "epoch": 0.384,
      "grad_norm": 72.0,
      "learning_rate": 1.918e-05,
      "loss": 19.3612,
      "mean_token_accuracy": 0.6901288989931345,
      "num_tokens": 4693812.0,
      "step": 960
    },
    {
      "entropy": 1.2153874900192023,
      "epoch": 0.388,
      "grad_norm": 76.0,
      "learning_rate": 1.938e-05,
      "loss": 19.4325,
      "mean_token_accuracy": 0.6911413293331862,
      "num_tokens": 4740309.0,
      "step": 970
    },
    {
      "entropy": 1.212491791881621,
      "epoch": 0.392,
      "grad_norm": 74.0,
      "learning_rate": 1.9580000000000002e-05,
      "loss": 19.43,
      "mean_token_accuracy": 0.6912798311561346,
      "num_tokens": 4790965.0,
      "step": 980
    },
    {
      "entropy": 1.1938600786030293,
      "epoch": 0.396,
      "grad_norm": 77.0,
      "learning_rate": 1.978e-05,
      "loss": 19.4531,
      "mean_token_accuracy": 0.6905827283859253,
      "num_tokens": 4839186.0,
      "step": 990
    },
    {
      "entropy": 1.2035906264558434,
      "epoch": 0.4,
      "grad_norm": 82.5,
      "learning_rate": 1.9980000000000002e-05,
      "loss": 19.4249,
      "mean_token_accuracy": 0.6903344210237264,
      "num_tokens": 4887094.0,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_biology_entropy": 1.1641007453799248,
      "eval_biology_loss": 1.1950935125350952,
      "eval_biology_mean_token_accuracy": 0.6969818140268326,
      "eval_biology_num_tokens": 4887094.0,
      "eval_biology_runtime": 34.3544,
      "eval_biology_samples_per_second": 14.554,
      "eval_biology_steps_per_second": 14.554,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_chemistry_entropy": 0.6638510838449001,
      "eval_chemistry_loss": 0.662416934967041,
      "eval_chemistry_mean_token_accuracy": 0.8127187831401825,
      "eval_chemistry_num_tokens": 4887094.0,
      "eval_chemistry_runtime": 38.9213,
      "eval_chemistry_samples_per_second": 12.846,
      "eval_chemistry_steps_per_second": 12.846,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 10000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 3.277009915508544e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}