{
  "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": 0.6712883992120624,
      "epoch": 0.004,
      "grad_norm": 304.0,
      "learning_rate": 6.000000000000001e-07,
      "loss": 28.4778,
      "mean_token_accuracy": 0.7057144690304995,
      "num_tokens": 89176.0,
      "step": 10
    },
    {
      "entropy": 0.7533201891928911,
      "epoch": 0.008,
      "grad_norm": 270.0,
      "learning_rate": 1.2666666666666669e-06,
      "loss": 30.5704,
      "mean_token_accuracy": 0.6818637173622847,
      "num_tokens": 173886.0,
      "step": 20
    },
    {
      "entropy": 0.7343124819919467,
      "epoch": 0.012,
      "grad_norm": 242.0,
      "learning_rate": 1.9333333333333336e-06,
      "loss": 28.2596,
      "mean_token_accuracy": 0.6965057767927647,
      "num_tokens": 258970.0,
      "step": 30
    },
    {
      "entropy": 0.7956574158743024,
      "epoch": 0.016,
      "grad_norm": 201.0,
      "learning_rate": 2.6e-06,
      "loss": 28.9995,
      "mean_token_accuracy": 0.6894745834171772,
      "num_tokens": 338352.0,
      "step": 40
    },
    {
      "entropy": 0.8114285696297884,
      "epoch": 0.02,
      "grad_norm": 197.0,
      "learning_rate": 3.266666666666667e-06,
      "loss": 26.2241,
      "mean_token_accuracy": 0.7063078790903091,
      "num_tokens": 424243.0,
      "step": 50
    },
    {
      "entropy": 0.8994641156867147,
      "epoch": 0.024,
      "grad_norm": 186.0,
      "learning_rate": 3.9333333333333335e-06,
      "loss": 26.076,
      "mean_token_accuracy": 0.6964696431532502,
      "num_tokens": 505613.0,
      "step": 60
    },
    {
      "entropy": 0.906761129759252,
      "epoch": 0.028,
      "grad_norm": 137.0,
      "learning_rate": 4.600000000000001e-06,
      "loss": 23.8081,
      "mean_token_accuracy": 0.7086012028157711,
      "num_tokens": 584157.0,
      "step": 70
    },
    {
      "entropy": 0.8810091400519013,
      "epoch": 0.032,
      "grad_norm": 74.0,
      "learning_rate": 5.2666666666666665e-06,
      "loss": 20.7997,
      "mean_token_accuracy": 0.7292992439121008,
      "num_tokens": 671193.0,
      "step": 80
    },
    {
      "entropy": 0.9790557857602835,
      "epoch": 0.036,
      "grad_norm": 124.0,
      "learning_rate": 5.933333333333335e-06,
      "loss": 21.1978,
      "mean_token_accuracy": 0.7139894777908922,
      "num_tokens": 753947.0,
      "step": 90
    },
    {
      "entropy": 0.9239323560148478,
      "epoch": 0.04,
      "grad_norm": 69.0,
      "learning_rate": 6.600000000000001e-06,
      "loss": 18.4839,
      "mean_token_accuracy": 0.739598892442882,
      "num_tokens": 835868.0,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_biology_entropy": 1.4848132659196853,
      "eval_biology_loss": 1.9657849073410034,
      "eval_biology_mean_token_accuracy": 0.606338870882988,
      "eval_biology_num_tokens": 835868.0,
      "eval_biology_runtime": 43.6264,
      "eval_biology_samples_per_second": 11.461,
      "eval_biology_steps_per_second": 11.461,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_chemistry_entropy": 1.0286893111467361,
      "eval_chemistry_loss": 1.3152016401290894,
      "eval_chemistry_mean_token_accuracy": 0.7144413039684295,
      "eval_chemistry_num_tokens": 835868.0,
      "eval_chemistry_runtime": 47.0241,
      "eval_chemistry_samples_per_second": 10.633,
      "eval_chemistry_steps_per_second": 10.633,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_math_entropy": 0.8130515841245651,
      "eval_math_loss": 1.1799616813659668,
      "eval_math_mean_token_accuracy": 0.739585277557373,
      "eval_math_num_tokens": 835868.0,
      "eval_math_runtime": 47.1539,
      "eval_math_samples_per_second": 10.604,
      "eval_math_steps_per_second": 10.604,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_physics_entropy": 0.9095318858623505,
      "eval_physics_loss": 1.1475844383239746,
      "eval_physics_mean_token_accuracy": 0.740034140408039,
      "eval_physics_num_tokens": 835868.0,
      "eval_physics_runtime": 52.2922,
      "eval_physics_samples_per_second": 9.562,
      "eval_physics_steps_per_second": 9.562,
      "step": 100
    },
    {
      "entropy": 0.8509915253147483,
      "epoch": 0.044,
      "grad_norm": 56.25,
      "learning_rate": 7.266666666666668e-06,
      "loss": 16.7359,
      "mean_token_accuracy": 0.7565285731106997,
      "num_tokens": 917534.0,
      "step": 110
    },
    {
      "entropy": 0.8670064052566886,
      "epoch": 0.048,
      "grad_norm": 49.75,
      "learning_rate": 7.933333333333334e-06,
      "loss": 16.3406,
      "mean_token_accuracy": 0.7602715287357569,
      "num_tokens": 1001008.0,
      "step": 120
    },
    {
      "entropy": 0.8608247054740786,
      "epoch": 0.052,
      "grad_norm": 44.0,
      "learning_rate": 8.6e-06,
      "loss": 15.9257,
      "mean_token_accuracy": 0.7605812683701515,
      "num_tokens": 1085098.0,
      "step": 130
    },
    {
      "entropy": 0.9198611425235868,
      "epoch": 0.056,
      "grad_norm": 45.75,
      "learning_rate": 9.266666666666667e-06,
      "loss": 16.7546,
      "mean_token_accuracy": 0.7533804941922426,
      "num_tokens": 1166615.0,
      "step": 140
    },
    {
      "entropy": 0.8417589077726007,
      "epoch": 0.06,
      "grad_norm": 34.25,
      "learning_rate": 9.933333333333334e-06,
      "loss": 15.0755,
      "mean_token_accuracy": 0.7657617449760437,
      "num_tokens": 1248387.0,
      "step": 150
    },
    {
      "entropy": 0.9038887821137905,
      "epoch": 0.064,
      "grad_norm": 37.25,
      "learning_rate": 1.0600000000000002e-05,
      "loss": 15.8985,
      "mean_token_accuracy": 0.7547885052859783,
      "num_tokens": 1331208.0,
      "step": 160
    },
    {
      "entropy": 0.8469558507204056,
      "epoch": 0.068,
      "grad_norm": 32.25,
      "learning_rate": 1.1266666666666668e-05,
      "loss": 14.5748,
      "mean_token_accuracy": 0.7721429634839296,
      "num_tokens": 1416841.0,
      "step": 170
    },
    {
      "entropy": 0.8490036362782121,
      "epoch": 0.072,
      "grad_norm": 41.75,
      "learning_rate": 1.1933333333333335e-05,
      "loss": 14.4396,
      "mean_token_accuracy": 0.7727861993014813,
      "num_tokens": 1499449.0,
      "step": 180
    },
    {
      "entropy": 0.8576838463544846,
      "epoch": 0.076,
      "grad_norm": 34.75,
      "learning_rate": 1.2600000000000001e-05,
      "loss": 14.2797,
      "mean_token_accuracy": 0.768672015145421,
      "num_tokens": 1585982.0,
      "step": 190
    },
    {
      "entropy": 0.8711228912696243,
      "epoch": 0.08,
      "grad_norm": 30.625,
      "learning_rate": 1.3266666666666668e-05,
      "loss": 14.4031,
      "mean_token_accuracy": 0.7655902322381735,
      "num_tokens": 1667532.0,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_biology_entropy": 1.4584133907556533,
      "eval_biology_loss": 1.5849124193191528,
      "eval_biology_mean_token_accuracy": 0.6393597082495689,
      "eval_biology_num_tokens": 1667532.0,
      "eval_biology_runtime": 42.746,
      "eval_biology_samples_per_second": 11.697,
      "eval_biology_steps_per_second": 11.697,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_chemistry_entropy": 1.0054012925028801,
      "eval_chemistry_loss": 1.029191255569458,
      "eval_chemistry_mean_token_accuracy": 0.7440219584703446,
      "eval_chemistry_num_tokens": 1667532.0,
      "eval_chemistry_runtime": 47.4172,
      "eval_chemistry_samples_per_second": 10.545,
      "eval_chemistry_steps_per_second": 10.545,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_math_entropy": 0.825192849367857,
      "eval_math_loss": 0.9638766646385193,
      "eval_math_mean_token_accuracy": 0.7583659777641296,
      "eval_math_num_tokens": 1667532.0,
      "eval_math_runtime": 47.5708,
      "eval_math_samples_per_second": 10.511,
      "eval_math_steps_per_second": 10.511,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_physics_entropy": 0.8860550626516343,
      "eval_physics_loss": 0.8900392055511475,
      "eval_physics_mean_token_accuracy": 0.7706669235229492,
      "eval_physics_num_tokens": 1667532.0,
      "eval_physics_runtime": 52.0022,
      "eval_physics_samples_per_second": 9.615,
      "eval_physics_steps_per_second": 9.615,
      "step": 200
    },
    {
      "entropy": 0.8298934359103441,
      "epoch": 0.084,
      "grad_norm": 28.125,
      "learning_rate": 1.3933333333333334e-05,
      "loss": 13.705,
      "mean_token_accuracy": 0.7773294027894735,
      "num_tokens": 1750813.0,
      "step": 210
    },
    {
      "entropy": 0.8596662495285272,
      "epoch": 0.088,
      "grad_norm": 33.75,
      "learning_rate": 1.46e-05,
      "loss": 13.8646,
      "mean_token_accuracy": 0.7744025643914938,
      "num_tokens": 1835489.0,
      "step": 220
    },
    {
      "entropy": 0.8833530260249972,
      "epoch": 0.092,
      "grad_norm": 36.25,
      "learning_rate": 1.5266666666666667e-05,
      "loss": 14.2962,
      "mean_token_accuracy": 0.7682582084089518,
      "num_tokens": 1915115.0,
      "step": 230
    },
    {
      "entropy": 0.8309201180934906,
      "epoch": 0.096,
      "grad_norm": 34.5,
      "learning_rate": 1.5933333333333336e-05,
      "loss": 13.3235,
      "mean_token_accuracy": 0.7823137249797583,
      "num_tokens": 2007176.0,
      "step": 240
    },
    {
      "entropy": 0.8500596512109041,
      "epoch": 0.1,
      "grad_norm": 29.5,
      "learning_rate": 1.66e-05,
      "loss": 13.5419,
      "mean_token_accuracy": 0.7775608815252781,
      "num_tokens": 2088696.0,
      "step": 250
    },
    {
      "entropy": 0.8657922638580203,
      "epoch": 0.104,
      "grad_norm": 27.625,
      "learning_rate": 1.726666666666667e-05,
      "loss": 13.7733,
      "mean_token_accuracy": 0.7756721317768097,
      "num_tokens": 2171335.0,
      "step": 260
    },
    {
      "entropy": 0.8355370864272118,
      "epoch": 0.108,
      "grad_norm": 43.75,
      "learning_rate": 1.7933333333333333e-05,
      "loss": 13.229,
      "mean_token_accuracy": 0.7806590050458908,
      "num_tokens": 2254168.0,
      "step": 270
    },
    {
      "entropy": 0.8173456467688084,
      "epoch": 0.112,
      "grad_norm": 35.0,
      "learning_rate": 1.86e-05,
      "loss": 12.7881,
      "mean_token_accuracy": 0.7856513697654008,
      "num_tokens": 2342259.0,
      "step": 280
    },
    {
      "entropy": 0.8355640687048436,
      "epoch": 0.116,
      "grad_norm": 32.0,
      "learning_rate": 1.926666666666667e-05,
      "loss": 13.1691,
      "mean_token_accuracy": 0.7810327019542456,
      "num_tokens": 2426289.0,
      "step": 290
    },
    {
      "entropy": 0.8739947421476245,
      "epoch": 0.12,
      "grad_norm": 26.0,
      "learning_rate": 1.9933333333333334e-05,
      "loss": 13.6561,
      "mean_token_accuracy": 0.7744034674018622,
      "num_tokens": 2504279.0,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_biology_entropy": 1.3556034941077233,
      "eval_biology_loss": 1.4841620922088623,
      "eval_biology_mean_token_accuracy": 0.65348690533638,
      "eval_biology_num_tokens": 2504279.0,
      "eval_biology_runtime": 42.283,
      "eval_biology_samples_per_second": 11.825,
      "eval_biology_steps_per_second": 11.825,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_chemistry_entropy": 0.9483438107967377,
      "eval_chemistry_loss": 0.9427877068519592,
      "eval_chemistry_mean_token_accuracy": 0.7558897666931153,
      "eval_chemistry_num_tokens": 2504279.0,
      "eval_chemistry_runtime": 46.6526,
      "eval_chemistry_samples_per_second": 10.718,
      "eval_chemistry_steps_per_second": 10.718,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_math_entropy": 0.8157476499080658,
      "eval_math_loss": 0.904818594455719,
      "eval_math_mean_token_accuracy": 0.7670594818592071,
      "eval_math_num_tokens": 2504279.0,
      "eval_math_runtime": 47.3869,
      "eval_math_samples_per_second": 10.551,
      "eval_math_steps_per_second": 10.551,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_physics_entropy": 0.8239805218577385,
      "eval_physics_loss": 0.8038909435272217,
      "eval_physics_mean_token_accuracy": 0.7841090430021286,
      "eval_physics_num_tokens": 2504279.0,
      "eval_physics_runtime": 52.6201,
      "eval_physics_samples_per_second": 9.502,
      "eval_physics_steps_per_second": 9.502,
      "step": 300
    },
    {
      "entropy": 0.8581649582833052,
      "epoch": 0.124,
      "grad_norm": 26.125,
      "learning_rate": 1.9933333333333334e-05,
      "loss": 13.3818,
      "mean_token_accuracy": 0.7762404128909111,
      "num_tokens": 2583599.0,
      "step": 310
    },
    {
      "entropy": 0.7946215324103832,
      "epoch": 0.128,
      "grad_norm": 25.25,
      "learning_rate": 1.985925925925926e-05,
      "loss": 12.3504,
      "mean_token_accuracy": 0.7905489154160023,
      "num_tokens": 2663987.0,
      "step": 320
    },
    {
      "entropy": 0.849458902142942,
      "epoch": 0.132,
      "grad_norm": 29.375,
      "learning_rate": 1.9785185185185187e-05,
      "loss": 13.2872,
      "mean_token_accuracy": 0.7787170443683863,
      "num_tokens": 2743933.0,
      "step": 330
    },
    {
      "entropy": 0.8214898691512644,
      "epoch": 0.136,
      "grad_norm": 22.125,
      "learning_rate": 1.971111111111111e-05,
      "loss": 12.8325,
      "mean_token_accuracy": 0.7875698886811733,
      "num_tokens": 2829720.0,
      "step": 340
    },
    {
      "entropy": 0.8233486831188201,
      "epoch": 0.14,
      "grad_norm": 21.625,
      "learning_rate": 1.963703703703704e-05,
      "loss": 12.8834,
      "mean_token_accuracy": 0.7814252562820911,
      "num_tokens": 2915814.0,
      "step": 350
    },
    {
      "entropy": 0.8093094614334404,
      "epoch": 0.144,
      "grad_norm": 30.5,
      "learning_rate": 1.9562962962962964e-05,
      "loss": 12.5969,
      "mean_token_accuracy": 0.7904934521764517,
      "num_tokens": 2997824.0,
      "step": 360
    },
    {
      "entropy": 0.8296919351443648,
      "epoch": 0.148,
      "grad_norm": 40.25,
      "learning_rate": 1.948888888888889e-05,
      "loss": 12.8145,
      "mean_token_accuracy": 0.7843083433806897,
      "num_tokens": 3076422.0,
      "step": 370
    },
    {
      "entropy": 0.7770398052409291,
      "epoch": 0.152,
      "grad_norm": 24.25,
      "learning_rate": 1.9414814814814817e-05,
      "loss": 12.0318,
      "mean_token_accuracy": 0.7944367878139019,
      "num_tokens": 3159946.0,
      "step": 380
    },
    {
      "entropy": 0.7682548280805349,
      "epoch": 0.156,
      "grad_norm": 24.75,
      "learning_rate": 1.9340740740740743e-05,
      "loss": 11.8293,
      "mean_token_accuracy": 0.7969654947519302,
      "num_tokens": 3247218.0,
      "step": 390
    },
    {
      "entropy": 0.7929054241627455,
      "epoch": 0.16,
      "grad_norm": 25.125,
      "learning_rate": 1.926666666666667e-05,
      "loss": 12.3015,
      "mean_token_accuracy": 0.7905295439064502,
      "num_tokens": 3330597.0,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_biology_entropy": 1.329050698697567,
      "eval_biology_loss": 1.4526971578598022,
      "eval_biology_mean_token_accuracy": 0.6555387216210365,
      "eval_biology_num_tokens": 3330597.0,
      "eval_biology_runtime": 42.8393,
      "eval_biology_samples_per_second": 11.672,
      "eval_biology_steps_per_second": 11.672,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_chemistry_entropy": 0.9211951863765716,
      "eval_chemistry_loss": 0.9123110175132751,
      "eval_chemistry_mean_token_accuracy": 0.759705729842186,
      "eval_chemistry_num_tokens": 3330597.0,
      "eval_chemistry_runtime": 46.4206,
      "eval_chemistry_samples_per_second": 10.771,
      "eval_chemistry_steps_per_second": 10.771,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_math_entropy": 0.7891712339818477,
      "eval_math_loss": 0.8866830468177795,
      "eval_math_mean_token_accuracy": 0.7700952163934708,
      "eval_math_num_tokens": 3330597.0,
      "eval_math_runtime": 47.5316,
      "eval_math_samples_per_second": 10.519,
      "eval_math_steps_per_second": 10.519,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_physics_entropy": 0.7942325170040131,
      "eval_physics_loss": 0.7685211300849915,
      "eval_physics_mean_token_accuracy": 0.7900554924607277,
      "eval_physics_num_tokens": 3330597.0,
      "eval_physics_runtime": 51.8291,
      "eval_physics_samples_per_second": 9.647,
      "eval_physics_steps_per_second": 9.647,
      "step": 400
    },
    {
      "entropy": 0.8409166993573308,
      "epoch": 0.164,
      "grad_norm": 27.875,
      "learning_rate": 1.9192592592592593e-05,
      "loss": 12.8131,
      "mean_token_accuracy": 0.7820813998579978,
      "num_tokens": 3414445.0,
      "step": 410
    },
    {
      "entropy": 0.7957731707952916,
      "epoch": 0.168,
      "grad_norm": 29.125,
      "learning_rate": 1.911851851851852e-05,
      "loss": 12.2831,
      "mean_token_accuracy": 0.7876646354794502,
      "num_tokens": 3496104.0,
      "step": 420
    },
    {
      "entropy": 0.8156736908480525,
      "epoch": 0.172,
      "grad_norm": 24.0,
      "learning_rate": 1.9044444444444446e-05,
      "loss": 12.6729,
      "mean_token_accuracy": 0.7858119580894709,
      "num_tokens": 3576979.0,
      "step": 430
    },
    {
      "entropy": 0.7964235540479422,
      "epoch": 0.176,
      "grad_norm": 29.0,
      "learning_rate": 1.8970370370370372e-05,
      "loss": 12.1643,
      "mean_token_accuracy": 0.789168506488204,
      "num_tokens": 3658264.0,
      "step": 440
    },
    {
      "entropy": 0.8107027977705001,
      "epoch": 0.18,
      "grad_norm": 32.0,
      "learning_rate": 1.8896296296296295e-05,
      "loss": 12.4439,
      "mean_token_accuracy": 0.78678366728127,
      "num_tokens": 3739142.0,
      "step": 450
    },
    {
      "entropy": 0.796150803938508,
      "epoch": 0.184,
      "grad_norm": 25.75,
      "learning_rate": 1.8822222222222225e-05,
      "loss": 12.3586,
      "mean_token_accuracy": 0.7868688877671957,
      "num_tokens": 3821352.0,
      "step": 460
    },
    {
      "entropy": 0.773661600612104,
      "epoch": 0.188,
      "grad_norm": 24.25,
      "learning_rate": 1.874814814814815e-05,
      "loss": 11.9009,
      "mean_token_accuracy": 0.7943602330982685,
      "num_tokens": 3905291.0,
      "step": 470
    },
    {
      "entropy": 0.7640144268982112,
      "epoch": 0.192,
      "grad_norm": 22.375,
      "learning_rate": 1.8674074074074075e-05,
      "loss": 11.9144,
      "mean_token_accuracy": 0.79531438536942,
      "num_tokens": 3995568.0,
      "step": 480
    },
    {
      "entropy": 0.7889644879847765,
      "epoch": 0.196,
      "grad_norm": 25.875,
      "learning_rate": 1.86e-05,
      "loss": 12.1612,
      "mean_token_accuracy": 0.7931702587753534,
      "num_tokens": 4075670.0,
      "step": 490
    },
    {
      "entropy": 0.7648822296410799,
      "epoch": 0.2,
      "grad_norm": 27.0,
      "learning_rate": 1.8525925925925928e-05,
      "loss": 11.7326,
      "mean_token_accuracy": 0.7970229849219322,
      "num_tokens": 4160918.0,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_biology_entropy": 1.3273861603736878,
      "eval_biology_loss": 1.429775595664978,
      "eval_biology_mean_token_accuracy": 0.6614253748059272,
      "eval_biology_num_tokens": 4160918.0,
      "eval_biology_runtime": 42.9809,
      "eval_biology_samples_per_second": 11.633,
      "eval_biology_steps_per_second": 11.633,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_chemistry_entropy": 0.9144625009298325,
      "eval_chemistry_loss": 0.8940703272819519,
      "eval_chemistry_mean_token_accuracy": 0.7638347852230072,
      "eval_chemistry_num_tokens": 4160918.0,
      "eval_chemistry_runtime": 46.9108,
      "eval_chemistry_samples_per_second": 10.659,
      "eval_chemistry_steps_per_second": 10.659,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_math_entropy": 0.7888483335673809,
      "eval_math_loss": 0.8778181672096252,
      "eval_math_mean_token_accuracy": 0.770329715013504,
      "eval_math_num_tokens": 4160918.0,
      "eval_math_runtime": 48.4026,
      "eval_math_samples_per_second": 10.33,
      "eval_math_steps_per_second": 10.33,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_physics_entropy": 0.7870894553661346,
      "eval_physics_loss": 0.7499192953109741,
      "eval_physics_mean_token_accuracy": 0.7939402560591697,
      "eval_physics_num_tokens": 4160918.0,
      "eval_physics_runtime": 51.8479,
      "eval_physics_samples_per_second": 9.644,
      "eval_physics_steps_per_second": 9.644,
      "step": 500
    },
    {
      "entropy": 0.7847399963065982,
      "epoch": 0.204,
      "grad_norm": 23.375,
      "learning_rate": 1.8451851851851855e-05,
      "loss": 11.9774,
      "mean_token_accuracy": 0.7954912718385458,
      "num_tokens": 4238810.0,
      "step": 510
    },
    {
      "entropy": 0.7635842556133866,
      "epoch": 0.208,
      "grad_norm": 23.75,
      "learning_rate": 1.8377777777777778e-05,
      "loss": 11.7668,
      "mean_token_accuracy": 0.7984023634344339,
      "num_tokens": 4321436.0,
      "step": 520
    },
    {
      "entropy": 0.7751439409330487,
      "epoch": 0.212,
      "grad_norm": 27.125,
      "learning_rate": 1.8303703703703704e-05,
      "loss": 11.8616,
      "mean_token_accuracy": 0.7927958484739065,
      "num_tokens": 4401710.0,
      "step": 530
    },
    {
      "entropy": 0.7921603286638856,
      "epoch": 0.216,
      "grad_norm": 23.5,
      "learning_rate": 1.822962962962963e-05,
      "loss": 12.2102,
      "mean_token_accuracy": 0.7888585310429335,
      "num_tokens": 4482225.0,
      "step": 540
    },
    {
      "entropy": 0.7579952264204621,
      "epoch": 0.22,
      "grad_norm": 24.25,
      "learning_rate": 1.8155555555555557e-05,
      "loss": 11.7695,
      "mean_token_accuracy": 0.7949905574321747,
      "num_tokens": 4564662.0,
      "step": 550
    },
    {
      "entropy": 0.7470487952232361,
      "epoch": 0.224,
      "grad_norm": 20.25,
      "learning_rate": 1.8081481481481484e-05,
      "loss": 11.4255,
      "mean_token_accuracy": 0.802862798050046,
      "num_tokens": 4648491.0,
      "step": 560
    },
    {
      "entropy": 0.7314794508740305,
      "epoch": 0.228,
      "grad_norm": 22.5,
      "learning_rate": 1.800740740740741e-05,
      "loss": 11.3663,
      "mean_token_accuracy": 0.8025615274906158,
      "num_tokens": 4731596.0,
      "step": 570
    },
    {
      "entropy": 0.7862403180450201,
      "epoch": 0.232,
      "grad_norm": 22.375,
      "learning_rate": 1.7933333333333333e-05,
      "loss": 12.1004,
      "mean_token_accuracy": 0.79150385633111,
      "num_tokens": 4813449.0,
      "step": 580
    },
    {
      "entropy": 0.7626398691907526,
      "epoch": 0.236,
      "grad_norm": 19.75,
      "learning_rate": 1.785925925925926e-05,
      "loss": 11.7449,
      "mean_token_accuracy": 0.7947790089994669,
      "num_tokens": 4899144.0,
      "step": 590
    },
    {
      "entropy": 0.7703196549788117,
      "epoch": 0.24,
      "grad_norm": 21.375,
      "learning_rate": 1.7785185185185186e-05,
      "loss": 11.8874,
      "mean_token_accuracy": 0.7943012371659279,
      "num_tokens": 4986175.0,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_biology_entropy": 1.305106226325035,
      "eval_biology_loss": 1.4206056594848633,
      "eval_biology_mean_token_accuracy": 0.6621832525730134,
      "eval_biology_num_tokens": 4986175.0,
      "eval_biology_runtime": 42.8609,
      "eval_biology_samples_per_second": 11.666,
      "eval_biology_steps_per_second": 11.666,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_chemistry_entropy": 0.8958090140223504,
      "eval_chemistry_loss": 0.8835439682006836,
      "eval_chemistry_mean_token_accuracy": 0.7649876139163971,
      "eval_chemistry_num_tokens": 4986175.0,
      "eval_chemistry_runtime": 46.9905,
      "eval_chemistry_samples_per_second": 10.64,
      "eval_chemistry_steps_per_second": 10.64,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_math_entropy": 0.7727740241885185,
      "eval_math_loss": 0.8744282126426697,
      "eval_math_mean_token_accuracy": 0.772739317893982,
      "eval_math_num_tokens": 4986175.0,
      "eval_math_runtime": 47.5429,
      "eval_math_samples_per_second": 10.517,
      "eval_math_steps_per_second": 10.517,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_physics_entropy": 0.7677388923168182,
      "eval_physics_loss": 0.7394809722900391,
      "eval_physics_mean_token_accuracy": 0.7961059678792953,
      "eval_physics_num_tokens": 4986175.0,
      "eval_physics_runtime": 52.1156,
      "eval_physics_samples_per_second": 9.594,
      "eval_physics_steps_per_second": 9.594,
      "step": 600
    },
    {
      "entropy": 0.7800249606370926,
      "epoch": 0.244,
      "grad_norm": 24.875,
      "learning_rate": 1.7711111111111113e-05,
      "loss": 12.0013,
      "mean_token_accuracy": 0.7928039100021124,
      "num_tokens": 5071994.0,
      "step": 610
    },
    {
      "entropy": 0.7350489016622305,
      "epoch": 0.248,
      "grad_norm": 22.0,
      "learning_rate": 1.763703703703704e-05,
      "loss": 11.3769,
      "mean_token_accuracy": 0.8017779812216759,
      "num_tokens": 5164741.0,
      "step": 620
    },
    {
      "entropy": 0.791132396273315,
      "epoch": 0.252,
      "grad_norm": 23.5,
      "learning_rate": 1.7562962962962962e-05,
      "loss": 12.1438,
      "mean_token_accuracy": 0.7909818843007088,
      "num_tokens": 5245878.0,
      "step": 630
    },
    {
      "entropy": 0.7686850160360337,
      "epoch": 0.256,
      "grad_norm": 26.125,
      "learning_rate": 1.7488888888888892e-05,
      "loss": 11.7679,
      "mean_token_accuracy": 0.7952951941639185,
      "num_tokens": 5329320.0,
      "step": 640
    },
    {
      "entropy": 0.7577827019616962,
      "epoch": 0.26,
      "grad_norm": 21.625,
      "learning_rate": 1.7414814814814815e-05,
      "loss": 11.6531,
      "mean_token_accuracy": 0.7984100170433521,
      "num_tokens": 5412557.0,
      "step": 650
    },
    {
      "entropy": 0.7561163729056716,
      "epoch": 0.264,
      "grad_norm": 25.75,
      "learning_rate": 1.7340740740740742e-05,
      "loss": 11.6519,
      "mean_token_accuracy": 0.7982414372265338,
      "num_tokens": 5497259.0,
      "step": 660
    },
    {
      "entropy": 0.7754932438954711,
      "epoch": 0.268,
      "grad_norm": 22.75,
      "learning_rate": 1.726666666666667e-05,
      "loss": 11.8623,
      "mean_token_accuracy": 0.7968760304152965,
      "num_tokens": 5576721.0,
      "step": 670
    },
    {
      "entropy": 0.7834382282570005,
      "epoch": 0.272,
      "grad_norm": 26.75,
      "learning_rate": 1.7192592592592595e-05,
      "loss": 12.0999,
      "mean_token_accuracy": 0.7898719735443592,
      "num_tokens": 5658796.0,
      "step": 680
    },
    {
      "entropy": 0.7922874016687274,
      "epoch": 0.276,
      "grad_norm": 24.375,
      "learning_rate": 1.711851851851852e-05,
      "loss": 12.242,
      "mean_token_accuracy": 0.7898238655179739,
      "num_tokens": 5737499.0,
      "step": 690
    },
    {
      "entropy": 0.7801830777898431,
      "epoch": 0.28,
      "grad_norm": 30.25,
      "learning_rate": 1.7044444444444445e-05,
      "loss": 11.8659,
      "mean_token_accuracy": 0.7971487943083048,
      "num_tokens": 5819027.0,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_biology_entropy": 1.3291225813031196,
      "eval_biology_loss": 1.412896990776062,
      "eval_biology_mean_token_accuracy": 0.6633186488747597,
      "eval_biology_num_tokens": 5819027.0,
      "eval_biology_runtime": 41.0252,
      "eval_biology_samples_per_second": 12.188,
      "eval_biology_steps_per_second": 12.188,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_chemistry_entropy": 0.9023040079474449,
      "eval_chemistry_loss": 0.875540018081665,
      "eval_chemistry_mean_token_accuracy": 0.7670781625509262,
      "eval_chemistry_num_tokens": 5819027.0,
      "eval_chemistry_runtime": 45.3343,
      "eval_chemistry_samples_per_second": 11.029,
      "eval_chemistry_steps_per_second": 11.029,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_math_entropy": 0.7758396743237972,
      "eval_math_loss": 0.8661128878593445,
      "eval_math_mean_token_accuracy": 0.7738337585926056,
      "eval_math_num_tokens": 5819027.0,
      "eval_math_runtime": 46.0356,
      "eval_math_samples_per_second": 10.861,
      "eval_math_steps_per_second": 10.861,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_physics_entropy": 0.7709107432365417,
      "eval_physics_loss": 0.7312194108963013,
      "eval_physics_mean_token_accuracy": 0.7975488212108612,
      "eval_physics_num_tokens": 5819027.0,
      "eval_physics_runtime": 50.9222,
      "eval_physics_samples_per_second": 9.819,
      "eval_physics_steps_per_second": 9.819,
      "step": 700
    },
    {
      "entropy": 0.8033936187624932,
      "epoch": 0.284,
      "grad_norm": 28.125,
      "learning_rate": 1.697037037037037e-05,
      "loss": 12.2242,
      "mean_token_accuracy": 0.7905067663639784,
      "num_tokens": 5898693.0,
      "step": 710
    },
    {
      "entropy": 0.7491526246070862,
      "epoch": 0.288,
      "grad_norm": 24.0,
      "learning_rate": 1.6896296296296298e-05,
      "loss": 11.654,
      "mean_token_accuracy": 0.7981848757714033,
      "num_tokens": 5980201.0,
      "step": 720
    },
    {
      "entropy": 0.7438665444031358,
      "epoch": 0.292,
      "grad_norm": 22.75,
      "learning_rate": 1.6822222222222224e-05,
      "loss": 11.3524,
      "mean_token_accuracy": 0.8035779967904091,
      "num_tokens": 6059507.0,
      "step": 730
    },
    {
      "entropy": 0.7584557346999645,
      "epoch": 0.296,
      "grad_norm": 24.375,
      "learning_rate": 1.6748148148148147e-05,
      "loss": 11.8562,
      "mean_token_accuracy": 0.7952129423618317,
      "num_tokens": 6139830.0,
      "step": 740
    },
    {
      "entropy": 0.7649100622162223,
      "epoch": 0.3,
      "grad_norm": 21.375,
      "learning_rate": 1.6674074074074077e-05,
      "loss": 11.6522,
      "mean_token_accuracy": 0.7979466877877712,
      "num_tokens": 6218134.0,
      "step": 750
    },
    {
      "entropy": 0.7744827693328261,
      "epoch": 0.304,
      "grad_norm": 23.0,
      "learning_rate": 1.66e-05,
      "loss": 11.7959,
      "mean_token_accuracy": 0.7969288475811481,
      "num_tokens": 6298858.0,
      "step": 760
    },
    {
      "entropy": 0.8085566008463502,
      "epoch": 0.308,
      "grad_norm": 29.125,
      "learning_rate": 1.6525925925925927e-05,
      "loss": 12.4325,
      "mean_token_accuracy": 0.7856661215424537,
      "num_tokens": 6375761.0,
      "step": 770
    },
    {
      "entropy": 0.7414015829563141,
      "epoch": 0.312,
      "grad_norm": 22.25,
      "learning_rate": 1.6451851851851853e-05,
      "loss": 11.3117,
      "mean_token_accuracy": 0.8043436016887426,
      "num_tokens": 6456502.0,
      "step": 780
    },
    {
      "entropy": 0.7605590077117086,
      "epoch": 0.316,
      "grad_norm": 24.875,
      "learning_rate": 1.637777777777778e-05,
      "loss": 11.7982,
      "mean_token_accuracy": 0.7984256561845541,
      "num_tokens": 6537521.0,
      "step": 790
    },
    {
      "entropy": 0.7728373639285564,
      "epoch": 0.32,
      "grad_norm": 21.0,
      "learning_rate": 1.6303703703703706e-05,
      "loss": 12.0483,
      "mean_token_accuracy": 0.7910115845501423,
      "num_tokens": 6622798.0,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_biology_entropy": 1.3250962881445885,
      "eval_biology_loss": 1.4059523344039917,
      "eval_biology_mean_token_accuracy": 0.6637759111523628,
      "eval_biology_num_tokens": 6622798.0,
      "eval_biology_runtime": 40.8068,
      "eval_biology_samples_per_second": 12.253,
      "eval_biology_steps_per_second": 12.253,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_chemistry_entropy": 0.8944689989089966,
      "eval_chemistry_loss": 0.8700199723243713,
      "eval_chemistry_mean_token_accuracy": 0.7685709017515182,
      "eval_chemistry_num_tokens": 6622798.0,
      "eval_chemistry_runtime": 45.5304,
      "eval_chemistry_samples_per_second": 10.982,
      "eval_chemistry_steps_per_second": 10.982,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_math_entropy": 0.7684379303753376,
      "eval_math_loss": 0.8683172464370728,
      "eval_math_mean_token_accuracy": 0.773405775308609,
      "eval_math_num_tokens": 6622798.0,
      "eval_math_runtime": 46.094,
      "eval_math_samples_per_second": 10.847,
      "eval_math_steps_per_second": 10.847,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_physics_entropy": 0.7618584106564522,
      "eval_physics_loss": 0.7243406772613525,
      "eval_physics_mean_token_accuracy": 0.7987258922457695,
      "eval_physics_num_tokens": 6622798.0,
      "eval_physics_runtime": 50.9642,
      "eval_physics_samples_per_second": 9.811,
      "eval_physics_steps_per_second": 9.811,
      "step": 800
    },
    {
      "entropy": 0.7659748939797282,
      "epoch": 0.324,
      "grad_norm": 25.125,
      "learning_rate": 1.622962962962963e-05,
      "loss": 11.7342,
      "mean_token_accuracy": 0.796695252135396,
      "num_tokens": 6699373.0,
      "step": 810
    },
    {
      "entropy": 0.7672781342640519,
      "epoch": 0.328,
      "grad_norm": 20.125,
      "learning_rate": 1.6155555555555556e-05,
      "loss": 11.8728,
      "mean_token_accuracy": 0.7931396309286356,
      "num_tokens": 6782985.0,
      "step": 820
    },
    {
      "entropy": 0.753248256072402,
      "epoch": 0.332,
      "grad_norm": 25.75,
      "learning_rate": 1.6081481481481482e-05,
      "loss": 11.524,
      "mean_token_accuracy": 0.7992388799786567,
      "num_tokens": 6868578.0,
      "step": 830
    },
    {
      "entropy": 0.7217373944818973,
      "epoch": 0.336,
      "grad_norm": 21.25,
      "learning_rate": 1.600740740740741e-05,
      "loss": 11.0518,
      "mean_token_accuracy": 0.8084735371172428,
      "num_tokens": 6953611.0,
      "step": 840
    },
    {
      "entropy": 0.7724356524646282,
      "epoch": 0.34,
      "grad_norm": 25.375,
      "learning_rate": 1.5933333333333336e-05,
      "loss": 11.9784,
      "mean_token_accuracy": 0.7944952342659235,
      "num_tokens": 7037831.0,
      "step": 850
    },
    {
      "entropy": 0.7315696366131306,
      "epoch": 0.344,
      "grad_norm": 28.75,
      "learning_rate": 1.5859259259259262e-05,
      "loss": 11.209,
      "mean_token_accuracy": 0.8057551696896553,
      "num_tokens": 7121949.0,
      "step": 860
    },
    {
      "entropy": 0.7565067631192506,
      "epoch": 0.348,
      "grad_norm": 21.75,
      "learning_rate": 1.5785185185185185e-05,
      "loss": 11.5165,
      "mean_token_accuracy": 0.801112212240696,
      "num_tokens": 7206056.0,
      "step": 870
    },
    {
      "entropy": 0.7334232210181654,
      "epoch": 0.352,
      "grad_norm": 26.25,
      "learning_rate": 1.571111111111111e-05,
      "loss": 11.4043,
      "mean_token_accuracy": 0.8033700969070197,
      "num_tokens": 7290378.0,
      "step": 880
    },
    {
      "entropy": 0.7614396458491683,
      "epoch": 0.356,
      "grad_norm": 18.75,
      "learning_rate": 1.5637037037037038e-05,
      "loss": 11.7008,
      "mean_token_accuracy": 0.7960948143154383,
      "num_tokens": 7367952.0,
      "step": 890
    },
    {
      "entropy": 0.7479236021637916,
      "epoch": 0.36,
      "grad_norm": 29.5,
      "learning_rate": 1.5562962962962965e-05,
      "loss": 11.5145,
      "mean_token_accuracy": 0.8022682599723339,
      "num_tokens": 7452920.0,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_biology_entropy": 1.3072672160863876,
      "eval_biology_loss": 1.402725100517273,
      "eval_biology_mean_token_accuracy": 0.6647561370134354,
      "eval_biology_num_tokens": 7452920.0,
      "eval_biology_runtime": 42.0955,
      "eval_biology_samples_per_second": 11.878,
      "eval_biology_steps_per_second": 11.878,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_chemistry_entropy": 0.889559015929699,
      "eval_chemistry_loss": 0.8674989938735962,
      "eval_chemistry_mean_token_accuracy": 0.768995472073555,
      "eval_chemistry_num_tokens": 7452920.0,
      "eval_chemistry_runtime": 45.4249,
      "eval_chemistry_samples_per_second": 11.007,
      "eval_chemistry_steps_per_second": 11.007,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_math_entropy": 0.7602320345044136,
      "eval_math_loss": 0.8669670224189758,
      "eval_math_mean_token_accuracy": 0.7743704164028168,
      "eval_math_num_tokens": 7452920.0,
      "eval_math_runtime": 45.9692,
      "eval_math_samples_per_second": 10.877,
      "eval_math_steps_per_second": 10.877,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_physics_entropy": 0.7541027455925942,
      "eval_physics_loss": 0.7185434699058533,
      "eval_physics_mean_token_accuracy": 0.7998301391601562,
      "eval_physics_num_tokens": 7452920.0,
      "eval_physics_runtime": 51.1617,
      "eval_physics_samples_per_second": 9.773,
      "eval_physics_steps_per_second": 9.773,
      "step": 900
    },
    {
      "entropy": 0.7335074551403522,
      "epoch": 0.364,
      "grad_norm": 22.125,
      "learning_rate": 1.548888888888889e-05,
      "loss": 11.2186,
      "mean_token_accuracy": 0.8046895559877157,
      "num_tokens": 7535961.0,
      "step": 910
    },
    {
      "entropy": 0.7247136728838086,
      "epoch": 0.368,
      "grad_norm": 22.25,
      "learning_rate": 1.5414814814814814e-05,
      "loss": 11.0732,
      "mean_token_accuracy": 0.8059736005961895,
      "num_tokens": 7621082.0,
      "step": 920
    },
    {
      "entropy": 0.7282988695427776,
      "epoch": 0.372,
      "grad_norm": 22.875,
      "learning_rate": 1.5340740740740744e-05,
      "loss": 11.1056,
      "mean_token_accuracy": 0.8030345071107149,
      "num_tokens": 7705651.0,
      "step": 930
    },
    {
      "entropy": 0.8100492168217898,
      "epoch": 0.376,
      "grad_norm": 22.75,
      "learning_rate": 1.5266666666666667e-05,
      "loss": 12.5658,
      "mean_token_accuracy": 0.7812250763177871,
      "num_tokens": 7787036.0,
      "step": 940
    },
    {
      "entropy": 0.7599826741963625,
      "epoch": 0.38,
      "grad_norm": 27.125,
      "learning_rate": 1.5192592592592594e-05,
      "loss": 11.5637,
      "mean_token_accuracy": 0.795925023406744,
      "num_tokens": 7869437.0,
      "step": 950
    },
    {
      "entropy": 0.734612544812262,
      "epoch": 0.384,
      "grad_norm": 19.625,
      "learning_rate": 1.5118518518518519e-05,
      "loss": 11.2725,
      "mean_token_accuracy": 0.8022117022424936,
      "num_tokens": 7955570.0,
      "step": 960
    },
    {
      "entropy": 0.7115672262385487,
      "epoch": 0.388,
      "grad_norm": 20.625,
      "learning_rate": 1.5044444444444445e-05,
      "loss": 10.9492,
      "mean_token_accuracy": 0.8076492041349411,
      "num_tokens": 8038818.0,
      "step": 970
    },
    {
      "entropy": 0.7394374579191207,
      "epoch": 0.392,
      "grad_norm": 20.0,
      "learning_rate": 1.497037037037037e-05,
      "loss": 11.3016,
      "mean_token_accuracy": 0.8001755453646183,
      "num_tokens": 8123527.0,
      "step": 980
    },
    {
      "entropy": 0.7394586138427257,
      "epoch": 0.396,
      "grad_norm": 23.0,
      "learning_rate": 1.4896296296296298e-05,
      "loss": 11.4531,
      "mean_token_accuracy": 0.8012358251959085,
      "num_tokens": 8206501.0,
      "step": 990
    },
    {
      "entropy": 0.7182500061579049,
      "epoch": 0.4,
      "grad_norm": 19.25,
      "learning_rate": 1.4822222222222225e-05,
      "loss": 11.0903,
      "mean_token_accuracy": 0.8073178566992283,
      "num_tokens": 8291049.0,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_biology_entropy": 1.307286373436451,
      "eval_biology_loss": 1.3969565629959106,
      "eval_biology_mean_token_accuracy": 0.6652992497086525,
      "eval_biology_num_tokens": 8291049.0,
      "eval_biology_runtime": 40.7973,
      "eval_biology_samples_per_second": 12.256,
      "eval_biology_steps_per_second": 12.256,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_chemistry_entropy": 0.8853131957352162,
      "eval_chemistry_loss": 0.8617890477180481,
      "eval_chemistry_mean_token_accuracy": 0.7699888870716095,
      "eval_chemistry_num_tokens": 8291049.0,
      "eval_chemistry_runtime": 46.893,
      "eval_chemistry_samples_per_second": 10.663,
      "eval_chemistry_steps_per_second": 10.663,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_math_entropy": 0.7565005440711975,
      "eval_math_loss": 0.865938127040863,
      "eval_math_mean_token_accuracy": 0.7747692105770111,
      "eval_math_num_tokens": 8291049.0,
      "eval_math_runtime": 47.0487,
      "eval_math_samples_per_second": 10.627,
      "eval_math_steps_per_second": 10.627,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_physics_entropy": 0.7521990032792092,
      "eval_physics_loss": 0.71455979347229,
      "eval_physics_mean_token_accuracy": 0.8008181518912315,
      "eval_physics_num_tokens": 8291049.0,
      "eval_physics_runtime": 51.6245,
      "eval_physics_samples_per_second": 9.685,
      "eval_physics_steps_per_second": 9.685,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 3000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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": 5.559510372210398e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}