{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 12588,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507937e-07,
      "loss": 1.297,
      "theoretical_loss": 3.515367210984307,
      "tokens_seen": 1650130944
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5873015873015873e-06,
      "loss": 1.328,
      "theoretical_loss": 3.515343236624406,
      "tokens_seen": 1650262016
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.3809523809523808e-06,
      "loss": 1.3087,
      "theoretical_loss": 3.5153192647017093,
      "tokens_seen": 1650393088
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.1746031746031746e-06,
      "loss": 1.3361,
      "theoretical_loss": 3.5152952952157754,
      "tokens_seen": 1650524160
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-06,
      "loss": 1.3018,
      "theoretical_loss": 3.515271328166164,
      "tokens_seen": 1650655232
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.7619047619047615e-06,
      "loss": 1.2823,
      "theoretical_loss": 3.515247363552433,
      "tokens_seen": 1650786304
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.555555555555556e-06,
      "loss": 1.2569,
      "theoretical_loss": 3.5152234013741417,
      "tokens_seen": 1650917376
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.349206349206349e-06,
      "loss": 1.2991,
      "theoretical_loss": 3.51519944163085,
      "tokens_seen": 1651048448
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.142857142857143e-06,
      "loss": 1.3684,
      "theoretical_loss": 3.515175484322116,
      "tokens_seen": 1651179520
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507936e-06,
      "loss": 1.3216,
      "theoretical_loss": 3.5151515294474995,
      "tokens_seen": 1651310592
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.73015873015873e-06,
      "loss": 1.2794,
      "theoretical_loss": 3.5151275770065604,
      "tokens_seen": 1651441664
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.523809523809523e-06,
      "loss": 1.2926,
      "theoretical_loss": 3.5151036269988585,
      "tokens_seen": 1651572736
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.0317460317460318e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.5150796794239527,
      "tokens_seen": 1651703808
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.515055734281404,
      "tokens_seen": 1651834880
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.5150317915707707,
      "tokens_seen": 1651965952
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.2698412698412699e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.5150078512916143,
      "tokens_seen": 1652097024
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.3492063492063492e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.5149839134434946,
      "tokens_seen": 1652228096
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.4285714285714285e-05,
      "loss": 1.3839,
      "theoretical_loss": 3.5149599780259715,
      "tokens_seen": 1652359168
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5079365079365079e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.5149360450386062,
      "tokens_seen": 1652490240
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.5149121144809587,
      "tokens_seen": 1652621312
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.5148881863525903,
      "tokens_seen": 1652752384
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.746031746031746e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.5148642606530602,
      "tokens_seen": 1652883456
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.8253968253968254e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.514840337381931,
      "tokens_seen": 1653014528
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.9047619047619046e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.514816416538763,
      "tokens_seen": 1653145600
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-05,
      "loss": 1.334,
      "theoretical_loss": 3.5147924981231173,
      "tokens_seen": 1653276672
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.0634920634920636e-05,
      "loss": 1.3821,
      "theoretical_loss": 3.514768582134555,
      "tokens_seen": 1653407744
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.514744668572638,
      "tokens_seen": 1653538816
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.514720757436927,
      "tokens_seen": 1653669888
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.3015873015873015e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.5146968487269845,
      "tokens_seen": 1653800960
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.380952380952381e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.5146729424423713,
      "tokens_seen": 1653932032
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.4603174603174602e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.5146490385826494,
      "tokens_seen": 1654063104
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.5396825396825397e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.5146251371473807,
      "tokens_seen": 1654194176
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.6190476190476192e-05,
      "loss": 1.3792,
      "theoretical_loss": 3.5146012381361276,
      "tokens_seen": 1654325248
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.6984126984126984e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.514577341548452,
      "tokens_seen": 1654456320
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.514553447383916,
      "tokens_seen": 1654587392
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.857142857142857e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.514529555642082,
      "tokens_seen": 1654718464
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9365079365079366e-05,
      "loss": 1.3522,
      "theoretical_loss": 3.514505666322513,
      "tokens_seen": 1654849536
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.0158730158730158e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.514481779424771,
      "tokens_seen": 1654980608
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.095238095238095e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.5144578949484186,
      "tokens_seen": 1655111680
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.1746031746031745e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.5144340128930196,
      "tokens_seen": 1655242752
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.253968253968254e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.5144101332581354,
      "tokens_seen": 1655373824
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 1.3599,
      "theoretical_loss": 3.5143862560433305,
      "tokens_seen": 1655504896
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.412698412698413e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.514362381248167,
      "tokens_seen": 1655635968
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.492063492063492e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.5143385088722088,
      "tokens_seen": 1655767040
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.571428571428572e-05,
      "loss": 1.3193,
      "theoretical_loss": 3.5143146389150193,
      "tokens_seen": 1655898112
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.650793650793651e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.514290771376162,
      "tokens_seen": 1656029184
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.730158730158731e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.5142669062552,
      "tokens_seen": 1656160256
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.809523809523809e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.5142430435516974,
      "tokens_seen": 1656291328
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.888888888888889e-05,
      "loss": 1.269,
      "theoretical_loss": 3.514219183265218,
      "tokens_seen": 1656422400
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.9931244850158691,
      "objective/train/avg_token_score": 0.006875487044453621,
      "objective/train/docs_used": 604805,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3661718368530273,
      "objective/train/original_loss": 1.3756071329116821,
      "objective/train/theoretical_loss": 3.5141953253953258,
      "objective/train/tokens_used": 27012576,
      "objective/train/unlikelihood_loss": 0.020264359191060066,
      "theoretical_loss": 3.5141953253953258,
      "tokens_seen": 1656553472
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.5141953253953258,
      "tokens_seen": 1656553472
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.047619047619048e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.5141714699415845,
      "tokens_seen": 1656684544
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.126984126984127e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.514147616903559,
      "tokens_seen": 1656815616
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.2063492063492065e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.5141237662808127,
      "tokens_seen": 1656946688
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.2857142857142856e-05,
      "loss": 1.3849,
      "theoretical_loss": 3.514099918072911,
      "tokens_seen": 1657077760
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.3650793650793655e-05,
      "loss": 1.25,
      "theoretical_loss": 3.5140760722794178,
      "tokens_seen": 1657208832
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.5140522288998977,
      "tokens_seen": 1657339904
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.523809523809524e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.514028387933916,
      "tokens_seen": 1657470976
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.603174603174603e-05,
      "loss": 1.242,
      "theoretical_loss": 3.5140045493810366,
      "tokens_seen": 1657602048
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.682539682539683e-05,
      "loss": 1.3436,
      "theoretical_loss": 3.5139807132408256,
      "tokens_seen": 1657733120
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.761904761904762e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.513956879512847,
      "tokens_seen": 1657864192
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.841269841269841e-05,
      "loss": 1.31,
      "theoretical_loss": 3.513933048196667,
      "tokens_seen": 1657995264
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.9206349206349204e-05,
      "loss": 1.3381,
      "theoretical_loss": 3.5139092192918495,
      "tokens_seen": 1658126336
    },
    {
      "epoch": 0.01,
      "learning_rate": 5e-05,
      "loss": 1.235,
      "theoretical_loss": 3.5138853927979614,
      "tokens_seen": 1658257408
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.0793650793650794e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.513861568714568,
      "tokens_seen": 1658388480
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.158730158730159e-05,
      "loss": 1.3739,
      "theoretical_loss": 3.513837747041234,
      "tokens_seen": 1658519552
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.2380952380952384e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.513813927777526,
      "tokens_seen": 1658650624
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.3174603174603176e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.5137901109230096,
      "tokens_seen": 1658781696
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.396825396825397e-05,
      "loss": 1.281,
      "theoretical_loss": 3.513766296477251,
      "tokens_seen": 1658912768
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.4761904761904766e-05,
      "loss": 1.3746,
      "theoretical_loss": 3.5137424844398164,
      "tokens_seen": 1659043840
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.555555555555556e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.5137186748102716,
      "tokens_seen": 1659174912
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.634920634920635e-05,
      "loss": 1.3507,
      "theoretical_loss": 3.5136948675881827,
      "tokens_seen": 1659305984
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.714285714285714e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.5136710627731165,
      "tokens_seen": 1659437056
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.793650793650795e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.51364726036464,
      "tokens_seen": 1659568128
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.873015873015873e-05,
      "loss": 1.4346,
      "theoretical_loss": 3.513623460362319,
      "tokens_seen": 1659699200
    },
    {
      "epoch": 0.01,
      "learning_rate": 5.9523809523809524e-05,
      "loss": 1.3381,
      "theoretical_loss": 3.5135996627657207,
      "tokens_seen": 1659830272
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.0317460317460316e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.5135758675744118,
      "tokens_seen": 1659961344
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.111111111111112e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.5135520747879596,
      "tokens_seen": 1660092416
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.19047619047619e-05,
      "loss": 1.3918,
      "theoretical_loss": 3.5135282844059317,
      "tokens_seen": 1660223488
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.26984126984127e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.5135044964278936,
      "tokens_seen": 1660354560
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.349206349206349e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.5134807108534147,
      "tokens_seen": 1660485632
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.428571428571429e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.5134569276820606,
      "tokens_seen": 1660616704
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.507936507936509e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.5134331469134006,
      "tokens_seen": 1660747776
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.587301587301587e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.5134093685470007,
      "tokens_seen": 1660878848
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.666666666666667e-05,
      "loss": 1.3488,
      "theoretical_loss": 3.5133855925824298,
      "tokens_seen": 1661009920
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.746031746031747e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.5133618190192553,
      "tokens_seen": 1661140992
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.825396825396825e-05,
      "loss": 1.3441,
      "theoretical_loss": 3.5133380478570455,
      "tokens_seen": 1661272064
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.904761904761905e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.5133142790953684,
      "tokens_seen": 1661403136
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.984126984126984e-05,
      "loss": 1.3335,
      "theoretical_loss": 3.513290512733792,
      "tokens_seen": 1661534208
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.063492063492065e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.513266748771885,
      "tokens_seen": 1661665280
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.142857142857143e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.5132429872092157,
      "tokens_seen": 1661796352
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.222222222222222e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.5132192280453527,
      "tokens_seen": 1661927424
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.301587301587302e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.5131954712798645,
      "tokens_seen": 1662058496
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.380952380952382e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.5131717169123196,
      "tokens_seen": 1662189568
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.460317460317461e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.513147964942288,
      "tokens_seen": 1662320640
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.53968253968254e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.513124215369338,
      "tokens_seen": 1662451712
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.619047619047618e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.513100468193038,
      "tokens_seen": 1662582784
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.6984126984127e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.5130767234129587,
      "tokens_seen": 1662713856
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.777777777777778e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.513052981028668,
      "tokens_seen": 1662844928
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.857142857142858e-05,
      "loss": 1.3632,
      "theoretical_loss": 3.513029241039737,
      "tokens_seen": 1662976000
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.9920444488525391,
      "objective/train/avg_token_score": 0.007955576293170452,
      "objective/train/docs_used": 607267,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3100258111953735,
      "objective/train/original_loss": 1.3211314678192139,
      "objective/train/theoretical_loss": 3.5130055034457337,
      "objective/train/tokens_used": 33566176,
      "objective/train/unlikelihood_loss": 0.02151995152235031,
      "theoretical_loss": 3.5130055034457337,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.01,
      "learning_rate": 7.936507936507937e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.5130055034457337,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.015873015873016e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.5129817682462283,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.095238095238096e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.512958035440791,
      "tokens_seen": 1663369216
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.174603174603175e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.512934305028991,
      "tokens_seen": 1663500288
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.253968253968255e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.512910577010399,
      "tokens_seen": 1663631360
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.333333333333334e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.512886851384585,
      "tokens_seen": 1663762432
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.412698412698413e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.5128631281511185,
      "tokens_seen": 1663893504
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.492063492063493e-05,
      "loss": 1.3684,
      "theoretical_loss": 3.5128394073095706,
      "tokens_seen": 1664024576
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.571428571428571e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.512815688859512,
      "tokens_seen": 1664155648
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.650793650793651e-05,
      "loss": 1.304,
      "theoretical_loss": 3.512791972800512,
      "tokens_seen": 1664286720
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.730158730158731e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.5127682591321427,
      "tokens_seen": 1664417792
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.80952380952381e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.5127445478539743,
      "tokens_seen": 1664548864
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.888888888888889e-05,
      "loss": 1.289,
      "theoretical_loss": 3.5127208389655777,
      "tokens_seen": 1664679936
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.968253968253969e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.5126971324665237,
      "tokens_seen": 1664811008
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.047619047619048e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.5126734283563836,
      "tokens_seen": 1664942080
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.126984126984128e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.5126497266347285,
      "tokens_seen": 1665073152
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.206349206349206e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.5126260273011303,
      "tokens_seen": 1665204224
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.285714285714286e-05,
      "loss": 1.3518,
      "theoretical_loss": 3.5126023303551595,
      "tokens_seen": 1665335296
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.365079365079366e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.5125786357963884,
      "tokens_seen": 1665466368
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.444444444444444e-05,
      "loss": 1.3967,
      "theoretical_loss": 3.5125549436243886,
      "tokens_seen": 1665597440
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.523809523809524e-05,
      "loss": 1.296,
      "theoretical_loss": 3.5125312538387314,
      "tokens_seen": 1665728512
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.603174603174604e-05,
      "loss": 1.3573,
      "theoretical_loss": 3.5125075664389893,
      "tokens_seen": 1665859584
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.682539682539682e-05,
      "loss": 1.3658,
      "theoretical_loss": 3.512483881424734,
      "tokens_seen": 1665990656
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.761904761904762e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.5124601987955377,
      "tokens_seen": 1666121728
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.841269841269841e-05,
      "loss": 1.3592,
      "theoretical_loss": 3.5124365185509725,
      "tokens_seen": 1666252800
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.920634920634922e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.5124128406906108,
      "tokens_seen": 1666383872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0001,
      "loss": 1.4048,
      "theoretical_loss": 3.5123891652140253,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.999197560584176e-05,
      "loss": 1.337,
      "theoretical_loss": 3.512365492120788,
      "tokens_seen": 1666646016
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.998395121168352e-05,
      "loss": 1.3757,
      "theoretical_loss": 3.5123418214104722,
      "tokens_seen": 1666777088
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.997592681752528e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.51231815308265,
      "tokens_seen": 1666908160
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.996790242336704e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.5122944871368946,
      "tokens_seen": 1667039232
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.99598780292088e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.5122708235727798,
      "tokens_seen": 1667170304
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.995185363505056e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.512247162389877,
      "tokens_seen": 1667301376
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.994382924089232e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.512223503587761,
      "tokens_seen": 1667432448
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.993580484673407e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.512199847166004,
      "tokens_seen": 1667563520
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.992778045257584e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.5121761931241804,
      "tokens_seen": 1667694592
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.991975605841759e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.512152541461863,
      "tokens_seen": 1667825664
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.991173166425936e-05,
      "loss": 1.3665,
      "theoretical_loss": 3.512128892178626,
      "tokens_seen": 1667956736
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.990370727010111e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.5121052452740424,
      "tokens_seen": 1668087808
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.989568287594288e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.512081600747687,
      "tokens_seen": 1668218880
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.988765848178463e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.512057958599133,
      "tokens_seen": 1668349952
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.987963408762638e-05,
      "loss": 1.3833,
      "theoretical_loss": 3.512034318827955,
      "tokens_seen": 1668481024
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.987160969346815e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.5120106814337273,
      "tokens_seen": 1668612096
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.98635852993099e-05,
      "loss": 1.3833,
      "theoretical_loss": 3.511987046416024,
      "tokens_seen": 1668743168
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.985556090515167e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.5119634137744193,
      "tokens_seen": 1668874240
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.984753651099342e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.5119397835084882,
      "tokens_seen": 1669005312
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.983951211683519e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.5119161556178047,
      "tokens_seen": 1669136384
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.983148772267694e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.511892530101944,
      "tokens_seen": 1669267456
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.982346332851871e-05,
      "loss": 1.3642,
      "theoretical_loss": 3.511868906960481,
      "tokens_seen": 1669398528
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.981543893436046e-05,
      "loss": 1.3893,
      "theoretical_loss": 3.5118452861929907,
      "tokens_seen": 1669529600
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.9960150122642517,
      "objective/train/avg_token_score": 0.003985044080764055,
      "objective/train/docs_used": 609762,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4583529233932495,
      "objective/train/original_loss": 1.4649226665496826,
      "objective/train/theoretical_loss": 3.5118216677990484,
      "objective/train/tokens_used": 40119776,
      "objective/train/unlikelihood_loss": 0.011458165012300014,
      "theoretical_loss": 3.5118216677990484,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.980741454020222e-05,
      "loss": 1.3846,
      "theoretical_loss": 3.5118216677990484,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.979939014604398e-05,
      "loss": 1.345,
      "theoretical_loss": 3.5117980517782286,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.979136575188574e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.5117744381301064,
      "tokens_seen": 1669922816
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.97833413577275e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.5117508268542585,
      "tokens_seen": 1670053888
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.977531696356926e-05,
      "loss": 1.4021,
      "theoretical_loss": 3.511727217950259,
      "tokens_seen": 1670184960
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.976729256941102e-05,
      "loss": 1.345,
      "theoretical_loss": 3.511703611417685,
      "tokens_seen": 1670316032
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.975926817525277e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.5116800072561114,
      "tokens_seen": 1670447104
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.975124378109453e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.5116564054651134,
      "tokens_seen": 1670578176
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.97432193869363e-05,
      "loss": 1.3499,
      "theoretical_loss": 3.5116328060442683,
      "tokens_seen": 1670709248
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.973519499277805e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.511609208993151,
      "tokens_seen": 1670840320
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.972717059861981e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.511585614311339,
      "tokens_seen": 1670971392
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.971914620446157e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.511562021998407,
      "tokens_seen": 1671102464
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.971112181030333e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.5115384320539325,
      "tokens_seen": 1671233536
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.970309741614509e-05,
      "loss": 1.3543,
      "theoretical_loss": 3.5115148444774915,
      "tokens_seen": 1671364608
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.969507302198684e-05,
      "loss": 1.3687,
      "theoretical_loss": 3.511491259268661,
      "tokens_seen": 1671495680
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.968704862782861e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.5114676764270176,
      "tokens_seen": 1671626752
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.967902423367036e-05,
      "loss": 1.3407,
      "theoretical_loss": 3.511444095952138,
      "tokens_seen": 1671757824
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.967099983951213e-05,
      "loss": 1.3959,
      "theoretical_loss": 3.5114205178435984,
      "tokens_seen": 1671888896
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.966297544535388e-05,
      "loss": 1.297,
      "theoretical_loss": 3.5113969421009776,
      "tokens_seen": 1672019968
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.965495105119565e-05,
      "loss": 1.349,
      "theoretical_loss": 3.5113733687238513,
      "tokens_seen": 1672151040
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.96469266570374e-05,
      "loss": 1.3683,
      "theoretical_loss": 3.5113497977117976,
      "tokens_seen": 1672282112
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.963890226287915e-05,
      "loss": 1.302,
      "theoretical_loss": 3.5113262290643936,
      "tokens_seen": 1672413184
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.963087786872092e-05,
      "loss": 1.339,
      "theoretical_loss": 3.5113026627812163,
      "tokens_seen": 1672544256
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.962285347456267e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.5112790988618436,
      "tokens_seen": 1672675328
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.961482908040444e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.5112555373058534,
      "tokens_seen": 1672806400
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.960680468624619e-05,
      "loss": 1.353,
      "theoretical_loss": 3.511231978112823,
      "tokens_seen": 1672937472
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.959878029208796e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.5112084212823316,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.959075589792971e-05,
      "loss": 1.3409,
      "theoretical_loss": 3.511184866813956,
      "tokens_seen": 1673199616
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.958273150377147e-05,
      "loss": 1.3938,
      "theoretical_loss": 3.5111613147072744,
      "tokens_seen": 1673330688
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.957470710961323e-05,
      "loss": 1.3197,
      "theoretical_loss": 3.5111377649618656,
      "tokens_seen": 1673461760
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.956668271545499e-05,
      "loss": 1.3336,
      "theoretical_loss": 3.5111142175773074,
      "tokens_seen": 1673592832
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.955865832129675e-05,
      "loss": 1.288,
      "theoretical_loss": 3.5110906725531787,
      "tokens_seen": 1673723904
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.95506339271385e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.5110671298890574,
      "tokens_seen": 1673854976
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.954260953298027e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.5110435895845233,
      "tokens_seen": 1673986048
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.953458513882203e-05,
      "loss": 1.3558,
      "theoretical_loss": 3.511020051639154,
      "tokens_seen": 1674117120
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.952656074466378e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.510996516052529,
      "tokens_seen": 1674248192
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.951853635050554e-05,
      "loss": 1.3837,
      "theoretical_loss": 3.5109729828242275,
      "tokens_seen": 1674379264
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.95105119563473e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.5109494519538282,
      "tokens_seen": 1674510336
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.950248756218906e-05,
      "loss": 1.4078,
      "theoretical_loss": 3.5109259234409107,
      "tokens_seen": 1674641408
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.949446316803082e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.5109023972850535,
      "tokens_seen": 1674772480
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.948643877387258e-05,
      "loss": 1.298,
      "theoretical_loss": 3.510878873485837,
      "tokens_seen": 1674903552
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.947841437971434e-05,
      "loss": 1.318,
      "theoretical_loss": 3.51085535204284,
      "tokens_seen": 1675034624
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.94703899855561e-05,
      "loss": 1.3537,
      "theoretical_loss": 3.5108318329556427,
      "tokens_seen": 1675165696
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.946236559139786e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.510808316223825,
      "tokens_seen": 1675296768
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.945434119723961e-05,
      "loss": 1.2197,
      "theoretical_loss": 3.510784801846966,
      "tokens_seen": 1675427840
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.944631680308138e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.5107612898246456,
      "tokens_seen": 1675558912
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.943829240892313e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.510737780156445,
      "tokens_seen": 1675689984
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.94302680147649e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.5107142728419434,
      "tokens_seen": 1675821056
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.942224362060665e-05,
      "loss": 1.3759,
      "theoretical_loss": 3.510690767880721,
      "tokens_seen": 1675952128
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.941421922644842e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.5106672652723594,
      "tokens_seen": 1676083200
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.9904825091362,
      "objective/train/avg_token_score": 0.009517486207187176,
      "objective/train/docs_used": 612232,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2982546091079712,
      "objective/train/original_loss": 1.3121106624603271,
      "objective/train/theoretical_loss": 3.5106437650164377,
      "objective/train/tokens_used": 46673376,
      "objective/train/unlikelihood_loss": 0.028360089287161827,
      "theoretical_loss": 3.5106437650164377,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.940619483229017e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.5106437650164377,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.939817043813192e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.5106202671125377,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.939014604397369e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.5105967715602393,
      "tokens_seen": 1676476416
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.938212164981544e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.5105732783591237,
      "tokens_seen": 1676607488
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.937409725565721e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.5105497875087717,
      "tokens_seen": 1676738560
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.936607286149896e-05,
      "loss": 1.287,
      "theoretical_loss": 3.510526299008764,
      "tokens_seen": 1676869632
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.935804846734073e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.5105028128586824,
      "tokens_seen": 1677000704
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.935002407318248e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.5104793290581084,
      "tokens_seen": 1677131776
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.934199967902424e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.510455847606622,
      "tokens_seen": 1677262848
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.9333975284866e-05,
      "loss": 1.3866,
      "theoretical_loss": 3.510432368503806,
      "tokens_seen": 1677393920
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.932595089070776e-05,
      "loss": 1.2053,
      "theoretical_loss": 3.5104088917492415,
      "tokens_seen": 1677524992
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.931792649654952e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.51038541734251,
      "tokens_seen": 1677656064
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.930990210239128e-05,
      "loss": 1.263,
      "theoretical_loss": 3.5103619452831936,
      "tokens_seen": 1677787136
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.930187770823304e-05,
      "loss": 1.369,
      "theoretical_loss": 3.510338475570874,
      "tokens_seen": 1677918208
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.92938533140748e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.5103150082051338,
      "tokens_seen": 1678049280
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.928582891991655e-05,
      "loss": 1.3743,
      "theoretical_loss": 3.5102915431855544,
      "tokens_seen": 1678180352
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.927780452575832e-05,
      "loss": 1.3359,
      "theoretical_loss": 3.510268080511718,
      "tokens_seen": 1678311424
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.926978013160007e-05,
      "loss": 1.3659,
      "theoretical_loss": 3.510244620183207,
      "tokens_seen": 1678442496
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.926175573744183e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.510221162199605,
      "tokens_seen": 1678573568
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.925373134328359e-05,
      "loss": 1.334,
      "theoretical_loss": 3.5101977065604926,
      "tokens_seen": 1678704640
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.924570694912535e-05,
      "loss": 1.3247,
      "theoretical_loss": 3.510174253265454,
      "tokens_seen": 1678835712
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.923768255496711e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.510150802314071,
      "tokens_seen": 1678966784
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.922965816080886e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.5101273537059274,
      "tokens_seen": 1679097856
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.922163376665063e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.5101039074406053,
      "tokens_seen": 1679228928
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.921360937249238e-05,
      "loss": 1.3335,
      "theoretical_loss": 3.5100804635176877,
      "tokens_seen": 1679360000
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.920558497833415e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.510057021936759,
      "tokens_seen": 1679491072
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.91975605841759e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.510033582697401,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.918953619001767e-05,
      "loss": 1.3386,
      "theoretical_loss": 3.510010145799198,
      "tokens_seen": 1679753216
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.918151179585942e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.5099867112417336,
      "tokens_seen": 1679884288
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.917348740170119e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.5099632790245905,
      "tokens_seen": 1680015360
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.916546300754294e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.5099398491473535,
      "tokens_seen": 1680146432
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.91574386133847e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.5099164216096055,
      "tokens_seen": 1680277504
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.914941421922646e-05,
      "loss": 1.2155,
      "theoretical_loss": 3.5098929964109313,
      "tokens_seen": 1680408576
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.914138982506821e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.5098695735509144,
      "tokens_seen": 1680539648
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.913336543090998e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.509846153029139,
      "tokens_seen": 1680670720
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.912534103675173e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.50982273484519,
      "tokens_seen": 1680801792
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.91173166425935e-05,
      "loss": 1.3859,
      "theoretical_loss": 3.5097993189986507,
      "tokens_seen": 1680932864
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.910929224843525e-05,
      "loss": 1.3758,
      "theoretical_loss": 3.5097759054891062,
      "tokens_seen": 1681063936
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.9101267854277e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.5097524943161407,
      "tokens_seen": 1681195008
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.909324346011877e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.5097290854793393,
      "tokens_seen": 1681326080
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.908521906596053e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.5097056789782863,
      "tokens_seen": 1681457152
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.907719467180229e-05,
      "loss": 1.3738,
      "theoretical_loss": 3.509682274812567,
      "tokens_seen": 1681588224
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.906917027764405e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.509658872981767,
      "tokens_seen": 1681719296
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.906114588348581e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.50963547348547,
      "tokens_seen": 1681850368
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.905312148932757e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.509612076323261,
      "tokens_seen": 1681981440
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.904509709516932e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.5095886814947277,
      "tokens_seen": 1682112512
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.903707270101109e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.509565288999453,
      "tokens_seen": 1682243584
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.902904830685284e-05,
      "loss": 1.3751,
      "theoretical_loss": 3.509541898837024,
      "tokens_seen": 1682374656
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.90210239126946e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.509518511007026,
      "tokens_seen": 1682505728
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.901299951853636e-05,
      "loss": 1.2931,
      "theoretical_loss": 3.5094951255090443,
      "tokens_seen": 1682636800
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.9924430251121521,
      "objective/train/avg_token_score": 0.00755696464329958,
      "objective/train/docs_used": 614786,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2526493072509766,
      "objective/train/original_loss": 1.2621850967407227,
      "objective/train/theoretical_loss": 3.5094717423426642,
      "objective/train/tokens_used": 53226976,
      "objective/train/unlikelihood_loss": 0.024380359798669815,
      "theoretical_loss": 3.5094717423426642,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.900497512437812e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.5094717423426642,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.899695073021988e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.5094483615074736,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.898892633606163e-05,
      "loss": 1.34,
      "theoretical_loss": 3.5094249830030564,
      "tokens_seen": 1683030016
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.89809019419034e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.509401606829,
      "tokens_seen": 1683161088
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.897287754774515e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.509378232984891,
      "tokens_seen": 1683292160
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.896485315358692e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.509354861470314,
      "tokens_seen": 1683423232
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.895682875942867e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.5093314922848577,
      "tokens_seen": 1683554304
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.894880436527044e-05,
      "loss": 1.3312,
      "theoretical_loss": 3.509308125428107,
      "tokens_seen": 1683685376
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.894077997111219e-05,
      "loss": 1.3576,
      "theoretical_loss": 3.509284760899649,
      "tokens_seen": 1683816448
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.893275557695394e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.5092613986990715,
      "tokens_seen": 1683947520
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.892473118279571e-05,
      "loss": 1.4084,
      "theoretical_loss": 3.50923803882596,
      "tokens_seen": 1684078592
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.891670678863746e-05,
      "loss": 1.3382,
      "theoretical_loss": 3.5092146812799023,
      "tokens_seen": 1684209664
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.890868239447923e-05,
      "loss": 1.3659,
      "theoretical_loss": 3.5091913260604857,
      "tokens_seen": 1684340736
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.890065800032098e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.5091679731672967,
      "tokens_seen": 1684471808
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.889263360616275e-05,
      "loss": 1.3693,
      "theoretical_loss": 3.5091446225999228,
      "tokens_seen": 1684602880
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.88846092120045e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.5091212743579514,
      "tokens_seen": 1684733952
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.887658481784626e-05,
      "loss": 1.3886,
      "theoretical_loss": 3.5090979284409705,
      "tokens_seen": 1684865024
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.886856042368802e-05,
      "loss": 1.4775,
      "theoretical_loss": 3.509074584848568,
      "tokens_seen": 1684996096
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.886053602952978e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.50905124358033,
      "tokens_seen": 1685127168
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.885251163537154e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.509027904635846,
      "tokens_seen": 1685258240
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.88444872412133e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.5090045680147037,
      "tokens_seen": 1685389312
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.883646284705506e-05,
      "loss": 1.3516,
      "theoretical_loss": 3.50898123371649,
      "tokens_seen": 1685520384
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.882843845289682e-05,
      "loss": 1.3789,
      "theoretical_loss": 3.5089579017407946,
      "tokens_seen": 1685651456
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.882041405873857e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.508934572087205,
      "tokens_seen": 1685782528
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.881238966458034e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.5089112447553097,
      "tokens_seen": 1685913600
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.880436527042209e-05,
      "loss": 1.3824,
      "theoretical_loss": 3.5088879197446974,
      "tokens_seen": 1686044672
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.879634087626386e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.5088645970549557,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.878831648210561e-05,
      "loss": 1.3427,
      "theoretical_loss": 3.5088412766856747,
      "tokens_seen": 1686306816
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.878029208794736e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.508817958636442,
      "tokens_seen": 1686437888
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.877226769378913e-05,
      "loss": 1.3626,
      "theoretical_loss": 3.5087946429068477,
      "tokens_seen": 1686568960
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.876424329963088e-05,
      "loss": 1.3814,
      "theoretical_loss": 3.5087713294964793,
      "tokens_seen": 1686700032
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.875621890547265e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.508748018404927,
      "tokens_seen": 1686831104
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.87481945113144e-05,
      "loss": 1.3629,
      "theoretical_loss": 3.50872470963178,
      "tokens_seen": 1686962176
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.874017011715617e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.5087014031766266,
      "tokens_seen": 1687093248
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.873214572299792e-05,
      "loss": 1.2398,
      "theoretical_loss": 3.5086780990390576,
      "tokens_seen": 1687224320
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.872412132883967e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.5086547972186617,
      "tokens_seen": 1687355392
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.871609693468144e-05,
      "loss": 1.3496,
      "theoretical_loss": 3.5086314977150286,
      "tokens_seen": 1687486464
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.87080725405232e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.508608200527749,
      "tokens_seen": 1687617536
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.870004814636496e-05,
      "loss": 1.3601,
      "theoretical_loss": 3.5085849056564107,
      "tokens_seen": 1687748608
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.869202375220671e-05,
      "loss": 1.288,
      "theoretical_loss": 3.508561613100605,
      "tokens_seen": 1687879680
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.868399935804847e-05,
      "loss": 1.4035,
      "theoretical_loss": 3.5085383228599225,
      "tokens_seen": 1688010752
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.867597496389023e-05,
      "loss": 1.3797,
      "theoretical_loss": 3.5085150349339522,
      "tokens_seen": 1688141824
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.866795056973199e-05,
      "loss": 1.3746,
      "theoretical_loss": 3.5084917493222845,
      "tokens_seen": 1688272896
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.865992617557375e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.5084684660245107,
      "tokens_seen": 1688403968
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.86519017814155e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.5084451850402205,
      "tokens_seen": 1688535040
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.864387738725727e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.508421906369004,
      "tokens_seen": 1688666112
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.863585299309903e-05,
      "loss": 1.28,
      "theoretical_loss": 3.5083986300104533,
      "tokens_seen": 1688797184
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.862782859894078e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.5083753559641577,
      "tokens_seen": 1688928256
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.861980420478255e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.508352084229709,
      "tokens_seen": 1689059328
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.86117798106243e-05,
      "loss": 1.3458,
      "theoretical_loss": 3.508328814806698,
      "tokens_seen": 1689190400
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.9927401542663574,
      "objective/train/avg_token_score": 0.00725984014570713,
      "objective/train/docs_used": 617011,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.447295904159546,
      "objective/train/original_loss": 1.4584932327270508,
      "objective/train/theoretical_loss": 3.5083055476947154,
      "objective/train/tokens_used": 59780576,
      "objective/train/unlikelihood_loss": 0.021061142906546593,
      "theoretical_loss": 3.5083055476947154,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.860375541646607e-05,
      "loss": 1.3823,
      "theoretical_loss": 3.5083055476947154,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.859573102230782e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.5082822828933526,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.858770662814957e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.5082590204022015,
      "tokens_seen": 1689583616
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.857968223399134e-05,
      "loss": 1.3292,
      "theoretical_loss": 3.5082357602208525,
      "tokens_seen": 1689714688
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.857165783983309e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.508212502348898,
      "tokens_seen": 1689845760
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.856363344567486e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.5081892467859292,
      "tokens_seen": 1689976832
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.855560905151661e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.5081659935315375,
      "tokens_seen": 1690107904
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.854758465735838e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.5081427425853153,
      "tokens_seen": 1690238976
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.853956026320013e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.508119493946854,
      "tokens_seen": 1690370048
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.853153586904188e-05,
      "loss": 1.3563,
      "theoretical_loss": 3.5080962476157462,
      "tokens_seen": 1690501120
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.852351147488365e-05,
      "loss": 1.4576,
      "theoretical_loss": 3.508073003591584,
      "tokens_seen": 1690632192
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.85154870807254e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.508049761873959,
      "tokens_seen": 1690763264
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.850746268656717e-05,
      "loss": 1.3602,
      "theoretical_loss": 3.5080265224624636,
      "tokens_seen": 1690894336
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.849943829240892e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.508003285356691,
      "tokens_seen": 1691025408
    },
    {
      "epoch": 0.02,
      "learning_rate": 9.849141389825068e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.507980050556233,
      "tokens_seen": 1691156480
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.848338950409244e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.507956818060683,
      "tokens_seen": 1691287552
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.84753651099342e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.507933587869633,
      "tokens_seen": 1691418624
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.846734071577596e-05,
      "loss": 1.4032,
      "theoretical_loss": 3.507910359982676,
      "tokens_seen": 1691549696
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.845931632161772e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.507887134399405,
      "tokens_seen": 1691680768
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.845129192745948e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.5078639111194136,
      "tokens_seen": 1691811840
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.844326753330124e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.5078406901422943,
      "tokens_seen": 1691942912
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.843524313914299e-05,
      "loss": 1.339,
      "theoretical_loss": 3.5078174714676407,
      "tokens_seen": 1692073984
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.842721874498476e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.507794255095046,
      "tokens_seen": 1692205056
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.841919435082651e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.5077710410241036,
      "tokens_seen": 1692336128
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.841116995666828e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.5077478292544075,
      "tokens_seen": 1692467200
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.840314556251003e-05,
      "loss": 1.292,
      "theoretical_loss": 3.5077246197855505,
      "tokens_seen": 1692598272
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.839512116835178e-05,
      "loss": 1.3546,
      "theoretical_loss": 3.507701412617128,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.838709677419355e-05,
      "loss": 1.3629,
      "theoretical_loss": 3.5076782077487323,
      "tokens_seen": 1692860416
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.83790723800353e-05,
      "loss": 1.322,
      "theoretical_loss": 3.5076550051799575,
      "tokens_seen": 1692991488
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.837104798587707e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.5076318049103987,
      "tokens_seen": 1693122560
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.836302359171882e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.50760860693965,
      "tokens_seen": 1693253632
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.835499919756059e-05,
      "loss": 1.3536,
      "theoretical_loss": 3.5075854112673044,
      "tokens_seen": 1693384704
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.834697480340234e-05,
      "loss": 1.302,
      "theoretical_loss": 3.507562217892958,
      "tokens_seen": 1693515776
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.83389504092441e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.507539026816204,
      "tokens_seen": 1693646848
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.833092601508586e-05,
      "loss": 1.3574,
      "theoretical_loss": 3.507515838036637,
      "tokens_seen": 1693777920
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.832290162092762e-05,
      "loss": 1.3849,
      "theoretical_loss": 3.5074926515538527,
      "tokens_seen": 1693908992
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.831487722676938e-05,
      "loss": 1.3131,
      "theoretical_loss": 3.5074694673674456,
      "tokens_seen": 1694040064
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.830685283261113e-05,
      "loss": 1.4134,
      "theoretical_loss": 3.5074462854770103,
      "tokens_seen": 1694171136
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.82988284384529e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.507423105882142,
      "tokens_seen": 1694302208
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.829080404429465e-05,
      "loss": 1.3751,
      "theoretical_loss": 3.5073999285824353,
      "tokens_seen": 1694433280
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.828277965013641e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.5073767535774865,
      "tokens_seen": 1694564352
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.827475525597817e-05,
      "loss": 1.3761,
      "theoretical_loss": 3.50735358086689,
      "tokens_seen": 1694695424
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.826673086181993e-05,
      "loss": 1.3201,
      "theoretical_loss": 3.507330410450242,
      "tokens_seen": 1694826496
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.82587064676617e-05,
      "loss": 1.375,
      "theoretical_loss": 3.507307242327137,
      "tokens_seen": 1694957568
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.825068207350345e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.5072840764971716,
      "tokens_seen": 1695088640
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.824265767934521e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.507260912959941,
      "tokens_seen": 1695219712
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.823463328518697e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.5072377517150413,
      "tokens_seen": 1695350784
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.822660889102873e-05,
      "loss": 1.319,
      "theoretical_loss": 3.507214592762068,
      "tokens_seen": 1695481856
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.821858449687049e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.507191436100618,
      "tokens_seen": 1695612928
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.821056010271224e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.507168281730287,
      "tokens_seen": 1695744000
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.9919037222862244,
      "objective/train/avg_token_score": 0.008096263743937016,
      "objective/train/docs_used": 619583,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.265417456626892,
      "objective/train/original_loss": 1.275237798690796,
      "objective/train/theoretical_loss": 3.507145129650671,
      "objective/train/tokens_used": 66334176,
      "objective/train/unlikelihood_loss": 0.025167396292090416,
      "theoretical_loss": 3.507145129650671,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.820253570855401e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.507145129650671,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.819451131439576e-05,
      "loss": 1.4031,
      "theoretical_loss": 3.507121979861367,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.818648692023753e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.50709883236197,
      "tokens_seen": 1696137216
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.817846252607928e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.5070756871520783,
      "tokens_seen": 1696268288
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.817043813192105e-05,
      "loss": 1.3567,
      "theoretical_loss": 3.5070525442312874,
      "tokens_seen": 1696399360
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.81624137377628e-05,
      "loss": 1.391,
      "theoretical_loss": 3.507029403599195,
      "tokens_seen": 1696530432
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.815438934360455e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.5070062652553977,
      "tokens_seen": 1696661504
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.814636494944632e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.5069831291994915,
      "tokens_seen": 1696792576
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.813834055528807e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.5069599954310746,
      "tokens_seen": 1696923648
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.813031616112984e-05,
      "loss": 1.2181,
      "theoretical_loss": 3.5069368639497434,
      "tokens_seen": 1697054720
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.812229176697159e-05,
      "loss": 1.382,
      "theoretical_loss": 3.506913734755096,
      "tokens_seen": 1697185792
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.811426737281336e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.506890607846729,
      "tokens_seen": 1697316864
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.810624297865511e-05,
      "loss": 1.2105,
      "theoretical_loss": 3.5068674832242404,
      "tokens_seen": 1697447936
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.809821858449687e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.5068443608872277,
      "tokens_seen": 1697579008
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.809019419033863e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.506821240835288,
      "tokens_seen": 1697710080
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.808216979618039e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.5067981230680196,
      "tokens_seen": 1697841152
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.807414540202215e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.506775007585021,
      "tokens_seen": 1697972224
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.80661210078639e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.5067518943858893,
      "tokens_seen": 1698103296
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.805809661370567e-05,
      "loss": 1.3855,
      "theoretical_loss": 3.5067287834702223,
      "tokens_seen": 1698234368
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.805007221954742e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.506705674837619,
      "tokens_seen": 1698365440
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.804204782538918e-05,
      "loss": 1.4161,
      "theoretical_loss": 3.5066825684876775,
      "tokens_seen": 1698496512
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.803402343123094e-05,
      "loss": 1.2931,
      "theoretical_loss": 3.5066594644199958,
      "tokens_seen": 1698627584
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.80259990370727e-05,
      "loss": 1.3969,
      "theoretical_loss": 3.5066363626341728,
      "tokens_seen": 1698758656
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.801797464291446e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.5066132631298075,
      "tokens_seen": 1698889728
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.800995024875622e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.5065901659064975,
      "tokens_seen": 1699020800
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.800192585459798e-05,
      "loss": 1.353,
      "theoretical_loss": 3.5065670709638423,
      "tokens_seen": 1699151872
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.799390146043974e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.5065439783014405,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.798587706628149e-05,
      "loss": 1.266,
      "theoretical_loss": 3.5065208879188914,
      "tokens_seen": 1699414016
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.797785267212326e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.506497799815794,
      "tokens_seen": 1699545088
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.796982827796501e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.5064747139917474,
      "tokens_seen": 1699676160
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.796180388380678e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.5064516304463513,
      "tokens_seen": 1699807232
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.795377948964853e-05,
      "loss": 1.3242,
      "theoretical_loss": 3.5064285491792044,
      "tokens_seen": 1699938304
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.79457550954903e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.506405470189907,
      "tokens_seen": 1700069376
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.793773070133205e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.5063823934780585,
      "tokens_seen": 1700200448
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.792970630717382e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.5063593190432574,
      "tokens_seen": 1700331520
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.792168191301557e-05,
      "loss": 1.2311,
      "theoretical_loss": 3.5063362468851054,
      "tokens_seen": 1700462592
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.791365751885732e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.506313177003202,
      "tokens_seen": 1700593664
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.790563312469909e-05,
      "loss": 1.3977,
      "theoretical_loss": 3.506290109397146,
      "tokens_seen": 1700724736
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.789760873054084e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.506267044066538,
      "tokens_seen": 1700855808
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.788958433638261e-05,
      "loss": 1.27,
      "theoretical_loss": 3.5062439810109796,
      "tokens_seen": 1700986880
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.788155994222436e-05,
      "loss": 1.4285,
      "theoretical_loss": 3.5062209202300694,
      "tokens_seen": 1701117952
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.787353554806613e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.5061978617234084,
      "tokens_seen": 1701249024
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.786551115390788e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.5061748054905975,
      "tokens_seen": 1701380096
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.785748675974964e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.506151751531237,
      "tokens_seen": 1701511168
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.78494623655914e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.5061286998449273,
      "tokens_seen": 1701642240
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.784143797143316e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.50610565043127,
      "tokens_seen": 1701773312
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.783341357727492e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.5060826032898653,
      "tokens_seen": 1701904384
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.782538918311668e-05,
      "loss": 1.3746,
      "theoretical_loss": 3.506059558420314,
      "tokens_seen": 1702035456
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.781736478895844e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.5060365158222186,
      "tokens_seen": 1702166528
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.78093403948002e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.506013475495179,
      "tokens_seen": 1702297600
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.9905288815498352,
      "objective/train/avg_token_score": 0.009471178986132145,
      "objective/train/docs_used": 621885,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2759546041488647,
      "objective/train/original_loss": 1.2892868518829346,
      "objective/train/theoretical_loss": 3.5059904374387973,
      "objective/train/tokens_used": 72887776,
      "objective/train/unlikelihood_loss": 0.026637762784957886,
      "theoretical_loss": 3.5059904374387973,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.780131600064195e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.5059904374387973,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.779329160648371e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.5059674016526747,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.778526721232547e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.5059443681364124,
      "tokens_seen": 1702690816
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.777724281816723e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.505921336889612,
      "tokens_seen": 1702821888
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.776921842400899e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.5058983079118757,
      "tokens_seen": 1702952960
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.776119402985075e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.505875281202805,
      "tokens_seen": 1703084032
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.775316963569251e-05,
      "loss": 1.316,
      "theoretical_loss": 3.5058522567620023,
      "tokens_seen": 1703215104
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.774514524153426e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.5058292345890694,
      "tokens_seen": 1703346176
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.773712084737603e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.505806214683608,
      "tokens_seen": 1703477248
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.772909645321778e-05,
      "loss": 1.4099,
      "theoretical_loss": 3.5057831970452207,
      "tokens_seen": 1703608320
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.772107205905955e-05,
      "loss": 1.3701,
      "theoretical_loss": 3.5057601816735096,
      "tokens_seen": 1703739392
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.77130476649013e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.5057371685680776,
      "tokens_seen": 1703870464
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.770502327074307e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.5057141577285273,
      "tokens_seen": 1704001536
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.769699887658482e-05,
      "loss": 1.2673,
      "theoretical_loss": 3.5056911491544605,
      "tokens_seen": 1704132608
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.768897448242657e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.5056681428454803,
      "tokens_seen": 1704263680
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.768095008826834e-05,
      "loss": 1.4026,
      "theoretical_loss": 3.5056451388011904,
      "tokens_seen": 1704394752
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.767292569411009e-05,
      "loss": 1.279,
      "theoretical_loss": 3.5056221370211924,
      "tokens_seen": 1704525824
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.766490129995186e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.50559913750509,
      "tokens_seen": 1704656896
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.765687690579361e-05,
      "loss": 1.3954,
      "theoretical_loss": 3.5055761402524865,
      "tokens_seen": 1704787968
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.764885251163538e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.5055531452629847,
      "tokens_seen": 1704919040
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.764082811747713e-05,
      "loss": 1.345,
      "theoretical_loss": 3.5055301525361884,
      "tokens_seen": 1705050112
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.763280372331889e-05,
      "loss": 1.2927,
      "theoretical_loss": 3.5055071620717007,
      "tokens_seen": 1705181184
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.762477932916065e-05,
      "loss": 1.3867,
      "theoretical_loss": 3.505484173869125,
      "tokens_seen": 1705312256
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.76167549350024e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.505461187928065,
      "tokens_seen": 1705443328
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.760873054084417e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.505438204248125,
      "tokens_seen": 1705574400
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.760070614668593e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.5054152228289084,
      "tokens_seen": 1705705472
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.759268175252769e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.505392243670019,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.758465735836945e-05,
      "loss": 1.3623,
      "theoretical_loss": 3.5053692667710608,
      "tokens_seen": 1705967616
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.757663296421121e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.5053462921316383,
      "tokens_seen": 1706098688
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.756860857005296e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.505323319751356,
      "tokens_seen": 1706229760
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.756058417589472e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.505300349629817,
      "tokens_seen": 1706360832
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.755255978173648e-05,
      "loss": 1.2105,
      "theoretical_loss": 3.5052773817666267,
      "tokens_seen": 1706491904
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.754453538757824e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.5052544161613897,
      "tokens_seen": 1706622976
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.753651099342e-05,
      "loss": 1.3635,
      "theoretical_loss": 3.5052314528137103,
      "tokens_seen": 1706754048
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.752848659926176e-05,
      "loss": 1.3632,
      "theoretical_loss": 3.5052084917231934,
      "tokens_seen": 1706885120
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.752046220510352e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.5051855328894437,
      "tokens_seen": 1707016192
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.751243781094528e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.5051625763120664,
      "tokens_seen": 1707147264
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.750441341678703e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.505139621990666,
      "tokens_seen": 1707278336
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.74963890226288e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.505116669924848,
      "tokens_seen": 1707409408
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.748836462847055e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.5050937201142176,
      "tokens_seen": 1707540480
    },
    {
      "epoch": 0.03,
      "learning_rate": 9.748034023431232e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.5050707725583803,
      "tokens_seen": 1707671552
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.747231584015407e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.5050478272569414,
      "tokens_seen": 1707802624
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.746429144599584e-05,
      "loss": 1.3207,
      "theoretical_loss": 3.505024884209506,
      "tokens_seen": 1707933696
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.745626705183759e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.50500194341568,
      "tokens_seen": 1708064768
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.744824265767934e-05,
      "loss": 1.262,
      "theoretical_loss": 3.5049790048750697,
      "tokens_seen": 1708195840
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.744021826352111e-05,
      "loss": 1.342,
      "theoretical_loss": 3.50495606858728,
      "tokens_seen": 1708326912
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.743219386936286e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.5049331345519175,
      "tokens_seen": 1708457984
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.742416947520463e-05,
      "loss": 1.2437,
      "theoretical_loss": 3.504910202768588,
      "tokens_seen": 1708589056
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.741614508104638e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.5048872732368976,
      "tokens_seen": 1708720128
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.740812068688815e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.5048643459564524,
      "tokens_seen": 1708851200
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.9949111342430115,
      "objective/train/avg_token_score": 0.005088924430310726,
      "objective/train/docs_used": 624223,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.339620590209961,
      "objective/train/original_loss": 1.3469386100769043,
      "objective/train/theoretical_loss": 3.5048414209268586,
      "objective/train/tokens_used": 79441376,
      "objective/train/unlikelihood_loss": 0.013174877502024174,
      "theoretical_loss": 3.5048414209268586,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.74000962927299e-05,
      "loss": 1.281,
      "theoretical_loss": 3.5048414209268586,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.739207189857166e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.5048184981477233,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.738404750441342e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.504795577618653,
      "tokens_seen": 1709244416
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.737602311025518e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.5047726593392534,
      "tokens_seen": 1709375488
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.736799871609694e-05,
      "loss": 1.302,
      "theoretical_loss": 3.5047497433091315,
      "tokens_seen": 1709506560
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.73599743219387e-05,
      "loss": 1.2498,
      "theoretical_loss": 3.5047268295278946,
      "tokens_seen": 1709637632
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.735194992778046e-05,
      "loss": 1.3298,
      "theoretical_loss": 3.5047039179951494,
      "tokens_seen": 1709768704
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.734392553362222e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.5046810087105027,
      "tokens_seen": 1709899776
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.733590113946397e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.504658101673562,
      "tokens_seen": 1710030848
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.732787674530574e-05,
      "loss": 1.3807,
      "theoretical_loss": 3.504635196883934,
      "tokens_seen": 1710161920
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.731985235114749e-05,
      "loss": 1.3211,
      "theoretical_loss": 3.504612294341227,
      "tokens_seen": 1710292992
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.731182795698925e-05,
      "loss": 1.3521,
      "theoretical_loss": 3.5045893940450465,
      "tokens_seen": 1710424064
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.730380356283101e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.504566495995002,
      "tokens_seen": 1710555136
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.729577916867277e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.5045436001907007,
      "tokens_seen": 1710686208
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.728775477451453e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.5045207066317494,
      "tokens_seen": 1710817280
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.72797303803563e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.5044978153177566,
      "tokens_seen": 1710948352
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.727170598619805e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.5044749262483297,
      "tokens_seen": 1711079424
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.72636815920398e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.504452039423078,
      "tokens_seen": 1711210496
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.725565719788157e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.5044291548416076,
      "tokens_seen": 1711341568
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.724763280372332e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.504406272503528,
      "tokens_seen": 1711472640
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.723960840956509e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.5043833924084473,
      "tokens_seen": 1711603712
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.723158401540684e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.504360514555974,
      "tokens_seen": 1711734784
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.722355962124861e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.5043376389457164,
      "tokens_seen": 1711865856
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.721553522709036e-05,
      "loss": 1.3653,
      "theoretical_loss": 3.504314765577283,
      "tokens_seen": 1711996928
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.720751083293211e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.5042918944502826,
      "tokens_seen": 1712128000
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.719948643877388e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.504269025564324,
      "tokens_seen": 1712259072
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.719146204461563e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.504246158919016,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.71834376504574e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.5042232945139675,
      "tokens_seen": 1712521216
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.717541325629915e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.504200432348788,
      "tokens_seen": 1712652288
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.716738886214092e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.504177572423086,
      "tokens_seen": 1712783360
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.715936446798267e-05,
      "loss": 1.288,
      "theoretical_loss": 3.504154714736472,
      "tokens_seen": 1712914432
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.715134007382443e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.504131859288554,
      "tokens_seen": 1713045504
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.714331567966619e-05,
      "loss": 1.3007,
      "theoretical_loss": 3.5041090060789424,
      "tokens_seen": 1713176576
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.713529128550795e-05,
      "loss": 1.3537,
      "theoretical_loss": 3.504086155107246,
      "tokens_seen": 1713307648
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.712726689134971e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.504063306373075,
      "tokens_seen": 1713438720
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.711924249719147e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.5040404598760393,
      "tokens_seen": 1713569792
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.711121810303323e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.504017615615748,
      "tokens_seen": 1713700864
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.710319370887499e-05,
      "loss": 1.332,
      "theoretical_loss": 3.5039947735918124,
      "tokens_seen": 1713831936
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.709516931471674e-05,
      "loss": 1.3618,
      "theoretical_loss": 3.5039719338038413,
      "tokens_seen": 1713963008
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.70871449205585e-05,
      "loss": 1.3639,
      "theoretical_loss": 3.503949096251445,
      "tokens_seen": 1714094080
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.707912052640026e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.503926260934234,
      "tokens_seen": 1714225152
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.707109613224202e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.503903427851819,
      "tokens_seen": 1714356224
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.706307173808378e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.5038805970038105,
      "tokens_seen": 1714487296
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.705504734392554e-05,
      "loss": 1.3197,
      "theoretical_loss": 3.5038577683898184,
      "tokens_seen": 1714618368
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.70470229497673e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.503834942009454,
      "tokens_seen": 1714749440
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.703899855560905e-05,
      "loss": 1.3796,
      "theoretical_loss": 3.5038121178623274,
      "tokens_seen": 1714880512
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.703097416145082e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.5037892959480494,
      "tokens_seen": 1715011584
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.702294976729257e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.503766476266232,
      "tokens_seen": 1715142656
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.701492537313434e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.503743658816485,
      "tokens_seen": 1715273728
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.700690097897609e-05,
      "loss": 1.3864,
      "theoretical_loss": 3.5037208435984204,
      "tokens_seen": 1715404800
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.9915130734443665,
      "objective/train/avg_token_score": 0.008486890234053135,
      "objective/train/docs_used": 626732,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2571014165878296,
      "objective/train/original_loss": 1.2687780857086182,
      "objective/train/theoretical_loss": 3.503698030611649,
      "objective/train/tokens_used": 85994976,
      "objective/train/unlikelihood_loss": 0.023852163925766945,
      "theoretical_loss": 3.503698030611649,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.699887658481786e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.503698030611649,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.699085219065961e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.503675219855782,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.698282779650138e-05,
      "loss": 1.3419,
      "theoretical_loss": 3.5036524113304317,
      "tokens_seen": 1715798016
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.697480340234313e-05,
      "loss": 1.3779,
      "theoretical_loss": 3.5036296050352087,
      "tokens_seen": 1715929088
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.696677900818488e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.5036068009697248,
      "tokens_seen": 1716060160
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.695875461402665e-05,
      "loss": 1.4053,
      "theoretical_loss": 3.5035839991335918,
      "tokens_seen": 1716191232
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.69507302198684e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.5035611995264215,
      "tokens_seen": 1716322304
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.694270582571017e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.503538402147826,
      "tokens_seen": 1716453376
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.693468143155192e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.503515606997417,
      "tokens_seen": 1716584448
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.692665703739369e-05,
      "loss": 1.3579,
      "theoretical_loss": 3.5034928140748067,
      "tokens_seen": 1716715520
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.691863264323544e-05,
      "loss": 1.313,
      "theoretical_loss": 3.5034700233796077,
      "tokens_seen": 1716846592
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.69106082490772e-05,
      "loss": 1.3732,
      "theoretical_loss": 3.5034472349114316,
      "tokens_seen": 1716977664
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.690258385491896e-05,
      "loss": 1.3153,
      "theoretical_loss": 3.5034244486698913,
      "tokens_seen": 1717108736
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.689455946076072e-05,
      "loss": 1.335,
      "theoretical_loss": 3.5034016646545996,
      "tokens_seen": 1717239808
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.688653506660248e-05,
      "loss": 1.3691,
      "theoretical_loss": 3.503378882865168,
      "tokens_seen": 1717370880
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.687851067244424e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.5033561033012104,
      "tokens_seen": 1717501952
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.6870486278286e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.5033333259623394,
      "tokens_seen": 1717633024
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.686246188412776e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.503310550848167,
      "tokens_seen": 1717764096
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.685443748996951e-05,
      "loss": 1.3818,
      "theoretical_loss": 3.503287777958307,
      "tokens_seen": 1717895168
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.684641309581128e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.503265007292372,
      "tokens_seen": 1718026240
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.683838870165303e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.5032422388499755,
      "tokens_seen": 1718157312
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.68303643074948e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.5032194726307306,
      "tokens_seen": 1718288384
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.682233991333655e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.5031967086342504,
      "tokens_seen": 1718419456
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.681431551917831e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.503173946860149,
      "tokens_seen": 1718550528
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.680629112502007e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.5031511873080396,
      "tokens_seen": 1718681600
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.679826673086182e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.503128429977536,
      "tokens_seen": 1718812672
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.679024233670359e-05,
      "loss": 1.3799,
      "theoretical_loss": 3.5031056748682516,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.678221794254534e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.503082921979801,
      "tokens_seen": 1719074816
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.677419354838711e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.5030601713117973,
      "tokens_seen": 1719205888
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.676616915422886e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.503037422863855,
      "tokens_seen": 1719336960
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.675814476007063e-05,
      "loss": 1.3651,
      "theoretical_loss": 3.5030146766355887,
      "tokens_seen": 1719468032
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.675012036591238e-05,
      "loss": 1.3672,
      "theoretical_loss": 3.5029919326266112,
      "tokens_seen": 1719599104
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.674209597175413e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.502969190836538,
      "tokens_seen": 1719730176
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.67340715775959e-05,
      "loss": 1.3982,
      "theoretical_loss": 3.5029464512649833,
      "tokens_seen": 1719861248
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.672604718343765e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.502923713911562,
      "tokens_seen": 1719992320
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.671802278927942e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.502900978775888,
      "tokens_seen": 1720123392
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.670999839512117e-05,
      "loss": 1.286,
      "theoretical_loss": 3.502878245857576,
      "tokens_seen": 1720254464
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.670197400096294e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.5028555151562415,
      "tokens_seen": 1720385536
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.669394960680469e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.502832786671499,
      "tokens_seen": 1720516608
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.668592521264645e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.5028100604029637,
      "tokens_seen": 1720647680
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.667790081848821e-05,
      "loss": 1.3637,
      "theoretical_loss": 3.5027873363502504,
      "tokens_seen": 1720778752
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.666987642432997e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.5027646145129747,
      "tokens_seen": 1720909824
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.666185203017173e-05,
      "loss": 1.4103,
      "theoretical_loss": 3.502741894890751,
      "tokens_seen": 1721040896
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.665382763601349e-05,
      "loss": 1.399,
      "theoretical_loss": 3.502719177483196,
      "tokens_seen": 1721171968
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.664580324185525e-05,
      "loss": 1.4261,
      "theoretical_loss": 3.502696462289925,
      "tokens_seen": 1721303040
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.6637778847697e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.5026737493105524,
      "tokens_seen": 1721434112
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.662975445353877e-05,
      "loss": 1.251,
      "theoretical_loss": 3.5026510385446947,
      "tokens_seen": 1721565184
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.662173005938053e-05,
      "loss": 1.3881,
      "theoretical_loss": 3.5026283299919676,
      "tokens_seen": 1721696256
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.661370566522228e-05,
      "loss": 1.257,
      "theoretical_loss": 3.5026056236519874,
      "tokens_seen": 1721827328
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.660568127106405e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.502582919524369,
      "tokens_seen": 1721958400
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.9925838708877563,
      "objective/train/avg_token_score": 0.00741605693474412,
      "objective/train/docs_used": 629016,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3224713802337646,
      "objective/train/original_loss": 1.3316936492919922,
      "objective/train/theoretical_loss": 3.50256021760873,
      "objective/train/tokens_used": 92548576,
      "objective/train/unlikelihood_loss": 0.021793292835354805,
      "theoretical_loss": 3.50256021760873,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.65976568769058e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.50256021760873,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.658963248274757e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.5025375179046847,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.658160808858932e-05,
      "loss": 1.365,
      "theoretical_loss": 3.502514820411851,
      "tokens_seen": 1722351616
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.657358369443108e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.5024921251298444,
      "tokens_seen": 1722482688
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.656555930027284e-05,
      "loss": 1.4736,
      "theoretical_loss": 3.502469432058281,
      "tokens_seen": 1722613760
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.655753490611459e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.5024467411967786,
      "tokens_seen": 1722744832
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.654951051195636e-05,
      "loss": 1.4182,
      "theoretical_loss": 3.502424052544953,
      "tokens_seen": 1722875904
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.654148611779811e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.502401366102421,
      "tokens_seen": 1723006976
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.653346172363988e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.502378681868799,
      "tokens_seen": 1723138048
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.652543732948163e-05,
      "loss": 1.3512,
      "theoretical_loss": 3.5023559998437053,
      "tokens_seen": 1723269120
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.65174129353234e-05,
      "loss": 1.324,
      "theoretical_loss": 3.5023333200267555,
      "tokens_seen": 1723400192
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.650938854116515e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.502310642417567,
      "tokens_seen": 1723531264
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.65013641470069e-05,
      "loss": 1.3312,
      "theoretical_loss": 3.5022879670157576,
      "tokens_seen": 1723662336
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.649333975284867e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.502265293820944,
      "tokens_seen": 1723793408
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.648531535869042e-05,
      "loss": 1.3335,
      "theoretical_loss": 3.5022426228327443,
      "tokens_seen": 1723924480
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.647729096453219e-05,
      "loss": 1.2391,
      "theoretical_loss": 3.5022199540507755,
      "tokens_seen": 1724055552
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.646926657037394e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.5021972874746554,
      "tokens_seen": 1724186624
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.646124217621571e-05,
      "loss": 1.382,
      "theoretical_loss": 3.5021746231040014,
      "tokens_seen": 1724317696
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.645321778205746e-05,
      "loss": 1.3297,
      "theoretical_loss": 3.502151960938431,
      "tokens_seen": 1724448768
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.644519338789922e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.502129300977563,
      "tokens_seen": 1724579840
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.643716899374098e-05,
      "loss": 1.3676,
      "theoretical_loss": 3.5021066432210146,
      "tokens_seen": 1724710912
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.642914459958274e-05,
      "loss": 1.336,
      "theoretical_loss": 3.5020839876684047,
      "tokens_seen": 1724841984
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.64211202054245e-05,
      "loss": 1.3756,
      "theoretical_loss": 3.5020613343193507,
      "tokens_seen": 1724973056
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.641309581126626e-05,
      "loss": 1.3447,
      "theoretical_loss": 3.502038683173471,
      "tokens_seen": 1725104128
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.640507141710802e-05,
      "loss": 1.341,
      "theoretical_loss": 3.502016034230384,
      "tokens_seen": 1725235200
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.639704702294978e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.5019933874897085,
      "tokens_seen": 1725366272
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.638902262879153e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.5019707429510625,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.63809982346333e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.501948100614065,
      "tokens_seen": 1725628416
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.637297384047505e-05,
      "loss": 1.29,
      "theoretical_loss": 3.5019254604783345,
      "tokens_seen": 1725759488
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.636494944631682e-05,
      "loss": 1.4098,
      "theoretical_loss": 3.5019028225434905,
      "tokens_seen": 1725890560
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.635692505215857e-05,
      "loss": 1.3932,
      "theoretical_loss": 3.5018801868091507,
      "tokens_seen": 1726021632
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.634890065800034e-05,
      "loss": 1.3699,
      "theoretical_loss": 3.501857553274935,
      "tokens_seen": 1726152704
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.634087626384209e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.5018349219404628,
      "tokens_seen": 1726283776
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.633285186968385e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.5018122928053526,
      "tokens_seen": 1726414848
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.632482747552561e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.5017896658692234,
      "tokens_seen": 1726545920
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.631680308136736e-05,
      "loss": 1.4245,
      "theoretical_loss": 3.501767041131696,
      "tokens_seen": 1726676992
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.630877868720913e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.5017444185923887,
      "tokens_seen": 1726808064
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.630075429305088e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.5017217982509212,
      "tokens_seen": 1726939136
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.629272989889265e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.5016991801069137,
      "tokens_seen": 1727070208
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.62847055047344e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.501676564159986,
      "tokens_seen": 1727201280
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.627668111057617e-05,
      "loss": 1.3714,
      "theoretical_loss": 3.5016539504097572,
      "tokens_seen": 1727332352
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.626865671641792e-05,
      "loss": 1.3708,
      "theoretical_loss": 3.501631338855848,
      "tokens_seen": 1727463424
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.626063232225967e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.501608729497878,
      "tokens_seen": 1727594496
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.625260792810144e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.501586122335468,
      "tokens_seen": 1727725568
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.62445835339432e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.5015635173682376,
      "tokens_seen": 1727856640
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.623655913978496e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.5015409145958074,
      "tokens_seen": 1727987712
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.622853474562671e-05,
      "loss": 1.3402,
      "theoretical_loss": 3.5015183140177975,
      "tokens_seen": 1728118784
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.622051035146848e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.5014957156338293,
      "tokens_seen": 1728249856
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.621248595731023e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.5014731194435225,
      "tokens_seen": 1728380928
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.620446156315199e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.5014505254464985,
      "tokens_seen": 1728512000
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.9941221475601196,
      "objective/train/avg_token_score": 0.005877892952412367,
      "objective/train/docs_used": 631392,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2947170734405518,
      "objective/train/original_loss": 1.3030240535736084,
      "objective/train/theoretical_loss": 3.5014279336423777,
      "objective/train/tokens_used": 99102176,
      "objective/train/unlikelihood_loss": 0.01648665778338909,
      "theoretical_loss": 3.5014279336423777,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.619643716899375e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.5014279336423777,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.61884127748355e-05,
      "loss": 1.3552,
      "theoretical_loss": 3.501405344030781,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.618038838067727e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.5013827566113296,
      "tokens_seen": 1728905216
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.617236398651903e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.501360171383645,
      "tokens_seen": 1729036288
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.616433959236079e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.5013375883473477,
      "tokens_seen": 1729167360
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.615631519820255e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.5013150075020594,
      "tokens_seen": 1729298432
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.61482908040443e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.5012924288474006,
      "tokens_seen": 1729429504
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.614026640988607e-05,
      "loss": 1.295,
      "theoretical_loss": 3.501269852382994,
      "tokens_seen": 1729560576
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.613224201572782e-05,
      "loss": 1.3718,
      "theoretical_loss": 3.501247278108461,
      "tokens_seen": 1729691648
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.612421762156959e-05,
      "loss": 1.342,
      "theoretical_loss": 3.5012247060234225,
      "tokens_seen": 1729822720
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.611619322741134e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.501202136127501,
      "tokens_seen": 1729953792
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.610816883325309e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.501179568420318,
      "tokens_seen": 1730084864
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.610014443909486e-05,
      "loss": 1.4285,
      "theoretical_loss": 3.501157002901495,
      "tokens_seen": 1730215936
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.609212004493661e-05,
      "loss": 1.282,
      "theoretical_loss": 3.5011344395706554,
      "tokens_seen": 1730347008
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.608409565077838e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.5011118784274204,
      "tokens_seen": 1730478080
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.607607125662013e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.501089319471412,
      "tokens_seen": 1730609152
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.60680468624619e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.501066762702253,
      "tokens_seen": 1730740224
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.606002246830365e-05,
      "loss": 1.249,
      "theoretical_loss": 3.5010442081195654,
      "tokens_seen": 1730871296
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.60519980741454e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.501021655722972,
      "tokens_seen": 1731002368
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.604397367998717e-05,
      "loss": 1.3371,
      "theoretical_loss": 3.5009991055120957,
      "tokens_seen": 1731133440
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.603594928582892e-05,
      "loss": 1.3937,
      "theoretical_loss": 3.500976557486559,
      "tokens_seen": 1731264512
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.602792489167069e-05,
      "loss": 1.368,
      "theoretical_loss": 3.5009540116459843,
      "tokens_seen": 1731395584
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.601990049751244e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.5009314679899948,
      "tokens_seen": 1731526656
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.60118761033542e-05,
      "loss": 1.2445,
      "theoretical_loss": 3.500908926518213,
      "tokens_seen": 1731657728
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.600385170919596e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.5008863872302634,
      "tokens_seen": 1731788800
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.599582731503772e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.500863850125768,
      "tokens_seen": 1731919872
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.598780292087948e-05,
      "loss": 1.3661,
      "theoretical_loss": 3.5008413152043496,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.597977852672124e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.5008187824656325,
      "tokens_seen": 1732182016
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.5971754132563e-05,
      "loss": 1.345,
      "theoretical_loss": 3.50079625190924,
      "tokens_seen": 1732313088
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.596372973840476e-05,
      "loss": 1.3764,
      "theoretical_loss": 3.5007737235347953,
      "tokens_seen": 1732444160
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.595570534424651e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.5007511973419225,
      "tokens_seen": 1732575232
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.594768095008828e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.5007286733302445,
      "tokens_seen": 1732706304
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.593965655593003e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.5007061514993865,
      "tokens_seen": 1732837376
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.59316321617718e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.500683631848971,
      "tokens_seen": 1732968448
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.592360776761355e-05,
      "loss": 1.3698,
      "theoretical_loss": 3.500661114378623,
      "tokens_seen": 1733099520
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.59155833734553e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.5006385990879663,
      "tokens_seen": 1733230592
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.590755897929707e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.5006160859766244,
      "tokens_seen": 1733361664
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.589953458513882e-05,
      "loss": 1.363,
      "theoretical_loss": 3.5005935750442223,
      "tokens_seen": 1733492736
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.589151019098059e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.5005710662903846,
      "tokens_seen": 1733623808
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.588348579682234e-05,
      "loss": 1.3747,
      "theoretical_loss": 3.500548559714735,
      "tokens_seen": 1733754880
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.587546140266411e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.5005260553168993,
      "tokens_seen": 1733885952
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.586743700850586e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.500503553096501,
      "tokens_seen": 1734017024
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.585941261434761e-05,
      "loss": 1.3618,
      "theoretical_loss": 3.500481053053165,
      "tokens_seen": 1734148096
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.585138822018938e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.500458555186517,
      "tokens_seen": 1734279168
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.584336382603113e-05,
      "loss": 1.31,
      "theoretical_loss": 3.500436059496181,
      "tokens_seen": 1734410240
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.58353394318729e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.5004135659817823,
      "tokens_seen": 1734541312
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.582731503771465e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.5003910746429456,
      "tokens_seen": 1734672384
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.581929064355641e-05,
      "loss": 1.4172,
      "theoretical_loss": 3.500368585479297,
      "tokens_seen": 1734803456
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.581126624939817e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.5003460984904615,
      "tokens_seen": 1734934528
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.580324185523993e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.5003236136760645,
      "tokens_seen": 1735065600
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.9936129450798035,
      "objective/train/avg_token_score": 0.006387016270309687,
      "objective/train/docs_used": 633815,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3433332443237305,
      "objective/train/original_loss": 1.3528486490249634,
      "objective/train/theoretical_loss": 3.500301131035731,
      "objective/train/tokens_used": 105655776,
      "objective/train/unlikelihood_loss": 0.017082227393984795,
      "theoretical_loss": 3.500301131035731,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.57952174610817e-05,
      "loss": 1.3386,
      "theoretical_loss": 3.500301131035731,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.578719306692345e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.5002786505690873,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.57791686727652e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.5002561722757592,
      "tokens_seen": 1735458816
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.577114427860697e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.5002336961553713,
      "tokens_seen": 1735589888
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.576311988444872e-05,
      "loss": 1.3727,
      "theoretical_loss": 3.5002112222075508,
      "tokens_seen": 1735720960
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.575509549029049e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.500188750431923,
      "tokens_seen": 1735852032
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.574707109613224e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.5001662808281138,
      "tokens_seen": 1735983104
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.5739046701974e-05,
      "loss": 1.3438,
      "theoretical_loss": 3.5001438133957494,
      "tokens_seen": 1736114176
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.573102230781576e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.500121348134457,
      "tokens_seen": 1736245248
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.572299791365751e-05,
      "loss": 1.4066,
      "theoretical_loss": 3.5000988850438617,
      "tokens_seen": 1736376320
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.571497351949928e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.5000764241235904,
      "tokens_seen": 1736507392
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.570694912534103e-05,
      "loss": 1.346,
      "theoretical_loss": 3.5000539653732696,
      "tokens_seen": 1736638464
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.56989247311828e-05,
      "loss": 1.3753,
      "theoretical_loss": 3.5000315087925253,
      "tokens_seen": 1736769536
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.569090033702455e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.5000090543809854,
      "tokens_seen": 1736900608
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.568287594286632e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.4999866021382764,
      "tokens_seen": 1737031680
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.567485154870807e-05,
      "loss": 1.3921,
      "theoretical_loss": 3.499964152064025,
      "tokens_seen": 1737162752
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.566682715454983e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.499941704157857,
      "tokens_seen": 1737293824
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.565880276039159e-05,
      "loss": 1.3646,
      "theoretical_loss": 3.499919258419401,
      "tokens_seen": 1737424896
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.565077836623335e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.4998968148482845,
      "tokens_seen": 1737555968
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.564275397207511e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.499874373444133,
      "tokens_seen": 1737687040
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.563472957791687e-05,
      "loss": 1.269,
      "theoretical_loss": 3.4998519342065753,
      "tokens_seen": 1737818112
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.562670518375863e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.499829497135238,
      "tokens_seen": 1737949184
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.561868078960038e-05,
      "loss": 1.2334,
      "theoretical_loss": 3.499807062229749,
      "tokens_seen": 1738080256
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.561065639544214e-05,
      "loss": 1.4615,
      "theoretical_loss": 3.4997846294897363,
      "tokens_seen": 1738211328
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.56026320012839e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.4997621989148273,
      "tokens_seen": 1738342400
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.559460760712566e-05,
      "loss": 1.3163,
      "theoretical_loss": 3.4997397705046493,
      "tokens_seen": 1738473472
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.558658321296742e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.4997173442588307,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.557855881880918e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.4996949201769993,
      "tokens_seen": 1738735616
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.557053442465094e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.4996724982587835,
      "tokens_seen": 1738866688
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.55625100304927e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.4996500785038114,
      "tokens_seen": 1738997760
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.555448563633445e-05,
      "loss": 1.3767,
      "theoretical_loss": 3.4996276609117105,
      "tokens_seen": 1739128832
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.554646124217622e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.4996052454821105,
      "tokens_seen": 1739259904
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.553843684801797e-05,
      "loss": 1.308,
      "theoretical_loss": 3.4995828322146387,
      "tokens_seen": 1739390976
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.553041245385974e-05,
      "loss": 1.3448,
      "theoretical_loss": 3.4995604211089244,
      "tokens_seen": 1739522048
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.552238805970149e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.499538012164596,
      "tokens_seen": 1739653120
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.551436366554326e-05,
      "loss": 1.3613,
      "theoretical_loss": 3.4995156053812817,
      "tokens_seen": 1739784192
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.550633927138501e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.499493200758611,
      "tokens_seen": 1739915264
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.549831487722676e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.4994707982962128,
      "tokens_seen": 1740046336
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.549029048306853e-05,
      "loss": 1.3948,
      "theoretical_loss": 3.4994483979937154,
      "tokens_seen": 1740177408
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.548226608891028e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.499425999850749,
      "tokens_seen": 1740308480
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.547424169475205e-05,
      "loss": 1.3323,
      "theoretical_loss": 3.499403603866942,
      "tokens_seen": 1740439552
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.54662173005938e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.4993812100419235,
      "tokens_seen": 1740570624
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.545819290643557e-05,
      "loss": 1.306,
      "theoretical_loss": 3.4993588183753235,
      "tokens_seen": 1740701696
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.545016851227732e-05,
      "loss": 1.3979,
      "theoretical_loss": 3.499336428866771,
      "tokens_seen": 1740832768
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.544214411811908e-05,
      "loss": 1.349,
      "theoretical_loss": 3.4993140415158956,
      "tokens_seen": 1740963840
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.543411972396084e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.4992916563223275,
      "tokens_seen": 1741094912
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.54260953298026e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.499269273285696,
      "tokens_seen": 1741225984
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.541807093564436e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.4992468924056306,
      "tokens_seen": 1741357056
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.541004654148612e-05,
      "loss": 1.3,
      "theoretical_loss": 3.499224513681762,
      "tokens_seen": 1741488128
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.540202214732788e-05,
      "loss": 1.319,
      "theoretical_loss": 3.499202137113719,
      "tokens_seen": 1741619200
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.9916802048683167,
      "objective/train/avg_token_score": 0.008319806307554245,
      "objective/train/docs_used": 636395,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2824292182922363,
      "objective/train/original_loss": 1.2926584482192993,
      "objective/train/theoretical_loss": 3.499179762701133,
      "objective/train/tokens_used": 112209376,
      "objective/train/unlikelihood_loss": 0.02861197106540203,
      "theoretical_loss": 3.499179762701133,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.539399775316964e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.499179762701133,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.53859733590114e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.4991573904436333,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.537794896485315e-05,
      "loss": 1.3292,
      "theoretical_loss": 3.4991350203408507,
      "tokens_seen": 1742012416
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.536992457069491e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.499112652392416,
      "tokens_seen": 1742143488
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.536190017653667e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.4990902865979585,
      "tokens_seen": 1742274560
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.535387578237843e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.499067922957109,
      "tokens_seen": 1742405632
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.53458513882202e-05,
      "loss": 1.3571,
      "theoretical_loss": 3.499045561469499,
      "tokens_seen": 1742536704
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.533782699406195e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.4990232021347585,
      "tokens_seen": 1742667776
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.532980259990371e-05,
      "loss": 1.357,
      "theoretical_loss": 3.4990008449525183,
      "tokens_seen": 1742798848
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.532177820574547e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.49897848992241,
      "tokens_seen": 1742929920
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.531375381158722e-05,
      "loss": 1.3572,
      "theoretical_loss": 3.498956137044064,
      "tokens_seen": 1743060992
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.530572941742899e-05,
      "loss": 1.3757,
      "theoretical_loss": 3.4989337863171115,
      "tokens_seen": 1743192064
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.529770502327074e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.498911437741184,
      "tokens_seen": 1743323136
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.528968062911251e-05,
      "loss": 1.3941,
      "theoretical_loss": 3.4988890913159123,
      "tokens_seen": 1743454208
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.528165623495426e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.4988667470409283,
      "tokens_seen": 1743585280
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.527363184079603e-05,
      "loss": 1.3611,
      "theoretical_loss": 3.4988444049158627,
      "tokens_seen": 1743716352
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.526560744663778e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.498822064940348,
      "tokens_seen": 1743847424
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.525758305247953e-05,
      "loss": 1.364,
      "theoretical_loss": 3.498799727114015,
      "tokens_seen": 1743978496
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.52495586583213e-05,
      "loss": 1.351,
      "theoretical_loss": 3.498777391436496,
      "tokens_seen": 1744109568
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.524153426416305e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.498755057907423,
      "tokens_seen": 1744240640
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.523350987000482e-05,
      "loss": 1.3406,
      "theoretical_loss": 3.498732726526427,
      "tokens_seen": 1744371712
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.522548547584657e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.498710397293141,
      "tokens_seen": 1744502784
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.521746108168834e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.498688070207197,
      "tokens_seen": 1744633856
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.520943668753009e-05,
      "loss": 1.3824,
      "theoretical_loss": 3.498665745268226,
      "tokens_seen": 1744764928
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.520141229337185e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.498643422475862,
      "tokens_seen": 1744896000
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.519338789921361e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.4986211018297357,
      "tokens_seen": 1745027072
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.518536350505537e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.4985987833294816,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.517733911089713e-05,
      "loss": 1.3261,
      "theoretical_loss": 3.49857646697473,
      "tokens_seen": 1745289216
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.516931471673889e-05,
      "loss": 1.321,
      "theoretical_loss": 3.498554152765115,
      "tokens_seen": 1745420288
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.516129032258065e-05,
      "loss": 1.3447,
      "theoretical_loss": 3.4985318407002692,
      "tokens_seen": 1745551360
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.51532659284224e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.498509530779825,
      "tokens_seen": 1745682432
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.514524153426416e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.498487223003415,
      "tokens_seen": 1745813504
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.513721714010593e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.498464917370673,
      "tokens_seen": 1745944576
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.512919274594768e-05,
      "loss": 1.3943,
      "theoretical_loss": 3.498442613881232,
      "tokens_seen": 1746075648
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.512116835178944e-05,
      "loss": 1.3493,
      "theoretical_loss": 3.498420312534725,
      "tokens_seen": 1746206720
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.51131439576312e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.4983980133307853,
      "tokens_seen": 1746337792
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.510511956347296e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.498375716269046,
      "tokens_seen": 1746468864
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.509709516931472e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.4983534213491407,
      "tokens_seen": 1746599936
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.508907077515648e-05,
      "loss": 1.351,
      "theoretical_loss": 3.4983311285707033,
      "tokens_seen": 1746731008
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.508104638099824e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.498308837933367,
      "tokens_seen": 1746862080
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.507302198683999e-05,
      "loss": 1.3314,
      "theoretical_loss": 3.4982865494367656,
      "tokens_seen": 1746993152
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.506499759268176e-05,
      "loss": 1.3,
      "theoretical_loss": 3.498264263080533,
      "tokens_seen": 1747124224
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.505697319852351e-05,
      "loss": 1.3839,
      "theoretical_loss": 3.4982419788643035,
      "tokens_seen": 1747255296
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.504894880436528e-05,
      "loss": 1.4037,
      "theoretical_loss": 3.498219696787711,
      "tokens_seen": 1747386368
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.504092441020703e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.4981974168503887,
      "tokens_seen": 1747517440
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.50329000160488e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.4981751390519715,
      "tokens_seen": 1747648512
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.502487562189055e-05,
      "loss": 1.331,
      "theoretical_loss": 3.498152863392094,
      "tokens_seen": 1747779584
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.50168512277323e-05,
      "loss": 1.306,
      "theoretical_loss": 3.4981305898703896,
      "tokens_seen": 1747910656
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.500882683357407e-05,
      "loss": 1.3828,
      "theoretical_loss": 3.4981083184864934,
      "tokens_seen": 1748041728
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.500080243941582e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.49808604924004,
      "tokens_seen": 1748172800
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.9924666285514832,
      "objective/train/avg_token_score": 0.007533431053161621,
      "objective/train/docs_used": 638722,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.234540343284607,
      "objective/train/original_loss": 1.2434945106506348,
      "objective/train/theoretical_loss": 3.4980637821306635,
      "objective/train/tokens_used": 118762976,
      "objective/train/unlikelihood_loss": 0.028133435174822807,
      "theoretical_loss": 3.4980637821306635,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.499277804525759e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.4980637821306635,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.498475365109934e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.498041517157999,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.497672925694111e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.4980192543216813,
      "tokens_seen": 1748566016
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.496870486278286e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.4979969936213458,
      "tokens_seen": 1748697088
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.496068046862462e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.4979747350566264,
      "tokens_seen": 1748828160
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.495265607446638e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.497952478627159,
      "tokens_seen": 1748959232
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.494463168030814e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.4979302243325785,
      "tokens_seen": 1749090304
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.49366072861499e-05,
      "loss": 1.3378,
      "theoretical_loss": 3.49790797217252,
      "tokens_seen": 1749221376
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.492858289199166e-05,
      "loss": 1.3838,
      "theoretical_loss": 3.497885722146619,
      "tokens_seen": 1749352448
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.492055849783342e-05,
      "loss": 1.277,
      "theoretical_loss": 3.497863474254511,
      "tokens_seen": 1749483520
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.491253410367518e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.4978412284958313,
      "tokens_seen": 1749614592
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.490450970951693e-05,
      "loss": 1.3379,
      "theoretical_loss": 3.497818984870216,
      "tokens_seen": 1749745664
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.48964853153587e-05,
      "loss": 1.3744,
      "theoretical_loss": 3.4977967433773003,
      "tokens_seen": 1749876736
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.488846092120045e-05,
      "loss": 1.3745,
      "theoretical_loss": 3.4977745040167205,
      "tokens_seen": 1750007808
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.488043652704221e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.497752266788112,
      "tokens_seen": 1750138880
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.487241213288397e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.497730031691111,
      "tokens_seen": 1750269952
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.486438773872573e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.497707798725353,
      "tokens_seen": 1750401024
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.485636334456749e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.4976855678904752,
      "tokens_seen": 1750532096
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.484833895040924e-05,
      "loss": 1.3709,
      "theoretical_loss": 3.497663339186113,
      "tokens_seen": 1750663168
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.484031455625101e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.4976411126119036,
      "tokens_seen": 1750794240
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.483229016209276e-05,
      "loss": 1.3319,
      "theoretical_loss": 3.497618888167482,
      "tokens_seen": 1750925312
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.482426576793453e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.497596665852486,
      "tokens_seen": 1751056384
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.481624137377628e-05,
      "loss": 1.3729,
      "theoretical_loss": 3.497574445666552,
      "tokens_seen": 1751187456
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.480821697961805e-05,
      "loss": 1.3592,
      "theoretical_loss": 3.497552227609316,
      "tokens_seen": 1751318528
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.48001925854598e-05,
      "loss": 1.3512,
      "theoretical_loss": 3.4975300116804156,
      "tokens_seen": 1751449600
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.479216819130155e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.4975077978794875,
      "tokens_seen": 1751580672
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.478414379714332e-05,
      "loss": 1.3297,
      "theoretical_loss": 3.497485586206168,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.477611940298507e-05,
      "loss": 1.4202,
      "theoretical_loss": 3.4974633766600944,
      "tokens_seen": 1751842816
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.476809500882684e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.497441169240904,
      "tokens_seen": 1751973888
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.476007061466859e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.4974189639482347,
      "tokens_seen": 1752104960
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.475204622051036e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.497396760781723,
      "tokens_seen": 1752236032
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.474402182635211e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.497374559741006,
      "tokens_seen": 1752367104
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.473599743219388e-05,
      "loss": 1.3933,
      "theoretical_loss": 3.497352360825722,
      "tokens_seen": 1752498176
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.472797303803563e-05,
      "loss": 1.3915,
      "theoretical_loss": 3.4973301640355086,
      "tokens_seen": 1752629248
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.471994864387739e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.497307969370002,
      "tokens_seen": 1752760320
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.471192424971915e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.497285776828842,
      "tokens_seen": 1752891392
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.47038998555609e-05,
      "loss": 1.4236,
      "theoretical_loss": 3.4972635864116652,
      "tokens_seen": 1753022464
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.469587546140267e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.49724139811811,
      "tokens_seen": 1753153536
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.468785106724443e-05,
      "loss": 1.3518,
      "theoretical_loss": 3.4972192119478143,
      "tokens_seen": 1753284608
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.467982667308619e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.4971970279004156,
      "tokens_seen": 1753415680
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.467180227892795e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.497174845975553,
      "tokens_seen": 1753546752
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.46637778847697e-05,
      "loss": 1.3871,
      "theoretical_loss": 3.4971526661728647,
      "tokens_seen": 1753677824
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.465575349061147e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.4971304884919885,
      "tokens_seen": 1753808896
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.464772909645322e-05,
      "loss": 1.3704,
      "theoretical_loss": 3.4971083129325633,
      "tokens_seen": 1753939968
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.463970470229499e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.497086139494228,
      "tokens_seen": 1754071040
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.463168030813674e-05,
      "loss": 1.372,
      "theoretical_loss": 3.4970639681766205,
      "tokens_seen": 1754202112
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.46236559139785e-05,
      "loss": 1.36,
      "theoretical_loss": 3.49704179897938,
      "tokens_seen": 1754333184
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.461563151982026e-05,
      "loss": 1.4013,
      "theoretical_loss": 3.4970196319021447,
      "tokens_seen": 1754464256
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.460760712566201e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.496997466944554,
      "tokens_seen": 1754595328
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.459958273150378e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.496975304106247,
      "tokens_seen": 1754726400
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.9905670285224915,
      "objective/train/avg_token_score": 0.009432926774024963,
      "objective/train/docs_used": 641116,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3168390989303589,
      "objective/train/original_loss": 1.3290144205093384,
      "objective/train/theoretical_loss": 3.496953143386863,
      "objective/train/tokens_used": 125316576,
      "objective/train/unlikelihood_loss": 0.02538631111383438,
      "theoretical_loss": 3.496953143386863,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.459155833734553e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.496953143386863,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.45835339431873e-05,
      "loss": 1.321,
      "theoretical_loss": 3.49693098478604,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.457550954902905e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.496908828303419,
      "tokens_seen": 1755119616
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.456748515487082e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.496886673938638,
      "tokens_seen": 1755250688
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.455946076071257e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.4968645216913368,
      "tokens_seen": 1755381760
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.455143636655432e-05,
      "loss": 1.3457,
      "theoretical_loss": 3.496842371561155,
      "tokens_seen": 1755512832
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.454341197239609e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.4968202235477324,
      "tokens_seen": 1755643904
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.453538757823784e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4967980776507086,
      "tokens_seen": 1755774976
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.452736318407961e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.4967759338697233,
      "tokens_seen": 1755906048
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.451933878992136e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.496753792204416,
      "tokens_seen": 1756037120
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.451131439576313e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.4967316526544274,
      "tokens_seen": 1756168192
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.450329000160488e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.496709515219397,
      "tokens_seen": 1756299264
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.449526560744664e-05,
      "loss": 1.3594,
      "theoretical_loss": 3.4966873798989653,
      "tokens_seen": 1756430336
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.44872412132884e-05,
      "loss": 1.3518,
      "theoretical_loss": 3.4966652466927726,
      "tokens_seen": 1756561408
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.447921681913016e-05,
      "loss": 1.3571,
      "theoretical_loss": 3.4966431156004587,
      "tokens_seen": 1756692480
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.447119242497192e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.4966209866216644,
      "tokens_seen": 1756823552
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.446316803081368e-05,
      "loss": 1.383,
      "theoretical_loss": 3.4965988597560305,
      "tokens_seen": 1756954624
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.445514363665544e-05,
      "loss": 1.399,
      "theoretical_loss": 3.496576735003197,
      "tokens_seen": 1757085696
    },
    {
      "epoch": 0.06,
      "learning_rate": 9.44471192424972e-05,
      "loss": 1.3826,
      "theoretical_loss": 3.4965546123628046,
      "tokens_seen": 1757216768
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.443909484833896e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.496532491834494,
      "tokens_seen": 1757347840
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.443107045418072e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.496510373417907,
      "tokens_seen": 1757478912
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.442304606002247e-05,
      "loss": 1.3402,
      "theoretical_loss": 3.4964882571126834,
      "tokens_seen": 1757609984
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.441502166586424e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.496466142918465,
      "tokens_seen": 1757741056
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.440699727170599e-05,
      "loss": 1.3713,
      "theoretical_loss": 3.4964440308348923,
      "tokens_seen": 1757872128
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.439897287754776e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.4964219208616067,
      "tokens_seen": 1758003200
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.439094848338951e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.4963998129982494,
      "tokens_seen": 1758134272
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.438292408923127e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.4963777072444624,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.437489969507303e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.496355603599887,
      "tokens_seen": 1758396416
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.436687530091478e-05,
      "loss": 1.2586,
      "theoretical_loss": 3.4963335020641635,
      "tokens_seen": 1758527488
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.435885090675655e-05,
      "loss": 1.3783,
      "theoretical_loss": 3.496311402636935,
      "tokens_seen": 1758658560
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.43508265125983e-05,
      "loss": 1.4027,
      "theoretical_loss": 3.4962893053178425,
      "tokens_seen": 1758789632
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.434280211844007e-05,
      "loss": 1.423,
      "theoretical_loss": 3.496267210106528,
      "tokens_seen": 1758920704
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.433477772428182e-05,
      "loss": 1.3658,
      "theoretical_loss": 3.496245117002634,
      "tokens_seen": 1759051776
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.432675333012359e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.4962230260058016,
      "tokens_seen": 1759182848
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.431872893596534e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.496200937115673,
      "tokens_seen": 1759313920
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.43107045418071e-05,
      "loss": 1.338,
      "theoretical_loss": 3.4961788503318907,
      "tokens_seen": 1759444992
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.430268014764886e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.4961567656540966,
      "tokens_seen": 1759576064
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.429465575349061e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.4961346830819338,
      "tokens_seen": 1759707136
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.428663135933238e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.4961126026150433,
      "tokens_seen": 1759838208
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.427860696517413e-05,
      "loss": 1.4042,
      "theoretical_loss": 3.496090524253069,
      "tokens_seen": 1759969280
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.42705825710159e-05,
      "loss": 1.4254,
      "theoretical_loss": 3.496068447995653,
      "tokens_seen": 1760100352
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.426255817685765e-05,
      "loss": 1.3734,
      "theoretical_loss": 3.4960463738424385,
      "tokens_seen": 1760231424
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.42545337826994e-05,
      "loss": 1.266,
      "theoretical_loss": 3.4960243017930672,
      "tokens_seen": 1760362496
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.424650938854117e-05,
      "loss": 1.3811,
      "theoretical_loss": 3.496002231847182,
      "tokens_seen": 1760493568
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.423848499438293e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.4959801640044272,
      "tokens_seen": 1760624640
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.423046060022469e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.4959580982644445,
      "tokens_seen": 1760755712
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.422243620606645e-05,
      "loss": 1.366,
      "theoretical_loss": 3.495936034626878,
      "tokens_seen": 1760886784
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.421441181190821e-05,
      "loss": 1.3413,
      "theoretical_loss": 3.49591397309137,
      "tokens_seen": 1761017856
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.420638741774997e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.4958919136575646,
      "tokens_seen": 1761148928
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.419836302359172e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.495869856325105,
      "tokens_seen": 1761280000
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.9927342534065247,
      "objective/train/avg_token_score": 0.007265761028975248,
      "objective/train/docs_used": 643454,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2792538404464722,
      "objective/train/original_loss": 1.2895350456237793,
      "objective/train/theoretical_loss": 3.4958478010936336,
      "objective/train/tokens_used": 131870176,
      "objective/train/unlikelihood_loss": 0.017868760973215103,
      "theoretical_loss": 3.4958478010936336,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.419033862943349e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.4958478010936336,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.418231423527524e-05,
      "loss": 1.3963,
      "theoretical_loss": 3.4958257479627957,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.4174289841117e-05,
      "loss": 1.3556,
      "theoretical_loss": 3.4958036969322337,
      "tokens_seen": 1761673216
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.416626544695876e-05,
      "loss": 1.3426,
      "theoretical_loss": 3.495781648001592,
      "tokens_seen": 1761804288
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.415824105280053e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.495759601170514,
      "tokens_seen": 1761935360
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.415021665864228e-05,
      "loss": 1.447,
      "theoretical_loss": 3.4957375564386437,
      "tokens_seen": 1762066432
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.414219226448404e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.495715513805625,
      "tokens_seen": 1762197504
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.41341678703258e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.4956934732711025,
      "tokens_seen": 1762328576
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.412614347616755e-05,
      "loss": 1.4279,
      "theoretical_loss": 3.4956714348347195,
      "tokens_seen": 1762459648
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.411811908200932e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.4956493984961208,
      "tokens_seen": 1762590720
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.411009468785107e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4956273642549514,
      "tokens_seen": 1762721792
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.410207029369284e-05,
      "loss": 1.3988,
      "theoretical_loss": 3.4956053321108547,
      "tokens_seen": 1762852864
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.409404589953459e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.4955833020634755,
      "tokens_seen": 1762983936
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.408602150537636e-05,
      "loss": 1.3994,
      "theoretical_loss": 3.495561274112458,
      "tokens_seen": 1763115008
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.407799711121811e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.4955392482574474,
      "tokens_seen": 1763246080
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.406997271705986e-05,
      "loss": 1.3653,
      "theoretical_loss": 3.495517224498089,
      "tokens_seen": 1763377152
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.406194832290163e-05,
      "loss": 1.357,
      "theoretical_loss": 3.495495202834027,
      "tokens_seen": 1763508224
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.405392392874338e-05,
      "loss": 1.5184,
      "theoretical_loss": 3.495473183264906,
      "tokens_seen": 1763639296
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.404589953458515e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.4954511657903717,
      "tokens_seen": 1763770368
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.40378751404269e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.495429150410069,
      "tokens_seen": 1763901440
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.402985074626867e-05,
      "loss": 1.402,
      "theoretical_loss": 3.495407137123643,
      "tokens_seen": 1764032512
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.402182635211042e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.495385125930739,
      "tokens_seen": 1764163584
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.401380195795218e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.4953631168310024,
      "tokens_seen": 1764294656
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.400577756379394e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.4953411098240785,
      "tokens_seen": 1764425728
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.39977531696357e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.4953191049096137,
      "tokens_seen": 1764556800
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.398972877547746e-05,
      "loss": 1.325,
      "theoretical_loss": 3.4952971020872527,
      "tokens_seen": 1764687872
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.398170438131922e-05,
      "loss": 1.3804,
      "theoretical_loss": 3.4952751013566417,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.397367998716098e-05,
      "loss": 1.3978,
      "theoretical_loss": 3.495253102717426,
      "tokens_seen": 1764950016
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.396565559300274e-05,
      "loss": 1.4073,
      "theoretical_loss": 3.495231106169252,
      "tokens_seen": 1765081088
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.395763119884449e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.4952091117117656,
      "tokens_seen": 1765212160
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.394960680468626e-05,
      "loss": 1.273,
      "theoretical_loss": 3.495187119344613,
      "tokens_seen": 1765343232
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.394158241052801e-05,
      "loss": 1.4184,
      "theoretical_loss": 3.4951651290674395,
      "tokens_seen": 1765474304
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.393355801636978e-05,
      "loss": 1.3397,
      "theoretical_loss": 3.4951431408798928,
      "tokens_seen": 1765605376
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.392553362221153e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.495121154781618,
      "tokens_seen": 1765736448
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.39175092280533e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.495099170772262,
      "tokens_seen": 1765867520
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.390948483389505e-05,
      "loss": 1.4076,
      "theoretical_loss": 3.4950771888514716,
      "tokens_seen": 1765998592
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.39014604397368e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.495055209018892,
      "tokens_seen": 1766129664
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.389343604557857e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.4950332312741716,
      "tokens_seen": 1766260736
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.388541165142032e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.495011255616957,
      "tokens_seen": 1766391808
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.387738725726209e-05,
      "loss": 1.322,
      "theoretical_loss": 3.4949892820468937,
      "tokens_seen": 1766522880
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.386936286310384e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.49496731056363,
      "tokens_seen": 1766653952
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.386133846894561e-05,
      "loss": 1.4372,
      "theoretical_loss": 3.4949453411668117,
      "tokens_seen": 1766785024
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.385331407478736e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.4949233738560874,
      "tokens_seen": 1766916096
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.384528968062911e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.4949014086311028,
      "tokens_seen": 1767047168
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.383726528647088e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.494879445491506,
      "tokens_seen": 1767178240
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.382924089231263e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.4948574844369444,
      "tokens_seen": 1767309312
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.38212164981544e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.494835525467065,
      "tokens_seen": 1767440384
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.381319210399615e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.494813568581516,
      "tokens_seen": 1767571456
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.380516770983792e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.494791613779944,
      "tokens_seen": 1767702528
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.379714331567967e-05,
      "loss": 1.3967,
      "theoretical_loss": 3.494769661061998,
      "tokens_seen": 1767833600
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.9928597807884216,
      "objective/train/avg_token_score": 0.00714022945612669,
      "objective/train/docs_used": 645687,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3427472114562988,
      "objective/train/original_loss": 1.353415608406067,
      "objective/train/theoretical_loss": 3.4947477104273243,
      "objective/train/tokens_used": 138423776,
      "objective/train/unlikelihood_loss": 0.021218040958046913,
      "theoretical_loss": 3.4947477104273243,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.378911892152144e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.4947477104273243,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.37810945273632e-05,
      "loss": 1.3403,
      "theoretical_loss": 3.494725761875572,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.377307013320495e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.494703815406388,
      "tokens_seen": 1768226816
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.376504573904671e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.494681871019422,
      "tokens_seen": 1768357888
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.375702134488847e-05,
      "loss": 1.3771,
      "theoretical_loss": 3.49465992871432,
      "tokens_seen": 1768488960
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.374899695073023e-05,
      "loss": 1.3707,
      "theoretical_loss": 3.494637988490732,
      "tokens_seen": 1768620032
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.374097255657199e-05,
      "loss": 1.4103,
      "theoretical_loss": 3.4946160503483057,
      "tokens_seen": 1768751104
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.373294816241375e-05,
      "loss": 1.3161,
      "theoretical_loss": 3.4945941142866888,
      "tokens_seen": 1768882176
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.37249237682555e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.494572180305531,
      "tokens_seen": 1769013248
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.371689937409726e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.4945502484044804,
      "tokens_seen": 1769144320
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.370887497993903e-05,
      "loss": 1.3328,
      "theoretical_loss": 3.4945283185831846,
      "tokens_seen": 1769275392
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.370085058578078e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.494506390841294,
      "tokens_seen": 1769406464
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.369282619162255e-05,
      "loss": 1.3815,
      "theoretical_loss": 3.4944844651784566,
      "tokens_seen": 1769537536
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.36848017974643e-05,
      "loss": 1.3394,
      "theoretical_loss": 3.494462541594321,
      "tokens_seen": 1769668608
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.367677740330607e-05,
      "loss": 1.3247,
      "theoretical_loss": 3.494440620088536,
      "tokens_seen": 1769799680
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.366875300914782e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.494418700660752,
      "tokens_seen": 1769930752
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.366072861498957e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.4943967833106173,
      "tokens_seen": 1770061824
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.365270422083134e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.494374868037781,
      "tokens_seen": 1770192896
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.364467982667309e-05,
      "loss": 1.2429,
      "theoretical_loss": 3.4943529548418923,
      "tokens_seen": 1770323968
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.363665543251486e-05,
      "loss": 1.3298,
      "theoretical_loss": 3.4943310437226014,
      "tokens_seen": 1770455040
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.362863103835661e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.4943091346795567,
      "tokens_seen": 1770586112
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.362060664419838e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.4942872277124093,
      "tokens_seen": 1770717184
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.361258225004013e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.494265322820808,
      "tokens_seen": 1770848256
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.360455785588188e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.4942434200044015,
      "tokens_seen": 1770979328
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.359653346172365e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.494221519262841,
      "tokens_seen": 1771110400
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.35885090675654e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.4941996205957766,
      "tokens_seen": 1771241472
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.358048467340717e-05,
      "loss": 1.3227,
      "theoretical_loss": 3.4941777240028573,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.357246027924892e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.494155829483734,
      "tokens_seen": 1771503616
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.356443588509069e-05,
      "loss": 1.307,
      "theoretical_loss": 3.4941339370380566,
      "tokens_seen": 1771634688
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.355641149093244e-05,
      "loss": 1.317,
      "theoretical_loss": 3.4941120466654745,
      "tokens_seen": 1771765760
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.35483870967742e-05,
      "loss": 1.3464,
      "theoretical_loss": 3.4940901583656396,
      "tokens_seen": 1771896832
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.354036270261596e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.4940682721382017,
      "tokens_seen": 1772027904
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.353233830845772e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.494046387982811,
      "tokens_seen": 1772158976
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.352431391429948e-05,
      "loss": 1.4264,
      "theoretical_loss": 3.494024505899118,
      "tokens_seen": 1772290048
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.351628952014124e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.4940026258867745,
      "tokens_seen": 1772421120
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.3508265125983e-05,
      "loss": 1.3568,
      "theoretical_loss": 3.4939807479454297,
      "tokens_seen": 1772552192
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.350024073182476e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.4939588720747357,
      "tokens_seen": 1772683264
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.349221633766652e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.493936998274343,
      "tokens_seen": 1772814336
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.348419194350828e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.493915126543902,
      "tokens_seen": 1772945408
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.347616754935003e-05,
      "loss": 1.32,
      "theoretical_loss": 3.493893256883065,
      "tokens_seen": 1773076480
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.34681431551918e-05,
      "loss": 1.318,
      "theoretical_loss": 3.493871389291482,
      "tokens_seen": 1773207552
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.346011876103355e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.4938495237688056,
      "tokens_seen": 1773338624
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.345209436687532e-05,
      "loss": 1.3604,
      "theoretical_loss": 3.493827660314686,
      "tokens_seen": 1773469696
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.344406997271707e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.493805798928775,
      "tokens_seen": 1773600768
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.343604557855882e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.493783939610725,
      "tokens_seen": 1773731840
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.342802118440059e-05,
      "loss": 1.4683,
      "theoretical_loss": 3.493762082360186,
      "tokens_seen": 1773862912
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.341999679024234e-05,
      "loss": 1.4073,
      "theoretical_loss": 3.4937402271768105,
      "tokens_seen": 1773993984
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.341197239608411e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.4937183740602507,
      "tokens_seen": 1774125056
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.340394800192586e-05,
      "loss": 1.3397,
      "theoretical_loss": 3.4936965230101578,
      "tokens_seen": 1774256128
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.339592360776763e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.4936746740261846,
      "tokens_seen": 1774387200
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.9899575114250183,
      "objective/train/avg_token_score": 0.010042516514658928,
      "objective/train/docs_used": 647978,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.339554786682129,
      "objective/train/original_loss": 1.3519911766052246,
      "objective/train/theoretical_loss": 3.493652827107982,
      "objective/train/tokens_used": 144977376,
      "objective/train/unlikelihood_loss": 0.028962889686226845,
      "theoretical_loss": 3.493652827107982,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.338789921360938e-05,
      "loss": 1.3539,
      "theoretical_loss": 3.493652827107982,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.337987481945113e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.493630982255203,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.33718504252929e-05,
      "loss": 1.2258,
      "theoretical_loss": 3.4936091394674995,
      "tokens_seen": 1774780416
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.336382603113465e-05,
      "loss": 1.3892,
      "theoretical_loss": 3.493587298744524,
      "tokens_seen": 1774911488
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.335580163697642e-05,
      "loss": 1.3215,
      "theoretical_loss": 3.4935654600859283,
      "tokens_seen": 1775042560
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.334777724281817e-05,
      "loss": 1.3131,
      "theoretical_loss": 3.4935436234913655,
      "tokens_seen": 1775173632
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.333975284865993e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.4935217889604875,
      "tokens_seen": 1775304704
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.33317284545017e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.493499956492948,
      "tokens_seen": 1775435776
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.332370406034345e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.493478126088398,
      "tokens_seen": 1775566848
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.331567966618521e-05,
      "loss": 1.353,
      "theoretical_loss": 3.4934562977464925,
      "tokens_seen": 1775697920
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.330765527202697e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.493434471466882,
      "tokens_seen": 1775828992
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.329963087786872e-05,
      "loss": 1.3842,
      "theoretical_loss": 3.4934126472492215,
      "tokens_seen": 1775960064
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.329160648371049e-05,
      "loss": 1.3369,
      "theoretical_loss": 3.4933908250931633,
      "tokens_seen": 1776091136
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.328358208955224e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.49336900499836,
      "tokens_seen": 1776222208
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.3275557695394e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.4933471869644652,
      "tokens_seen": 1776353280
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.326753330123576e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.4933253709911325,
      "tokens_seen": 1776484352
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.325950890707753e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.493303557078015,
      "tokens_seen": 1776615424
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.325148451291928e-05,
      "loss": 1.3724,
      "theoretical_loss": 3.4932817452247664,
      "tokens_seen": 1776746496
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.324346011876103e-05,
      "loss": 1.291,
      "theoretical_loss": 3.4932599354310403,
      "tokens_seen": 1776877568
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.32354357246028e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.4932381276964897,
      "tokens_seen": 1777008640
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.322741133044455e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.4932163220207686,
      "tokens_seen": 1777139712
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.321938693628632e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.4931945184035307,
      "tokens_seen": 1777270784
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.321136254212807e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.4931727168444304,
      "tokens_seen": 1777401856
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.320333814796983e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.4931509173431214,
      "tokens_seen": 1777532928
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.319531375381159e-05,
      "loss": 1.3756,
      "theoretical_loss": 3.4931291198992582,
      "tokens_seen": 1777664000
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.318728935965335e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.493107324512494,
      "tokens_seen": 1777795072
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.317926496549511e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.4930855311824835,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.317124057133686e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.493063739908881,
      "tokens_seen": 1778057216
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.316321617717863e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.493041950691341,
      "tokens_seen": 1778188288
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.315519178302038e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.4930201635295175,
      "tokens_seen": 1778319360
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.314716738886214e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.4929983784230663,
      "tokens_seen": 1778450432
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.31391429947039e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.4929765953716405,
      "tokens_seen": 1778581504
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.313111860054566e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.4929548143748956,
      "tokens_seen": 1778712576
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.312309420638742e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.492933035432486,
      "tokens_seen": 1778843648
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.311506981222918e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.4929112585440674,
      "tokens_seen": 1778974720
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.310704541807093e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.492889483709294,
      "tokens_seen": 1779105792
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.30990210239127e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.492867710927821,
      "tokens_seen": 1779236864
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.309099662975445e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.4928459401993033,
      "tokens_seen": 1779367936
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.308297223559622e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.492824171523397,
      "tokens_seen": 1779499008
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.307494784143797e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.4928024048997566,
      "tokens_seen": 1779630080
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.306692344727974e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.4927806403280384,
      "tokens_seen": 1779761152
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.305889905312149e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.492758877807897,
      "tokens_seen": 1779892224
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.305087465896324e-05,
      "loss": 1.3816,
      "theoretical_loss": 3.4927371173389874,
      "tokens_seen": 1780023296
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.304285026480501e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.4927153589209667,
      "tokens_seen": 1780154368
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.303482587064676e-05,
      "loss": 1.315,
      "theoretical_loss": 3.49269360255349,
      "tokens_seen": 1780285440
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.302680147648853e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.492671848236213,
      "tokens_seen": 1780416512
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.301877708233028e-05,
      "loss": 1.244,
      "theoretical_loss": 3.4926500959687914,
      "tokens_seen": 1780547584
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.301075268817204e-05,
      "loss": 1.3649,
      "theoretical_loss": 3.4926283457508815,
      "tokens_seen": 1780678656
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.30027282940138e-05,
      "loss": 1.3833,
      "theoretical_loss": 3.4926065975821396,
      "tokens_seen": 1780809728
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.299470389985556e-05,
      "loss": 1.2633,
      "theoretical_loss": 3.4925848514622215,
      "tokens_seen": 1780940800
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.9912664294242859,
      "objective/train/avg_token_score": 0.008733604103326797,
      "objective/train/docs_used": 650314,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.351446509361267,
      "objective/train/original_loss": 1.3623491525650024,
      "objective/train/theoretical_loss": 3.492563107390783,
      "objective/train/tokens_used": 151530976,
      "objective/train/unlikelihood_loss": 0.032497338950634,
      "theoretical_loss": 3.492563107390783,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.298667950569732e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.492563107390783,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.297865511153908e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.4925413653674813,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.297063071738084e-05,
      "loss": 1.3574,
      "theoretical_loss": 3.4925196253919717,
      "tokens_seen": 1781334016
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.29626063232226e-05,
      "loss": 1.4011,
      "theoretical_loss": 3.4924978874639114,
      "tokens_seen": 1781465088
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.295458192906435e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.4924761515829577,
      "tokens_seen": 1781596160
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.294655753490612e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.4924544177487657,
      "tokens_seen": 1781727232
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.293853314074787e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.492432685960993,
      "tokens_seen": 1781858304
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.293050874658963e-05,
      "loss": 1.418,
      "theoretical_loss": 3.492410956219296,
      "tokens_seen": 1781989376
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.292248435243139e-05,
      "loss": 1.355,
      "theoretical_loss": 3.492389228523332,
      "tokens_seen": 1782120448
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.291445995827315e-05,
      "loss": 1.3084,
      "theoretical_loss": 3.4923675028727574,
      "tokens_seen": 1782251520
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.290643556411491e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.49234577926723,
      "tokens_seen": 1782382592
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.289841116995666e-05,
      "loss": 1.3732,
      "theoretical_loss": 3.4923240577064067,
      "tokens_seen": 1782513664
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.289038677579843e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.492302338189944,
      "tokens_seen": 1782644736
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.288236238164018e-05,
      "loss": 1.3119,
      "theoretical_loss": 3.4922806207174997,
      "tokens_seen": 1782775808
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.287433798748195e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.4922589052887316,
      "tokens_seen": 1782906880
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.28663135933237e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.492237191903297,
      "tokens_seen": 1783037952
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.285828919916547e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.492215480560853,
      "tokens_seen": 1783169024
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.285026480500722e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.4921937712610576,
      "tokens_seen": 1783300096
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.284224041084899e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.492172064003568,
      "tokens_seen": 1783431168
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.283421601669074e-05,
      "loss": 1.306,
      "theoretical_loss": 3.4921503587880425,
      "tokens_seen": 1783562240
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.28261916225325e-05,
      "loss": 1.309,
      "theoretical_loss": 3.492128655614139,
      "tokens_seen": 1783693312
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.281816722837426e-05,
      "loss": 1.3954,
      "theoretical_loss": 3.4921069544815153,
      "tokens_seen": 1783824384
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.281014283421601e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.4920852553898296,
      "tokens_seen": 1783955456
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.280211844005778e-05,
      "loss": 1.4386,
      "theoretical_loss": 3.4920635583387396,
      "tokens_seen": 1784086528
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.279409404589953e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.492041863327904,
      "tokens_seen": 1784217600
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.27860696517413e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.49202017035698,
      "tokens_seen": 1784348672
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.277804525758305e-05,
      "loss": 1.3741,
      "theoretical_loss": 3.491998479425628,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.27700208634248e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.4919767905335046,
      "tokens_seen": 1784610816
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.276199646926657e-05,
      "loss": 1.3707,
      "theoretical_loss": 3.491955103680269,
      "tokens_seen": 1784741888
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.275397207510833e-05,
      "loss": 1.4396,
      "theoretical_loss": 3.49193341886558,
      "tokens_seen": 1784872960
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.274594768095009e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.4919117360890963,
      "tokens_seen": 1785004032
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.273792328679185e-05,
      "loss": 1.4431,
      "theoretical_loss": 3.491890055350477,
      "tokens_seen": 1785135104
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.272989889263361e-05,
      "loss": 1.3871,
      "theoretical_loss": 3.4918683766493794,
      "tokens_seen": 1785266176
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.272187449847537e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.491846699985464,
      "tokens_seen": 1785397248
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.271385010431712e-05,
      "loss": 1.3886,
      "theoretical_loss": 3.491825025358389,
      "tokens_seen": 1785528320
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.270582571015889e-05,
      "loss": 1.3759,
      "theoretical_loss": 3.4918033527678145,
      "tokens_seen": 1785659392
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.269780131600064e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.4917816822133982,
      "tokens_seen": 1785790464
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.26897769218424e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.491760013694801,
      "tokens_seen": 1785921536
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.268175252768416e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.4917383472116814,
      "tokens_seen": 1786052608
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.267372813352592e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.4917166827636987,
      "tokens_seen": 1786183680
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.266570373936768e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.4916950203505124,
      "tokens_seen": 1786314752
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.265767934520943e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.4916733599717826,
      "tokens_seen": 1786445824
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.26496549510512e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.491651701627169,
      "tokens_seen": 1786576896
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.264163055689295e-05,
      "loss": 1.4082,
      "theoretical_loss": 3.491630045316331,
      "tokens_seen": 1786707968
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.263360616273472e-05,
      "loss": 1.4336,
      "theoretical_loss": 3.491608391038928,
      "tokens_seen": 1786839040
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.262558176857647e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.491586738794621,
      "tokens_seen": 1786970112
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.261755737441824e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.4915650885830694,
      "tokens_seen": 1787101184
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.260953298025999e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.4915434404039334,
      "tokens_seen": 1787232256
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.260150858610174e-05,
      "loss": 1.3367,
      "theoretical_loss": 3.491521794256873,
      "tokens_seen": 1787363328
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.259348419194351e-05,
      "loss": 1.303,
      "theoretical_loss": 3.4915001501415484,
      "tokens_seen": 1787494400
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.987833559513092,
      "objective/train/avg_token_score": 0.012166428379714489,
      "objective/train/docs_used": 652772,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3535284996032715,
      "objective/train/original_loss": 1.3691428899765015,
      "objective/train/theoretical_loss": 3.4914785080576207,
      "objective/train/tokens_used": 158084576,
      "objective/train/unlikelihood_loss": 0.03675776720046997,
      "theoretical_loss": 3.4914785080576207,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.258545979778526e-05,
      "loss": 1.3936,
      "theoretical_loss": 3.4914785080576207,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.257743540362703e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.4914568680047493,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.256941100946878e-05,
      "loss": 1.3575,
      "theoretical_loss": 3.491435229982595,
      "tokens_seen": 1787887616
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.256138661531055e-05,
      "loss": 1.3935,
      "theoretical_loss": 3.491413593990819,
      "tokens_seen": 1788018688
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.25533622211523e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.4913919600290813,
      "tokens_seen": 1788149760
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.254533782699407e-05,
      "loss": 1.28,
      "theoretical_loss": 3.4913703280970427,
      "tokens_seen": 1788280832
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.253731343283582e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.4913486981943644,
      "tokens_seen": 1788411904
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.252928903867758e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.491327070320707,
      "tokens_seen": 1788542976
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.252126464451934e-05,
      "loss": 1.327,
      "theoretical_loss": 3.4913054444757314,
      "tokens_seen": 1788674048
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.25132402503611e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.491283820659099,
      "tokens_seen": 1788805120
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.250521585620286e-05,
      "loss": 1.3812,
      "theoretical_loss": 3.491262198870471,
      "tokens_seen": 1788936192
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.249719146204462e-05,
      "loss": 1.298,
      "theoretical_loss": 3.4912405791095082,
      "tokens_seen": 1789067264
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.248916706788638e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.4912189613758726,
      "tokens_seen": 1789198336
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.248114267372814e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.491197345669225,
      "tokens_seen": 1789329408
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.247311827956989e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.4911757319892267,
      "tokens_seen": 1789460480
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.246509388541166e-05,
      "loss": 1.289,
      "theoretical_loss": 3.4911541203355396,
      "tokens_seen": 1789591552
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.245706949125341e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.4911325107078257,
      "tokens_seen": 1789722624
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.244904509709518e-05,
      "loss": 1.3455,
      "theoretical_loss": 3.4911109031057457,
      "tokens_seen": 1789853696
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.244102070293693e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.491089297528963,
      "tokens_seen": 1789984768
    },
    {
      "epoch": 0.08,
      "learning_rate": 9.24329963087787e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.491067693977138,
      "tokens_seen": 1790115840
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.242497191462045e-05,
      "loss": 1.3774,
      "theoretical_loss": 3.491046092449933,
      "tokens_seen": 1790246912
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.24169475204622e-05,
      "loss": 1.4128,
      "theoretical_loss": 3.4910244929470107,
      "tokens_seen": 1790377984
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.240892312630397e-05,
      "loss": 1.2355,
      "theoretical_loss": 3.4910028954680326,
      "tokens_seen": 1790509056
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.240089873214572e-05,
      "loss": 1.3772,
      "theoretical_loss": 3.490981300012661,
      "tokens_seen": 1790640128
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.239287433798749e-05,
      "loss": 1.262,
      "theoretical_loss": 3.4909597065805587,
      "tokens_seen": 1790771200
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.238484994382924e-05,
      "loss": 1.3822,
      "theoretical_loss": 3.490938115171387,
      "tokens_seen": 1790902272
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.237682554967101e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.49091652578481,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.236880115551276e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.4908949384204884,
      "tokens_seen": 1791164416
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.236077676135451e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.4908733530780856,
      "tokens_seen": 1791295488
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.235275236719628e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.490851769757265,
      "tokens_seen": 1791426560
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.234472797303803e-05,
      "loss": 1.2144,
      "theoretical_loss": 3.4908301884576884,
      "tokens_seen": 1791557632
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.23367035788798e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.490808609179019,
      "tokens_seen": 1791688704
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.232867918472155e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.4907870319209198,
      "tokens_seen": 1791819776
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.232065479056332e-05,
      "loss": 1.293,
      "theoretical_loss": 3.4907654566830537,
      "tokens_seen": 1791950848
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.231263039640507e-05,
      "loss": 1.3692,
      "theoretical_loss": 3.4907438834650844,
      "tokens_seen": 1792081920
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.230460600224683e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.4907223122666737,
      "tokens_seen": 1792212992
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.229658160808859e-05,
      "loss": 1.38,
      "theoretical_loss": 3.490700743087486,
      "tokens_seen": 1792344064
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.228855721393035e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.4906791759271845,
      "tokens_seen": 1792475136
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.228053281977211e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.4906576107854326,
      "tokens_seen": 1792606208
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.227250842561387e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.4906360476618934,
      "tokens_seen": 1792737280
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.226448403145563e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.4906144865562307,
      "tokens_seen": 1792868352
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.225645963729739e-05,
      "loss": 1.289,
      "theoretical_loss": 3.490592927468108,
      "tokens_seen": 1792999424
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.224843524313914e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.49057137039719,
      "tokens_seen": 1793130496
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.22404108489809e-05,
      "loss": 1.2275,
      "theoretical_loss": 3.4905498153431394,
      "tokens_seen": 1793261568
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.223238645482266e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.49052826230562,
      "tokens_seen": 1793392640
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.222436206066443e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.490506711284297,
      "tokens_seen": 1793523712
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.221633766650618e-05,
      "loss": 1.3817,
      "theoretical_loss": 3.490485162278834,
      "tokens_seen": 1793654784
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.220831327234795e-05,
      "loss": 1.3369,
      "theoretical_loss": 3.4904636152888946,
      "tokens_seen": 1793785856
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.22002888781897e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.490442070314143,
      "tokens_seen": 1793916928
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.219226448403146e-05,
      "loss": 1.246,
      "theoretical_loss": 3.490420527354244,
      "tokens_seen": 1794048000
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.9925491809844971,
      "objective/train/avg_token_score": 0.007450864650309086,
      "objective/train/docs_used": 655001,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.308032512664795,
      "objective/train/original_loss": 1.31708824634552,
      "objective/train/theoretical_loss": 3.4903989864088625,
      "objective/train/tokens_used": 164638176,
      "objective/train/unlikelihood_loss": 0.025070594623684883,
      "theoretical_loss": 3.4903989864088625,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.218424008987322e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.4903989864088625,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.217621569571497e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.4903774474776617,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.216819130155674e-05,
      "loss": 1.2296,
      "theoretical_loss": 3.4903559105603073,
      "tokens_seen": 1794441216
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.216016690739849e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.4903343756564627,
      "tokens_seen": 1794572288
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.215214251324026e-05,
      "loss": 1.3771,
      "theoretical_loss": 3.4903128427657935,
      "tokens_seen": 1794703360
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.214411811908201e-05,
      "loss": 1.303,
      "theoretical_loss": 3.4902913118879644,
      "tokens_seen": 1794834432
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.213609372492378e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.4902697830226406,
      "tokens_seen": 1794965504
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.212806933076553e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.490248256169486,
      "tokens_seen": 1795096576
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.212004493660728e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.4902267313281667,
      "tokens_seen": 1795227648
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.211202054244905e-05,
      "loss": 1.2272,
      "theoretical_loss": 3.490205208498347,
      "tokens_seen": 1795358720
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.21039961482908e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.490183687679693,
      "tokens_seen": 1795489792
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.209597175413257e-05,
      "loss": 1.1874,
      "theoretical_loss": 3.490162168871869,
      "tokens_seen": 1795620864
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.208794735997432e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.490140652074541,
      "tokens_seen": 1795751936
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.207992296581609e-05,
      "loss": 1.3962,
      "theoretical_loss": 3.4901191372873743,
      "tokens_seen": 1795883008
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.207189857165784e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.4900976245100344,
      "tokens_seen": 1796014080
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.20638741774996e-05,
      "loss": 1.34,
      "theoretical_loss": 3.4900761137421865,
      "tokens_seen": 1796145152
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.205584978334136e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.490054604983497,
      "tokens_seen": 1796276224
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.204782538918312e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.490033098233631,
      "tokens_seen": 1796407296
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.203980099502488e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.4900115934922544,
      "tokens_seen": 1796538368
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.203177660086664e-05,
      "loss": 1.3744,
      "theoretical_loss": 3.4899900907590338,
      "tokens_seen": 1796669440
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.20237522067084e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.4899685900336346,
      "tokens_seen": 1796800512
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.201572781255016e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.489947091315723,
      "tokens_seen": 1796931584
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.200770341839191e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.489925594604965,
      "tokens_seen": 1797062656
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.199967902423368e-05,
      "loss": 1.4098,
      "theoretical_loss": 3.489904099901027,
      "tokens_seen": 1797193728
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.199165463007543e-05,
      "loss": 1.3297,
      "theoretical_loss": 3.489882607203576,
      "tokens_seen": 1797324800
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.19836302359172e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.4898611165122766,
      "tokens_seen": 1797455872
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.197560584175895e-05,
      "loss": 1.3563,
      "theoretical_loss": 3.489839627826797,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.196758144760072e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.489818141146803,
      "tokens_seen": 1797718016
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.195955705344247e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.4897966564719614,
      "tokens_seen": 1797849088
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.195153265928422e-05,
      "loss": 1.3643,
      "theoretical_loss": 3.489775173801939,
      "tokens_seen": 1797980160
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.194350826512599e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.489753693136402,
      "tokens_seen": 1798111232
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.193548387096774e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.4897322144750182,
      "tokens_seen": 1798242304
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.192745947680951e-05,
      "loss": 1.3792,
      "theoretical_loss": 3.489710737817454,
      "tokens_seen": 1798373376
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.191943508265126e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.4896892631633762,
      "tokens_seen": 1798504448
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.191141068849303e-05,
      "loss": 1.3724,
      "theoretical_loss": 3.4896677905124527,
      "tokens_seen": 1798635520
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.190338629433478e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.4896463198643497,
      "tokens_seen": 1798766592
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.189536190017655e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.4896248512187356,
      "tokens_seen": 1798897664
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.18873375060183e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.4896033845752763,
      "tokens_seen": 1799028736
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.187931311186005e-05,
      "loss": 1.258,
      "theoretical_loss": 3.4895819199336406,
      "tokens_seen": 1799159808
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.187128871770182e-05,
      "loss": 1.2393,
      "theoretical_loss": 3.4895604572934955,
      "tokens_seen": 1799290880
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.186326432354357e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.4895389966545087,
      "tokens_seen": 1799421952
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.185523992938534e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.4895175380163472,
      "tokens_seen": 1799553024
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.18472155352271e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.4894960813786797,
      "tokens_seen": 1799684096
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.183919114106886e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.489474626741174,
      "tokens_seen": 1799815168
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.183116674691061e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.4894531741034966,
      "tokens_seen": 1799946240
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.182314235275237e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.489431723465317,
      "tokens_seen": 1800077312
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.181511795859413e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.489410274826303,
      "tokens_seen": 1800208384
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.180709356443589e-05,
      "loss": 1.3782,
      "theoretical_loss": 3.4893888281861223,
      "tokens_seen": 1800339456
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.179906917027765e-05,
      "loss": 1.3735,
      "theoretical_loss": 3.489367383544443,
      "tokens_seen": 1800470528
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.17910447761194e-05,
      "loss": 1.319,
      "theoretical_loss": 3.489345940900934,
      "tokens_seen": 1800601600
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.9920512437820435,
      "objective/train/avg_token_score": 0.007948792539536953,
      "objective/train/docs_used": 657521,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.380130648612976,
      "objective/train/original_loss": 1.3918962478637695,
      "objective/train/theoretical_loss": 3.4893245002552638,
      "objective/train/tokens_used": 171191776,
      "objective/train/unlikelihood_loss": 0.019094528630375862,
      "theoretical_loss": 3.4893245002552638,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.178302038196117e-05,
      "loss": 1.3878,
      "theoretical_loss": 3.4893245002552638,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.177499598780293e-05,
      "loss": 1.285,
      "theoretical_loss": 3.4893030616070995,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.176697159364468e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.4892816249561114,
      "tokens_seen": 1800994816
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.175894719948645e-05,
      "loss": 1.2931,
      "theoretical_loss": 3.489260190301967,
      "tokens_seen": 1801125888
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.17509228053282e-05,
      "loss": 1.3287,
      "theoretical_loss": 3.4892387576443356,
      "tokens_seen": 1801256960
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.174289841116997e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.4892173269828852,
      "tokens_seen": 1801388032
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.173487401701172e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.489195898317286,
      "tokens_seen": 1801519104
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.172684962285349e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.489174471647205,
      "tokens_seen": 1801650176
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.171882522869524e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.489153046972313,
      "tokens_seen": 1801781248
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.171080083453699e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.4891316242922787,
      "tokens_seen": 1801912320
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.170277644037876e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.4891102036067703,
      "tokens_seen": 1802043392
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.169475204622051e-05,
      "loss": 1.286,
      "theoretical_loss": 3.4890887849154577,
      "tokens_seen": 1802174464
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.168672765206228e-05,
      "loss": 1.2244,
      "theoretical_loss": 3.4890673682180107,
      "tokens_seen": 1802305536
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.167870325790403e-05,
      "loss": 1.2068,
      "theoretical_loss": 3.489045953514098,
      "tokens_seen": 1802436608
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.16706788637458e-05,
      "loss": 1.3866,
      "theoretical_loss": 3.4890245408033893,
      "tokens_seen": 1802567680
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.166265446958755e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.489003130085554,
      "tokens_seen": 1802698752
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.16546300754293e-05,
      "loss": 1.3382,
      "theoretical_loss": 3.4889817213602625,
      "tokens_seen": 1802829824
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.164660568127107e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.4889603146271835,
      "tokens_seen": 1802960896
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.163858128711282e-05,
      "loss": 1.3646,
      "theoretical_loss": 3.488938909885988,
      "tokens_seen": 1803091968
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.163055689295459e-05,
      "loss": 1.322,
      "theoretical_loss": 3.4889175071363443,
      "tokens_seen": 1803223040
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.162253249879634e-05,
      "loss": 1.3471,
      "theoretical_loss": 3.488896106377924,
      "tokens_seen": 1803354112
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.161450810463811e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.4888747076103956,
      "tokens_seen": 1803485184
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.160648371047986e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.4888533108334303,
      "tokens_seen": 1803616256
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.159845931632163e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.4888319160466983,
      "tokens_seen": 1803747328
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.159043492216338e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.488810523249869,
      "tokens_seen": 1803878400
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.158241052800514e-05,
      "loss": 1.3727,
      "theoretical_loss": 3.4887891324426135,
      "tokens_seen": 1804009472
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.15743861338469e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.488767743624602,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.156636173968866e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.488746356795505,
      "tokens_seen": 1804271616
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.155833734553042e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.4887249719549933,
      "tokens_seen": 1804402688
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.155031295137218e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.4887035891027374,
      "tokens_seen": 1804533760
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.154228855721394e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.4886822082384077,
      "tokens_seen": 1804664832
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.15342641630557e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.4886608293616757,
      "tokens_seen": 1804795904
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.152623976889745e-05,
      "loss": 1.2056,
      "theoretical_loss": 3.4886394524722117,
      "tokens_seen": 1804926976
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.151821537473922e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.4886180775696873,
      "tokens_seen": 1805058048
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.151019098058097e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.488596704653773,
      "tokens_seen": 1805189120
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.150216658642274e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.4885753337241403,
      "tokens_seen": 1805320192
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.149414219226449e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.48855396478046,
      "tokens_seen": 1805451264
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.148611779810626e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.4885325978224038,
      "tokens_seen": 1805582336
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.147809340394801e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.4885112328496426,
      "tokens_seen": 1805713408
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.147006900978976e-05,
      "loss": 1.3338,
      "theoretical_loss": 3.4884898698618487,
      "tokens_seen": 1805844480
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.146204461563153e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.488468508858693,
      "tokens_seen": 1805975552
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.145402022147328e-05,
      "loss": 1.276,
      "theoretical_loss": 3.488447149839847,
      "tokens_seen": 1806106624
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.144599582731505e-05,
      "loss": 1.329,
      "theoretical_loss": 3.488425792804983,
      "tokens_seen": 1806237696
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.14379714331568e-05,
      "loss": 1.2322,
      "theoretical_loss": 3.488404437753772,
      "tokens_seen": 1806368768
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.142994703899857e-05,
      "loss": 1.3979,
      "theoretical_loss": 3.4883830846858856,
      "tokens_seen": 1806499840
    },
    {
      "epoch": 0.09,
      "learning_rate": 9.142192264484032e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.488361733600997,
      "tokens_seen": 1806630912
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.141389825068207e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.4883403844987777,
      "tokens_seen": 1806761984
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.140587385652384e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.4883190373788997,
      "tokens_seen": 1806893056
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.13978494623656e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.4882976922410345,
      "tokens_seen": 1807024128
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.138982506820736e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.488276349084855,
      "tokens_seen": 1807155200
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.9913113117218018,
      "objective/train/avg_token_score": 0.008688726462423801,
      "objective/train/docs_used": 659683,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2719718217849731,
      "objective/train/original_loss": 1.283247947692871,
      "objective/train/theoretical_loss": 3.4882550079100336,
      "objective/train/tokens_used": 177745376,
      "objective/train/unlikelihood_loss": 0.027942229062318802,
      "theoretical_loss": 3.4882550079100336,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.138180067404911e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.4882550079100336,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.137377627989088e-05,
      "loss": 1.3536,
      "theoretical_loss": 3.4882336687162425,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.136575188573263e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.4882123315031537,
      "tokens_seen": 1807548416
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.135772749157439e-05,
      "loss": 1.3482,
      "theoretical_loss": 3.4881909962704407,
      "tokens_seen": 1807679488
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.134970309741615e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.488169663017775,
      "tokens_seen": 1807810560
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.13416787032579e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.4881483317448305,
      "tokens_seen": 1807941632
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.133365430909967e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.488127002451279,
      "tokens_seen": 1808072704
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.132562991494143e-05,
      "loss": 1.3684,
      "theoretical_loss": 3.488105675136794,
      "tokens_seen": 1808203776
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.131760552078319e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.4880843498010483,
      "tokens_seen": 1808334848
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.130958112662495e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.4880630264437142,
      "tokens_seen": 1808465920
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.13015567324667e-05,
      "loss": 1.3733,
      "theoretical_loss": 3.488041705064466,
      "tokens_seen": 1808596992
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.129353233830847e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.488020385662976,
      "tokens_seen": 1808728064
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.128550794415022e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.4879990682389175,
      "tokens_seen": 1808859136
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.127748354999199e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.4879777527919638,
      "tokens_seen": 1808990208
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.126945915583374e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.4879564393217892,
      "tokens_seen": 1809121280
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.12614347616755e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.4879351278280657,
      "tokens_seen": 1809252352
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.125341036751726e-05,
      "loss": 1.3664,
      "theoretical_loss": 3.487913818310468,
      "tokens_seen": 1809383424
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.124538597335903e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.4878925107686696,
      "tokens_seen": 1809514496
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.123736157920078e-05,
      "loss": 1.3664,
      "theoretical_loss": 3.4878712052023433,
      "tokens_seen": 1809645568
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.122933718504253e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.487849901611164,
      "tokens_seen": 1809776640
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.12213127908843e-05,
      "loss": 1.285,
      "theoretical_loss": 3.4878285999948053,
      "tokens_seen": 1809907712
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.121328839672605e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.487807300352941,
      "tokens_seen": 1810038784
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.120526400256782e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.4877860026852443,
      "tokens_seen": 1810169856
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.119723960840957e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.487764706991391,
      "tokens_seen": 1810300928
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.118921521425134e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.487743413271054,
      "tokens_seen": 1810432000
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.118119082009309e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.487722121523907,
      "tokens_seen": 1810563072
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.117316642593484e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.4877008317496268,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.116514203177661e-05,
      "loss": 1.353,
      "theoretical_loss": 3.487679543947885,
      "tokens_seen": 1810825216
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.115711763761836e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.4876582581183575,
      "tokens_seen": 1810956288
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.114909324346013e-05,
      "loss": 1.3494,
      "theoretical_loss": 3.4876369742607194,
      "tokens_seen": 1811087360
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.114106884930188e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.4876156923746437,
      "tokens_seen": 1811218432
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.113304445514365e-05,
      "loss": 1.3427,
      "theoretical_loss": 3.4875944124598064,
      "tokens_seen": 1811349504
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.11250200609854e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.4875731345158822,
      "tokens_seen": 1811480576
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.111699566682716e-05,
      "loss": 1.336,
      "theoretical_loss": 3.4875518585425453,
      "tokens_seen": 1811611648
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.110897127266892e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.4875305845394715,
      "tokens_seen": 1811742720
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.110094687851068e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.487509312506335,
      "tokens_seen": 1811873792
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.109292248435244e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.487488042442811,
      "tokens_seen": 1812004864
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.10848980901942e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.487466774348575,
      "tokens_seen": 1812135936
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.107687369603596e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.4874455082233027,
      "tokens_seen": 1812267008
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.106884930187772e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.4874242440666685,
      "tokens_seen": 1812398080
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.106082490771947e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.4874029818783487,
      "tokens_seen": 1812529152
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.105280051356124e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.487381721658018,
      "tokens_seen": 1812660224
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.104477611940299e-05,
      "loss": 1.3293,
      "theoretical_loss": 3.487360463405352,
      "tokens_seen": 1812791296
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.103675172524476e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.4873392071200264,
      "tokens_seen": 1812922368
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.102872733108651e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.4873179528017175,
      "tokens_seen": 1813053440
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.102070293692828e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.487296700450101,
      "tokens_seen": 1813184512
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.101267854277003e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.4872754500648524,
      "tokens_seen": 1813315584
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.100465414861178e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.487254201645648,
      "tokens_seen": 1813446656
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.099662975445355e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.487232955192163,
      "tokens_seen": 1813577728
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.09886053602953e-05,
      "loss": 1.3396,
      "theoretical_loss": 3.487211710704074,
      "tokens_seen": 1813708800
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.9918608069419861,
      "objective/train/avg_token_score": 0.008139228448271751,
      "objective/train/docs_used": 661946,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.224936604499817,
      "objective/train/original_loss": 1.235142707824707,
      "objective/train/theoretical_loss": 3.4871904681810575,
      "objective/train/tokens_used": 184298976,
      "objective/train/unlikelihood_loss": 0.023813558742403984,
      "theoretical_loss": 3.4871904681810575,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.098058096613707e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.4871904681810575,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.097255657197882e-05,
      "loss": 1.407,
      "theoretical_loss": 3.4871692276227897,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.096453217782059e-05,
      "loss": 1.3476,
      "theoretical_loss": 3.487147989028946,
      "tokens_seen": 1814102016
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.095650778366234e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.487126752399204,
      "tokens_seen": 1814233088
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.094848338950411e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.48710551773324,
      "tokens_seen": 1814364160
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.094045899534586e-05,
      "loss": 1.3788,
      "theoretical_loss": 3.48708428503073,
      "tokens_seen": 1814495232
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.093243460118761e-05,
      "loss": 1.3964,
      "theoretical_loss": 3.487063054291351,
      "tokens_seen": 1814626304
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.092441020702938e-05,
      "loss": 1.294,
      "theoretical_loss": 3.4870418255147793,
      "tokens_seen": 1814757376
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.091638581287113e-05,
      "loss": 1.395,
      "theoretical_loss": 3.4870205987006924,
      "tokens_seen": 1814888448
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.09083614187129e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.486999373848767,
      "tokens_seen": 1815019520
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.090033702455465e-05,
      "loss": 1.3553,
      "theoretical_loss": 3.486978150958679,
      "tokens_seen": 1815150592
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.089231263039642e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.486956930030107,
      "tokens_seen": 1815281664
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.088428823623817e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.486935711062727,
      "tokens_seen": 1815412736
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.087626384207993e-05,
      "loss": 1.3563,
      "theoretical_loss": 3.486914494056217,
      "tokens_seen": 1815543808
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.08682394479217e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.4868932790102534,
      "tokens_seen": 1815674880
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.086021505376345e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.486872065924514,
      "tokens_seen": 1815805952
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.085219065960521e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.4868508547986754,
      "tokens_seen": 1815937024
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.084416626544697e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.4868296456324166,
      "tokens_seen": 1816068096
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.083614187128873e-05,
      "loss": 1.2512,
      "theoretical_loss": 3.486808438425414,
      "tokens_seen": 1816199168
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.082811747713049e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.4867872331773455,
      "tokens_seen": 1816330240
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.082009308297224e-05,
      "loss": 1.3708,
      "theoretical_loss": 3.486766029887889,
      "tokens_seen": 1816461312
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.0812068688814e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.4867448285567226,
      "tokens_seen": 1816592384
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.080404429465576e-05,
      "loss": 1.316,
      "theoretical_loss": 3.486723629183523,
      "tokens_seen": 1816723456
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.079601990049753e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.486702431767969,
      "tokens_seen": 1816854528
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.078799550633928e-05,
      "loss": 1.279,
      "theoretical_loss": 3.486681236309739,
      "tokens_seen": 1816985600
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.077997111218105e-05,
      "loss": 1.3949,
      "theoretical_loss": 3.48666004280851,
      "tokens_seen": 1817116672
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.07719467180228e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.486638851263961,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.076392232386455e-05,
      "loss": 1.3367,
      "theoretical_loss": 3.48661766167577,
      "tokens_seen": 1817378816
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.075589792970632e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.486596474043615,
      "tokens_seen": 1817509888
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.074787353554807e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.486575288367175,
      "tokens_seen": 1817640960
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.073984914138984e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.486554104646128,
      "tokens_seen": 1817772032
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.073182474723159e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.4865329228801523,
      "tokens_seen": 1817903104
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.072380035307334e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.4865117430689274,
      "tokens_seen": 1818034176
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.071577595891511e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.486490565212131,
      "tokens_seen": 1818165248
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.070775156475686e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.486469389309443,
      "tokens_seen": 1818296320
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.069972717059863e-05,
      "loss": 1.3527,
      "theoretical_loss": 3.4864482153605416,
      "tokens_seen": 1818427392
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.069170277644038e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.4864270433651052,
      "tokens_seen": 1818558464
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.068367838228215e-05,
      "loss": 1.3722,
      "theoretical_loss": 3.486405873322814,
      "tokens_seen": 1818689536
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.06756539881239e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.486384705233346,
      "tokens_seen": 1818820608
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.066762959396566e-05,
      "loss": 1.3501,
      "theoretical_loss": 3.486363539096381,
      "tokens_seen": 1818951680
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.065960519980742e-05,
      "loss": 1.3817,
      "theoretical_loss": 3.4863423749115983,
      "tokens_seen": 1819082752
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.065158080564918e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.4863212126786762,
      "tokens_seen": 1819213824
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.064355641149094e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.4863000523972953,
      "tokens_seen": 1819344896
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.06355320173327e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.4862788940671345,
      "tokens_seen": 1819475968
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.062750762317445e-05,
      "loss": 1.4016,
      "theoretical_loss": 3.486257737687873,
      "tokens_seen": 1819607040
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.061948322901622e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.486236583259191,
      "tokens_seen": 1819738112
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.061145883485797e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.486215430780769,
      "tokens_seen": 1819869184
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.060343444069974e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.4861942802522843,
      "tokens_seen": 1820000256
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.059541004654149e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.486173131673419,
      "tokens_seen": 1820131328
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.058738565238326e-05,
      "loss": 1.3741,
      "theoretical_loss": 3.4861519850438514,
      "tokens_seen": 1820262400
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.9930785298347473,
      "objective/train/avg_token_score": 0.006921479944139719,
      "objective/train/docs_used": 664244,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3643888235092163,
      "objective/train/original_loss": 1.374588966369629,
      "objective/train/theoretical_loss": 3.486130840363263,
      "objective/train/tokens_used": 190852576,
      "objective/train/unlikelihood_loss": 0.01842253841459751,
      "theoretical_loss": 3.486130840363263,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.057936125822501e-05,
      "loss": 1.3236,
      "theoretical_loss": 3.486130840363263,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.057133686406676e-05,
      "loss": 1.3808,
      "theoretical_loss": 3.486109697631333,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.056331246990853e-05,
      "loss": 1.3596,
      "theoretical_loss": 3.4860885568477413,
      "tokens_seen": 1820655616
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.055528807575028e-05,
      "loss": 1.209,
      "theoretical_loss": 3.486067418012169,
      "tokens_seen": 1820786688
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.054726368159205e-05,
      "loss": 1.4282,
      "theoretical_loss": 3.4860462811242954,
      "tokens_seen": 1820917760
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.05392392874338e-05,
      "loss": 1.3525,
      "theoretical_loss": 3.4860251461838017,
      "tokens_seen": 1821048832
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.053121489327556e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.486004013190368,
      "tokens_seen": 1821179904
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.052319049911732e-05,
      "loss": 1.3714,
      "theoretical_loss": 3.4859828821436745,
      "tokens_seen": 1821310976
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.051516610495908e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.4859617530434024,
      "tokens_seen": 1821442048
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.050714171080084e-05,
      "loss": 1.311,
      "theoretical_loss": 3.4859406258892314,
      "tokens_seen": 1821573120
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.04991173166426e-05,
      "loss": 1.303,
      "theoretical_loss": 3.4859195006808434,
      "tokens_seen": 1821704192
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.049109292248436e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.485898377417919,
      "tokens_seen": 1821835264
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.048306852832611e-05,
      "loss": 1.3673,
      "theoretical_loss": 3.485877256100138,
      "tokens_seen": 1821966336
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.047504413416787e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.4858561367271825,
      "tokens_seen": 1822097408
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.046701974000963e-05,
      "loss": 1.4003,
      "theoretical_loss": 3.4858350192987335,
      "tokens_seen": 1822228480
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.045899534585139e-05,
      "loss": 1.3367,
      "theoretical_loss": 3.4858139038144715,
      "tokens_seen": 1822359552
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.045097095169315e-05,
      "loss": 1.3852,
      "theoretical_loss": 3.485792790274078,
      "tokens_seen": 1822490624
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.044294655753491e-05,
      "loss": 1.4027,
      "theoretical_loss": 3.4857716786772346,
      "tokens_seen": 1822621696
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.043492216337666e-05,
      "loss": 1.387,
      "theoretical_loss": 3.4857505690236223,
      "tokens_seen": 1822752768
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.042689776921843e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.485729461312922,
      "tokens_seen": 1822883840
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.041887337506018e-05,
      "loss": 1.3456,
      "theoretical_loss": 3.485708355544816,
      "tokens_seen": 1823014912
    },
    {
      "epoch": 0.1,
      "learning_rate": 9.041084898090195e-05,
      "loss": 1.298,
      "theoretical_loss": 3.4856872517189856,
      "tokens_seen": 1823145984
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.04028245867437e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.4856661498351125,
      "tokens_seen": 1823277056
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.039480019258547e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.485645049892878,
      "tokens_seen": 1823408128
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.038677579842722e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.4856239518919647,
      "tokens_seen": 1823539200
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.037875140426897e-05,
      "loss": 1.3597,
      "theoretical_loss": 3.485602855832054,
      "tokens_seen": 1823670272
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.037072701011074e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.4855817617128277,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.036270261595249e-05,
      "loss": 1.3565,
      "theoretical_loss": 3.4855606695339683,
      "tokens_seen": 1823932416
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.035467822179426e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.4855395792951573,
      "tokens_seen": 1824063488
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.034665382763601e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.4855184909960775,
      "tokens_seen": 1824194560
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.033862943347777e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.4854974046364102,
      "tokens_seen": 1824325632
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.033060503931953e-05,
      "loss": 1.3439,
      "theoretical_loss": 3.4854763202158385,
      "tokens_seen": 1824456704
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.032258064516129e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.485455237734045,
      "tokens_seen": 1824587776
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.031455625100305e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.485434157190711,
      "tokens_seen": 1824718848
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.03065318568448e-05,
      "loss": 1.2129,
      "theoretical_loss": 3.4854130785855206,
      "tokens_seen": 1824849920
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.029850746268657e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.485392001918155,
      "tokens_seen": 1824980992
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.029048306852833e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.485370927188298,
      "tokens_seen": 1825112064
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.028245867437008e-05,
      "loss": 1.3499,
      "theoretical_loss": 3.485349854395632,
      "tokens_seen": 1825243136
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.027443428021185e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.4853287835398388,
      "tokens_seen": 1825374208
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.02664098860536e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.485307714620603,
      "tokens_seen": 1825505280
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.025838549189537e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.4852866476376065,
      "tokens_seen": 1825636352
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.025036109773712e-05,
      "loss": 1.3338,
      "theoretical_loss": 3.4852655825905328,
      "tokens_seen": 1825767424
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.024233670357888e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.485244519479065,
      "tokens_seen": 1825898496
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.023431230942064e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.4852234583028863,
      "tokens_seen": 1826029568
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.022628791526239e-05,
      "loss": 1.395,
      "theoretical_loss": 3.48520239906168,
      "tokens_seen": 1826160640
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.021826352110416e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.485181341755129,
      "tokens_seen": 1826291712
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.021023912694591e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.4851602863829174,
      "tokens_seen": 1826422784
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.020221473278768e-05,
      "loss": 1.27,
      "theoretical_loss": 3.4851392329447286,
      "tokens_seen": 1826553856
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.019419033862943e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.485118181440246,
      "tokens_seen": 1826684928
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.01861659444712e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.485097131869153,
      "tokens_seen": 1826816000
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.9935856461524963,
      "objective/train/avg_token_score": 0.006414412520825863,
      "objective/train/docs_used": 666029,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2992877960205078,
      "objective/train/original_loss": 1.3070013523101807,
      "objective/train/theoretical_loss": 3.485076084231134,
      "objective/train/tokens_used": 197406176,
      "objective/train/unlikelihood_loss": 0.021732643246650696,
      "theoretical_loss": 3.485076084231134,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.017814155031295e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.485076084231134,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.01701171561547e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.4850550385258723,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.016209276199647e-05,
      "loss": 1.4052,
      "theoretical_loss": 3.485033994753052,
      "tokens_seen": 1827209216
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.015406836783822e-05,
      "loss": 1.297,
      "theoretical_loss": 3.4850129529123564,
      "tokens_seen": 1827340288
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.014604397367999e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.484991913003471,
      "tokens_seen": 1827471360
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.013801957952174e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.484970875026079,
      "tokens_seen": 1827602432
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.012999518536351e-05,
      "loss": 1.3521,
      "theoretical_loss": 3.4849498389798637,
      "tokens_seen": 1827733504
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.012197079120526e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.484928804864511,
      "tokens_seen": 1827864576
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.011394639704702e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.484907772679705,
      "tokens_seen": 1827995648
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.010592200288878e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.484886742425129,
      "tokens_seen": 1828126720
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.009789760873054e-05,
      "loss": 1.3466,
      "theoretical_loss": 3.4848657141004677,
      "tokens_seen": 1828257792
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.00898732145723e-05,
      "loss": 1.3312,
      "theoretical_loss": 3.4848446877054062,
      "tokens_seen": 1828388864
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.008184882041406e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.4848236632396294,
      "tokens_seen": 1828519936
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.007382442625582e-05,
      "loss": 1.378,
      "theoretical_loss": 3.484802640702821,
      "tokens_seen": 1828651008
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.006580003209758e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.484781620094667,
      "tokens_seen": 1828782080
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.005777563793933e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.4847606014148518,
      "tokens_seen": 1828913152
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.00497512437811e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.4847395846630596,
      "tokens_seen": 1829044224
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.004172684962285e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.484718569838976,
      "tokens_seen": 1829175296
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.003370245546462e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.484697556942286,
      "tokens_seen": 1829306368
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.002567806130637e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.4846765459726745,
      "tokens_seen": 1829437440
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.001765366714814e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.484655536929827,
      "tokens_seen": 1829568512
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.000962927298989e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.484634529813429,
      "tokens_seen": 1829699584
    },
    {
      "epoch": 0.11,
      "learning_rate": 9.000160487883166e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.4846135246231658,
      "tokens_seen": 1829830656
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.999358048467341e-05,
      "loss": 1.3602,
      "theoretical_loss": 3.484592521358722,
      "tokens_seen": 1829961728
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.998555609051516e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.4845715200197844,
      "tokens_seen": 1830092800
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.997753169635693e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.484550520606038,
      "tokens_seen": 1830223872
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.996950730219868e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.484529523117168,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.996148290804045e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.484508527552861,
      "tokens_seen": 1830486016
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.99534585138822e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.4844875339128016,
      "tokens_seen": 1830617088
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.994543411972397e-05,
      "loss": 1.3328,
      "theoretical_loss": 3.484466542196677,
      "tokens_seen": 1830748160
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.993740972556572e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.484445552404173,
      "tokens_seen": 1830879232
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.992938533140747e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.484424564534975,
      "tokens_seen": 1831010304
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.992136093724924e-05,
      "loss": 1.3694,
      "theoretical_loss": 3.4844035785887693,
      "tokens_seen": 1831141376
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.9913336543091e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.4843825945652425,
      "tokens_seen": 1831272448
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.990531214893276e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.4843616124640797,
      "tokens_seen": 1831403520
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.989728775477451e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.484340632284969,
      "tokens_seen": 1831534592
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.988926336061628e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.484319654027595,
      "tokens_seen": 1831665664
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.988123896645803e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.484298677691645,
      "tokens_seen": 1831796736
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.987321457229979e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.484277703276806,
      "tokens_seen": 1831927808
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.986519017814155e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.4842567307827643,
      "tokens_seen": 1832058880
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.98571657839833e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.484235760209206,
      "tokens_seen": 1832189952
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.984914138982507e-05,
      "loss": 1.3715,
      "theoretical_loss": 3.4842147915558184,
      "tokens_seen": 1832321024
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.984111699566683e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.4841938248222886,
      "tokens_seen": 1832452096
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.983309260150859e-05,
      "loss": 1.3576,
      "theoretical_loss": 3.484172860008303,
      "tokens_seen": 1832583168
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.982506820735035e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.4841518971135486,
      "tokens_seen": 1832714240
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.98170438131921e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.484130936137713,
      "tokens_seen": 1832845312
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.980901941903387e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.484109977080483,
      "tokens_seen": 1832976384
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.980099502487562e-05,
      "loss": 1.259,
      "theoretical_loss": 3.4840890199415453,
      "tokens_seen": 1833107456
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.979297063071739e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.484068064720588,
      "tokens_seen": 1833238528
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.978494623655914e-05,
      "loss": 1.3828,
      "theoretical_loss": 3.484047111417298,
      "tokens_seen": 1833369600
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.9920766353607178,
      "objective/train/avg_token_score": 0.00792335532605648,
      "objective/train/docs_used": 668449,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.339785099029541,
      "objective/train/original_loss": 1.3505280017852783,
      "objective/train/theoretical_loss": 3.4840261600313633,
      "objective/train/tokens_used": 203959776,
      "objective/train/unlikelihood_loss": 0.022513849660754204,
      "theoretical_loss": 3.4840261600313633,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.97769218424009e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.4840261600313633,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.976889744824266e-05,
      "loss": 1.3666,
      "theoretical_loss": 3.4840052105624704,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.976087305408441e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.483984263010308,
      "tokens_seen": 1833762816
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.975284865992618e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.4839633173745628,
      "tokens_seen": 1833893888
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.974482426576793e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.483942373654923,
      "tokens_seen": 1834024960
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.97367998716097e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.4839214318510763,
      "tokens_seen": 1834156032
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.972877547745145e-05,
      "loss": 1.3586,
      "theoretical_loss": 3.483900491962711,
      "tokens_seen": 1834287104
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.972075108329322e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.4838795539895138,
      "tokens_seen": 1834418176
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.971272668913497e-05,
      "loss": 1.3873,
      "theoretical_loss": 3.483858617931174,
      "tokens_seen": 1834549248
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.970470229497674e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.483837683787379,
      "tokens_seen": 1834680320
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.969667790081849e-05,
      "loss": 1.2181,
      "theoretical_loss": 3.4838167515578173,
      "tokens_seen": 1834811392
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.968865350666024e-05,
      "loss": 1.2104,
      "theoretical_loss": 3.483795821242177,
      "tokens_seen": 1834942464
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.968062911250201e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.4837748928401466,
      "tokens_seen": 1835073536
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.967260471834376e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.483753966351414,
      "tokens_seen": 1835204608
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.966458032418553e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.4837330417756682,
      "tokens_seen": 1835335680
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.965655593002728e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.4837121191125977,
      "tokens_seen": 1835466752
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.964853153586905e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.483691198361891,
      "tokens_seen": 1835597824
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.96405071417108e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.4836702795232357,
      "tokens_seen": 1835728896
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.963248274755256e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.483649362596322,
      "tokens_seen": 1835859968
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.962445835339432e-05,
      "loss": 1.3656,
      "theoretical_loss": 3.4836284475808386,
      "tokens_seen": 1835991040
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.961643395923608e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.4836075344764734,
      "tokens_seen": 1836122112
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.960840956507784e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.4835866232829167,
      "tokens_seen": 1836253184
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.96003851709196e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.483565713999856,
      "tokens_seen": 1836384256
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.959236077676136e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.4835448066269814,
      "tokens_seen": 1836515328
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.958433638260312e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.4835239011639825,
      "tokens_seen": 1836646400
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.957631198844487e-05,
      "loss": 1.3359,
      "theoretical_loss": 3.4835029976105476,
      "tokens_seen": 1836777472
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.956828759428664e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.483482095966366,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.956026320012839e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.483461196231128,
      "tokens_seen": 1837039616
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.955223880597016e-05,
      "loss": 1.3719,
      "theoretical_loss": 3.4834402984045223,
      "tokens_seen": 1837170688
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.954421441181191e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.4834194024862386,
      "tokens_seen": 1837301760
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.953619001765368e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.483398508475967,
      "tokens_seen": 1837432832
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.952816562349543e-05,
      "loss": 1.3772,
      "theoretical_loss": 3.4833776163733967,
      "tokens_seen": 1837563904
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.952014122933718e-05,
      "loss": 1.4184,
      "theoretical_loss": 3.4833567261782177,
      "tokens_seen": 1837694976
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.951211683517895e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.4833358378901194,
      "tokens_seen": 1837826048
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.95040924410207e-05,
      "loss": 1.3406,
      "theoretical_loss": 3.483314951508792,
      "tokens_seen": 1837957120
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.949606804686247e-05,
      "loss": 1.3438,
      "theoretical_loss": 3.483294067033926,
      "tokens_seen": 1838088192
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.948804365270422e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.48327318446521,
      "tokens_seen": 1838219264
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.948001925854599e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.4832523038023355,
      "tokens_seen": 1838350336
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.947199486438774e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.483231425044993,
      "tokens_seen": 1838481408
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.94639704702295e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.4832105481928712,
      "tokens_seen": 1838612480
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.945594607607126e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.4831896732456613,
      "tokens_seen": 1838743552
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.944792168191301e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.483168800203054,
      "tokens_seen": 1838874624
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.943989728775478e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.4831479290647396,
      "tokens_seen": 1839005696
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.943187289359653e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.4831270598304087,
      "tokens_seen": 1839136768
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.94238484994383e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.4831061924997515,
      "tokens_seen": 1839267840
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.941582410528005e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.483085327072459,
      "tokens_seen": 1839398912
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.94077997111218e-05,
      "loss": 1.3292,
      "theoretical_loss": 3.4830644635482217,
      "tokens_seen": 1839529984
    },
    {
      "epoch": 0.11,
      "learning_rate": 8.939977531696357e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.483043601926731,
      "tokens_seen": 1839661056
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.939175092280533e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4830227422076776,
      "tokens_seen": 1839792128
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.93837265286471e-05,
      "loss": 1.281,
      "theoretical_loss": 3.4830018843907524,
      "tokens_seen": 1839923200
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.9910498261451721,
      "objective/train/avg_token_score": 0.008950138464570045,
      "objective/train/docs_used": 670853,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2321091890335083,
      "objective/train/original_loss": 1.244242548942566,
      "objective/train/theoretical_loss": 3.4829810284756464,
      "objective/train/tokens_used": 210513376,
      "objective/train/unlikelihood_loss": 0.02576657570898533,
      "theoretical_loss": 3.4829810284756464,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.937570213448885e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.4829810284756464,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.936767774033061e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.4829601744620513,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.935965334617237e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.482939322349657,
      "tokens_seen": 1840316416
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.935162895201413e-05,
      "loss": 1.2645,
      "theoretical_loss": 3.4829184721381568,
      "tokens_seen": 1840447488
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.934360455785589e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.48289762382724,
      "tokens_seen": 1840578560
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.933558016369764e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.482876777416599,
      "tokens_seen": 1840709632
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.93275557695394e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.4828559329059257,
      "tokens_seen": 1840840704
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.931953137538116e-05,
      "loss": 1.3616,
      "theoretical_loss": 3.482835090294911,
      "tokens_seen": 1840971776
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.931150698122293e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.4828142495832464,
      "tokens_seen": 1841102848
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.930348258706468e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.482793410770624,
      "tokens_seen": 1841233920
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.929545819290645e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.482772573856736,
      "tokens_seen": 1841364992
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.92874337987482e-05,
      "loss": 1.3637,
      "theoretical_loss": 3.4827517388412734,
      "tokens_seen": 1841496064
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.927940940458995e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.482730905723929,
      "tokens_seen": 1841627136
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.927138501043172e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.4827100745043937,
      "tokens_seen": 1841758208
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.926336061627347e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.4826892451823603,
      "tokens_seen": 1841889280
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.925533622211524e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.4826684177575213,
      "tokens_seen": 1842020352
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.924731182795699e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.4826475922295677,
      "tokens_seen": 1842151424
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.923928743379876e-05,
      "loss": 1.3494,
      "theoretical_loss": 3.4826267685981933,
      "tokens_seen": 1842282496
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.923126303964051e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.4826059468630897,
      "tokens_seen": 1842413568
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.922323864548226e-05,
      "loss": 1.3418,
      "theoretical_loss": 3.482585127023949,
      "tokens_seen": 1842544640
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.921521425132403e-05,
      "loss": 1.3706,
      "theoretical_loss": 3.4825643090804643,
      "tokens_seen": 1842675712
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.920718985716578e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.4825434930323276,
      "tokens_seen": 1842806784
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.919916546300755e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.4825226788792323,
      "tokens_seen": 1842937856
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.91911410688493e-05,
      "loss": 1.3695,
      "theoretical_loss": 3.4825018666208702,
      "tokens_seen": 1843068928
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.918311667469107e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.482481056256935,
      "tokens_seen": 1843200000
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.917509228053282e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.4824602477871194,
      "tokens_seen": 1843331072
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.916706788637458e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.482439441211116,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.915904349221634e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.4824186365286174,
      "tokens_seen": 1843593216
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.91510190980581e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.4823978337393173,
      "tokens_seen": 1843724288
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.914299470389986e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.4823770328429093,
      "tokens_seen": 1843855360
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.913497030974162e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.4823562338390857,
      "tokens_seen": 1843986432
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.912694591558338e-05,
      "loss": 1.3574,
      "theoretical_loss": 3.4823354367275403,
      "tokens_seen": 1844117504
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.911892152142514e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.482314641507966,
      "tokens_seen": 1844248576
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.911089712726689e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.4822938481800567,
      "tokens_seen": 1844379648
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.910287273310866e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.482273056743506,
      "tokens_seen": 1844510720
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.909484833895041e-05,
      "loss": 1.284,
      "theoretical_loss": 3.4822522671980067,
      "tokens_seen": 1844641792
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.908682394479218e-05,
      "loss": 1.3595,
      "theoretical_loss": 3.4822314795432536,
      "tokens_seen": 1844772864
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.907879955063393e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.4822106937789394,
      "tokens_seen": 1844903936
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.90707751564757e-05,
      "loss": 1.324,
      "theoretical_loss": 3.4821899099047586,
      "tokens_seen": 1845035008
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.906275076231745e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.482169127920405,
      "tokens_seen": 1845166080
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.905472636815922e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.482148347825572,
      "tokens_seen": 1845297152
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.904670197400097e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.482127569619954,
      "tokens_seen": 1845428224
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.903867757984272e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.482106793303245,
      "tokens_seen": 1845559296
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.903065318568449e-05,
      "loss": 1.3642,
      "theoretical_loss": 3.482086018875139,
      "tokens_seen": 1845690368
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.902262879152624e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.4820652463353303,
      "tokens_seen": 1845821440
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.901460439736801e-05,
      "loss": 1.3688,
      "theoretical_loss": 3.4820444756835136,
      "tokens_seen": 1845952512
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.900658000320976e-05,
      "loss": 1.3696,
      "theoretical_loss": 3.482023706919383,
      "tokens_seen": 1846083584
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.899855560905153e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.4820029400426327,
      "tokens_seen": 1846214656
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.899053121489328e-05,
      "loss": 1.2542,
      "theoretical_loss": 3.481982175052957,
      "tokens_seen": 1846345728
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.898250682073503e-05,
      "loss": 1.3755,
      "theoretical_loss": 3.481961411950051,
      "tokens_seen": 1846476800
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.9902796149253845,
      "objective/train/avg_token_score": 0.009720345959067345,
      "objective/train/docs_used": 673286,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3855735063552856,
      "objective/train/original_loss": 1.4005990028381348,
      "objective/train/theoretical_loss": 3.4819406507336095,
      "objective/train/tokens_used": 217066976,
      "objective/train/unlikelihood_loss": 0.02520112879574299,
      "theoretical_loss": 3.4819406507336095,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.89744824265768e-05,
      "loss": 1.3539,
      "theoretical_loss": 3.4819406507336095,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.896645803241855e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4819198914033267,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.895843363826032e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.4818991339588976,
      "tokens_seen": 1846870016
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.895040924410207e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.481878378400017,
      "tokens_seen": 1847001088
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.894238484994384e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.4818576247263806,
      "tokens_seen": 1847132160
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.89343604557856e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.4818368729376816,
      "tokens_seen": 1847263232
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.892633606162735e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.4818161230336173,
      "tokens_seen": 1847394304
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.891831166746911e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4817953750138813,
      "tokens_seen": 1847525376
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.891028727331087e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.481774628878169,
      "tokens_seen": 1847656448
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.890226287915263e-05,
      "loss": 1.3885,
      "theoretical_loss": 3.4817538846261757,
      "tokens_seen": 1847787520
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.889423848499439e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.4817331422575974,
      "tokens_seen": 1847918592
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.888621409083615e-05,
      "loss": 1.349,
      "theoretical_loss": 3.4817124017721293,
      "tokens_seen": 1848049664
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.88781896966779e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.4816916631694665,
      "tokens_seen": 1848180736
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.887016530251966e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.4816709264493046,
      "tokens_seen": 1848311808
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.886214090836143e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.4816501916113394,
      "tokens_seen": 1848442880
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.885411651420318e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.481629458655267,
      "tokens_seen": 1848573952
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.884609212004495e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4816087275807828,
      "tokens_seen": 1848705024
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.88380677258867e-05,
      "loss": 1.3942,
      "theoretical_loss": 3.4815879983875826,
      "tokens_seen": 1848836096
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.883004333172847e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.4815672710753622,
      "tokens_seen": 1848967168
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.882201893757022e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.481546545643818,
      "tokens_seen": 1849098240
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.881399454341197e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.481525822092646,
      "tokens_seen": 1849229312
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.880597014925374e-05,
      "loss": 1.3416,
      "theoretical_loss": 3.4815051004215416,
      "tokens_seen": 1849360384
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.879794575509549e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.481484380630202,
      "tokens_seen": 1849491456
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.878992136093726e-05,
      "loss": 1.3537,
      "theoretical_loss": 3.4814636627183235,
      "tokens_seen": 1849622528
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.878189696677901e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.481442946685602,
      "tokens_seen": 1849753600
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.877387257262078e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.481422232531733,
      "tokens_seen": 1849884672
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.876584817846253e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.481401520256415,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.87578237843043e-05,
      "loss": 1.367,
      "theoretical_loss": 3.481380809859343,
      "tokens_seen": 1850146816
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.874979939014605e-05,
      "loss": 1.4265,
      "theoretical_loss": 3.481360101340214,
      "tokens_seen": 1850277888
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.87417749959878e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.4813393946987254,
      "tokens_seen": 1850408960
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.873375060182957e-05,
      "loss": 1.288,
      "theoretical_loss": 3.481318689934573,
      "tokens_seen": 1850540032
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.872572620767132e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.481297987047454,
      "tokens_seen": 1850671104
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.871770181351309e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.4812772860370655,
      "tokens_seen": 1850802176
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.870967741935484e-05,
      "loss": 1.3504,
      "theoretical_loss": 3.4812565869031045,
      "tokens_seen": 1850933248
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.870165302519661e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.481235889645268,
      "tokens_seen": 1851064320
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.869362863103836e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.4812151942632523,
      "tokens_seen": 1851195392
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.868560423688012e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.4811945007567564,
      "tokens_seen": 1851326464
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.867757984272188e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.4811738091254756,
      "tokens_seen": 1851457536
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.866955544856364e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.4811531193691083,
      "tokens_seen": 1851588608
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.86615310544054e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.481132431487352,
      "tokens_seen": 1851719680
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.865350666024716e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.481111745479904,
      "tokens_seen": 1851850752
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.864548226608892e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.4810910613464614,
      "tokens_seen": 1851981824
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.863745787193068e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.481070379086722,
      "tokens_seen": 1852112896
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.862943347777243e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.4810496987003843,
      "tokens_seen": 1852243968
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.86214090836142e-05,
      "loss": 1.283,
      "theoretical_loss": 3.4810290201871448,
      "tokens_seen": 1852375040
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.861338468945595e-05,
      "loss": 1.35,
      "theoretical_loss": 3.4810083435467023,
      "tokens_seen": 1852506112
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.860536029529772e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.480987668778754,
      "tokens_seen": 1852637184
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.859733590113947e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.4809669958829983,
      "tokens_seen": 1852768256
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.858931150698124e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.4809463248591332,
      "tokens_seen": 1852899328
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.858128711282299e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.480925655706857,
      "tokens_seen": 1853030400
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.9897891879081726,
      "objective/train/avg_token_score": 0.010210780426859856,
      "objective/train/docs_used": 675642,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3331942558288574,
      "objective/train/original_loss": 1.3470110893249512,
      "objective/train/theoretical_loss": 3.480904988425867,
      "objective/train/tokens_used": 223620576,
      "objective/train/unlikelihood_loss": 0.029013056308031082,
      "theoretical_loss": 3.480904988425867,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.857326271866474e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.480904988425867,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.856523832450651e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.4808843230158626,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.855721393034826e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.480863659476541,
      "tokens_seen": 1853423616
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.854918953619003e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.480842997807602,
      "tokens_seen": 1853554688
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.854116514203178e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4808223380087426,
      "tokens_seen": 1853685760
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.853314074787355e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.4808016800796624,
      "tokens_seen": 1853816832
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.85251163537153e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.480781024020059,
      "tokens_seen": 1853947904
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.851709195955705e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.480760369829632,
      "tokens_seen": 1854078976
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.850906756539882e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.4807397175080794,
      "tokens_seen": 1854210048
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.850104317124057e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.4807190670551007,
      "tokens_seen": 1854341120
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.849301877708234e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.480698418470394,
      "tokens_seen": 1854472192
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.84849943829241e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.480677771753659,
      "tokens_seen": 1854603264
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.847696998876586e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.4806571269045943,
      "tokens_seen": 1854734336
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.846894559460761e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.480636483922899,
      "tokens_seen": 1854865408
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.846092120044937e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.4806158428082723,
      "tokens_seen": 1854996480
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.845289680629113e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.480595203560414,
      "tokens_seen": 1855127552
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.844487241213289e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.480574566179022,
      "tokens_seen": 1855258624
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.843684801797465e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.480553930663797,
      "tokens_seen": 1855389696
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.842882362381641e-05,
      "loss": 1.3404,
      "theoretical_loss": 3.4805332970144383,
      "tokens_seen": 1855520768
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.842079922965817e-05,
      "loss": 1.302,
      "theoretical_loss": 3.4805126652306444,
      "tokens_seen": 1855651840
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.841277483549993e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.480492035312116,
      "tokens_seen": 1855782912
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.84047504413417e-05,
      "loss": 1.22,
      "theoretical_loss": 3.480471407258552,
      "tokens_seen": 1855913984
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.839672604718345e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.480450781069652,
      "tokens_seen": 1856045056
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.83887016530252e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.4804301567451166,
      "tokens_seen": 1856176128
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.838067725886697e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.480409534284645,
      "tokens_seen": 1856307200
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.837265286470872e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.4803889136879373,
      "tokens_seen": 1856438272
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.836462847055049e-05,
      "loss": 1.3594,
      "theoretical_loss": 3.480368294954694,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.835660407639224e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.4803476780846143,
      "tokens_seen": 1856700416
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.8348579682234e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.4803270630773984,
      "tokens_seen": 1856831488
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.834055528807576e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.4803064499327476,
      "tokens_seen": 1856962560
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.833253089391751e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.480285838650361,
      "tokens_seen": 1857093632
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.832450649975928e-05,
      "loss": 1.3404,
      "theoretical_loss": 3.480265229229939,
      "tokens_seen": 1857224704
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.831648210560103e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.4802446216711824,
      "tokens_seen": 1857355776
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.83084577114428e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.480224015973792,
      "tokens_seen": 1857486848
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.830043331728455e-05,
      "loss": 1.3427,
      "theoretical_loss": 3.4802034121374676,
      "tokens_seen": 1857617920
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.829240892312632e-05,
      "loss": 1.3726,
      "theoretical_loss": 3.48018281016191,
      "tokens_seen": 1857748992
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.828438452896807e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.480162210046821,
      "tokens_seen": 1857880064
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.827636013480982e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.480141611791899,
      "tokens_seen": 1858011136
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.826833574065159e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.480121015396847,
      "tokens_seen": 1858142208
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.826031134649334e-05,
      "loss": 1.3591,
      "theoretical_loss": 3.480100420861365,
      "tokens_seen": 1858273280
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.825228695233511e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.480079828185154,
      "tokens_seen": 1858404352
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.824426255817686e-05,
      "loss": 1.332,
      "theoretical_loss": 3.4800592373679153,
      "tokens_seen": 1858535424
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.823623816401863e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.4800386484093497,
      "tokens_seen": 1858666496
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.822821376986038e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.480018061309158,
      "tokens_seen": 1858797568
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.822018937570214e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.4799974760670427,
      "tokens_seen": 1858928640
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.82121649815439e-05,
      "loss": 1.2249,
      "theoretical_loss": 3.4799768926827035,
      "tokens_seen": 1859059712
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.820414058738566e-05,
      "loss": 1.3808,
      "theoretical_loss": 3.479956311155843,
      "tokens_seen": 1859190784
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.819611619322742e-05,
      "loss": 1.1989,
      "theoretical_loss": 3.4799357314861625,
      "tokens_seen": 1859321856
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.818809179906918e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.479915153673363,
      "tokens_seen": 1859452928
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.818006740491094e-05,
      "loss": 1.25,
      "theoretical_loss": 3.479894577717146,
      "tokens_seen": 1859584000
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.9955404996871948,
      "objective/train/avg_token_score": 0.0044595031067729,
      "objective/train/docs_used": 678160,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.266061782836914,
      "objective/train/original_loss": 1.272694706916809,
      "objective/train/theoretical_loss": 3.4798740036172138,
      "objective/train/tokens_used": 230174176,
      "objective/train/unlikelihood_loss": 0.012577024288475513,
      "theoretical_loss": 3.4798740036172138,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.81720430107527e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.4798740036172138,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.816401861659445e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.4798534313732676,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.815599422243622e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.4798328609850095,
      "tokens_seen": 1859977216
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.814796982827797e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.4798122924521415,
      "tokens_seen": 1860108288
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.813994543411974e-05,
      "loss": 1.3382,
      "theoretical_loss": 3.4797917257743647,
      "tokens_seen": 1860239360
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.813192103996149e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.479771160951382,
      "tokens_seen": 1860370432
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.812389664580326e-05,
      "loss": 1.266,
      "theoretical_loss": 3.4797505979828953,
      "tokens_seen": 1860501504
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.811587225164501e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.479730036868606,
      "tokens_seen": 1860632576
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.810784785748678e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.4797094776082176,
      "tokens_seen": 1860763648
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.809982346332853e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.479688920201431,
      "tokens_seen": 1860894720
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.809179906917028e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.4796683646479494,
      "tokens_seen": 1861025792
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.808377467501205e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.4796478109474753,
      "tokens_seen": 1861156864
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.80757502808538e-05,
      "loss": 1.3261,
      "theoretical_loss": 3.4796272590997104,
      "tokens_seen": 1861287936
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.806772588669557e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.479606709104358,
      "tokens_seen": 1861419008
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.805970149253732e-05,
      "loss": 1.3495,
      "theoretical_loss": 3.4795861609611203,
      "tokens_seen": 1861550080
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.805167709837908e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.4795656146697,
      "tokens_seen": 1861681152
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.804365270422084e-05,
      "loss": 1.362,
      "theoretical_loss": 3.4795450702298005,
      "tokens_seen": 1861812224
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.80356283100626e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4795245276411233,
      "tokens_seen": 1861943296
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.802760391590436e-05,
      "loss": 1.4213,
      "theoretical_loss": 3.4795039869033726,
      "tokens_seen": 1862074368
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.801957952174611e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.479483448016251,
      "tokens_seen": 1862205440
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.801155512758788e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.4794629109794606,
      "tokens_seen": 1862336512
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.800353073342963e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.479442375792706,
      "tokens_seen": 1862467584
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.799550633927139e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.4794218424556895,
      "tokens_seen": 1862598656
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.798748194511315e-05,
      "loss": 1.294,
      "theoretical_loss": 3.479401310968114,
      "tokens_seen": 1862729728
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.797945755095491e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.4793807813296835,
      "tokens_seen": 1862860800
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.797143315679667e-05,
      "loss": 1.2245,
      "theoretical_loss": 3.4793602535401016,
      "tokens_seen": 1862991872
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.796340876263843e-05,
      "loss": 1.292,
      "theoretical_loss": 3.4793397275990703,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.795538436848018e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.4793192035062948,
      "tokens_seen": 1863254016
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.794735997432195e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.4792986812614775,
      "tokens_seen": 1863385088
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.79393355801637e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.479278160864323,
      "tokens_seen": 1863516160
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.793131118600547e-05,
      "loss": 1.3416,
      "theoretical_loss": 3.4792576423145345,
      "tokens_seen": 1863647232
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.792328679184722e-05,
      "loss": 1.318,
      "theoretical_loss": 3.4792371256118155,
      "tokens_seen": 1863778304
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.791526239768899e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.47921661075587,
      "tokens_seen": 1863909376
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.790723800353074e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.479196097746402,
      "tokens_seen": 1864040448
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.789921360937249e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.479175586583116,
      "tokens_seen": 1864171520
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.789118921521426e-05,
      "loss": 1.3319,
      "theoretical_loss": 3.479155077265715,
      "tokens_seen": 1864302592
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.788316482105601e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.479134569793904,
      "tokens_seen": 1864433664
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.787514042689778e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.479114064167387,
      "tokens_seen": 1864564736
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.786711603273953e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.4790935603858677,
      "tokens_seen": 1864695808
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.785909163858129e-05,
      "loss": 1.3571,
      "theoretical_loss": 3.4790730584490515,
      "tokens_seen": 1864826880
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.785106724442305e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.479052558356642,
      "tokens_seen": 1864957952
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.78430428502648e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.4790320601083433,
      "tokens_seen": 1865089024
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.783501845610657e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.4790115637038608,
      "tokens_seen": 1865220096
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.782699406194833e-05,
      "loss": 1.265,
      "theoretical_loss": 3.478991069142899,
      "tokens_seen": 1865351168
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.781896966779009e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.478970576425162,
      "tokens_seen": 1865482240
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.781094527363185e-05,
      "loss": 1.3325,
      "theoretical_loss": 3.478950085550355,
      "tokens_seen": 1865613312
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.78029208794736e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.478929596518183,
      "tokens_seen": 1865744384
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.779489648531536e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.4789091093283497,
      "tokens_seen": 1865875456
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.778687209115712e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.478888623980562,
      "tokens_seen": 1866006528
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.777884769699888e-05,
      "loss": 1.3702,
      "theoretical_loss": 3.478868140474523,
      "tokens_seen": 1866137600
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.9927670359611511,
      "objective/train/avg_token_score": 0.007232929579913616,
      "objective/train/docs_used": 681054,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3302700519561768,
      "objective/train/original_loss": 1.3410484790802002,
      "objective/train/theoretical_loss": 3.478847658809938,
      "objective/train/tokens_used": 236727776,
      "objective/train/unlikelihood_loss": 0.01834222860634327,
      "theoretical_loss": 3.478847658809938,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.777082330284064e-05,
      "loss": 1.3292,
      "theoretical_loss": 3.478847658809938,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.776279890868239e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.4788271789865135,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.775477451452416e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.478806701003954,
      "tokens_seen": 1866530816
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.774675012036591e-05,
      "loss": 1.2434,
      "theoretical_loss": 3.4787862248619645,
      "tokens_seen": 1866661888
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.773872572620768e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.4787657505602505,
      "tokens_seen": 1866792960
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.773070133204943e-05,
      "loss": 1.3623,
      "theoretical_loss": 3.4787452780985175,
      "tokens_seen": 1866924032
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.77226769378912e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.478724807476471,
      "tokens_seen": 1867055104
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.771465254373295e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.4787043386938166,
      "tokens_seen": 1867186176
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.77066281495747e-05,
      "loss": 1.4082,
      "theoretical_loss": 3.4786838717502593,
      "tokens_seen": 1867317248
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.769860375541647e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.4786634066455058,
      "tokens_seen": 1867448320
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.769057936125822e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.4786429433792607,
      "tokens_seen": 1867579392
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.768255496709999e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.4786224819512306,
      "tokens_seen": 1867710464
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.767453057294174e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.478602022361122,
      "tokens_seen": 1867841536
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.76665061787835e-05,
      "loss": 1.3458,
      "theoretical_loss": 3.478581564608639,
      "tokens_seen": 1867972608
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.765848178462526e-05,
      "loss": 1.3598,
      "theoretical_loss": 3.47856110869349,
      "tokens_seen": 1868103680
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.765045739046702e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.478540654615379,
      "tokens_seen": 1868234752
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.764243299630878e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.478520202374013,
      "tokens_seen": 1868365824
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.763440860215054e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.478499751969098,
      "tokens_seen": 1868496896
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.76263842079923e-05,
      "loss": 1.3698,
      "theoretical_loss": 3.478479303400341,
      "tokens_seen": 1868627968
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.761835981383406e-05,
      "loss": 1.3435,
      "theoretical_loss": 3.478458856667448,
      "tokens_seen": 1868759040
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.761033541967581e-05,
      "loss": 1.4536,
      "theoretical_loss": 3.4784384117701244,
      "tokens_seen": 1868890112
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.760231102551758e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.4784179687080785,
      "tokens_seen": 1869021184
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.759428663135933e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.4783975274810155,
      "tokens_seen": 1869152256
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.75862622372011e-05,
      "loss": 1.2819,
      "theoretical_loss": 3.478377088088642,
      "tokens_seen": 1869283328
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.757823784304285e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.478356650530666,
      "tokens_seen": 1869414400
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.75702134488846e-05,
      "loss": 1.367,
      "theoretical_loss": 3.4783362148067933,
      "tokens_seen": 1869545472
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.756218905472637e-05,
      "loss": 1.3335,
      "theoretical_loss": 3.4783157809167307,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.755416466056812e-05,
      "loss": 1.432,
      "theoretical_loss": 3.4782953488601853,
      "tokens_seen": 1869807616
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.754614026640989e-05,
      "loss": 1.3432,
      "theoretical_loss": 3.478274918636864,
      "tokens_seen": 1869938688
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.753811587225164e-05,
      "loss": 1.2054,
      "theoretical_loss": 3.4782544902464743,
      "tokens_seen": 1870069760
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.753009147809341e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.4782340636887223,
      "tokens_seen": 1870200832
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.752206708393516e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.4782136389633163,
      "tokens_seen": 1870331904
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.751404268977691e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.4781932160699625,
      "tokens_seen": 1870462976
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.750601829561868e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.478172795008369,
      "tokens_seen": 1870594048
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.749799390146043e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.478152375778243,
      "tokens_seen": 1870725120
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.74899695073022e-05,
      "loss": 1.3634,
      "theoretical_loss": 3.4781319583792913,
      "tokens_seen": 1870856192
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.748194511314395e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.478111542811222,
      "tokens_seen": 1870987264
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.747392071898572e-05,
      "loss": 1.3868,
      "theoretical_loss": 3.4780911290737433,
      "tokens_seen": 1871118336
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.746589632482747e-05,
      "loss": 1.3795,
      "theoretical_loss": 3.4780707171665615,
      "tokens_seen": 1871249408
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.745787193066924e-05,
      "loss": 1.3658,
      "theoretical_loss": 3.478050307089385,
      "tokens_seen": 1871380480
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.7449847536511e-05,
      "loss": 1.3619,
      "theoretical_loss": 3.478029898841922,
      "tokens_seen": 1871511552
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.744182314235275e-05,
      "loss": 1.319,
      "theoretical_loss": 3.4780094924238796,
      "tokens_seen": 1871642624
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.743379874819451e-05,
      "loss": 1.3785,
      "theoretical_loss": 3.477989087834966,
      "tokens_seen": 1871773696
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.742577435403627e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.477968685074889,
      "tokens_seen": 1871904768
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.741774995987803e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.477948284143357,
      "tokens_seen": 1872035840
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.740972556571979e-05,
      "loss": 1.336,
      "theoretical_loss": 3.477927885040078,
      "tokens_seen": 1872166912
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.740170117156155e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.4779074877647593,
      "tokens_seen": 1872297984
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.73936767774033e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.477887092317111,
      "tokens_seen": 1872429056
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.738565238324506e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.4778666986968396,
      "tokens_seen": 1872560128
    },
    {
      "epoch": 0.13,
      "learning_rate": 8.737762798908683e-05,
      "loss": 1.3446,
      "theoretical_loss": 3.477846306903655,
      "tokens_seen": 1872691200
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.9942508935928345,
      "objective/train/avg_token_score": 0.005749125499278307,
      "objective/train/docs_used": 683408,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3061952590942383,
      "objective/train/original_loss": 1.3137394189834595,
      "objective/train/theoretical_loss": 3.477825916937264,
      "objective/train/tokens_used": 243281376,
      "objective/train/unlikelihood_loss": 0.017372548580169678,
      "theoretical_loss": 3.477825916937264,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.736960359492858e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.477825916937264,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.736157920077035e-05,
      "loss": 1.3937,
      "theoretical_loss": 3.4778055287973766,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.73535548066121e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.4777851424837007,
      "tokens_seen": 1873084416
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.734553041245387e-05,
      "loss": 1.3747,
      "theoretical_loss": 3.4777647579959448,
      "tokens_seen": 1873215488
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.733750601829562e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.4777443753338178,
      "tokens_seen": 1873346560
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.732948162413737e-05,
      "loss": 1.34,
      "theoretical_loss": 3.477723994497029,
      "tokens_seen": 1873477632
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.732145722997914e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.4777036154852863,
      "tokens_seen": 1873608704
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.731343283582089e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.477683238298299,
      "tokens_seen": 1873739776
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.730540844166266e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.4776628629357766,
      "tokens_seen": 1873870848
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.729738404750441e-05,
      "loss": 1.2114,
      "theoretical_loss": 3.477642489397428,
      "tokens_seen": 1874001920
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.728935965334618e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.4776221176829614,
      "tokens_seen": 1874132992
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.728133525918793e-05,
      "loss": 1.3653,
      "theoretical_loss": 3.4776017477920864,
      "tokens_seen": 1874264064
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.727331086502968e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.477581379724513,
      "tokens_seen": 1874395136
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.726528647087145e-05,
      "loss": 1.355,
      "theoretical_loss": 3.4775610134799497,
      "tokens_seen": 1874526208
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.72572620767132e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.4775406490581062,
      "tokens_seen": 1874657280
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.724923768255497e-05,
      "loss": 1.347,
      "theoretical_loss": 3.4775202864586916,
      "tokens_seen": 1874788352
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.724121328839672e-05,
      "loss": 1.3775,
      "theoretical_loss": 3.4774999256814163,
      "tokens_seen": 1874919424
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.723318889423849e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.477479566725989,
      "tokens_seen": 1875050496
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.722516450008024e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.4774592095921197,
      "tokens_seen": 1875181568
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.7217140105922e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.477438854279518,
      "tokens_seen": 1875312640
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.720911571176376e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.477418500787894,
      "tokens_seen": 1875443712
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.720109131760552e-05,
      "loss": 1.3247,
      "theoretical_loss": 3.477398149116957,
      "tokens_seen": 1875574784
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.719306692344728e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.477377799266417,
      "tokens_seen": 1875705856
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.718504252928904e-05,
      "loss": 1.261,
      "theoretical_loss": 3.477357451235984,
      "tokens_seen": 1875836928
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.71770181351308e-05,
      "loss": 1.3561,
      "theoretical_loss": 3.477337105025369,
      "tokens_seen": 1875968000
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.716899374097256e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.477316760634281,
      "tokens_seen": 1876099072
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.716096934681432e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.477296418062431,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.715294495265608e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.477276077309528,
      "tokens_seen": 1876361216
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.714492055849783e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.477255738375283,
      "tokens_seen": 1876492288
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.71368961643396e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.4772354012594073,
      "tokens_seen": 1876623360
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.712887177018135e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.4772150659616097,
      "tokens_seen": 1876754432
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.712084737602312e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.4771947324816024,
      "tokens_seen": 1876885504
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.711282298186487e-05,
      "loss": 1.3201,
      "theoretical_loss": 3.477174400819094,
      "tokens_seen": 1877016576
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.710479858770664e-05,
      "loss": 1.345,
      "theoretical_loss": 3.477154070973797,
      "tokens_seen": 1877147648
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.709677419354839e-05,
      "loss": 1.3359,
      "theoretical_loss": 3.477133742945421,
      "tokens_seen": 1877278720
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.708874979939014e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.4771134167336775,
      "tokens_seen": 1877409792
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.708072540523191e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.4770930923382766,
      "tokens_seen": 1877540864
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.707270101107366e-05,
      "loss": 1.4048,
      "theoretical_loss": 3.47707276975893,
      "tokens_seen": 1877671936
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.706467661691543e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.477052448995348,
      "tokens_seen": 1877803008
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.705665222275718e-05,
      "loss": 1.33,
      "theoretical_loss": 3.4770321300472418,
      "tokens_seen": 1877934080
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.704862782859895e-05,
      "loss": 1.3597,
      "theoretical_loss": 3.4770118129143226,
      "tokens_seen": 1878065152
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.70406034344407e-05,
      "loss": 1.3506,
      "theoretical_loss": 3.476991497596302,
      "tokens_seen": 1878196224
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.703257904028245e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.47697118409289,
      "tokens_seen": 1878327296
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.702455464612422e-05,
      "loss": 1.2949,
      "theoretical_loss": 3.4769508724038,
      "tokens_seen": 1878458368
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.701653025196597e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.476930562528741,
      "tokens_seen": 1878589440
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.700850585780774e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.476910254467426,
      "tokens_seen": 1878720512
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.70004814636495e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.476889948219566,
      "tokens_seen": 1878851584
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.699245706949126e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.4768696437848723,
      "tokens_seen": 1878982656
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.698443267533301e-05,
      "loss": 1.3635,
      "theoretical_loss": 3.4768493411630574,
      "tokens_seen": 1879113728
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.697640828117477e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.476829040353832,
      "tokens_seen": 1879244800
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.9950573444366455,
      "objective/train/avg_token_score": 0.004942657891660929,
      "objective/train/docs_used": 685672,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2672836780548096,
      "objective/train/original_loss": 1.273506999015808,
      "objective/train/theoretical_loss": 3.4768087413569084,
      "objective/train/tokens_used": 249834976,
      "objective/train/unlikelihood_loss": 0.01577509008347988,
      "theoretical_loss": 3.4768087413569084,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.696838388701653e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.4768087413569084,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.696035949285829e-05,
      "loss": 1.3565,
      "theoretical_loss": 3.4767884441719987,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.695233509870005e-05,
      "loss": 1.255,
      "theoretical_loss": 3.476768148798814,
      "tokens_seen": 1879638016
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.69443107045418e-05,
      "loss": 1.3816,
      "theoretical_loss": 3.4767478552370674,
      "tokens_seen": 1879769088
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.693628631038357e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.4767275634864694,
      "tokens_seen": 1879900160
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.692826191622533e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.476707273546733,
      "tokens_seen": 1880031232
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.692023752206708e-05,
      "loss": 1.311,
      "theoretical_loss": 3.476686985417571,
      "tokens_seen": 1880162304
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.691221312790885e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.4766666990986947,
      "tokens_seen": 1880293376
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.69041887337506e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.4766464145898164,
      "tokens_seen": 1880424448
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.689616433959237e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.476626131890649,
      "tokens_seen": 1880555520
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.688813994543412e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.4766058510009046,
      "tokens_seen": 1880686592
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.688011555127589e-05,
      "loss": 1.3087,
      "theoretical_loss": 3.476585571920295,
      "tokens_seen": 1880817664
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.687209115711764e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.476565294648535,
      "tokens_seen": 1880948736
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.68640667629594e-05,
      "loss": 1.3644,
      "theoretical_loss": 3.476545019185334,
      "tokens_seen": 1881079808
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.685604236880116e-05,
      "loss": 1.325,
      "theoretical_loss": 3.4765247455304076,
      "tokens_seen": 1881210880
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.684801797464291e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.476504473683467,
      "tokens_seen": 1881341952
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.683999358048468e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.476484203644225,
      "tokens_seen": 1881473024
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.683196918632643e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.4764639354123954,
      "tokens_seen": 1881604096
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.68239447921682e-05,
      "loss": 1.248,
      "theoretical_loss": 3.4764436689876903,
      "tokens_seen": 1881735168
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.681592039800995e-05,
      "loss": 1.3541,
      "theoretical_loss": 3.476423404369823,
      "tokens_seen": 1881866240
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.680789600385172e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.4764031415585066,
      "tokens_seen": 1881997312
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.679987160969347e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.4763828805534547,
      "tokens_seen": 1882128384
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.679184721553522e-05,
      "loss": 1.4175,
      "theoretical_loss": 3.4763626213543795,
      "tokens_seen": 1882259456
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.678382282137699e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.4763423639609954,
      "tokens_seen": 1882390528
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.677579842721874e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.4763221083730147,
      "tokens_seen": 1882521600
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.676777403306051e-05,
      "loss": 1.4167,
      "theoretical_loss": 3.4763018545901514,
      "tokens_seen": 1882652672
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.675974963890226e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.476281602612119,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.675172524474403e-05,
      "loss": 1.3751,
      "theoretical_loss": 3.476261352438631,
      "tokens_seen": 1882914816
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.674370085058578e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.476241104069401,
      "tokens_seen": 1883045888
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.673567645642754e-05,
      "loss": 1.3543,
      "theoretical_loss": 3.4762208575041424,
      "tokens_seen": 1883176960
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.67276520622693e-05,
      "loss": 1.3945,
      "theoretical_loss": 3.476200612742569,
      "tokens_seen": 1883308032
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.671962766811106e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.476180369784395,
      "tokens_seen": 1883439104
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.671160327395282e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.476160128629334,
      "tokens_seen": 1883570176
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.670357887979458e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.4761398892771,
      "tokens_seen": 1883701248
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.669555448563634e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.476119651727407,
      "tokens_seen": 1883832320
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.66875300914781e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.4760994159799687,
      "tokens_seen": 1883963392
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.667950569731985e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.4760791820345,
      "tokens_seen": 1884094464
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.667148130316162e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.476058949890714,
      "tokens_seen": 1884225536
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.666345690900337e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.4760387195483258,
      "tokens_seen": 1884356608
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.665543251484514e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.47601849100705,
      "tokens_seen": 1884487680
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.664740812068689e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.4759982642666003,
      "tokens_seen": 1884618752
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.663938372652866e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.475978039326691,
      "tokens_seen": 1884749824
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.663135933237041e-05,
      "loss": 1.3563,
      "theoretical_loss": 3.4759578161870373,
      "tokens_seen": 1884880896
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.662333493821216e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.4759375948473537,
      "tokens_seen": 1885011968
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.661531054405393e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.475917375307354,
      "tokens_seen": 1885143040
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.660728614989568e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.475897157566754,
      "tokens_seen": 1885274112
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.659926175573745e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.475876941625268,
      "tokens_seen": 1885405184
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.65912373615792e-05,
      "loss": 1.3561,
      "theoretical_loss": 3.4758567274826104,
      "tokens_seen": 1885536256
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.658321296742097e-05,
      "loss": 1.3677,
      "theoretical_loss": 3.4758365151384973,
      "tokens_seen": 1885667328
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.657518857326272e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.4758163045926427,
      "tokens_seen": 1885798400
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.9930122494697571,
      "objective/train/avg_token_score": 0.006987700238823891,
      "objective/train/docs_used": 687569,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.30894935131073,
      "objective/train/original_loss": 1.319667100906372,
      "objective/train/theoretical_loss": 3.475796095844762,
      "objective/train/tokens_used": 256388576,
      "objective/train/unlikelihood_loss": 0.02034331113100052,
      "theoretical_loss": 3.475796095844762,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.656716417910447e-05,
      "loss": 1.3678,
      "theoretical_loss": 3.475796095844762,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.655913978494624e-05,
      "loss": 1.3851,
      "theoretical_loss": 3.47577588889457,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.6551115390788e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.475755683741782,
      "tokens_seen": 1886191616
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.654309099662976e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.4757354803861134,
      "tokens_seen": 1886322688
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.653506660247151e-05,
      "loss": 1.4133,
      "theoretical_loss": 3.475715278827279,
      "tokens_seen": 1886453760
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.652704220831328e-05,
      "loss": 1.4306,
      "theoretical_loss": 3.4756950790649954,
      "tokens_seen": 1886584832
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.651901781415503e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.4756748810989766,
      "tokens_seen": 1886715904
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.65109934199968e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.475654684928939,
      "tokens_seen": 1886846976
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.650296902583855e-05,
      "loss": 1.3356,
      "theoretical_loss": 3.4756344905545977,
      "tokens_seen": 1886978048
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.649494463168031e-05,
      "loss": 1.3789,
      "theoretical_loss": 3.4756142979756683,
      "tokens_seen": 1887109120
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.648692023752207e-05,
      "loss": 1.3541,
      "theoretical_loss": 3.475594107191867,
      "tokens_seen": 1887240192
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.647889584336383e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.475573918202909,
      "tokens_seen": 1887371264
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.64708714492056e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.475553731008511,
      "tokens_seen": 1887502336
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.646284705504735e-05,
      "loss": 1.3553,
      "theoretical_loss": 3.4755335456083873,
      "tokens_seen": 1887633408
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.645482266088911e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.4755133620022556,
      "tokens_seen": 1887764480
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.644679826673087e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.4754931801898308,
      "tokens_seen": 1887895552
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.643877387257262e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.4754730001708287,
      "tokens_seen": 1888026624
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.643074947841439e-05,
      "loss": 1.3562,
      "theoretical_loss": 3.475452821944967,
      "tokens_seen": 1888157696
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.642272508425614e-05,
      "loss": 1.241,
      "theoretical_loss": 3.47543264551196,
      "tokens_seen": 1888288768
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.64147006900979e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.4754124708715257,
      "tokens_seen": 1888419840
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.640667629593966e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.4753922980233796,
      "tokens_seen": 1888550912
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.639865190178143e-05,
      "loss": 1.3927,
      "theoretical_loss": 3.4753721269672373,
      "tokens_seen": 1888681984
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.639062750762318e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.4753519577028165,
      "tokens_seen": 1888813056
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.638260311346493e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.4753317902298337,
      "tokens_seen": 1888944128
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.63745787193067e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.475311624548005,
      "tokens_seen": 1889075200
    },
    {
      "epoch": 0.14,
      "learning_rate": 8.636655432514845e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.475291460657047,
      "tokens_seen": 1889206272
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.635852993099022e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.4752712985566765,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.635050553683197e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.4752511382466107,
      "tokens_seen": 1889468416
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.634248114267374e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.475230979726566,
      "tokens_seen": 1889599488
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.633445674851549e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.4752108229962593,
      "tokens_seen": 1889730560
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.632643235435724e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.475190668055408,
      "tokens_seen": 1889861632
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.631840796019901e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.475170514903729,
      "tokens_seen": 1889992704
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.631038356604076e-05,
      "loss": 1.3346,
      "theoretical_loss": 3.475150363540939,
      "tokens_seen": 1890123776
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.630235917188253e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.4751302139667555,
      "tokens_seen": 1890254848
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.629433477772428e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.4751100661808954,
      "tokens_seen": 1890385920
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.628631038356605e-05,
      "loss": 1.3982,
      "theoretical_loss": 3.475089920183077,
      "tokens_seen": 1890516992
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.62782859894078e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.4750697759730165,
      "tokens_seen": 1890648064
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.627026159524956e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.475049633550432,
      "tokens_seen": 1890779136
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.626223720109132e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.4750294929150405,
      "tokens_seen": 1890910208
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.625421280693308e-05,
      "loss": 1.3469,
      "theoretical_loss": 3.4750093540665605,
      "tokens_seen": 1891041280
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.624618841277484e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.4749892170047083,
      "tokens_seen": 1891172352
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.62381640186166e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.474969081729203,
      "tokens_seen": 1891303424
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.623013962445836e-05,
      "loss": 1.3435,
      "theoretical_loss": 3.4749489482397613,
      "tokens_seen": 1891434496
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.622211523030012e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.474928816536101,
      "tokens_seen": 1891565568
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.621409083614188e-05,
      "loss": 1.2259,
      "theoretical_loss": 3.4749086866179413,
      "tokens_seen": 1891696640
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.620606644198364e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.4748885584849987,
      "tokens_seen": 1891827712
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.619804204782539e-05,
      "loss": 1.3087,
      "theoretical_loss": 3.474868432136991,
      "tokens_seen": 1891958784
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.619001765366716e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.474848307573638,
      "tokens_seen": 1892089856
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.618199325950891e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.474828184794656,
      "tokens_seen": 1892220928
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.617396886535068e-05,
      "loss": 1.3349,
      "theoretical_loss": 3.474808063799764,
      "tokens_seen": 1892352000
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.9939754605293274,
      "objective/train/avg_token_score": 0.006024450995028019,
      "objective/train/docs_used": 690244,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.392919659614563,
      "objective/train/original_loss": 1.4019582271575928,
      "objective/train/theoretical_loss": 3.474787944588681,
      "objective/train/tokens_used": 262942176,
      "objective/train/unlikelihood_loss": 0.015826327726244926,
      "theoretical_loss": 3.474787944588681,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.616594447119243e-05,
      "loss": 1.348,
      "theoretical_loss": 3.474787944588681,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.61579200770342e-05,
      "loss": 1.392,
      "theoretical_loss": 3.4747678271611244,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.614989568287595e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.4747477115168124,
      "tokens_seen": 1892745216
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.61418712887177e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.474727597655464,
      "tokens_seen": 1892876288
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.613384689455947e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.4747074855767974,
      "tokens_seen": 1893007360
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.612582250040122e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.4746873752805314,
      "tokens_seen": 1893138432
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.611779810624299e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.474667266766385,
      "tokens_seen": 1893269504
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.610977371208474e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.4746471600340763,
      "tokens_seen": 1893400576
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.610174931792651e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.474627055083324,
      "tokens_seen": 1893531648
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.609372492376826e-05,
      "loss": 1.3701,
      "theoretical_loss": 3.474606951913848,
      "tokens_seen": 1893662720
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.608570052961001e-05,
      "loss": 1.3815,
      "theoretical_loss": 3.4745868505253656,
      "tokens_seen": 1893793792
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.607767613545178e-05,
      "loss": 1.295,
      "theoretical_loss": 3.474566750917597,
      "tokens_seen": 1893924864
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.606965174129353e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.4745466530902607,
      "tokens_seen": 1894055936
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.60616273471353e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.4745265570430757,
      "tokens_seen": 1894187008
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.605360295297705e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.4745064627757616,
      "tokens_seen": 1894318080
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.604557855881882e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.4744863702880378,
      "tokens_seen": 1894449152
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.603755416466057e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.474466279579623,
      "tokens_seen": 1894580224
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.602952977050233e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.4744461906502364,
      "tokens_seen": 1894711296
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.60215053763441e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.474426103499598,
      "tokens_seen": 1894842368
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.601348098218585e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.4744060181274268,
      "tokens_seen": 1894973440
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.600545658802761e-05,
      "loss": 1.3494,
      "theoretical_loss": 3.474385934533443,
      "tokens_seen": 1895104512
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.599743219386937e-05,
      "loss": 1.3536,
      "theoretical_loss": 3.4743658527173658,
      "tokens_seen": 1895235584
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.598940779971113e-05,
      "loss": 1.32,
      "theoretical_loss": 3.4743457726789146,
      "tokens_seen": 1895366656
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.598138340555289e-05,
      "loss": 1.3706,
      "theoretical_loss": 3.4743256944178094,
      "tokens_seen": 1895497728
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.597335901139464e-05,
      "loss": 1.3635,
      "theoretical_loss": 3.47430561793377,
      "tokens_seen": 1895628800
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.59653346172364e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.4742855432265163,
      "tokens_seen": 1895759872
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.595731022307816e-05,
      "loss": 1.3422,
      "theoretical_loss": 3.4742654702957685,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.594928582891993e-05,
      "loss": 1.31,
      "theoretical_loss": 3.474245399141246,
      "tokens_seen": 1896022016
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.594126143476168e-05,
      "loss": 1.3426,
      "theoretical_loss": 3.474225329762669,
      "tokens_seen": 1896153088
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.593323704060345e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.4742052621597574,
      "tokens_seen": 1896284160
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.59252126464452e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.4741851963322326,
      "tokens_seen": 1896415232
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.591718825228697e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.474165132279813,
      "tokens_seen": 1896546304
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.590916385812872e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.47414507000222,
      "tokens_seen": 1896677376
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.590113946397047e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.4741250094991747,
      "tokens_seen": 1896808448
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.589311506981224e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.474104950770396,
      "tokens_seen": 1896939520
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.588509067565399e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.474084893815605,
      "tokens_seen": 1897070592
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.587706628149576e-05,
      "loss": 1.327,
      "theoretical_loss": 3.474064838634522,
      "tokens_seen": 1897201664
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.586904188733751e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.4740447852268685,
      "tokens_seen": 1897332736
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.586101749317928e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.4740247335923646,
      "tokens_seen": 1897463808
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.585299309902103e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.4740046837307306,
      "tokens_seen": 1897594880
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.584496870486278e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.4739846356416875,
      "tokens_seen": 1897725952
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.583694431070455e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.4739645893249573,
      "tokens_seen": 1897857024
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.58289199165463e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.4739445447802595,
      "tokens_seen": 1897988096
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.582089552238807e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.473924502007316,
      "tokens_seen": 1898119168
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.581287112822982e-05,
      "loss": 1.3475,
      "theoretical_loss": 3.473904461005847,
      "tokens_seen": 1898250240
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.580484673407159e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.4738844217755744,
      "tokens_seen": 1898381312
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.579682233991334e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.4738643843162196,
      "tokens_seen": 1898512384
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.57887979457551e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.473844348627503,
      "tokens_seen": 1898643456
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.578077355159686e-05,
      "loss": 1.3161,
      "theoretical_loss": 3.4738243147091463,
      "tokens_seen": 1898774528
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.577274915743862e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.473804282560871,
      "tokens_seen": 1898905600
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.991893470287323,
      "objective/train/avg_token_score": 0.008106522262096405,
      "objective/train/docs_used": 692658,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3104873895645142,
      "objective/train/original_loss": 1.3214442729949951,
      "objective/train/theoretical_loss": 3.4737842521823983,
      "objective/train/tokens_used": 269495776,
      "objective/train/unlikelihood_loss": 0.0227174274623394,
      "theoretical_loss": 3.4737842521823983,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.576472476328038e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.4737842521823983,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.575670036912214e-05,
      "loss": 1.3632,
      "theoretical_loss": 3.47376422357345,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.57486759749639e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.4737441967337475,
      "tokens_seen": 1899298816
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.574065158080566e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.4737241716630125,
      "tokens_seen": 1899429888
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.573262718664741e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.4737041483609663,
      "tokens_seen": 1899560960
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.572460279248918e-05,
      "loss": 1.2446,
      "theoretical_loss": 3.4736841268273313,
      "tokens_seen": 1899692032
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.571657839833093e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.473664107061829,
      "tokens_seen": 1899823104
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.57085540041727e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.473644089064181,
      "tokens_seen": 1899954176
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.570052961001445e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.47362407283411,
      "tokens_seen": 1900085248
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.569250521585622e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.473604058371337,
      "tokens_seen": 1900216320
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.568448082169797e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.473584045675585,
      "tokens_seen": 1900347392
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.567645642753972e-05,
      "loss": 1.3084,
      "theoretical_loss": 3.473564034746576,
      "tokens_seen": 1900478464
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.566843203338149e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.4735440255840313,
      "tokens_seen": 1900609536
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.566040763922324e-05,
      "loss": 1.3709,
      "theoretical_loss": 3.4735240181876743,
      "tokens_seen": 1900740608
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.565238324506501e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.4735040125572265,
      "tokens_seen": 1900871680
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.564435885090676e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.473484008692411,
      "tokens_seen": 1901002752
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.563633445674853e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4734640065929496,
      "tokens_seen": 1901133824
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.562831006259028e-05,
      "loss": 1.3639,
      "theoretical_loss": 3.473444006258565,
      "tokens_seen": 1901264896
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.562028566843204e-05,
      "loss": 1.3475,
      "theoretical_loss": 3.47342400768898,
      "tokens_seen": 1901395968
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.56122612742738e-05,
      "loss": 1.362,
      "theoretical_loss": 3.4734040108839173,
      "tokens_seen": 1901527040
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.560423688011555e-05,
      "loss": 1.3496,
      "theoretical_loss": 3.4733840158430995,
      "tokens_seen": 1901658112
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.559621248595732e-05,
      "loss": 1.3183,
      "theoretical_loss": 3.4733640225662485,
      "tokens_seen": 1901789184
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.558818809179907e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.473344031053089,
      "tokens_seen": 1901920256
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.558016369764084e-05,
      "loss": 1.3859,
      "theoretical_loss": 3.473324041303342,
      "tokens_seen": 1902051328
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.55721393034826e-05,
      "loss": 1.3339,
      "theoretical_loss": 3.473304053316731,
      "tokens_seen": 1902182400
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.556411490932436e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.4732840670929797,
      "tokens_seen": 1902313472
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.555609051516611e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.4732640826318106,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.554806612100787e-05,
      "loss": 1.3371,
      "theoretical_loss": 3.473244099932947,
      "tokens_seen": 1902575616
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.554004172684963e-05,
      "loss": 1.1958,
      "theoretical_loss": 3.4732241189961117,
      "tokens_seen": 1902706688
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.553201733269139e-05,
      "loss": 1.3791,
      "theoretical_loss": 3.473204139821029,
      "tokens_seen": 1902837760
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.552399293853315e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.473184162407421,
      "tokens_seen": 1902968832
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.551596854437491e-05,
      "loss": 1.238,
      "theoretical_loss": 3.473164186755012,
      "tokens_seen": 1903099904
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.550794415021667e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.4731442128635246,
      "tokens_seen": 1903230976
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.549991975605843e-05,
      "loss": 1.303,
      "theoretical_loss": 3.473124240732683,
      "tokens_seen": 1903362048
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.549189536190018e-05,
      "loss": 1.3863,
      "theoretical_loss": 3.473104270362211,
      "tokens_seen": 1903493120
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.548387096774195e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.473084301751831,
      "tokens_seen": 1903624192
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.54758465735837e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.4730643349012684,
      "tokens_seen": 1903755264
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.546782217942547e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.4730443698102453,
      "tokens_seen": 1903886336
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.545979778526722e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.473024406478487,
      "tokens_seen": 1904017408
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.545177339110899e-05,
      "loss": 1.3443,
      "theoretical_loss": 3.4730044449057162,
      "tokens_seen": 1904148480
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.544374899695074e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.472984485091657,
      "tokens_seen": 1904279552
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.543572460279249e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.4729645270360345,
      "tokens_seen": 1904410624
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.542770020863426e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.472944570738571,
      "tokens_seen": 1904541696
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.541967581447601e-05,
      "loss": 1.216,
      "theoretical_loss": 3.472924616198993,
      "tokens_seen": 1904672768
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.541165142031778e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.4729046634170224,
      "tokens_seen": 1904803840
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.540362702615953e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.4728847123923847,
      "tokens_seen": 1904934912
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.53956026320013e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.4728647631248037,
      "tokens_seen": 1905065984
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.538757823784305e-05,
      "loss": 1.3757,
      "theoretical_loss": 3.4728448156140046,
      "tokens_seen": 1905197056
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.53795538436848e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.4728248698597106,
      "tokens_seen": 1905328128
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.537152944952657e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.472804925861648,
      "tokens_seen": 1905459200
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.9944093227386475,
      "objective/train/avg_token_score": 0.005590598098933697,
      "objective/train/docs_used": 694687,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2861883640289307,
      "objective/train/original_loss": 1.293397307395935,
      "objective/train/theoretical_loss": 3.472784983619539,
      "objective/train/tokens_used": 276049376,
      "objective/train/unlikelihood_loss": 0.015973342582583427,
      "theoretical_loss": 3.472784983619539,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.536350505536833e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.472784983619539,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.15,
      "learning_rate": 8.535548066121009e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.47276504313311,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.534745626705184e-05,
      "loss": 1.347,
      "theoretical_loss": 3.472745104402085,
      "tokens_seen": 1905852416
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.533943187289361e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.472725167426189,
      "tokens_seen": 1905983488
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.533140747873536e-05,
      "loss": 1.3852,
      "theoretical_loss": 3.472705232205147,
      "tokens_seen": 1906114560
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.532338308457712e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.472685298738684,
      "tokens_seen": 1906245632
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.531535869041888e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.472665367026524,
      "tokens_seen": 1906376704
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.530733429626064e-05,
      "loss": 1.287,
      "theoretical_loss": 3.4726454370683926,
      "tokens_seen": 1906507776
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.52993099021024e-05,
      "loss": 1.3163,
      "theoretical_loss": 3.4726255088640148,
      "tokens_seen": 1906638848
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.529128550794416e-05,
      "loss": 1.3962,
      "theoretical_loss": 3.472605582413116,
      "tokens_seen": 1906769920
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.528326111378591e-05,
      "loss": 1.3293,
      "theoretical_loss": 3.4725856577154213,
      "tokens_seen": 1906900992
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.527523671962768e-05,
      "loss": 1.3958,
      "theoretical_loss": 3.4725657347706558,
      "tokens_seen": 1907032064
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.526721232546943e-05,
      "loss": 1.335,
      "theoretical_loss": 3.4725458135785447,
      "tokens_seen": 1907163136
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.52591879313112e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.4725258941388137,
      "tokens_seen": 1907294208
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.525116353715295e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.472505976451188,
      "tokens_seen": 1907425280
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.524313914299472e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.4724860605153935,
      "tokens_seen": 1907556352
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.523511474883647e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.472466146331155,
      "tokens_seen": 1907687424
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.522709035467822e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.4724462338981983,
      "tokens_seen": 1907818496
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.521906596051999e-05,
      "loss": 1.3646,
      "theoretical_loss": 3.47242632321625,
      "tokens_seen": 1907949568
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.521104156636174e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.472406414285035,
      "tokens_seen": 1908080640
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.520301717220351e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.4723865071042797,
      "tokens_seen": 1908211712
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.519499277804526e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.472366601673709,
      "tokens_seen": 1908342784
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.518696838388702e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.47234669799305,
      "tokens_seen": 1908473856
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.517894398972878e-05,
      "loss": 1.4014,
      "theoretical_loss": 3.4723267960620277,
      "tokens_seen": 1908604928
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.517091959557054e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.4723068958803687,
      "tokens_seen": 1908736000
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.51628952014123e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.472286997447799,
      "tokens_seen": 1908867072
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.515487080725406e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.4722671007640447,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.514684641309582e-05,
      "loss": 1.3538,
      "theoretical_loss": 3.4722472058288325,
      "tokens_seen": 1909129216
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.513882201893758e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.472227312641888,
      "tokens_seen": 1909260288
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.513079762477933e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.4722074212029383,
      "tokens_seen": 1909391360
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.51227732306211e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.4721875315117092,
      "tokens_seen": 1909522432
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.511474883646285e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.472167643567927,
      "tokens_seen": 1909653504
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.510672444230461e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.472147757371319,
      "tokens_seen": 1909784576
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.509870004814637e-05,
      "loss": 1.2041,
      "theoretical_loss": 3.472127872921612,
      "tokens_seen": 1909915648
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.509067565398812e-05,
      "loss": 1.3435,
      "theoretical_loss": 3.4721079902185314,
      "tokens_seen": 1910046720
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.508265125982989e-05,
      "loss": 1.3849,
      "theoretical_loss": 3.4720881092618052,
      "tokens_seen": 1910177792
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.507462686567164e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.4720682300511596,
      "tokens_seen": 1910308864
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.506660247151341e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.4720483525863215,
      "tokens_seen": 1910439936
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.505857807735516e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.4720284768670178,
      "tokens_seen": 1910571008
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.505055368319693e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.472008602892976,
      "tokens_seen": 1910702080
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.504252928903868e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.471988730663922,
      "tokens_seen": 1910833152
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.503450489488043e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.471968860179584,
      "tokens_seen": 1910964224
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.50264805007222e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.471948991439689,
      "tokens_seen": 1911095296
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.501845610656395e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.471929124443964,
      "tokens_seen": 1911226368
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.501043171240572e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.471909259192136,
      "tokens_seen": 1911357440
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.500240731824747e-05,
      "loss": 1.305,
      "theoretical_loss": 3.4718893956839327,
      "tokens_seen": 1911488512
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.499438292408923e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.4718695339190813,
      "tokens_seen": 1911619584
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.498635852993099e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.4718496738973097,
      "tokens_seen": 1911750656
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.497833413577275e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.471829815618345,
      "tokens_seen": 1911881728
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.497030974161451e-05,
      "loss": 1.3687,
      "theoretical_loss": 3.471809959081915,
      "tokens_seen": 1912012800
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.9946847558021545,
      "objective/train/avg_token_score": 0.005315225105732679,
      "objective/train/docs_used": 697086,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3265072107315063,
      "objective/train/original_loss": 1.3332548141479492,
      "objective/train/theoretical_loss": 3.4717901042877477,
      "objective/train/tokens_used": 282602976,
      "objective/train/unlikelihood_loss": 0.01651288941502571,
      "theoretical_loss": 3.4717901042877477,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.496228534745627e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.4717901042877477,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.495426095329802e-05,
      "loss": 1.3584,
      "theoretical_loss": 3.47177025123557,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.494623655913979e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.47175039992511,
      "tokens_seen": 1912406016
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.493821216498154e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.471730550356096,
      "tokens_seen": 1912537088
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.49301877708233e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.471710702528256,
      "tokens_seen": 1912668160
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.492216337666506e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.4716908564413167,
      "tokens_seen": 1912799232
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.491413898250683e-05,
      "loss": 1.3,
      "theoretical_loss": 3.4716710120950074,
      "tokens_seen": 1912930304
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.490611458834858e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.4716511694890553,
      "tokens_seen": 1913061376
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.489809019419033e-05,
      "loss": 1.278,
      "theoretical_loss": 3.4716313286231895,
      "tokens_seen": 1913192448
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.48900658000321e-05,
      "loss": 1.3516,
      "theoretical_loss": 3.4716114894971373,
      "tokens_seen": 1913323520
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.488204140587385e-05,
      "loss": 1.283,
      "theoretical_loss": 3.471591652110627,
      "tokens_seen": 1913454592
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.487401701171562e-05,
      "loss": 1.3064,
      "theoretical_loss": 3.471571816463388,
      "tokens_seen": 1913585664
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.486599261755737e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.4715519825551477,
      "tokens_seen": 1913716736
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.485796822339914e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.471532150385635,
      "tokens_seen": 1913847808
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.484994382924089e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.471512319954578,
      "tokens_seen": 1913978880
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.484191943508264e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.4714924912617056,
      "tokens_seen": 1914109952
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.483389504092441e-05,
      "loss": 1.2172,
      "theoretical_loss": 3.4714726643067464,
      "tokens_seen": 1914241024
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.482587064676616e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.4714528390894293,
      "tokens_seen": 1914372096
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.481784625260793e-05,
      "loss": 1.3806,
      "theoretical_loss": 3.471433015609482,
      "tokens_seen": 1914503168
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.480982185844968e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.471413193866635,
      "tokens_seen": 1914634240
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.480179746429145e-05,
      "loss": 1.2164,
      "theoretical_loss": 3.4713933738606157,
      "tokens_seen": 1914765312
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.47937730701332e-05,
      "loss": 1.334,
      "theoretical_loss": 3.471373555591154,
      "tokens_seen": 1914896384
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.478574867597496e-05,
      "loss": 1.3478,
      "theoretical_loss": 3.471353739057978,
      "tokens_seen": 1915027456
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.477772428181672e-05,
      "loss": 1.3853,
      "theoretical_loss": 3.4713339242608177,
      "tokens_seen": 1915158528
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.476969988765848e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.471314111199402,
      "tokens_seen": 1915289600
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.476167549350024e-05,
      "loss": 1.318,
      "theoretical_loss": 3.4712942998734597,
      "tokens_seen": 1915420672
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.4753651099342e-05,
      "loss": 1.325,
      "theoretical_loss": 3.47127449028272,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.474562670518376e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.4712546824269124,
      "tokens_seen": 1915682816
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.473760231102552e-05,
      "loss": 1.3509,
      "theoretical_loss": 3.4712348763057665,
      "tokens_seen": 1915813888
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.472957791686727e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.471215071919012,
      "tokens_seen": 1915944960
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.472155352270904e-05,
      "loss": 1.2723,
      "theoretical_loss": 3.4711952692663774,
      "tokens_seen": 1916076032
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.471352912855079e-05,
      "loss": 1.3626,
      "theoretical_loss": 3.4711754683475933,
      "tokens_seen": 1916207104
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.470550473439256e-05,
      "loss": 1.3499,
      "theoretical_loss": 3.471155669162388,
      "tokens_seen": 1916338176
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.469748034023431e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.4711358717104925,
      "tokens_seen": 1916469248
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.468945594607608e-05,
      "loss": 1.305,
      "theoretical_loss": 3.4711160759916355,
      "tokens_seen": 1916600320
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.468143155191783e-05,
      "loss": 1.308,
      "theoretical_loss": 3.4710962820055475,
      "tokens_seen": 1916731392
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.467340715775958e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.471076489751958,
      "tokens_seen": 1916862464
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.466538276360135e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.4710566992305973,
      "tokens_seen": 1916993536
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.46573583694431e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.471036910441195,
      "tokens_seen": 1917124608
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.464933397528487e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.471017123383481,
      "tokens_seen": 1917255680
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.464130958112662e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.470997338057186,
      "tokens_seen": 1917386752
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.463328518696839e-05,
      "loss": 1.346,
      "theoretical_loss": 3.4709775544620394,
      "tokens_seen": 1917517824
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.462526079281014e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.470957772597772,
      "tokens_seen": 1917648896
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.461723639865191e-05,
      "loss": 1.3519,
      "theoretical_loss": 3.4709379924641137,
      "tokens_seen": 1917779968
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.460921200449366e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.470918214060795,
      "tokens_seen": 1917911040
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.460118761033541e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.4708984373875467,
      "tokens_seen": 1918042112
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.459316321617718e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.4708786624440986,
      "tokens_seen": 1918173184
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.458513882201893e-05,
      "loss": 1.309,
      "theoretical_loss": 3.470858889230181,
      "tokens_seen": 1918304256
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.45771144278607e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.470839117745525,
      "tokens_seen": 1918435328
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.456909003370245e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.4708193479898615,
      "tokens_seen": 1918566400
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.9923063516616821,
      "objective/train/avg_token_score": 0.007693628314882517,
      "objective/train/docs_used": 699453,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3235951662063599,
      "objective/train/original_loss": 1.3338947296142578,
      "objective/train/theoretical_loss": 3.4707995799629208,
      "objective/train/tokens_used": 289156576,
      "objective/train/unlikelihood_loss": 0.022165171802043915,
      "theoretical_loss": 3.4707995799629208,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.456106563954422e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.4707995799629208,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.455304124538597e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.470779813664434,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.454501685122773e-05,
      "loss": 1.308,
      "theoretical_loss": 3.4707600490941313,
      "tokens_seen": 1918959616
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.45369924570695e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.4707402862517442,
      "tokens_seen": 1919090688
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.452896806291125e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.470720525137003,
      "tokens_seen": 1919221760
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.452094366875301e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.4707007657496387,
      "tokens_seen": 1919352832
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.451291927459477e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.4706810080893833,
      "tokens_seen": 1919483904
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.450489488043653e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.4706612521559674,
      "tokens_seen": 1919614976
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.449687048627829e-05,
      "loss": 1.2586,
      "theoretical_loss": 3.4706414979491216,
      "tokens_seen": 1919746048
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.448884609212004e-05,
      "loss": 1.298,
      "theoretical_loss": 3.470621745468578,
      "tokens_seen": 1919877120
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.44808216979618e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.470601994714068,
      "tokens_seen": 1920008192
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.447279730380356e-05,
      "loss": 1.3467,
      "theoretical_loss": 3.4705822456853213,
      "tokens_seen": 1920139264
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.446477290964533e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.4705624983820718,
      "tokens_seen": 1920270336
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.445674851548708e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.470542752804049,
      "tokens_seen": 1920401408
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.444872412132885e-05,
      "loss": 1.1788,
      "theoretical_loss": 3.470523008950985,
      "tokens_seen": 1920532480
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.44406997271706e-05,
      "loss": 1.327,
      "theoretical_loss": 3.470503266822612,
      "tokens_seen": 1920663552
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.443267533301235e-05,
      "loss": 1.3636,
      "theoretical_loss": 3.470483526418661,
      "tokens_seen": 1920794624
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.442465093885412e-05,
      "loss": 1.255,
      "theoretical_loss": 3.470463787738864,
      "tokens_seen": 1920925696
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.441662654469587e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.4704440507829526,
      "tokens_seen": 1921056768
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.440860215053764e-05,
      "loss": 1.316,
      "theoretical_loss": 3.4704243155506584,
      "tokens_seen": 1921187840
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.440057775637939e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.4704045820417138,
      "tokens_seen": 1921318912
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.439255336222116e-05,
      "loss": 1.327,
      "theoretical_loss": 3.4703848502558508,
      "tokens_seen": 1921449984
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.438452896806291e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.4703651201928016,
      "tokens_seen": 1921581056
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.437650457390466e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.4703453918522973,
      "tokens_seen": 1921712128
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.436848017974643e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.470325665234071,
      "tokens_seen": 1921843200
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.436045578558818e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.4703059403378544,
      "tokens_seen": 1921974272
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.435243139142995e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.4702862171633795,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.43444069972717e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.47026649571038,
      "tokens_seen": 1922236416
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.433638260311347e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.470246775978586,
      "tokens_seen": 1922367488
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.432835820895522e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.4702270579677323,
      "tokens_seen": 1922498560
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.432033381479699e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.4702073416775496,
      "tokens_seen": 1922629632
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.431230942063874e-05,
      "loss": 1.3507,
      "theoretical_loss": 3.470187627107772,
      "tokens_seen": 1922760704
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.43042850264805e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.4701679142581305,
      "tokens_seen": 1922891776
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.429626063232226e-05,
      "loss": 1.1826,
      "theoretical_loss": 3.4701482031283595,
      "tokens_seen": 1923022848
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.428823623816402e-05,
      "loss": 1.414,
      "theoretical_loss": 3.4701284937181898,
      "tokens_seen": 1923153920
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.428021184400578e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.470108786027356,
      "tokens_seen": 1923284992
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.427218744984754e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.4700890800555895,
      "tokens_seen": 1923416064
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.42641630556893e-05,
      "loss": 1.207,
      "theoretical_loss": 3.4700693758026246,
      "tokens_seen": 1923547136
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.425613866153106e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.4700496732681936,
      "tokens_seen": 1923678208
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.424811426737281e-05,
      "loss": 1.2275,
      "theoretical_loss": 3.470029972452029,
      "tokens_seen": 1923809280
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.424008987321458e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.4700102733538642,
      "tokens_seen": 1923940352
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.423206547905633e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.4699905759734335,
      "tokens_seen": 1924071424
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.42240410848981e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.4699708803104685,
      "tokens_seen": 1924202496
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.421601669073985e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.4699511863647032,
      "tokens_seen": 1924333568
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.420799229658162e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.4699314941358708,
      "tokens_seen": 1924464640
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.419996790242337e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.469911803623705,
      "tokens_seen": 1924595712
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.419194350826512e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.4698921148279394,
      "tokens_seen": 1924726784
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.418391911410689e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.4698724277483066,
      "tokens_seen": 1924857856
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.417589471994864e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.469852742384541,
      "tokens_seen": 1924988928
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.416787032579041e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.469833058736376,
      "tokens_seen": 1925120000
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.9933785796165466,
      "objective/train/avg_token_score": 0.006621368229389191,
      "objective/train/docs_used": 701973,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3796347379684448,
      "objective/train/original_loss": 1.388584017753601,
      "objective/train/theoretical_loss": 3.469813376803545,
      "objective/train/tokens_used": 295710176,
      "objective/train/unlikelihood_loss": 0.01968778856098652,
      "theoretical_loss": 3.469813376803545,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.415984593163216e-05,
      "loss": 1.3722,
      "theoretical_loss": 3.469813376803545,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.415182153747393e-05,
      "loss": 1.2294,
      "theoretical_loss": 3.469793696585782,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.414379714331568e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.469774018082821,
      "tokens_seen": 1925513216
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.413577274915743e-05,
      "loss": 1.3358,
      "theoretical_loss": 3.469754341294396,
      "tokens_seen": 1925644288
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.41277483549992e-05,
      "loss": 1.387,
      "theoretical_loss": 3.46973466622024,
      "tokens_seen": 1925775360
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.411972396084095e-05,
      "loss": 1.3875,
      "theoretical_loss": 3.469714992860088,
      "tokens_seen": 1925906432
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.411169956668272e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.4696953212136736,
      "tokens_seen": 1926037504
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.410367517252447e-05,
      "loss": 1.311,
      "theoretical_loss": 3.469675651280731,
      "tokens_seen": 1926168576
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.409565077836624e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.469655983060994,
      "tokens_seen": 1926299648
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.4087626384208e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.4696363165541975,
      "tokens_seen": 1926430720
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.407960199004975e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.4696166517600755,
      "tokens_seen": 1926561792
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.407157759589151e-05,
      "loss": 1.3752,
      "theoretical_loss": 3.469596988678362,
      "tokens_seen": 1926692864
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.406355320173327e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.4695773273087918,
      "tokens_seen": 1926823936
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.405552880757503e-05,
      "loss": 1.323,
      "theoretical_loss": 3.469557667651099,
      "tokens_seen": 1926955008
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.404750441341679e-05,
      "loss": 1.3521,
      "theoretical_loss": 3.469538009705019,
      "tokens_seen": 1927086080
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.403948001925855e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.4695183534702854,
      "tokens_seen": 1927217152
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.403145562510031e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.4694986989466328,
      "tokens_seen": 1927348224
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.402343123094206e-05,
      "loss": 1.285,
      "theoretical_loss": 3.469479046133797,
      "tokens_seen": 1927479296
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.401540683678383e-05,
      "loss": 1.3674,
      "theoretical_loss": 3.4694593950315116,
      "tokens_seen": 1927610368
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.400738244262558e-05,
      "loss": 1.345,
      "theoretical_loss": 3.4694397456395123,
      "tokens_seen": 1927741440
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.399935804846735e-05,
      "loss": 1.3847,
      "theoretical_loss": 3.469420097957533,
      "tokens_seen": 1927872512
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.39913336543091e-05,
      "loss": 1.2197,
      "theoretical_loss": 3.46940045198531,
      "tokens_seen": 1928003584
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.398330926015087e-05,
      "loss": 1.3386,
      "theoretical_loss": 3.469380807722577,
      "tokens_seen": 1928134656
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.397528486599262e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.4693611651690692,
      "tokens_seen": 1928265728
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.396726047183439e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.4693415243245225,
      "tokens_seen": 1928396800
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.395923607767614e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.4693218851886716,
      "tokens_seen": 1928527872
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.395121168351789e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.4693022477612514,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.394318728935966e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.469282612041998,
      "tokens_seen": 1928790016
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.393516289520141e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.4692629780306463,
      "tokens_seen": 1928921088
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.392713850104318e-05,
      "loss": 1.3659,
      "theoretical_loss": 3.469243345726931,
      "tokens_seen": 1929052160
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.391911410688493e-05,
      "loss": 1.256,
      "theoretical_loss": 3.4692237151305894,
      "tokens_seen": 1929183232
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.39110897127267e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.469204086241355,
      "tokens_seen": 1929314304
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.390306531856845e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.469184459058965,
      "tokens_seen": 1929445376
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.38950409244102e-05,
      "loss": 1.409,
      "theoretical_loss": 3.4691648335831538,
      "tokens_seen": 1929576448
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.388701653025197e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.469145209813658,
      "tokens_seen": 1929707520
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.387899213609372e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.469125587750213,
      "tokens_seen": 1929838592
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.387096774193549e-05,
      "loss": 1.3662,
      "theoretical_loss": 3.4691059673925544,
      "tokens_seen": 1929969664
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.386294334777724e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.4690863487404187,
      "tokens_seen": 1930100736
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.385491895361901e-05,
      "loss": 1.3467,
      "theoretical_loss": 3.469066731793541,
      "tokens_seen": 1930231808
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.384689455946076e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.469047116551658,
      "tokens_seen": 1930362880
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.383887016530252e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.4690275030145052,
      "tokens_seen": 1930493952
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.383084577114428e-05,
      "loss": 1.3677,
      "theoretical_loss": 3.4690078911818194,
      "tokens_seen": 1930625024
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.382282137698604e-05,
      "loss": 1.329,
      "theoretical_loss": 3.4689882810533366,
      "tokens_seen": 1930756096
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.38147969828278e-05,
      "loss": 1.3657,
      "theoretical_loss": 3.4689686726287925,
      "tokens_seen": 1930887168
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.380677258866956e-05,
      "loss": 1.3482,
      "theoretical_loss": 3.468949065907924,
      "tokens_seen": 1931018240
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.379874819451132e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.4689294608904664,
      "tokens_seen": 1931149312
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.379072380035308e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.4689098575761577,
      "tokens_seen": 1931280384
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.378269940619483e-05,
      "loss": 1.315,
      "theoretical_loss": 3.468890255964733,
      "tokens_seen": 1931411456
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.37746750120366e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.46887065605593,
      "tokens_seen": 1931542528
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.376665061787835e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.468851057849484,
      "tokens_seen": 1931673600
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.9932452440261841,
      "objective/train/avg_token_score": 0.006754749454557896,
      "objective/train/docs_used": 704289,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4284248352050781,
      "objective/train/original_loss": 1.4395936727523804,
      "objective/train/theoretical_loss": 3.4688314613451325,
      "objective/train/tokens_used": 302263776,
      "objective/train/unlikelihood_loss": 0.01620173081755638,
      "theoretical_loss": 3.4688314613451325,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.375862622372012e-05,
      "loss": 1.4136,
      "theoretical_loss": 3.4688314613451325,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.375060182956187e-05,
      "loss": 1.3548,
      "theoretical_loss": 3.4688118665426124,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.374257743540364e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.46879227344166,
      "tokens_seen": 1932066816
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.373455304124539e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.4687726820420117,
      "tokens_seen": 1932197888
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.372652864708714e-05,
      "loss": 1.2052,
      "theoretical_loss": 3.4687530923434053,
      "tokens_seen": 1932328960
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.371850425292891e-05,
      "loss": 1.3607,
      "theoretical_loss": 3.468733504345577,
      "tokens_seen": 1932460032
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.371047985877066e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.4687139180482647,
      "tokens_seen": 1932591104
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.370245546461243e-05,
      "loss": 1.259,
      "theoretical_loss": 3.4686943334512046,
      "tokens_seen": 1932722176
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.369443107045418e-05,
      "loss": 1.3341,
      "theoretical_loss": 3.4686747505541335,
      "tokens_seen": 1932853248
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.368640667629595e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.46865516935679,
      "tokens_seen": 1932984320
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.36783822821377e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.46863558985891,
      "tokens_seen": 1933115392
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.367035788797947e-05,
      "loss": 1.3333,
      "theoretical_loss": 3.4686160120602314,
      "tokens_seen": 1933246464
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.366233349382122e-05,
      "loss": 1.3524,
      "theoretical_loss": 3.468596435960491,
      "tokens_seen": 1933377536
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.365430909966297e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.468576861559427,
      "tokens_seen": 1933508608
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.364628470550474e-05,
      "loss": 1.261,
      "theoretical_loss": 3.4685572888567764,
      "tokens_seen": 1933639680
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.36382603113465e-05,
      "loss": 1.329,
      "theoretical_loss": 3.468537717852277,
      "tokens_seen": 1933770752
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.363023591718826e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.468518148545666,
      "tokens_seen": 1933901824
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.362221152303001e-05,
      "loss": 1.3775,
      "theoretical_loss": 3.4684985809366813,
      "tokens_seen": 1934032896
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.361418712887178e-05,
      "loss": 1.3803,
      "theoretical_loss": 3.4684790150250606,
      "tokens_seen": 1934163968
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.360616273471353e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.468459450810541,
      "tokens_seen": 1934295040
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.359813834055529e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.4684398882928615,
      "tokens_seen": 1934426112
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.359011394639705e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.468420327471759,
      "tokens_seen": 1934557184
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.358208955223881e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.4684007683469718,
      "tokens_seen": 1934688256
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.357406515808057e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.4683812109182384,
      "tokens_seen": 1934819328
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.356604076392233e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.4683616551852956,
      "tokens_seen": 1934950400
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.35580163697641e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.4683421011478828,
      "tokens_seen": 1935081472
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.354999197560585e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.4683225488057374,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.35419675814476e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.4683029981585975,
      "tokens_seen": 1935343616
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.353394318728937e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4682834492062016,
      "tokens_seen": 1935474688
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.352591879313112e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.468263901948289,
      "tokens_seen": 1935605760
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.351789439897289e-05,
      "loss": 1.3409,
      "theoretical_loss": 3.468244356384596,
      "tokens_seen": 1935736832
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.350987000481464e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.468224812514863,
      "tokens_seen": 1935867904
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.35018456106564e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.468205270338828,
      "tokens_seen": 1935998976
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.349382121649816e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.468185729856229,
      "tokens_seen": 1936130048
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.348579682233991e-05,
      "loss": 1.3349,
      "theoretical_loss": 3.4681661910668047,
      "tokens_seen": 1936261120
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.347777242818168e-05,
      "loss": 1.2803,
      "theoretical_loss": 3.4681466539702948,
      "tokens_seen": 1936392192
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.346974803402343e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.4681271185664366,
      "tokens_seen": 1936523264
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.34617236398652e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.4681075848549696,
      "tokens_seen": 1936654336
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.345369924570695e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.4680880528356326,
      "tokens_seen": 1936785408
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.344567485154872e-05,
      "loss": 1.305,
      "theoretical_loss": 3.4680685225081653,
      "tokens_seen": 1936916480
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.343765045739047e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.468048993872306,
      "tokens_seen": 1937047552
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.342962606323223e-05,
      "loss": 1.3429,
      "theoretical_loss": 3.468029466927793,
      "tokens_seen": 1937178624
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.342160166907399e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.468009941674366,
      "tokens_seen": 1937309696
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.341357727491575e-05,
      "loss": 1.3236,
      "theoretical_loss": 3.4679904181117642,
      "tokens_seen": 1937440768
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.340555288075751e-05,
      "loss": 1.324,
      "theoretical_loss": 3.467970896239727,
      "tokens_seen": 1937571840
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.339752848659926e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.467951376057994,
      "tokens_seen": 1937702912
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.338950409244103e-05,
      "loss": 1.347,
      "theoretical_loss": 3.467931857566304,
      "tokens_seen": 1937833984
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.338147969828278e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.4679123407643955,
      "tokens_seen": 1937965056
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.337345530412455e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.46789282565201,
      "tokens_seen": 1938096128
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.33654309099663e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.467873312228885,
      "tokens_seen": 1938227200
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.9916789531707764,
      "objective/train/avg_token_score": 0.0083210663869977,
      "objective/train/docs_used": 706849,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2584171295166016,
      "objective/train/original_loss": 1.2713639736175537,
      "objective/train/theoretical_loss": 3.467853800494761,
      "objective/train/tokens_used": 308817376,
      "objective/train/unlikelihood_loss": 0.022031841799616814,
      "theoretical_loss": 3.467853800494761,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.335740651580806e-05,
      "loss": 1.229,
      "theoretical_loss": 3.467853800494761,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.334938212164982e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.4678342904493777,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.334135772749158e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.4678147820924745,
      "tokens_seen": 1938620416
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.333333333333334e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.467795275423791,
      "tokens_seen": 1938751488
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.33253089391751e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.4677757704430676,
      "tokens_seen": 1938882560
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.331728454501686e-05,
      "loss": 1.273,
      "theoretical_loss": 3.467756267150044,
      "tokens_seen": 1939013632
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.330926015085862e-05,
      "loss": 1.3325,
      "theoretical_loss": 3.467736765544459,
      "tokens_seen": 1939144704
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.330123575670037e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.467717265626054,
      "tokens_seen": 1939275776
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.329321136254214e-05,
      "loss": 1.2144,
      "theoretical_loss": 3.4676977673945677,
      "tokens_seen": 1939406848
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.328518696838389e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.467678270849741,
      "tokens_seen": 1939537920
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.327716257422566e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.467658775991315,
      "tokens_seen": 1939668992
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.326913818006741e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.4676392828190274,
      "tokens_seen": 1939800064
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.326111378590918e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.4676197913326208,
      "tokens_seen": 1939931136
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.325308939175093e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.467600301531834,
      "tokens_seen": 1940062208
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.324506499759268e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.467580813416408,
      "tokens_seen": 1940193280
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.323704060343445e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.4675613269860834,
      "tokens_seen": 1940324352
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.32290162092762e-05,
      "loss": 1.3413,
      "theoretical_loss": 3.4675418422405997,
      "tokens_seen": 1940455424
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.322099181511797e-05,
      "loss": 1.3798,
      "theoretical_loss": 3.4675223591796986,
      "tokens_seen": 1940586496
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.321296742095972e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.4675028778031205,
      "tokens_seen": 1940717568
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.320494302680149e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.4674833981106055,
      "tokens_seen": 1940848640
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.319691863264324e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.4674639201018946,
      "tokens_seen": 1940979712
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.3188894238485e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.4674444437767282,
      "tokens_seen": 1941110784
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.318086984432676e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.467424969134848,
      "tokens_seen": 1941241856
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.317284545016852e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.467405496175994,
      "tokens_seen": 1941372928
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.316482105601028e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.4673860248999073,
      "tokens_seen": 1941504000
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.315679666185203e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.4673665553063295,
      "tokens_seen": 1941635072
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.31487722676938e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.4673470873950007,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.314074787353555e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.467327621165663,
      "tokens_seen": 1941897216
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.313272347937731e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.4673081566180572,
      "tokens_seen": 1942028288
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.312469908521907e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.4672886937519243,
      "tokens_seen": 1942159360
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.311667469106083e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.4672692325670056,
      "tokens_seen": 1942290432
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.31086502969026e-05,
      "loss": 1.2,
      "theoretical_loss": 3.467249773063042,
      "tokens_seen": 1942421504
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.310062590274435e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.4672303152397763,
      "tokens_seen": 1942552576
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.309260150858611e-05,
      "loss": 1.3579,
      "theoretical_loss": 3.467210859096948,
      "tokens_seen": 1942683648
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.308457711442787e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.467191404634301,
      "tokens_seen": 1942814720
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.307655272026962e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.4671719518515745,
      "tokens_seen": 1942945792
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.306852832611139e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.4671525007485116,
      "tokens_seen": 1943076864
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.306050393195314e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.4671330513248533,
      "tokens_seen": 1943207936
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.305247953779491e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.4671136035803416,
      "tokens_seen": 1943339008
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.304445514363666e-05,
      "loss": 1.3007,
      "theoretical_loss": 3.4670941575147185,
      "tokens_seen": 1943470080
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.303643074947843e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.4670747131277255,
      "tokens_seen": 1943601152
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.302840635532018e-05,
      "loss": 1.3586,
      "theoretical_loss": 3.4670552704191047,
      "tokens_seen": 1943732224
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.302038196116195e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.467035829388598,
      "tokens_seen": 1943863296
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.30123575670037e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.4670163900359476,
      "tokens_seen": 1943994368
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.300433317284545e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.466996952360895,
      "tokens_seen": 1944125440
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.299630877868722e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.4669775163631833,
      "tokens_seen": 1944256512
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.298828438452897e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.466958082042553,
      "tokens_seen": 1944387584
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.298025999037074e-05,
      "loss": 1.288,
      "theoretical_loss": 3.466938649398749,
      "tokens_seen": 1944518656
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.297223559621249e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.466919218431511,
      "tokens_seen": 1944649728
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.296421120205426e-05,
      "loss": 1.307,
      "theoretical_loss": 3.466899789140583,
      "tokens_seen": 1944780800
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.9927396774291992,
      "objective/train/avg_token_score": 0.007260307669639587,
      "objective/train/docs_used": 709206,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3710139989852905,
      "objective/train/original_loss": 1.3811849355697632,
      "objective/train/theoretical_loss": 3.466880361525707,
      "objective/train/tokens_used": 315370976,
      "objective/train/unlikelihood_loss": 0.019459262490272522,
      "theoretical_loss": 3.466880361525707,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.295618680789601e-05,
      "loss": 1.3682,
      "theoretical_loss": 3.466880361525707,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.294816241373777e-05,
      "loss": 1.4073,
      "theoretical_loss": 3.4668609355866247,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.294013801957953e-05,
      "loss": 1.328,
      "theoretical_loss": 3.4668415113230795,
      "tokens_seen": 1945174016
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.293211362542129e-05,
      "loss": 1.3696,
      "theoretical_loss": 3.466822088734814,
      "tokens_seen": 1945305088
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.292408923126305e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.4668026678215704,
      "tokens_seen": 1945436160
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.29160648371048e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.466783248583092,
      "tokens_seen": 1945567232
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.290804044294657e-05,
      "loss": 1.3632,
      "theoretical_loss": 3.4667638310191204,
      "tokens_seen": 1945698304
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.290001604878832e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.4667444151294005,
      "tokens_seen": 1945829376
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.289199165463008e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.466725000913673,
      "tokens_seen": 1945960448
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.288396726047184e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.466705588371682,
      "tokens_seen": 1946091520
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.28759428663136e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.46668617750317,
      "tokens_seen": 1946222592
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.286791847215536e-05,
      "loss": 1.327,
      "theoretical_loss": 3.4666667683078805,
      "tokens_seen": 1946353664
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.285989407799712e-05,
      "loss": 1.3529,
      "theoretical_loss": 3.466647360785556,
      "tokens_seen": 1946484736
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.285186968383888e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.4666279549359404,
      "tokens_seen": 1946615808
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.284384528968064e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.4666085507587767,
      "tokens_seen": 1946746880
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.283582089552239e-05,
      "loss": 1.3498,
      "theoretical_loss": 3.466589148253808,
      "tokens_seen": 1946877952
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.282779650136416e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.4665697474207775,
      "tokens_seen": 1947009024
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.281977210720591e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.466550348259428,
      "tokens_seen": 1947140096
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.281174771304768e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.4665309507695046,
      "tokens_seen": 1947271168
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.280372331888943e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.46651155495075,
      "tokens_seen": 1947402240
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.27956989247312e-05,
      "loss": 1.3965,
      "theoretical_loss": 3.466492160802907,
      "tokens_seen": 1947533312
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.278767453057295e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.4664727683257204,
      "tokens_seen": 1947664384
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.27796501364147e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.466453377518933,
      "tokens_seen": 1947795456
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.277162574225647e-05,
      "loss": 1.3904,
      "theoretical_loss": 3.466433988382289,
      "tokens_seen": 1947926528
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.276360134809822e-05,
      "loss": 1.3609,
      "theoretical_loss": 3.466414600915532,
      "tokens_seen": 1948057600
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.275557695393999e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.466395215118406,
      "tokens_seen": 1948188672
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.274755255978174e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.4663758309906543,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.273952816562351e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.4663564485320215,
      "tokens_seen": 1948450816
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.273150377146526e-05,
      "loss": 1.355,
      "theoretical_loss": 3.466337067742252,
      "tokens_seen": 1948581888
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.272347937730703e-05,
      "loss": 1.3897,
      "theoretical_loss": 3.4663176886210887,
      "tokens_seen": 1948712960
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.271545498314878e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.4662983111682766,
      "tokens_seen": 1948844032
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.270743058899054e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.4662789353835595,
      "tokens_seen": 1948975104
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.26994061948323e-05,
      "loss": 1.3443,
      "theoretical_loss": 3.466259561266682,
      "tokens_seen": 1949106176
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.269138180067406e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.4662401888173875,
      "tokens_seen": 1949237248
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.268335740651582e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.4662208180354215,
      "tokens_seen": 1949368320
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.267533301235758e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.4662014489205277,
      "tokens_seen": 1949499392
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.266730861819934e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.466182081472451,
      "tokens_seen": 1949630464
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.26592842240411e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.4661627156909356,
      "tokens_seen": 1949761536
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.265125982988285e-05,
      "loss": 1.249,
      "theoretical_loss": 3.4661433515757256,
      "tokens_seen": 1949892608
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.264323543572461e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.4661239891265665,
      "tokens_seen": 1950023680
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.263521104156637e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.466104628343203,
      "tokens_seen": 1950154752
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.262718664740813e-05,
      "loss": 1.3836,
      "theoretical_loss": 3.466085269225379,
      "tokens_seen": 1950285824
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.261916225324989e-05,
      "loss": 1.3487,
      "theoretical_loss": 3.46606591177284,
      "tokens_seen": 1950416896
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.261113785909164e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.4660465559853306,
      "tokens_seen": 1950547968
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.260311346493341e-05,
      "loss": 1.3489,
      "theoretical_loss": 3.4660272018625955,
      "tokens_seen": 1950679040
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.259508907077516e-05,
      "loss": 1.3899,
      "theoretical_loss": 3.4660078494043804,
      "tokens_seen": 1950810112
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.258706467661693e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.4659884986104297,
      "tokens_seen": 1950941184
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.257904028245868e-05,
      "loss": 1.3632,
      "theoretical_loss": 3.4659691494804887,
      "tokens_seen": 1951072256
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.257101588830045e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.4659498020143023,
      "tokens_seen": 1951203328
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.25629914941422e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.4659304562116158,
      "tokens_seen": 1951334400
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.9904803037643433,
      "objective/train/avg_token_score": 0.009519672021269798,
      "objective/train/docs_used": 711575,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2500711679458618,
      "objective/train/original_loss": 1.260582685470581,
      "objective/train/theoretical_loss": 3.465911112072175,
      "objective/train/tokens_used": 321924576,
      "objective/train/unlikelihood_loss": 0.028411181643605232,
      "theoretical_loss": 3.465911112072175,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.255496709998395e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.465911112072175,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.254694270582572e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.465891769595724,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.253891831166747e-05,
      "loss": 1.328,
      "theoretical_loss": 3.465872428782009,
      "tokens_seen": 1951727616
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.253089391750924e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.4658530896307758,
      "tokens_seen": 1951858688
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.252286952335099e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.465833752141769,
      "tokens_seen": 1951989760
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.251484512919275e-05,
      "loss": 1.3876,
      "theoretical_loss": 3.465814416314734,
      "tokens_seen": 1952120832
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.250682073503451e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.465795082149418,
      "tokens_seen": 1952251904
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.249879634087627e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.4657757496455646,
      "tokens_seen": 1952382976
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.249077194671803e-05,
      "loss": 1.3339,
      "theoretical_loss": 3.4657564188029206,
      "tokens_seen": 1952514048
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.248274755255979e-05,
      "loss": 1.3087,
      "theoretical_loss": 3.465737089621232,
      "tokens_seen": 1952645120
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.247472315840154e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.465717762100244,
      "tokens_seen": 1952776192
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.24666987642433e-05,
      "loss": 1.3547,
      "theoretical_loss": 3.465698436239702,
      "tokens_seen": 1952907264
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.245867437008506e-05,
      "loss": 1.311,
      "theoretical_loss": 3.465679112039353,
      "tokens_seen": 1953038336
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.245064997592683e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.4656597894989427,
      "tokens_seen": 1953169408
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.244262558176858e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.4656404686182167,
      "tokens_seen": 1953300480
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.243460118761035e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.4656211493969216,
      "tokens_seen": 1953431552
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.24265767934521e-05,
      "loss": 1.3489,
      "theoretical_loss": 3.465601831834803,
      "tokens_seen": 1953562624
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.241855239929385e-05,
      "loss": 1.3339,
      "theoretical_loss": 3.465582515931608,
      "tokens_seen": 1953693696
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.241052800513562e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.4655632016870817,
      "tokens_seen": 1953824768
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.240250361097737e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.465543889100971,
      "tokens_seen": 1953955840
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.239447921681914e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.4655245781730226,
      "tokens_seen": 1954086912
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.238645482266089e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.465505268902982,
      "tokens_seen": 1954217984
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.237843042850264e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.4654859612905966,
      "tokens_seen": 1954349056
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.237040603434441e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.4654666553356126,
      "tokens_seen": 1954480128
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.236238164018616e-05,
      "loss": 1.3392,
      "theoretical_loss": 3.465447351037776,
      "tokens_seen": 1954611200
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.235435724602793e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.4654280483968347,
      "tokens_seen": 1954742272
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.234633285186968e-05,
      "loss": 1.4457,
      "theoretical_loss": 3.4654087474125346,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.233830845771145e-05,
      "loss": 1.3243,
      "theoretical_loss": 3.4653894480846223,
      "tokens_seen": 1955004416
    },
    {
      "epoch": 0.18,
      "learning_rate": 8.23302840635532e-05,
      "loss": 1.3884,
      "theoretical_loss": 3.465370150412845,
      "tokens_seen": 1955135488
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.232225966939496e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.4653508543969496,
      "tokens_seen": 1955266560
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.231423527523672e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.4653315600366827,
      "tokens_seen": 1955397632
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.230621088107848e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.4653122673317913,
      "tokens_seen": 1955528704
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.229818648692024e-05,
      "loss": 1.3242,
      "theoretical_loss": 3.465292976282023,
      "tokens_seen": 1955659776
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.2290162092762e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.465273686887124,
      "tokens_seen": 1955790848
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.228213769860375e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.4652543991468425,
      "tokens_seen": 1955921920
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.227411330444552e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.465235113060925,
      "tokens_seen": 1956052992
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.226608891028727e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.4652158286291184,
      "tokens_seen": 1956184064
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.225806451612904e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.465196545851171,
      "tokens_seen": 1956315136
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.225004012197079e-05,
      "loss": 1.3236,
      "theoretical_loss": 3.4651772647268295,
      "tokens_seen": 1956446208
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.224201572781256e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.4651579852558414,
      "tokens_seen": 1956577280
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.223399133365431e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.4651387074379545,
      "tokens_seen": 1956708352
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.222596693949606e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.4651194312729157,
      "tokens_seen": 1956839424
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.221794254533783e-05,
      "loss": 1.3556,
      "theoretical_loss": 3.4651001567604736,
      "tokens_seen": 1956970496
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.220991815117958e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.465080883900375,
      "tokens_seen": 1957101568
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.220189375702135e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.4650616126923675,
      "tokens_seen": 1957232640
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.21938693628631e-05,
      "loss": 1.3579,
      "theoretical_loss": 3.4650423431362,
      "tokens_seen": 1957363712
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.218584496870485e-05,
      "loss": 1.3713,
      "theoretical_loss": 3.4650230752316187,
      "tokens_seen": 1957494784
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.217782057454662e-05,
      "loss": 1.303,
      "theoretical_loss": 3.4650038089783726,
      "tokens_seen": 1957625856
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.216979618038837e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.4649845443762093,
      "tokens_seen": 1957756928
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.216177178623014e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.4649652814248775,
      "tokens_seen": 1957888000
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.9926722049713135,
      "objective/train/avg_token_score": 0.007327770348638296,
      "objective/train/docs_used": 713999,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3536126613616943,
      "objective/train/original_loss": 1.3644202947616577,
      "objective/train/theoretical_loss": 3.4649460201241236,
      "objective/train/tokens_used": 328478176,
      "objective/train/unlikelihood_loss": 0.01878575049340725,
      "theoretical_loss": 3.4649460201241236,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.21537473920719e-05,
      "loss": 1.3392,
      "theoretical_loss": 3.4649460201241236,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.214572299791366e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.4649267604736966,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.213769860375541e-05,
      "loss": 1.359,
      "theoretical_loss": 3.4649075024733453,
      "tokens_seen": 1958281216
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.212967420959717e-05,
      "loss": 1.3699,
      "theoretical_loss": 3.4648882461228174,
      "tokens_seen": 1958412288
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.212164981543893e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.4648689914218602,
      "tokens_seen": 1958543360
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.211362542128069e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.4648497383702237,
      "tokens_seen": 1958674432
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.210560102712245e-05,
      "loss": 1.3798,
      "theoretical_loss": 3.4648304869676547,
      "tokens_seen": 1958805504
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.209757663296421e-05,
      "loss": 1.4322,
      "theoretical_loss": 3.464811237213903,
      "tokens_seen": 1958936576
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.208955223880597e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.4647919891087158,
      "tokens_seen": 1959067648
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.208152784464773e-05,
      "loss": 1.3519,
      "theoretical_loss": 3.464772742651842,
      "tokens_seen": 1959198720
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.20735034504895e-05,
      "loss": 1.3776,
      "theoretical_loss": 3.4647534978430308,
      "tokens_seen": 1959329792
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.206547905633125e-05,
      "loss": 1.3447,
      "theoretical_loss": 3.464734254682031,
      "tokens_seen": 1959460864
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.2057454662173e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.4647150131685898,
      "tokens_seen": 1959591936
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.204943026801477e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.4646957733024575,
      "tokens_seen": 1959723008
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.204140587385652e-05,
      "loss": 1.3865,
      "theoretical_loss": 3.4646765350833824,
      "tokens_seen": 1959854080
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.203338147969829e-05,
      "loss": 1.309,
      "theoretical_loss": 3.4646572985111126,
      "tokens_seen": 1959985152
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.202535708554004e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.464638063585398,
      "tokens_seen": 1960116224
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.20173326913818e-05,
      "loss": 1.257,
      "theoretical_loss": 3.464618830305988,
      "tokens_seen": 1960247296
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.200930829722356e-05,
      "loss": 1.3272,
      "theoretical_loss": 3.4645995986726303,
      "tokens_seen": 1960378368
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.200128390306531e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.4645803686850747,
      "tokens_seen": 1960509440
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.199325950890708e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.4645611403430703,
      "tokens_seen": 1960640512
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.198523511474883e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.464541913646366,
      "tokens_seen": 1960771584
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.19772107205906e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.4645226885947116,
      "tokens_seen": 1960902656
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.196918632643235e-05,
      "loss": 1.3742,
      "theoretical_loss": 3.464503465187856,
      "tokens_seen": 1961033728
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.196116193227412e-05,
      "loss": 1.3323,
      "theoretical_loss": 3.464484243425548,
      "tokens_seen": 1961164800
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.195313753811587e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.464465023307538,
      "tokens_seen": 1961295872
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.194511314395762e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.4644458048335753,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.193708874979939e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.464426588003409,
      "tokens_seen": 1961558016
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.192906435564114e-05,
      "loss": 1.323,
      "theoretical_loss": 3.4644073728167895,
      "tokens_seen": 1961689088
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.192103996148291e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.464388159273465,
      "tokens_seen": 1961820160
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.191301556732466e-05,
      "loss": 1.3596,
      "theoretical_loss": 3.4643689473731865,
      "tokens_seen": 1961951232
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.190499117316643e-05,
      "loss": 1.3541,
      "theoretical_loss": 3.464349737115703,
      "tokens_seen": 1962082304
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.189696677900818e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.464330528500765,
      "tokens_seen": 1962213376
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.188894238484994e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.464311321528122,
      "tokens_seen": 1962344448
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.18809179906917e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.464292116197523,
      "tokens_seen": 1962475520
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.187289359653346e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.4642729125087195,
      "tokens_seen": 1962606592
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.186486920237522e-05,
      "loss": 1.2433,
      "theoretical_loss": 3.4642537104614606,
      "tokens_seen": 1962737664
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.185684480821698e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.4642345100554963,
      "tokens_seen": 1962868736
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.184882041405874e-05,
      "loss": 1.2927,
      "theoretical_loss": 3.464215311290577,
      "tokens_seen": 1962999808
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.18407960199005e-05,
      "loss": 1.3511,
      "theoretical_loss": 3.4641961141664535,
      "tokens_seen": 1963130880
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.183277162574225e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.464176918682875,
      "tokens_seen": 1963261952
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.182474723158402e-05,
      "loss": 1.3383,
      "theoretical_loss": 3.464157724839592,
      "tokens_seen": 1963393024
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.181672283742577e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.4641385326363556,
      "tokens_seen": 1963524096
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.180869844326754e-05,
      "loss": 1.3485,
      "theoretical_loss": 3.464119342072915,
      "tokens_seen": 1963655168
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.180067404910929e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.4641001531490216,
      "tokens_seen": 1963786240
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.179264965495106e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.464080965864426,
      "tokens_seen": 1963917312
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.178462526079281e-05,
      "loss": 1.3467,
      "theoretical_loss": 3.464061780218878,
      "tokens_seen": 1964048384
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.177660086663458e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.4640425962121286,
      "tokens_seen": 1964179456
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.176857647247633e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.464023413843928,
      "tokens_seen": 1964310528
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.176055207831808e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.464004233114028,
      "tokens_seen": 1964441600
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.9915092587471008,
      "objective/train/avg_token_score": 0.008490774780511856,
      "objective/train/docs_used": 716326,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.293127417564392,
      "objective/train/original_loss": 1.304544448852539,
      "objective/train/theoretical_loss": 3.463985054022179,
      "objective/train/tokens_used": 335031776,
      "objective/train/unlikelihood_loss": 0.025795821100473404,
      "theoretical_loss": 3.463985054022179,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.175252768415985e-05,
      "loss": 1.3287,
      "theoretical_loss": 3.463985054022179,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.17445032900016e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.463965876568131,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.173647889584337e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.463946700751636,
      "tokens_seen": 1964834816
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.172845450168512e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.463927526572444,
      "tokens_seen": 1964965888
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.172043010752689e-05,
      "loss": 1.3747,
      "theoretical_loss": 3.4639083540303064,
      "tokens_seen": 1965096960
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.171240571336864e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.4638891831249747,
      "tokens_seen": 1965228032
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.17043813192104e-05,
      "loss": 1.3,
      "theoretical_loss": 3.4638700138561993,
      "tokens_seen": 1965359104
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.169635692505216e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.463850846223732,
      "tokens_seen": 1965490176
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.168833253089391e-05,
      "loss": 1.2264,
      "theoretical_loss": 3.4638316802273232,
      "tokens_seen": 1965621248
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.168030813673568e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.4638125158667252,
      "tokens_seen": 1965752320
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.167228374257743e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.4637933531416882,
      "tokens_seen": 1965883392
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.16642593484192e-05,
      "loss": 1.316,
      "theoretical_loss": 3.4637741920519645,
      "tokens_seen": 1966014464
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.165623495426095e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.463755032597305,
      "tokens_seen": 1966145536
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.164821056010271e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.463735874777462,
      "tokens_seen": 1966276608
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.164018616594447e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.4637167185921856,
      "tokens_seen": 1966407680
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.163216177178623e-05,
      "loss": 1.3293,
      "theoretical_loss": 3.4636975640412286,
      "tokens_seen": 1966538752
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.1624137377628e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.4636784111243424,
      "tokens_seen": 1966669824
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.161611298346975e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.463659259841279,
      "tokens_seen": 1966800896
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.160808858931151e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.463640110191789,
      "tokens_seen": 1966931968
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.160006419515327e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.463620962175625,
      "tokens_seen": 1967063040
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.159203980099502e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.463601815792539,
      "tokens_seen": 1967194112
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.158401540683679e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.463582671042283,
      "tokens_seen": 1967325184
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.157599101267854e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.463563527924608,
      "tokens_seen": 1967456256
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.15679666185203e-05,
      "loss": 1.4287,
      "theoretical_loss": 3.463544386439268,
      "tokens_seen": 1967587328
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.155994222436206e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.4635252465860127,
      "tokens_seen": 1967718400
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.155191783020383e-05,
      "loss": 1.3493,
      "theoretical_loss": 3.4635061083645957,
      "tokens_seen": 1967849472
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.154389343604558e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.463486971774769,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.153586904188733e-05,
      "loss": 1.3766,
      "theoretical_loss": 3.463467836816285,
      "tokens_seen": 1968111616
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.15278446477291e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.463448703488895,
      "tokens_seen": 1968242688
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.151982025357085e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.4634295717923527,
      "tokens_seen": 1968373760
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.151179585941262e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.4634104417264098,
      "tokens_seen": 1968504832
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.150377146525437e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.4633913132908187,
      "tokens_seen": 1968635904
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.149574707109614e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.463372186485332,
      "tokens_seen": 1968766976
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.148772267693789e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.4633530613097028,
      "tokens_seen": 1968898048
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.147969828277966e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.463333937763683,
      "tokens_seen": 1969029120
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.147167388862141e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.463314815847026,
      "tokens_seen": 1969160192
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.146364949446317e-05,
      "loss": 1.3521,
      "theoretical_loss": 3.4632956955594834,
      "tokens_seen": 1969291264
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.145562510030493e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.463276576900809,
      "tokens_seen": 1969422336
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.144760070614668e-05,
      "loss": 1.3633,
      "theoretical_loss": 3.4632574598707553,
      "tokens_seen": 1969553408
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.143957631198845e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.463238344469075,
      "tokens_seen": 1969684480
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.14315519178302e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.463219230695522,
      "tokens_seen": 1969815552
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.142352752367197e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.4632001185498478,
      "tokens_seen": 1969946624
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.141550312951372e-05,
      "loss": 1.3448,
      "theoretical_loss": 3.4631810080318064,
      "tokens_seen": 1970077696
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.140747873535548e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.4631618991411512,
      "tokens_seen": 1970208768
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.139945434119724e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.463142791877634,
      "tokens_seen": 1970339840
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.1391429947039e-05,
      "loss": 1.2293,
      "theoretical_loss": 3.46312368624101,
      "tokens_seen": 1970470912
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.138340555288076e-05,
      "loss": 1.31,
      "theoretical_loss": 3.463104582231031,
      "tokens_seen": 1970601984
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.137538115872252e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.4630854798474506,
      "tokens_seen": 1970733056
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.136735676456428e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.4630663790900225,
      "tokens_seen": 1970864128
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.135933237040604e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.4630472799585,
      "tokens_seen": 1970995200
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.9945093393325806,
      "objective/train/avg_token_score": 0.005490629933774471,
      "objective/train/docs_used": 718798,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.400763750076294,
      "objective/train/original_loss": 1.4092763662338257,
      "objective/train/theoretical_loss": 3.4630281824526365,
      "objective/train/tokens_used": 341585376,
      "objective/train/unlikelihood_loss": 0.013902870938181877,
      "theoretical_loss": 3.4630281824526365,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.135130797624779e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.4630281824526365,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.134328358208956e-05,
      "loss": 1.3975,
      "theoretical_loss": 3.4630090865721854,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.133525918793131e-05,
      "loss": 1.3582,
      "theoretical_loss": 3.462989992316901,
      "tokens_seen": 1971388416
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.132723479377308e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.462970899686536,
      "tokens_seen": 1971519488
    },
    {
      "epoch": 0.19,
      "learning_rate": 8.131921039961483e-05,
      "loss": 1.3707,
      "theoretical_loss": 3.4629518086808444,
      "tokens_seen": 1971650560
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.13111860054566e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.4629327192995802,
      "tokens_seen": 1971781632
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.130316161129835e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.4629136315424973,
      "tokens_seen": 1971912704
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.12951372171401e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.462894545409349,
      "tokens_seen": 1972043776
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.128711282298187e-05,
      "loss": 1.29,
      "theoretical_loss": 3.46287546089989,
      "tokens_seen": 1972174848
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.127908842882362e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.462856378013874,
      "tokens_seen": 1972305920
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.127106403466539e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.4628372967510543,
      "tokens_seen": 1972436992
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.126303964050714e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.4628182171111863,
      "tokens_seen": 1972568064
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.125501524634891e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.462799139094023,
      "tokens_seen": 1972699136
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.124699085219066e-05,
      "loss": 1.367,
      "theoretical_loss": 3.462780062699319,
      "tokens_seen": 1972830208
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.123896645803242e-05,
      "loss": 1.416,
      "theoretical_loss": 3.4627609879268286,
      "tokens_seen": 1972961280
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.123094206387418e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.4627419147763057,
      "tokens_seen": 1973092352
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.122291766971594e-05,
      "loss": 1.3733,
      "theoretical_loss": 3.4627228432475055,
      "tokens_seen": 1973223424
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.12148932755577e-05,
      "loss": 1.3797,
      "theoretical_loss": 3.4627037733401815,
      "tokens_seen": 1973354496
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.120686888139945e-05,
      "loss": 1.272,
      "theoretical_loss": 3.462684705054089,
      "tokens_seen": 1973485568
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.119884448724122e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.4626656383889816,
      "tokens_seen": 1973616640
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.119082009308297e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.4626465733446143,
      "tokens_seen": 1973747712
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.118279569892473e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.462627509920742,
      "tokens_seen": 1973878784
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.11747713047665e-05,
      "loss": 1.3896,
      "theoretical_loss": 3.462608448117119,
      "tokens_seen": 1974009856
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.116674691060825e-05,
      "loss": 1.3685,
      "theoretical_loss": 3.4625893879335,
      "tokens_seen": 1974140928
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.115872251645001e-05,
      "loss": 1.321,
      "theoretical_loss": 3.4625703293696395,
      "tokens_seen": 1974272000
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.115069812229177e-05,
      "loss": 1.2431,
      "theoretical_loss": 3.4625512724252934,
      "tokens_seen": 1974403072
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.114267372813353e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.4625322171002155,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.113464933397529e-05,
      "loss": 1.4328,
      "theoretical_loss": 3.4625131633941617,
      "tokens_seen": 1974665216
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.112662493981705e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.462494111306886,
      "tokens_seen": 1974796288
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.111860054565881e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.4624750608381434,
      "tokens_seen": 1974927360
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.111057615150056e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.4624560119876904,
      "tokens_seen": 1975058432
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.110255175734233e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.4624369647552804,
      "tokens_seen": 1975189504
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.109452736318408e-05,
      "loss": 1.3702,
      "theoretical_loss": 3.4624179191406697,
      "tokens_seen": 1975320576
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.108650296902585e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.462398875143613,
      "tokens_seen": 1975451648
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.10784785748676e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.462379832763866,
      "tokens_seen": 1975582720
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.107045418070937e-05,
      "loss": 1.336,
      "theoretical_loss": 3.462360792001184,
      "tokens_seen": 1975713792
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.106242978655112e-05,
      "loss": 1.3441,
      "theoretical_loss": 3.4623417528553224,
      "tokens_seen": 1975844864
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.105440539239287e-05,
      "loss": 1.3768,
      "theoretical_loss": 3.462322715326037,
      "tokens_seen": 1975975936
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.104638099823464e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.4623036794130826,
      "tokens_seen": 1976107008
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.103835660407639e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.462284645116215,
      "tokens_seen": 1976238080
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.103033220991816e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.46226561243519,
      "tokens_seen": 1976369152
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.102230781575991e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.462246581369763,
      "tokens_seen": 1976500224
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.101428342160168e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.46222755191969,
      "tokens_seen": 1976631296
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.100625902744343e-05,
      "loss": 1.3614,
      "theoretical_loss": 3.462208524084727,
      "tokens_seen": 1976762368
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.099823463328519e-05,
      "loss": 1.2995,
      "theoretical_loss": 3.46218949786463,
      "tokens_seen": 1976893440
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.099021023912695e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.4621704732591536,
      "tokens_seen": 1977024512
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.09821858449687e-05,
      "loss": 1.308,
      "theoretical_loss": 3.4621514502680553,
      "tokens_seen": 1977155584
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.097416145081047e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.4621324288910897,
      "tokens_seen": 1977286656
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.096613705665222e-05,
      "loss": 1.2311,
      "theoretical_loss": 3.4621134091280137,
      "tokens_seen": 1977417728
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.095811266249399e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.4620943909785837,
      "tokens_seen": 1977548800
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.991405189037323,
      "objective/train/avg_token_score": 0.00859482679516077,
      "objective/train/docs_used": 721216,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3776800632476807,
      "objective/train/original_loss": 1.390610694885254,
      "objective/train/theoretical_loss": 3.4620753744425556,
      "objective/train/tokens_used": 348138976,
      "objective/train/unlikelihood_loss": 0.021327294409275055,
      "theoretical_loss": 3.4620753744425556,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.095008826833574e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.4620753744425556,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.09420638741775e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.4620563595196847,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.093403948001926e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.462037346209729,
      "tokens_seen": 1977942016
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.092601508586102e-05,
      "loss": 1.2324,
      "theoretical_loss": 3.462018334512443,
      "tokens_seen": 1978073088
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.091799069170278e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.461999324427584,
      "tokens_seen": 1978204160
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.090996629754454e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.461980315954909,
      "tokens_seen": 1978335232
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.09019419033863e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.461961309094173,
      "tokens_seen": 1978466304
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.089391750922806e-05,
      "loss": 1.4139,
      "theoretical_loss": 3.4619423038451336,
      "tokens_seen": 1978597376
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.088589311506981e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.4619233002075473,
      "tokens_seen": 1978728448
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.087786872091158e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.46190429818117,
      "tokens_seen": 1978859520
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.086984432675333e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.4618852977657593,
      "tokens_seen": 1978990592
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.08618199325951e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.461866298961072,
      "tokens_seen": 1979121664
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.085379553843685e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.461847301766864,
      "tokens_seen": 1979252736
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.084577114427862e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.461828306182893,
      "tokens_seen": 1979383808
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.083774675012037e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.461809312208915,
      "tokens_seen": 1979514880
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.082972235596214e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.4617903198446878,
      "tokens_seen": 1979645952
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.082169796180389e-05,
      "loss": 1.237,
      "theoretical_loss": 3.461771329089968,
      "tokens_seen": 1979777024
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.081367356764564e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.4617523399445123,
      "tokens_seen": 1979908096
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.080564917348741e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.4617333524080784,
      "tokens_seen": 1980039168
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.079762477932916e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.461714366480424,
      "tokens_seen": 1980170240
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.078960038517093e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.461695382161305,
      "tokens_seen": 1980301312
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.078157599101268e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.4616763994504796,
      "tokens_seen": 1980432384
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.077355159685445e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.4616574183477047,
      "tokens_seen": 1980563456
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.07655272026962e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.461638438852738,
      "tokens_seen": 1980694528
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.075750280853796e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.461619460965336,
      "tokens_seen": 1980825600
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.074947841437972e-05,
      "loss": 1.358,
      "theoretical_loss": 3.4616004846852575,
      "tokens_seen": 1980956672
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.074145402022148e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.4615815100122593,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.073342962606324e-05,
      "loss": 1.3499,
      "theoretical_loss": 3.461562536946099,
      "tokens_seen": 1981218816
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.0725405231905e-05,
      "loss": 1.3261,
      "theoretical_loss": 3.4615435654865347,
      "tokens_seen": 1981349888
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.071738083774676e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.4615245956333234,
      "tokens_seen": 1981480960
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.070935644358851e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.4615056273862232,
      "tokens_seen": 1981612032
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.070133204943027e-05,
      "loss": 1.264,
      "theoretical_loss": 3.461486660744992,
      "tokens_seen": 1981743104
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.069330765527203e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.4614676957093873,
      "tokens_seen": 1981874176
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.068528326111379e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.4614487322791674,
      "tokens_seen": 1982005248
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.067725886695555e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.46142977045409,
      "tokens_seen": 1982136320
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.066923447279731e-05,
      "loss": 1.25,
      "theoretical_loss": 3.461410810233913,
      "tokens_seen": 1982267392
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.066121007863907e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.4613918516183952,
      "tokens_seen": 1982398464
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.065318568448083e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.461372894607294,
      "tokens_seen": 1982529536
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.064516129032258e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.461353939200367,
      "tokens_seen": 1982660608
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.063713689616435e-05,
      "loss": 1.3381,
      "theoretical_loss": 3.461334985397374,
      "tokens_seen": 1982791680
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.06291125020061e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.461316033198072,
      "tokens_seen": 1982922752
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.062108810784787e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.4612970826022194,
      "tokens_seen": 1983053824
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.061306371368962e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.461278133609575,
      "tokens_seen": 1983184896
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.060503931953139e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.4612591862198974,
      "tokens_seen": 1983315968
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.059701492537314e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.4612402404329448,
      "tokens_seen": 1983447040
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.058899053121489e-05,
      "loss": 1.3631,
      "theoretical_loss": 3.4612212962484756,
      "tokens_seen": 1983578112
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.058096613705666e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4612023536662484,
      "tokens_seen": 1983709184
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.057294174289841e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.4611834126860224,
      "tokens_seen": 1983840256
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.056491734874018e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.461164473307555,
      "tokens_seen": 1983971328
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.055689295458193e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.4611455355306067,
      "tokens_seen": 1984102400
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.9923150539398193,
      "objective/train/avg_token_score": 0.0076849013566970825,
      "objective/train/docs_used": 723589,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2110234498977661,
      "objective/train/original_loss": 1.223958969116211,
      "objective/train/theoretical_loss": 3.4611265993549347,
      "objective/train/tokens_used": 354692576,
      "objective/train/unlikelihood_loss": 0.020189402624964714,
      "theoretical_loss": 3.4611265993549347,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.05488685604237e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.4611265993549347,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.054084416626545e-05,
      "loss": 1.3499,
      "theoretical_loss": 3.4611076647802985,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.053281977210722e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.461088731806457,
      "tokens_seen": 1984495616
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.052479537794897e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.4610698004331693,
      "tokens_seen": 1984626688
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.051677098379073e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.461050870660194,
      "tokens_seen": 1984757760
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.050874658963249e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.46103194248729,
      "tokens_seen": 1984888832
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.050072219547425e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.461013015914218,
      "tokens_seen": 1985019904
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.049269780131601e-05,
      "loss": 1.3902,
      "theoretical_loss": 3.4609940909407353,
      "tokens_seen": 1985150976
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.048467340715777e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.4609751675666014,
      "tokens_seen": 1985282048
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.047664901299953e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.460956245791576,
      "tokens_seen": 1985413120
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.046862461884128e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.4609373256154186,
      "tokens_seen": 1985544192
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.046060022468304e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.4609184070378882,
      "tokens_seen": 1985675264
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.04525758305248e-05,
      "loss": 1.349,
      "theoretical_loss": 3.460899490058744,
      "tokens_seen": 1985806336
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.044455143636656e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.4608805746777467,
      "tokens_seen": 1985937408
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.043652704220832e-05,
      "loss": 1.3742,
      "theoretical_loss": 3.460861660894654,
      "tokens_seen": 1986068480
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.042850264805008e-05,
      "loss": 1.3292,
      "theoretical_loss": 3.4608427487092266,
      "tokens_seen": 1986199552
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.042047825389184e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.4608238381212235,
      "tokens_seen": 1986330624
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.04124538597336e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.4608049291304055,
      "tokens_seen": 1986461696
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.040442946557535e-05,
      "loss": 1.3525,
      "theoretical_loss": 3.4607860217365314,
      "tokens_seen": 1986592768
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.039640507141712e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.460767115939361,
      "tokens_seen": 1986723840
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.038838067725887e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.460748211738654,
      "tokens_seen": 1986854912
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.038035628310064e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.4607293091341713,
      "tokens_seen": 1986985984
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.037233188894239e-05,
      "loss": 1.3298,
      "theoretical_loss": 3.4607104081256717,
      "tokens_seen": 1987117056
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.036430749478416e-05,
      "loss": 1.3436,
      "theoretical_loss": 3.4606915087129155,
      "tokens_seen": 1987248128
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.035628310062591e-05,
      "loss": 1.367,
      "theoretical_loss": 3.4606726108956627,
      "tokens_seen": 1987379200
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.034825870646766e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.460653714673674,
      "tokens_seen": 1987510272
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.034023431230943e-05,
      "loss": 1.2381,
      "theoretical_loss": 3.4606348200467094,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.033220991815118e-05,
      "loss": 1.304,
      "theoretical_loss": 3.460615927014528,
      "tokens_seen": 1987772416
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.032418552399295e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.4605970355768916,
      "tokens_seen": 1987903488
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.03161611298347e-05,
      "loss": 1.3597,
      "theoretical_loss": 3.4605781457335594,
      "tokens_seen": 1988034560
    },
    {
      "epoch": 0.2,
      "learning_rate": 8.030813673567647e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.4605592574842916,
      "tokens_seen": 1988165632
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.030011234151822e-05,
      "loss": 1.3776,
      "theoretical_loss": 3.46054037082885,
      "tokens_seen": 1988296704
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.029208794735998e-05,
      "loss": 1.297,
      "theoretical_loss": 3.4605214857669937,
      "tokens_seen": 1988427776
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.028406355320174e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.4605026022984835,
      "tokens_seen": 1988558848
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.02760391590435e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.4604837204230803,
      "tokens_seen": 1988689920
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.026801476488526e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.4604648401405447,
      "tokens_seen": 1988820992
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.025999037072702e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.460445961450637,
      "tokens_seen": 1988952064
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.025196597656878e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.460427084353118,
      "tokens_seen": 1989083136
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.024394158241054e-05,
      "loss": 1.344,
      "theoretical_loss": 3.4604082088477486,
      "tokens_seen": 1989214208
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.023591718825229e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.4603893349342894,
      "tokens_seen": 1989345280
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.022789279409406e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.4603704626125023,
      "tokens_seen": 1989476352
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.021986839993581e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.4603515918821466,
      "tokens_seen": 1989607424
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.021184400577757e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.4603327227429843,
      "tokens_seen": 1989738496
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.020381961161933e-05,
      "loss": 1.259,
      "theoretical_loss": 3.460313855194776,
      "tokens_seen": 1989869568
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.01957952174611e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.4602949892372834,
      "tokens_seen": 1990000640
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.018777082330285e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.4602761248702665,
      "tokens_seen": 1990131712
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.017974642914461e-05,
      "loss": 1.3407,
      "theoretical_loss": 3.4602572620934877,
      "tokens_seen": 1990262784
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.017172203498637e-05,
      "loss": 1.331,
      "theoretical_loss": 3.4602384009067073,
      "tokens_seen": 1990393856
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.016369764082812e-05,
      "loss": 1.323,
      "theoretical_loss": 3.4602195413096877,
      "tokens_seen": 1990524928
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.015567324666989e-05,
      "loss": 1.3795,
      "theoretical_loss": 3.4602006833021886,
      "tokens_seen": 1990656000
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.9909166693687439,
      "objective/train/avg_token_score": 0.009083281271159649,
      "objective/train/docs_used": 726036,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3317525386810303,
      "objective/train/original_loss": 1.3440974950790405,
      "objective/train/theoretical_loss": 3.4601818268839732,
      "objective/train/tokens_used": 361246176,
      "objective/train/unlikelihood_loss": 0.025575045496225357,
      "theoretical_loss": 3.4601818268839732,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.014764885251164e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.4601818268839732,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.013962445835341e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.4601629720548015,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.013160006419516e-05,
      "loss": 1.4005,
      "theoretical_loss": 3.4601441188144353,
      "tokens_seen": 1991049216
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.012357567003693e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.4601252671626366,
      "tokens_seen": 1991180288
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.011555127587868e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.4601064170991673,
      "tokens_seen": 1991311360
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.010752688172043e-05,
      "loss": 1.4128,
      "theoretical_loss": 3.4600875686237877,
      "tokens_seen": 1991442432
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.00995024875622e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.4600687217362607,
      "tokens_seen": 1991573504
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.009147809340395e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.4600498764363476,
      "tokens_seen": 1991704576
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.008345369924572e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.46003103272381,
      "tokens_seen": 1991835648
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.007542930508747e-05,
      "loss": 1.3574,
      "theoretical_loss": 3.4600121905984103,
      "tokens_seen": 1991966720
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.006740491092924e-05,
      "loss": 1.307,
      "theoretical_loss": 3.45999335005991,
      "tokens_seen": 1992097792
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.005938051677099e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.4599745111080713,
      "tokens_seen": 1992228864
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.005135612261275e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.4599556737426553,
      "tokens_seen": 1992359936
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.004333172845451e-05,
      "loss": 1.319,
      "theoretical_loss": 3.4599368379634257,
      "tokens_seen": 1992491008
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.003530733429627e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.4599180037701434,
      "tokens_seen": 1992622080
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.002728294013803e-05,
      "loss": 1.3394,
      "theoretical_loss": 3.4598991711625704,
      "tokens_seen": 1992753152
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.001925854597979e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.45988034014047,
      "tokens_seen": 1992884224
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.001123415182155e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.4598615107036035,
      "tokens_seen": 1993015296
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.00032097576633e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.4598426828517335,
      "tokens_seen": 1993146368
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.999518536350506e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.459823856584623,
      "tokens_seen": 1993277440
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.998716096934683e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.459805031902033,
      "tokens_seen": 1993408512
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.997913657518858e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.4597862088037266,
      "tokens_seen": 1993539584
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.997111218103034e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.4597673872894674,
      "tokens_seen": 1993670656
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.99630877868721e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.4597485673590165,
      "tokens_seen": 1993801728
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.995506339271386e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.459729749012137,
      "tokens_seen": 1993932800
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.994703899855562e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.459710932248592,
      "tokens_seen": 1994063872
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.993901460439737e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.459692117068144,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.993099021023914e-05,
      "loss": 1.348,
      "theoretical_loss": 3.459673303470555,
      "tokens_seen": 1994326016
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.992296581608089e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.4596544914555887,
      "tokens_seen": 1994457088
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.991494142192266e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.4596356810230082,
      "tokens_seen": 1994588160
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.990691702776441e-05,
      "loss": 1.3416,
      "theoretical_loss": 3.4596168721725755,
      "tokens_seen": 1994719232
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.989889263360616e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.4595980649040543,
      "tokens_seen": 1994850304
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.989086823944793e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.459579259217207,
      "tokens_seen": 1994981376
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.988284384528968e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.4595604551117973,
      "tokens_seen": 1995112448
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.987481945113145e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.459541652587588,
      "tokens_seen": 1995243520
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.98667950569732e-05,
      "loss": 1.4091,
      "theoretical_loss": 3.4595228516443424,
      "tokens_seen": 1995374592
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.985877066281497e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.4595040522818232,
      "tokens_seen": 1995505664
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.985074626865672e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.459485254499795,
      "tokens_seen": 1995636736
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.984272187449848e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.45946645829802,
      "tokens_seen": 1995767808
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.983469748034024e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.4594476636762614,
      "tokens_seen": 1995898880
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.9826673086182e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.4594288706342837,
      "tokens_seen": 1996029952
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.981864869202376e-05,
      "loss": 1.1761,
      "theoretical_loss": 3.4594100791718496,
      "tokens_seen": 1996161024
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.981062429786552e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.459391289288723,
      "tokens_seen": 1996292096
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.980259990370727e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.459372500984667,
      "tokens_seen": 1996423168
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.979457550954904e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.459353714259446,
      "tokens_seen": 1996554240
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.978655111539079e-05,
      "loss": 1.3323,
      "theoretical_loss": 3.4593349291128233,
      "tokens_seen": 1996685312
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.977852672123256e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.459316145544562,
      "tokens_seen": 1996816384
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.977050232707431e-05,
      "loss": 1.3668,
      "theoretical_loss": 3.459297363554427,
      "tokens_seen": 1996947456
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.976247793291608e-05,
      "loss": 1.3215,
      "theoretical_loss": 3.459278583142181,
      "tokens_seen": 1997078528
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.975445353875783e-05,
      "loss": 1.3554,
      "theoretical_loss": 3.4592598043075893,
      "tokens_seen": 1997209600
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.9915937185287476,
      "objective/train/avg_token_score": 0.008406254462897778,
      "objective/train/docs_used": 728467,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2727662324905396,
      "objective/train/original_loss": 1.2843021154403687,
      "objective/train/theoretical_loss": 3.459241027050415,
      "objective/train/tokens_used": 367799776,
      "objective/train/unlikelihood_loss": 0.02523968555033207,
      "theoretical_loss": 3.459241027050415,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.974642914459958e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.459241027050415,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.973840475044135e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.459222251370422,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.97303803562831e-05,
      "loss": 1.3593,
      "theoretical_loss": 3.459203477267374,
      "tokens_seen": 1997602816
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.972235596212487e-05,
      "loss": 1.3557,
      "theoretical_loss": 3.4591847047410362,
      "tokens_seen": 1997733888
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.971433156796662e-05,
      "loss": 1.2486,
      "theoretical_loss": 3.459165933791172,
      "tokens_seen": 1997864960
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.970630717380837e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.4591471644175456,
      "tokens_seen": 1997996032
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.969828277965014e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.4591283966199216,
      "tokens_seen": 1998127104
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.96902583854919e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.459109630398064,
      "tokens_seen": 1998258176
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.968223399133366e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.4590908657517376,
      "tokens_seen": 1998389248
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.967420959717541e-05,
      "loss": 1.349,
      "theoretical_loss": 3.459072102680706,
      "tokens_seen": 1998520320
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.966618520301718e-05,
      "loss": 1.353,
      "theoretical_loss": 3.4590533411847346,
      "tokens_seen": 1998651392
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.965816080885893e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.4590345812635865,
      "tokens_seen": 1998782464
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.965013641470069e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.459015822917028,
      "tokens_seen": 1998913536
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.964211202054245e-05,
      "loss": 1.3298,
      "theoretical_loss": 3.4589970661448226,
      "tokens_seen": 1999044608
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.96340876263842e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.458978310946735,
      "tokens_seen": 1999175680
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.962606323222597e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.4589595573225305,
      "tokens_seen": 1999306752
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.961803883806773e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.4589408052719737,
      "tokens_seen": 1999437824
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.961001444390948e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.4589220547948285,
      "tokens_seen": 1999568896
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.960199004975125e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.458903305890861,
      "tokens_seen": 1999699968
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.9593965655593e-05,
      "loss": 1.4172,
      "theoretical_loss": 3.458884558559835,
      "tokens_seen": 1999831040
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.958594126143477e-05,
      "loss": 1.3799,
      "theoretical_loss": 3.4588658128015166,
      "tokens_seen": 1999962112
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.957791686727652e-05,
      "loss": 1.231,
      "theoretical_loss": 3.45884706861567,
      "tokens_seen": 2000093184
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.956989247311829e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.4588283260020605,
      "tokens_seen": 2000224256
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.956186807896004e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.458809584960453,
      "tokens_seen": 2000355328
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.955384368480179e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.458790845490613,
      "tokens_seen": 2000486400
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.954581929064356e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.4587721075923055,
      "tokens_seen": 2000617472
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.953779489648531e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.458753371265296,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.952977050232708e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.4587346365093494,
      "tokens_seen": 2000879616
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.952174610816883e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.458715903324231,
      "tokens_seen": 2001010688
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.951372171401059e-05,
      "loss": 1.2426,
      "theoretical_loss": 3.4586971717097077,
      "tokens_seen": 2001141760
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.950569731985235e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.4586784416655423,
      "tokens_seen": 2001272832
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.94976729256941e-05,
      "loss": 1.3604,
      "theoretical_loss": 3.4586597131915027,
      "tokens_seen": 2001403904
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.948964853153587e-05,
      "loss": 1.3495,
      "theoretical_loss": 3.458640986287353,
      "tokens_seen": 2001534976
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.948162413737762e-05,
      "loss": 1.3445,
      "theoretical_loss": 3.45862226095286,
      "tokens_seen": 2001666048
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.947359974321939e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.458603537187788,
      "tokens_seen": 2001797120
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.946557534906114e-05,
      "loss": 1.3197,
      "theoretical_loss": 3.458584814991904,
      "tokens_seen": 2001928192
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.94575509549029e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.458566094364973,
      "tokens_seen": 2002059264
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.944952656074466e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.458547375306761,
      "tokens_seen": 2002190336
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.944150216658642e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.4585286578170336,
      "tokens_seen": 2002321408
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.943347777242818e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.4585099418955574,
      "tokens_seen": 2002452480
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.942545337826994e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.4584912275420976,
      "tokens_seen": 2002583552
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.94174289841117e-05,
      "loss": 1.3538,
      "theoretical_loss": 3.4584725147564206,
      "tokens_seen": 2002714624
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.940940458995346e-05,
      "loss": 1.3215,
      "theoretical_loss": 3.4584538035382923,
      "tokens_seen": 2002845696
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.940138019579521e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.4584350938874797,
      "tokens_seen": 2002976768
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.939335580163698e-05,
      "loss": 1.3421,
      "theoretical_loss": 3.4584163858037478,
      "tokens_seen": 2003107840
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.938533140747873e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.458397679286863,
      "tokens_seen": 2003238912
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.93773070133205e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.458378974336592,
      "tokens_seen": 2003369984
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.936928261916225e-05,
      "loss": 1.3538,
      "theoretical_loss": 3.458360270952701,
      "tokens_seen": 2003501056
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.936125822500402e-05,
      "loss": 1.304,
      "theoretical_loss": 3.458341569134956,
      "tokens_seen": 2003632128
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.935323383084577e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.458322868883124,
      "tokens_seen": 2003763200
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.9897269606590271,
      "objective/train/avg_token_score": 0.01027307566255331,
      "objective/train/docs_used": 730811,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.372354507446289,
      "objective/train/original_loss": 1.3877522945404053,
      "objective/train/theoretical_loss": 3.4583041701969712,
      "objective/train/tokens_used": 374353376,
      "objective/train/unlikelihood_loss": 0.027070851996541023,
      "theoretical_loss": 3.4583041701969712,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.934520943668752e-05,
      "loss": 1.333,
      "theoretical_loss": 3.4583041701969712,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.933718504252929e-05,
      "loss": 1.3876,
      "theoretical_loss": 3.4582854730762635,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.932916064837104e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.4582667775207687,
      "tokens_seen": 2004156416
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.932113625421281e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.4582480835302527,
      "tokens_seen": 2004287488
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.931311186005456e-05,
      "loss": 1.3402,
      "theoretical_loss": 3.4582293911044824,
      "tokens_seen": 2004418560
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.930508746589633e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.4582107002432236,
      "tokens_seen": 2004549632
    },
    {
      "epoch": 0.21,
      "learning_rate": 7.929706307173808e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.4581920109462443,
      "tokens_seen": 2004680704
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.928903867757984e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.4581733232133107,
      "tokens_seen": 2004811776
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.92810142834216e-05,
      "loss": 1.3456,
      "theoretical_loss": 3.4581546370441902,
      "tokens_seen": 2004942848
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.927298988926336e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.4581359524386492,
      "tokens_seen": 2005073920
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.926496549510512e-05,
      "loss": 1.4042,
      "theoretical_loss": 3.458117269396455,
      "tokens_seen": 2005204992
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.925694110094687e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.458098587917374,
      "tokens_seen": 2005336064
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.924891670678864e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.4580799080011735,
      "tokens_seen": 2005467136
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.92408923126304e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.4580612296476207,
      "tokens_seen": 2005598208
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.923286791847216e-05,
      "loss": 1.3333,
      "theoretical_loss": 3.458042552856484,
      "tokens_seen": 2005729280
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.922484352431391e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.4580238776275283,
      "tokens_seen": 2005860352
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.921681913015567e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.4580052039605222,
      "tokens_seen": 2005991424
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.920879473599743e-05,
      "loss": 1.3413,
      "theoretical_loss": 3.4579865318552336,
      "tokens_seen": 2006122496
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.920077034183919e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.4579678613114284,
      "tokens_seen": 2006253568
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.919274594768095e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.457949192328875,
      "tokens_seen": 2006384640
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.918472155352271e-05,
      "loss": 1.3328,
      "theoretical_loss": 3.457930524907341,
      "tokens_seen": 2006515712
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.917669715936447e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.457911859046593,
      "tokens_seen": 2006646784
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.916867276520623e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.4578931947464,
      "tokens_seen": 2006777856
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.916064837104798e-05,
      "loss": 1.2307,
      "theoretical_loss": 3.457874532006527,
      "tokens_seen": 2006908928
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.915262397688975e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.457855870826745,
      "tokens_seen": 2007040000
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.91445995827315e-05,
      "loss": 1.333,
      "theoretical_loss": 3.4578372112068196,
      "tokens_seen": 2007171072
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.913657518857327e-05,
      "loss": 1.3607,
      "theoretical_loss": 3.4578185531465184,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.912855079441502e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.4577998966456107,
      "tokens_seen": 2007433216
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.912052640025679e-05,
      "loss": 1.3215,
      "theoretical_loss": 3.4577812417038634,
      "tokens_seen": 2007564288
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.911250200609854e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.457762588321044,
      "tokens_seen": 2007695360
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.910447761194029e-05,
      "loss": 1.295,
      "theoretical_loss": 3.4577439364969216,
      "tokens_seen": 2007826432
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.909645321778206e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.4577252862312635,
      "tokens_seen": 2007957504
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.908842882362381e-05,
      "loss": 1.1959,
      "theoretical_loss": 3.457706637523838,
      "tokens_seen": 2008088576
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.908040442946558e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.457687990374413,
      "tokens_seen": 2008219648
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.907238003530733e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.457669344782757,
      "tokens_seen": 2008350720
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.90643556411491e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.457650700748638,
      "tokens_seen": 2008481792
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.905633124699085e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.457632058271824,
      "tokens_seen": 2008612864
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.90483068528326e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.4576134173520834,
      "tokens_seen": 2008743936
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.904028245867437e-05,
      "loss": 1.27,
      "theoretical_loss": 3.457594777989185,
      "tokens_seen": 2008875008
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.903225806451613e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.457576140182897,
      "tokens_seen": 2009006080
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.902423367035789e-05,
      "loss": 1.311,
      "theoretical_loss": 3.4575575039329873,
      "tokens_seen": 2009137152
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.901620927619964e-05,
      "loss": 1.2927,
      "theoretical_loss": 3.457538869239225,
      "tokens_seen": 2009268224
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.900818488204141e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.457520236101379,
      "tokens_seen": 2009399296
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.900016048788316e-05,
      "loss": 1.3763,
      "theoretical_loss": 3.4575016045192175,
      "tokens_seen": 2009530368
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.899213609372492e-05,
      "loss": 1.3453,
      "theoretical_loss": 3.4574829744925086,
      "tokens_seen": 2009661440
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.898411169956668e-05,
      "loss": 1.3679,
      "theoretical_loss": 3.457464346021022,
      "tokens_seen": 2009792512
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.897608730540844e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.457445719104526,
      "tokens_seen": 2009923584
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.89680629112502e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.457427093742789,
      "tokens_seen": 2010054656
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.896003851709196e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.4574084699355807,
      "tokens_seen": 2010185728
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.895201412293372e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.4573898476826694,
      "tokens_seen": 2010316800
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.9927845001220703,
      "objective/train/avg_token_score": 0.007215526886284351,
      "objective/train/docs_used": 733267,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2716056108474731,
      "objective/train/original_loss": 1.2806960344314575,
      "objective/train/theoretical_loss": 3.457371226983824,
      "objective/train/tokens_used": 380906976,
      "objective/train/unlikelihood_loss": 0.028046388179063797,
      "theoretical_loss": 3.457371226983824,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.894398972877548e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.457371226983824,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.893596533461724e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.457352607838814,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.8927940940459e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.4573339902474083,
      "tokens_seen": 2010710016
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.891991654630075e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.457315374209376,
      "tokens_seen": 2010841088
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.891189215214252e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.4572967597244855,
      "tokens_seen": 2010972160
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.890386775798427e-05,
      "loss": 1.3554,
      "theoretical_loss": 3.4572781467925067,
      "tokens_seen": 2011103232
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.889584336382604e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.4572595354132094,
      "tokens_seen": 2011234304
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.888781896966779e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.457240925586362,
      "tokens_seen": 2011365376
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.887979457550956e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.457222317311734,
      "tokens_seen": 2011496448
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.887177018135131e-05,
      "loss": 1.3211,
      "theoretical_loss": 3.457203710589095,
      "tokens_seen": 2011627520
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.886374578719306e-05,
      "loss": 1.2146,
      "theoretical_loss": 3.4571851054182146,
      "tokens_seen": 2011758592
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.885572139303483e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.457166501798862,
      "tokens_seen": 2011889664
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.884769699887658e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.4571478997308063,
      "tokens_seen": 2012020736
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.883967260471835e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.4571292992138183,
      "tokens_seen": 2012151808
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.88316482105601e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.4571107002476666,
      "tokens_seen": 2012282880
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.882362381640187e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.4570921028321218,
      "tokens_seen": 2012413952
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.881559942224362e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.4570735069669523,
      "tokens_seen": 2012545024
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.880757502808538e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.4570549126519294,
      "tokens_seen": 2012676096
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.879955063392714e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.457036319886822,
      "tokens_seen": 2012807168
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.87915262397689e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.4570177286714,
      "tokens_seen": 2012938240
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.878350184561066e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.4569991390054335,
      "tokens_seen": 2013069312
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.877547745145242e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.4569805508886935,
      "tokens_seen": 2013200384
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.876745305729418e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.456961964320948,
      "tokens_seen": 2013331456
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.875942866313593e-05,
      "loss": 1.3446,
      "theoretical_loss": 3.4569433793019693,
      "tokens_seen": 2013462528
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.875140426897769e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.4569247958315255,
      "tokens_seen": 2013593600
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.874337987481945e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.456906213909388,
      "tokens_seen": 2013724672
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.873535548066121e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.456887633535327,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.872733108650297e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.456869054709112,
      "tokens_seen": 2013986816
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.871930669234473e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.4568504774305144,
      "tokens_seen": 2014117888
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.87112822981865e-05,
      "loss": 1.36,
      "theoretical_loss": 3.456831901699304,
      "tokens_seen": 2014248960
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.870325790402825e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.4568133275152513,
      "tokens_seen": 2014380032
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.869523350987e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.4567947548781266,
      "tokens_seen": 2014511104
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.868720911571177e-05,
      "loss": 1.3564,
      "theoretical_loss": 3.4567761837877002,
      "tokens_seen": 2014642176
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.867918472155352e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.456757614243744,
      "tokens_seen": 2014773248
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.867116032739529e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.4567390462460272,
      "tokens_seen": 2014904320
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.866313593323704e-05,
      "loss": 1.2048,
      "theoretical_loss": 3.456720479794321,
      "tokens_seen": 2015035392
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.865511153907881e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.456701914888396,
      "tokens_seen": 2015166464
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.864708714492056e-05,
      "loss": 1.259,
      "theoretical_loss": 3.4566833515280235,
      "tokens_seen": 2015297536
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.863906275076233e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.456664789712973,
      "tokens_seen": 2015428608
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.863103835660408e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.456646229443017,
      "tokens_seen": 2015559680
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.862301396244583e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.4566276707179258,
      "tokens_seen": 2015690752
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.86149895682876e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.45660911353747,
      "tokens_seen": 2015821824
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.860696517412935e-05,
      "loss": 1.3087,
      "theoretical_loss": 3.456590557901421,
      "tokens_seen": 2015952896
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.859894077997112e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.4565720038095495,
      "tokens_seen": 2016083968
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.859091638581287e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.4565534512616267,
      "tokens_seen": 2016215040
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.858289199165464e-05,
      "loss": 1.2941,
      "theoretical_loss": 3.4565349002574246,
      "tokens_seen": 2016346112
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.857486759749639e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.456516350796713,
      "tokens_seen": 2016477184
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.856684320333815e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.4564978028792646,
      "tokens_seen": 2016608256
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.855881880917991e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.4564792565048497,
      "tokens_seen": 2016739328
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.855079441502167e-05,
      "loss": 1.3371,
      "theoretical_loss": 3.45646071167324,
      "tokens_seen": 2016870400
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.9912194013595581,
      "objective/train/avg_token_score": 0.00878061167895794,
      "objective/train/docs_used": 735776,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4352633953094482,
      "objective/train/original_loss": 1.4497721195220947,
      "objective/train/theoretical_loss": 3.456442168384207,
      "objective/train/tokens_used": 387460576,
      "objective/train/unlikelihood_loss": 0.022441977635025978,
      "theoretical_loss": 3.456442168384207,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.854277002086343e-05,
      "loss": 1.4238,
      "theoretical_loss": 3.456442168384207,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.853474562670519e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.456423626637522,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.852672123254695e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.456405086432956,
      "tokens_seen": 2017263616
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.85186968383887e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.4563865477702818,
      "tokens_seen": 2017394688
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.851067244423046e-05,
      "loss": 1.361,
      "theoretical_loss": 3.4563680106492702,
      "tokens_seen": 2017525760
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.850264805007222e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.456349475069693,
      "tokens_seen": 2017656832
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.849462365591398e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.4563309410313217,
      "tokens_seen": 2017787904
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.848659926175574e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.456312408533928,
      "tokens_seen": 2017918976
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.84785748675975e-05,
      "loss": 1.2494,
      "theoretical_loss": 3.4562938775772842,
      "tokens_seen": 2018050048
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.847055047343926e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.456275348161162,
      "tokens_seen": 2018181120
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.846252607928102e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.4562568202853328,
      "tokens_seen": 2018312192
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.845450168512277e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.456238293949569,
      "tokens_seen": 2018443264
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.844647729096454e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.4562197691536425,
      "tokens_seen": 2018574336
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.843845289680629e-05,
      "loss": 1.3313,
      "theoretical_loss": 3.4562012458973257,
      "tokens_seen": 2018705408
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.843042850264806e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.4561827241803895,
      "tokens_seen": 2018836480
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.842240410848981e-05,
      "loss": 1.3384,
      "theoretical_loss": 3.4561642040026075,
      "tokens_seen": 2018967552
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.841437971433158e-05,
      "loss": 1.283,
      "theoretical_loss": 3.456145685363751,
      "tokens_seen": 2019098624
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.840635532017333e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.456127168263593,
      "tokens_seen": 2019229696
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.839833092601508e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.4561086527019045,
      "tokens_seen": 2019360768
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.839030653185685e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.4560901386784586,
      "tokens_seen": 2019491840
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.83822821376986e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.456071626193028,
      "tokens_seen": 2019622912
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.837425774354037e-05,
      "loss": 1.339,
      "theoretical_loss": 3.4560531152453846,
      "tokens_seen": 2019753984
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.836623334938212e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.4560346058353018,
      "tokens_seen": 2019885056
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.835820895522389e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.4560160979625505,
      "tokens_seen": 2020016128
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.835018456106564e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.455997591626905,
      "tokens_seen": 2020147200
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.83421601669074e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.455979086828136,
      "tokens_seen": 2020278272
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.833413577274916e-05,
      "loss": 1.245,
      "theoretical_loss": 3.455960583566018,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.832611137859092e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.4559420818403233,
      "tokens_seen": 2020540416
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.831808698443268e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.455923581650824,
      "tokens_seen": 2020671488
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.831006259027444e-05,
      "loss": 1.341,
      "theoretical_loss": 3.455905082997293,
      "tokens_seen": 2020802560
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.83020381961162e-05,
      "loss": 1.3524,
      "theoretical_loss": 3.4558865858795036,
      "tokens_seen": 2020933632
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.829401380195796e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.4558680902972285,
      "tokens_seen": 2021064704
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.828598940779972e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.455849596250241,
      "tokens_seen": 2021195776
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.827796501364148e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.455831103738314,
      "tokens_seen": 2021326848
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.826994061948323e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.4558126127612194,
      "tokens_seen": 2021457920
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.8261916225325e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.455794123318732,
      "tokens_seen": 2021588992
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.825389183116675e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.4557756354106246,
      "tokens_seen": 2021720064
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.824586743700851e-05,
      "loss": 1.325,
      "theoretical_loss": 3.4557571490366694,
      "tokens_seen": 2021851136
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.823784304285027e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.45573866419664,
      "tokens_seen": 2021982208
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.822981864869203e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.455720180890311,
      "tokens_seen": 2022113280
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.822179425453379e-05,
      "loss": 1.3971,
      "theoretical_loss": 3.455701699117454,
      "tokens_seen": 2022244352
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.821376986037554e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.4556832188778435,
      "tokens_seen": 2022375424
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.820574546621731e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.455664740171253,
      "tokens_seen": 2022506496
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.819772107205906e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.4556462629974547,
      "tokens_seen": 2022637568
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.818969667790083e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.455627787356223,
      "tokens_seen": 2022768640
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.818167228374258e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.455609313247332,
      "tokens_seen": 2022899712
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.817364788958435e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.4555908406705544,
      "tokens_seen": 2023030784
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.81656234954261e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.455572369625665,
      "tokens_seen": 2023161856
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.815759910126785e-05,
      "loss": 1.3356,
      "theoretical_loss": 3.4555539001124362,
      "tokens_seen": 2023292928
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.814957470710962e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.4555354321306426,
      "tokens_seen": 2023424000
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.9941638112068176,
      "objective/train/avg_token_score": 0.005836206488311291,
      "objective/train/docs_used": 738179,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3090972900390625,
      "objective/train/original_loss": 1.3170045614242554,
      "objective/train/theoretical_loss": 3.455516965680058,
      "objective/train/tokens_used": 394014176,
      "objective/train/unlikelihood_loss": 0.01857689581811428,
      "theoretical_loss": 3.455516965680058,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.814155031295137e-05,
      "loss": 1.2197,
      "theoretical_loss": 3.455516965680058,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.813352591879314e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.455498500760456,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.812550152463489e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.455480037371611,
      "tokens_seen": 2023817216
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.811747713047666e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.4554615755132962,
      "tokens_seen": 2023948288
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.810945273631841e-05,
      "loss": 1.3421,
      "theoretical_loss": 3.455443115185286,
      "tokens_seen": 2024079360
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.810142834216017e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.455424656387355,
      "tokens_seen": 2024210432
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.809340394800193e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.4554061991192766,
      "tokens_seen": 2024341504
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.808537955384369e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.4553877433808253,
      "tokens_seen": 2024472576
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.807735515968545e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.4553692891717747,
      "tokens_seen": 2024603648
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.80693307655272e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4553508364919,
      "tokens_seen": 2024734720
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.806130637136897e-05,
      "loss": 1.4021,
      "theoretical_loss": 3.455332385340975,
      "tokens_seen": 2024865792
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.805328197721073e-05,
      "loss": 1.321,
      "theoretical_loss": 3.455313935718774,
      "tokens_seen": 2024996864
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.804525758305248e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.455295487625072,
      "tokens_seen": 2025127936
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.803723318889425e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.4552770410596425,
      "tokens_seen": 2025259008
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.8029208794736e-05,
      "loss": 1.326,
      "theoretical_loss": 3.4552585960222606,
      "tokens_seen": 2025390080
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.802118440057776e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.455240152512701,
      "tokens_seen": 2025521152
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.801316000641952e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.4552217105307377,
      "tokens_seen": 2025652224
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.800513561226128e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.4552032700761455,
      "tokens_seen": 2025783296
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.799711121810304e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.4551848311487,
      "tokens_seen": 2025914368
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.79890868239448e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.4551663937481747,
      "tokens_seen": 2026045440
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.798106242978656e-05,
      "loss": 1.2219,
      "theoretical_loss": 3.455147957874345,
      "tokens_seen": 2026176512
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.797303803562831e-05,
      "loss": 1.3369,
      "theoretical_loss": 3.4551295235269857,
      "tokens_seen": 2026307584
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.796501364147008e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.4551110907058713,
      "tokens_seen": 2026438656
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.795698924731183e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.455092659410777,
      "tokens_seen": 2026569728
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.79489648531536e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.4550742296414785,
      "tokens_seen": 2026700800
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.794094045899535e-05,
      "loss": 1.2414,
      "theoretical_loss": 3.4550558013977497,
      "tokens_seen": 2026831872
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.793291606483712e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.455037374679366,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.792489167067887e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.455018949486103,
      "tokens_seen": 2027094016
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.791686727652062e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.455000525817735,
      "tokens_seen": 2027225088
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.790884288236239e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.4549821036740385,
      "tokens_seen": 2027356160
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.790081848820414e-05,
      "loss": 1.2931,
      "theoretical_loss": 3.454963683054787,
      "tokens_seen": 2027487232
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.789279409404591e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.4549452639597575,
      "tokens_seen": 2027618304
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.788476969988766e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.4549268463887244,
      "tokens_seen": 2027749376
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.787674530572943e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.4549084303414634,
      "tokens_seen": 2027880448
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.786872091157118e-05,
      "loss": 1.335,
      "theoretical_loss": 3.45489001581775,
      "tokens_seen": 2028011520
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.786069651741294e-05,
      "loss": 1.404,
      "theoretical_loss": 3.4548716028173594,
      "tokens_seen": 2028142592
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.78526721232547e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.4548531913400677,
      "tokens_seen": 2028273664
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.784464772909646e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.45483478138565,
      "tokens_seen": 2028404736
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.783662333493822e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.454816372953882,
      "tokens_seen": 2028535808
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.782859894077998e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.4547979660445396,
      "tokens_seen": 2028666880
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.782057454662174e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.4547795606573986,
      "tokens_seen": 2028797952
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.78125501524635e-05,
      "loss": 1.303,
      "theoretical_loss": 3.4547611567922347,
      "tokens_seen": 2028929024
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.780452575830525e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.454742754448823,
      "tokens_seen": 2029060096
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.779650136414702e-05,
      "loss": 1.3803,
      "theoretical_loss": 3.4547243536269407,
      "tokens_seen": 2029191168
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.778847696998877e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.454705954326363,
      "tokens_seen": 2029322240
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.778045257583053e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.4546875565468658,
      "tokens_seen": 2029453312
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.777242818167229e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.4546691602882253,
      "tokens_seen": 2029584384
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.776440378751405e-05,
      "loss": 1.257,
      "theoretical_loss": 3.454650765550218,
      "tokens_seen": 2029715456
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.775637939335581e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.4546323723326187,
      "tokens_seen": 2029846528
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.774835499919756e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.4546139806352048,
      "tokens_seen": 2029977600
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.9942470788955688,
      "objective/train/avg_token_score": 0.005752995144575834,
      "objective/train/docs_used": 740596,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2560597658157349,
      "objective/train/original_loss": 1.2645281553268433,
      "objective/train/theoretical_loss": 3.4545955904577523,
      "objective/train/tokens_used": 400567776,
      "objective/train/unlikelihood_loss": 0.015021015889942646,
      "theoretical_loss": 3.4545955904577523,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.774033060503933e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.4545955904577523,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.773230621088108e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.4545772018000376,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.772428181672285e-05,
      "loss": 1.328,
      "theoretical_loss": 3.4545588146618362,
      "tokens_seen": 2030370816
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.77162574225646e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.454540429042925,
      "tokens_seen": 2030501888
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.770823302840637e-05,
      "loss": 1.2218,
      "theoretical_loss": 3.4545220449430807,
      "tokens_seen": 2030632960
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.770020863424812e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.4545036623620793,
      "tokens_seen": 2030764032
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.769218424008987e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.4544852812996973,
      "tokens_seen": 2030895104
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.768415984593164e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.4544669017557115,
      "tokens_seen": 2031026176
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.76761354517734e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.4544485237298987,
      "tokens_seen": 2031157248
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.766811105761516e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.454430147222035,
      "tokens_seen": 2031288320
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.766008666345691e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.4544117722318965,
      "tokens_seen": 2031419392
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.765206226929868e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.4543933987592617,
      "tokens_seen": 2031550464
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.764403787514043e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.4543750268039064,
      "tokens_seen": 2031681536
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.76360134809822e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.454356656365607,
      "tokens_seen": 2031812608
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.762798908682395e-05,
      "loss": 1.317,
      "theoretical_loss": 3.4543382874441413,
      "tokens_seen": 2031943680
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.76199646926657e-05,
      "loss": 1.286,
      "theoretical_loss": 3.454319920039285,
      "tokens_seen": 2032074752
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.761194029850747e-05,
      "loss": 1.3934,
      "theoretical_loss": 3.4543015541508164,
      "tokens_seen": 2032205824
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.760391590434923e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.4542831897785122,
      "tokens_seen": 2032336896
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.759589151019099e-05,
      "loss": 1.305,
      "theoretical_loss": 3.4542648269221488,
      "tokens_seen": 2032467968
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.758786711603275e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.4542464655815035,
      "tokens_seen": 2032599040
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.757984272187451e-05,
      "loss": 1.3346,
      "theoretical_loss": 3.454228105756354,
      "tokens_seen": 2032730112
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.757181832771627e-05,
      "loss": 1.237,
      "theoretical_loss": 3.4542097474464772,
      "tokens_seen": 2032861184
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.756379393355802e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.4541913906516504,
      "tokens_seen": 2032992256
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.755576953939979e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.4541730353716504,
      "tokens_seen": 2033123328
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.754774514524154e-05,
      "loss": 1.332,
      "theoretical_loss": 3.454154681606256,
      "tokens_seen": 2033254400
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.75397207510833e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.4541363293552427,
      "tokens_seen": 2033385472
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.753169635692506e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.4541179786183887,
      "tokens_seen": 2033516544
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.752367196276682e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.454099629395472,
      "tokens_seen": 2033647616
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.751564756860858e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.4540812816862694,
      "tokens_seen": 2033778688
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.750762317445033e-05,
      "loss": 1.3494,
      "theoretical_loss": 3.454062935490559,
      "tokens_seen": 2033909760
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.74995987802921e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.4540445908081185,
      "tokens_seen": 2034040832
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.749157438613385e-05,
      "loss": 1.2578,
      "theoretical_loss": 3.4540262476387253,
      "tokens_seen": 2034171904
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.748354999197562e-05,
      "loss": 1.3264,
      "theoretical_loss": 3.4540079059821576,
      "tokens_seen": 2034302976
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.747552559781737e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.453989565838192,
      "tokens_seen": 2034434048
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.746750120365914e-05,
      "loss": 1.2949,
      "theoretical_loss": 3.4539712272066074,
      "tokens_seen": 2034565120
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.745947680950089e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.453952890087181,
      "tokens_seen": 2034696192
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.745145241534264e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.453934554479691,
      "tokens_seen": 2034827264
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.744342802118441e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.453916220383916,
      "tokens_seen": 2034958336
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.743540362702616e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.4538978877996325,
      "tokens_seen": 2035089408
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.742737923286793e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.45387955672662,
      "tokens_seen": 2035220480
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.741935483870968e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.4538612271646563,
      "tokens_seen": 2035351552
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.741133044455145e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.453842899113519,
      "tokens_seen": 2035482624
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.74033060503932e-05,
      "loss": 1.258,
      "theoretical_loss": 3.453824572572987,
      "tokens_seen": 2035613696
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.739528165623496e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.4538062475428375,
      "tokens_seen": 2035744768
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.738725726207672e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.4537879240228495,
      "tokens_seen": 2035875840
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.737923286791848e-05,
      "loss": 1.3472,
      "theoretical_loss": 3.4537696020128017,
      "tokens_seen": 2036006912
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.737120847376024e-05,
      "loss": 1.242,
      "theoretical_loss": 3.4537512815124716,
      "tokens_seen": 2036137984
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.7363184079602e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.4537329625216384,
      "tokens_seen": 2036269056
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.735515968544376e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.4537146450400797,
      "tokens_seen": 2036400128
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.734713529128552e-05,
      "loss": 1.3469,
      "theoretical_loss": 3.4536963290675753,
      "tokens_seen": 2036531200
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.9946879148483276,
      "objective/train/avg_token_score": 0.005312114953994751,
      "objective/train/docs_used": 742794,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4194769859313965,
      "objective/train/original_loss": 1.4273796081542969,
      "objective/train/theoretical_loss": 3.4536780146039026,
      "objective/train/tokens_used": 407121376,
      "objective/train/unlikelihood_loss": 0.013317558914422989,
      "theoretical_loss": 3.4536780146039026,
      "tokens_seen": 2036662272
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.733911089712728e-05,
      "loss": 1.3705,
      "theoretical_loss": 3.4536780146039026,
      "tokens_seen": 2036662272
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.733108650296904e-05,
      "loss": 1.4034,
      "theoretical_loss": 3.4536597016488404,
      "tokens_seen": 2036793344
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.732306210881079e-05,
      "loss": 1.283,
      "theoretical_loss": 3.453641390202168,
      "tokens_seen": 2036924416
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.731503771465256e-05,
      "loss": 1.289,
      "theoretical_loss": 3.4536230802636636,
      "tokens_seen": 2037055488
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.730701332049431e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.453604771833106,
      "tokens_seen": 2037186560
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.729898892633608e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.4535864649102743,
      "tokens_seen": 2037317632
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.729096453217783e-05,
      "loss": 1.327,
      "theoretical_loss": 3.4535681594949468,
      "tokens_seen": 2037448704
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.72829401380196e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.4535498555869033,
      "tokens_seen": 2037579776
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.727491574386135e-05,
      "loss": 1.3633,
      "theoretical_loss": 3.453531553185922,
      "tokens_seen": 2037710848
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.72668913497031e-05,
      "loss": 1.2239,
      "theoretical_loss": 3.4535132522917813,
      "tokens_seen": 2037841920
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.725886695554487e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.4534949529042613,
      "tokens_seen": 2037972992
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.725084256138662e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.453476655023141,
      "tokens_seen": 2038104064
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.724281816722839e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.4534583586481995,
      "tokens_seen": 2038235136
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.723479377307014e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.4534400637792153,
      "tokens_seen": 2038366208
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.72267693789119e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.453421770415968,
      "tokens_seen": 2038497280
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.721874498475366e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.4534034785582373,
      "tokens_seen": 2038628352
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.721072059059541e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.453385188205802,
      "tokens_seen": 2038759424
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.720269619643718e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.453366899358442,
      "tokens_seen": 2038890496
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.719467180227893e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.453348612015936,
      "tokens_seen": 2039021568
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.71866474081207e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.4533303261780635,
      "tokens_seen": 2039152640
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.717862301396245e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.4533120418446046,
      "tokens_seen": 2039283712
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.71705986198042e-05,
      "loss": 1.3608,
      "theoretical_loss": 3.453293759015338,
      "tokens_seen": 2039414784
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.716257422564597e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.4532754776900445,
      "tokens_seen": 2039545856
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.715454983148773e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.4532571978685027,
      "tokens_seen": 2039676928
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.714652543732949e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.4532389195504933,
      "tokens_seen": 2039808000
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.713850104317125e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.4532206427357948,
      "tokens_seen": 2039939072
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.7130476649013e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.4532023674241876,
      "tokens_seen": 2040070144
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.712245225485477e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.4531840936154516,
      "tokens_seen": 2040201216
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.711442786069652e-05,
      "loss": 1.3247,
      "theoretical_loss": 3.4531658213093666,
      "tokens_seen": 2040332288
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.710640346653829e-05,
      "loss": 1.218,
      "theoretical_loss": 3.4531475505057125,
      "tokens_seen": 2040463360
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.709837907238004e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.453129281204269,
      "tokens_seen": 2040594432
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.70903546782218e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.4531110134048166,
      "tokens_seen": 2040725504
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.708233028406356e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.4530927471071347,
      "tokens_seen": 2040856576
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.707430588990531e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.453074482311004,
      "tokens_seen": 2040987648
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.706628149574708e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.4530562190162044,
      "tokens_seen": 2041118720
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.705825710158883e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.4530379572225165,
      "tokens_seen": 2041249792
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.70502327074306e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.45301969692972,
      "tokens_seen": 2041380864
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.704220831327235e-05,
      "loss": 1.3043,
      "theoretical_loss": 3.4530014381375955,
      "tokens_seen": 2041511936
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.70341839191141e-05,
      "loss": 1.3808,
      "theoretical_loss": 3.452983180845923,
      "tokens_seen": 2041643008
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.702615952495587e-05,
      "loss": 1.3823,
      "theoretical_loss": 3.4529649250544834,
      "tokens_seen": 2041774080
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.701813513079762e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.4529466707630565,
      "tokens_seen": 2041905152
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.701011073663939e-05,
      "loss": 1.3106,
      "theoretical_loss": 3.4529284179714237,
      "tokens_seen": 2042036224
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.700208634248114e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.4529101666793642,
      "tokens_seen": 2042167296
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.699406194832291e-05,
      "loss": 1.27,
      "theoretical_loss": 3.4528919168866596,
      "tokens_seen": 2042298368
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.698603755416466e-05,
      "loss": 1.3397,
      "theoretical_loss": 3.4528736685930905,
      "tokens_seen": 2042429440
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.697801316000642e-05,
      "loss": 1.3572,
      "theoretical_loss": 3.4528554217984366,
      "tokens_seen": 2042560512
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.696998876584818e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.45283717650248,
      "tokens_seen": 2042691584
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.696196437168994e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.452818932705,
      "tokens_seen": 2042822656
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.69539399775317e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.4528006904057786,
      "tokens_seen": 2042953728
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.694591558337346e-05,
      "loss": 1.3359,
      "theoretical_loss": 3.4527824496045962,
      "tokens_seen": 2043084800
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.9943661689758301,
      "objective/train/avg_token_score": 0.00563386082649231,
      "objective/train/docs_used": 745253,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3212183713912964,
      "objective/train/original_loss": 1.3294398784637451,
      "objective/train/theoretical_loss": 3.4527642103012335,
      "objective/train/tokens_used": 413674976,
      "objective/train/unlikelihood_loss": 0.013941488228738308,
      "theoretical_loss": 3.4527642103012335,
      "tokens_seen": 2043215872
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.693789118921521e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.4527642103012335,
      "tokens_seen": 2043215872
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.692986679505698e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.452745972495472,
      "tokens_seen": 2043346944
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.692184240089873e-05,
      "loss": 1.2449,
      "theoretical_loss": 3.452727736187091,
      "tokens_seen": 2043478016
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.69138180067405e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.4527095013758737,
      "tokens_seen": 2043609088
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.690579361258225e-05,
      "loss": 1.2441,
      "theoretical_loss": 3.4526912680616006,
      "tokens_seen": 2043740160
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.689776921842402e-05,
      "loss": 1.2414,
      "theoretical_loss": 3.452673036244052,
      "tokens_seen": 2043871232
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.688974482426577e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.45265480592301,
      "tokens_seen": 2044002304
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.688172043010752e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.4526365770982546,
      "tokens_seen": 2044133376
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.687369603594929e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.452618349769568,
      "tokens_seen": 2044264448
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.686567164179104e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.4526001239367323,
      "tokens_seen": 2044395520
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.685764724763281e-05,
      "loss": 1.226,
      "theoretical_loss": 3.4525818995995268,
      "tokens_seen": 2044526592
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.684962285347456e-05,
      "loss": 1.3736,
      "theoretical_loss": 3.452563676757735,
      "tokens_seen": 2044657664
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.684159845931632e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.452545455411136,
      "tokens_seen": 2044788736
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.683357406515808e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.452527235559513,
      "tokens_seen": 2044919808
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.682554967099984e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.452509017202648,
      "tokens_seen": 2045050880
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.68175252768416e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.4524908003403207,
      "tokens_seen": 2045181952
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.680950088268335e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.4524725849723144,
      "tokens_seen": 2045313024
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.680147648852512e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.4524543710984097,
      "tokens_seen": 2045444096
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.679345209436687e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.452436158718389,
      "tokens_seen": 2045575168
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.678542770020863e-05,
      "loss": 1.2449,
      "theoretical_loss": 3.452417947832034,
      "tokens_seen": 2045706240
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.67774033060504e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.4523997384391256,
      "tokens_seen": 2045837312
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.676937891189215e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.4523815305394465,
      "tokens_seen": 2045968384
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.676135451773391e-05,
      "loss": 1.366,
      "theoretical_loss": 3.452363324132779,
      "tokens_seen": 2046099456
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.675333012357567e-05,
      "loss": 1.3599,
      "theoretical_loss": 3.452345119218904,
      "tokens_seen": 2046230528
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.674530572941742e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.452326915797604,
      "tokens_seen": 2046361600
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.673728133525919e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.4523087138686614,
      "tokens_seen": 2046492672
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.672925694110094e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.4522905134318576,
      "tokens_seen": 2046623744
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.672123254694271e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.4522723144869754,
      "tokens_seen": 2046754816
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.671320815278446e-05,
      "loss": 1.3793,
      "theoretical_loss": 3.4522541170337964,
      "tokens_seen": 2046885888
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.670518375862623e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.452235921072103,
      "tokens_seen": 2047016960
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.669715936446798e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.4522177266016776,
      "tokens_seen": 2047148032
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.668913497030975e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.452199533622302,
      "tokens_seen": 2047279104
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.66811105761515e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.4521813421337595,
      "tokens_seen": 2047410176
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.667308618199325e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.4521631521358316,
      "tokens_seen": 2047541248
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.666506178783502e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.4521449636283013,
      "tokens_seen": 2047672320
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.665703739367677e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.4521267766109514,
      "tokens_seen": 2047803392
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.664901299951854e-05,
      "loss": 1.3332,
      "theoretical_loss": 3.4521085910835634,
      "tokens_seen": 2047934464
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.664098860536029e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.4520904070459206,
      "tokens_seen": 2048065536
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.663296421120206e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.4520722244978055,
      "tokens_seen": 2048196608
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.662493981704381e-05,
      "loss": 1.3349,
      "theoretical_loss": 3.4520540434390012,
      "tokens_seen": 2048327680
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.661691542288557e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.4520358638692894,
      "tokens_seen": 2048458752
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.660889102872733e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.4520176857884532,
      "tokens_seen": 2048589824
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.660086663456909e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.451999509196276,
      "tokens_seen": 2048720896
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.659284224041085e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.4519813340925407,
      "tokens_seen": 2048851968
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.65848178462526e-05,
      "loss": 1.3397,
      "theoretical_loss": 3.4519631604770296,
      "tokens_seen": 2048983040
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.657679345209437e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.451944988349526,
      "tokens_seen": 2049114112
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.656876905793612e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.4519268177098126,
      "tokens_seen": 2049245184
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.656074466377788e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.451908648557673,
      "tokens_seen": 2049376256
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.655272026961964e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.451890480892889,
      "tokens_seen": 2049507328
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.65446958754614e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.4518723147152457,
      "tokens_seen": 2049638400
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.9918603897094727,
      "objective/train/avg_token_score": 0.008139638230204582,
      "objective/train/docs_used": 747702,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3274792432785034,
      "objective/train/original_loss": 1.338588833808899,
      "objective/train/theoretical_loss": 3.4518541500245243,
      "objective/train/tokens_used": 420228576,
      "objective/train/unlikelihood_loss": 0.022673005238175392,
      "theoretical_loss": 3.4518541500245243,
      "tokens_seen": 2049769472
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.653667148130316e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.4518541500245243,
      "tokens_seen": 2049769472
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.652864708714492e-05,
      "loss": 1.281,
      "theoretical_loss": 3.4518359868205097,
      "tokens_seen": 2049900544
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.652062269298668e-05,
      "loss": 1.268,
      "theoretical_loss": 3.4518178251029843,
      "tokens_seen": 2050031616
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.651259829882844e-05,
      "loss": 1.363,
      "theoretical_loss": 3.4517996648717313,
      "tokens_seen": 2050162688
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.650457390467019e-05,
      "loss": 1.4074,
      "theoretical_loss": 3.4517815061265344,
      "tokens_seen": 2050293760
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.649654951051196e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.4517633488671766,
      "tokens_seen": 2050424832
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.648852511635371e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.4517451930934424,
      "tokens_seen": 2050555904
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.648050072219548e-05,
      "loss": 1.3394,
      "theoretical_loss": 3.451727038805114,
      "tokens_seen": 2050686976
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.647247632803723e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.451708886001976,
      "tokens_seen": 2050818048
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.6464451933879e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.451690734683811,
      "tokens_seen": 2050949120
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.645642753972075e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.451672584850403,
      "tokens_seen": 2051080192
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.64484031455625e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.451654436501536,
      "tokens_seen": 2051211264
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.644037875140427e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.451636289636993,
      "tokens_seen": 2051342336
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.643235435724602e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.4516181442565594,
      "tokens_seen": 2051473408
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.642432996308779e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.451600000360017,
      "tokens_seen": 2051604480
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.641630556892954e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.451581857947151,
      "tokens_seen": 2051735552
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.640828117477131e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.451563717017745,
      "tokens_seen": 2051866624
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.640025678061306e-05,
      "loss": 1.305,
      "theoretical_loss": 3.4515455775715824,
      "tokens_seen": 2051997696
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.639223238645483e-05,
      "loss": 1.3523,
      "theoretical_loss": 3.451527439608448,
      "tokens_seen": 2052128768
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.638420799229658e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.451509303128125,
      "tokens_seen": 2052259840
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.637618359813834e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.4514911681303984,
      "tokens_seen": 2052390912
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.63681592039801e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.4514730346150517,
      "tokens_seen": 2052521984
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.636013480982186e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.4514549025818693,
      "tokens_seen": 2052653056
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.635211041566362e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.451436772030635,
      "tokens_seen": 2052784128
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.634408602150538e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.451418642961134,
      "tokens_seen": 2052915200
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.633606162734714e-05,
      "loss": 1.3106,
      "theoretical_loss": 3.45140051537315,
      "tokens_seen": 2053046272
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.63280372331889e-05,
      "loss": 1.3515,
      "theoretical_loss": 3.451382389266467,
      "tokens_seen": 2053177344
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.632001283903065e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.45136426464087,
      "tokens_seen": 2053308416
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.631198844487241e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.4513461414961433,
      "tokens_seen": 2053439488
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.630396405071417e-05,
      "loss": 1.3525,
      "theoretical_loss": 3.4513280198320713,
      "tokens_seen": 2053570560
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.629593965655593e-05,
      "loss": 1.3312,
      "theoretical_loss": 3.4513098996484386,
      "tokens_seen": 2053701632
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.628791526239769e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.45129178094503,
      "tokens_seen": 2053832704
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.627989086823945e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.4512736637216292,
      "tokens_seen": 2053963776
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.627186647408121e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.4512555479780223,
      "tokens_seen": 2054094848
    },
    {
      "epoch": 0.24,
      "learning_rate": 7.626384207992296e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.451237433713993,
      "tokens_seen": 2054225920
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.625581768576473e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.4512193209293267,
      "tokens_seen": 2054356992
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.624779329160648e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.4512012096238074,
      "tokens_seen": 2054488064
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.623976889744825e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.4511830997972206,
      "tokens_seen": 2054619136
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.623174450329e-05,
      "loss": 1.298,
      "theoretical_loss": 3.4511649914493505,
      "tokens_seen": 2054750208
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.622372010913177e-05,
      "loss": 1.4195,
      "theoretical_loss": 3.451146884579983,
      "tokens_seen": 2054881280
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.621569571497352e-05,
      "loss": 1.3404,
      "theoretical_loss": 3.451128779188903,
      "tokens_seen": 2055012352
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.620767132081527e-05,
      "loss": 1.3593,
      "theoretical_loss": 3.4511106752758947,
      "tokens_seen": 2055143424
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.619964692665704e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.4510925728407438,
      "tokens_seen": 2055274496
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.619162253249879e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.451074471883236,
      "tokens_seen": 2055405568
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.618359813834056e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.4510563724031544,
      "tokens_seen": 2055536640
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.617557374418231e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.451038274400286,
      "tokens_seen": 2055667712
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.616754935002408e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4510201778744163,
      "tokens_seen": 2055798784
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.615952495586583e-05,
      "loss": 1.337,
      "theoretical_loss": 3.451002082825329,
      "tokens_seen": 2055929856
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.615150056170759e-05,
      "loss": 1.3564,
      "theoretical_loss": 3.450983989252811,
      "tokens_seen": 2056060928
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.614347616754935e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.450965897156647,
      "tokens_seen": 2056192000
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.9949284791946411,
      "objective/train/avg_token_score": 0.005071508698165417,
      "objective/train/docs_used": 750166,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2640128135681152,
      "objective/train/original_loss": 1.2708607912063599,
      "objective/train/theoretical_loss": 3.450947806536622,
      "objective/train/tokens_used": 426782176,
      "objective/train/unlikelihood_loss": 0.015937235206365585,
      "theoretical_loss": 3.450947806536622,
      "tokens_seen": 2056323072
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.61354517733911e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.450947806536622,
      "tokens_seen": 2056323072
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.612742737923287e-05,
      "loss": 1.3698,
      "theoretical_loss": 3.4509297173925226,
      "tokens_seen": 2056454144
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.611940298507463e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.4509116297241333,
      "tokens_seen": 2056585216
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.611137859091639e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.4508935435312402,
      "tokens_seen": 2056716288
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.610335419675815e-05,
      "loss": 1.252,
      "theoretical_loss": 3.450875458813629,
      "tokens_seen": 2056847360
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.609532980259991e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.450857375571085,
      "tokens_seen": 2056978432
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.608730540844167e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.4508392938033943,
      "tokens_seen": 2057109504
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.607928101428342e-05,
      "loss": 1.3466,
      "theoretical_loss": 3.4508212135103418,
      "tokens_seen": 2057240576
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.607125662012518e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.4508031346917143,
      "tokens_seen": 2057371648
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.606323222596694e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.4507850573472973,
      "tokens_seen": 2057502720
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.60552078318087e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.4507669814768764,
      "tokens_seen": 2057633792
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.604718343765046e-05,
      "loss": 1.3418,
      "theoretical_loss": 3.450748907080238,
      "tokens_seen": 2057764864
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.603915904349222e-05,
      "loss": 1.2313,
      "theoretical_loss": 3.4507308341571674,
      "tokens_seen": 2057895936
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.603113464933398e-05,
      "loss": 1.3547,
      "theoretical_loss": 3.450712762707452,
      "tokens_seen": 2058027008
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.602311025517573e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.450694692730876,
      "tokens_seen": 2058158080
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.60150858610175e-05,
      "loss": 1.3588,
      "theoretical_loss": 3.450676624227227,
      "tokens_seen": 2058289152
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.600706146685925e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.45065855719629,
      "tokens_seen": 2058420224
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.599903707270102e-05,
      "loss": 1.272,
      "theoretical_loss": 3.4506404916378526,
      "tokens_seen": 2058551296
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.599101267854277e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.4506224275517,
      "tokens_seen": 2058682368
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.598298828438454e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.450604364937618,
      "tokens_seen": 2058813440
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.597496389022629e-05,
      "loss": 1.3804,
      "theoretical_loss": 3.4505863037953945,
      "tokens_seen": 2058944512
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.596693949606804e-05,
      "loss": 1.2311,
      "theoretical_loss": 3.4505682441248147,
      "tokens_seen": 2059075584
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.595891510190981e-05,
      "loss": 1.1973,
      "theoretical_loss": 3.4505501859256653,
      "tokens_seen": 2059206656
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.595089070775156e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.450532129197733,
      "tokens_seen": 2059337728
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.594286631359333e-05,
      "loss": 1.306,
      "theoretical_loss": 3.450514073940804,
      "tokens_seen": 2059468800
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.593484191943508e-05,
      "loss": 1.3402,
      "theoretical_loss": 3.450496020154665,
      "tokens_seen": 2059599872
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.592681752527685e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.4504779678391024,
      "tokens_seen": 2059730944
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.59187931311186e-05,
      "loss": 1.3542,
      "theoretical_loss": 3.4504599169939034,
      "tokens_seen": 2059862016
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.591076873696036e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.4504418676188537,
      "tokens_seen": 2059993088
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.590274434280212e-05,
      "loss": 1.2578,
      "theoretical_loss": 3.450423819713741,
      "tokens_seen": 2060124160
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.589471994864388e-05,
      "loss": 1.277,
      "theoretical_loss": 3.450405773278352,
      "tokens_seen": 2060255232
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.588669555448564e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.4503877283124726,
      "tokens_seen": 2060386304
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.58786711603274e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.450369684815891,
      "tokens_seen": 2060517376
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.587064676616916e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.450351642788393,
      "tokens_seen": 2060648448
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.586262237201092e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.450333602229766,
      "tokens_seen": 2060779520
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.585459797785267e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.4503155631397977,
      "tokens_seen": 2060910592
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.584657358369444e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.4502975255182733,
      "tokens_seen": 2061041664
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.583854918953619e-05,
      "loss": 1.3403,
      "theoretical_loss": 3.4502794893649815,
      "tokens_seen": 2061172736
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.583052479537795e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.4502614546797092,
      "tokens_seen": 2061303808
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.582250040121971e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.4502434214622433,
      "tokens_seen": 2061434880
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.581447600706147e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.4502253897123705,
      "tokens_seen": 2061565952
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.580645161290323e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.4502073594298794,
      "tokens_seen": 2061697024
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.579842721874498e-05,
      "loss": 1.3064,
      "theoretical_loss": 3.450189330614556,
      "tokens_seen": 2061828096
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.579040282458675e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.4501713032661883,
      "tokens_seen": 2061959168
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.57823784304285e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.4501532773845636,
      "tokens_seen": 2062090240
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.577435403627027e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.4501352529694693,
      "tokens_seen": 2062221312
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.576632964211202e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.450117230020693,
      "tokens_seen": 2062352384
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.575830524795379e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.450099208538022,
      "tokens_seen": 2062483456
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.575028085379554e-05,
      "loss": 1.244,
      "theoretical_loss": 3.450081188521244,
      "tokens_seen": 2062614528
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.574225645963731e-05,
      "loss": 1.2172,
      "theoretical_loss": 3.4500631699701465,
      "tokens_seen": 2062745600
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.990867018699646,
      "objective/train/avg_token_score": 0.009132950566709042,
      "objective/train/docs_used": 752534,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2888686656951904,
      "objective/train/original_loss": 1.3014006614685059,
      "objective/train/theoretical_loss": 3.4500451528845177,
      "objective/train/tokens_used": 433335776,
      "objective/train/unlikelihood_loss": 0.02659301832318306,
      "theoretical_loss": 3.4500451528845177,
      "tokens_seen": 2062876672
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.573423206547906e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.4500451528845177,
      "tokens_seen": 2062876672
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.572620767132081e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.450027137264145,
      "tokens_seen": 2063007744
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.571818327716258e-05,
      "loss": 1.3448,
      "theoretical_loss": 3.4500091231088157,
      "tokens_seen": 2063138816
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.571015888300433e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.449991110418318,
      "tokens_seen": 2063269888
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.57021344888461e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.44997309919244,
      "tokens_seen": 2063400960
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.569411009468785e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.4499550894309694,
      "tokens_seen": 2063532032
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.568608570052962e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.449937081133694,
      "tokens_seen": 2063663104
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.567806130637137e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.449919074300401,
      "tokens_seen": 2063794176
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.567003691221313e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.4499010689308798,
      "tokens_seen": 2063925248
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.566201251805489e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.4498830650249177,
      "tokens_seen": 2064056320
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.565398812389665e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.4498650625823033,
      "tokens_seen": 2064187392
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.564596372973841e-05,
      "loss": 1.3705,
      "theoretical_loss": 3.4498470616028243,
      "tokens_seen": 2064318464
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.563793933558017e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.449829062086269,
      "tokens_seen": 2064449536
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.562991494142193e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.449811064032426,
      "tokens_seen": 2064580608
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.562189054726369e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.4497930674410826,
      "tokens_seen": 2064711680
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.561386615310544e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.449775072312028,
      "tokens_seen": 2064842752
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.56058417589472e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.449757078645051,
      "tokens_seen": 2064973824
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.559781736478896e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.449739086439939,
      "tokens_seen": 2065104896
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.558979297063073e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.4497210956964803,
      "tokens_seen": 2065235968
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.558176857647248e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.4497031064144648,
      "tokens_seen": 2065367040
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.557374418231424e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.44968511859368,
      "tokens_seen": 2065498112
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.5565719788156e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.4496671322339143,
      "tokens_seen": 2065629184
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.555769539399775e-05,
      "loss": 1.266,
      "theoretical_loss": 3.4496491473349566,
      "tokens_seen": 2065760256
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.554967099983952e-05,
      "loss": 1.3851,
      "theoretical_loss": 3.449631163896596,
      "tokens_seen": 2065891328
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.554164660568127e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.4496131819186213,
      "tokens_seen": 2066022400
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.553362221152304e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.4495952014008204,
      "tokens_seen": 2066153472
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.552559781736479e-05,
      "loss": 1.232,
      "theoretical_loss": 3.449577222342983,
      "tokens_seen": 2066284544
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.551757342320656e-05,
      "loss": 1.2931,
      "theoretical_loss": 3.449559244744897,
      "tokens_seen": 2066415616
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.550954902904831e-05,
      "loss": 1.3726,
      "theoretical_loss": 3.4495412686063522,
      "tokens_seen": 2066546688
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.550152463489006e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.449523293927137,
      "tokens_seen": 2066677760
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.549350024073183e-05,
      "loss": 1.237,
      "theoretical_loss": 3.449505320707041,
      "tokens_seen": 2066808832
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.548547584657358e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.449487348945853,
      "tokens_seen": 2066939904
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.547745145241535e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.4494693786433612,
      "tokens_seen": 2067070976
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.54694270582571e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.449451409799356,
      "tokens_seen": 2067202048
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.546140266409887e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.4494334424136257,
      "tokens_seen": 2067333120
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.545337826994062e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.4494154764859593,
      "tokens_seen": 2067464192
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.544535387578239e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.449397512016147,
      "tokens_seen": 2067595264
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.543732948162414e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.449379549003978,
      "tokens_seen": 2067726336
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.54293050874659e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.449361587449241,
      "tokens_seen": 2067857408
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.542128069330766e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.4493436273517255,
      "tokens_seen": 2067988480
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.541325629914942e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.4493256687112215,
      "tokens_seen": 2068119552
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.540523190499118e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4493077115275175,
      "tokens_seen": 2068250624
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.539720751083294e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.449289755800404,
      "tokens_seen": 2068381696
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.53891831166747e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.4492718015296697,
      "tokens_seen": 2068512768
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.538115872251646e-05,
      "loss": 1.381,
      "theoretical_loss": 3.449253848715105,
      "tokens_seen": 2068643840
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.537313432835821e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.4492358973564987,
      "tokens_seen": 2068774912
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.536510993419998e-05,
      "loss": 1.3227,
      "theoretical_loss": 3.4492179474536413,
      "tokens_seen": 2068905984
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.535708554004173e-05,
      "loss": 1.3482,
      "theoretical_loss": 3.4491999990063222,
      "tokens_seen": 2069037056
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.53490611458835e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.449182052014331,
      "tokens_seen": 2069168128
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.534103675172525e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.4491641064774576,
      "tokens_seen": 2069299200
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.989462673664093,
      "objective/train/avg_token_score": 0.010537360794842243,
      "objective/train/docs_used": 754612,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3068878650665283,
      "objective/train/original_loss": 1.3221567869186401,
      "objective/train/theoretical_loss": 3.4491461623954924,
      "objective/train/tokens_used": 439889376,
      "objective/train/unlikelihood_loss": 0.025783000513911247,
      "theoretical_loss": 3.4491461623954924,
      "tokens_seen": 2069430272
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.533301235756701e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.4491461623954924,
      "tokens_seen": 2069430272
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.532498796340877e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.449128219768225,
      "tokens_seen": 2069561344
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.531696356925052e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.4491102785954446,
      "tokens_seen": 2069692416
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.530893917509229e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.449092338876942,
      "tokens_seen": 2069823488
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.530091478093404e-05,
      "loss": 1.3369,
      "theoretical_loss": 3.449074400612507,
      "tokens_seen": 2069954560
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.529289038677581e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.44905646380193,
      "tokens_seen": 2070085632
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.528486599261756e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.4490385284450005,
      "tokens_seen": 2070216704
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.527684159845933e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.449020594541509,
      "tokens_seen": 2070347776
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.526881720430108e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.449002662091246,
      "tokens_seen": 2070478848
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.526079281014283e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.448984731094001,
      "tokens_seen": 2070609920
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.52527684159846e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.448966801549565,
      "tokens_seen": 2070740992
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.524474402182635e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.448948873457728,
      "tokens_seen": 2070872064
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.523671962766812e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.4489309468182814,
      "tokens_seen": 2071003136
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.522869523350987e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.4489130216310135,
      "tokens_seen": 2071134208
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.522067083935164e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.4488950978957167,
      "tokens_seen": 2071265280
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.521264644519339e-05,
      "loss": 1.359,
      "theoretical_loss": 3.448877175612181,
      "tokens_seen": 2071396352
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.520462205103515e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.4488592547801966,
      "tokens_seen": 2071527424
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.519659765687691e-05,
      "loss": 1.3261,
      "theoretical_loss": 3.448841335399554,
      "tokens_seen": 2071658496
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.518857326271867e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.4488234174700443,
      "tokens_seen": 2071789568
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.518054886856043e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.448805500991458,
      "tokens_seen": 2071920640
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.517252447440219e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.4487875859635864,
      "tokens_seen": 2072051712
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.516450008024395e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.448769672386219,
      "tokens_seen": 2072182784
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.51564756860857e-05,
      "loss": 1.2194,
      "theoretical_loss": 3.4487517602591478,
      "tokens_seen": 2072313856
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.514845129192747e-05,
      "loss": 1.3627,
      "theoretical_loss": 3.4487338495821627,
      "tokens_seen": 2072444928
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.514042689776923e-05,
      "loss": 1.3418,
      "theoretical_loss": 3.4487159403550556,
      "tokens_seen": 2072576000
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.513240250361098e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.448698032577617,
      "tokens_seen": 2072707072
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.512437810945275e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.448680126249638,
      "tokens_seen": 2072838144
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.51163537152945e-05,
      "loss": 1.3931,
      "theoretical_loss": 3.448662221370909,
      "tokens_seen": 2072969216
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.510832932113627e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.4486443179412216,
      "tokens_seen": 2073100288
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.510030492697802e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.4486264159603675,
      "tokens_seen": 2073231360
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.509228053281979e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.4486085154281367,
      "tokens_seen": 2073362432
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.508425613866154e-05,
      "loss": 1.253,
      "theoretical_loss": 3.4485906163443216,
      "tokens_seen": 2073493504
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.507623174450329e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.4485727187087125,
      "tokens_seen": 2073624576
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.506820735034506e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.4485548225211007,
      "tokens_seen": 2073755648
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.506018295618681e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.4485369277812783,
      "tokens_seen": 2073886720
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.505215856202858e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.4485190344890366,
      "tokens_seen": 2074017792
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.504413416787033e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.448501142644166,
      "tokens_seen": 2074148864
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.50361097737121e-05,
      "loss": 1.296,
      "theoretical_loss": 3.4484832522464597,
      "tokens_seen": 2074279936
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.502808537955385e-05,
      "loss": 1.3422,
      "theoretical_loss": 3.448465363295707,
      "tokens_seen": 2074411008
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.50200609853956e-05,
      "loss": 1.3582,
      "theoretical_loss": 3.4484474757917014,
      "tokens_seen": 2074542080
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.501203659123737e-05,
      "loss": 1.3508,
      "theoretical_loss": 3.4484295897342334,
      "tokens_seen": 2074673152
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.500401219707912e-05,
      "loss": 1.2055,
      "theoretical_loss": 3.448411705123095,
      "tokens_seen": 2074804224
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.499598780292089e-05,
      "loss": 1.3495,
      "theoretical_loss": 3.4483938219580783,
      "tokens_seen": 2074935296
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.498796340876264e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.448375940238974,
      "tokens_seen": 2075066368
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.497993901460441e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.448358059965575,
      "tokens_seen": 2075197440
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.497191462044616e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.4483401811376724,
      "tokens_seen": 2075328512
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.496389022628792e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.4483223037550585,
      "tokens_seen": 2075459584
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.495586583212968e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.4483044278175248,
      "tokens_seen": 2075590656
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.494784143797144e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.448286553324863,
      "tokens_seen": 2075721728
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.49398170438132e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.4482686802768656,
      "tokens_seen": 2075852800
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.990453839302063,
      "objective/train/avg_token_score": 0.009546210989356041,
      "objective/train/docs_used": 756913,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1569746732711792,
      "objective/train/original_loss": 1.1686919927597046,
      "objective/train/theoretical_loss": 3.4482508086733246,
      "objective/train/tokens_used": 446442976,
      "objective/train/unlikelihood_loss": 0.032265596091747284,
      "theoretical_loss": 3.4482508086733246,
      "tokens_seen": 2075983872
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.493179264965496e-05,
      "loss": 1.2128,
      "theoretical_loss": 3.4482508086733246,
      "tokens_seen": 2075983872
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.492376825549672e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.4482329385140327,
      "tokens_seen": 2076114944
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.491574386133848e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.4482150697987803,
      "tokens_seen": 2076246016
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.490771946718023e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.448197202527361,
      "tokens_seen": 2076377088
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.4899695073022e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.4481793366995666,
      "tokens_seen": 2076508160
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.489167067886375e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.4481614723151894,
      "tokens_seen": 2076639232
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.488364628470552e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.4481436093740214,
      "tokens_seen": 2076770304
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.487562189054727e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.4481257478758556,
      "tokens_seen": 2076901376
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.486759749638904e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.448107887820483,
      "tokens_seen": 2077032448
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.485957310223079e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.4480900292076977,
      "tokens_seen": 2077163520
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.485154870807254e-05,
      "loss": 1.235,
      "theoretical_loss": 3.4480721720372913,
      "tokens_seen": 2077294592
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.484352431391431e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.4480543163090562,
      "tokens_seen": 2077425664
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.483549991975606e-05,
      "loss": 1.257,
      "theoretical_loss": 3.4480364620227855,
      "tokens_seen": 2077556736
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.482747552559783e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.448018609178271,
      "tokens_seen": 2077687808
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.481945113143958e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.448000757775306,
      "tokens_seen": 2077818880
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.481142673728135e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.447982907813683,
      "tokens_seen": 2077949952
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.48034023431231e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.4479650592931947,
      "tokens_seen": 2078081024
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.479537794896487e-05,
      "loss": 1.3314,
      "theoretical_loss": 3.447947212213634,
      "tokens_seen": 2078212096
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.478735355480662e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.447929366574793,
      "tokens_seen": 2078343168
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.477932916064837e-05,
      "loss": 1.293,
      "theoretical_loss": 3.4479115223764656,
      "tokens_seen": 2078474240
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.477130476649014e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.447893679618444,
      "tokens_seen": 2078605312
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.47632803723319e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.4478758383005212,
      "tokens_seen": 2078736384
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.475525597817366e-05,
      "loss": 1.2449,
      "theoretical_loss": 3.4478579984224904,
      "tokens_seen": 2078867456
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.474723158401541e-05,
      "loss": 1.3087,
      "theoretical_loss": 3.4478401599841444,
      "tokens_seen": 2078998528
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.473920718985718e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.4478223229852767,
      "tokens_seen": 2079129600
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.473118279569893e-05,
      "loss": 1.3496,
      "theoretical_loss": 3.44780448742568,
      "tokens_seen": 2079260672
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.472315840154069e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.447786653305147,
      "tokens_seen": 2079391744
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.471513400738245e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.4477688206234722,
      "tokens_seen": 2079522816
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.47071096132242e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.4477509893804474,
      "tokens_seen": 2079653888
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.469908521906597e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.4477331595758667,
      "tokens_seen": 2079784960
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.469106082490773e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.4477153312095234,
      "tokens_seen": 2079916032
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.468303643074949e-05,
      "loss": 1.3106,
      "theoretical_loss": 3.4476975042812104,
      "tokens_seen": 2080047104
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.467501203659125e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.447679678790722,
      "tokens_seen": 2080178176
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.4666987642433e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.4476618547378504,
      "tokens_seen": 2080309248
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.465896324827477e-05,
      "loss": 1.3193,
      "theoretical_loss": 3.44764403212239,
      "tokens_seen": 2080440320
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.465093885411652e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.447626210944134,
      "tokens_seen": 2080571392
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.464291445995829e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.447608391202876,
      "tokens_seen": 2080702464
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.463489006580004e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.44759057289841,
      "tokens_seen": 2080833536
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.46268656716418e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.447572756030529,
      "tokens_seen": 2080964608
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.461884127748356e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.4475549405990273,
      "tokens_seen": 2081095680
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.461081688332531e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.447537126603698,
      "tokens_seen": 2081226752
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.460279248916708e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.447519314044335,
      "tokens_seen": 2081357824
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.459476809500883e-05,
      "loss": 1.3409,
      "theoretical_loss": 3.4475015029207334,
      "tokens_seen": 2081488896
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.45867437008506e-05,
      "loss": 1.2645,
      "theoretical_loss": 3.4474836932326856,
      "tokens_seen": 2081619968
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.457871930669235e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.4474658849799855,
      "tokens_seen": 2081751040
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.457069491253412e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.447448078162428,
      "tokens_seen": 2081882112
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.456267051837587e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.4474302727798065,
      "tokens_seen": 2082013184
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.455464612421762e-05,
      "loss": 1.3325,
      "theoretical_loss": 3.4474124688319154,
      "tokens_seen": 2082144256
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.454662173005939e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.4473946663185484,
      "tokens_seen": 2082275328
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.453859733590114e-05,
      "loss": 1.3446,
      "theoretical_loss": 3.4473768652395,
      "tokens_seen": 2082406400
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.9873658418655396,
      "objective/train/avg_token_score": 0.01263416837900877,
      "objective/train/docs_used": 759184,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2322478294372559,
      "objective/train/original_loss": 1.2462577819824219,
      "objective/train/theoretical_loss": 3.447359065594564,
      "objective/train/tokens_used": 452996576,
      "objective/train/unlikelihood_loss": 0.04016647860407829,
      "theoretical_loss": 3.447359065594564,
      "tokens_seen": 2082537472
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.453057294174291e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.447359065594564,
      "tokens_seen": 2082537472
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.452254854758466e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.4473412673835346,
      "tokens_seen": 2082668544
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.451452415342643e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.4473234706062064,
      "tokens_seen": 2082799616
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.450649975926818e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.447305675262373,
      "tokens_seen": 2082930688
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.449847536510994e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.44728788135183,
      "tokens_seen": 2083061760
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.44904509709517e-05,
      "loss": 1.329,
      "theoretical_loss": 3.447270088874371,
      "tokens_seen": 2083192832
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.448242657679346e-05,
      "loss": 1.3746,
      "theoretical_loss": 3.4472522978297904,
      "tokens_seen": 2083323904
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.447440218263522e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.447234508217883,
      "tokens_seen": 2083454976
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.446637778847698e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.4472167200384423,
      "tokens_seen": 2083586048
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.445835339431873e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.447198933291264,
      "tokens_seen": 2083717120
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.44503290001605e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.447181147976142,
      "tokens_seen": 2083848192
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.444230460600225e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.447163364092872,
      "tokens_seen": 2083979264
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.443428021184402e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.4471455816412475,
      "tokens_seen": 2084110336
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.442625581768577e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.4471278006210637,
      "tokens_seen": 2084241408
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.441823142352754e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.447110021032115,
      "tokens_seen": 2084372480
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.441020702936929e-05,
      "loss": 1.293,
      "theoretical_loss": 3.447092242874197,
      "tokens_seen": 2084503552
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.440218263521104e-05,
      "loss": 1.3682,
      "theoretical_loss": 3.4470744661471038,
      "tokens_seen": 2084634624
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.439415824105281e-05,
      "loss": 1.25,
      "theoretical_loss": 3.447056690850631,
      "tokens_seen": 2084765696
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.438613384689456e-05,
      "loss": 1.354,
      "theoretical_loss": 3.447038916984572,
      "tokens_seen": 2084896768
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.437810945273633e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.4470211445487235,
      "tokens_seen": 2085027840
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.437008505857808e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.44700337354288,
      "tokens_seen": 2085158912
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.436206066441983e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.4469856039668363,
      "tokens_seen": 2085289984
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.43540362702616e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.446967835820388,
      "tokens_seen": 2085421056
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.434601187610335e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.4469500691033295,
      "tokens_seen": 2085552128
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.433798748194512e-05,
      "loss": 1.3756,
      "theoretical_loss": 3.4469323038154567,
      "tokens_seen": 2085683200
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.432996308778687e-05,
      "loss": 1.3548,
      "theoretical_loss": 3.4469145399565644,
      "tokens_seen": 2085814272
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.432193869362864e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.4468967775264474,
      "tokens_seen": 2085945344
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.43139142994704e-05,
      "loss": 1.342,
      "theoretical_loss": 3.4468790165249024,
      "tokens_seen": 2086076416
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.430588990531215e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.4468612569517236,
      "tokens_seen": 2086207488
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.429786551115391e-05,
      "loss": 1.3796,
      "theoretical_loss": 3.4468434988067065,
      "tokens_seen": 2086338560
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.428984111699567e-05,
      "loss": 1.3755,
      "theoretical_loss": 3.446825742089647,
      "tokens_seen": 2086469632
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.428181672283743e-05,
      "loss": 1.255,
      "theoretical_loss": 3.4468079868003407,
      "tokens_seen": 2086600704
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.427379232867919e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.446790232938582,
      "tokens_seen": 2086731776
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.426576793452094e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.446772480504168,
      "tokens_seen": 2086862848
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.425774354036271e-05,
      "loss": 1.3423,
      "theoretical_loss": 3.4467547294968934,
      "tokens_seen": 2086993920
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.424971914620446e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.446736979916554,
      "tokens_seen": 2087124992
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.424169475204623e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.4467192317629456,
      "tokens_seen": 2087256064
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.423367035788798e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.446701485035864,
      "tokens_seen": 2087387136
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.422564596372975e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.4466837397351044,
      "tokens_seen": 2087518208
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.42176215695715e-05,
      "loss": 1.3432,
      "theoretical_loss": 3.4466659958604633,
      "tokens_seen": 2087649280
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.420959717541325e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.446648253411736,
      "tokens_seen": 2087780352
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.420157278125502e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.4466305123887193,
      "tokens_seen": 2087911424
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.419354838709677e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.446612772791208,
      "tokens_seen": 2088042496
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.418552399293854e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.4465950346189986,
      "tokens_seen": 2088173568
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.417749959878029e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.4465772978718876,
      "tokens_seen": 2088304640
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.416947520462205e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.4465595625496706,
      "tokens_seen": 2088435712
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.416145081046381e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.4465418286521436,
      "tokens_seen": 2088566784
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.415342641630557e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.446524096179103,
      "tokens_seen": 2088697856
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.414540202214733e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.446506365130345,
      "tokens_seen": 2088828928
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.413737762798909e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.4464886355056654,
      "tokens_seen": 2088960000
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.9929499626159668,
      "objective/train/avg_token_score": 0.007049982901662588,
      "objective/train/docs_used": 761616,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2772737741470337,
      "objective/train/original_loss": 1.2876113653182983,
      "objective/train/theoretical_loss": 3.446470907304861,
      "objective/train/tokens_used": 459550176,
      "objective/train/unlikelihood_loss": 0.018986672163009644,
      "theoretical_loss": 3.446470907304861,
      "tokens_seen": 2089091072
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.412935323383084e-05,
      "loss": 1.294,
      "theoretical_loss": 3.446470907304861,
      "tokens_seen": 2089091072
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.41213288396726e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.446453180527728,
      "tokens_seen": 2089222144
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.411330444551436e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.446435455174063,
      "tokens_seen": 2089353216
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.410528005135612e-05,
      "loss": 1.3211,
      "theoretical_loss": 3.4464177312436615,
      "tokens_seen": 2089484288
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.409725565719788e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.446400008736321,
      "tokens_seen": 2089615360
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.408923126303964e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.4463822876518377,
      "tokens_seen": 2089746432
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.40812068688814e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.446364567990008,
      "tokens_seen": 2089877504
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.407318247472315e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.446346849750628,
      "tokens_seen": 2090008576
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.406515808056492e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.446329132933495,
      "tokens_seen": 2090139648
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.405713368640667e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.4463114175384053,
      "tokens_seen": 2090270720
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.404910929224844e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.446293703565156,
      "tokens_seen": 2090401792
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.404108489809019e-05,
      "loss": 1.321,
      "theoretical_loss": 3.4462759910135436,
      "tokens_seen": 2090532864
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.403306050393196e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.4462582798833648,
      "tokens_seen": 2090663936
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.402503610977371e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.4462405701744165,
      "tokens_seen": 2090795008
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.401701171561546e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.4462228618864956,
      "tokens_seen": 2090926080
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.400898732145723e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.446205155019399,
      "tokens_seen": 2091057152
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.400096292729898e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.4461874495729234,
      "tokens_seen": 2091188224
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.399293853314075e-05,
      "loss": 1.3319,
      "theoretical_loss": 3.4461697455468663,
      "tokens_seen": 2091319296
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.39849141389825e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.446152042941024,
      "tokens_seen": 2091450368
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.397688974482427e-05,
      "loss": 1.4021,
      "theoretical_loss": 3.4461343417551946,
      "tokens_seen": 2091581440
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.396886535066602e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.4461166419891747,
      "tokens_seen": 2091712512
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.396084095650778e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.446098943642761,
      "tokens_seen": 2091843584
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.395281656234954e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.4460812467157513,
      "tokens_seen": 2091974656
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.39447921681913e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.446063551207942,
      "tokens_seen": 2092105728
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.393676777403306e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.446045857119132,
      "tokens_seen": 2092236800
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.392874337987482e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.446028164449117,
      "tokens_seen": 2092367872
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.392071898571658e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.4460104731976955,
      "tokens_seen": 2092498944
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.391269459155834e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.445992783364664,
      "tokens_seen": 2092630016
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.390467019740009e-05,
      "loss": 1.3298,
      "theoretical_loss": 3.4459750949498207,
      "tokens_seen": 2092761088
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.389664580324186e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.445957407952963,
      "tokens_seen": 2092892160
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.388862140908361e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.4459397223738875,
      "tokens_seen": 2093023232
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.388059701492537e-05,
      "loss": 1.3097,
      "theoretical_loss": 3.4459220382123927,
      "tokens_seen": 2093154304
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.387257262076713e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.4459043554682762,
      "tokens_seen": 2093285376
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.38645482266089e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.4458866741413354,
      "tokens_seen": 2093416448
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.385652383245065e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.4458689942313683,
      "tokens_seen": 2093547520
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.384849943829241e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.445851315738172,
      "tokens_seen": 2093678592
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.384047504413417e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.445833638661545,
      "tokens_seen": 2093809664
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.383245064997592e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.4458159630012846,
      "tokens_seen": 2093940736
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.382442625581769e-05,
      "loss": 1.2448,
      "theoretical_loss": 3.4457982887571887,
      "tokens_seen": 2094071808
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.381640186165944e-05,
      "loss": 1.3677,
      "theoretical_loss": 3.445780615929056,
      "tokens_seen": 2094202880
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.380837746750121e-05,
      "loss": 1.2222,
      "theoretical_loss": 3.4457629445166833,
      "tokens_seen": 2094333952
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.380035307334296e-05,
      "loss": 1.3163,
      "theoretical_loss": 3.4457452745198696,
      "tokens_seen": 2094465024
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.379232867918473e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.445727605938412,
      "tokens_seen": 2094596096
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.378430428502648e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.445709938772109,
      "tokens_seen": 2094727168
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.377627989086823e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.445692273020759,
      "tokens_seen": 2094858240
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.376825549671e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.44567460868416,
      "tokens_seen": 2094989312
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.376023110255175e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.44565694576211,
      "tokens_seen": 2095120384
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.375220670839352e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.445639284254407,
      "tokens_seen": 2095251456
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.374418231423527e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.44562162416085,
      "tokens_seen": 2095382528
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.373615792007704e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.4456039654812365,
      "tokens_seen": 2095513600
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.9952130317687988,
      "objective/train/avg_token_score": 0.004787010606378317,
      "objective/train/docs_used": 764034,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2309796810150146,
      "objective/train/original_loss": 1.2377663850784302,
      "objective/train/theoretical_loss": 3.4455863082153657,
      "objective/train/tokens_used": 466103776,
      "objective/train/unlikelihood_loss": 0.012290958315134048,
      "theoretical_loss": 3.4455863082153657,
      "tokens_seen": 2095644672
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.372813352591879e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.4455863082153657,
      "tokens_seen": 2095644672
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.372010913176055e-05,
      "loss": 1.361,
      "theoretical_loss": 3.4455686523630358,
      "tokens_seen": 2095775744
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.371208473760231e-05,
      "loss": 1.2803,
      "theoretical_loss": 3.4455509979240446,
      "tokens_seen": 2095906816
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.370406034344407e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.445533344898191,
      "tokens_seen": 2096037888
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.369603594928583e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.445515693285274,
      "tokens_seen": 2096168960
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.368801155512759e-05,
      "loss": 1.3554,
      "theoretical_loss": 3.4454980430850917,
      "tokens_seen": 2096300032
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.367998716096935e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.4454803942974426,
      "tokens_seen": 2096431104
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.36719627668111e-05,
      "loss": 1.3903,
      "theoretical_loss": 3.445462746922126,
      "tokens_seen": 2096562176
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.366393837265286e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.4454451009589393,
      "tokens_seen": 2096693248
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.365591397849463e-05,
      "loss": 1.3871,
      "theoretical_loss": 3.4454274564076823,
      "tokens_seen": 2096824320
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.364788958433638e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.445409813268154,
      "tokens_seen": 2096955392
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.363986519017815e-05,
      "loss": 1.3531,
      "theoretical_loss": 3.445392171540152,
      "tokens_seen": 2097086464
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.36318407960199e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.445374531223476,
      "tokens_seen": 2097217536
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.362381640186166e-05,
      "loss": 1.3778,
      "theoretical_loss": 3.445356892317925,
      "tokens_seen": 2097348608
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.361579200770342e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.445339254823298,
      "tokens_seen": 2097479680
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.360776761354517e-05,
      "loss": 1.208,
      "theoretical_loss": 3.4453216187393934,
      "tokens_seen": 2097610752
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.359974321938694e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.445303984066011,
      "tokens_seen": 2097741824
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.359171882522869e-05,
      "loss": 1.2437,
      "theoretical_loss": 3.4452863508029488,
      "tokens_seen": 2097872896
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.358369443107046e-05,
      "loss": 1.32,
      "theoretical_loss": 3.4452687189500066,
      "tokens_seen": 2098003968
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.357567003691221e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.445251088506984,
      "tokens_seen": 2098135040
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.356764564275398e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.4452334594736786,
      "tokens_seen": 2098266112
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.355962124859573e-05,
      "loss": 1.3985,
      "theoretical_loss": 3.4452158318498913,
      "tokens_seen": 2098397184
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.35515968544375e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.445198205635421,
      "tokens_seen": 2098528256
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.354357246027925e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.4451805808300664,
      "tokens_seen": 2098659328
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.3535548066121e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.4451629574336273,
      "tokens_seen": 2098790400
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.352752367196277e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.445145335445903,
      "tokens_seen": 2098921472
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.351949927780452e-05,
      "loss": 1.2547,
      "theoretical_loss": 3.4451277148666932,
      "tokens_seen": 2099052544
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.351147488364629e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.4451100956957967,
      "tokens_seen": 2099183616
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.350345048948804e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.445092477933014,
      "tokens_seen": 2099314688
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.349542609532981e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.4450748615781435,
      "tokens_seen": 2099445760
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.348740170117156e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.4450572466309852,
      "tokens_seen": 2099576832
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.347937730701332e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.4450396330913398,
      "tokens_seen": 2099707904
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.347135291285508e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.4450220209590054,
      "tokens_seen": 2099838976
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.346332851869684e-05,
      "loss": 1.2547,
      "theoretical_loss": 3.445004410233783,
      "tokens_seen": 2099970048
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.34553041245386e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.4449868009154714,
      "tokens_seen": 2100101120
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.344727973038036e-05,
      "loss": 1.259,
      "theoretical_loss": 3.444969193003871,
      "tokens_seen": 2100232192
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.343925533622212e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.4449515864987816,
      "tokens_seen": 2100363264
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.343123094206388e-05,
      "loss": 1.3518,
      "theoretical_loss": 3.4449339814000024,
      "tokens_seen": 2100494336
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.342320654790563e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.4449163777073344,
      "tokens_seen": 2100625408
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.34151821537474e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.444898775420577,
      "tokens_seen": 2100756480
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.340715775958915e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.44488117453953,
      "tokens_seen": 2100887552
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.339913336543092e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.4448635750639935,
      "tokens_seen": 2101018624
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.339110897127267e-05,
      "loss": 1.263,
      "theoretical_loss": 3.4448459769937685,
      "tokens_seen": 2101149696
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.338308457711443e-05,
      "loss": 1.257,
      "theoretical_loss": 3.4448283803286537,
      "tokens_seen": 2101280768
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.337506018295619e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.4448107850684506,
      "tokens_seen": 2101411840
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.336703578879794e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.444793191212958,
      "tokens_seen": 2101542912
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.335901139463971e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.4447755987619777,
      "tokens_seen": 2101673984
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.335098700048146e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.4447580077153095,
      "tokens_seen": 2101805056
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.334296260632323e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.4447404180727528,
      "tokens_seen": 2101936128
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.333493821216498e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.4447228298341086,
      "tokens_seen": 2102067200
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.9941034913063049,
      "objective/train/avg_token_score": 0.005896490067243576,
      "objective/train/docs_used": 765966,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2985467910766602,
      "objective/train/original_loss": 1.3068296909332275,
      "objective/train/theoretical_loss": 3.444705242999178,
      "objective/train/tokens_used": 472657376,
      "objective/train/unlikelihood_loss": 0.01847456768155098,
      "theoretical_loss": 3.444705242999178,
      "tokens_seen": 2102198272
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.332691381800675e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.444705242999178,
      "tokens_seen": 2102198272
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.33188894238485e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.4446876575677603,
      "tokens_seen": 2102329344
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.331086502969025e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.444670073539657,
      "tokens_seen": 2102460416
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.330284063553202e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.4446524909146676,
      "tokens_seen": 2102591488
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.329481624137377e-05,
      "loss": 1.2298,
      "theoretical_loss": 3.444634909692594,
      "tokens_seen": 2102722560
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.328679184721554e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.4446173298732354,
      "tokens_seen": 2102853632
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.32787674530573e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.4445997514563937,
      "tokens_seen": 2102984704
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.327074305889906e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.444582174441869,
      "tokens_seen": 2103115776
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.326271866474081e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.444564598829462,
      "tokens_seen": 2103246848
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.325469427058258e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.4445470246189736,
      "tokens_seen": 2103377920
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.324666987642433e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.4445294518102045,
      "tokens_seen": 2103508992
    },
    {
      "epoch": 0.27,
      "learning_rate": 7.323864548226609e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.4445118804029558,
      "tokens_seen": 2103640064
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.323062108810785e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.4444943103970282,
      "tokens_seen": 2103771136
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.32225966939496e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.444476741792223,
      "tokens_seen": 2103902208
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.321457229979137e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.4444591745883404,
      "tokens_seen": 2104033280
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.320654790563313e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.4444416087851826,
      "tokens_seen": 2104164352
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.319852351147489e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.44442404438255,
      "tokens_seen": 2104295424
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.319049911731665e-05,
      "loss": 1.2542,
      "theoretical_loss": 3.444406481380243,
      "tokens_seen": 2104426496
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.31824747231584e-05,
      "loss": 1.1968,
      "theoretical_loss": 3.444388919778064,
      "tokens_seen": 2104557568
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.317445032900017e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.4443713595758134,
      "tokens_seen": 2104688640
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.316642593484192e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.444353800773293,
      "tokens_seen": 2104819712
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.315840154068369e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.4443362433703038,
      "tokens_seen": 2104950784
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.315037714652544e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.444318687366647,
      "tokens_seen": 2105081856
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.31423527523672e-05,
      "loss": 1.257,
      "theoretical_loss": 3.4443011327621242,
      "tokens_seen": 2105212928
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.313432835820896e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.4442835795565365,
      "tokens_seen": 2105344000
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.312630396405071e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.444266027749685,
      "tokens_seen": 2105475072
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.311827956989248e-05,
      "loss": 1.308,
      "theoretical_loss": 3.4442484773413726,
      "tokens_seen": 2105606144
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.311025517573423e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.4442309283313994,
      "tokens_seen": 2105737216
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.3102230781576e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.4442133807195674,
      "tokens_seen": 2105868288
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.309420638741775e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.4441958345056785,
      "tokens_seen": 2105999360
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.308618199325952e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.4441782896895337,
      "tokens_seen": 2106130432
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.307815759910127e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.444160746270935,
      "tokens_seen": 2106261504
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.307013320494302e-05,
      "loss": 1.252,
      "theoretical_loss": 3.4441432042496842,
      "tokens_seen": 2106392576
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.306210881078479e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.4441256636255835,
      "tokens_seen": 2106523648
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.305408441662654e-05,
      "loss": 1.311,
      "theoretical_loss": 3.4441081243984337,
      "tokens_seen": 2106654720
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.304606002246831e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.444090586568038,
      "tokens_seen": 2106785792
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.303803562831006e-05,
      "loss": 1.2645,
      "theoretical_loss": 3.4440730501341967,
      "tokens_seen": 2106916864
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.303001123415183e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.4440555150967125,
      "tokens_seen": 2107047936
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.302198683999358e-05,
      "loss": 1.229,
      "theoretical_loss": 3.444037981455387,
      "tokens_seen": 2107179008
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.301396244583534e-05,
      "loss": 1.191,
      "theoretical_loss": 3.444020449210023,
      "tokens_seen": 2107310080
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.30059380516771e-05,
      "loss": 1.2214,
      "theoretical_loss": 3.444002918360422,
      "tokens_seen": 2107441152
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.299791365751886e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.4439853889063863,
      "tokens_seen": 2107572224
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.298988926336062e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.4439678608477178,
      "tokens_seen": 2107703296
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.298186486920238e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.4439503341842186,
      "tokens_seen": 2107834368
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.297384047504414e-05,
      "loss": 1.256,
      "theoretical_loss": 3.443932808915691,
      "tokens_seen": 2107965440
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.29658160808859e-05,
      "loss": 1.2586,
      "theoretical_loss": 3.4439152850419377,
      "tokens_seen": 2108096512
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.295779168672765e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.4438977625627603,
      "tokens_seen": 2108227584
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.294976729256942e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.4438802414779612,
      "tokens_seen": 2108358656
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.294174289841117e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.4438627217873434,
      "tokens_seen": 2108489728
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.293371850425294e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.4438452034907083,
      "tokens_seen": 2108620800
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.9933674931526184,
      "objective/train/avg_token_score": 0.0066325124353170395,
      "objective/train/docs_used": 768317,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1539947986602783,
      "objective/train/original_loss": 1.1629319190979004,
      "objective/train/theoretical_loss": 3.443827686587859,
      "objective/train/tokens_used": 479210976,
      "objective/train/unlikelihood_loss": 0.020673779770731926,
      "theoretical_loss": 3.443827686587859,
      "tokens_seen": 2108751872
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.292569411009469e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.443827686587859,
      "tokens_seen": 2108751872
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.291766971593646e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.4438101710785984,
      "tokens_seen": 2108882944
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.290964532177821e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4437926569627284,
      "tokens_seen": 2109014016
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.290162092761998e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.4437751442400515,
      "tokens_seen": 2109145088
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.289359653346173e-05,
      "loss": 1.2423,
      "theoretical_loss": 3.443757632910371,
      "tokens_seen": 2109276160
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.288557213930348e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.4437401229734883,
      "tokens_seen": 2109407232
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.287754774514525e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.4437226144292072,
      "tokens_seen": 2109538304
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.2869523350987e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.4437051072773297,
      "tokens_seen": 2109669376
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.286149895682877e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.4436876015176594,
      "tokens_seen": 2109800448
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.285347456267052e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.4436700971499983,
      "tokens_seen": 2109931520
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.284545016851229e-05,
      "loss": 1.2175,
      "theoretical_loss": 3.44365259417415,
      "tokens_seen": 2110062592
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.283742577435404e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.4436350925899166,
      "tokens_seen": 2110193664
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.28294013801958e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.4436175923971017,
      "tokens_seen": 2110324736
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.282137698603756e-05,
      "loss": 1.255,
      "theoretical_loss": 3.443600093595508,
      "tokens_seen": 2110455808
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.281335259187931e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.443582596184938,
      "tokens_seen": 2110586880
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.280532819772108e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.4435651001651957,
      "tokens_seen": 2110717952
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.279730380356283e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.443547605536083,
      "tokens_seen": 2110849024
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.27892794094046e-05,
      "loss": 1.2188,
      "theoretical_loss": 3.4435301122974047,
      "tokens_seen": 2110980096
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.278125501524635e-05,
      "loss": 1.1447,
      "theoretical_loss": 3.4435126204489626,
      "tokens_seen": 2111111168
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.27732306210881e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.44349512999056,
      "tokens_seen": 2111242240
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.276520622692987e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.443477640922001,
      "tokens_seen": 2111373312
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.275718183277163e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.4434601532430884,
      "tokens_seen": 2111504384
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.274915743861339e-05,
      "loss": 1.1936,
      "theoretical_loss": 3.443442666953625,
      "tokens_seen": 2111635456
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.274113304445515e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.4434251820534145,
      "tokens_seen": 2111766528
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.273310865029691e-05,
      "loss": 1.2158,
      "theoretical_loss": 3.443407698542261,
      "tokens_seen": 2111897600
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.272508425613867e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.4433902164199672,
      "tokens_seen": 2112028672
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.271705986198042e-05,
      "loss": 1.28,
      "theoretical_loss": 3.443372735686337,
      "tokens_seen": 2112159744
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.270903546782219e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.4433552563411736,
      "tokens_seen": 2112290816
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.270101107366394e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.443337778384281,
      "tokens_seen": 2112421888
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.26929866795057e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.4433203018154623,
      "tokens_seen": 2112552960
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.268496228534746e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.443302826634521,
      "tokens_seen": 2112684032
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.267693789118923e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.4432853528412615,
      "tokens_seen": 2112815104
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.266891349703098e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.443267880435487,
      "tokens_seen": 2112946176
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.266088910287273e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.4432504094170016,
      "tokens_seen": 2113077248
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.26528647087145e-05,
      "loss": 1.1928,
      "theoretical_loss": 3.443232939785609,
      "tokens_seen": 2113208320
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.264484031455625e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.443215471541113,
      "tokens_seen": 2113339392
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.263681592039802e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.4431980046833175,
      "tokens_seen": 2113470464
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.262879152623977e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.443180539212026,
      "tokens_seen": 2113601536
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.262076713208154e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.4431630751270434,
      "tokens_seen": 2113732608
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.261274273792329e-05,
      "loss": 1.2122,
      "theoretical_loss": 3.4431456124281734,
      "tokens_seen": 2113863680
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.260471834376506e-05,
      "loss": 1.3713,
      "theoretical_loss": 3.443128151115219,
      "tokens_seen": 2113994752
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.259669394960681e-05,
      "loss": 1.303,
      "theoretical_loss": 3.443110691187986,
      "tokens_seen": 2114125824
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.258866955544856e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.4430932326462775,
      "tokens_seen": 2114256896
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.258064516129033e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.4430757754898975,
      "tokens_seen": 2114387968
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.257262076713208e-05,
      "loss": 1.278,
      "theoretical_loss": 3.4430583197186504,
      "tokens_seen": 2114519040
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.256459637297385e-05,
      "loss": 1.304,
      "theoretical_loss": 3.443040865332341,
      "tokens_seen": 2114650112
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.25565719788156e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.4430234123307732,
      "tokens_seen": 2114781184
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.254854758465737e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.4430059607137515,
      "tokens_seen": 2114912256
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.254052319049912e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.4429885104810793,
      "tokens_seen": 2115043328
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.253249879634088e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.442971061632562,
      "tokens_seen": 2115174400
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.9916481971740723,
      "objective/train/avg_token_score": 0.008351797237992287,
      "objective/train/docs_used": 770743,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2003364562988281,
      "objective/train/original_loss": 1.2103271484375,
      "objective/train/theoretical_loss": 3.4429536141680046,
      "objective/train/tokens_used": 485764576,
      "objective/train/unlikelihood_loss": 0.028402267023921013,
      "theoretical_loss": 3.4429536141680046,
      "tokens_seen": 2115305472
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.252447440218264e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.4429536141680046,
      "tokens_seen": 2115305472
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.25164500080244e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.4429361680872104,
      "tokens_seen": 2115436544
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.250842561386616e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.442918723389984,
      "tokens_seen": 2115567616
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.250040121970792e-05,
      "loss": 1.2182,
      "theoretical_loss": 3.4429012800761303,
      "tokens_seen": 2115698688
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.249237682554968e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.4428838381454545,
      "tokens_seen": 2115829760
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.248435243139144e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.4428663975977596,
      "tokens_seen": 2115960832
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.247632803723319e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.4428489584328523,
      "tokens_seen": 2116091904
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.246830364307496e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.442831520650536,
      "tokens_seen": 2116222976
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.246027924891671e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.442814084250616,
      "tokens_seen": 2116354048
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.245225485475848e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.4427966492328967,
      "tokens_seen": 2116485120
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.244423046060023e-05,
      "loss": 1.2215,
      "theoretical_loss": 3.4427792155971835,
      "tokens_seen": 2116616192
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.2436206066442e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.442761783343281,
      "tokens_seen": 2116747264
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.242818167228375e-05,
      "loss": 1.3193,
      "theoretical_loss": 3.4427443524709935,
      "tokens_seen": 2116878336
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.24201572781255e-05,
      "loss": 1.2645,
      "theoretical_loss": 3.442726922980127,
      "tokens_seen": 2117009408
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.241213288396727e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.4427094948704857,
      "tokens_seen": 2117140480
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.240410848980902e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.4426920681418753,
      "tokens_seen": 2117271552
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.239608409565079e-05,
      "loss": 1.2321,
      "theoretical_loss": 3.4426746427941004,
      "tokens_seen": 2117402624
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.238805970149254e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.4426572188269664,
      "tokens_seen": 2117533696
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.238003530733431e-05,
      "loss": 1.2206,
      "theoretical_loss": 3.442639796240278,
      "tokens_seen": 2117664768
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.237201091317606e-05,
      "loss": 1.256,
      "theoretical_loss": 3.4426223750338414,
      "tokens_seen": 2117795840
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.236398651901781e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.442604955207461,
      "tokens_seen": 2117926912
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.235596212485958e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.442587536760942,
      "tokens_seen": 2118057984
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.234793773070133e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.44257011969409,
      "tokens_seen": 2118189056
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.23399133365431e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.44255270400671,
      "tokens_seen": 2118320128
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.233188894238485e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.4425352896986086,
      "tokens_seen": 2118451200
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.232386454822662e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.44251787676959,
      "tokens_seen": 2118582272
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.231584015406837e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.4425004652194593,
      "tokens_seen": 2118713344
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.230781575991014e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.4424830550480237,
      "tokens_seen": 2118844416
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.22997913657519e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.4424656462550876,
      "tokens_seen": 2118975488
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.229176697159365e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.4424482388404565,
      "tokens_seen": 2119106560
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.228374257743541e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.442430832803937,
      "tokens_seen": 2119237632
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.227571818327717e-05,
      "loss": 1.342,
      "theoretical_loss": 3.4424134281453336,
      "tokens_seen": 2119368704
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.226769378911893e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.4423960248644523,
      "tokens_seen": 2119499776
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.225966939496069e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.4423786229610998,
      "tokens_seen": 2119630848
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.225164500080245e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.442361222435081,
      "tokens_seen": 2119761920
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.22436206066442e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.442343823286201,
      "tokens_seen": 2119892992
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.223559621248596e-05,
      "loss": 1.28,
      "theoretical_loss": 3.442326425514268,
      "tokens_seen": 2120024064
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.222757181832773e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.4423090291190857,
      "tokens_seen": 2120155136
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.221954742416948e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.442291634100461,
      "tokens_seen": 2120286208
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.221152303001125e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.4422742404581994,
      "tokens_seen": 2120417280
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.2203498635853e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.442256848192107,
      "tokens_seen": 2120548352
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.219547424169477e-05,
      "loss": 1.2198,
      "theoretical_loss": 3.4422394573019908,
      "tokens_seen": 2120679424
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.218744984753652e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.4422220677876565,
      "tokens_seen": 2120810496
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.217942545337827e-05,
      "loss": 1.3954,
      "theoretical_loss": 3.4422046796489094,
      "tokens_seen": 2120941568
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.217140105922004e-05,
      "loss": 1.2122,
      "theoretical_loss": 3.4421872928855564,
      "tokens_seen": 2121072640
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.216337666506179e-05,
      "loss": 1.1962,
      "theoretical_loss": 3.4421699074974033,
      "tokens_seen": 2121203712
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.215535227090356e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.442152523484257,
      "tokens_seen": 2121334784
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.214732787674531e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.4421351408459238,
      "tokens_seen": 2121465856
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.213930348258708e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.4421177595822092,
      "tokens_seen": 2121596928
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.213127908842883e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.4421003796929197,
      "tokens_seen": 2121728000
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.9931816458702087,
      "objective/train/avg_token_score": 0.006818348541855812,
      "objective/train/docs_used": 773131,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2505849599838257,
      "objective/train/original_loss": 1.2599543333053589,
      "objective/train/theoretical_loss": 3.442083001177863,
      "objective/train/tokens_used": 492318176,
      "objective/train/unlikelihood_loss": 0.01911384053528309,
      "theoretical_loss": 3.442083001177863,
      "tokens_seen": 2121859072
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.212325469427058e-05,
      "loss": 1.2137,
      "theoretical_loss": 3.442083001177863,
      "tokens_seen": 2121859072
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.211523030011235e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.442065624036844,
      "tokens_seen": 2121990144
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.21072059059541e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.4420482482696704,
      "tokens_seen": 2122121216
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.209918151179587e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.4420308738761474,
      "tokens_seen": 2122252288
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.209115711763762e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.442013500856083,
      "tokens_seen": 2122383360
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.208313272347939e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.441996129209283,
      "tokens_seen": 2122514432
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.207510832932114e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.441978758935554,
      "tokens_seen": 2122645504
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.20670839351629e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.4419613900347032,
      "tokens_seen": 2122776576
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.205905954100466e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.4419440225065374,
      "tokens_seen": 2122907648
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.205103514684642e-05,
      "loss": 1.3392,
      "theoretical_loss": 3.441926656350862,
      "tokens_seen": 2123038720
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.204301075268818e-05,
      "loss": 1.2423,
      "theoretical_loss": 3.441909291567486,
      "tokens_seen": 2123169792
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.203498635852994e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.441891928156215,
      "tokens_seen": 2123300864
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.20269619643717e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.441874566116855,
      "tokens_seen": 2123431936
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.201893757021346e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.4418572054492147,
      "tokens_seen": 2123563008
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.201091317605521e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.4418398461531003,
      "tokens_seen": 2123694080
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.200288878189698e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.441822488228319,
      "tokens_seen": 2123825152
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.199486438773873e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.441805131674677,
      "tokens_seen": 2123956224
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.19868399935805e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.4417877764919824,
      "tokens_seen": 2124087296
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.197881559942225e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.441770422680042,
      "tokens_seen": 2124218368
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.197079120526402e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.4417530702386627,
      "tokens_seen": 2124349440
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.196276681110577e-05,
      "loss": 1.3946,
      "theoretical_loss": 3.441735719167652,
      "tokens_seen": 2124480512
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.195474241694754e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.441718369466817,
      "tokens_seen": 2124611584
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.194671802278929e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.4417010211359655,
      "tokens_seen": 2124742656
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.193869362863104e-05,
      "loss": 1.262,
      "theoretical_loss": 3.441683674174904,
      "tokens_seen": 2124873728
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.193066923447281e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.44166632858344,
      "tokens_seen": 2125004800
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.192264484031456e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.4416489843613816,
      "tokens_seen": 2125135872
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.191462044615633e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.441631641508536,
      "tokens_seen": 2125266944
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.190659605199808e-05,
      "loss": 1.3662,
      "theoretical_loss": 3.44161430002471,
      "tokens_seen": 2125398016
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.189857165783985e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.4415969599097114,
      "tokens_seen": 2125529088
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.18905472636816e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.4415796211633483,
      "tokens_seen": 2125660160
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.188252286952335e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.441562283785428,
      "tokens_seen": 2125791232
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.187449847536512e-05,
      "loss": 1.3547,
      "theoretical_loss": 3.4415449477757574,
      "tokens_seen": 2125922304
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.186647408120687e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.441527613134145,
      "tokens_seen": 2126053376
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.185844968704864e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.441510279860399,
      "tokens_seen": 2126184448
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.18504252928904e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.441492947954326,
      "tokens_seen": 2126315520
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.184240089873216e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.4414756174157337,
      "tokens_seen": 2126446592
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.183437650457391e-05,
      "loss": 1.323,
      "theoretical_loss": 3.441458288244431,
      "tokens_seen": 2126577664
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.182635211041567e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.441440960440225,
      "tokens_seen": 2126708736
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.181832771625743e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.4414236340029243,
      "tokens_seen": 2126839808
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.181030332209919e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.4414063089323363,
      "tokens_seen": 2126970880
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.180227892794095e-05,
      "loss": 1.2103,
      "theoretical_loss": 3.4413889852282686,
      "tokens_seen": 2127101952
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.17942545337827e-05,
      "loss": 1.357,
      "theoretical_loss": 3.44137166289053,
      "tokens_seen": 2127233024
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.178623013962446e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.4413543419189283,
      "tokens_seen": 2127364096
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.177820574546623e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.441337022313272,
      "tokens_seen": 2127495168
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.177018135130798e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.441319704073368,
      "tokens_seen": 2127626240
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.176215695714975e-05,
      "loss": 1.346,
      "theoretical_loss": 3.4413023871990256,
      "tokens_seen": 2127757312
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.17541325629915e-05,
      "loss": 1.1898,
      "theoretical_loss": 3.4412850716900527,
      "tokens_seen": 2127888384
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.174610816883327e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.4412677575462576,
      "tokens_seen": 2128019456
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.173808377467502e-05,
      "loss": 1.25,
      "theoretical_loss": 3.441250444767448,
      "tokens_seen": 2128150528
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.173005938051677e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.441233133353433,
      "tokens_seen": 2128281600
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.9931174516677856,
      "objective/train/avg_token_score": 0.006882534362375736,
      "objective/train/docs_used": 775219,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.275771975517273,
      "objective/train/original_loss": 1.2837028503417969,
      "objective/train/theoretical_loss": 3.441215823304021,
      "objective/train/tokens_used": 498871776,
      "objective/train/unlikelihood_loss": 0.02251371741294861,
      "theoretical_loss": 3.441215823304021,
      "tokens_seen": 2128412672
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.172203498635854e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.441215823304021,
      "tokens_seen": 2128412672
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.171401059220029e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.4411985146190203,
      "tokens_seen": 2128543744
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.170598619804206e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.4411812072982384,
      "tokens_seen": 2128674816
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.169796180388381e-05,
      "loss": 1.275,
      "theoretical_loss": 3.441163901341485,
      "tokens_seen": 2128805888
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.168993740972557e-05,
      "loss": 1.2496,
      "theoretical_loss": 3.441146596748568,
      "tokens_seen": 2128936960
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.168191301556733e-05,
      "loss": 1.1913,
      "theoretical_loss": 3.4411292935192965,
      "tokens_seen": 2129068032
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.167388862140908e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.4411119916534787,
      "tokens_seen": 2129199104
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.166586422725085e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.441094691150923,
      "tokens_seen": 2129330176
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.16578398330926e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.4410773920114384,
      "tokens_seen": 2129461248
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.164981543893436e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.4410600942348335,
      "tokens_seen": 2129592320
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.164179104477612e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.4410427978209173,
      "tokens_seen": 2129723392
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.163376665061788e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.4410255027694987,
      "tokens_seen": 2129854464
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.162574225645964e-05,
      "loss": 1.3629,
      "theoretical_loss": 3.441008209080386,
      "tokens_seen": 2129985536
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.16177178623014e-05,
      "loss": 1.4019,
      "theoretical_loss": 3.440990916753388,
      "tokens_seen": 2130116608
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.160969346814316e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.440973625788314,
      "tokens_seen": 2130247680
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.160166907398492e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.440956336184974,
      "tokens_seen": 2130378752
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.159364467982667e-05,
      "loss": 1.3631,
      "theoretical_loss": 3.440939047943175,
      "tokens_seen": 2130509824
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.158562028566844e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.4409217610627265,
      "tokens_seen": 2130640896
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.157759589151019e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.4409044755434386,
      "tokens_seen": 2130771968
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.156957149735196e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.4408871913851193,
      "tokens_seen": 2130903040
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.156154710319371e-05,
      "loss": 1.23,
      "theoretical_loss": 3.4408699085875787,
      "tokens_seen": 2131034112
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.155352270903546e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.4408526271506252,
      "tokens_seen": 2131165184
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.154549831487723e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.440835347074068,
      "tokens_seen": 2131296256
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.153747392071898e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.4408180683577174,
      "tokens_seen": 2131427328
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.152944952656075e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.4408007910013816,
      "tokens_seen": 2131558400
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.15214251324025e-05,
      "loss": 1.3603,
      "theoretical_loss": 3.4407835150048705,
      "tokens_seen": 2131689472
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.151340073824427e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.440766240367993,
      "tokens_seen": 2131820544
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.150537634408602e-05,
      "loss": 1.2205,
      "theoretical_loss": 3.440748967090559,
      "tokens_seen": 2131951616
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.149735194992778e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.440731695172377,
      "tokens_seen": 2132082688
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.148932755576954e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.440714424613258,
      "tokens_seen": 2132213760
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.14813031616113e-05,
      "loss": 1.3359,
      "theoretical_loss": 3.4406971554130106,
      "tokens_seen": 2132344832
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.147327876745306e-05,
      "loss": 1.4035,
      "theoretical_loss": 3.440679887571444,
      "tokens_seen": 2132475904
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.146525437329482e-05,
      "loss": 1.3439,
      "theoretical_loss": 3.440662621088369,
      "tokens_seen": 2132606976
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.145722997913657e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.440645355963594,
      "tokens_seen": 2132738048
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.144920558497834e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.44062809219693,
      "tokens_seen": 2132869120
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.144118119082009e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.440610829788185,
      "tokens_seen": 2133000192
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.143315679666185e-05,
      "loss": 1.2289,
      "theoretical_loss": 3.44059356873717,
      "tokens_seen": 2133131264
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.142513240250361e-05,
      "loss": 1.321,
      "theoretical_loss": 3.4405763090436947,
      "tokens_seen": 2133262336
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.141710800834537e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.440559050707569,
      "tokens_seen": 2133393408
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.140908361418713e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.440541793728602,
      "tokens_seen": 2133524480
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.140105922002888e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.4405245381066045,
      "tokens_seen": 2133655552
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.139303482587065e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.4405072838413853,
      "tokens_seen": 2133786624
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.13850104317124e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.440490030932756,
      "tokens_seen": 2133917696
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.137698603755417e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.440472779380525,
      "tokens_seen": 2134048768
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.136896164339592e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.440455529184504,
      "tokens_seen": 2134179840
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.136093724923767e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.4404382803445017,
      "tokens_seen": 2134310912
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.135291285507944e-05,
      "loss": 1.2285,
      "theoretical_loss": 3.440421032860329,
      "tokens_seen": 2134441984
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.13448884609212e-05,
      "loss": 1.3889,
      "theoretical_loss": 3.440403786731795,
      "tokens_seen": 2134573056
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.133686406676296e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.440386541958712,
      "tokens_seen": 2134704128
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.132883967260471e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.4403692985408885,
      "tokens_seen": 2134835200
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.992087721824646,
      "objective/train/avg_token_score": 0.00791224092245102,
      "objective/train/docs_used": 777595,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2920936346054077,
      "objective/train/original_loss": 1.3026537895202637,
      "objective/train/theoretical_loss": 3.440352056478135,
      "objective/train/tokens_used": 505425376,
      "objective/train/unlikelihood_loss": 0.022452739998698235,
      "theoretical_loss": 3.440352056478135,
      "tokens_seen": 2134966272
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.132081527844648e-05,
      "loss": 1.3007,
      "theoretical_loss": 3.440352056478135,
      "tokens_seen": 2134966272
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.131279088428823e-05,
      "loss": 1.249,
      "theoretical_loss": 3.4403348157702625,
      "tokens_seen": 2135097344
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.130476649013e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.4403175764170806,
      "tokens_seen": 2135228416
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.129674209597175e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.4403003384184005,
      "tokens_seen": 2135359488
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.12887177018135e-05,
      "loss": 1.2064,
      "theoretical_loss": 3.4402831017740323,
      "tokens_seen": 2135490560
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.128069330765527e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.440265866483786,
      "tokens_seen": 2135621632
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.127266891349703e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.440248632547473,
      "tokens_seen": 2135752704
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.126464451933879e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.4402313999649037,
      "tokens_seen": 2135883776
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.125662012518055e-05,
      "loss": 1.247,
      "theoretical_loss": 3.4402141687358876,
      "tokens_seen": 2136014848
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.124859573102231e-05,
      "loss": 1.3849,
      "theoretical_loss": 3.4401969388602365,
      "tokens_seen": 2136145920
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.124057133686407e-05,
      "loss": 1.246,
      "theoretical_loss": 3.4401797103377607,
      "tokens_seen": 2136276992
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.123254694270582e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.440162483168271,
      "tokens_seen": 2136408064
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.122452254854759e-05,
      "loss": 1.279,
      "theoretical_loss": 3.4401452573515785,
      "tokens_seen": 2136539136
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.121649815438934e-05,
      "loss": 1.276,
      "theoretical_loss": 3.440128032887493,
      "tokens_seen": 2136670208
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.12084737602311e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.4401108097758257,
      "tokens_seen": 2136801280
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.120044936607286e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.440093588016388,
      "tokens_seen": 2136932352
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.119242497191462e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.4400763676089907,
      "tokens_seen": 2137063424
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.118440057775638e-05,
      "loss": 1.2448,
      "theoretical_loss": 3.440059148553444,
      "tokens_seen": 2137194496
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.117637618359813e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.44004193084956,
      "tokens_seen": 2137325568
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.11683517894399e-05,
      "loss": 1.3894,
      "theoretical_loss": 3.4400247144971488,
      "tokens_seen": 2137456640
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.116032739528165e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.440007499496021,
      "tokens_seen": 2137587712
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.115230300112342e-05,
      "loss": 1.276,
      "theoretical_loss": 3.4399902858459894,
      "tokens_seen": 2137718784
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.114427860696517e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.439973073546864,
      "tokens_seen": 2137849856
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.113625421280694e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.439955862598456,
      "tokens_seen": 2137980928
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.112822981864869e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.439938653000577,
      "tokens_seen": 2138112000
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.112020542449044e-05,
      "loss": 1.3441,
      "theoretical_loss": 3.4399214447530375,
      "tokens_seen": 2138243072
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.111218103033221e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.4399042378556497,
      "tokens_seen": 2138374144
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.110415663617396e-05,
      "loss": 1.287,
      "theoretical_loss": 3.4398870323082242,
      "tokens_seen": 2138505216
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.109613224201573e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.439869828110573,
      "tokens_seen": 2138636288
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.108810784785748e-05,
      "loss": 1.301,
      "theoretical_loss": 3.439852625262507,
      "tokens_seen": 2138767360
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.108008345369925e-05,
      "loss": 1.152,
      "theoretical_loss": 3.4398354237638378,
      "tokens_seen": 2138898432
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.1072059059541e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.4398182236143766,
      "tokens_seen": 2139029504
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.106403466538276e-05,
      "loss": 1.356,
      "theoretical_loss": 3.439801024813935,
      "tokens_seen": 2139160576
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.105601027122452e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.439783827362325,
      "tokens_seen": 2139291648
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.104798587706628e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.439766631259358,
      "tokens_seen": 2139422720
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.103996148290804e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.4397494365048455,
      "tokens_seen": 2139553792
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.10319370887498e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.4397322430985993,
      "tokens_seen": 2139684864
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.102391269459156e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.43971505104043,
      "tokens_seen": 2139815936
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.101588830043332e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.439697860330152,
      "tokens_seen": 2139947008
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.100786390627508e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.439680670967574,
      "tokens_seen": 2140078080
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.099983951211684e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.4396634829525095,
      "tokens_seen": 2140209152
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.099181511795859e-05,
      "loss": 1.2369,
      "theoretical_loss": 3.4396462962847707,
      "tokens_seen": 2140340224
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.098379072380036e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.4396291109641677,
      "tokens_seen": 2140471296
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.097576632964211e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.4396119269905143,
      "tokens_seen": 2140602368
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.096774193548388e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.439594744363622,
      "tokens_seen": 2140733440
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.095971754132563e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.4395775630833016,
      "tokens_seen": 2140864512
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.09516931471674e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.439560383149367,
      "tokens_seen": 2140995584
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.094366875300915e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.439543204561629,
      "tokens_seen": 2141126656
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.09356443588509e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.4395260273199,
      "tokens_seen": 2141257728
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.092761996469267e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.439508851423992,
      "tokens_seen": 2141388800
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.9912110567092896,
      "objective/train/avg_token_score": 0.008788884617388248,
      "objective/train/docs_used": 779818,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2728849649429321,
      "objective/train/original_loss": 1.2834006547927856,
      "objective/train/theoretical_loss": 3.4394916768737174,
      "objective/train/tokens_used": 511978976,
      "objective/train/unlikelihood_loss": 0.02810192108154297,
      "theoretical_loss": 3.4394916768737174,
      "tokens_seen": 2141519872
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.091959557053442e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.4394916768737174,
      "tokens_seen": 2141519872
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.091157117637619e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.4394745036688885,
      "tokens_seen": 2141650944
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.090354678221794e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.439457331809318,
      "tokens_seen": 2141782016
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.089552238805971e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.4394401612948173,
      "tokens_seen": 2141913088
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.088749799390146e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.4394229921251993,
      "tokens_seen": 2142044160
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.087947359974321e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.439405824300276,
      "tokens_seen": 2142175232
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.087144920558498e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.43938865781986,
      "tokens_seen": 2142306304
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.086342481142673e-05,
      "loss": 1.3571,
      "theoretical_loss": 3.4393714926837635,
      "tokens_seen": 2142437376
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.08554004172685e-05,
      "loss": 1.2285,
      "theoretical_loss": 3.4393543288918,
      "tokens_seen": 2142568448
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.084737602311025e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.4393371664437806,
      "tokens_seen": 2142699520
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.083935162895202e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.4393200053395185,
      "tokens_seen": 2142830592
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.083132723479377e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.439302845578827,
      "tokens_seen": 2142961664
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.082330284063553e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.439285687161518,
      "tokens_seen": 2143092736
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.081527844647729e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.439268530087404,
      "tokens_seen": 2143223808
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.080725405231905e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.439251374356298,
      "tokens_seen": 2143354880
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.079922965816081e-05,
      "loss": 1.2007,
      "theoretical_loss": 3.4392342199680126,
      "tokens_seen": 2143485952
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.079120526400257e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.439217066922361,
      "tokens_seen": 2143617024
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.078318086984433e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.439199915219156,
      "tokens_seen": 2143748096
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.077515647568609e-05,
      "loss": 1.2355,
      "theoretical_loss": 3.4391827648582103,
      "tokens_seen": 2143879168
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.076713208152784e-05,
      "loss": 1.2335,
      "theoretical_loss": 3.439165615839336,
      "tokens_seen": 2144010240
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.07591076873696e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.4391484681623474,
      "tokens_seen": 2144141312
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.075108329321136e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.439131321827057,
      "tokens_seen": 2144272384
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.074305889905313e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.4391141768332774,
      "tokens_seen": 2144403456
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.073503450489488e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.439097033180822,
      "tokens_seen": 2144534528
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.072701011073665e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.4390798908695035,
      "tokens_seen": 2144665600
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.07189857165784e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.4390627498991355,
      "tokens_seen": 2144796672
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.071096132242017e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.4390456102695315,
      "tokens_seen": 2144927744
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.070293692826192e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.439028471980504,
      "tokens_seen": 2145058816
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.069491253410367e-05,
      "loss": 1.3153,
      "theoretical_loss": 3.439011335031866,
      "tokens_seen": 2145189888
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.068688813994544e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.4389941994234317,
      "tokens_seen": 2145320960
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.067886374578719e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.4389770651550142,
      "tokens_seen": 2145452032
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.067083935162896e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.4389599322264264,
      "tokens_seen": 2145583104
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.066281495747071e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.4389428006374816,
      "tokens_seen": 2145714176
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.065479056331248e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.4389256703879942,
      "tokens_seen": 2145845248
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.064676616915423e-05,
      "loss": 1.3614,
      "theoretical_loss": 3.438908541477776,
      "tokens_seen": 2145976320
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.063874177499598e-05,
      "loss": 1.2188,
      "theoretical_loss": 3.438891413906642,
      "tokens_seen": 2146107392
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.063071738083775e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.438874287674406,
      "tokens_seen": 2146238464
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.06226929866795e-05,
      "loss": 1.25,
      "theoretical_loss": 3.43885716278088,
      "tokens_seen": 2146369536
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.061466859252127e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.4388400392258784,
      "tokens_seen": 2146500608
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.060664419836302e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.4388229170092153,
      "tokens_seen": 2146631680
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.059861980420479e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.438805796130704,
      "tokens_seen": 2146762752
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.059059541004654e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.438788676590158,
      "tokens_seen": 2146893824
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.05825710158883e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.4387715583873915,
      "tokens_seen": 2147024896
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.057454662173006e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.4387544415222178,
      "tokens_seen": 2147155968
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.056652222757182e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.4387373259944507,
      "tokens_seen": 2147287040
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.055849783341358e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.438720211803905,
      "tokens_seen": 2147418112
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.055047343925534e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.438703098950394,
      "tokens_seen": 2147549184
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.05424490450971e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.4386859874337317,
      "tokens_seen": 2147680256
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.053442465093886e-05,
      "loss": 1.2473,
      "theoretical_loss": 3.4386688772537317,
      "tokens_seen": 2147811328
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.052640025678061e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.4386517684102085,
      "tokens_seen": 2147942400
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.9943053126335144,
      "objective/train/avg_token_score": 0.0056946841068565845,
      "objective/train/docs_used": 782164,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3371922969818115,
      "objective/train/original_loss": 1.346187949180603,
      "objective/train/theoretical_loss": 3.4386346609029763,
      "objective/train/tokens_used": 518532576,
      "objective/train/unlikelihood_loss": 0.014336034655570984,
      "theoretical_loss": 3.4386346609029763,
      "tokens_seen": 2148073472
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.051837586262238e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.4386346609029763,
      "tokens_seen": 2148073472
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.051035146846413e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.4386175547318487,
      "tokens_seen": 2148204544
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.05023270743059e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.43860044989664,
      "tokens_seen": 2148335616
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.049430268014765e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.4385833463971647,
      "tokens_seen": 2148466688
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.048627828598942e-05,
      "loss": 1.273,
      "theoretical_loss": 3.438566244233236,
      "tokens_seen": 2148597760
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.047825389183117e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.4385491434046696,
      "tokens_seen": 2148728832
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.047022949767292e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.4385320439112794,
      "tokens_seen": 2148859904
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.046220510351469e-05,
      "loss": 1.24,
      "theoretical_loss": 3.438514945752879,
      "tokens_seen": 2148990976
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.045418070935644e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.4384978489292832,
      "tokens_seen": 2149122048
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.044615631519821e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.4384807534403063,
      "tokens_seen": 2149253120
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.043813192103996e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.438463659285763,
      "tokens_seen": 2149384192
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.043010752688173e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.4384465664654673,
      "tokens_seen": 2149515264
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.042208313272348e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.4384294749792343,
      "tokens_seen": 2149646336
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.041405873856523e-05,
      "loss": 1.2402,
      "theoretical_loss": 3.4384123848268784,
      "tokens_seen": 2149777408
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.0406034344407e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.4383952960082134,
      "tokens_seen": 2149908480
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.039800995024875e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.438378208523055,
      "tokens_seen": 2150039552
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.038998555609052e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.438361122371217,
      "tokens_seen": 2150170624
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.038196116193227e-05,
      "loss": 1.248,
      "theoretical_loss": 3.4383440375525147,
      "tokens_seen": 2150301696
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.037393676777404e-05,
      "loss": 1.2594,
      "theoretical_loss": 3.4383269540667625,
      "tokens_seen": 2150432768
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.03659123736158e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.4383098719137757,
      "tokens_seen": 2150563840
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.035788797945756e-05,
      "loss": 1.292,
      "theoretical_loss": 3.4382927910933683,
      "tokens_seen": 2150694912
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.034986358529931e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.438275711605356,
      "tokens_seen": 2150825984
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.034183919114107e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.4382586334495526,
      "tokens_seen": 2150957056
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.033381479698283e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.4382415566257745,
      "tokens_seen": 2151088128
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.032579040282459e-05,
      "loss": 1.3824,
      "theoretical_loss": 3.438224481133835,
      "tokens_seen": 2151219200
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.031776600866635e-05,
      "loss": 1.2445,
      "theoretical_loss": 3.43820740697355,
      "tokens_seen": 2151350272
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.03097416145081e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.4381903341447346,
      "tokens_seen": 2151481344
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.030171722034987e-05,
      "loss": 1.2101,
      "theoretical_loss": 3.4381732626472035,
      "tokens_seen": 2151612416
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.029369282619163e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.438156192480772,
      "tokens_seen": 2151743488
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.028566843203338e-05,
      "loss": 1.3571,
      "theoretical_loss": 3.4381391236452554,
      "tokens_seen": 2151874560
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.027764403787515e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.4381220561404686,
      "tokens_seen": 2152005632
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.02696196437169e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.438104989966227,
      "tokens_seen": 2152136704
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.026159524955867e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.4380879251223453,
      "tokens_seen": 2152267776
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.025357085540042e-05,
      "loss": 1.2542,
      "theoretical_loss": 3.4380708616086393,
      "tokens_seen": 2152398848
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.024554646124219e-05,
      "loss": 1.2298,
      "theoretical_loss": 3.4380537994249245,
      "tokens_seen": 2152529920
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.023752206708394e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.438036738571016,
      "tokens_seen": 2152660992
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.022949767292569e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.4380196790467292,
      "tokens_seen": 2152792064
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.022147327876746e-05,
      "loss": 1.3106,
      "theoretical_loss": 3.4380026208518792,
      "tokens_seen": 2152923136
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.021344888460921e-05,
      "loss": 1.29,
      "theoretical_loss": 3.4379855639862824,
      "tokens_seen": 2153054208
    },
    {
      "epoch": 0.3,
      "learning_rate": 7.020542449045098e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.437968508449753,
      "tokens_seen": 2153185280
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.019740009629273e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.437951454242108,
      "tokens_seen": 2153316352
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.01893757021345e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4379344013631616,
      "tokens_seen": 2153447424
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.018135130797625e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.4379173498127304,
      "tokens_seen": 2153578496
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.0173326913818e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.43790029959063,
      "tokens_seen": 2153709568
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.016530251965977e-05,
      "loss": 1.289,
      "theoretical_loss": 3.437883250696675,
      "tokens_seen": 2153840640
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.015727812550152e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.4378662031306826,
      "tokens_seen": 2153971712
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.014925373134329e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.437849156892468,
      "tokens_seen": 2154102784
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.014122933718504e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.4378321119818462,
      "tokens_seen": 2154233856
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.013320494302681e-05,
      "loss": 1.33,
      "theoretical_loss": 3.4378150683986344,
      "tokens_seen": 2154364928
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.012518054886856e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.4377980261426475,
      "tokens_seen": 2154496000
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.9932624697685242,
      "objective/train/avg_token_score": 0.006737549789249897,
      "objective/train/docs_used": 784476,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3465124368667603,
      "objective/train/original_loss": 1.3562757968902588,
      "objective/train/theoretical_loss": 3.437780985213702,
      "objective/train/tokens_used": 525086176,
      "objective/train/unlikelihood_loss": 0.018097007647156715,
      "theoretical_loss": 3.437780985213702,
      "tokens_seen": 2154627072
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.011715615471032e-05,
      "loss": 1.268,
      "theoretical_loss": 3.437780985213702,
      "tokens_seen": 2154627072
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.010913176055208e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.4377639456116134,
      "tokens_seen": 2154758144
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.010110736639384e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.4377469073361975,
      "tokens_seen": 2154889216
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.00930829722356e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.4377298703872707,
      "tokens_seen": 2155020288
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.008505857807736e-05,
      "loss": 1.2241,
      "theoretical_loss": 3.4377128347646497,
      "tokens_seen": 2155151360
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.007703418391912e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.4376958004681493,
      "tokens_seen": 2155282432
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.006900978976088e-05,
      "loss": 1.253,
      "theoretical_loss": 3.4376787674975864,
      "tokens_seen": 2155413504
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.006098539560264e-05,
      "loss": 1.3629,
      "theoretical_loss": 3.4376617358527772,
      "tokens_seen": 2155544576
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.00529610014444e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.437644705533537,
      "tokens_seen": 2155675648
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.004493660728615e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.437627676539684,
      "tokens_seen": 2155806720
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.003691221312792e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.4376106488710327,
      "tokens_seen": 2155937792
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.002888781896967e-05,
      "loss": 1.293,
      "theoretical_loss": 3.4375936225273995,
      "tokens_seen": 2156068864
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.002086342481144e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.4375765975086017,
      "tokens_seen": 2156199936
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.001283903065319e-05,
      "loss": 1.1881,
      "theoretical_loss": 3.437559573814455,
      "tokens_seen": 2156331008
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.000481463649496e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.437542551444776,
      "tokens_seen": 2156462080
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.999679024233671e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.4375255303993812,
      "tokens_seen": 2156593152
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.998876584817846e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.4375085106780876,
      "tokens_seen": 2156724224
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.998074145402023e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.437491492280711,
      "tokens_seen": 2156855296
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.997271705986198e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.437474475207068,
      "tokens_seen": 2156986368
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.996469266570375e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.4374574594569753,
      "tokens_seen": 2157117440
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.99566682715455e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.4374404450302496,
      "tokens_seen": 2157248512
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.994864387738727e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.4374234319267076,
      "tokens_seen": 2157379584
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.994061948322902e-05,
      "loss": 1.3575,
      "theoretical_loss": 3.437406420146167,
      "tokens_seen": 2157510656
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.993259508907077e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.4373894096884423,
      "tokens_seen": 2157641728
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.992457069491254e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.437372400553352,
      "tokens_seen": 2157772800
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.99165463007543e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.437355392740713,
      "tokens_seen": 2157903872
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.990852190659606e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.4373383862503415,
      "tokens_seen": 2158034944
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.990049751243781e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.437321381082054,
      "tokens_seen": 2158166016
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.989247311827958e-05,
      "loss": 1.274,
      "theoretical_loss": 3.4373043772356686,
      "tokens_seen": 2158297088
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.988444872412133e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.4372873747110013,
      "tokens_seen": 2158428160
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.987642432996309e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.43727037350787,
      "tokens_seen": 2158559232
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.986839993580485e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.437253373626091,
      "tokens_seen": 2158690304
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.98603755416466e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.4372363750654813,
      "tokens_seen": 2158821376
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.985235114748837e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.437219377825859,
      "tokens_seen": 2158952448
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.984432675333013e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.43720238190704,
      "tokens_seen": 2159083520
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.98363023591719e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.4371853873088423,
      "tokens_seen": 2159214592
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.982827796501365e-05,
      "loss": 1.2927,
      "theoretical_loss": 3.437168394031083,
      "tokens_seen": 2159345664
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.98202535708554e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.4371514020735794,
      "tokens_seen": 2159476736
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.981222917669717e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.437134411436148,
      "tokens_seen": 2159607808
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.980420478253892e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.4371174221186074,
      "tokens_seen": 2159738880
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.979618038838069e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.4371004341207745,
      "tokens_seen": 2159869952
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.978815599422244e-05,
      "loss": 1.2709,
      "theoretical_loss": 3.4370834474424665,
      "tokens_seen": 2160001024
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.97801316000642e-05,
      "loss": 1.3247,
      "theoretical_loss": 3.437066462083501,
      "tokens_seen": 2160132096
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.977210720590596e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.4370494780436953,
      "tokens_seen": 2160263168
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.976408281174773e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.437032495322867,
      "tokens_seen": 2160394240
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.975605841758948e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.4370155139208336,
      "tokens_seen": 2160525312
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.974803402343123e-05,
      "loss": 1.4097,
      "theoretical_loss": 3.436998533837413,
      "tokens_seen": 2160656384
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.9740009629273e-05,
      "loss": 1.2941,
      "theoretical_loss": 3.4369815550724225,
      "tokens_seen": 2160787456
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.973198523511475e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.43696457762568,
      "tokens_seen": 2160918528
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.972396084095652e-05,
      "loss": 1.2016,
      "theoretical_loss": 3.436947601497003,
      "tokens_seen": 2161049600
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.9920182228088379,
      "objective/train/avg_token_score": 0.007981769740581512,
      "objective/train/docs_used": 786923,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1435245275497437,
      "objective/train/original_loss": 1.1522362232208252,
      "objective/train/theoretical_loss": 3.436930626686209,
      "objective/train/tokens_used": 531639776,
      "objective/train/unlikelihood_loss": 0.02994629740715027,
      "theoretical_loss": 3.436930626686209,
      "tokens_seen": 2161180672
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.971593644679827e-05,
      "loss": 1.238,
      "theoretical_loss": 3.436930626686209,
      "tokens_seen": 2161180672
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.970791205264004e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.436913653193117,
      "tokens_seen": 2161311744
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.969988765848179e-05,
      "loss": 1.1902,
      "theoretical_loss": 3.436896681017543,
      "tokens_seen": 2161442816
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.969186326432354e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.4368797101593063,
      "tokens_seen": 2161573888
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.968383887016531e-05,
      "loss": 1.291,
      "theoretical_loss": 3.4368627406182237,
      "tokens_seen": 2161704960
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.967581447600706e-05,
      "loss": 1.2109,
      "theoretical_loss": 3.436845772394114,
      "tokens_seen": 2161836032
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.966779008184883e-05,
      "loss": 1.3543,
      "theoretical_loss": 3.436828805486795,
      "tokens_seen": 2161967104
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.965976568769058e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4368118398960843,
      "tokens_seen": 2162098176
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.965174129353235e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.4367948756218,
      "tokens_seen": 2162229248
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.96437168993741e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.4367779126637603,
      "tokens_seen": 2162360320
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.963569250521586e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.4367609510217836,
      "tokens_seen": 2162491392
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.962766811105762e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.436743990695687,
      "tokens_seen": 2162622464
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.961964371689938e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.4367270316852903,
      "tokens_seen": 2162753536
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.961161932274114e-05,
      "loss": 1.34,
      "theoretical_loss": 3.43671007399041,
      "tokens_seen": 2162884608
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.96035949285829e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.4366931176108655,
      "tokens_seen": 2163015680
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.959557053442466e-05,
      "loss": 1.231,
      "theoretical_loss": 3.436676162546475,
      "tokens_seen": 2163146752
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.958754614026642e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.4366592087970558,
      "tokens_seen": 2163277824
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.957952174610817e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.4366422563624277,
      "tokens_seen": 2163408896
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.957149735194994e-05,
      "loss": 1.2226,
      "theoretical_loss": 3.436625305242408,
      "tokens_seen": 2163539968
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.956347295779169e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.4366083554368156,
      "tokens_seen": 2163671040
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.955544856363346e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.4365914069454684,
      "tokens_seen": 2163802112
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.954742416947521e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.436574459768186,
      "tokens_seen": 2163933184
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.953939977531698e-05,
      "loss": 1.4043,
      "theoretical_loss": 3.436557513904786,
      "tokens_seen": 2164064256
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.953137538115873e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.436540569355087,
      "tokens_seen": 2164195328
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.952335098700048e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.436523626118908,
      "tokens_seen": 2164326400
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.951532659284225e-05,
      "loss": 1.3626,
      "theoretical_loss": 3.436506684196068,
      "tokens_seen": 2164457472
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.9507302198684e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.4364897435863844,
      "tokens_seen": 2164588544
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.949927780452577e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.436472804289677,
      "tokens_seen": 2164719616
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.949125341036752e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.4364558663057645,
      "tokens_seen": 2164850688
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.948322901620929e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.4364389296344653,
      "tokens_seen": 2164981760
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.947520462205104e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.436421994275598,
      "tokens_seen": 2165112832
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.94671802278928e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.436405060228982,
      "tokens_seen": 2165243904
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.945915583373456e-05,
      "loss": 1.3619,
      "theoretical_loss": 3.436388127494436,
      "tokens_seen": 2165374976
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.945113143957631e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.4363711960717787,
      "tokens_seen": 2165506048
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.944310704541808e-05,
      "loss": 1.2278,
      "theoretical_loss": 3.4363542659608295,
      "tokens_seen": 2165637120
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.943508265125983e-05,
      "loss": 1.358,
      "theoretical_loss": 3.4363373371614068,
      "tokens_seen": 2165768192
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.94270582571016e-05,
      "loss": 1.258,
      "theoretical_loss": 3.43632040967333,
      "tokens_seen": 2165899264
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.941903386294335e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.4363034834964186,
      "tokens_seen": 2166030336
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.941100946878512e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.436286558630491,
      "tokens_seen": 2166161408
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.940298507462687e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.436269635075367,
      "tokens_seen": 2166292480
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.939496068046863e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.436252712830865,
      "tokens_seen": 2166423552
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.93869362863104e-05,
      "loss": 1.2205,
      "theoretical_loss": 3.436235791896805,
      "tokens_seen": 2166554624
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.937891189215215e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.4362188722730056,
      "tokens_seen": 2166685696
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.937088749799391e-05,
      "loss": 1.322,
      "theoretical_loss": 3.4362019539592863,
      "tokens_seen": 2166816768
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.936286310383567e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.4361850369554667,
      "tokens_seen": 2166947840
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.935483870967743e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.4361681212613657,
      "tokens_seen": 2167078912
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.934681431551919e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.4361512068768034,
      "tokens_seen": 2167209984
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.933878992136094e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.436134293801598,
      "tokens_seen": 2167341056
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.93307655272027e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.4361173820355706,
      "tokens_seen": 2167472128
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.932274113304446e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.4361004715785395,
      "tokens_seen": 2167603200
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.9908237457275391,
      "objective/train/avg_token_score": 0.009176228195428848,
      "objective/train/docs_used": 789244,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.271486759185791,
      "objective/train/original_loss": 1.283355712890625,
      "objective/train/theoretical_loss": 3.436083562430324,
      "objective/train/tokens_used": 538193376,
      "objective/train/unlikelihood_loss": 0.02557038888335228,
      "theoretical_loss": 3.436083562430324,
      "tokens_seen": 2167734272
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.931471673888623e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.436083562430324,
      "tokens_seen": 2167734272
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.930669234472798e-05,
      "loss": 1.2282,
      "theoretical_loss": 3.4360666545907446,
      "tokens_seen": 2167865344
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.929866795056975e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.4360497480596206,
      "tokens_seen": 2167996416
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.92906435564115e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.436032842836772,
      "tokens_seen": 2168127488
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.928261916225325e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.4360159389220177,
      "tokens_seen": 2168258560
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.927459476809502e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.4359990363151773,
      "tokens_seen": 2168389632
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.926657037393677e-05,
      "loss": 1.247,
      "theoretical_loss": 3.4359821350160717,
      "tokens_seen": 2168520704
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.925854597977854e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.4359652350245202,
      "tokens_seen": 2168651776
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.925052158562029e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.435948336340342,
      "tokens_seen": 2168782848
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.924249719146206e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.4359314389633573,
      "tokens_seen": 2168913920
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.923447279730381e-05,
      "loss": 1.3327,
      "theoretical_loss": 3.4359145428933866,
      "tokens_seen": 2169044992
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.922644840314556e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.435897648130249,
      "tokens_seen": 2169176064
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.921842400898733e-05,
      "loss": 1.3043,
      "theoretical_loss": 3.435880754673765,
      "tokens_seen": 2169307136
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.921039961482908e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.435863862523754,
      "tokens_seen": 2169438208
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.920237522067085e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.435846971680037,
      "tokens_seen": 2169569280
    },
    {
      "epoch": 0.31,
      "learning_rate": 6.91943508265126e-05,
      "loss": 1.2136,
      "theoretical_loss": 3.4358300821424335,
      "tokens_seen": 2169700352
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.918632643235437e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.435813193910763,
      "tokens_seen": 2169831424
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.917830203819612e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.4357963069848467,
      "tokens_seen": 2169962496
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.917027764403788e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.4357794213645043,
      "tokens_seen": 2170093568
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.916225324987964e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.4357625370495564,
      "tokens_seen": 2170224640
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.91542288557214e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.4357456540398226,
      "tokens_seen": 2170355712
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.914620446156316e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.4357287723351235,
      "tokens_seen": 2170486784
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.913818006740492e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.4357118919352794,
      "tokens_seen": 2170617856
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.913015567324668e-05,
      "loss": 1.3531,
      "theoretical_loss": 3.4356950128401107,
      "tokens_seen": 2170748928
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.912213127908844e-05,
      "loss": 1.3561,
      "theoretical_loss": 3.435678135049438,
      "tokens_seen": 2170880000
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.911410688493019e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.4356612585630812,
      "tokens_seen": 2171011072
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.910608249077196e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.435644383380861,
      "tokens_seen": 2171142144
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.909805809661371e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.4356275095025985,
      "tokens_seen": 2171273216
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.909003370245548e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.4356106369281134,
      "tokens_seen": 2171404288
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.908200930829723e-05,
      "loss": 1.3657,
      "theoretical_loss": 3.4355937656572264,
      "tokens_seen": 2171535360
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.907398491413898e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.435576895689759,
      "tokens_seen": 2171666432
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.906596051998075e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.4355600270255304,
      "tokens_seen": 2171797504
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.90579361258225e-05,
      "loss": 1.3558,
      "theoretical_loss": 3.4355431596643626,
      "tokens_seen": 2171928576
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.904991173166427e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.4355262936060753,
      "tokens_seen": 2172059648
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.904188733750602e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.43550942885049,
      "tokens_seen": 2172190720
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.903386294334779e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.4354925653974266,
      "tokens_seen": 2172321792
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.902583854918954e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.435475703246707,
      "tokens_seen": 2172452864
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.90178141550313e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.4354588423981514,
      "tokens_seen": 2172583936
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.900978976087306e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.4354419828515805,
      "tokens_seen": 2172715008
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.900176536671482e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.435425124606816,
      "tokens_seen": 2172846080
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.899374097255658e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.4354082676636786,
      "tokens_seen": 2172977152
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.898571657839833e-05,
      "loss": 1.1484,
      "theoretical_loss": 3.4353914120219886,
      "tokens_seen": 2173108224
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.897769218424009e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.4353745576815675,
      "tokens_seen": 2173239296
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.896966779008185e-05,
      "loss": 1.3471,
      "theoretical_loss": 3.4353577046422368,
      "tokens_seen": 2173370368
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.896164339592361e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.435340852903817,
      "tokens_seen": 2173501440
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.895361900176537e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.43532400246613,
      "tokens_seen": 2173632512
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.894559460760713e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.4353071533289956,
      "tokens_seen": 2173763584
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.89375702134489e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.435290305492236,
      "tokens_seen": 2173894656
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.892954581929065e-05,
      "loss": 1.26,
      "theoretical_loss": 3.435273458955673,
      "tokens_seen": 2174025728
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.89215214251324e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.4352566137191265,
      "tokens_seen": 2174156800
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.9926750063896179,
      "objective/train/avg_token_score": 0.0073250047862529755,
      "objective/train/docs_used": 791682,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3066558837890625,
      "objective/train/original_loss": 1.315915584564209,
      "objective/train/theoretical_loss": 3.435239769782419,
      "objective/train/tokens_used": 544746976,
      "objective/train/unlikelihood_loss": 0.023318398743867874,
      "theoretical_loss": 3.435239769782419,
      "tokens_seen": 2174287872
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.891349703097417e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.435239769782419,
      "tokens_seen": 2174287872
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.890547263681592e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.4352229271453707,
      "tokens_seen": 2174418944
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.889744824265769e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.435206085807804,
      "tokens_seen": 2174550016
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.888942384849944e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.43518924576954,
      "tokens_seen": 2174681088
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.88813994543412e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.4351724070304,
      "tokens_seen": 2174812160
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.887337506018296e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.4351555695902056,
      "tokens_seen": 2174943232
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.886535066602471e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.4351387334487784,
      "tokens_seen": 2175074304
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.885732627186648e-05,
      "loss": 1.2819,
      "theoretical_loss": 3.4351218986059395,
      "tokens_seen": 2175205376
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.884930187770823e-05,
      "loss": 1.2259,
      "theoretical_loss": 3.4351050650615114,
      "tokens_seen": 2175336448
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.884127748355e-05,
      "loss": 1.2332,
      "theoretical_loss": 3.435088232815315,
      "tokens_seen": 2175467520
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.883325308939175e-05,
      "loss": 1.1811,
      "theoretical_loss": 3.435071401867172,
      "tokens_seen": 2175598592
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.88252286952335e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.4350545722169046,
      "tokens_seen": 2175729664
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.881720430107527e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.435037743864334,
      "tokens_seen": 2175860736
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.880917990691703e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.4350209168092825,
      "tokens_seen": 2175991808
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.880115551275879e-05,
      "loss": 1.2141,
      "theoretical_loss": 3.4350040910515713,
      "tokens_seen": 2176122880
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.879313111860055e-05,
      "loss": 1.2098,
      "theoretical_loss": 3.4349872665910226,
      "tokens_seen": 2176253952
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.87851067244423e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.4349704434274586,
      "tokens_seen": 2176385024
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.877708233028407e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.4349536215607004,
      "tokens_seen": 2176516096
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.876905793612582e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.4349368009905707,
      "tokens_seen": 2176647168
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.876103354196759e-05,
      "loss": 1.2141,
      "theoretical_loss": 3.4349199817168907,
      "tokens_seen": 2176778240
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.875300914780934e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.4349031637394836,
      "tokens_seen": 2176909312
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.87449847536511e-05,
      "loss": 1.3488,
      "theoretical_loss": 3.4348863470581703,
      "tokens_seen": 2177040384
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.873696035949286e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.4348695316727733,
      "tokens_seen": 2177171456
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.872893596533461e-05,
      "loss": 1.3406,
      "theoretical_loss": 3.4348527175831154,
      "tokens_seen": 2177302528
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.872091157117638e-05,
      "loss": 1.3584,
      "theoretical_loss": 3.4348359047890176,
      "tokens_seen": 2177433600
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.871288717701813e-05,
      "loss": 1.286,
      "theoretical_loss": 3.4348190932903027,
      "tokens_seen": 2177564672
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.87048627828599e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.434802283086793,
      "tokens_seen": 2177695744
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.869683838870165e-05,
      "loss": 1.244,
      "theoretical_loss": 3.4347854741783106,
      "tokens_seen": 2177826816
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.86888139945434e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.4347686665646777,
      "tokens_seen": 2177957888
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.868078960038517e-05,
      "loss": 1.265,
      "theoretical_loss": 3.4347518602457168,
      "tokens_seen": 2178088960
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.867276520622692e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.4347350552212506,
      "tokens_seen": 2178220032
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.866474081206869e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.4347182514911005,
      "tokens_seen": 2178351104
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.865671641791044e-05,
      "loss": 1.1986,
      "theoretical_loss": 3.43470144905509,
      "tokens_seen": 2178482176
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.864869202375221e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.4346846479130413,
      "tokens_seen": 2178613248
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.864066762959396e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.4346678480647768,
      "tokens_seen": 2178744320
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.863264323543572e-05,
      "loss": 1.2138,
      "theoretical_loss": 3.4346510495101183,
      "tokens_seen": 2178875392
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.862461884127748e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.4346342522488897,
      "tokens_seen": 2179006464
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.861659444711924e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.4346174562809133,
      "tokens_seen": 2179137536
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.8608570052961e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.434600661606011,
      "tokens_seen": 2179268608
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.860054565880276e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.4345838682240064,
      "tokens_seen": 2179399680
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.859252126464452e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.4345670761347216,
      "tokens_seen": 2179530752
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.858449687048628e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.4345502853379792,
      "tokens_seen": 2179661824
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.857647247632803e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.4345334958336027,
      "tokens_seen": 2179792896
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.85684480821698e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.4345167076214143,
      "tokens_seen": 2179923968
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.856042368801155e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.4344999207012377,
      "tokens_seen": 2180055040
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.855239929385332e-05,
      "loss": 1.3902,
      "theoretical_loss": 3.4344831350728944,
      "tokens_seen": 2180186112
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.854437489969507e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.4344663507362085,
      "tokens_seen": 2180317184
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.853635050553684e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.434449567691003,
      "tokens_seen": 2180448256
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.852832611137859e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.4344327859371004,
      "tokens_seen": 2180579328
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.852030171722034e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.4344160054743234,
      "tokens_seen": 2180710400
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.9938268065452576,
      "objective/train/avg_token_score": 0.006173178553581238,
      "objective/train/docs_used": 794081,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2653727531433105,
      "objective/train/original_loss": 1.2741312980651855,
      "objective/train/theoretical_loss": 3.434399226302496,
      "objective/train/tokens_used": 551300576,
      "objective/train/unlikelihood_loss": 0.01693965680897236,
      "theoretical_loss": 3.434399226302496,
      "tokens_seen": 2180841472
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.851227732306211e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.434399226302496,
      "tokens_seen": 2180841472
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.850425292890386e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.4343824484214407,
      "tokens_seen": 2180972544
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.849622853474563e-05,
      "loss": 1.3119,
      "theoretical_loss": 3.4343656718309807,
      "tokens_seen": 2181103616
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.848820414058738e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.434348896530939,
      "tokens_seen": 2181234688
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.848017974642915e-05,
      "loss": 1.2257,
      "theoretical_loss": 3.4343321225211394,
      "tokens_seen": 2181365760
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.84721553522709e-05,
      "loss": 1.332,
      "theoretical_loss": 3.434315349801405,
      "tokens_seen": 2181496832
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.846413095811267e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.4342985783715587,
      "tokens_seen": 2181627904
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.845610656395442e-05,
      "loss": 1.3526,
      "theoretical_loss": 3.434281808231424,
      "tokens_seen": 2181758976
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.844808216979617e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.434265039380824,
      "tokens_seen": 2181890048
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.844005777563794e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.4342482718195826,
      "tokens_seen": 2182021120
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.84320333814797e-05,
      "loss": 1.2198,
      "theoretical_loss": 3.434231505547523,
      "tokens_seen": 2182152192
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.842400898732146e-05,
      "loss": 1.253,
      "theoretical_loss": 3.4342147405644683,
      "tokens_seen": 2182283264
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.841598459316321e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.4341979768702426,
      "tokens_seen": 2182414336
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.840796019900498e-05,
      "loss": 1.3332,
      "theoretical_loss": 3.434181214464669,
      "tokens_seen": 2182545408
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.839993580484673e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.4341644533475715,
      "tokens_seen": 2182676480
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.839191141068849e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.434147693518773,
      "tokens_seen": 2182807552
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.838388701653025e-05,
      "loss": 1.3328,
      "theoretical_loss": 3.4341309349780977,
      "tokens_seen": 2182938624
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.8375862622372e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.4341141777253688,
      "tokens_seen": 2183069696
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.836783822821377e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.4340974217604106,
      "tokens_seen": 2183200768
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.835981383405553e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.4340806670830464,
      "tokens_seen": 2183331840
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.835178943989729e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.4340639136931,
      "tokens_seen": 2183462912
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.834376504573905e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.4340471615903954,
      "tokens_seen": 2183593984
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.83357406515808e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.434030410774756,
      "tokens_seen": 2183725056
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.832771625742257e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.434013661246006,
      "tokens_seen": 2183856128
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.831969186326432e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.43399691300397,
      "tokens_seen": 2183987200
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.831166746910609e-05,
      "loss": 1.3627,
      "theoretical_loss": 3.43398016604847,
      "tokens_seen": 2184118272
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.830364307494784e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.433963420379332,
      "tokens_seen": 2184249344
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.82956186807896e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.433946675996379,
      "tokens_seen": 2184380416
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.828759428663136e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.4339299328994355,
      "tokens_seen": 2184511488
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.827956989247311e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.433913191088325,
      "tokens_seen": 2184642560
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.827154549831488e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.433896450562872,
      "tokens_seen": 2184773632
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.826352110415663e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.4338797113229007,
      "tokens_seen": 2184904704
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.82554967099984e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.4338629733682353,
      "tokens_seen": 2185035776
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.824747231584015e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.4338462366986997,
      "tokens_seen": 2185166848
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.823944792168192e-05,
      "loss": 1.344,
      "theoretical_loss": 3.4338295013141185,
      "tokens_seen": 2185297920
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.823142352752367e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.4338127672143157,
      "tokens_seen": 2185428992
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.822339913336542e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.4337960343991156,
      "tokens_seen": 2185560064
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.821537473920719e-05,
      "loss": 1.3283,
      "theoretical_loss": 3.4337793028683428,
      "tokens_seen": 2185691136
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.820735034504894e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.4337625726218217,
      "tokens_seen": 2185822208
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.819932595089071e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.4337458436593766,
      "tokens_seen": 2185953280
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.819130155673246e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.433729115980832,
      "tokens_seen": 2186084352
    },
    {
      "epoch": 0.32,
      "learning_rate": 6.818327716257423e-05,
      "loss": 1.354,
      "theoretical_loss": 3.433712389586012,
      "tokens_seen": 2186215424
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.817525276841598e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.433695664474742,
      "tokens_seen": 2186346496
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.816722837425775e-05,
      "loss": 1.34,
      "theoretical_loss": 3.433678940646846,
      "tokens_seen": 2186477568
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.81592039800995e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.4336622181021488,
      "tokens_seen": 2186608640
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.815117958594126e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.4336454968404744,
      "tokens_seen": 2186739712
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.814315519178302e-05,
      "loss": 1.3502,
      "theoretical_loss": 3.4336287768616485,
      "tokens_seen": 2186870784
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.813513079762478e-05,
      "loss": 1.273,
      "theoretical_loss": 3.4336120581654948,
      "tokens_seen": 2187001856
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.812710640346654e-05,
      "loss": 1.3668,
      "theoretical_loss": 3.4335953407518387,
      "tokens_seen": 2187132928
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.81190820093083e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.4335786246205044,
      "tokens_seen": 2187264000
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.9942664504051208,
      "objective/train/avg_token_score": 0.005733543075621128,
      "objective/train/docs_used": 796380,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.391446828842163,
      "objective/train/original_loss": 1.400647521018982,
      "objective/train/theoretical_loss": 3.4335619097713175,
      "objective/train/tokens_used": 557854176,
      "objective/train/unlikelihood_loss": 0.013015193864703178,
      "theoretical_loss": 3.4335619097713175,
      "tokens_seen": 2187395072
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.811105761515006e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.4335619097713175,
      "tokens_seen": 2187395072
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.810303322099182e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.4335451962041024,
      "tokens_seen": 2187526144
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.809500882683357e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.4335284839186837,
      "tokens_seen": 2187657216
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.808698443267534e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.433511772914887,
      "tokens_seen": 2187788288
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.807896003851709e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.4334950631925363,
      "tokens_seen": 2187919360
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.807093564435886e-05,
      "loss": 1.313,
      "theoretical_loss": 3.4334783547514576,
      "tokens_seen": 2188050432
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.806291125020061e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.4334616475914754,
      "tokens_seen": 2188181504
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.805488685604238e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.4334449417124144,
      "tokens_seen": 2188312576
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.804686246188413e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.4334282371141,
      "tokens_seen": 2188443648
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.803883806772588e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.4334115337963578,
      "tokens_seen": 2188574720
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.803081367356765e-05,
      "loss": 1.326,
      "theoretical_loss": 3.433394831759012,
      "tokens_seen": 2188705792
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.80227892794094e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.4333781310018887,
      "tokens_seen": 2188836864
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.801476488525117e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.433361431524813,
      "tokens_seen": 2188967936
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.800674049109292e-05,
      "loss": 1.328,
      "theoretical_loss": 3.4333447333276093,
      "tokens_seen": 2189099008
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.799871609693469e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.4333280364101038,
      "tokens_seen": 2189230080
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.799069170277644e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.433311340772121,
      "tokens_seen": 2189361152
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.79826673086182e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.433294646413487,
      "tokens_seen": 2189492224
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.797464291445996e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.4332779533340263,
      "tokens_seen": 2189623296
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.796661852030171e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.433261261533566,
      "tokens_seen": 2189754368
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.795859412614348e-05,
      "loss": 1.3197,
      "theoretical_loss": 3.4332445710119295,
      "tokens_seen": 2189885440
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.795056973198523e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.4332278817689437,
      "tokens_seen": 2190016512
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.7942545337827e-05,
      "loss": 1.3487,
      "theoretical_loss": 3.433211193804434,
      "tokens_seen": 2190147584
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.793452094366875e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.433194507118225,
      "tokens_seen": 2190278656
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.792649654951051e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.433177821710143,
      "tokens_seen": 2190409728
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.791847215535227e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.433161137580014,
      "tokens_seen": 2190540800
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.791044776119403e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.4331444547276635,
      "tokens_seen": 2190671872
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.79024233670358e-05,
      "loss": 1.253,
      "theoretical_loss": 3.433127773152916,
      "tokens_seen": 2190802944
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.789439897287755e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.433111092855599,
      "tokens_seen": 2190934016
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.788637457871931e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.433094413835537,
      "tokens_seen": 2191065088
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.787835018456107e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.4330777360925566,
      "tokens_seen": 2191196160
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.787032579040283e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.433061059626483,
      "tokens_seen": 2191327232
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.786230139624459e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.4330443844371423,
      "tokens_seen": 2191458304
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.785427700208634e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.4330277105243603,
      "tokens_seen": 2191589376
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.78462526079281e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.4330110378879635,
      "tokens_seen": 2191720448
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.783822821376986e-05,
      "loss": 1.355,
      "theoretical_loss": 3.432994366527777,
      "tokens_seen": 2191851520
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.783020381961163e-05,
      "loss": 1.301,
      "theoretical_loss": 3.4329776964436274,
      "tokens_seen": 2191982592
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.782217942545338e-05,
      "loss": 1.2512,
      "theoretical_loss": 3.432961027635341,
      "tokens_seen": 2192113664
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.781415503129515e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.4329443601027427,
      "tokens_seen": 2192244736
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.78061306371369e-05,
      "loss": 1.1982,
      "theoretical_loss": 3.43292769384566,
      "tokens_seen": 2192375808
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.779810624297865e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.4329110288639177,
      "tokens_seen": 2192506880
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.779008184882042e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.432894365157343,
      "tokens_seen": 2192637952
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.778205745466217e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.432877702725762,
      "tokens_seen": 2192769024
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.777403306050394e-05,
      "loss": 1.3163,
      "theoretical_loss": 3.4328610415690006,
      "tokens_seen": 2192900096
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.776600866634569e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.432844381686885,
      "tokens_seen": 2193031168
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.775798427218746e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.4328277230792423,
      "tokens_seen": 2193162240
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.774995987802921e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.4328110657458977,
      "tokens_seen": 2193293312
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.774193548387096e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.4327944096866783,
      "tokens_seen": 2193424384
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.773391108971273e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.43277775490141,
      "tokens_seen": 2193555456
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.772588669555448e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.43276110138992,
      "tokens_seen": 2193686528
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.771786230139625e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.4327444491520342,
      "tokens_seen": 2193817600
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.9935293197631836,
      "objective/train/avg_token_score": 0.006470712833106518,
      "objective/train/docs_used": 798764,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3236756324768066,
      "objective/train/original_loss": 1.3330610990524292,
      "objective/train/theoretical_loss": 3.4327277981875794,
      "objective/train/tokens_used": 564407776,
      "objective/train/unlikelihood_loss": 0.015855329111218452,
      "theoretical_loss": 3.4327277981875794,
      "tokens_seen": 2193948672
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.7709837907238e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.4327277981875794,
      "tokens_seen": 2193948672
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.770181351307977e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.432711148496382,
      "tokens_seen": 2194079744
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.769378911892152e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.4326945000782687,
      "tokens_seen": 2194210816
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.768576472476328e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.432677852933066,
      "tokens_seen": 2194341888
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.767774033060504e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.4326612070606006,
      "tokens_seen": 2194472960
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.76697159364468e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.432644562460699,
      "tokens_seen": 2194604032
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.766169154228856e-05,
      "loss": 1.2433,
      "theoretical_loss": 3.4326279191331883,
      "tokens_seen": 2194735104
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.765366714813032e-05,
      "loss": 1.3511,
      "theoretical_loss": 3.432611277077895,
      "tokens_seen": 2194866176
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.764564275397208e-05,
      "loss": 1.3325,
      "theoretical_loss": 3.432594636294646,
      "tokens_seen": 2194997248
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.763761835981384e-05,
      "loss": 1.2673,
      "theoretical_loss": 3.4325779967832677,
      "tokens_seen": 2195128320
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.762959396565559e-05,
      "loss": 1.277,
      "theoretical_loss": 3.4325613585435875,
      "tokens_seen": 2195259392
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.762156957149736e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.4325447215754323,
      "tokens_seen": 2195390464
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.761354517733911e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.4325280858786287,
      "tokens_seen": 2195521536
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.760552078318088e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.4325114514530037,
      "tokens_seen": 2195652608
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.759749638902263e-05,
      "loss": 1.383,
      "theoretical_loss": 3.4324948182983843,
      "tokens_seen": 2195783680
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.75894719948644e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.4324781864145977,
      "tokens_seen": 2195914752
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.758144760070615e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.432461555801471,
      "tokens_seen": 2196045824
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.75734232065479e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.432444926458831,
      "tokens_seen": 2196176896
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.756539881238967e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.4324282983865047,
      "tokens_seen": 2196307968
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.755737441823142e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.43241167158432,
      "tokens_seen": 2196439040
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.754935002407319e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.432395046052103,
      "tokens_seen": 2196570112
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.754132562991494e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.432378421789682,
      "tokens_seen": 2196701184
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.753330123575671e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.432361798796884,
      "tokens_seen": 2196832256
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.752527684159846e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.4323451770735356,
      "tokens_seen": 2196963328
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.751725244744023e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.4323285566194652,
      "tokens_seen": 2197094400
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.750922805328198e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.4323119374344992,
      "tokens_seen": 2197225472
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.750120365912373e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.4322953195184653,
      "tokens_seen": 2197356544
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.74931792649655e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.432278702871191,
      "tokens_seen": 2197487616
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.748515487080725e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.432262087492504,
      "tokens_seen": 2197618688
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.747713047664902e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.4322454733822316,
      "tokens_seen": 2197749760
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.746910608249077e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.432228860540201,
      "tokens_seen": 2197880832
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.746108168833254e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.4322122489662403,
      "tokens_seen": 2198011904
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.74530572941743e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.432195638660177,
      "tokens_seen": 2198142976
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.744503290001605e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.432179029621838,
      "tokens_seen": 2198274048
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.743700850585781e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.4321624218510522,
      "tokens_seen": 2198405120
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.742898411169957e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.4321458153476465,
      "tokens_seen": 2198536192
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.742095971754133e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.432129210111449,
      "tokens_seen": 2198667264
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.741293532338309e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.4321126061422866,
      "tokens_seen": 2198798336
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.740491092922485e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.4320960034399883,
      "tokens_seen": 2198929408
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.73968865350666e-05,
      "loss": 1.2144,
      "theoretical_loss": 3.432079402004381,
      "tokens_seen": 2199060480
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.738886214090836e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.4320628018352934,
      "tokens_seen": 2199191552
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.738083774675013e-05,
      "loss": 1.2465,
      "theoretical_loss": 3.4320462029325527,
      "tokens_seen": 2199322624
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.737281335259188e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.432029605295987,
      "tokens_seen": 2199453696
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.736478895843365e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.432013008925425,
      "tokens_seen": 2199584768
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.73567645642754e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.4319964138206935,
      "tokens_seen": 2199715840
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.734874017011717e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.4319798199816214,
      "tokens_seen": 2199846912
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.734071577595892e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.431963227408036,
      "tokens_seen": 2199977984
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.733269138180067e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.4319466360997666,
      "tokens_seen": 2200109056
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.732466698764244e-05,
      "loss": 1.3456,
      "theoretical_loss": 3.43193004605664,
      "tokens_seen": 2200240128
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.731664259348419e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.4319134572784855,
      "tokens_seen": 2200371200
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.9917894601821899,
      "objective/train/avg_token_score": 0.0082105016335845,
      "objective/train/docs_used": 801168,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2335906028747559,
      "objective/train/original_loss": 1.2447292804718018,
      "objective/train/theoretical_loss": 3.4318968697651306,
      "objective/train/tokens_used": 570961376,
      "objective/train/unlikelihood_loss": 0.02282066084444523,
      "theoretical_loss": 3.4318968697651306,
      "tokens_seen": 2200502272
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.730861819932596e-05,
      "loss": 1.2485,
      "theoretical_loss": 3.4318968697651306,
      "tokens_seen": 2200502272
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.730059380516771e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.4318802835164037,
      "tokens_seen": 2200633344
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.729256941100948e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.431863698532133,
      "tokens_seen": 2200764416
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.728454501685123e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.4318471148121468,
      "tokens_seen": 2200895488
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.727652062269298e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.431830532356274,
      "tokens_seen": 2201026560
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.726849622853475e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.431813951164342,
      "tokens_seen": 2201157632
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.72604718343765e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.43179737123618,
      "tokens_seen": 2201288704
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.725244744021827e-05,
      "loss": 1.2398,
      "theoretical_loss": 3.431780792571616,
      "tokens_seen": 2201419776
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.724442304606002e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.431764215170479,
      "tokens_seen": 2201550848
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.723639865190179e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.431747639032597,
      "tokens_seen": 2201681920
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.722837425774354e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.4317310641577987,
      "tokens_seen": 2201812992
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.722034986358531e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.4317144905459127,
      "tokens_seen": 2201944064
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.721232546942706e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.4316979181967673,
      "tokens_seen": 2202075136
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.720430107526882e-05,
      "loss": 1.286,
      "theoretical_loss": 3.4316813471101915,
      "tokens_seen": 2202206208
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.719627668111058e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.431664777286014,
      "tokens_seen": 2202337280
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.718825228695234e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.431648208724063,
      "tokens_seen": 2202468352
    },
    {
      "epoch": 0.33,
      "learning_rate": 6.71802278927941e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.431631641424168,
      "tokens_seen": 2202599424
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.717220349863586e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.431615075386157,
      "tokens_seen": 2202730496
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.716417910447762e-05,
      "loss": 1.45,
      "theoretical_loss": 3.4315985106098594,
      "tokens_seen": 2202861568
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.715615471031938e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.431581947095104,
      "tokens_seen": 2202992640
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.714813031616113e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.43156538484172,
      "tokens_seen": 2203123712
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.71401059220029e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.431548823849535,
      "tokens_seen": 2203254784
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.713208152784465e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.431532264118379,
      "tokens_seen": 2203385856
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.712405713368642e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.4315157056480805,
      "tokens_seen": 2203516928
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.711603273952817e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.431499148438469,
      "tokens_seen": 2203648000
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.710800834536994e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.4314825924893735,
      "tokens_seen": 2203779072
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.709998395121169e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.4314660378006225,
      "tokens_seen": 2203910144
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.709195955705344e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.4314494843720458,
      "tokens_seen": 2204041216
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.708393516289521e-05,
      "loss": 1.2327,
      "theoretical_loss": 3.431432932203472,
      "tokens_seen": 2204172288
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.707591076873696e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.4314163812947305,
      "tokens_seen": 2204303360
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.706788637457873e-05,
      "loss": 1.3655,
      "theoretical_loss": 3.4313998316456504,
      "tokens_seen": 2204434432
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.705986198042048e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.431383283256061,
      "tokens_seen": 2204565504
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.705183758626225e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.4313667361257916,
      "tokens_seen": 2204696576
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.7043813192104e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.4313501902546717,
      "tokens_seen": 2204827648
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.703578879794575e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.43133364564253,
      "tokens_seen": 2204958720
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.702776440378752e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.4313171022891966,
      "tokens_seen": 2205089792
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.701974000962927e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.431300560194501,
      "tokens_seen": 2205220864
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.701171561547104e-05,
      "loss": 1.3885,
      "theoretical_loss": 3.4312840193582717,
      "tokens_seen": 2205351936
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.70036912213128e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.431267479780339,
      "tokens_seen": 2205483008
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.699566682715456e-05,
      "loss": 1.3164,
      "theoretical_loss": 3.431250941460532,
      "tokens_seen": 2205614080
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.698764243299631e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.4312344043986807,
      "tokens_seen": 2205745152
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.697961803883807e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.431217868594614,
      "tokens_seen": 2205876224
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.697159364467983e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.431201334048162,
      "tokens_seen": 2206007296
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.696356925052159e-05,
      "loss": 1.283,
      "theoretical_loss": 3.4311848007591546,
      "tokens_seen": 2206138368
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.695554485636335e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.4311682687274208,
      "tokens_seen": 2206269440
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.694752046220511e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.4311517379527903,
      "tokens_seen": 2206400512
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.693949606804687e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.4311352084350935,
      "tokens_seen": 2206531584
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.693147167388863e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.4311186801741598,
      "tokens_seen": 2206662656
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.69234472797304e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.4311021531698187,
      "tokens_seen": 2206793728
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.691542288557215e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.4310856274219006,
      "tokens_seen": 2206924800
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.9934345483779907,
      "objective/train/avg_token_score": 0.006565408315509558,
      "objective/train/docs_used": 803600,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3014805316925049,
      "objective/train/original_loss": 1.3105437755584717,
      "objective/train/theoretical_loss": 3.431069102930235,
      "objective/train/tokens_used": 577514976,
      "objective/train/unlikelihood_loss": 0.02175849676132202,
      "theoretical_loss": 3.431069102930235,
      "tokens_seen": 2207055872
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.69073984914139e-05,
      "loss": 1.3499,
      "theoretical_loss": 3.431069102930235,
      "tokens_seen": 2207055872
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.689937409725567e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.4310525796946516,
      "tokens_seen": 2207186944
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.689134970309742e-05,
      "loss": 1.2209,
      "theoretical_loss": 3.431036057714981,
      "tokens_seen": 2207318016
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.688332530893919e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.4310195369910526,
      "tokens_seen": 2207449088
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.687530091478094e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.431003017522697,
      "tokens_seen": 2207580160
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.68672765206227e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.4309864993097436,
      "tokens_seen": 2207711232
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.685925212646446e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.430969982352023,
      "tokens_seen": 2207842304
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.685122773230621e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.4309534666493646,
      "tokens_seen": 2207973376
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.684320333814798e-05,
      "loss": 1.3543,
      "theoretical_loss": 3.4309369522015993,
      "tokens_seen": 2208104448
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.683517894398973e-05,
      "loss": 1.2379,
      "theoretical_loss": 3.430920439008557,
      "tokens_seen": 2208235520
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.68271545498315e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.4309039270700676,
      "tokens_seen": 2208366592
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.681913015567325e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.430887416385962,
      "tokens_seen": 2208497664
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.681110576151502e-05,
      "loss": 1.286,
      "theoretical_loss": 3.43087090695607,
      "tokens_seen": 2208628736
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.680308136735677e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.4308543987802214,
      "tokens_seen": 2208759808
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.679505697319852e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.430837891858247,
      "tokens_seen": 2208890880
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.678703257904029e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.430821386189978,
      "tokens_seen": 2209021952
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.677900818488204e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.4308048817752437,
      "tokens_seen": 2209153024
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.677098379072381e-05,
      "loss": 1.3879,
      "theoretical_loss": 3.430788378613875,
      "tokens_seen": 2209284096
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.676295939656556e-05,
      "loss": 1.352,
      "theoretical_loss": 3.4307718767057023,
      "tokens_seen": 2209415168
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.675493500240733e-05,
      "loss": 1.3519,
      "theoretical_loss": 3.4307553760505556,
      "tokens_seen": 2209546240
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.674691060824908e-05,
      "loss": 1.2423,
      "theoretical_loss": 3.4307388766482667,
      "tokens_seen": 2209677312
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.673888621409084e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.4307223784986647,
      "tokens_seen": 2209808384
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.67308618199326e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.430705881601581,
      "tokens_seen": 2209939456
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.672283742577436e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.430689385956846,
      "tokens_seen": 2210070528
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.671481303161612e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.430672891564291,
      "tokens_seen": 2210201600
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.670678863745788e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.4306563984237455,
      "tokens_seen": 2210332672
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.669876424329964e-05,
      "loss": 1.296,
      "theoretical_loss": 3.430639906535041,
      "tokens_seen": 2210463744
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.66907398491414e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.4306234158980082,
      "tokens_seen": 2210594816
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.668271545498315e-05,
      "loss": 1.2819,
      "theoretical_loss": 3.4306069265124783,
      "tokens_seen": 2210725888
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.667469106082492e-05,
      "loss": 1.2395,
      "theoretical_loss": 3.4305904383782813,
      "tokens_seen": 2210856960
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.666666666666667e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.4305739514952482,
      "tokens_seen": 2210988032
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.665864227250844e-05,
      "loss": 1.3631,
      "theoretical_loss": 3.4305574658632105,
      "tokens_seen": 2211119104
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.665061787835019e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.4305409814819985,
      "tokens_seen": 2211250176
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.664259348419196e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.430524498351444,
      "tokens_seen": 2211381248
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.663456909003371e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.430508016471377,
      "tokens_seen": 2211512320
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.662654469587546e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.430491535841629,
      "tokens_seen": 2211643392
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.661852030171723e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.4304750564620305,
      "tokens_seen": 2211774464
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.661049590755898e-05,
      "loss": 1.3997,
      "theoretical_loss": 3.4304585783324133,
      "tokens_seen": 2211905536
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.660247151340075e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.4304421014526083,
      "tokens_seen": 2212036608
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.65944471192425e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.430425625822447,
      "tokens_seen": 2212167680
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.658642272508427e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.4304091514417605,
      "tokens_seen": 2212298752
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.657839833092602e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.430392678310379,
      "tokens_seen": 2212429824
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.657037393676779e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.430376206428135,
      "tokens_seen": 2212560896
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.656234954260954e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.4303597357948594,
      "tokens_seen": 2212691968
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.65543251484513e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.4303432664103832,
      "tokens_seen": 2212823040
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.654630075429306e-05,
      "loss": 1.3723,
      "theoretical_loss": 3.4303267982745385,
      "tokens_seen": 2212954112
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.653827636013481e-05,
      "loss": 1.245,
      "theoretical_loss": 3.4303103313871555,
      "tokens_seen": 2213085184
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.653025196597658e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.430293865748067,
      "tokens_seen": 2213216256
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.652222757181833e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.4302774013571034,
      "tokens_seen": 2213347328
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.65142031776601e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.4302609382140967,
      "tokens_seen": 2213478400
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.9922282099723816,
      "objective/train/avg_token_score": 0.007771804928779602,
      "objective/train/docs_used": 805966,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3015726804733276,
      "objective/train/original_loss": 1.31229829788208,
      "objective/train/theoretical_loss": 3.430244476318878,
      "objective/train/tokens_used": 584068576,
      "objective/train/unlikelihood_loss": 0.021249018609523773,
      "theoretical_loss": 3.430244476318878,
      "tokens_seen": 2213609472
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.650617878350185e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.430244476318878,
      "tokens_seen": 2213609472
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.649815438934361e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4302280156712794,
      "tokens_seen": 2213740544
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.649012999518537e-05,
      "loss": 1.3262,
      "theoretical_loss": 3.430211556271132,
      "tokens_seen": 2213871616
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.648210560102713e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.4301950981182676,
      "tokens_seen": 2214002688
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.64740812068689e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.4301786412125184,
      "tokens_seen": 2214133760
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.646605681271065e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.4301621855537148,
      "tokens_seen": 2214264832
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.645803241855241e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.43014573114169,
      "tokens_seen": 2214395904
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.645000802439417e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.4301292779762744,
      "tokens_seen": 2214526976
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.644198363023592e-05,
      "loss": 1.33,
      "theoretical_loss": 3.4301128260573006,
      "tokens_seen": 2214658048
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.643395923607769e-05,
      "loss": 1.3593,
      "theoretical_loss": 3.4300963753846005,
      "tokens_seen": 2214789120
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.642593484191944e-05,
      "loss": 1.2014,
      "theoretical_loss": 3.4300799259580055,
      "tokens_seen": 2214920192
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.64179104477612e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.430063477777347,
      "tokens_seen": 2215051264
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.640988605360296e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.4300470308424584,
      "tokens_seen": 2215182336
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.640186165944471e-05,
      "loss": 1.2155,
      "theoretical_loss": 3.4300305851531703,
      "tokens_seen": 2215313408
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.639383726528648e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.4300141407093157,
      "tokens_seen": 2215444480
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.638581287112823e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.429997697510726,
      "tokens_seen": 2215575552
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.637778847697e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.429981255557233,
      "tokens_seen": 2215706624
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.636976408281175e-05,
      "loss": 1.4009,
      "theoretical_loss": 3.4299648148486686,
      "tokens_seen": 2215837696
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.636173968865352e-05,
      "loss": 1.3452,
      "theoretical_loss": 3.429948375384866,
      "tokens_seen": 2215968768
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.635371529449527e-05,
      "loss": 1.215,
      "theoretical_loss": 3.4299319371656565,
      "tokens_seen": 2216099840
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.634569090033703e-05,
      "loss": 1.341,
      "theoretical_loss": 3.429915500190873,
      "tokens_seen": 2216230912
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.633766650617879e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.4298990644603466,
      "tokens_seen": 2216361984
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.632964211202055e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.4298826299739105,
      "tokens_seen": 2216493056
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.632161771786231e-05,
      "loss": 1.3039,
      "theoretical_loss": 3.429866196731396,
      "tokens_seen": 2216624128
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.631359332370407e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.429849764732637,
      "tokens_seen": 2216755200
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.630556892954582e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.4298333339774643,
      "tokens_seen": 2216886272
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.629754453538758e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.429816904465711,
      "tokens_seen": 2217017344
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.628952014122934e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.429800476197209,
      "tokens_seen": 2217148416
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.62814957470711e-05,
      "loss": 1.2186,
      "theoretical_loss": 3.4297840491717912,
      "tokens_seen": 2217279488
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.627347135291286e-05,
      "loss": 1.3512,
      "theoretical_loss": 3.42976762338929,
      "tokens_seen": 2217410560
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.626544695875462e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.4297511988495377,
      "tokens_seen": 2217541632
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.625742256459638e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.429734775552367,
      "tokens_seen": 2217672704
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.624939817043813e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.42971835349761,
      "tokens_seen": 2217803776
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.62413737762799e-05,
      "loss": 1.228,
      "theoretical_loss": 3.4297019326851004,
      "tokens_seen": 2217934848
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.623334938212165e-05,
      "loss": 1.3243,
      "theoretical_loss": 3.4296855131146695,
      "tokens_seen": 2218065920
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.622532498796342e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.429669094786151,
      "tokens_seen": 2218196992
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.621730059380517e-05,
      "loss": 1.3097,
      "theoretical_loss": 3.429652677699377,
      "tokens_seen": 2218328064
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.620927619964692e-05,
      "loss": 1.325,
      "theoretical_loss": 3.4296362618541805,
      "tokens_seen": 2218459136
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.620125180548869e-05,
      "loss": 1.27,
      "theoretical_loss": 3.429619847250394,
      "tokens_seen": 2218590208
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.619322741133044e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.4296034338878503,
      "tokens_seen": 2218721280
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.618520301717221e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.4295870217663826,
      "tokens_seen": 2218852352
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.617717862301396e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.4295706108858233,
      "tokens_seen": 2218983424
    },
    {
      "epoch": 0.34,
      "learning_rate": 6.616915422885573e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.4295542012460056,
      "tokens_seen": 2219114496
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.616112983469748e-05,
      "loss": 1.1828,
      "theoretical_loss": 3.4295377928467623,
      "tokens_seen": 2219245568
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.615310544053924e-05,
      "loss": 1.3382,
      "theoretical_loss": 3.4295213856879263,
      "tokens_seen": 2219376640
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.6145081046381e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.4295049797693307,
      "tokens_seen": 2219507712
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.613705665222276e-05,
      "loss": 1.229,
      "theoretical_loss": 3.4294885750908084,
      "tokens_seen": 2219638784
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.612903225806452e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.4294721716521925,
      "tokens_seen": 2219769856
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.612100786390628e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.429455769453316,
      "tokens_seen": 2219900928
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.611298346974803e-05,
      "loss": 1.23,
      "theoretical_loss": 3.4294393684940125,
      "tokens_seen": 2220032000
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.989526629447937,
      "objective/train/avg_token_score": 0.010473337024450302,
      "objective/train/docs_used": 808099,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2616746425628662,
      "objective/train/original_loss": 1.2737996578216553,
      "objective/train/theoretical_loss": 3.4294229687741145,
      "objective/train/tokens_used": 590622176,
      "objective/train/unlikelihood_loss": 0.03540501371026039,
      "theoretical_loss": 3.4294229687741145,
      "tokens_seen": 2220163072
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.61049590755898e-05,
      "loss": 1.2166,
      "theoretical_loss": 3.4294229687741145,
      "tokens_seen": 2220163072
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.609693468143155e-05,
      "loss": 1.3236,
      "theoretical_loss": 3.4294065702934553,
      "tokens_seen": 2220294144
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.608891028727332e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.4293901730518686,
      "tokens_seen": 2220425216
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.608088589311507e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.4293737770491868,
      "tokens_seen": 2220556288
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.607286149895684e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.429357382285244,
      "tokens_seen": 2220687360
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.606483710479859e-05,
      "loss": 1.3506,
      "theoretical_loss": 3.429340988759873,
      "tokens_seen": 2220818432
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.605681271064034e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.429324596472907,
      "tokens_seen": 2220949504
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.604878831648211e-05,
      "loss": 1.351,
      "theoretical_loss": 3.4293082054241797,
      "tokens_seen": 2221080576
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.604076392232386e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.429291815613525,
      "tokens_seen": 2221211648
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.603273952816563e-05,
      "loss": 1.2496,
      "theoretical_loss": 3.4292754270407753,
      "tokens_seen": 2221342720
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.602471513400738e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.4292590397057645,
      "tokens_seen": 2221473792
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.601669073984913e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.4292426536083265,
      "tokens_seen": 2221604864
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.60086663456909e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.429226268748294,
      "tokens_seen": 2221735936
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.600064195153265e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.429209885125502,
      "tokens_seen": 2221867008
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.599261755737442e-05,
      "loss": 1.2407,
      "theoretical_loss": 3.429193502739782,
      "tokens_seen": 2221998080
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.598459316321617e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.4291771215909694,
      "tokens_seen": 2222129152
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.597656876905794e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.4291607416788974,
      "tokens_seen": 2222260224
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.59685443748997e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.4291443630033993,
      "tokens_seen": 2222391296
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.596051998074145e-05,
      "loss": 1.3452,
      "theoretical_loss": 3.4291279855643086,
      "tokens_seen": 2222522368
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.595249558658321e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.42911160936146,
      "tokens_seen": 2222653440
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.594447119242497e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.429095234394687,
      "tokens_seen": 2222784512
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.593644679826673e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.429078860663823,
      "tokens_seen": 2222915584
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.592842240410849e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.429062488168702,
      "tokens_seen": 2223046656
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.592039800995025e-05,
      "loss": 1.2486,
      "theoretical_loss": 3.429046116909158,
      "tokens_seen": 2223177728
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.5912373615792e-05,
      "loss": 1.224,
      "theoretical_loss": 3.429029746885025,
      "tokens_seen": 2223308800
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.590434922163376e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.4290133780961365,
      "tokens_seen": 2223439872
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.589632482747553e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.428997010542327,
      "tokens_seen": 2223570944
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.588830043331728e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.4289806442234303,
      "tokens_seen": 2223702016
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.588027603915905e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.4289642791392803,
      "tokens_seen": 2223833088
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.58722516450008e-05,
      "loss": 1.3935,
      "theoretical_loss": 3.4289479152897115,
      "tokens_seen": 2223964160
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.586422725084257e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.4289315526745576,
      "tokens_seen": 2224095232
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.585620285668432e-05,
      "loss": 1.3164,
      "theoretical_loss": 3.428915191293653,
      "tokens_seen": 2224226304
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.584817846252607e-05,
      "loss": 1.3994,
      "theoretical_loss": 3.428898831146832,
      "tokens_seen": 2224357376
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.584015406836784e-05,
      "loss": 1.2073,
      "theoretical_loss": 3.428882472233928,
      "tokens_seen": 2224488448
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.583212967420959e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.4288661145547765,
      "tokens_seen": 2224619520
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.582410528005136e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.4288497581092106,
      "tokens_seen": 2224750592
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.581608088589311e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.428833402897065,
      "tokens_seen": 2224881664
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.580805649173488e-05,
      "loss": 1.3649,
      "theoretical_loss": 3.4288170489181744,
      "tokens_seen": 2225012736
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.580003209757663e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.428800696172373,
      "tokens_seen": 2225143808
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.579200770341838e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.428784344659495,
      "tokens_seen": 2225274880
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.578398330926015e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.4287679943793745,
      "tokens_seen": 2225405952
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.57759589151019e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.4287516453318467,
      "tokens_seen": 2225537024
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.576793452094367e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.428735297516746,
      "tokens_seen": 2225668096
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.575991012678542e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.4287189509339067,
      "tokens_seen": 2225799168
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.575188573262719e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.4287026055831626,
      "tokens_seen": 2225930240
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.574386133846894e-05,
      "loss": 1.2238,
      "theoretical_loss": 3.4286862614643496,
      "tokens_seen": 2226061312
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.57358369443107e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.428669918577302,
      "tokens_seen": 2226192384
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.572781255015246e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.428653576921854,
      "tokens_seen": 2226323456
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.571978815599422e-05,
      "loss": 1.3667,
      "theoretical_loss": 3.42863723649784,
      "tokens_seen": 2226454528
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.571176376183598e-05,
      "loss": 1.1956,
      "theoretical_loss": 3.428620897305096,
      "tokens_seen": 2226585600
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.9919843077659607,
      "objective/train/avg_token_score": 0.008015724830329418,
      "objective/train/docs_used": 810670,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.257180094718933,
      "objective/train/original_loss": 1.2687060832977295,
      "objective/train/theoretical_loss": 3.428604559343456,
      "objective/train/tokens_used": 597175776,
      "objective/train/unlikelihood_loss": 0.02167501300573349,
      "theoretical_loss": 3.428604559343456,
      "tokens_seen": 2226716672
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.570373936767774e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.428604559343456,
      "tokens_seen": 2226716672
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.56957149735195e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.4285882226127544,
      "tokens_seen": 2226847744
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.568769057936126e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.428571887112826,
      "tokens_seen": 2226978816
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.567966618520301e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.4285555528435068,
      "tokens_seen": 2227109888
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.567164179104478e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.4285392198046303,
      "tokens_seen": 2227240960
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.566361739688653e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.4285228879960323,
      "tokens_seen": 2227372032
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.56555930027283e-05,
      "loss": 1.291,
      "theoretical_loss": 3.4285065574175473,
      "tokens_seen": 2227503104
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.564756860857005e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.428490228069011,
      "tokens_seen": 2227634176
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.563954421441182e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.428473899950257,
      "tokens_seen": 2227765248
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.563151982025357e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.4284575730611215,
      "tokens_seen": 2227896320
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.562349542609534e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.428441247401439,
      "tokens_seen": 2228027392
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.561547103193709e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.428424922971045,
      "tokens_seen": 2228158464
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.560744663777884e-05,
      "loss": 1.2021,
      "theoretical_loss": 3.428408599769775,
      "tokens_seen": 2228289536
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.559942224362061e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.428392277797463,
      "tokens_seen": 2228420608
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.559139784946236e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.428375957053945,
      "tokens_seen": 2228551680
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.558337345530413e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.4283596375390557,
      "tokens_seen": 2228682752
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.557534906114588e-05,
      "loss": 1.2209,
      "theoretical_loss": 3.4283433192526314,
      "tokens_seen": 2228813824
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.556732466698765e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.4283270021945063,
      "tokens_seen": 2228944896
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.55593002728294e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.428310686364516,
      "tokens_seen": 2229075968
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.555127587867115e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.428294371762496,
      "tokens_seen": 2229207040
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.554325148451292e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.4282780583882815,
      "tokens_seen": 2229338112
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.553522709035467e-05,
      "loss": 1.3396,
      "theoretical_loss": 3.4282617462417084,
      "tokens_seen": 2229469184
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.552720269619644e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.428245435322612,
      "tokens_seen": 2229600256
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.55191783020382e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.4282291256308275,
      "tokens_seen": 2229731328
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.551115390787996e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.42821281716619,
      "tokens_seen": 2229862400
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.550312951372171e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.428196509928536,
      "tokens_seen": 2229993472
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.549510511956347e-05,
      "loss": 1.288,
      "theoretical_loss": 3.4281802039177007,
      "tokens_seen": 2230124544
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.548708072540523e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.4281638991335197,
      "tokens_seen": 2230255616
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.547905633124699e-05,
      "loss": 1.265,
      "theoretical_loss": 3.4281475955758287,
      "tokens_seen": 2230386688
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.547103193708875e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.428131293244463,
      "tokens_seen": 2230517760
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.54630075429305e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.4281149921392586,
      "tokens_seen": 2230648832
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.545498314877227e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.428098692260051,
      "tokens_seen": 2230779904
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.544695875461403e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.4280823936066764,
      "tokens_seen": 2230910976
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.543893436045578e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.4280660961789704,
      "tokens_seen": 2231042048
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.543090996629755e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.428049799976769,
      "tokens_seen": 2231173120
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.54228855721393e-05,
      "loss": 1.2209,
      "theoretical_loss": 3.4280335049999073,
      "tokens_seen": 2231304192
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.541486117798107e-05,
      "loss": 1.2138,
      "theoretical_loss": 3.428017211248222,
      "tokens_seen": 2231435264
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.540683678382282e-05,
      "loss": 1.3193,
      "theoretical_loss": 3.428000918721549,
      "tokens_seen": 2231566336
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.539881238966459e-05,
      "loss": 1.3531,
      "theoretical_loss": 3.4279846274197237,
      "tokens_seen": 2231697408
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.539078799550634e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.427968337342583,
      "tokens_seen": 2231828480
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.538276360134809e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.4279520484899617,
      "tokens_seen": 2231959552
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.537473920718986e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.427935760861697,
      "tokens_seen": 2232090624
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.536671481303161e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.427919474457624,
      "tokens_seen": 2232221696
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.535869041887338e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.4279031892775795,
      "tokens_seen": 2232352768
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.535066602471513e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.4278869053213996,
      "tokens_seen": 2232483840
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.53426416305569e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.4278706225889204,
      "tokens_seen": 2232614912
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.533461723639865e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.4278543410799776,
      "tokens_seen": 2232745984
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.532659284224042e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.427838060794408,
      "tokens_seen": 2232877056
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.531856844808217e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.4278217817320478,
      "tokens_seen": 2233008128
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.531054405392392e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.4278055038927335,
      "tokens_seen": 2233139200
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.9938318133354187,
      "objective/train/avg_token_score": 0.006168222054839134,
      "objective/train/docs_used": 813220,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2396076917648315,
      "objective/train/original_loss": 1.2475056648254395,
      "objective/train/theoretical_loss": 3.4277892272763006,
      "objective/train/tokens_used": 603729376,
      "objective/train/unlikelihood_loss": 0.018593713641166687,
      "theoretical_loss": 3.4277892272763006,
      "tokens_seen": 2233270272
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.530251965976569e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.4277892272763006,
      "tokens_seen": 2233270272
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.529449526560744e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.4277729518825866,
      "tokens_seen": 2233401344
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.528647087144921e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.427756677711427,
      "tokens_seen": 2233532416
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.527844647729096e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.427740404762659,
      "tokens_seen": 2233663488
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.527042208313273e-05,
      "loss": 1.2172,
      "theoretical_loss": 3.427724133036118,
      "tokens_seen": 2233794560
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.526239768897448e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.4277078625316415,
      "tokens_seen": 2233925632
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.525437329481624e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.4276915932490657,
      "tokens_seen": 2234056704
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.5246348900658e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.4276753251882273,
      "tokens_seen": 2234187776
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.523832450649976e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.4276590583489623,
      "tokens_seen": 2234318848
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.523030011234152e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.4276427927311075,
      "tokens_seen": 2234449920
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.522227571818328e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.4276265283345,
      "tokens_seen": 2234580992
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.521425132402504e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.4276102651589766,
      "tokens_seen": 2234712064
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.52062269298668e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4275940032043732,
      "tokens_seen": 2234843136
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.519820253570855e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.427577742470527,
      "tokens_seen": 2234974208
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.519017814155032e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.427561482957275,
      "tokens_seen": 2235105280
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.518215374739207e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.427545224664454,
      "tokens_seen": 2235236352
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.517412935323384e-05,
      "loss": 1.345,
      "theoretical_loss": 3.4275289675919,
      "tokens_seen": 2235367424
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.516610495907559e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.427512711739451,
      "tokens_seen": 2235498496
    },
    {
      "epoch": 0.35,
      "learning_rate": 6.515808056491736e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.427496457106943,
      "tokens_seen": 2235629568
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.515005617075911e-05,
      "loss": 1.333,
      "theoretical_loss": 3.4274802036942136,
      "tokens_seen": 2235760640
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.514203177660086e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.4274639515010996,
      "tokens_seen": 2235891712
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.513400738244263e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.4274477005274377,
      "tokens_seen": 2236022784
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.512598298828438e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.427431450773065,
      "tokens_seen": 2236153856
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.511795859412615e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.427415202237819,
      "tokens_seen": 2236284928
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.51099341999679e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.4273989549215362,
      "tokens_seen": 2236416000
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.510190980580967e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.4273827088240543,
      "tokens_seen": 2236547072
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.509388541165142e-05,
      "loss": 1.3786,
      "theoretical_loss": 3.4273664639452095,
      "tokens_seen": 2236678144
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.508586101749317e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.42735022028484,
      "tokens_seen": 2236809216
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.507783662333494e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.4273339778427827,
      "tokens_seen": 2236940288
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.50698122291767e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.4273177366188747,
      "tokens_seen": 2237071360
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.506178783501846e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.4273014966129534,
      "tokens_seen": 2237202432
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.505376344086021e-05,
      "loss": 1.3634,
      "theoretical_loss": 3.4272852578248556,
      "tokens_seen": 2237333504
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.504573904670198e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.4272690202544194,
      "tokens_seen": 2237464576
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.503771465254373e-05,
      "loss": 1.252,
      "theoretical_loss": 3.427252783901482,
      "tokens_seen": 2237595648
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.50296902583855e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.4272365487658805,
      "tokens_seen": 2237726720
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.502166586422725e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.4272203148474527,
      "tokens_seen": 2237857792
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.501364147006901e-05,
      "loss": 1.3713,
      "theoretical_loss": 3.427204082146036,
      "tokens_seen": 2237988864
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.500561707591077e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.4271878506614675,
      "tokens_seen": 2238119936
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.499759268175253e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.427171620393585,
      "tokens_seen": 2238251008
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.49895682875943e-05,
      "loss": 1.3568,
      "theoretical_loss": 3.427155391342226,
      "tokens_seen": 2238382080
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.498154389343605e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.427139163507228,
      "tokens_seen": 2238513152
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.497351949927781e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.427122936888429,
      "tokens_seen": 2238644224
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.496549510511957e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.427106711485666,
      "tokens_seen": 2238775296
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.495747071096132e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.427090487298778,
      "tokens_seen": 2238906368
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.494944631680309e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.427074264327601,
      "tokens_seen": 2239037440
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.494142192264484e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.4270580425719737,
      "tokens_seen": 2239168512
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.49333975284866e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.427041822031734,
      "tokens_seen": 2239299584
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.492537313432836e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.427025602706719,
      "tokens_seen": 2239430656
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.491734874017013e-05,
      "loss": 1.259,
      "theoretical_loss": 3.4270093845967673,
      "tokens_seen": 2239561728
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.490932434601188e-05,
      "loss": 1.2495,
      "theoretical_loss": 3.4269931677017165,
      "tokens_seen": 2239692800
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.9927558302879333,
      "objective/train/avg_token_score": 0.007244119420647621,
      "objective/train/docs_used": 815683,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3097038269042969,
      "objective/train/original_loss": 1.3185343742370605,
      "objective/train/theoretical_loss": 3.426976952021404,
      "objective/train/tokens_used": 610282976,
      "objective/train/unlikelihood_loss": 0.022534752264618874,
      "theoretical_loss": 3.426976952021404,
      "tokens_seen": 2239823872
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.490129995185363e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.426976952021404,
      "tokens_seen": 2239823872
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.48932755576954e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.4269607375556683,
      "tokens_seen": 2239954944
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.488525116353715e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.4269445243043477,
      "tokens_seen": 2240086016
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.487722676937892e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.4269283122672793,
      "tokens_seen": 2240217088
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.486920237522067e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.4269121014443016,
      "tokens_seen": 2240348160
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.486117798106244e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.426895891835253,
      "tokens_seen": 2240479232
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.485315358690419e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.426879683439971,
      "tokens_seen": 2240610304
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.484512919274594e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.4268634762582937,
      "tokens_seen": 2240741376
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.483710479858771e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.426847270290059,
      "tokens_seen": 2240872448
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.482908040442946e-05,
      "loss": 1.3097,
      "theoretical_loss": 3.4268310655351066,
      "tokens_seen": 2241003520
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.482105601027123e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.426814861993273,
      "tokens_seen": 2241134592
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.481303161611298e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.4267986596643976,
      "tokens_seen": 2241265664
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.480500722195475e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.4267824585483178,
      "tokens_seen": 2241396736
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.47969828277965e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.4267662586448724,
      "tokens_seen": 2241527808
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.478895843363826e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.4267500599538994,
      "tokens_seen": 2241658880
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.478093403948002e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.4267338624752375,
      "tokens_seen": 2241789952
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.477290964532178e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.426717666208725,
      "tokens_seen": 2241921024
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.476488525116354e-05,
      "loss": 1.3575,
      "theoretical_loss": 3.4267014711542005,
      "tokens_seen": 2242052096
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.47568608570053e-05,
      "loss": 1.2265,
      "theoretical_loss": 3.4266852773115017,
      "tokens_seen": 2242183168
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.474883646284706e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.4266690846804684,
      "tokens_seen": 2242314240
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.474081206868882e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.426652893260938,
      "tokens_seen": 2242445312
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.473278767453057e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4266367030527496,
      "tokens_seen": 2242576384
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.472476328037234e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.4266205140557413,
      "tokens_seen": 2242707456
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.471673888621409e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.4266043262697523,
      "tokens_seen": 2242838528
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.470871449205586e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.426588139694621,
      "tokens_seen": 2242969600
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.470069009789761e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.4265719543301856,
      "tokens_seen": 2243100672
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.469266570373938e-05,
      "loss": 1.1941,
      "theoretical_loss": 3.426555770176286,
      "tokens_seen": 2243231744
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.468464130958113e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.4265395872327593,
      "tokens_seen": 2243362816
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.46766169154229e-05,
      "loss": 1.3201,
      "theoretical_loss": 3.426523405499446,
      "tokens_seen": 2243493888
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.466859252126465e-05,
      "loss": 1.3561,
      "theoretical_loss": 3.4265072249761834,
      "tokens_seen": 2243624960
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.46605681271064e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.4264910456628117,
      "tokens_seen": 2243756032
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.465254373294817e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.4264748675591687,
      "tokens_seen": 2243887104
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.464451933878992e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.4264586906650933,
      "tokens_seen": 2244018176
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.463649494463169e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.426442514980425,
      "tokens_seen": 2244149248
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.462847055047344e-05,
      "loss": 1.3818,
      "theoretical_loss": 3.426426340505003,
      "tokens_seen": 2244280320
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.462044615631521e-05,
      "loss": 1.3319,
      "theoretical_loss": 3.4264101672386653,
      "tokens_seen": 2244411392
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.461242176215696e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.4263939951812517,
      "tokens_seen": 2244542464
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.460439736799871e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.426377824332601,
      "tokens_seen": 2244673536
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.459637297384048e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.426361654692552,
      "tokens_seen": 2244804608
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.458834857968223e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.426345486260944,
      "tokens_seen": 2244935680
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.4580324185524e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.4263293190376167,
      "tokens_seen": 2245066752
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.457229979136575e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.4263131530224085,
      "tokens_seen": 2245197824
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.456427539720752e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.426296988215159,
      "tokens_seen": 2245328896
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.455625100304927e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.4262808246157066,
      "tokens_seen": 2245459968
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.454822660889103e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.426264662223892,
      "tokens_seen": 2245591040
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.45402022147328e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.4262485010395536,
      "tokens_seen": 2245722112
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.453217782057455e-05,
      "loss": 1.3574,
      "theoretical_loss": 3.426232341062531,
      "tokens_seen": 2245853184
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.452415342641631e-05,
      "loss": 1.3358,
      "theoretical_loss": 3.4262161822926633,
      "tokens_seen": 2245984256
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.451612903225807e-05,
      "loss": 1.2255,
      "theoretical_loss": 3.42620002472979,
      "tokens_seen": 2246115328
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.450810463809983e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.4261838683737507,
      "tokens_seen": 2246246400
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.9919527173042297,
      "objective/train/avg_token_score": 0.008047258481383324,
      "objective/train/docs_used": 817879,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2910064458847046,
      "objective/train/original_loss": 1.3017232418060303,
      "objective/train/theoretical_loss": 3.4261677132243844,
      "objective/train/tokens_used": 616836576,
      "objective/train/unlikelihood_loss": 0.022911157459020615,
      "theoretical_loss": 3.4261677132243844,
      "tokens_seen": 2246377472
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.450008024394159e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.4261677132243844,
      "tokens_seen": 2246377472
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.449205584978334e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.426151559281531,
      "tokens_seen": 2246508544
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.448403145562511e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.42613540654503,
      "tokens_seen": 2246639616
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.447600706146686e-05,
      "loss": 1.3889,
      "theoretical_loss": 3.4261192550147204,
      "tokens_seen": 2246770688
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.446798266730863e-05,
      "loss": 1.3576,
      "theoretical_loss": 3.4261031046904424,
      "tokens_seen": 2246901760
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.445995827315038e-05,
      "loss": 1.276,
      "theoretical_loss": 3.4260869555720355,
      "tokens_seen": 2247032832
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.445193387899215e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.4260708076593396,
      "tokens_seen": 2247163904
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.44439094848339e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.426054660952193,
      "tokens_seen": 2247294976
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.443588509067565e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.426038515450437,
      "tokens_seen": 2247426048
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.442786069651742e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.426022371153911,
      "tokens_seen": 2247557120
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.441983630235917e-05,
      "loss": 1.2085,
      "theoretical_loss": 3.426006228062454,
      "tokens_seen": 2247688192
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.441181190820094e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.425990086175906,
      "tokens_seen": 2247819264
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.440378751404269e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.425973945494108,
      "tokens_seen": 2247950336
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.439576311988446e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.425957806016898,
      "tokens_seen": 2248081408
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.438773872572621e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.425941667744117,
      "tokens_seen": 2248212480
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.437971433156798e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.4259255306756042,
      "tokens_seen": 2248343552
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.437168993740973e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.4259093948112005,
      "tokens_seen": 2248474624
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.436366554325149e-05,
      "loss": 1.3264,
      "theoretical_loss": 3.4258932601507457,
      "tokens_seen": 2248605696
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.435564114909325e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.425877126694079,
      "tokens_seen": 2248736768
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.4347616754935e-05,
      "loss": 1.241,
      "theoretical_loss": 3.425860994441041,
      "tokens_seen": 2248867840
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.433959236077677e-05,
      "loss": 1.343,
      "theoretical_loss": 3.4258448633914718,
      "tokens_seen": 2248998912
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.433156796661852e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.4258287335452113,
      "tokens_seen": 2249129984
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.432354357246029e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.4258126049020996,
      "tokens_seen": 2249261056
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.431551917830204e-05,
      "loss": 1.2206,
      "theoretical_loss": 3.425796477461977,
      "tokens_seen": 2249392128
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.43074947841438e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.4257803512246836,
      "tokens_seen": 2249523200
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.429947038998556e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.4257642261900596,
      "tokens_seen": 2249654272
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.429144599582732e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.4257481023579457,
      "tokens_seen": 2249785344
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.428342160166908e-05,
      "loss": 1.3127,
      "theoretical_loss": 3.425731979728181,
      "tokens_seen": 2249916416
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.427539720751084e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.425715858300607,
      "tokens_seen": 2250047488
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.42673728133526e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.425699738075063,
      "tokens_seen": 2250178560
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.425934841919436e-05,
      "loss": 1.3547,
      "theoretical_loss": 3.425683619051391,
      "tokens_seen": 2250309632
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.425132402503611e-05,
      "loss": 1.2208,
      "theoretical_loss": 3.4256675012294293,
      "tokens_seen": 2250440704
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.424329963087788e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.42565138460902,
      "tokens_seen": 2250571776
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.423527523671963e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.425635269190003,
      "tokens_seen": 2250702848
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.42272508425614e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.4256191549722184,
      "tokens_seen": 2250833920
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.421922644840315e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.4256030419555072,
      "tokens_seen": 2250964992
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.421120205424492e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.4255869301397097,
      "tokens_seen": 2251096064
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.420317766008667e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.4255708195246664,
      "tokens_seen": 2251227136
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.419515326592842e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.425554710110218,
      "tokens_seen": 2251358208
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.418712887177019e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.4255386018962057,
      "tokens_seen": 2251489280
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.417910447761194e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.425522494882469,
      "tokens_seen": 2251620352
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.417108008345371e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.42550638906885,
      "tokens_seen": 2251751424
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.416305568929546e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.425490284455188,
      "tokens_seen": 2251882496
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.415503129513723e-05,
      "loss": 1.287,
      "theoretical_loss": 3.4254741810413254,
      "tokens_seen": 2252013568
    },
    {
      "epoch": 0.36,
      "learning_rate": 6.414700690097898e-05,
      "loss": 1.269,
      "theoretical_loss": 3.4254580788271016,
      "tokens_seen": 2252144640
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.413898250682074e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.4254419778123575,
      "tokens_seen": 2252275712
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.41309581126625e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.425425877996935,
      "tokens_seen": 2252406784
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.412293371850426e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.425409779380674,
      "tokens_seen": 2252537856
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.411490932434602e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.4253936819634156,
      "tokens_seen": 2252668928
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.410688493018777e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.425377585745001,
      "tokens_seen": 2252800000
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.9918113946914673,
      "objective/train/avg_token_score": 0.008188593201339245,
      "objective/train/docs_used": 820362,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3687012195587158,
      "objective/train/original_loss": 1.3802019357681274,
      "objective/train/theoretical_loss": 3.4253614907252707,
      "objective/train/tokens_used": 623390176,
      "objective/train/unlikelihood_loss": 0.023289231583476067,
      "theoretical_loss": 3.4253614907252707,
      "tokens_seen": 2252931072
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.409886053602954e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.4253614907252707,
      "tokens_seen": 2252931072
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.40908361418713e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.4253453969040666,
      "tokens_seen": 2253062144
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.408281174771305e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.425329304281229,
      "tokens_seen": 2253193216
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.407478735355481e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.425313212856599,
      "tokens_seen": 2253324288
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.406676295939657e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.425297122630018,
      "tokens_seen": 2253455360
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.405873856523833e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.425281033601327,
      "tokens_seen": 2253586432
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.405071417108009e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.425264945770367,
      "tokens_seen": 2253717504
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.404268977692185e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.4252488591369796,
      "tokens_seen": 2253848576
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.403466538276361e-05,
      "loss": 1.365,
      "theoretical_loss": 3.4252327737010053,
      "tokens_seen": 2253979648
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.402664098860537e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.425216689462286,
      "tokens_seen": 2254110720
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.401861659444713e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.425200606420663,
      "tokens_seen": 2254241792
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.401059220028888e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.425184524575977,
      "tokens_seen": 2254372864
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.400256780613065e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.4251684439280696,
      "tokens_seen": 2254503936
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.39945434119724e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.4251523644767827,
      "tokens_seen": 2254635008
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.398651901781417e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.425136286221957,
      "tokens_seen": 2254766080
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.397849462365592e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.4251202091634343,
      "tokens_seen": 2254897152
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.397047022949769e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.425104133301056,
      "tokens_seen": 2255028224
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.396244583533944e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.4250880586346635,
      "tokens_seen": 2255159296
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.395442144118119e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.4250719851640983,
      "tokens_seen": 2255290368
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.394639704702296e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.425055912889202,
      "tokens_seen": 2255421440
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.393837265286471e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.4250398418098165,
      "tokens_seen": 2255552512
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.393034825870648e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.4250237719257823,
      "tokens_seen": 2255683584
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.392232386454823e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.4250077032369424,
      "tokens_seen": 2255814656
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.391429947039e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.424991635743138,
      "tokens_seen": 2255945728
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.390627507623175e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.4249755694442108,
      "tokens_seen": 2256076800
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.38982506820735e-05,
      "loss": 1.25,
      "theoretical_loss": 3.424959504340002,
      "tokens_seen": 2256207872
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.389022628791527e-05,
      "loss": 1.262,
      "theoretical_loss": 3.4249434404303543,
      "tokens_seen": 2256338944
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.388220189375703e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.4249273777151084,
      "tokens_seen": 2256470016
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.387417749959879e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.424911316194107,
      "tokens_seen": 2256601088
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.386615310544055e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.4248952558671917,
      "tokens_seen": 2256732160
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.385812871128231e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.4248791967342043,
      "tokens_seen": 2256863232
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.385010431712406e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.424863138794987,
      "tokens_seen": 2256994304
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.384207992296582e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.4248470820493813,
      "tokens_seen": 2257125376
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.383405552880758e-05,
      "loss": 1.3187,
      "theoretical_loss": 3.424831026497229,
      "tokens_seen": 2257256448
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.382603113464934e-05,
      "loss": 1.299,
      "theoretical_loss": 3.424814972138373,
      "tokens_seen": 2257387520
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.38180067404911e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.4247989189726544,
      "tokens_seen": 2257518592
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.380998234633286e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.424782866999916,
      "tokens_seen": 2257649664
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.380195795217462e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.424766816219999,
      "tokens_seen": 2257780736
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.379393355801638e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.424750766632747,
      "tokens_seen": 2257911808
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.378590916385813e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.4247347182380006,
      "tokens_seen": 2258042880
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.37778847696999e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.4247186710356026,
      "tokens_seen": 2258173952
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.376986037554165e-05,
      "loss": 1.3482,
      "theoretical_loss": 3.4247026250253954,
      "tokens_seen": 2258305024
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.376183598138342e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.424686580207221,
      "tokens_seen": 2258436096
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.375381158722517e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.424670536580922,
      "tokens_seen": 2258567168
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.374578719306694e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.42465449414634,
      "tokens_seen": 2258698240
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.373776279890869e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.424638452903318,
      "tokens_seen": 2258829312
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.372973840475044e-05,
      "loss": 1.2222,
      "theoretical_loss": 3.424622412851698,
      "tokens_seen": 2258960384
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.372171401059221e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.424606373991322,
      "tokens_seen": 2259091456
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.371368961643396e-05,
      "loss": 1.2168,
      "theoretical_loss": 3.424590336322034,
      "tokens_seen": 2259222528
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.370566522227573e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.424574299843675,
      "tokens_seen": 2259353600
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.9913434982299805,
      "objective/train/avg_token_score": 0.008656485006213188,
      "objective/train/docs_used": 822867,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3387336730957031,
      "objective/train/original_loss": 1.3522007465362549,
      "objective/train/theoretical_loss": 3.4245582645560875,
      "objective/train/tokens_used": 629943776,
      "objective/train/unlikelihood_loss": 0.022395966574549675,
      "theoretical_loss": 3.4245582645560875,
      "tokens_seen": 2259484672
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.369764082811748e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.4245582645560875,
      "tokens_seen": 2259484672
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.368961643395925e-05,
      "loss": 1.2429,
      "theoretical_loss": 3.4245422304591147,
      "tokens_seen": 2259615744
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.3681592039801e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.424526197552599,
      "tokens_seen": 2259746816
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.367356764564276e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.4245101658363826,
      "tokens_seen": 2259877888
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.366554325148452e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.424494135310308,
      "tokens_seen": 2260008960
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.365751885732628e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.4244781059742184,
      "tokens_seen": 2260140032
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.364949446316804e-05,
      "loss": 1.235,
      "theoretical_loss": 3.4244620778279566,
      "tokens_seen": 2260271104
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.36414700690098e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.4244460508713646,
      "tokens_seen": 2260402176
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.363344567485155e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.424430025104285,
      "tokens_seen": 2260533248
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.362542128069332e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.4244140005265615,
      "tokens_seen": 2260664320
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.361739688653507e-05,
      "loss": 1.3773,
      "theoretical_loss": 3.424397977138036,
      "tokens_seen": 2260795392
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.360937249237683e-05,
      "loss": 1.3384,
      "theoretical_loss": 3.4243819549385517,
      "tokens_seen": 2260926464
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.360134809821859e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.4243659339279517,
      "tokens_seen": 2261057536
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.359332370406035e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.4243499141060783,
      "tokens_seen": 2261188608
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.358529930990211e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.424333895472775,
      "tokens_seen": 2261319680
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.357727491574386e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.4243178780278836,
      "tokens_seen": 2261450752
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.356925052158563e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.424301861771249,
      "tokens_seen": 2261581824
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.356122612742738e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.424285846702712,
      "tokens_seen": 2261712896
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.355320173326915e-05,
      "loss": 1.3671,
      "theoretical_loss": 3.4242698328221173,
      "tokens_seen": 2261843968
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.35451773391109e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.424253820129307,
      "tokens_seen": 2261975040
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.353715294495265e-05,
      "loss": 1.1953,
      "theoretical_loss": 3.4242378086241247,
      "tokens_seen": 2262106112
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.352912855079442e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.424221798306413,
      "tokens_seen": 2262237184
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.352110415663617e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.424205789176016,
      "tokens_seen": 2262368256
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.351307976247794e-05,
      "loss": 1.3855,
      "theoretical_loss": 3.4241897812327755,
      "tokens_seen": 2262499328
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.35050553683197e-05,
      "loss": 1.298,
      "theoretical_loss": 3.424173774476536,
      "tokens_seen": 2262630400
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.349703097416146e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.4241577689071394,
      "tokens_seen": 2262761472
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.348900658000321e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.42414176452443,
      "tokens_seen": 2262892544
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.348098218584497e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.424125761328251,
      "tokens_seen": 2263023616
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.347295779168673e-05,
      "loss": 1.2179,
      "theoretical_loss": 3.4241097593184455,
      "tokens_seen": 2263154688
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.346493339752849e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.424093758494857,
      "tokens_seen": 2263285760
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.345690900337025e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.4240777588573286,
      "tokens_seen": 2263416832
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.3448884609212e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.4240617604057038,
      "tokens_seen": 2263547904
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.344086021505376e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.424045763139826,
      "tokens_seen": 2263678976
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.343283582089553e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4240297670595385,
      "tokens_seen": 2263810048
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.342481142673728e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.4240137721646855,
      "tokens_seen": 2263941120
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.341678703257905e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.42399777845511,
      "tokens_seen": 2264072192
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.34087626384208e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4239817859306556,
      "tokens_seen": 2264203264
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.340073824426257e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.423965794591166,
      "tokens_seen": 2264334336
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.339271385010432e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.4239498044364844,
      "tokens_seen": 2264465408
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.338468945594607e-05,
      "loss": 1.2539,
      "theoretical_loss": 3.423933815466455,
      "tokens_seen": 2264596480
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.337666506178784e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.4239178276809215,
      "tokens_seen": 2264727552
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.336864066762959e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.4239018410797266,
      "tokens_seen": 2264858624
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.336061627347136e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.423885855662715,
      "tokens_seen": 2264989696
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.335259187931311e-05,
      "loss": 1.3498,
      "theoretical_loss": 3.423869871429731,
      "tokens_seen": 2265120768
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.334456748515486e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.4238538883806173,
      "tokens_seen": 2265251840
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.333654309099663e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.423837906515218,
      "tokens_seen": 2265382912
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.332851869683838e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.423821925833377,
      "tokens_seen": 2265513984
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.332049430268015e-05,
      "loss": 1.285,
      "theoretical_loss": 3.4238059463349377,
      "tokens_seen": 2265645056
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.33124699085219e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.423789968019745,
      "tokens_seen": 2265776128
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.330444551436366e-05,
      "loss": 1.3508,
      "theoretical_loss": 3.4237739908876423,
      "tokens_seen": 2265907200
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.9939123392105103,
      "objective/train/avg_token_score": 0.006087609101086855,
      "objective/train/docs_used": 825438,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2394496202468872,
      "objective/train/original_loss": 1.2478358745574951,
      "objective/train/theoretical_loss": 3.423758014938474,
      "objective/train/tokens_used": 636497376,
      "objective/train/unlikelihood_loss": 0.01812308095395565,
      "theoretical_loss": 3.423758014938474,
      "tokens_seen": 2266038272
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.329642112020542e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.423758014938474,
      "tokens_seen": 2266038272
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.328839672604718e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.4237420401720833,
      "tokens_seen": 2266169344
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.328037233188894e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.4237260665883147,
      "tokens_seen": 2266300416
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.32723479377307e-05,
      "loss": 1.4009,
      "theoretical_loss": 3.423710094187012,
      "tokens_seen": 2266431488
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.326432354357246e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.42369412296802,
      "tokens_seen": 2266562560
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.325629914941422e-05,
      "loss": 1.3466,
      "theoretical_loss": 3.423678152931182,
      "tokens_seen": 2266693632
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.324827475525597e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.423662184076343,
      "tokens_seen": 2266824704
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.324025036109774e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.4236462164033457,
      "tokens_seen": 2266955776
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.323222596693949e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.423630249912036,
      "tokens_seen": 2267086848
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.322420157278126e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.423614284602257,
      "tokens_seen": 2267217920
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.321617717862301e-05,
      "loss": 1.299,
      "theoretical_loss": 3.423598320473854,
      "tokens_seen": 2267348992
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.320815278446478e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.42358235752667,
      "tokens_seen": 2267480064
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.320012839030653e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.42356639576055,
      "tokens_seen": 2267611136
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.319210399614828e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.423550435175339,
      "tokens_seen": 2267742208
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.318407960199005e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.42353447577088,
      "tokens_seen": 2267873280
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.31760552078318e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.4235185175470186,
      "tokens_seen": 2268004352
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.316803081367357e-05,
      "loss": 1.3734,
      "theoretical_loss": 3.4235025605035987,
      "tokens_seen": 2268135424
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.316000641951532e-05,
      "loss": 1.318,
      "theoretical_loss": 3.423486604640465,
      "tokens_seen": 2268266496
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.315198202535709e-05,
      "loss": 1.267,
      "theoretical_loss": 3.423470649957461,
      "tokens_seen": 2268397568
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.314395763119884e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.4234546964544332,
      "tokens_seen": 2268528640
    },
    {
      "epoch": 0.37,
      "learning_rate": 6.31359332370406e-05,
      "loss": 1.2163,
      "theoretical_loss": 3.4234387441312246,
      "tokens_seen": 2268659712
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.312790884288236e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.4234227929876804,
      "tokens_seen": 2268790784
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.311988444872411e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.4234068430236446,
      "tokens_seen": 2268921856
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.311186005456588e-05,
      "loss": 1.288,
      "theoretical_loss": 3.423390894238963,
      "tokens_seen": 2269052928
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.310383566040763e-05,
      "loss": 1.2028,
      "theoretical_loss": 3.423374946633479,
      "tokens_seen": 2269184000
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.30958112662494e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.4233590002070384,
      "tokens_seen": 2269315072
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.308778687209115e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.4233430549594854,
      "tokens_seen": 2269446144
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.307976247793292e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.4233271108906647,
      "tokens_seen": 2269577216
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.307173808377467e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.423311168000421,
      "tokens_seen": 2269708288
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.306371368961643e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.4232952262886,
      "tokens_seen": 2269839360
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.30556892954582e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.423279285755045,
      "tokens_seen": 2269970432
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.304766490129995e-05,
      "loss": 1.214,
      "theoretical_loss": 3.4232633463996023,
      "tokens_seen": 2270101504
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.303964050714171e-05,
      "loss": 1.2209,
      "theoretical_loss": 3.4232474082221165,
      "tokens_seen": 2270232576
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.303161611298347e-05,
      "loss": 1.2327,
      "theoretical_loss": 3.4232314712224317,
      "tokens_seen": 2270363648
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.302359171882523e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.423215535400394,
      "tokens_seen": 2270494720
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.301556732466699e-05,
      "loss": 1.301,
      "theoretical_loss": 3.423199600755848,
      "tokens_seen": 2270625792
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.300754293050874e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.4231836672886384,
      "tokens_seen": 2270756864
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.29995185363505e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.4231677349986107,
      "tokens_seen": 2270887936
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.299149414219226e-05,
      "loss": 1.1577,
      "theoretical_loss": 3.4231518038856095,
      "tokens_seen": 2271019008
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.298346974803403e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.42313587394948,
      "tokens_seen": 2271150080
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.297544535387578e-05,
      "loss": 1.363,
      "theoretical_loss": 3.4231199451900682,
      "tokens_seen": 2271281152
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.296742095971755e-05,
      "loss": 1.2288,
      "theoretical_loss": 3.4231040176072187,
      "tokens_seen": 2271412224
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.29593965655593e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.4230880912007766,
      "tokens_seen": 2271543296
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.295137217140105e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.423072165970587,
      "tokens_seen": 2271674368
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.294334777724282e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.423056241916495,
      "tokens_seen": 2271805440
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.293532338308457e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.423040319038347,
      "tokens_seen": 2271936512
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.292729898892634e-05,
      "loss": 1.2402,
      "theoretical_loss": 3.423024397335987,
      "tokens_seen": 2272067584
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.291927459476809e-05,
      "loss": 1.311,
      "theoretical_loss": 3.4230084768092612,
      "tokens_seen": 2272198656
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.291125020060986e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.4229925574580147,
      "tokens_seen": 2272329728
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.290322580645161e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.4229766392820933,
      "tokens_seen": 2272460800
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.9928696155548096,
      "objective/train/avg_token_score": 0.007130359299480915,
      "objective/train/docs_used": 827716,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2788420915603638,
      "objective/train/original_loss": 1.2889883518218994,
      "objective/train/theoretical_loss": 3.4229607222813416,
      "objective/train/tokens_used": 643050976,
      "objective/train/unlikelihood_loss": 0.01785312592983246,
      "theoretical_loss": 3.4229607222813416,
      "tokens_seen": 2272591872
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.289520141229336e-05,
      "loss": 1.235,
      "theoretical_loss": 3.4229607222813416,
      "tokens_seen": 2272591872
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.288717701813513e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.4229448064556056,
      "tokens_seen": 2272722944
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.287915262397688e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.422928891804731,
      "tokens_seen": 2272854016
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.287112822981865e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.422912978328563,
      "tokens_seen": 2272985088
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.28631038356604e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.4228970660269473,
      "tokens_seen": 2273116160
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.285507944150217e-05,
      "loss": 1.267,
      "theoretical_loss": 3.422881154899729,
      "tokens_seen": 2273247232
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.284705504734392e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.4228652449467547,
      "tokens_seen": 2273378304
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.283903065318568e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.4228493361678693,
      "tokens_seen": 2273509376
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.283100625902744e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.4228334285629187,
      "tokens_seen": 2273640448
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.28229818648692e-05,
      "loss": 1.3824,
      "theoretical_loss": 3.4228175221317483,
      "tokens_seen": 2273771520
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.281495747071096e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.4228016168742044,
      "tokens_seen": 2273902592
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.280693307655272e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.4227857127901324,
      "tokens_seen": 2274033664
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.279890868239448e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.4227698098793784,
      "tokens_seen": 2274164736
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.279088428823624e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.4227539081417877,
      "tokens_seen": 2274295808
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.2782859894078e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.4227380075772067,
      "tokens_seen": 2274426880
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.277483549991976e-05,
      "loss": 1.3727,
      "theoretical_loss": 3.422722108185481,
      "tokens_seen": 2274557952
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.276681110576151e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.4227062099664565,
      "tokens_seen": 2274689024
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.275878671160328e-05,
      "loss": 1.2138,
      "theoretical_loss": 3.4226903129199786,
      "tokens_seen": 2274820096
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.275076231744503e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.4226744170458945,
      "tokens_seen": 2274951168
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.27427379232868e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.4226585223440495,
      "tokens_seen": 2275082240
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.273471352912855e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.4226426288142893,
      "tokens_seen": 2275213312
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.272668913497032e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.422626736456461,
      "tokens_seen": 2275344384
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.271866474081207e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4226108452704094,
      "tokens_seen": 2275475456
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.271064034665382e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.422594955255981,
      "tokens_seen": 2275606528
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.270261595249559e-05,
      "loss": 1.338,
      "theoretical_loss": 3.422579066413023,
      "tokens_seen": 2275737600
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.269459155833734e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.42256317874138,
      "tokens_seen": 2275868672
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.268656716417911e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.4225472922408993,
      "tokens_seen": 2275999744
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.267854277002086e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.4225314069114265,
      "tokens_seen": 2276130816
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.267051837586263e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.4225155227528083,
      "tokens_seen": 2276261888
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.266249398170438e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.422499639764891,
      "tokens_seen": 2276392960
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.265446958754613e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.4224837579475205,
      "tokens_seen": 2276524032
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.26464451933879e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.422467877300543,
      "tokens_seen": 2276655104
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.263842079922965e-05,
      "loss": 1.299,
      "theoretical_loss": 3.4224519978238055,
      "tokens_seen": 2276786176
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.263039640507142e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.4224361195171547,
      "tokens_seen": 2276917248
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.262237201091317e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.422420242380436,
      "tokens_seen": 2277048320
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.261434761675494e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.422404366413496,
      "tokens_seen": 2277179392
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.26063232225967e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.422388491616182,
      "tokens_seen": 2277310464
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.259829882843845e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.42237261798834,
      "tokens_seen": 2277441536
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.259027443428021e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4223567455298167,
      "tokens_seen": 2277572608
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.258225004012197e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.422340874240458,
      "tokens_seen": 2277703680
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.257422564596373e-05,
      "loss": 1.3577,
      "theoretical_loss": 3.4223250041201116,
      "tokens_seen": 2277834752
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.256620125180549e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.4223091351686237,
      "tokens_seen": 2277965824
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.255817685764725e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.4222932673858404,
      "tokens_seen": 2278096896
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.255015246348901e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.4222774007716095,
      "tokens_seen": 2278227968
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.254212806933076e-05,
      "loss": 1.2173,
      "theoretical_loss": 3.4222615353257764,
      "tokens_seen": 2278359040
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.253410367517253e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.422245671048189,
      "tokens_seen": 2278490112
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.252607928101428e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.422229807938694,
      "tokens_seen": 2278621184
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.251805488685605e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.422213945997137,
      "tokens_seen": 2278752256
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.25100304926978e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.422198085223366,
      "tokens_seen": 2278883328
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.250200609853957e-05,
      "loss": 1.2247,
      "theoretical_loss": 3.422182225617228,
      "tokens_seen": 2279014400
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.9944167137145996,
      "objective/train/avg_token_score": 0.005583287216722965,
      "objective/train/docs_used": 830070,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.309889316558838,
      "objective/train/original_loss": 1.3184558153152466,
      "objective/train/theoretical_loss": 3.4221663671785687,
      "objective/train/tokens_used": 649604576,
      "objective/train/unlikelihood_loss": 0.014190822839736938,
      "theoretical_loss": 3.4221663671785687,
      "tokens_seen": 2279145472
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.249398170438132e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.4221663671785687,
      "tokens_seen": 2279145472
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.248595731022309e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.422150509907236,
      "tokens_seen": 2279276544
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.247793291606484e-05,
      "loss": 1.2044,
      "theoretical_loss": 3.422134653803077,
      "tokens_seen": 2279407616
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.246990852190659e-05,
      "loss": 1.3436,
      "theoretical_loss": 3.422118798865938,
      "tokens_seen": 2279538688
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.246188412774836e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.422102945095666,
      "tokens_seen": 2279669760
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.245385973359011e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.4220870924921085,
      "tokens_seen": 2279800832
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.244583533943188e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.4220712410551126,
      "tokens_seen": 2279931904
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.243781094527363e-05,
      "loss": 1.2288,
      "theoretical_loss": 3.422055390784525,
      "tokens_seen": 2280062976
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.24297865511154e-05,
      "loss": 1.268,
      "theoretical_loss": 3.4220395416801934,
      "tokens_seen": 2280194048
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.242176215695715e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.4220236937419646,
      "tokens_seen": 2280325120
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.24137377627989e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.4220078469696853,
      "tokens_seen": 2280456192
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.240571336864067e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.4219920013632037,
      "tokens_seen": 2280587264
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.239768897448242e-05,
      "loss": 1.257,
      "theoretical_loss": 3.421976156922366,
      "tokens_seen": 2280718336
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.238966458032419e-05,
      "loss": 1.3341,
      "theoretical_loss": 3.4219603136470207,
      "tokens_seen": 2280849408
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.238164018616594e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.4219444715370138,
      "tokens_seen": 2280980480
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.237361579200771e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.4219286305921934,
      "tokens_seen": 2281111552
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.236559139784946e-05,
      "loss": 1.3381,
      "theoretical_loss": 3.421912790812407,
      "tokens_seen": 2281242624
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.235756700369122e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.421896952197501,
      "tokens_seen": 2281373696
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.234954260953298e-05,
      "loss": 1.1979,
      "theoretical_loss": 3.421881114747324,
      "tokens_seen": 2281504768
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.234151821537474e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.421865278461723,
      "tokens_seen": 2281635840
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.23334938212165e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.421849443340545,
      "tokens_seen": 2281766912
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.232546942705826e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.421833609383638,
      "tokens_seen": 2281897984
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.231744503290002e-05,
      "loss": 1.312,
      "theoretical_loss": 3.4218177765908497,
      "tokens_seen": 2282029056
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.230942063874178e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.421801944962027,
      "tokens_seen": 2282160128
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.230139624458353e-05,
      "loss": 1.1934,
      "theoretical_loss": 3.421786114497018,
      "tokens_seen": 2282291200
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.22933718504253e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.4217702851956706,
      "tokens_seen": 2282422272
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.228534745626705e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.4217544570578315,
      "tokens_seen": 2282553344
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.227732306210882e-05,
      "loss": 1.2494,
      "theoretical_loss": 3.4217386300833494,
      "tokens_seen": 2282684416
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.226929866795057e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.421722804272071,
      "tokens_seen": 2282815488
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.226127427379234e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.421706979623845,
      "tokens_seen": 2282946560
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.225324987963409e-05,
      "loss": 1.2593,
      "theoretical_loss": 3.421691156138518,
      "tokens_seen": 2283077632
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.224522548547584e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.4216753338159394,
      "tokens_seen": 2283208704
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.223720109131761e-05,
      "loss": 1.2158,
      "theoretical_loss": 3.4216595126559555,
      "tokens_seen": 2283339776
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.222917669715936e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.421643692658415,
      "tokens_seen": 2283470848
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.222115230300113e-05,
      "loss": 1.3544,
      "theoretical_loss": 3.4216278738231654,
      "tokens_seen": 2283601920
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.221312790884288e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.421612056150055,
      "tokens_seen": 2283732992
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.220510351468465e-05,
      "loss": 1.278,
      "theoretical_loss": 3.421596239638931,
      "tokens_seen": 2283864064
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.21970791205264e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.421580424289642,
      "tokens_seen": 2283995136
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.218905472636816e-05,
      "loss": 1.2538,
      "theoretical_loss": 3.4215646101020356,
      "tokens_seen": 2284126208
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.218103033220992e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.4215487970759604,
      "tokens_seen": 2284257280
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.217300593805168e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.4215329852112637,
      "tokens_seen": 2284388352
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.216498154389344e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.4215171745077946,
      "tokens_seen": 2284519424
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.21569571497352e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.4215013649654,
      "tokens_seen": 2284650496
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.214893275557696e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.4214855565839284,
      "tokens_seen": 2284781568
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.214090836141871e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.4214697493632285,
      "tokens_seen": 2284912640
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.213288396726048e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.4214539433031477,
      "tokens_seen": 2285043712
    },
    {
      "epoch": 0.38,
      "learning_rate": 6.212485957310223e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.421438138403535,
      "tokens_seen": 2285174784
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.211683517894399e-05,
      "loss": 1.317,
      "theoretical_loss": 3.4214223346642383,
      "tokens_seen": 2285305856
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.210881078478575e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.4214065320851055,
      "tokens_seen": 2285436928
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.210078639062751e-05,
      "loss": 1.2165,
      "theoretical_loss": 3.4213907306659856,
      "tokens_seen": 2285568000
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.9909844398498535,
      "objective/train/avg_token_score": 0.0090155228972435,
      "objective/train/docs_used": 832464,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2537413835525513,
      "objective/train/original_loss": 1.265404224395752,
      "objective/train/theoretical_loss": 3.4213749304067265,
      "objective/train/tokens_used": 656158176,
      "objective/train/unlikelihood_loss": 0.025877809152007103,
      "theoretical_loss": 3.4213749304067265,
      "tokens_seen": 2285699072
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.209276199646927e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.4213749304067265,
      "tokens_seen": 2285699072
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.208473760231103e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.4213591313071765,
      "tokens_seen": 2285830144
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.20767132081528e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.4213433333671843,
      "tokens_seen": 2285961216
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.206868881399455e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.421327536586598,
      "tokens_seen": 2286092288
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.20606644198363e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.4213117409652662,
      "tokens_seen": 2286223360
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.205264002567807e-05,
      "loss": 1.3312,
      "theoretical_loss": 3.4212959465030375,
      "tokens_seen": 2286354432
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.204461563151982e-05,
      "loss": 1.272,
      "theoretical_loss": 3.4212801531997603,
      "tokens_seen": 2286485504
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.203659123736159e-05,
      "loss": 1.287,
      "theoretical_loss": 3.421264361055283,
      "tokens_seen": 2286616576
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.202856684320334e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.421248570069454,
      "tokens_seen": 2286747648
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.20205424490451e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.4212327802421227,
      "tokens_seen": 2286878720
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.201251805488686e-05,
      "loss": 1.342,
      "theoretical_loss": 3.4212169915731367,
      "tokens_seen": 2287009792
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.200449366072861e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4212012040623456,
      "tokens_seen": 2287140864
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.199646926657038e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.421185417709597,
      "tokens_seen": 2287271936
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.198844487241213e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.4211696325147405,
      "tokens_seen": 2287403008
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.19804204782539e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.4211538484776245,
      "tokens_seen": 2287534080
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.197239608409565e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.4211380655980976,
      "tokens_seen": 2287665152
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.196437168993742e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.421122283876009,
      "tokens_seen": 2287796224
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.195634729577917e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.421106503311207,
      "tokens_seen": 2287927296
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.194832290162093e-05,
      "loss": 1.3406,
      "theoretical_loss": 3.421090723903541,
      "tokens_seen": 2288058368
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.194029850746269e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.4210749456528595,
      "tokens_seen": 2288189440
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.193227411330445e-05,
      "loss": 1.2186,
      "theoretical_loss": 3.4210591685590113,
      "tokens_seen": 2288320512
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.192424971914621e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.4210433926218458,
      "tokens_seen": 2288451584
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.191622532498797e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.4210276178412116,
      "tokens_seen": 2288582656
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.190820093082973e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.4210118442169573,
      "tokens_seen": 2288713728
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.190017653667148e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.4209960717489327,
      "tokens_seen": 2288844800
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.189215214251324e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.4209803004369865,
      "tokens_seen": 2288975872
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.1884127748355e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.420964530280968,
      "tokens_seen": 2289106944
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.187610335419676e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.4209487612807257,
      "tokens_seen": 2289238016
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.186807896003852e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.4209329934361095,
      "tokens_seen": 2289369088
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.186005456588028e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.4209172267469676,
      "tokens_seen": 2289500160
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.185203017172204e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.42090146121315,
      "tokens_seen": 2289631232
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.18440057775638e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.420885696834506,
      "tokens_seen": 2289762304
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.183598138340556e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.4208699336108843,
      "tokens_seen": 2289893376
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.182795698924732e-05,
      "loss": 1.3472,
      "theoretical_loss": 3.420854171542134,
      "tokens_seen": 2290024448
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.181993259508907e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.420838410628105,
      "tokens_seen": 2290155520
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.181190820093084e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.4208226508686463,
      "tokens_seen": 2290286592
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.180388380677259e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.4208068922636077,
      "tokens_seen": 2290417664
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.179585941261436e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.4207911348128377,
      "tokens_seen": 2290548736
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.178783501845611e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.4207753785161867,
      "tokens_seen": 2290679808
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.177981062429788e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.4207596233735034,
      "tokens_seen": 2290810880
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.177178623013963e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.420743869384638,
      "tokens_seen": 2290941952
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.176376183598138e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.420728116549439,
      "tokens_seen": 2291073024
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.175573744182315e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.4207123648677564,
      "tokens_seen": 2291204096
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.17477130476649e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.42069661433944,
      "tokens_seen": 2291335168
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.173968865350667e-05,
      "loss": 1.309,
      "theoretical_loss": 3.420680864964339,
      "tokens_seen": 2291466240
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.173166425934842e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.4206651167423034,
      "tokens_seen": 2291597312
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.172363986519019e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.420649369673183,
      "tokens_seen": 2291728384
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.171561547103194e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.4206336237568267,
      "tokens_seen": 2291859456
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.17075910768737e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.4206178789930846,
      "tokens_seen": 2291990528
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.169956668271546e-05,
      "loss": 1.266,
      "theoretical_loss": 3.420602135381806,
      "tokens_seen": 2292121600
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.9938550591468811,
      "objective/train/avg_token_score": 0.0061449627391994,
      "objective/train/docs_used": 834793,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2349987030029297,
      "objective/train/original_loss": 1.2439334392547607,
      "objective/train/theoretical_loss": 3.4205863929228415,
      "objective/train/tokens_used": 662711776,
      "objective/train/unlikelihood_loss": 0.016227662563323975,
      "theoretical_loss": 3.4205863929228415,
      "tokens_seen": 2292252672
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.169154228855722e-05,
      "loss": 1.2077,
      "theoretical_loss": 3.4205863929228415,
      "tokens_seen": 2292252672
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.168351789439898e-05,
      "loss": 1.2145,
      "theoretical_loss": 3.4205706516160403,
      "tokens_seen": 2292383744
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.167549350024074e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.4205549114612523,
      "tokens_seen": 2292514816
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.16674691060825e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.4205391724583274,
      "tokens_seen": 2292645888
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.165944471192425e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.420523434607116,
      "tokens_seen": 2292776960
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.165142031776601e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.4205076979074667,
      "tokens_seen": 2292908032
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.164339592360777e-05,
      "loss": 1.2214,
      "theoretical_loss": 3.42049196235923,
      "tokens_seen": 2293039104
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.163537152944953e-05,
      "loss": 1.1945,
      "theoretical_loss": 3.420476227962256,
      "tokens_seen": 2293170176
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.16273471352913e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.4204604947163952,
      "tokens_seen": 2293301248
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.161932274113305e-05,
      "loss": 1.252,
      "theoretical_loss": 3.420444762621497,
      "tokens_seen": 2293432320
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.161129834697481e-05,
      "loss": 1.3523,
      "theoretical_loss": 3.4204290316774113,
      "tokens_seen": 2293563392
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.160327395281657e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.420413301883988,
      "tokens_seen": 2293694464
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.159524955865832e-05,
      "loss": 1.1922,
      "theoretical_loss": 3.420397573241078,
      "tokens_seen": 2293825536
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.158722516450009e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.4203818457485307,
      "tokens_seen": 2293956608
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.157920077034184e-05,
      "loss": 1.3272,
      "theoretical_loss": 3.4203661194061965,
      "tokens_seen": 2294087680
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.157117637618361e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.420350394213926,
      "tokens_seen": 2294218752
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.156315198202536e-05,
      "loss": 1.261,
      "theoretical_loss": 3.4203346701715684,
      "tokens_seen": 2294349824
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.155512758786713e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.420318947278975,
      "tokens_seen": 2294480896
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.154710319370888e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.420303225535995,
      "tokens_seen": 2294611968
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.153907879955065e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.4202875049424795,
      "tokens_seen": 2294743040
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.15310544053924e-05,
      "loss": 1.3402,
      "theoretical_loss": 3.4202717854982785,
      "tokens_seen": 2294874112
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.152303001123415e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.4202560672032423,
      "tokens_seen": 2295005184
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.151500561707592e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.420240350057221,
      "tokens_seen": 2295136256
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.150698122291767e-05,
      "loss": 1.2037,
      "theoretical_loss": 3.420224634060066,
      "tokens_seen": 2295267328
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.149895682875944e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.420208919211627,
      "tokens_seen": 2295398400
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.149093243460119e-05,
      "loss": 1.1999,
      "theoretical_loss": 3.420193205511754,
      "tokens_seen": 2295529472
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.148290804044296e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.420177492960298,
      "tokens_seen": 2295660544
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.147488364628471e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.42016178155711,
      "tokens_seen": 2295791616
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.146685925212647e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.4201460713020397,
      "tokens_seen": 2295922688
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.145883485796823e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.4201303621949384,
      "tokens_seen": 2296053760
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.145081046380999e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.4201146542356557,
      "tokens_seen": 2296184832
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.144278606965175e-05,
      "loss": 1.3349,
      "theoretical_loss": 3.420098947424043,
      "tokens_seen": 2296315904
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.14347616754935e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.4200832417599507,
      "tokens_seen": 2296446976
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.142673728133527e-05,
      "loss": 1.4049,
      "theoretical_loss": 3.42006753724323,
      "tokens_seen": 2296578048
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.141871288717702e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.4200518338737305,
      "tokens_seen": 2296709120
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.141068849301878e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.4200361316513037,
      "tokens_seen": 2296840192
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.140266409886054e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.4200204305758004,
      "tokens_seen": 2296971264
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.13946397047023e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.420004730647071,
      "tokens_seen": 2297102336
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.138661531054406e-05,
      "loss": 1.255,
      "theoretical_loss": 3.419989031864967,
      "tokens_seen": 2297233408
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.137859091638582e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.419973334229338,
      "tokens_seen": 2297364480
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.137056652222758e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.419957637740036,
      "tokens_seen": 2297495552
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.136254212806934e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.4199419423969117,
      "tokens_seen": 2297626624
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.135451773391109e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.4199262481998156,
      "tokens_seen": 2297757696
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.134649333975286e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.419910555148599,
      "tokens_seen": 2297888768
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.133846894559461e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.4198948632431128,
      "tokens_seen": 2298019840
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.133044455143638e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.419879172483208,
      "tokens_seen": 2298150912
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.132242015727813e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.4198634828687355,
      "tokens_seen": 2298281984
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.13143957631199e-05,
      "loss": 1.2723,
      "theoretical_loss": 3.4198477943995464,
      "tokens_seen": 2298413056
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.130637136896165e-05,
      "loss": 1.2492,
      "theoretical_loss": 3.419832107075492,
      "tokens_seen": 2298544128
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.12983469748034e-05,
      "loss": 1.3509,
      "theoretical_loss": 3.4198164208964235,
      "tokens_seen": 2298675200
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.9941117167472839,
      "objective/train/avg_token_score": 0.005888289771974087,
      "objective/train/docs_used": 837138,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.176972508430481,
      "objective/train/original_loss": 1.183982253074646,
      "objective/train/theoretical_loss": 3.419800735862192,
      "objective/train/tokens_used": 669265376,
      "objective/train/unlikelihood_loss": 0.019639795646071434,
      "theoretical_loss": 3.419800735862192,
      "tokens_seen": 2298806272
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.129032258064517e-05,
      "loss": 1.1901,
      "theoretical_loss": 3.419800735862192,
      "tokens_seen": 2298806272
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.128229818648692e-05,
      "loss": 1.3835,
      "theoretical_loss": 3.4197850519726485,
      "tokens_seen": 2298937344
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.127427379232869e-05,
      "loss": 1.2055,
      "theoretical_loss": 3.4197693692276436,
      "tokens_seen": 2299068416
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.126624939817044e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.41975368762703,
      "tokens_seen": 2299199488
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.125822500401221e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.4197380071706576,
      "tokens_seen": 2299330560
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.125020060985396e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.419722327858379,
      "tokens_seen": 2299461632
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.124217621569572e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.4197066496900437,
      "tokens_seen": 2299592704
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.123415182153748e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.4196909726655047,
      "tokens_seen": 2299723776
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.122612742737924e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.4196752967846127,
      "tokens_seen": 2299854848
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.1218103033221e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.419659622047219,
      "tokens_seen": 2299985920
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.121007863906276e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.4196439484531753,
      "tokens_seen": 2300116992
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.120205424490452e-05,
      "loss": 1.3409,
      "theoretical_loss": 3.4196282760023324,
      "tokens_seen": 2300248064
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.119402985074628e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.419612604694543,
      "tokens_seen": 2300379136
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.118600545658804e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.419596934529658,
      "tokens_seen": 2300510208
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.11779810624298e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.4195812655075284,
      "tokens_seen": 2300641280
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.116995666827155e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.4195655976280066,
      "tokens_seen": 2300772352
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.116193227411331e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.419549930890944,
      "tokens_seen": 2300903424
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.115390787995507e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.4195342652961918,
      "tokens_seen": 2301034496
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.114588348579683e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.4195186008436016,
      "tokens_seen": 2301165568
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.113785909163859e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.4195029375330255,
      "tokens_seen": 2301296640
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.112983469748035e-05,
      "loss": 1.3853,
      "theoretical_loss": 3.419487275364315,
      "tokens_seen": 2301427712
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.112181030332211e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.4194716143373225,
      "tokens_seen": 2301558784
    },
    {
      "epoch": 0.39,
      "learning_rate": 6.111378590916386e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.4194559544518985,
      "tokens_seen": 2301689856
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.110576151500563e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.419440295707896,
      "tokens_seen": 2301820928
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.109773712084738e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.4194246381051663,
      "tokens_seen": 2301952000
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.108971272668915e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.419408981643561,
      "tokens_seen": 2302083072
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.10816883325309e-05,
      "loss": 1.3087,
      "theoretical_loss": 3.4193933263229317,
      "tokens_seen": 2302214144
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.107366393837267e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.4193776721431313,
      "tokens_seen": 2302345216
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.106563954421442e-05,
      "loss": 1.296,
      "theoretical_loss": 3.4193620191040113,
      "tokens_seen": 2302476288
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.105761515005617e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.4193463672054234,
      "tokens_seen": 2302607360
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.104959075589794e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.41933071644722,
      "tokens_seen": 2302738432
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.104156636173969e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.4193150668292525,
      "tokens_seen": 2302869504
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.103354196758145e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.4192994183513736,
      "tokens_seen": 2303000576
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.1025517573423206e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.4192837710134345,
      "tokens_seen": 2303131648
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.101749317926497e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.4192681248152885,
      "tokens_seen": 2303262720
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.1009468785106726e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.419252479756787,
      "tokens_seen": 2303393792
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.100144439094848e-05,
      "loss": 1.327,
      "theoretical_loss": 3.419236835837782,
      "tokens_seen": 2303524864
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0993419996790246e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.419221193058126,
      "tokens_seen": 2303655936
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0985395602632e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.4192055514176714,
      "tokens_seen": 2303787008
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0977371208473766e-05,
      "loss": 1.258,
      "theoretical_loss": 3.4191899109162693,
      "tokens_seen": 2303918080
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.096934681431552e-05,
      "loss": 1.3646,
      "theoretical_loss": 3.419174271553773,
      "tokens_seen": 2304049152
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0961322420157285e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.4191586333300354,
      "tokens_seen": 2304180224
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.095329802599904e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.4191429962449074,
      "tokens_seen": 2304311296
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.094527363184079e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.419127360298242,
      "tokens_seen": 2304442368
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.093724923768256e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.4191117254898913,
      "tokens_seen": 2304573440
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.092922484352431e-05,
      "loss": 1.3565,
      "theoretical_loss": 3.419096091819708,
      "tokens_seen": 2304704512
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.092120044936608e-05,
      "loss": 1.307,
      "theoretical_loss": 3.4190804592875446,
      "tokens_seen": 2304835584
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.091317605520783e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.419064827893253,
      "tokens_seen": 2304966656
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.09051516610496e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.4190491976366864,
      "tokens_seen": 2305097728
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.089712726689135e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.4190335685176967,
      "tokens_seen": 2305228800
    },
    {
      "debugging/Compilability": 1.0,
      "debugging/distinct-1-grams": 0.7431796110913657,
      "debugging/entropy-1-grams": 5.572979440278162,
      "debugging/length": 451.2962962962963,
      "debugging/num_segments": 27,
      "debugging/score": 0.010468038290363413,
      "debugging/score_std": 0.015814533663842854,
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.9937704801559448,
      "objective/train/avg_token_score": 0.0062295314855873585,
      "objective/train/docs_used": 839437,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3646734952926636,
      "objective/train/original_loss": 1.3745945692062378,
      "objective/train/theoretical_loss": 3.4190179405361367,
      "objective/train/tokens_used": 675818976,
      "objective/train/unlikelihood_loss": 0.014110151678323746,
      "theoretical_loss": 3.4190179405361367,
      "tokens_seen": 2305359872
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.088910287273312e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.4190179405361367,
      "tokens_seen": 2305359872
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.088107847857487e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.419002313691859,
      "tokens_seen": 2305490944
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0873054084416624e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.4189866879847157,
      "tokens_seen": 2305622016
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.086502969025839e-05,
      "loss": 1.3799,
      "theoretical_loss": 3.4189710634145607,
      "tokens_seen": 2305753088
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0857005296100144e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.418955439981245,
      "tokens_seen": 2305884160
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.084898090194191e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.418939817684622,
      "tokens_seen": 2306015232
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0840956507783664e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.418924196524545,
      "tokens_seen": 2306146304
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.083293211362543e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.418908576500866,
      "tokens_seen": 2306277376
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0824907719467183e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.418892957613438,
      "tokens_seen": 2306408448
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0816883325308937e-05,
      "loss": 1.277,
      "theoretical_loss": 3.4188773398621137,
      "tokens_seen": 2306539520
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.08088589311507e-05,
      "loss": 1.2186,
      "theoretical_loss": 3.4188617232467458,
      "tokens_seen": 2306670592
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0800834536992456e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.4188461077671874,
      "tokens_seen": 2306801664
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.079281014283422e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.418830493423291,
      "tokens_seen": 2306932736
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0784785748675976e-05,
      "loss": 1.3734,
      "theoretical_loss": 3.4188148802149096,
      "tokens_seen": 2307063808
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.077676135451774e-05,
      "loss": 1.3784,
      "theoretical_loss": 3.418799268141896,
      "tokens_seen": 2307194880
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0768736960359496e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.418783657204104,
      "tokens_seen": 2307325952
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.076071256620125e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.418768047401386,
      "tokens_seen": 2307457024
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0752688172043016e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.4187524387335944,
      "tokens_seen": 2307588096
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.074466377788477e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.418736831200583,
      "tokens_seen": 2307719168
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0736639383726536e-05,
      "loss": 1.273,
      "theoretical_loss": 3.4187212248022045,
      "tokens_seen": 2307850240
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.072861498956829e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.418705619538312,
      "tokens_seen": 2307981312
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0720590595410055e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.418690015408759,
      "tokens_seen": 2308112384
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.071256620125181e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.418674412413398,
      "tokens_seen": 2308243456
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.070454180709356e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.418658810552083,
      "tokens_seen": 2308374528
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.069651741293533e-05,
      "loss": 1.3692,
      "theoretical_loss": 3.418643209824666,
      "tokens_seen": 2308505600
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.068849301877708e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.4186276102310016,
      "tokens_seen": 2308636672
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.068046862461885e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.418612011770942,
      "tokens_seen": 2308767744
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.06724442304606e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.4185964144443406,
      "tokens_seen": 2308898816
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.066441983630237e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.4185808182510513,
      "tokens_seen": 2309029888
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.065639544214412e-05,
      "loss": 1.255,
      "theoretical_loss": 3.418565223190927,
      "tokens_seen": 2309160960
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0648371047985874e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.418549629263821,
      "tokens_seen": 2309292032
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.064034665382764e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.4185340364695866,
      "tokens_seen": 2309423104
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0632322259669394e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.418518444808078,
      "tokens_seen": 2309554176
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.062429786551116e-05,
      "loss": 1.3799,
      "theoretical_loss": 3.4185028542791476,
      "tokens_seen": 2309685248
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0616273471352914e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.4184872648826494,
      "tokens_seen": 2309816320
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.060824907719468e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.4184716766184366,
      "tokens_seen": 2309947392
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0600224683036434e-05,
      "loss": 1.3594,
      "theoretical_loss": 3.418456089486363,
      "tokens_seen": 2310078464
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.05922002888782e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.4184405034862824,
      "tokens_seen": 2310209536
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0584175894719954e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.4184249186180478,
      "tokens_seen": 2310340608
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.057615150056171e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.4184093348815128,
      "tokens_seen": 2310471680
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.056812710640347e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.4183937522765317,
      "tokens_seen": 2310602752
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0560102712245226e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.4183781708029572,
      "tokens_seen": 2310733824
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.055207831808699e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.418362590460644,
      "tokens_seen": 2310864896
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0544053923928746e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.4183470112494447,
      "tokens_seen": 2310995968
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.053602952977051e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.418331433169214,
      "tokens_seen": 2311127040
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0528005135612266e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.4183158562198055,
      "tokens_seen": 2311258112
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.051998074145402e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.4183002804010725,
      "tokens_seen": 2311389184
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0511956347295786e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.4182847057128694,
      "tokens_seen": 2311520256
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.050393195313754e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.4182691321550496,
      "tokens_seen": 2311651328
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0495907558979306e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.418253559727467,
      "tokens_seen": 2311782400
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.9930840730667114,
      "objective/train/avg_token_score": 0.006915949285030365,
      "objective/train/docs_used": 841951,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2644598484039307,
      "objective/train/original_loss": 1.2732954025268555,
      "objective/train/theoretical_loss": 3.4182379884299756,
      "objective/train/tokens_used": 682372576,
      "objective/train/unlikelihood_loss": 0.01842862367630005,
      "theoretical_loss": 3.4182379884299756,
      "tokens_seen": 2311913472
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.048788316482106e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.4182379884299756,
      "tokens_seen": 2311913472
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0479858770662826e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.4182224182624297,
      "tokens_seen": 2312044544
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.047183437650458e-05,
      "loss": 1.3131,
      "theoretical_loss": 3.4182068492246827,
      "tokens_seen": 2312175616
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.046380998234633e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.418191281316589,
      "tokens_seen": 2312306688
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.04557855881881e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.418175714538002,
      "tokens_seen": 2312437760
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.044776119402985e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.4181601488887763,
      "tokens_seen": 2312568832
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.043973679987162e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.4181445843687657,
      "tokens_seen": 2312699904
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.043171240571337e-05,
      "loss": 1.1728,
      "theoretical_loss": 3.418129020977825,
      "tokens_seen": 2312830976
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.042368801155514e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.4181134587158066,
      "tokens_seen": 2312962048
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.041566361739689e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.418097897582567,
      "tokens_seen": 2313093120
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0407639223238644e-05,
      "loss": 1.32,
      "theoretical_loss": 3.418082337577958,
      "tokens_seen": 2313224192
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.039961482908041e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.4180667787018355,
      "tokens_seen": 2313355264
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0391590434922164e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.418051220954053,
      "tokens_seen": 2313486336
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.038356604076393e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.418035664334465,
      "tokens_seen": 2313617408
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0375541646605684e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.4180201088429256,
      "tokens_seen": 2313748480
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.036751725244745e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.418004554479289,
      "tokens_seen": 2313879552
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0359492858289204e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.41798900124341,
      "tokens_seen": 2314010624
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.035146846413096e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.417973449135143,
      "tokens_seen": 2314141696
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0343444069972724e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.4179578981543415,
      "tokens_seen": 2314272768
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.033541967581448e-05,
      "loss": 1.3541,
      "theoretical_loss": 3.41794234830086,
      "tokens_seen": 2314403840
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0327395281656243e-05,
      "loss": 1.287,
      "theoretical_loss": 3.4179267995745546,
      "tokens_seen": 2314534912
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0319370887497997e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.4179112519752777,
      "tokens_seen": 2314665984
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0311346493339756e-05,
      "loss": 1.284,
      "theoretical_loss": 3.417895705502885,
      "tokens_seen": 2314797056
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0303322099181516e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.417880160157231,
      "tokens_seen": 2314928128
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.029529770502327e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.4178646159381696,
      "tokens_seen": 2315059200
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0287273310865036e-05,
      "loss": 1.3518,
      "theoretical_loss": 3.4178490728455557,
      "tokens_seen": 2315190272
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.027924891670679e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.4178335308792436,
      "tokens_seen": 2315321344
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0271224522548556e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.417817990039089,
      "tokens_seen": 2315452416
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.026320012839031e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.4178024503249453,
      "tokens_seen": 2315583488
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.025517573423207e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.417786911736668,
      "tokens_seen": 2315714560
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.024715134007383e-05,
      "loss": 1.313,
      "theoretical_loss": 3.4177713742741114,
      "tokens_seen": 2315845632
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.023912694591559e-05,
      "loss": 1.2266,
      "theoretical_loss": 3.4177558379371304,
      "tokens_seen": 2315976704
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.023110255175735e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.4177403027255795,
      "tokens_seen": 2316107776
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.02230781575991e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.4177247686393137,
      "tokens_seen": 2316238848
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.021505376344086e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.4177092356781884,
      "tokens_seen": 2316369920
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.020702936928262e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.4176937038420574,
      "tokens_seen": 2316500992
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.019900497512438e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.417678173130776,
      "tokens_seen": 2316632064
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.019098058096614e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.4176626435441992,
      "tokens_seen": 2316763136
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.01829561868079e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.4176471150821826,
      "tokens_seen": 2316894208
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.017493179264966e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4176315877445793,
      "tokens_seen": 2317025280
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0166907398491414e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.417616061531246,
      "tokens_seen": 2317156352
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0158883004333174e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.417600536442037,
      "tokens_seen": 2317287424
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0150858610174934e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.417585012476808,
      "tokens_seen": 2317418496
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0142834216016694e-05,
      "loss": 1.307,
      "theoretical_loss": 3.4175694896354125,
      "tokens_seen": 2317549568
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0134809821858454e-05,
      "loss": 1.299,
      "theoretical_loss": 3.4175539679177076,
      "tokens_seen": 2317680640
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.0126785427700214e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.4175384473235466,
      "tokens_seen": 2317811712
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.011876103354197e-05,
      "loss": 1.3801,
      "theoretical_loss": 3.417522927852786,
      "tokens_seen": 2317942784
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.011073663938373e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.4175074095052804,
      "tokens_seen": 2318073856
    },
    {
      "epoch": 0.4,
      "learning_rate": 6.010271224522549e-05,
      "loss": 1.3358,
      "theoretical_loss": 3.417491892280885,
      "tokens_seen": 2318204928
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.009468785106725e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.417476376179455,
      "tokens_seen": 2318336000
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.9904979467391968,
      "objective/train/avg_token_score": 0.009502057917416096,
      "objective/train/docs_used": 844398,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3198580741882324,
      "objective/train/original_loss": 1.3330411911010742,
      "objective/train/theoretical_loss": 3.4174608612008455,
      "objective/train/tokens_used": 688926176,
      "objective/train/unlikelihood_loss": 0.025220731273293495,
      "theoretical_loss": 3.4174608612008455,
      "tokens_seen": 2318467072
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.008666345690901e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.4174608612008455,
      "tokens_seen": 2318467072
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.007863906275077e-05,
      "loss": 1.3541,
      "theoretical_loss": 3.4174453473449122,
      "tokens_seen": 2318598144
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.0070614668592527e-05,
      "loss": 1.279,
      "theoretical_loss": 3.41742983461151,
      "tokens_seen": 2318729216
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.006259027443428e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.4174143230004947,
      "tokens_seen": 2318860288
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.005456588027604e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.4173988125117214,
      "tokens_seen": 2318991360
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.00465414861178e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.417383303145045,
      "tokens_seen": 2319122432
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.003851709195956e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.417367794900322,
      "tokens_seen": 2319253504
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.003049269780132e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.4173522877774065,
      "tokens_seen": 2319384576
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.002246830364307e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.417336781776155,
      "tokens_seen": 2319515648
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.001444390948484e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.417321276896423,
      "tokens_seen": 2319646720
    },
    {
      "epoch": 0.41,
      "learning_rate": 6.000641951532659e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.4173057731380654,
      "tokens_seen": 2319777792
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.999839512116835e-05,
      "loss": 1.3493,
      "theoretical_loss": 3.417290270500938,
      "tokens_seen": 2319908864
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.999037072701011e-05,
      "loss": 1.3576,
      "theoretical_loss": 3.417274768984896,
      "tokens_seen": 2320039936
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.998234633285187e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.4172592685897962,
      "tokens_seen": 2320171008
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.997432193869363e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.417243769315493,
      "tokens_seen": 2320302080
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9966297544535385e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.4172282711618425,
      "tokens_seen": 2320433152
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.995827315037715e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.4172127741287004,
      "tokens_seen": 2320564224
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9950248756218905e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.417197278215922,
      "tokens_seen": 2320695296
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.994222436206067e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.4171817834233638,
      "tokens_seen": 2320826368
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9934199967902425e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.417166289750881,
      "tokens_seen": 2320957440
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.992617557374418e-05,
      "loss": 1.361,
      "theoretical_loss": 3.4171507971983295,
      "tokens_seen": 2321088512
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9918151179585944e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.4171353057655653,
      "tokens_seen": 2321219584
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.99101267854277e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.417119815452444,
      "tokens_seen": 2321350656
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9902102391269464e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.4171043262588214,
      "tokens_seen": 2321481728
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.989407799711122e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.417088838184554,
      "tokens_seen": 2321612800
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9886053602952984e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.4170733512294964,
      "tokens_seen": 2321743872
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.987802920879474e-05,
      "loss": 1.2149,
      "theoretical_loss": 3.417057865393506,
      "tokens_seen": 2321874944
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.987000481463649e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.4170423806764383,
      "tokens_seen": 2322006016
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.986198042047826e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.417026897078149,
      "tokens_seen": 2322137088
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.985395602632001e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.4170114145984942,
      "tokens_seen": 2322268160
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.984593163216178e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.41699593323733,
      "tokens_seen": 2322399232
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.983790723800353e-05,
      "loss": 1.2236,
      "theoretical_loss": 3.4169804529945127,
      "tokens_seen": 2322530304
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.98298828438453e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.416964973869898,
      "tokens_seen": 2322661376
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.982185844968705e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.4169494958633426,
      "tokens_seen": 2322792448
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.98138340555288e-05,
      "loss": 1.4392,
      "theoretical_loss": 3.416934018974702,
      "tokens_seen": 2322923520
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.980580966137057e-05,
      "loss": 1.3708,
      "theoretical_loss": 3.4169185432038334,
      "tokens_seen": 2323054592
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.979778526721232e-05,
      "loss": 1.3384,
      "theoretical_loss": 3.4169030685505914,
      "tokens_seen": 2323185664
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.978976087305409e-05,
      "loss": 1.233,
      "theoretical_loss": 3.416887595014834,
      "tokens_seen": 2323316736
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.978173647889584e-05,
      "loss": 1.253,
      "theoretical_loss": 3.4168721225964163,
      "tokens_seen": 2323447808
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.977371208473761e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.416856651295195,
      "tokens_seen": 2323578880
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.976568769057936e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.4168411811110255,
      "tokens_seen": 2323709952
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9757663296421115e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.416825712043766,
      "tokens_seen": 2323841024
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.974963890226288e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.4168102440932717,
      "tokens_seen": 2323972096
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9741614508104635e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.4167947772593985,
      "tokens_seen": 2324103168
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.97335901139464e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.416779311542004,
      "tokens_seen": 2324234240
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9725565719788155e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.4167638469409445,
      "tokens_seen": 2324365312
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.971754132562992e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.4167483834560755,
      "tokens_seen": 2324496384
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9709516931471675e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.416732921087254,
      "tokens_seen": 2324627456
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.970149253731343e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.416717459834337,
      "tokens_seen": 2324758528
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9693468143155195e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.416701999697181,
      "tokens_seen": 2324889600
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.9931883811950684,
      "objective/train/avg_token_score": 0.006811574567109346,
      "objective/train/docs_used": 846164,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2964158058166504,
      "objective/train/original_loss": 1.3061633110046387,
      "objective/train/theoretical_loss": 3.4166865406756415,
      "objective/train/tokens_used": 695479776,
      "objective/train/unlikelihood_loss": 0.019905351102352142,
      "theoretical_loss": 3.4166865406756415,
      "tokens_seen": 2325020672
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.968544374899695e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.4166865406756415,
      "tokens_seen": 2325020672
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9677419354838715e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.4166710827695765,
      "tokens_seen": 2325151744
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.966939496068047e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.416655625978842,
      "tokens_seen": 2325282816
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9661370566522234e-05,
      "loss": 1.2465,
      "theoretical_loss": 3.4166401703032943,
      "tokens_seen": 2325413888
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.965334617236399e-05,
      "loss": 1.2965,
      "theoretical_loss": 3.416624715742791,
      "tokens_seen": 2325544960
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9645321778205754e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.416609262297188,
      "tokens_seen": 2325676032
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.963729738404751e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.416593809966342,
      "tokens_seen": 2325807104
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.962927298988926e-05,
      "loss": 1.338,
      "theoretical_loss": 3.416578358750111,
      "tokens_seen": 2325938176
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.962124859573103e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.41656290864835,
      "tokens_seen": 2326069248
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.961322420157278e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.4165474596609173,
      "tokens_seen": 2326200320
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.960519980741455e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.416532011787669,
      "tokens_seen": 2326331392
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.95971754132563e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.4165165650284623,
      "tokens_seen": 2326462464
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.958915101909807e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.416501119383154,
      "tokens_seen": 2326593536
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.958112662493982e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.4164856748516006,
      "tokens_seen": 2326724608
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.957310223078157e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.4164702314336597,
      "tokens_seen": 2326855680
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.956507783662334e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.4164547891291877,
      "tokens_seen": 2326986752
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.955705344246509e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.4164393479380424,
      "tokens_seen": 2327117824
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.954902904830686e-05,
      "loss": 1.3534,
      "theoretical_loss": 3.4164239078600804,
      "tokens_seen": 2327248896
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.954100465414861e-05,
      "loss": 1.293,
      "theoretical_loss": 3.4164084688951584,
      "tokens_seen": 2327379968
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.953298025999038e-05,
      "loss": 1.226,
      "theoretical_loss": 3.4163930310431336,
      "tokens_seen": 2327511040
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.952495586583213e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.4163775943038637,
      "tokens_seen": 2327642112
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9516931471673886e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.4163621586772055,
      "tokens_seen": 2327773184
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.950890707751565e-05,
      "loss": 1.2512,
      "theoretical_loss": 3.416346724163016,
      "tokens_seen": 2327904256
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9500882683357405e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.4163312907611525,
      "tokens_seen": 2328035328
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.949285828919917e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.416315858471472,
      "tokens_seen": 2328166400
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9484833895040925e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.4163004272938324,
      "tokens_seen": 2328297472
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.947680950088269e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.4162849972280904,
      "tokens_seen": 2328428544
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9468785106724445e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.4162695682741036,
      "tokens_seen": 2328559616
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.94607607125662e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.416254140431729,
      "tokens_seen": 2328690688
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9452736318407965e-05,
      "loss": 1.278,
      "theoretical_loss": 3.416238713700824,
      "tokens_seen": 2328821760
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.944471192424972e-05,
      "loss": 1.3336,
      "theoretical_loss": 3.4162232880812464,
      "tokens_seen": 2328952832
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9436687530091485e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.416207863572853,
      "tokens_seen": 2329083904
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.942866313593324e-05,
      "loss": 1.338,
      "theoretical_loss": 3.416192440175502,
      "tokens_seen": 2329214976
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9420638741775004e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.41617701788905,
      "tokens_seen": 2329346048
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.941261434761676e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.416161596713355,
      "tokens_seen": 2329477120
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.940458995345851e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.416146176648274,
      "tokens_seen": 2329608192
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.939656555930028e-05,
      "loss": 1.2174,
      "theoretical_loss": 3.416130757693666,
      "tokens_seen": 2329739264
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.938854116514203e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.4161153398493864,
      "tokens_seen": 2329870336
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.93805167709838e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.4160999231152944,
      "tokens_seen": 2330001408
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.937249237682555e-05,
      "loss": 1.3871,
      "theoretical_loss": 3.416084507491247,
      "tokens_seen": 2330132480
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.936446798266732e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.416069092977102,
      "tokens_seen": 2330263552
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.935644358850907e-05,
      "loss": 1.26,
      "theoretical_loss": 3.416053679572716,
      "tokens_seen": 2330394624
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.934841919435082e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.4160382672779486,
      "tokens_seen": 2330525696
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.934039480019259e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.4160228560926567,
      "tokens_seen": 2330656768
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.933237040603434e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.416007446016698,
      "tokens_seen": 2330787840
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.932434601187611e-05,
      "loss": 1.2725,
      "theoretical_loss": 3.4159920370499295,
      "tokens_seen": 2330918912
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.931632161771786e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.41597662919221,
      "tokens_seen": 2331049984
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.930829722355963e-05,
      "loss": 1.2202,
      "theoretical_loss": 3.4159612224433973,
      "tokens_seen": 2331181056
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.930027282940138e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.4159458168033487,
      "tokens_seen": 2331312128
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.929224843524315e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.4159304122719227,
      "tokens_seen": 2331443200
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.9901067018508911,
      "objective/train/avg_token_score": 0.009893345646560192,
      "objective/train/docs_used": 848727,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2684025764465332,
      "objective/train/original_loss": 1.2827916145324707,
      "objective/train/theoretical_loss": 3.4159150088489767,
      "objective/train/tokens_used": 702033376,
      "objective/train/unlikelihood_loss": 0.026813924312591553,
      "theoretical_loss": 3.4159150088489767,
      "tokens_seen": 2331574272
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.92842240410849e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.4159150088489767,
      "tokens_seen": 2331574272
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9276199646926656e-05,
      "loss": 1.352,
      "theoretical_loss": 3.4158996065343685,
      "tokens_seen": 2331705344
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.926817525276842e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.4158842053279566,
      "tokens_seen": 2331836416
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9260150858610175e-05,
      "loss": 1.2393,
      "theoretical_loss": 3.4158688052295987,
      "tokens_seen": 2331967488
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.925212646445194e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.415853406239153,
      "tokens_seen": 2332098560
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9244102070293695e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.415838008356477,
      "tokens_seen": 2332229632
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.923607767613546e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.4158226115814294,
      "tokens_seen": 2332360704
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9228053281977215e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.415807215913868,
      "tokens_seen": 2332491776
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.922002888781897e-05,
      "loss": 1.2214,
      "theoretical_loss": 3.415791821353651,
      "tokens_seen": 2332622848
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9212004493660735e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.415776427900636,
      "tokens_seen": 2332753920
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.920398009950249e-05,
      "loss": 1.2379,
      "theoretical_loss": 3.4157610355546817,
      "tokens_seen": 2332884992
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9195955705344255e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.4157456443156464,
      "tokens_seen": 2333016064
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.918793131118601e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.4157302541833885,
      "tokens_seen": 2333147136
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.9179906917027774e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.4157148651577653,
      "tokens_seen": 2333278208
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.917188252286953e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.4156994772386358,
      "tokens_seen": 2333409280
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.916385812871128e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.4156840904258585,
      "tokens_seen": 2333540352
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.915583373455305e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.4156687047192906,
      "tokens_seen": 2333671424
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.91478093403948e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.4156533201187917,
      "tokens_seen": 2333802496
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.913978494623657e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.4156379366242193,
      "tokens_seen": 2333933568
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.913176055207832e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.4156225542354326,
      "tokens_seen": 2334064640
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.912373615792009e-05,
      "loss": 1.271,
      "theoretical_loss": 3.4156071729522894,
      "tokens_seen": 2334195712
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.911571176376184e-05,
      "loss": 1.2335,
      "theoretical_loss": 3.4155917927746486,
      "tokens_seen": 2334326784
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.910768736960359e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.415576413702368,
      "tokens_seen": 2334457856
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.909966297544536e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.415561035735307,
      "tokens_seen": 2334588928
    },
    {
      "epoch": 0.41,
      "learning_rate": 5.909163858128711e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.4155456588733233,
      "tokens_seen": 2334720000
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.908361418712888e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.4155302831162757,
      "tokens_seen": 2334851072
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.907558979297063e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.415514908464023,
      "tokens_seen": 2334982144
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.90675653988124e-05,
      "loss": 1.1889,
      "theoretical_loss": 3.4154995349164237,
      "tokens_seen": 2335113216
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.905954100465415e-05,
      "loss": 1.3878,
      "theoretical_loss": 3.4154841624733367,
      "tokens_seen": 2335244288
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.9051516610495906e-05,
      "loss": 1.3579,
      "theoretical_loss": 3.41546879113462,
      "tokens_seen": 2335375360
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.904349221633767e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.415453420900133,
      "tokens_seen": 2335506432
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.9035467822179426e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.4154380517697343,
      "tokens_seen": 2335637504
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.902744342802119e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.415422683743282,
      "tokens_seen": 2335768576
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.9019419033862945e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.415407316820635,
      "tokens_seen": 2335899648
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.901139463970471e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.4153919510016526,
      "tokens_seen": 2336030720
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.9003370245546465e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.415376586286194,
      "tokens_seen": 2336161792
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8995345851388225e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.415361222674117,
      "tokens_seen": 2336292864
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8987321457229985e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.4153458601652806,
      "tokens_seen": 2336423936
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.897929706307174e-05,
      "loss": 1.336,
      "theoretical_loss": 3.415330498759544,
      "tokens_seen": 2336555008
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8971272668913505e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.4153151384567666,
      "tokens_seen": 2336686080
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.896324827475526e-05,
      "loss": 1.3601,
      "theoretical_loss": 3.4152997792568063,
      "tokens_seen": 2336817152
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8955223880597025e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.415284421159523,
      "tokens_seen": 2336948224
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.894719948643878e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.415269064164775,
      "tokens_seen": 2337079296
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.893917509228054e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.415253708272422,
      "tokens_seen": 2337210368
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.89311506981223e-05,
      "loss": 1.3441,
      "theoretical_loss": 3.415238353482323,
      "tokens_seen": 2337341440
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.892312630396405e-05,
      "loss": 1.3483,
      "theoretical_loss": 3.415222999794336,
      "tokens_seen": 2337472512
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.891510190980582e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.4152076472083213,
      "tokens_seen": 2337603584
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.890707751564757e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.4151922957241374,
      "tokens_seen": 2337734656
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.889905312148933e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.415176945341644,
      "tokens_seen": 2337865728
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.889102872733109e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.4151615960607,
      "tokens_seen": 2337996800
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.9922761917114258,
      "objective/train/avg_token_score": 0.007723750080913305,
      "objective/train/docs_used": 851124,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3660627603530884,
      "objective/train/original_loss": 1.3770129680633545,
      "objective/train/theoretical_loss": 3.4151462478811636,
      "objective/train/tokens_used": 708586976,
      "objective/train/unlikelihood_loss": 0.019970623776316643,
      "theoretical_loss": 3.4151462478811636,
      "tokens_seen": 2338127872
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.888300433317285e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.4151462478811636,
      "tokens_seen": 2338127872
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.887497993901461e-05,
      "loss": 1.291,
      "theoretical_loss": 3.415130900802896,
      "tokens_seen": 2338258944
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.886695554485636e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.415115554825755,
      "tokens_seen": 2338390016
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.885893115069813e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.4151002099496,
      "tokens_seen": 2338521088
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.885090675653988e-05,
      "loss": 1.3732,
      "theoretical_loss": 3.415084866174291,
      "tokens_seen": 2338652160
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.884288236238164e-05,
      "loss": 1.315,
      "theoretical_loss": 3.415069523499687,
      "tokens_seen": 2338783232
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.88348579682234e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.4150541819256475,
      "tokens_seen": 2338914304
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.882683357406516e-05,
      "loss": 1.2135,
      "theoretical_loss": 3.415038841452031,
      "tokens_seen": 2339045376
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.881880917990692e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.4150235020786983,
      "tokens_seen": 2339176448
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8810784785748676e-05,
      "loss": 1.273,
      "theoretical_loss": 3.4150081638055076,
      "tokens_seen": 2339307520
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8802760391590436e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.414992826632319,
      "tokens_seen": 2339438592
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8794735997432196e-05,
      "loss": 1.4048,
      "theoretical_loss": 3.414977490558992,
      "tokens_seen": 2339569664
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8786711603273956e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.414962155585386,
      "tokens_seen": 2339700736
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8778687209115716e-05,
      "loss": 1.2403,
      "theoretical_loss": 3.4149468217113608,
      "tokens_seen": 2339831808
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8770662814957475e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.4149314889367752,
      "tokens_seen": 2339962880
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8762638420799235e-05,
      "loss": 1.282,
      "theoretical_loss": 3.4149161572614894,
      "tokens_seen": 2340093952
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.875461402664099e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.414900826685363,
      "tokens_seen": 2340225024
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.874658963248275e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.414885497208256,
      "tokens_seen": 2340356096
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.873856523832451e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.414870168830027,
      "tokens_seen": 2340487168
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.873054084416627e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.414854841550537,
      "tokens_seen": 2340618240
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.872251645000803e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.4148395153696445,
      "tokens_seen": 2340749312
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.871449205584979e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.41482419028721,
      "tokens_seen": 2340880384
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.870646766169154e-05,
      "loss": 1.3457,
      "theoretical_loss": 3.414808866303093,
      "tokens_seen": 2341011456
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.869844326753331e-05,
      "loss": 1.2248,
      "theoretical_loss": 3.414793543417153,
      "tokens_seen": 2341142528
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.869041887337506e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.414778221629251,
      "tokens_seen": 2341273600
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.868239447921682e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.4147629009392455,
      "tokens_seen": 2341404672
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.867437008505858e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.414747581346997,
      "tokens_seen": 2341535744
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8666345690900334e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.4147322628523655,
      "tokens_seen": 2341666816
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.86583212967421e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.4147169454552104,
      "tokens_seen": 2341797888
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8650296902583854e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.4147016291553918,
      "tokens_seen": 2341928960
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.864227250842562e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.4146863139527706,
      "tokens_seen": 2342060032
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8634248114267374e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4146709998472056,
      "tokens_seen": 2342191104
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8626223720109133e-05,
      "loss": 1.3834,
      "theoretical_loss": 3.4146556868385574,
      "tokens_seen": 2342322176
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.861819932595089e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.414640374926686,
      "tokens_seen": 2342453248
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8610174931792646e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.4146250641114513,
      "tokens_seen": 2342584320
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.860215053763441e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.414609754392714,
      "tokens_seen": 2342715392
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8594126143476166e-05,
      "loss": 1.363,
      "theoretical_loss": 3.414594445770333,
      "tokens_seen": 2342846464
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.858610174931793e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.41457913824417,
      "tokens_seen": 2342977536
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8578077355159686e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.414563831814084,
      "tokens_seen": 2343108608
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.857005296100144e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.414548526479936,
      "tokens_seen": 2343239680
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8562028566843206e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.4145332222415856,
      "tokens_seen": 2343370752
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.855400417268496e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.4145179190988935,
      "tokens_seen": 2343501824
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8545979778526726e-05,
      "loss": 1.3797,
      "theoretical_loss": 3.4145026170517196,
      "tokens_seen": 2343632896
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.853795538436848e-05,
      "loss": 1.296,
      "theoretical_loss": 3.4144873160999243,
      "tokens_seen": 2343763968
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8529930990210246e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.4144720162433684,
      "tokens_seen": 2343895040
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8521906596052e-05,
      "loss": 1.3,
      "theoretical_loss": 3.414456717481912,
      "tokens_seen": 2344026112
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.851388220189375e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.4144414198154154,
      "tokens_seen": 2344157184
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.850585780773552e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.4144261232437385,
      "tokens_seen": 2344288256
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.849783341357727e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.4144108277667424,
      "tokens_seen": 2344419328
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.848980901941904e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.4143955333842877,
      "tokens_seen": 2344550400
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.9915265440940857,
      "objective/train/avg_token_score": 0.008473447524011135,
      "objective/train/docs_used": 853299,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2087483406066895,
      "objective/train/original_loss": 1.2207425832748413,
      "objective/train/theoretical_loss": 3.414380240096235,
      "objective/train/tokens_used": 715140576,
      "objective/train/unlikelihood_loss": 0.024549948051571846,
      "theoretical_loss": 3.414380240096235,
      "tokens_seen": 2344681472
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.848178462526079e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.414380240096235,
      "tokens_seen": 2344681472
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.847376023110256e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.414364947902444,
      "tokens_seen": 2344812544
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.846573583694431e-05,
      "loss": 1.245,
      "theoretical_loss": 3.4143496568027754,
      "tokens_seen": 2344943616
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8457711442786064e-05,
      "loss": 1.3817,
      "theoretical_loss": 3.4143343667970907,
      "tokens_seen": 2345074688
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.844968704862783e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.4143190778852492,
      "tokens_seen": 2345205760
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8441662654469584e-05,
      "loss": 1.2014,
      "theoretical_loss": 3.4143037900671125,
      "tokens_seen": 2345336832
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.843363826031135e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.414288503342541,
      "tokens_seen": 2345467904
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8425613866153104e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.4142732177113952,
      "tokens_seen": 2345598976
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.841758947199487e-05,
      "loss": 1.1981,
      "theoretical_loss": 3.414257933173536,
      "tokens_seen": 2345730048
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8409565077836624e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.414242649728824,
      "tokens_seen": 2345861120
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.840154068367838e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.4142273673771197,
      "tokens_seen": 2345992192
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8393516289520144e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.414212086118284,
      "tokens_seen": 2346123264
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.83854918953619e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.4141968059521783,
      "tokens_seen": 2346254336
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8377467501203663e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.4141815268786626,
      "tokens_seen": 2346385408
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8369443107045417e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.4141662488975983,
      "tokens_seen": 2346516480
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.836141871288718e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.414150972008846,
      "tokens_seen": 2346647552
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8353394318728936e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.4141356962122664,
      "tokens_seen": 2346778624
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.83453699245707e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.4141204215077208,
      "tokens_seen": 2346909696
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8337345530412456e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.4141051478950706,
      "tokens_seen": 2347040768
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.832932113625421e-05,
      "loss": 1.3333,
      "theoretical_loss": 3.4140898753741755,
      "tokens_seen": 2347171840
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8321296742095976e-05,
      "loss": 1.32,
      "theoretical_loss": 3.4140746039448975,
      "tokens_seen": 2347302912
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.831327234793773e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.4140593336070975,
      "tokens_seen": 2347433984
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8305247953779496e-05,
      "loss": 1.2662,
      "theoretical_loss": 3.414044064360636,
      "tokens_seen": 2347565056
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.829722355962125e-05,
      "loss": 1.3421,
      "theoretical_loss": 3.4140287962053746,
      "tokens_seen": 2347696128
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8289199165463016e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.414013529141174,
      "tokens_seen": 2347827200
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.828117477130477e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.4139982631678962,
      "tokens_seen": 2347958272
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.827315037714652e-05,
      "loss": 1.313,
      "theoretical_loss": 3.4139829982854017,
      "tokens_seen": 2348089344
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.826512598298829e-05,
      "loss": 1.2034,
      "theoretical_loss": 3.4139677344935517,
      "tokens_seen": 2348220416
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.825710158883004e-05,
      "loss": 1.2016,
      "theoretical_loss": 3.413952471792207,
      "tokens_seen": 2348351488
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.824907719467181e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.4139372101812295,
      "tokens_seen": 2348482560
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.824105280051356e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.4139219496604802,
      "tokens_seen": 2348613632
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.823302840635533e-05,
      "loss": 1.3522,
      "theoretical_loss": 3.4139066902298207,
      "tokens_seen": 2348744704
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.822500401219708e-05,
      "loss": 1.2042,
      "theoretical_loss": 3.4138914318891116,
      "tokens_seen": 2348875776
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8216979618038834e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.413876174638215,
      "tokens_seen": 2349006848
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.82089552238806e-05,
      "loss": 1.321,
      "theoretical_loss": 3.413860918476992,
      "tokens_seen": 2349137920
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8200930829722354e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.4138456634053034,
      "tokens_seen": 2349268992
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.819290643556412e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.4138304094230114,
      "tokens_seen": 2349400064
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8184882041405874e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.4138151565299775,
      "tokens_seen": 2349531136
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.817685764724764e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.413799904726062,
      "tokens_seen": 2349662208
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8168833253089394e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.4137846540111276,
      "tokens_seen": 2349793280
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.816080885893115e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.4137694043850355,
      "tokens_seen": 2349924352
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8152784464772914e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.413754155847647,
      "tokens_seen": 2350055424
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.814476007061467e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.4137389083988237,
      "tokens_seen": 2350186496
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8136735676456433e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.4137236620384277,
      "tokens_seen": 2350317568
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.812871128229819e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.4137084167663194,
      "tokens_seen": 2350448640
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.812068688813995e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.4136931725823616,
      "tokens_seen": 2350579712
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8112662493981706e-05,
      "loss": 1.3583,
      "theoretical_loss": 3.4136779294864157,
      "tokens_seen": 2350710784
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.810463809982346e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.413662687478343,
      "tokens_seen": 2350841856
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.8096613705665226e-05,
      "loss": 1.332,
      "theoretical_loss": 3.413647446558006,
      "tokens_seen": 2350972928
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.808858931150698e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.413632206725265,
      "tokens_seen": 2351104000
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.991912305355072,
      "objective/train/avg_token_score": 0.008087667636573315,
      "objective/train/docs_used": 855824,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.306010365486145,
      "objective/train/original_loss": 1.317354440689087,
      "objective/train/theoretical_loss": 3.4136169679799835,
      "objective/train/tokens_used": 721694176,
      "objective/train/unlikelihood_loss": 0.022028451785445213,
      "theoretical_loss": 3.4136169679799835,
      "tokens_seen": 2351235072
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.8080564917348746e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.4136169679799835,
      "tokens_seen": 2351235072
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.80725405231905e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.4136017303220223,
      "tokens_seen": 2351366144
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.8064516129032266e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.413586493751243,
      "tokens_seen": 2351497216
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.805649173487402e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.413571258267508,
      "tokens_seen": 2351628288
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.8048467340715786e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.413556023870679,
      "tokens_seen": 2351759360
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.804044294655754e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.413540790560618,
      "tokens_seen": 2351890432
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.803241855239929e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.4135255583371866,
      "tokens_seen": 2352021504
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.802439415824106e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.413510327200247,
      "tokens_seen": 2352152576
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.801636976408281e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.413495097149661,
      "tokens_seen": 2352283648
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.800834536992458e-05,
      "loss": 1.4111,
      "theoretical_loss": 3.413479868185291,
      "tokens_seen": 2352414720
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.800032097576633e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.4134646403069984,
      "tokens_seen": 2352545792
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.79922965816081e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.4134494135146456,
      "tokens_seen": 2352676864
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.798427218744985e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.4134341878080945,
      "tokens_seen": 2352807936
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7976247793291605e-05,
      "loss": 1.273,
      "theoretical_loss": 3.4134189631872074,
      "tokens_seen": 2352939008
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.796822339913337e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.413403739651846,
      "tokens_seen": 2353070080
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7960199004975124e-05,
      "loss": 1.253,
      "theoretical_loss": 3.4133885172018728,
      "tokens_seen": 2353201152
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.795217461081689e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.4133732958371503,
      "tokens_seen": 2353332224
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7944150216658644e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.4133580755575403,
      "tokens_seen": 2353463296
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.793612582250041e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.4133428563629047,
      "tokens_seen": 2353594368
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7928101428342164e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.413327638253106,
      "tokens_seen": 2353725440
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.792007703418392e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.4133124212280066,
      "tokens_seen": 2353856512
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7912052640025684e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.413297205287469,
      "tokens_seen": 2353987584
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.790402824586744e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.413281990431355,
      "tokens_seen": 2354118656
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7896003851709204e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.413266776659527,
      "tokens_seen": 2354249728
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.788797945755096e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.413251563971848,
      "tokens_seen": 2354380800
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.787995506339272e-05,
      "loss": 1.2949,
      "theoretical_loss": 3.4132363523681795,
      "tokens_seen": 2354511872
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7871930669234477e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.4132211418483847,
      "tokens_seen": 2354642944
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.786390627507623e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.4132059324123256,
      "tokens_seen": 2354774016
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7855881880917996e-05,
      "loss": 1.225,
      "theoretical_loss": 3.4131907240598647,
      "tokens_seen": 2354905088
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.784785748675975e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.4131755167908646,
      "tokens_seen": 2355036160
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7839833092601516e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.4131603106051878,
      "tokens_seen": 2355167232
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.783180869844327e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.413145105502697,
      "tokens_seen": 2355298304
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7823784304285036e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.4131299014832543,
      "tokens_seen": 2355429376
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.781575991012679e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4131146985467224,
      "tokens_seen": 2355560448
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.780773551596854e-05,
      "loss": 1.291,
      "theoretical_loss": 3.4130994966929644,
      "tokens_seen": 2355691520
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.779971112181031e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.4130842959218426,
      "tokens_seen": 2355822592
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.779168672765206e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.4130690962332197,
      "tokens_seen": 2355953664
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.778366233349383e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.4130538976269587,
      "tokens_seen": 2356084736
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.777563793933558e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.4130387001029217,
      "tokens_seen": 2356215808
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.776761354517735e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.4130235036609715,
      "tokens_seen": 2356346880
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.77595891510191e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.413008308300972,
      "tokens_seen": 2356477952
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.775156475686087e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.4129931140227843,
      "tokens_seen": 2356609024
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.774354036270262e-05,
      "loss": 1.2453,
      "theoretical_loss": 3.4129779208262723,
      "tokens_seen": 2356740096
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7735515968544375e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.4129627287112987,
      "tokens_seen": 2356871168
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.772749157438614e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.4129475376777263,
      "tokens_seen": 2357002240
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7719467180227894e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.412932347725418,
      "tokens_seen": 2357133312
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.771144278606966e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.4129171588542366,
      "tokens_seen": 2357264384
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7703418391911414e-05,
      "loss": 1.3471,
      "theoretical_loss": 3.412901971064045,
      "tokens_seen": 2357395456
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.769539399775318e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.4128867843547064,
      "tokens_seen": 2357526528
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7687369603594934e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.412871598726084,
      "tokens_seen": 2357657600
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.9918233156204224,
      "objective/train/avg_token_score": 0.008176704868674278,
      "objective/train/docs_used": 858137,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2057864665985107,
      "objective/train/original_loss": 1.2163323163986206,
      "objective/train/theoretical_loss": 3.41285641417804,
      "objective/train/tokens_used": 728247776,
      "objective/train/unlikelihood_loss": 0.02831781655550003,
      "theoretical_loss": 3.41285641417804,
      "tokens_seen": 2357788672
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.767934520943669e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.41285641417804,
      "tokens_seen": 2357788672
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7671320815278454e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.4128412307104377,
      "tokens_seen": 2357919744
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.766329642112021e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.412826048323141,
      "tokens_seen": 2358050816
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7655272026961974e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.412810867016012,
      "tokens_seen": 2358181888
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.764724763280373e-05,
      "loss": 1.313,
      "theoretical_loss": 3.4127956867889147,
      "tokens_seen": 2358312960
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.763922323864549e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.4127805076417115,
      "tokens_seen": 2358444032
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7631198844487247e-05,
      "loss": 1.2633,
      "theoretical_loss": 3.412765329574266,
      "tokens_seen": 2358575104
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7623174450329e-05,
      "loss": 1.239,
      "theoretical_loss": 3.412750152586441,
      "tokens_seen": 2358706176
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7615150056170766e-05,
      "loss": 1.3565,
      "theoretical_loss": 3.4127349766781,
      "tokens_seen": 2358837248
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.760712566201252e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.4127198018491067,
      "tokens_seen": 2358968320
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7599101267854286e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.4127046280993234,
      "tokens_seen": 2359099392
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.759107687369604e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.412689455428614,
      "tokens_seen": 2359230464
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.75830524795378e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.4126742838368416,
      "tokens_seen": 2359361536
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.757502808537956e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.4126591133238695,
      "tokens_seen": 2359492608
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.756700369122131e-05,
      "loss": 1.308,
      "theoretical_loss": 3.4126439438895613,
      "tokens_seen": 2359623680
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.755897929706308e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.41262877553378,
      "tokens_seen": 2359754752
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.755095490290483e-05,
      "loss": 1.346,
      "theoretical_loss": 3.41261360825639,
      "tokens_seen": 2359885824
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.754293050874659e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.412598442057254,
      "tokens_seen": 2360016896
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.753490611458835e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.4125832769362345,
      "tokens_seen": 2360147968
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.752688172043011e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.4125681128931973,
      "tokens_seen": 2360279040
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.751885732627187e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.4125529499280036,
      "tokens_seen": 2360410112
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7510832932113625e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.4125377880405185,
      "tokens_seen": 2360541184
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.750280853795539e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.412522627230605,
      "tokens_seen": 2360672256
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7494784143797145e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.4125074674981266,
      "tokens_seen": 2360803328
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7486759749638905e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.412492308842947,
      "tokens_seen": 2360934400
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7478735355480664e-05,
      "loss": 1.284,
      "theoretical_loss": 3.41247715126493,
      "tokens_seen": 2361065472
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7470710961322424e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.412461994763939,
      "tokens_seen": 2361196544
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7462686567164184e-05,
      "loss": 1.2594,
      "theoretical_loss": 3.412446839339838,
      "tokens_seen": 2361327616
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.745466217300594e-05,
      "loss": 1.2186,
      "theoretical_loss": 3.4124316849924905,
      "tokens_seen": 2361458688
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.74466377788477e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.4124165317217603,
      "tokens_seen": 2361589760
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.743861338468946e-05,
      "loss": 1.2051,
      "theoretical_loss": 3.4124013795275117,
      "tokens_seen": 2361720832
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.743058899053122e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.4123862284096074,
      "tokens_seen": 2361851904
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.742256459637298e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.4123710783679115,
      "tokens_seen": 2361982976
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.741454020221474e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.4123559294022883,
      "tokens_seen": 2362114048
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.74065158080565e-05,
      "loss": 1.3333,
      "theoretical_loss": 3.4123407815126017,
      "tokens_seen": 2362245120
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.739849141389826e-05,
      "loss": 1.2171,
      "theoretical_loss": 3.4123256346987154,
      "tokens_seen": 2362376192
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.739046701974001e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.412310488960493,
      "tokens_seen": 2362507264
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.738244262558177e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.412295344297799,
      "tokens_seen": 2362638336
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.737441823142353e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.412280200710497,
      "tokens_seen": 2362769408
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.736639383726529e-05,
      "loss": 1.3485,
      "theoretical_loss": 3.412265058198451,
      "tokens_seen": 2362900480
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.735836944310705e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.412249916761525,
      "tokens_seen": 2363031552
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.73503450489488e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.4122347763995835,
      "tokens_seen": 2363162624
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.734232065479057e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.4122196371124898,
      "tokens_seen": 2363293696
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.733429626063232e-05,
      "loss": 1.295,
      "theoretical_loss": 3.4122044989001084,
      "tokens_seen": 2363424768
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.732627186647408e-05,
      "loss": 1.3593,
      "theoretical_loss": 3.4121893617623034,
      "tokens_seen": 2363555840
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.731824747231584e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.412174225698939,
      "tokens_seen": 2363686912
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.73102230781576e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.4121590907098795,
      "tokens_seen": 2363817984
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.730219868399936e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.4121439567949885,
      "tokens_seen": 2363949056
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7294174289841115e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.4121288239541308,
      "tokens_seen": 2364080128
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.728614989568288e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.4121136921871704,
      "tokens_seen": 2364211200
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.9936105012893677,
      "objective/train/avg_token_score": 0.0063894521445035934,
      "objective/train/docs_used": 860660,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2686890363693237,
      "objective/train/original_loss": 1.2767412662506104,
      "objective/train/theoretical_loss": 3.4120985614939716,
      "objective/train/tokens_used": 734801376,
      "objective/train/unlikelihood_loss": 0.017710207030177116,
      "theoretical_loss": 3.4120985614939716,
      "tokens_seen": 2364342272
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7278125501524635e-05,
      "loss": 1.267,
      "theoretical_loss": 3.4120985614939716,
      "tokens_seen": 2364342272
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7270101107366395e-05,
      "loss": 1.3703,
      "theoretical_loss": 3.4120834318743984,
      "tokens_seen": 2364473344
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7262076713208155e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.412068303328316,
      "tokens_seen": 2364604416
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.725405231904991e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.4120531758555877,
      "tokens_seen": 2364735488
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7246027924891675e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.4120380494560782,
      "tokens_seen": 2364866560
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.723800353073343e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.412022924129652,
      "tokens_seen": 2364997632
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7229979136575194e-05,
      "loss": 1.301,
      "theoretical_loss": 3.4120077998761733,
      "tokens_seen": 2365128704
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.722195474241695e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.411992676695507,
      "tokens_seen": 2365259776
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.721393034825871e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.411977554587517,
      "tokens_seen": 2365390848
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.720590595410047e-05,
      "loss": 1.342,
      "theoretical_loss": 3.4119624335520684,
      "tokens_seen": 2365521920
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.719788155994222e-05,
      "loss": 1.3384,
      "theoretical_loss": 3.411947313589025,
      "tokens_seen": 2365652992
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.718985716578399e-05,
      "loss": 1.1842,
      "theoretical_loss": 3.411932194698252,
      "tokens_seen": 2365784064
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.718183277162574e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.411917076879613,
      "tokens_seen": 2365915136
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.717380837746751e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.411901960132974,
      "tokens_seen": 2366046208
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.716578398330926e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.4118868444581985,
      "tokens_seen": 2366177280
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.715775958915101e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.411871729855151,
      "tokens_seen": 2366308352
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.714973519499278e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.411856616323697,
      "tokens_seen": 2366439424
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.714171080083453e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.411841503863701,
      "tokens_seen": 2366570496
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.71336864066763e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.411826392475027,
      "tokens_seen": 2366701568
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.712566201251805e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.411811282157541,
      "tokens_seen": 2366832640
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.711763761835982e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.411796172911106,
      "tokens_seen": 2366963712
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.710961322420157e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.4117810647355884,
      "tokens_seen": 2367094784
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.710158883004334e-05,
      "loss": 1.301,
      "theoretical_loss": 3.4117659576308523,
      "tokens_seen": 2367225856
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.709356443588509e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.4117508515967625,
      "tokens_seen": 2367356928
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.7085540041726846e-05,
      "loss": 1.1888,
      "theoretical_loss": 3.411735746633184,
      "tokens_seen": 2367488000
    },
    {
      "epoch": 0.43,
      "learning_rate": 5.707751564756861e-05,
      "loss": 1.32,
      "theoretical_loss": 3.411720642739981,
      "tokens_seen": 2367619072
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.7069491253410365e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.41170553991702,
      "tokens_seen": 2367750144
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.706146685925213e-05,
      "loss": 1.259,
      "theoretical_loss": 3.411690438164164,
      "tokens_seen": 2367881216
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.7053442465093885e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.411675337481279,
      "tokens_seen": 2368012288
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.704541807093565e-05,
      "loss": 1.1906,
      "theoretical_loss": 3.4116602378682304,
      "tokens_seen": 2368143360
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.7037393676777405e-05,
      "loss": 1.3949,
      "theoretical_loss": 3.4116451393248823,
      "tokens_seen": 2368274432
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.702936928261916e-05,
      "loss": 1.2538,
      "theoretical_loss": 3.4116300418511,
      "tokens_seen": 2368405504
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.7021344888460925e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.411614945446749,
      "tokens_seen": 2368536576
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.701332049430268e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.4115998501116938,
      "tokens_seen": 2368667648
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.7005296100144445e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.4115847558457997,
      "tokens_seen": 2368798720
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.69972717059862e-05,
      "loss": 1.3469,
      "theoretical_loss": 3.411569662648932,
      "tokens_seen": 2368929792
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6989247311827965e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.411554570520955,
      "tokens_seen": 2369060864
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.698122291766972e-05,
      "loss": 1.1841,
      "theoretical_loss": 3.411539479461735,
      "tokens_seen": 2369191936
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.697319852351147e-05,
      "loss": 1.331,
      "theoretical_loss": 3.4115243894711367,
      "tokens_seen": 2369323008
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.696517412935324e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.4115093005490253,
      "tokens_seen": 2369454080
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.695714973519499e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.411494212695266,
      "tokens_seen": 2369585152
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.694912534103676e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.4114791259097244,
      "tokens_seen": 2369716224
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.694110094687851e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.411464040192265,
      "tokens_seen": 2369847296
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.693307655272028e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.4114489555427543,
      "tokens_seen": 2369978368
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.692505215856203e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.411433871961057,
      "tokens_seen": 2370109440
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.691702776440378e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.411418789447038,
      "tokens_seen": 2370240512
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.690900337024555e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.4114037080005635,
      "tokens_seen": 2370371584
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.69009789760873e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.4113886276214984,
      "tokens_seen": 2370502656
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.689295458192907e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.411373548309708,
      "tokens_seen": 2370633728
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.688493018777082e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.4113584700650588,
      "tokens_seen": 2370764800
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.9924556016921997,
      "objective/train/avg_token_score": 0.0075443582609295845,
      "objective/train/docs_used": 863154,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2047375440597534,
      "objective/train/original_loss": 1.2144237756729126,
      "objective/train/theoretical_loss": 3.4113433928874146,
      "objective/train/tokens_used": 741354976,
      "objective/train/unlikelihood_loss": 0.022444510832428932,
      "theoretical_loss": 3.4113433928874146,
      "tokens_seen": 2370895872
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.687690579361259e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.4113433928874146,
      "tokens_seen": 2370895872
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.686888139945434e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.411328316776643,
      "tokens_seen": 2371026944
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6860857005296096e-05,
      "loss": 1.3655,
      "theoretical_loss": 3.4113132417326075,
      "tokens_seen": 2371158016
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.685283261113786e-05,
      "loss": 1.333,
      "theoretical_loss": 3.411298167755175,
      "tokens_seen": 2371289088
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6844808216979616e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.4112830948442108,
      "tokens_seen": 2371420160
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.683678382282138e-05,
      "loss": 1.3568,
      "theoretical_loss": 3.4112680229995798,
      "tokens_seen": 2371551232
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6828759428663136e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.4112529522211483,
      "tokens_seen": 2371682304
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.68207350345049e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.4112378825087823,
      "tokens_seen": 2371813376
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6812710640346655e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.411222813862347,
      "tokens_seen": 2371944448
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.680468624618841e-05,
      "loss": 1.277,
      "theoretical_loss": 3.4112077462817076,
      "tokens_seen": 2372075520
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6796661852030175e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.4111926797667307,
      "tokens_seen": 2372206592
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.678863745787193e-05,
      "loss": 1.2143,
      "theoretical_loss": 3.411177614317282,
      "tokens_seen": 2372337664
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6780613063713695e-05,
      "loss": 1.2151,
      "theoretical_loss": 3.4111625499332265,
      "tokens_seen": 2372468736
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.677258866955545e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.4111474866144307,
      "tokens_seen": 2372599808
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6764564275397215e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.4111324243607606,
      "tokens_seen": 2372730880
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.675653988123897e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.4111173631720813,
      "tokens_seen": 2372861952
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6748515487080735e-05,
      "loss": 1.2253,
      "theoretical_loss": 3.411102303048259,
      "tokens_seen": 2372993024
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.674049109292249e-05,
      "loss": 1.3792,
      "theoretical_loss": 3.4110872439891597,
      "tokens_seen": 2373124096
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.673246669876424e-05,
      "loss": 1.3932,
      "theoretical_loss": 3.41107218599465,
      "tokens_seen": 2373255168
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.672444230460601e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.411057129064594,
      "tokens_seen": 2373386240
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.671641791044776e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.4110420731988595,
      "tokens_seen": 2373517312
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.670839351628953e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.411027018397312,
      "tokens_seen": 2373648384
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.670036912213128e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.411011964659817,
      "tokens_seen": 2373779456
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.669234472797305e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.410996911986241,
      "tokens_seen": 2373910528
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.66843203338148e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.41098186037645,
      "tokens_seen": 2374041600
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6676295939656553e-05,
      "loss": 1.2298,
      "theoretical_loss": 3.41096680983031,
      "tokens_seen": 2374172672
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.666827154549832e-05,
      "loss": 1.2227,
      "theoretical_loss": 3.4109517603476873,
      "tokens_seen": 2374303744
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.666024715134007e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.4109367119284477,
      "tokens_seen": 2374434816
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.665222275718184e-05,
      "loss": 1.3097,
      "theoretical_loss": 3.4109216645724576,
      "tokens_seen": 2374565888
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.664419836302359e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.4109066182795833,
      "tokens_seen": 2374696960
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.663617396886536e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.410891573049691,
      "tokens_seen": 2374828032
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.662814957470711e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.4108765288826466,
      "tokens_seen": 2374959104
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6620125180548866e-05,
      "loss": 1.3829,
      "theoretical_loss": 3.410861485778317,
      "tokens_seen": 2375090176
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.661210078639063e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.4108464437365678,
      "tokens_seen": 2375221248
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6604076392232386e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.4108314027572657,
      "tokens_seen": 2375352320
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.659605199807415e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.4108163628402766,
      "tokens_seen": 2375483392
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6588027603915906e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.4108013239854675,
      "tokens_seen": 2375614464
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.658000320975767e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.4107862861927045,
      "tokens_seen": 2375745536
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6571978815599425e-05,
      "loss": 1.285,
      "theoretical_loss": 3.4107712494618543,
      "tokens_seen": 2375876608
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.656395442144118e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.4107562137927827,
      "tokens_seen": 2376007680
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6555930027282945e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.410741179185356,
      "tokens_seen": 2376138752
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.65479056331247e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.410726145639442,
      "tokens_seen": 2376269824
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6539881238966465e-05,
      "loss": 1.2074,
      "theoretical_loss": 3.410711113154906,
      "tokens_seen": 2376400896
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.653185684480822e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.410696081731615,
      "tokens_seen": 2376531968
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6523832450649985e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.4106810513694357,
      "tokens_seen": 2376663040
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.651580805649174e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.4106660220682343,
      "tokens_seen": 2376794112
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.650778366233349e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.4106509938278773,
      "tokens_seen": 2376925184
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.649975926817526e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.4106359666482318,
      "tokens_seen": 2377056256
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.649173487401701e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.4106209405291645,
      "tokens_seen": 2377187328
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.648371047985878e-05,
      "loss": 1.301,
      "theoretical_loss": 3.4106059154705415,
      "tokens_seen": 2377318400
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.9902986288070679,
      "objective/train/avg_token_score": 0.009701413102447987,
      "objective/train/docs_used": 865580,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3254047632217407,
      "objective/train/original_loss": 1.3394651412963867,
      "objective/train/theoretical_loss": 3.41059089147223,
      "objective/train/tokens_used": 747908576,
      "objective/train/unlikelihood_loss": 0.026240406557917595,
      "theoretical_loss": 3.41059089147223,
      "tokens_seen": 2377449472
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.647568608570053e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.41059089147223,
      "tokens_seen": 2377449472
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.64676616915423e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.410575868534096,
      "tokens_seen": 2377580544
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.645963729738405e-05,
      "loss": 1.2139,
      "theoretical_loss": 3.4105608466560073,
      "tokens_seen": 2377711616
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.645161290322582e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.41054582583783,
      "tokens_seen": 2377842688
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.644358850906757e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.4105308060794313,
      "tokens_seen": 2377973760
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6435564114909324e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.4105157873806773,
      "tokens_seen": 2378104832
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.642753972075109e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.410500769741436,
      "tokens_seen": 2378235904
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.641951532659284e-05,
      "loss": 1.2949,
      "theoretical_loss": 3.4104857531615727,
      "tokens_seen": 2378366976
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.641149093243461e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.410470737640956,
      "tokens_seen": 2378498048
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.640346653827636e-05,
      "loss": 1.287,
      "theoretical_loss": 3.4104557231794512,
      "tokens_seen": 2378629120
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.639544214411813e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.4104407097769265,
      "tokens_seen": 2378760192
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.638741774995988e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.4104256974332485,
      "tokens_seen": 2378891264
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6379393355801636e-05,
      "loss": 1.3471,
      "theoretical_loss": 3.410410686148284,
      "tokens_seen": 2379022336
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.63713689616434e-05,
      "loss": 1.3687,
      "theoretical_loss": 3.4103956759219,
      "tokens_seen": 2379153408
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6363344567485156e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.410380666753964,
      "tokens_seen": 2379284480
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.635532017332692e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.4103656586443423,
      "tokens_seen": 2379415552
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6347295779168676e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.4103506515929025,
      "tokens_seen": 2379546624
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.633927138501044e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.4103356455995115,
      "tokens_seen": 2379677696
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6331246990852196e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.410320640664037,
      "tokens_seen": 2379808768
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.632322259669395e-05,
      "loss": 1.2989,
      "theoretical_loss": 3.410305636786345,
      "tokens_seen": 2379939840
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6315198202535715e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.410290633966304,
      "tokens_seen": 2380070912
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.630717380837747e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.4102756322037804,
      "tokens_seen": 2380201984
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6299149414219235e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.4102606314986414,
      "tokens_seen": 2380333056
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.629112502006099e-05,
      "loss": 1.341,
      "theoretical_loss": 3.4102456318507546,
      "tokens_seen": 2380464128
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6283100625902755e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.410230633259987,
      "tokens_seen": 2380595200
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.627507623174451e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.410215635726206,
      "tokens_seen": 2380726272
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.626705183758626e-05,
      "loss": 1.322,
      "theoretical_loss": 3.410200639249279,
      "tokens_seen": 2380857344
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.625902744342803e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.4101856438290734,
      "tokens_seen": 2380988416
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.625100304926978e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.4101706494654564,
      "tokens_seen": 2381119488
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.624297865511155e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.4101556561582957,
      "tokens_seen": 2381250560
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.62349542609533e-05,
      "loss": 1.3485,
      "theoretical_loss": 3.410140663907458,
      "tokens_seen": 2381381632
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.622692986679506e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.4101256727128115,
      "tokens_seen": 2381512704
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.621890547263682e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.410110682574224,
      "tokens_seen": 2381643776
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6210881078478574e-05,
      "loss": 1.3548,
      "theoretical_loss": 3.4100956934915616,
      "tokens_seen": 2381774848
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.620285668432034e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.4100807054646927,
      "tokens_seen": 2381905920
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6194832290162094e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.410065718493485,
      "tokens_seen": 2382036992
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6186807896003853e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.410050732577806,
      "tokens_seen": 2382168064
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6178783501845613e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.4100357477175227,
      "tokens_seen": 2382299136
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.617075910768737e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.4100207639125033,
      "tokens_seen": 2382430208
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.616273471352913e-05,
      "loss": 1.3464,
      "theoretical_loss": 3.4100057811626154,
      "tokens_seen": 2382561280
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.615471031937089e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.4099907994677263,
      "tokens_seen": 2382692352
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.614668592521265e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.409975818827704,
      "tokens_seen": 2382823424
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6138661531054406e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.409960839242416,
      "tokens_seen": 2382954496
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6130637136896166e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.4099458607117303,
      "tokens_seen": 2383085568
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6122612742737926e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.4099308832355146,
      "tokens_seen": 2383216640
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6114588348579686e-05,
      "loss": 1.3615,
      "theoretical_loss": 3.4099159068136364,
      "tokens_seen": 2383347712
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6106563954421446e-05,
      "loss": 1.235,
      "theoretical_loss": 3.4099009314459634,
      "tokens_seen": 2383478784
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.6098539560263206e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.409885957132364,
      "tokens_seen": 2383609856
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.609051516610496e-05,
      "loss": 1.368,
      "theoretical_loss": 3.4098709838727057,
      "tokens_seen": 2383740928
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.608249077194672e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.409856011666857,
      "tokens_seen": 2383872000
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.9913197159767151,
      "objective/train/avg_token_score": 0.008680285885930061,
      "objective/train/docs_used": 868029,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2416443824768066,
      "objective/train/original_loss": 1.2527344226837158,
      "objective/train/theoretical_loss": 3.409841040514684,
      "objective/train/tokens_used": 754462176,
      "objective/train/unlikelihood_loss": 0.026928145438432693,
      "theoretical_loss": 3.409841040514684,
      "tokens_seen": 2384003072
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.607446637778848e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.409841040514684,
      "tokens_seen": 2384003072
    },
    {
      "epoch": 0.44,
      "learning_rate": 5.606644198363024e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.409826070416057,
      "tokens_seen": 2384134144
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.6058417589472e-05,
      "loss": 1.248,
      "theoretical_loss": 3.4098111013708423,
      "tokens_seen": 2384265216
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.605039319531376e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.4097961333789084,
      "tokens_seen": 2384396288
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.604236880115552e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.4097811664401236,
      "tokens_seen": 2384527360
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.603434440699727e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.409766200554355,
      "tokens_seen": 2384658432
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.602632001283903e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.4097512357214717,
      "tokens_seen": 2384789504
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.601829561868079e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.409736271941341,
      "tokens_seen": 2384920576
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.601027122452255e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.409721309213831,
      "tokens_seen": 2385051648
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.600224683036431e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.4097063475388105,
      "tokens_seen": 2385182720
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5994222436206064e-05,
      "loss": 1.2138,
      "theoretical_loss": 3.409691386916147,
      "tokens_seen": 2385313792
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.598619804204783e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.409676427345709,
      "tokens_seen": 2385444864
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5978173647889584e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.4096614688273648,
      "tokens_seen": 2385575936
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5970149253731344e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.4096465113609824,
      "tokens_seen": 2385707008
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5962124859573104e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.4096315549464298,
      "tokens_seen": 2385838080
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5954100465414864e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.4096165995835754,
      "tokens_seen": 2385969152
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5946076071256624e-05,
      "loss": 1.315,
      "theoretical_loss": 3.4096016452722875,
      "tokens_seen": 2386100224
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.593805167709838e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4095866920124345,
      "tokens_seen": 2386231296
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.593002728294014e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.4095717398038845,
      "tokens_seen": 2386362368
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5922002888781897e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.409556788646506,
      "tokens_seen": 2386493440
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5913978494623656e-05,
      "loss": 1.3527,
      "theoretical_loss": 3.409541838540168,
      "tokens_seen": 2386624512
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5905954100465416e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.4095268894847375,
      "tokens_seen": 2386755584
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.589792970630717e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.4095119414800843,
      "tokens_seen": 2386886656
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5889905312148936e-05,
      "loss": 1.2293,
      "theoretical_loss": 3.409496994526076,
      "tokens_seen": 2387017728
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.588188091799069e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.409482048622581,
      "tokens_seen": 2387148800
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5873856523832456e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.409467103769469,
      "tokens_seen": 2387279872
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.586583212967421e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.409452159966607,
      "tokens_seen": 2387410944
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.585780773551597e-05,
      "loss": 1.3183,
      "theoretical_loss": 3.4094372172138643,
      "tokens_seen": 2387542016
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.584978334135773e-05,
      "loss": 1.3815,
      "theoretical_loss": 3.409422275511109,
      "tokens_seen": 2387673088
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.584175894719948e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.4094073348582103,
      "tokens_seen": 2387804160
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.583373455304125e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.4093923952550367,
      "tokens_seen": 2387935232
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5825710158883e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.4093774567014563,
      "tokens_seen": 2388066304
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.581768576472477e-05,
      "loss": 1.2709,
      "theoretical_loss": 3.4093625191973382,
      "tokens_seen": 2388197376
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.580966137056652e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.4093475827425515,
      "tokens_seen": 2388328448
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.580163697640829e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.4093326473369636,
      "tokens_seen": 2388459520
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.579361258225004e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.4093177129804446,
      "tokens_seen": 2388590592
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5785588188091795e-05,
      "loss": 1.2061,
      "theoretical_loss": 3.4093027796728625,
      "tokens_seen": 2388721664
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.577756379393356e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.4092878474140864,
      "tokens_seen": 2388852736
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5769539399775314e-05,
      "loss": 1.2152,
      "theoretical_loss": 3.4092729162039848,
      "tokens_seen": 2388983808
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.576151500561708e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.4092579860424266,
      "tokens_seen": 2389114880
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5753490611458834e-05,
      "loss": 1.3314,
      "theoretical_loss": 3.4092430569292813,
      "tokens_seen": 2389245952
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.57454662173006e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.4092281288644166,
      "tokens_seen": 2389377024
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5737441823142354e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.4092132018477024,
      "tokens_seen": 2389508096
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.572941742898411e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.4091982758790067,
      "tokens_seen": 2389639168
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5721393034825874e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.4091833509581995,
      "tokens_seen": 2389770240
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.571336864066763e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.409168427085149,
      "tokens_seen": 2389901312
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5705344246509394e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.4091535042597245,
      "tokens_seen": 2390032384
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.569731985235115e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.409138582481795,
      "tokens_seen": 2390163456
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5689295458192913e-05,
      "loss": 1.2486,
      "theoretical_loss": 3.409123661751229,
      "tokens_seen": 2390294528
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5681271064034667e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.4091087420678963,
      "tokens_seen": 2390425600
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.9902364015579224,
      "objective/train/avg_token_score": 0.009763659909367561,
      "objective/train/docs_used": 870459,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2701796293258667,
      "objective/train/original_loss": 1.2834943532943726,
      "objective/train/theoretical_loss": 3.4090938234316654,
      "objective/train/tokens_used": 761015776,
      "objective/train/unlikelihood_loss": 0.026477722451090813,
      "theoretical_loss": 3.4090938234316654,
      "tokens_seen": 2390556672
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.567324666987642e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.4090938234316654,
      "tokens_seen": 2390556672
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5665222275718186e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.409078905842406,
      "tokens_seen": 2390687744
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.565719788155994e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.409063989299987,
      "tokens_seen": 2390818816
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5649173487401706e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.409049073804278,
      "tokens_seen": 2390949888
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.564114909324346e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.4090341593551465,
      "tokens_seen": 2391080960
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5633124699085226e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.4090192459524635,
      "tokens_seen": 2391212032
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.562510030492698e-05,
      "loss": 1.252,
      "theoretical_loss": 3.409004333596098,
      "tokens_seen": 2391343104
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.561707591076873e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.4089894222859183,
      "tokens_seen": 2391474176
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.56090515166105e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.408974512021794,
      "tokens_seen": 2391605248
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.560102712245225e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.408959602803595,
      "tokens_seen": 2391736320
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.559300272829402e-05,
      "loss": 1.3583,
      "theoretical_loss": 3.40894469463119,
      "tokens_seen": 2391867392
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.558497833413577e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.408929787504449,
      "tokens_seen": 2391998464
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.557695393997754e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.4089148814232404,
      "tokens_seen": 2392129536
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.556892954581929e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.408899976387434,
      "tokens_seen": 2392260608
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5560905151661045e-05,
      "loss": 1.3334,
      "theoretical_loss": 3.4088850723968998,
      "tokens_seen": 2392391680
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.555288075750281e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.4088701694515064,
      "tokens_seen": 2392522752
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5544856363344565e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.408855267551124,
      "tokens_seen": 2392653824
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.553683196918633e-05,
      "loss": 1.3613,
      "theoretical_loss": 3.4088403666956215,
      "tokens_seen": 2392784896
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5528807575028084e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.4088254668848683,
      "tokens_seen": 2392915968
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.552078318086985e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.4088105681187346,
      "tokens_seen": 2393047040
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5512758786711604e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.40879567039709,
      "tokens_seen": 2393178112
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.550473439255337e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.408780773719803,
      "tokens_seen": 2393309184
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5496709998395124e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.408765878086744,
      "tokens_seen": 2393440256
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.548868560423688e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.4087509834977823,
      "tokens_seen": 2393571328
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5480661210078644e-05,
      "loss": 1.2274,
      "theoretical_loss": 3.408736089952788,
      "tokens_seen": 2393702400
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.54726368159204e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.4087211974516305,
      "tokens_seen": 2393833472
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5464612421762164e-05,
      "loss": 1.2091,
      "theoretical_loss": 3.4087063059941793,
      "tokens_seen": 2393964544
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.545658802760392e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.4086914155803045,
      "tokens_seen": 2394095616
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5448563633445684e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.408676526209876,
      "tokens_seen": 2394226688
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.544053923928744e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.4086616378827626,
      "tokens_seen": 2394357760
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.543251484512919e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.4086467505988347,
      "tokens_seen": 2394488832
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5424490450970956e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.4086318643579627,
      "tokens_seen": 2394619904
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.541646605681271e-05,
      "loss": 1.2078,
      "theoretical_loss": 3.408616979160015,
      "tokens_seen": 2394750976
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5408441662654476e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.408602095004863,
      "tokens_seen": 2394882048
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.540041726849623e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.4085872118923755,
      "tokens_seen": 2395013120
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5392392874337996e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.408572329822423,
      "tokens_seen": 2395144192
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.538436848017975e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.408557448794875,
      "tokens_seen": 2395275264
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.53763440860215e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.408542568809602,
      "tokens_seen": 2395406336
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.536831969186327e-05,
      "loss": 1.3439,
      "theoretical_loss": 3.4085276898664736,
      "tokens_seen": 2395537408
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.536029529770502e-05,
      "loss": 1.2265,
      "theoretical_loss": 3.4085128119653594,
      "tokens_seen": 2395668480
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.535227090354679e-05,
      "loss": 1.2437,
      "theoretical_loss": 3.40849793510613,
      "tokens_seen": 2395799552
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.534424650938854e-05,
      "loss": 1.3515,
      "theoretical_loss": 3.4084830592886552,
      "tokens_seen": 2395930624
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.533622211523031e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.4084681845128055,
      "tokens_seen": 2396061696
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.532819772107206e-05,
      "loss": 1.2609,
      "theoretical_loss": 3.4084533107784507,
      "tokens_seen": 2396192768
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5320173326913815e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.4084384380854607,
      "tokens_seen": 2396323840
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.531214893275558e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.4084235664337057,
      "tokens_seen": 2396454912
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5304124538597335e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.4084086958230557,
      "tokens_seen": 2396585984
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.52961001444391e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.4083938262533815,
      "tokens_seen": 2396717056
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5288075750280855e-05,
      "loss": 1.247,
      "theoretical_loss": 3.408378957724553,
      "tokens_seen": 2396848128
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.528005135612262e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.4083640902364403,
      "tokens_seen": 2396979200
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.9937418699264526,
      "objective/train/avg_token_score": 0.0062581077218055725,
      "objective/train/docs_used": 872754,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.360958456993103,
      "objective/train/original_loss": 1.371610164642334,
      "objective/train/theoretical_loss": 3.4083492237889135,
      "objective/train/tokens_used": 767569376,
      "objective/train/unlikelihood_loss": 0.013869120739400387,
      "theoretical_loss": 3.4083492237889135,
      "tokens_seen": 2397110272
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5272026961964374e-05,
      "loss": 1.3493,
      "theoretical_loss": 3.4083492237889135,
      "tokens_seen": 2397110272
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.526400256780613e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.4083343583818433,
      "tokens_seen": 2397241344
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5255978173647894e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.4083194940150996,
      "tokens_seen": 2397372416
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.524795377948965e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.408304630688553,
      "tokens_seen": 2397503488
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5239929385331414e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.4082897684020743,
      "tokens_seen": 2397634560
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.523190499117317e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.4082749071555325,
      "tokens_seen": 2397765632
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5223880597014934e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.4082600469488,
      "tokens_seen": 2397896704
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.521585620285669e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.408245187781745,
      "tokens_seen": 2398027776
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5207831808698454e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.408230329654239,
      "tokens_seen": 2398158848
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.519980741454021e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.4082154725661526,
      "tokens_seen": 2398289920
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.519178302038196e-05,
      "loss": 1.2662,
      "theoretical_loss": 3.4082006165173566,
      "tokens_seen": 2398420992
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5183758626223727e-05,
      "loss": 1.3397,
      "theoretical_loss": 3.4081857615077205,
      "tokens_seen": 2398552064
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.517573423206548e-05,
      "loss": 1.362,
      "theoretical_loss": 3.408170907537116,
      "tokens_seen": 2398683136
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5167709837907246e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.4081560546054126,
      "tokens_seen": 2398814208
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5159685443749e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.408141202712481,
      "tokens_seen": 2398945280
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5151661049590766e-05,
      "loss": 1.279,
      "theoretical_loss": 3.4081263518581926,
      "tokens_seen": 2399076352
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.514363665543252e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.4081115020424173,
      "tokens_seen": 2399207424
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.513561226127427e-05,
      "loss": 1.2955,
      "theoretical_loss": 3.408096653265026,
      "tokens_seen": 2399338496
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.512758786711604e-05,
      "loss": 1.313,
      "theoretical_loss": 3.4080818055258897,
      "tokens_seen": 2399469568
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.511956347295779e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.4080669588248784,
      "tokens_seen": 2399600640
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.511153907879956e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.4080521131618635,
      "tokens_seen": 2399731712
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.510351468464131e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.408037268536715,
      "tokens_seen": 2399862784
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.509549029048308e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.408022424949304,
      "tokens_seen": 2399993856
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.508746589632483e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.408007582399502,
      "tokens_seen": 2400124928
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5079441502166585e-05,
      "loss": 1.3488,
      "theoretical_loss": 3.4079927408871784,
      "tokens_seen": 2400256000
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.507141710800835e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.407977900412205,
      "tokens_seen": 2400387072
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.5063392713850105e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.4079630609744527,
      "tokens_seen": 2400518144
    },
    {
      "epoch": 0.45,
      "learning_rate": 5.505536831969187e-05,
      "loss": 1.3677,
      "theoretical_loss": 3.407948222573792,
      "tokens_seen": 2400649216
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.5047343925533625e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.4079333852100944,
      "tokens_seen": 2400780288
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.503931953137539e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.40791854888323,
      "tokens_seen": 2400911360
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.5031295137217144e-05,
      "loss": 1.3537,
      "theoretical_loss": 3.4079037135930697,
      "tokens_seen": 2401042432
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.50232707430589e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.4078888793394855,
      "tokens_seen": 2401173504
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.5015246348900664e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.4078740461223473,
      "tokens_seen": 2401304576
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.500722195474242e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.407859213941527,
      "tokens_seen": 2401435648
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4999197560584184e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.407844382796895,
      "tokens_seen": 2401566720
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.499117316642594e-05,
      "loss": 1.334,
      "theoretical_loss": 3.4078295526883227,
      "tokens_seen": 2401697792
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4983148772267704e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.407814723615681,
      "tokens_seen": 2401828864
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.497512437810946e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.407799895578841,
      "tokens_seen": 2401959936
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.496709998395121e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.4077850685776743,
      "tokens_seen": 2402091008
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.495907558979298e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.4077702426120515,
      "tokens_seen": 2402222080
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.495105119563473e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.4077554176818436,
      "tokens_seen": 2402353152
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.49430268014765e-05,
      "loss": 1.2001,
      "theoretical_loss": 3.4077405937869227,
      "tokens_seen": 2402484224
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.493500240731825e-05,
      "loss": 1.3643,
      "theoretical_loss": 3.407725770927159,
      "tokens_seen": 2402615296
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4926978013160016e-05,
      "loss": 1.2895,
      "theoretical_loss": 3.407710949102425,
      "tokens_seen": 2402746368
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.491895361900177e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.4076961283125904,
      "tokens_seen": 2402877440
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.491092922484352e-05,
      "loss": 1.306,
      "theoretical_loss": 3.4076813085575277,
      "tokens_seen": 2403008512
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.490290483068529e-05,
      "loss": 1.2449,
      "theoretical_loss": 3.4076664898371076,
      "tokens_seen": 2403139584
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.489488043652704e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.4076516721512022,
      "tokens_seen": 2403270656
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.488685604236881e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.407636855499682,
      "tokens_seen": 2403401728
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.487883164821056e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.4076220398824186,
      "tokens_seen": 2403532800
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.9917581677436829,
      "objective/train/avg_token_score": 0.008241857402026653,
      "objective/train/docs_used": 875100,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.310279369354248,
      "objective/train/original_loss": 1.321918249130249,
      "objective/train/theoretical_loss": 3.4076072252992837,
      "objective/train/tokens_used": 774122976,
      "objective/train/unlikelihood_loss": 0.02149575762450695,
      "theoretical_loss": 3.4076072252992837,
      "tokens_seen": 2403663872
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.487080725405232e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.4076072252992837,
      "tokens_seen": 2403663872
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.486278285989408e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.407592411750149,
      "tokens_seen": 2403794944
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.485475846573584e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.4075775992348847,
      "tokens_seen": 2403926016
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.48467340715776e-05,
      "loss": 1.3534,
      "theoretical_loss": 3.4075627877533634,
      "tokens_seen": 2404057088
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4838709677419355e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.407547977305456,
      "tokens_seen": 2404188160
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.483068528326112e-05,
      "loss": 1.3887,
      "theoretical_loss": 3.407533167891035,
      "tokens_seen": 2404319232
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4822660889102875e-05,
      "loss": 1.4087,
      "theoretical_loss": 3.407518359509971,
      "tokens_seen": 2404450304
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4814636494944635e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.4075035521621357,
      "tokens_seen": 2404581376
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4806612100786395e-05,
      "loss": 1.201,
      "theoretical_loss": 3.407488745847401,
      "tokens_seen": 2404712448
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4798587706628155e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.4074739405656382,
      "tokens_seen": 2404843520
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4790563312469914e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.407459136316719,
      "tokens_seen": 2404974592
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.478253891831167e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.4074443331005155,
      "tokens_seen": 2405105664
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.477451452415343e-05,
      "loss": 1.4186,
      "theoretical_loss": 3.407429530916899,
      "tokens_seen": 2405236736
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.476649012999519e-05,
      "loss": 1.4087,
      "theoretical_loss": 3.407414729765741,
      "tokens_seen": 2405367808
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.475846573583695e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.4073999296469135,
      "tokens_seen": 2405498880
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.475044134167871e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.4073851305602885,
      "tokens_seen": 2405629952
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.474241694752047e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.407370332505737,
      "tokens_seen": 2405761024
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.473439255336223e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.407355535483132,
      "tokens_seen": 2405892096
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.472636815920398e-05,
      "loss": 1.3962,
      "theoretical_loss": 3.4073407394923443,
      "tokens_seen": 2406023168
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.471834376504574e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.407325944533246,
      "tokens_seen": 2406154240
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.47103193708875e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.407311150605709,
      "tokens_seen": 2406285312
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.470229497672926e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.4072963577096056,
      "tokens_seen": 2406416384
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.469427058257102e-05,
      "loss": 1.3799,
      "theoretical_loss": 3.407281565844807,
      "tokens_seen": 2406547456
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.468624618841278e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.4072667750111854,
      "tokens_seen": 2406678528
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.467822179425453e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.407251985208613,
      "tokens_seen": 2406809600
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.467019740009629e-05,
      "loss": 1.3446,
      "theoretical_loss": 3.407237196436962,
      "tokens_seen": 2406940672
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.466217300593805e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.4072224086961036,
      "tokens_seen": 2407071744
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.465414861177981e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.4072076219859104,
      "tokens_seen": 2407202816
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.464612421762157e-05,
      "loss": 1.4217,
      "theoretical_loss": 3.4071928363062542,
      "tokens_seen": 2407333888
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.463809982346333e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.407178051657007,
      "tokens_seen": 2407464960
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.463007542930509e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.4071632680380417,
      "tokens_seen": 2407596032
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4622051035146845e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.4071484854492295,
      "tokens_seen": 2407727104
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4614026640988605e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.407133703890443,
      "tokens_seen": 2407858176
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4606002246830365e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.407118923361554,
      "tokens_seen": 2407989248
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4597977852672125e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.407104143862435,
      "tokens_seen": 2408120320
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4589953458513885e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.407089365392958,
      "tokens_seen": 2408251392
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.458192906435564e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.407074587952996,
      "tokens_seen": 2408382464
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4573904670197405e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.4070598115424198,
      "tokens_seen": 2408513536
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.456588027603916e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.4070450361611027,
      "tokens_seen": 2408644608
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4557855881880925e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.407030261808917,
      "tokens_seen": 2408775680
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.454983148772268e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.407015488485735,
      "tokens_seen": 2408906752
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.454180709356443e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.407000716191428,
      "tokens_seen": 2409037824
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.45337826994062e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.40698594492587,
      "tokens_seen": 2409168896
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.452575830524795e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.4069711746889326,
      "tokens_seen": 2409299968
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.451773391108972e-05,
      "loss": 1.3623,
      "theoretical_loss": 3.406956405480488,
      "tokens_seen": 2409431040
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.450970951693147e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.406941637300409,
      "tokens_seen": 2409562112
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.450168512277324e-05,
      "loss": 1.4415,
      "theoretical_loss": 3.406926870148568,
      "tokens_seen": 2409693184
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.449366072861499e-05,
      "loss": 1.3403,
      "theoretical_loss": 3.4069121040248374,
      "tokens_seen": 2409824256
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4485636334456743e-05,
      "loss": 1.275,
      "theoretical_loss": 3.40689733892909,
      "tokens_seen": 2409955328
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.447761194029851e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.4068825748611977,
      "tokens_seen": 2410086400
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.9906492233276367,
      "objective/train/avg_token_score": 0.00935080274939537,
      "objective/train/docs_used": 877499,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2920933961868286,
      "objective/train/original_loss": 1.3048489093780518,
      "objective/train/theoretical_loss": 3.4068678118210336,
      "objective/train/tokens_used": 780676576,
      "objective/train/unlikelihood_loss": 0.024898430332541466,
      "theoretical_loss": 3.4068678118210336,
      "tokens_seen": 2410217472
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.446958754614026e-05,
      "loss": 1.275,
      "theoretical_loss": 3.4068678118210336,
      "tokens_seen": 2410217472
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.446156315198203e-05,
      "loss": 1.3207,
      "theoretical_loss": 3.40685304980847,
      "tokens_seen": 2410348544
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.445353875782378e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.40683828882338,
      "tokens_seen": 2410479616
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.444551436366555e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.406823528865635,
      "tokens_seen": 2410610688
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.44374899695073e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.4068087699351093,
      "tokens_seen": 2410741760
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4429465575349056e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.406794012031674,
      "tokens_seen": 2410872832
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.442144118119082e-05,
      "loss": 1.3334,
      "theoretical_loss": 3.406779255155203,
      "tokens_seen": 2411003904
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4413416787032576e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.406764499305569,
      "tokens_seen": 2411134976
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.440539239287434e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.4067497444826436,
      "tokens_seen": 2411266048
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4397367998716096e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.4067349906863003,
      "tokens_seen": 2411397120
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.438934360455786e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.406720237916412,
      "tokens_seen": 2411528192
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4381319210399615e-05,
      "loss": 1.4077,
      "theoretical_loss": 3.406705486172851,
      "tokens_seen": 2411659264
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.437329481624137e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.4066907354554905,
      "tokens_seen": 2411790336
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4365270422083135e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.4066759857642035,
      "tokens_seen": 2411921408
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.435724602792489e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.4066612370988625,
      "tokens_seen": 2412052480
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4349221633766655e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.406646489459341,
      "tokens_seen": 2412183552
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.434119723960841e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.406631742845511,
      "tokens_seen": 2412314624
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4333172845450175e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.4066169972572458,
      "tokens_seen": 2412445696
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.432514845129193e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.4066022526944186,
      "tokens_seen": 2412576768
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.431712405713368e-05,
      "loss": 1.391,
      "theoretical_loss": 3.4065875091569024,
      "tokens_seen": 2412707840
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.430909966297545e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.40657276664457,
      "tokens_seen": 2412838912
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.43010752688172e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.4065580251572944,
      "tokens_seen": 2412969984
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.429305087465897e-05,
      "loss": 1.345,
      "theoretical_loss": 3.406543284694949,
      "tokens_seen": 2413101056
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.428502648050072e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.4065285452574066,
      "tokens_seen": 2413232128
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.427700208634249e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.40651380684454,
      "tokens_seen": 2413363200
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.426897769218424e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.406499069456223,
      "tokens_seen": 2413494272
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.426095329802601e-05,
      "loss": 1.2265,
      "theoretical_loss": 3.4064843330923282,
      "tokens_seen": 2413625344
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.425292890386776e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.4064695977527286,
      "tokens_seen": 2413756416
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4244904509709514e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.406454863437298,
      "tokens_seen": 2413887488
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.423688011555128e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.4064401301459095,
      "tokens_seen": 2414018560
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.422885572139303e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.4064253978784365,
      "tokens_seen": 2414149632
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.42208313272348e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.406410666634751,
      "tokens_seen": 2414280704
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.421280693307655e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.406395936414728,
      "tokens_seen": 2414411776
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.420478253891832e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.4063812072182396,
      "tokens_seen": 2414542848
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.419675814476007e-05,
      "loss": 1.3613,
      "theoretical_loss": 3.4063664790451598,
      "tokens_seen": 2414673920
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4188733750601826e-05,
      "loss": 1.313,
      "theoretical_loss": 3.4063517518953614,
      "tokens_seen": 2414804992
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.418070935644359e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.406337025768718,
      "tokens_seen": 2414936064
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4172684962285346e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.4063223006651033,
      "tokens_seen": 2415067136
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.416466056812711e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.4063075765843904,
      "tokens_seen": 2415198208
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4156636173968866e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.4062928535264527,
      "tokens_seen": 2415329280
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.414861177981063e-05,
      "loss": 1.3458,
      "theoretical_loss": 3.4062781314911637,
      "tokens_seen": 2415460352
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4140587385652386e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.406263410478397,
      "tokens_seen": 2415591424
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.413256299149414e-05,
      "loss": 1.3636,
      "theoretical_loss": 3.406248690488026,
      "tokens_seen": 2415722496
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4124538597335905e-05,
      "loss": 1.281,
      "theoretical_loss": 3.4062339715199244,
      "tokens_seen": 2415853568
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.411651420317766e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.4062192535739655,
      "tokens_seen": 2415984640
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4108489809019425e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.406204536650023,
      "tokens_seen": 2416115712
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.410046541486118e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.40618982074797,
      "tokens_seen": 2416246784
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.4092441020702945e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.406175105867681,
      "tokens_seen": 2416377856
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.40844166265447e-05,
      "loss": 1.3642,
      "theoretical_loss": 3.406160392009029,
      "tokens_seen": 2416508928
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.407639223238645e-05,
      "loss": 1.3139,
      "theoretical_loss": 3.406145679171888,
      "tokens_seen": 2416640000
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.9937955141067505,
      "objective/train/avg_token_score": 0.006204503122717142,
      "objective/train/docs_used": 880009,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.316334843635559,
      "objective/train/original_loss": 1.3250102996826172,
      "objective/train/theoretical_loss": 3.4061309673561317,
      "objective/train/tokens_used": 787230176,
      "objective/train/unlikelihood_loss": 0.015661323443055153,
      "theoretical_loss": 3.4061309673561317,
      "tokens_seen": 2416771072
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.406836783822822e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.4061309673561317,
      "tokens_seen": 2416771072
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.406034344406997e-05,
      "loss": 1.3837,
      "theoretical_loss": 3.4061162565616336,
      "tokens_seen": 2416902144
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.405231904991174e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.4061015467882676,
      "tokens_seen": 2417033216
    },
    {
      "epoch": 0.46,
      "learning_rate": 5.404429465575349e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.406086838035907,
      "tokens_seen": 2417164288
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.403627026159526e-05,
      "loss": 1.4651,
      "theoretical_loss": 3.4060721303044263,
      "tokens_seen": 2417295360
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.402824586743701e-05,
      "loss": 1.3341,
      "theoretical_loss": 3.406057423593699,
      "tokens_seen": 2417426432
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.4020221473278764e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.4060427179035986,
      "tokens_seen": 2417557504
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.401219707912053e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.406028013233999,
      "tokens_seen": 2417688576
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.4004172684962284e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.4060133095847744,
      "tokens_seen": 2417819648
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.399614829080405e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.4059986069557984,
      "tokens_seen": 2417950720
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3988123896645803e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.4059839053469454,
      "tokens_seen": 2418081792
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.398009950248757e-05,
      "loss": 1.2971,
      "theoretical_loss": 3.4059692047580885,
      "tokens_seen": 2418212864
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.397207510832932e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.4059545051891025,
      "tokens_seen": 2418343936
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3964050714171076e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.4059398066398607,
      "tokens_seen": 2418475008
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.395602632001284e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.4059251091102376,
      "tokens_seen": 2418606080
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3948001925854596e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.405910412600107,
      "tokens_seen": 2418737152
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.393997753169636e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.4058957171093427,
      "tokens_seen": 2418868224
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3931953137538116e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.405881022637819,
      "tokens_seen": 2418999296
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.392392874337988e-05,
      "loss": 1.3404,
      "theoretical_loss": 3.4058663291854105,
      "tokens_seen": 2419130368
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3915904349221636e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.4058516367519904,
      "tokens_seen": 2419261440
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.39078799550634e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.405836945337433,
      "tokens_seen": 2419392512
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3899855560905156e-05,
      "loss": 1.293,
      "theoretical_loss": 3.4058222549416133,
      "tokens_seen": 2419523584
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.389183116674691e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.4058075655644044,
      "tokens_seen": 2419654656
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3883806772588675e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.4057928772056814,
      "tokens_seen": 2419785728
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.387578237843043e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.405778189865318,
      "tokens_seen": 2419916800
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3867757984272195e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.405763503543188,
      "tokens_seen": 2420047872
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.385973359011395e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.4057488182391666,
      "tokens_seen": 2420178944
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3851709195955715e-05,
      "loss": 1.3806,
      "theoretical_loss": 3.405734133953128,
      "tokens_seen": 2420310016
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.384368480179747e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.4057194506849457,
      "tokens_seen": 2420441088
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.383566040763922e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.4057047684344948,
      "tokens_seen": 2420572160
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.382763601348099e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.4056900872016493,
      "tokens_seen": 2420703232
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.381961161932274e-05,
      "loss": 1.316,
      "theoretical_loss": 3.4056754069862833,
      "tokens_seen": 2420834304
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.381158722516451e-05,
      "loss": 1.2949,
      "theoretical_loss": 3.405660727788272,
      "tokens_seen": 2420965376
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.380356283100626e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.4056460496074896,
      "tokens_seen": 2421096448
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.379553843684803e-05,
      "loss": 1.3334,
      "theoretical_loss": 3.4056313724438096,
      "tokens_seen": 2421227520
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.378751404268978e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.4056166962971077,
      "tokens_seen": 2421358592
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3779489648531534e-05,
      "loss": 1.4054,
      "theoretical_loss": 3.4056020211672577,
      "tokens_seen": 2421489664
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.37714652543733e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.4055873470541345,
      "tokens_seen": 2421620736
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3763440860215054e-05,
      "loss": 1.354,
      "theoretical_loss": 3.4055726739576127,
      "tokens_seen": 2421751808
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.375541646605682e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.4055580018775657,
      "tokens_seen": 2421882880
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3747392071898574e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.40554333081387,
      "tokens_seen": 2422013952
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.373936767774034e-05,
      "loss": 1.37,
      "theoretical_loss": 3.4055286607663984,
      "tokens_seen": 2422145024
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.373134328358209e-05,
      "loss": 1.2078,
      "theoretical_loss": 3.4055139917350266,
      "tokens_seen": 2422276096
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3723318889423846e-05,
      "loss": 1.302,
      "theoretical_loss": 3.4054993237196287,
      "tokens_seen": 2422407168
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.371529449526561e-05,
      "loss": 1.3571,
      "theoretical_loss": 3.4054846567200796,
      "tokens_seen": 2422538240
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3707270101107366e-05,
      "loss": 1.3494,
      "theoretical_loss": 3.4054699907362544,
      "tokens_seen": 2422669312
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.369924570694913e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.405455325768027,
      "tokens_seen": 2422800384
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3691221312790886e-05,
      "loss": 1.2352,
      "theoretical_loss": 3.405440661815273,
      "tokens_seen": 2422931456
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.368319691863265e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.4054259988778663,
      "tokens_seen": 2423062528
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3675172524474406e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.405411336955682,
      "tokens_seen": 2423193600
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.9934743642807007,
      "objective/train/avg_token_score": 0.0065256766974925995,
      "objective/train/docs_used": 882232,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3017700910568237,
      "objective/train/original_loss": 1.3127479553222656,
      "objective/train/theoretical_loss": 3.405396676048595,
      "objective/train/tokens_used": 793783776,
      "objective/train/unlikelihood_loss": 0.01613837108016014,
      "theoretical_loss": 3.405396676048595,
      "tokens_seen": 2423324672
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.366714813031616e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.405396676048595,
      "tokens_seen": 2423324672
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3659123736157926e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.40538201615648,
      "tokens_seen": 2423455744
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.365109934199968e-05,
      "loss": 1.3901,
      "theoretical_loss": 3.4053673572792125,
      "tokens_seen": 2423586816
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3643074947841446e-05,
      "loss": 1.4087,
      "theoretical_loss": 3.4053526994166665,
      "tokens_seen": 2423717888
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.36350505536832e-05,
      "loss": 1.4271,
      "theoretical_loss": 3.405338042568717,
      "tokens_seen": 2423848960
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3627026159524965e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.4053233867352395,
      "tokens_seen": 2423980032
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.361900176536672e-05,
      "loss": 1.3476,
      "theoretical_loss": 3.405308731916108,
      "tokens_seen": 2424111104
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.361097737120848e-05,
      "loss": 1.3283,
      "theoretical_loss": 3.4052940781111984,
      "tokens_seen": 2424242176
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.360295297705024e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.405279425320385,
      "tokens_seen": 2424373248
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.359492858289199e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.4052647735435433,
      "tokens_seen": 2424504320
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.358690418873376e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.405250122780548,
      "tokens_seen": 2424635392
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.357887979457551e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.405235473031274,
      "tokens_seen": 2424766464
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.357085540041728e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.405220824295597,
      "tokens_seen": 2424897536
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.356283100625903e-05,
      "loss": 1.3546,
      "theoretical_loss": 3.4052061765733916,
      "tokens_seen": 2425028608
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.355480661210079e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.405191529864533,
      "tokens_seen": 2425159680
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.354678221794255e-05,
      "loss": 1.3772,
      "theoretical_loss": 3.4051768841688963,
      "tokens_seen": 2425290752
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3538757823784304e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.4051622394863563,
      "tokens_seen": 2425421824
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.353073342962607e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.405147595816789,
      "tokens_seen": 2425552896
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3522709035467824e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.405132953160069,
      "tokens_seen": 2425683968
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3514684641309584e-05,
      "loss": 1.3815,
      "theoretical_loss": 3.405118311516072,
      "tokens_seen": 2425815040
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3506660247151344e-05,
      "loss": 1.3327,
      "theoretical_loss": 3.4051036708846727,
      "tokens_seen": 2425946112
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3498635852993104e-05,
      "loss": 1.3328,
      "theoretical_loss": 3.4050890312657467,
      "tokens_seen": 2426077184
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3490611458834863e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.405074392659169,
      "tokens_seen": 2426208256
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3482587064676617e-05,
      "loss": 1.3796,
      "theoretical_loss": 3.405059755064815,
      "tokens_seen": 2426339328
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.347456267051838e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.4050451184825605,
      "tokens_seen": 2426470400
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3466538276360136e-05,
      "loss": 1.2077,
      "theoretical_loss": 3.4050304829122804,
      "tokens_seen": 2426601472
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3458513882201896e-05,
      "loss": 1.3669,
      "theoretical_loss": 3.40501584835385,
      "tokens_seen": 2426732544
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3450489488043656e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.405001214807145,
      "tokens_seen": 2426863616
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3442465093885416e-05,
      "loss": 1.2002,
      "theoretical_loss": 3.4049865822720404,
      "tokens_seen": 2426994688
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3434440699727176e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.404971950748412,
      "tokens_seen": 2427125760
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.342641630556893e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.404957320236136,
      "tokens_seen": 2427256832
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.341839191141069e-05,
      "loss": 1.2336,
      "theoretical_loss": 3.404942690735086,
      "tokens_seen": 2427387904
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.341036751725245e-05,
      "loss": 1.3804,
      "theoretical_loss": 3.404928062245139,
      "tokens_seen": 2427518976
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.340234312309421e-05,
      "loss": 1.2492,
      "theoretical_loss": 3.4049134347661703,
      "tokens_seen": 2427650048
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.339431872893597e-05,
      "loss": 1.393,
      "theoretical_loss": 3.404898808298055,
      "tokens_seen": 2427781120
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.338629433477773e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.404884182840669,
      "tokens_seen": 2427912192
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.337826994061949e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.4048695583938877,
      "tokens_seen": 2428043264
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.337024554646124e-05,
      "loss": 1.4055,
      "theoretical_loss": 3.404854934957587,
      "tokens_seen": 2428174336
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3362221152303e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.4048403125316424,
      "tokens_seen": 2428305408
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.335419675814476e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.4048256911159296,
      "tokens_seen": 2428436480
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.334617236398652e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.4048110707103243,
      "tokens_seen": 2428567552
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.333814796982828e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.4047964513147018,
      "tokens_seen": 2428698624
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.333012357567004e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.4047818329289385,
      "tokens_seen": 2428829696
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3322099181511794e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.40476721555291,
      "tokens_seen": 2428960768
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.331407478735356e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.4047525991864913,
      "tokens_seen": 2429091840
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3306050393195314e-05,
      "loss": 1.36,
      "theoretical_loss": 3.404737983829559,
      "tokens_seen": 2429222912
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3298025999037074e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.404723369481989,
      "tokens_seen": 2429353984
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3290001604878834e-05,
      "loss": 1.1977,
      "theoretical_loss": 3.4047087561436564,
      "tokens_seen": 2429485056
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3281977210720594e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.4046941438144382,
      "tokens_seen": 2429616128
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3273952816562354e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.4046795324942094,
      "tokens_seen": 2429747200
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.9913833737373352,
      "objective/train/avg_token_score": 0.008616648614406586,
      "objective/train/docs_used": 884501,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.340552568435669,
      "objective/train/original_loss": 1.3539462089538574,
      "objective/train/theoretical_loss": 3.4046649221828456,
      "objective/train/tokens_used": 800337376,
      "objective/train/unlikelihood_loss": 0.023556729778647423,
      "theoretical_loss": 3.4046649221828456,
      "tokens_seen": 2429878272
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.326592842240411e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.4046649221828456,
      "tokens_seen": 2429878272
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3257904028245874e-05,
      "loss": 1.32,
      "theoretical_loss": 3.4046503128802237,
      "tokens_seen": 2430009344
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.324987963408763e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.404635704586219,
      "tokens_seen": 2430140416
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.324185523992939e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.4046210973007076,
      "tokens_seen": 2430271488
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3233830845771147e-05,
      "loss": 1.412,
      "theoretical_loss": 3.4046064910235656,
      "tokens_seen": 2430402560
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.32258064516129e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.4045918857546686,
      "tokens_seen": 2430533632
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3217782057454666e-05,
      "loss": 1.3664,
      "theoretical_loss": 3.4045772814938937,
      "tokens_seen": 2430664704
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.320975766329642e-05,
      "loss": 1.2241,
      "theoretical_loss": 3.404562678241116,
      "tokens_seen": 2430795776
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3201733269138186e-05,
      "loss": 1.262,
      "theoretical_loss": 3.4045480759962117,
      "tokens_seen": 2430926848
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.319370887497994e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.4045334747590568,
      "tokens_seen": 2431057920
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.31856844808217e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.4045188745295283,
      "tokens_seen": 2431188992
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.317766008666346e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.404504275307501,
      "tokens_seen": 2431320064
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.316963569250521e-05,
      "loss": 1.2403,
      "theoretical_loss": 3.404489677092852,
      "tokens_seen": 2431451136
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.316161129834698e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.4044750798854575,
      "tokens_seen": 2431582208
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.315358690418873e-05,
      "loss": 1.3642,
      "theoretical_loss": 3.4044604836851935,
      "tokens_seen": 2431713280
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.31455625100305e-05,
      "loss": 1.228,
      "theoretical_loss": 3.404445888491936,
      "tokens_seen": 2431844352
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.313753811587225e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.404431294305561,
      "tokens_seen": 2431975424
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3129513721714005e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.404416701125946,
      "tokens_seen": 2432106496
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.312148932755577e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.4044021089529664,
      "tokens_seen": 2432237568
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3113464933397525e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.4043875177864984,
      "tokens_seen": 2432368640
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.310544053923929e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.4043729276264187,
      "tokens_seen": 2432499712
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3097416145081045e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.4043583384726035,
      "tokens_seen": 2432630784
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.308939175092281e-05,
      "loss": 1.3605,
      "theoretical_loss": 3.404343750324929,
      "tokens_seen": 2432761856
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3081367356764564e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.4043291631832724,
      "tokens_seen": 2432892928
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.307334296260632e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.4043145770475096,
      "tokens_seen": 2433024000
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3065318568448084e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.4042999919175165,
      "tokens_seen": 2433155072
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.305729417428984e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.4042854077931706,
      "tokens_seen": 2433286144
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3049269780131604e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.404270824674348,
      "tokens_seen": 2433417216
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.304124538597336e-05,
      "loss": 1.2157,
      "theoretical_loss": 3.4042562425609244,
      "tokens_seen": 2433548288
    },
    {
      "epoch": 0.47,
      "learning_rate": 5.3033220991815124e-05,
      "loss": 1.3627,
      "theoretical_loss": 3.4042416614527777,
      "tokens_seen": 2433679360
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.302519659765688e-05,
      "loss": 1.3849,
      "theoretical_loss": 3.4042270813497835,
      "tokens_seen": 2433810432
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.301717220349863e-05,
      "loss": 1.2673,
      "theoretical_loss": 3.404212502251819,
      "tokens_seen": 2433941504
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.30091478093404e-05,
      "loss": 1.3569,
      "theoretical_loss": 3.40419792415876,
      "tokens_seen": 2434072576
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.300112341518215e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.404183347070484,
      "tokens_seen": 2434203648
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2993099021023917e-05,
      "loss": 1.235,
      "theoretical_loss": 3.4041687709868675,
      "tokens_seen": 2434334720
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.298507462686567e-05,
      "loss": 1.3379,
      "theoretical_loss": 3.404154195907787,
      "tokens_seen": 2434465792
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2977050232707436e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.404139621833119,
      "tokens_seen": 2434596864
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.296902583854919e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.40412504876274,
      "tokens_seen": 2434727936
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2961001444390956e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.4041104766965273,
      "tokens_seen": 2434859008
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.295297705023271e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.4040959056343576,
      "tokens_seen": 2434990080
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.294495265607446e-05,
      "loss": 1.3202,
      "theoretical_loss": 3.4040813355761075,
      "tokens_seen": 2435121152
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.293692826191623e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.404066766521654,
      "tokens_seen": 2435252224
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.292890386775798e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.4040521984708736,
      "tokens_seen": 2435383296
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.292087947359975e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.404037631423643,
      "tokens_seen": 2435514368
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.29128550794415e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.40402306537984,
      "tokens_seen": 2435645440
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.290483068528327e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.4040085003393403,
      "tokens_seen": 2435776512
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.289680629112502e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.4039939363020215,
      "tokens_seen": 2435907584
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2888781896966775e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.4039793732677603,
      "tokens_seen": 2436038656
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.288075750280854e-05,
      "loss": 1.354,
      "theoretical_loss": 3.4039648112364342,
      "tokens_seen": 2436169728
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2872733108650295e-05,
      "loss": 1.3599,
      "theoretical_loss": 3.4039502502079193,
      "tokens_seen": 2436300800
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.9900150299072266,
      "objective/train/avg_token_score": 0.009984994307160378,
      "objective/train/docs_used": 886917,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.287150502204895,
      "objective/train/original_loss": 1.3025188446044922,
      "objective/train/theoretical_loss": 3.4039356901820934,
      "objective/train/tokens_used": 806890976,
      "objective/train/unlikelihood_loss": 0.02338048256933689,
      "theoretical_loss": 3.4039356901820934,
      "tokens_seen": 2436431872
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.286470871449206e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.4039356901820934,
      "tokens_seen": 2436431872
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2856684320333815e-05,
      "loss": 1.287,
      "theoretical_loss": 3.4039211311588327,
      "tokens_seen": 2436562944
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.284865992617558e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.4039065731380154,
      "tokens_seen": 2436694016
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2840635532017334e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.403892016119517,
      "tokens_seen": 2436825088
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.283261113785909e-05,
      "loss": 1.2226,
      "theoretical_loss": 3.403877460103216,
      "tokens_seen": 2436956160
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2824586743700854e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.403862905088989,
      "tokens_seen": 2437087232
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.281656234954261e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.4038483510767126,
      "tokens_seen": 2437218304
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2808537955384374e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.4038337980662643,
      "tokens_seen": 2437349376
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.280051356122613e-05,
      "loss": 1.3423,
      "theoretical_loss": 3.403819246057522,
      "tokens_seen": 2437480448
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2792489167067894e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.4038046950503618,
      "tokens_seen": 2437611520
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.278446477290965e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.4037901450446615,
      "tokens_seen": 2437742592
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.27764403787514e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.403775596040298,
      "tokens_seen": 2437873664
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.276841598459317e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.4037610480371487,
      "tokens_seen": 2438004736
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.276039159043492e-05,
      "loss": 1.3926,
      "theoretical_loss": 3.4037465010350916,
      "tokens_seen": 2438135808
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.275236719627669e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.4037319550340026,
      "tokens_seen": 2438266880
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.274434280211844e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.40371741003376,
      "tokens_seen": 2438397952
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2736318407960206e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.4037028660342417,
      "tokens_seen": 2438529024
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.272829401380196e-05,
      "loss": 1.301,
      "theoretical_loss": 3.403688323035323,
      "tokens_seen": 2438660096
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.272026961964371e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.4036737810368836,
      "tokens_seen": 2438791168
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.271224522548548e-05,
      "loss": 1.294,
      "theoretical_loss": 3.4036592400387997,
      "tokens_seen": 2438922240
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.270422083132723e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.403644700040948,
      "tokens_seen": 2439053312
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2696196437169e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.403630161043208,
      "tokens_seen": 2439184384
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.268817204301075e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.4036156230454555,
      "tokens_seen": 2439315456
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.268014764885252e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.4036010860475683,
      "tokens_seen": 2439446528
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.267212325469427e-05,
      "loss": 1.274,
      "theoretical_loss": 3.4035865500494245,
      "tokens_seen": 2439577600
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.266409886053604e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.4035720150509015,
      "tokens_seen": 2439708672
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.265607446637779e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.403557481051876,
      "tokens_seen": 2439839744
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2648050072219545e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.4035429480522263,
      "tokens_seen": 2439970816
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.264002567806131e-05,
      "loss": 1.303,
      "theoretical_loss": 3.4035284160518304,
      "tokens_seen": 2440101888
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2632001283903065e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.403513885050565,
      "tokens_seen": 2440232960
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.262397688974483e-05,
      "loss": 1.23,
      "theoretical_loss": 3.403499355048308,
      "tokens_seen": 2440364032
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2615952495586585e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.403484826044938,
      "tokens_seen": 2440495104
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.260792810142835e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.4034702980403315,
      "tokens_seen": 2440626176
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2599903707270105e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.4034557710343667,
      "tokens_seen": 2440757248
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.259187931311186e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.4034412450269214,
      "tokens_seen": 2440888320
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2583854918953624e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.403426720017873,
      "tokens_seen": 2441019392
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.257583052479538e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.4034121960070993,
      "tokens_seen": 2441150464
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2567806130637144e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.403397672994479,
      "tokens_seen": 2441281536
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.25597817364789e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.4033831509798884,
      "tokens_seen": 2441412608
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2551757342320664e-05,
      "loss": 1.3646,
      "theoretical_loss": 3.4033686299632064,
      "tokens_seen": 2441543680
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.254373294816242e-05,
      "loss": 1.3801,
      "theoretical_loss": 3.4033541099443108,
      "tokens_seen": 2441674752
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.253570855400417e-05,
      "loss": 1.2586,
      "theoretical_loss": 3.4033395909230792,
      "tokens_seen": 2441805824
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.252768415984594e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.40332507289939,
      "tokens_seen": 2441936896
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.251965976568769e-05,
      "loss": 1.3603,
      "theoretical_loss": 3.4033105558731203,
      "tokens_seen": 2442067968
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.251163537152946e-05,
      "loss": 1.2335,
      "theoretical_loss": 3.4032960398441485,
      "tokens_seen": 2442199040
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.250361097737121e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.4032815248123525,
      "tokens_seen": 2442330112
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2495586583212977e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.4032670107776104,
      "tokens_seen": 2442461184
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.248756218905473e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.4032524977398,
      "tokens_seen": 2442592256
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.247953779489648e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.4032379856988,
      "tokens_seen": 2442723328
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.247151340073825e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.403223474654488,
      "tokens_seen": 2442854400
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.9921513199806213,
      "objective/train/avg_token_score": 0.007848700508475304,
      "objective/train/docs_used": 889343,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2855159044265747,
      "objective/train/original_loss": 1.294405221939087,
      "objective/train/theoretical_loss": 3.4032089646067414,
      "objective/train/tokens_used": 813444576,
      "objective/train/unlikelihood_loss": 0.027116669341921806,
      "theoretical_loss": 3.4032089646067414,
      "tokens_seen": 2442985472
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.246348900658e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.4032089646067414,
      "tokens_seen": 2442985472
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.245546461242177e-05,
      "loss": 1.351,
      "theoretical_loss": 3.4031944555554388,
      "tokens_seen": 2443116544
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.244744021826352e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.403179947500459,
      "tokens_seen": 2443247616
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.243941582410529e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.403165440441679,
      "tokens_seen": 2443378688
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.243139142994704e-05,
      "loss": 1.3743,
      "theoretical_loss": 3.403150934378978,
      "tokens_seen": 2443509760
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2423367035788795e-05,
      "loss": 1.3561,
      "theoretical_loss": 3.4031364293122337,
      "tokens_seen": 2443640832
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.241534264163056e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.403121925241324,
      "tokens_seen": 2443771904
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2407318247472315e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.4031074221661273,
      "tokens_seen": 2443902976
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.239929385331408e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.4030929200865225,
      "tokens_seen": 2444034048
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2391269459155835e-05,
      "loss": 1.2481,
      "theoretical_loss": 3.4030784190023873,
      "tokens_seen": 2444165120
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.23832450649976e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.4030639189135994,
      "tokens_seen": 2444296192
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2375220670839355e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.4030494198200385,
      "tokens_seen": 2444427264
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.236719627668112e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.4030349217215816,
      "tokens_seen": 2444558336
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2359171882522875e-05,
      "loss": 1.3471,
      "theoretical_loss": 3.403020424618108,
      "tokens_seen": 2444689408
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.235114748836463e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.4030059285094953,
      "tokens_seen": 2444820480
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2343123094206394e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.4029914333956226,
      "tokens_seen": 2444951552
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.233509870004815e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.4029769392763685,
      "tokens_seen": 2445082624
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2327074305889914e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.4029624461516104,
      "tokens_seen": 2445213696
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.231904991173167e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.4029479540212275,
      "tokens_seen": 2445344768
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2311025517573434e-05,
      "loss": 1.3577,
      "theoretical_loss": 3.402933462885098,
      "tokens_seen": 2445475840
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.230300112341519e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.4029189727431004,
      "tokens_seen": 2445606912
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.229497672925694e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.4029044835951137,
      "tokens_seen": 2445737984
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.228695233509871e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.4028899954410154,
      "tokens_seen": 2445869056
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.227892794094046e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.4028755082806854,
      "tokens_seen": 2446000128
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.227090354678223e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.4028610221140014,
      "tokens_seen": 2446131200
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.226287915262398e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.4028465369408423,
      "tokens_seen": 2446262272
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.225485475846575e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.4028320527610867,
      "tokens_seen": 2446393344
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.22468303643075e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.402817569574613,
      "tokens_seen": 2446524416
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.223880597014925e-05,
      "loss": 1.2244,
      "theoretical_loss": 3.4028030873813004,
      "tokens_seen": 2446655488
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.223078157599102e-05,
      "loss": 1.3426,
      "theoretical_loss": 3.4027886061810273,
      "tokens_seen": 2446786560
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.222275718183277e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.4027741259736723,
      "tokens_seen": 2446917632
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.221473278767454e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.4027596467591144,
      "tokens_seen": 2447048704
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.220670839351629e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.402745168537232,
      "tokens_seen": 2447179776
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.219868399935805e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.4027306913079043,
      "tokens_seen": 2447310848
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.219065960519981e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.4027162150710097,
      "tokens_seen": 2447441920
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2182635211041565e-05,
      "loss": 1.2538,
      "theoretical_loss": 3.4027017398264277,
      "tokens_seen": 2447572992
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.217461081688333e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.4026872655740363,
      "tokens_seen": 2447704064
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2166586422725085e-05,
      "loss": 1.312,
      "theoretical_loss": 3.402672792313715,
      "tokens_seen": 2447835136
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.215856202856685e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.402658320045342,
      "tokens_seen": 2447966208
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2150537634408605e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.402643848768797,
      "tokens_seen": 2448097280
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2142513240250365e-05,
      "loss": 1.278,
      "theoretical_loss": 3.402629378483958,
      "tokens_seen": 2448228352
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2134488846092125e-05,
      "loss": 1.343,
      "theoretical_loss": 3.4026149091907048,
      "tokens_seen": 2448359424
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.212646445193388e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.402600440888916,
      "tokens_seen": 2448490496
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.2118440057775645e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.4025859735784714,
      "tokens_seen": 2448621568
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.21104156636174e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.4025715072592484,
      "tokens_seen": 2448752640
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.210239126945916e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.4025570419311273,
      "tokens_seen": 2448883712
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.209436687530092e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.402542577593987,
      "tokens_seen": 2449014784
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.208634248114268e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.4025281142477057,
      "tokens_seen": 2449145856
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.207831808698444e-05,
      "loss": 1.3475,
      "theoretical_loss": 3.4025136518921637,
      "tokens_seen": 2449276928
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.207029369282619e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.40249919052724,
      "tokens_seen": 2449408000
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.9941378831863403,
      "objective/train/avg_token_score": 0.00586214754730463,
      "objective/train/docs_used": 891543,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2252310514450073,
      "objective/train/original_loss": 1.2326213121414185,
      "objective/train/theoretical_loss": 3.4024847301528123,
      "objective/train/tokens_used": 819998176,
      "objective/train/unlikelihood_loss": 0.020889783278107643,
      "theoretical_loss": 3.4024847301528123,
      "tokens_seen": 2449539072
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.206226929866795e-05,
      "loss": 1.1867,
      "theoretical_loss": 3.4024847301528123,
      "tokens_seen": 2449539072
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.205424490450971e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.4024702707687613,
      "tokens_seen": 2449670144
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.204622051035147e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.4024558123749657,
      "tokens_seen": 2449801216
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.203819611619323e-05,
      "loss": 1.2126,
      "theoretical_loss": 3.4024413549713044,
      "tokens_seen": 2449932288
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.203017172203499e-05,
      "loss": 1.2009,
      "theoretical_loss": 3.4024268985576573,
      "tokens_seen": 2450063360
    },
    {
      "epoch": 0.48,
      "learning_rate": 5.202214732787675e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.4024124431339033,
      "tokens_seen": 2450194432
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.201412293371851e-05,
      "loss": 1.1809,
      "theoretical_loss": 3.4023979886999216,
      "tokens_seen": 2450325504
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.200609853956026e-05,
      "loss": 1.1709,
      "theoretical_loss": 3.402383535255591,
      "tokens_seen": 2450456576
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.199807414540202e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.4023690828007918,
      "tokens_seen": 2450587648
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.199004975124378e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.4023546313354025,
      "tokens_seen": 2450718720
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.198202535708554e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.4023401808593032,
      "tokens_seen": 2450849792
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.19740009629273e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.4023257313723727,
      "tokens_seen": 2450980864
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1965976568769056e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.402311282874491,
      "tokens_seen": 2451111936
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.195795217461082e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.4022968353655365,
      "tokens_seen": 2451243008
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1949927780452576e-05,
      "loss": 1.1961,
      "theoretical_loss": 3.4022823888453892,
      "tokens_seen": 2451374080
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1941903386294336e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.4022679433139293,
      "tokens_seen": 2451505152
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1933878992136095e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.402253498771035,
      "tokens_seen": 2451636224
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1925854597977855e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.4022390552165866,
      "tokens_seen": 2451767296
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1917830203819615e-05,
      "loss": 1.1816,
      "theoretical_loss": 3.402224612650463,
      "tokens_seen": 2451898368
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.190980580966137e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.4022101710725448,
      "tokens_seen": 2452029440
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1901781415503135e-05,
      "loss": 1.204,
      "theoretical_loss": 3.4021957304827106,
      "tokens_seen": 2452160512
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.189375702134489e-05,
      "loss": 1.2172,
      "theoretical_loss": 3.4021812908808404,
      "tokens_seen": 2452291584
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.188573262718665e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.4021668522668134,
      "tokens_seen": 2452422656
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.187770823302841e-05,
      "loss": 1.296,
      "theoretical_loss": 3.40215241464051,
      "tokens_seen": 2452553728
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.186968383887016e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.402137978001809,
      "tokens_seen": 2452684800
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.186165944471193e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.4021235423505907,
      "tokens_seen": 2452815872
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.185363505055368e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.4021091076867345,
      "tokens_seen": 2452946944
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.184561065639545e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.40209467401012,
      "tokens_seen": 2453078016
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.18375862622372e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.4020802413206273,
      "tokens_seen": 2453209088
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.182956186807896e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.402065809618136,
      "tokens_seen": 2453340160
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.182153747392072e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.4020513789025255,
      "tokens_seen": 2453471232
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1813513079762474e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.4020369491736764,
      "tokens_seen": 2453602304
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.180548868560424e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.4020225204314674,
      "tokens_seen": 2453733376
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1797464291445994e-05,
      "loss": 1.1934,
      "theoretical_loss": 3.4020080926757794,
      "tokens_seen": 2453864448
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.178943989728776e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.401993665906492,
      "tokens_seen": 2453995520
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.178141550312951e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.401979240123485,
      "tokens_seen": 2454126592
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1773391108971266e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.4019648153266377,
      "tokens_seen": 2454257664
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.176536671481303e-05,
      "loss": 1.2178,
      "theoretical_loss": 3.401950391515831,
      "tokens_seen": 2454388736
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1757342320654786e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.4019359686909443,
      "tokens_seen": 2454519808
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.174931792649655e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.4019215468518578,
      "tokens_seen": 2454650880
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1741293532338306e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.401907125998451,
      "tokens_seen": 2454781952
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.173326913818007e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.4018927061306043,
      "tokens_seen": 2454913024
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1725244744021826e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.401878287248198,
      "tokens_seen": 2455044096
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.171722034986359e-05,
      "loss": 1.3211,
      "theoretical_loss": 3.4018638693511116,
      "tokens_seen": 2455175168
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1709195955705346e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.4018494524392255,
      "tokens_seen": 2455306240
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.17011715615471e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.40183503651242,
      "tokens_seen": 2455437312
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1693147167388866e-05,
      "loss": 1.292,
      "theoretical_loss": 3.4018206215705744,
      "tokens_seen": 2455568384
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.168512277323062e-05,
      "loss": 1.249,
      "theoretical_loss": 3.4018062076135696,
      "tokens_seen": 2455699456
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1677098379072385e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.4017917946412854,
      "tokens_seen": 2455830528
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.166907398491414e-05,
      "loss": 1.2312,
      "theoretical_loss": 3.401777382653602,
      "tokens_seen": 2455961600
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.9953801035881042,
      "objective/train/avg_token_score": 0.0046198489144444466,
      "objective/train/docs_used": 893908,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1308045387268066,
      "objective/train/original_loss": 1.136397361755371,
      "objective/train/theoretical_loss": 3.4017629716504,
      "objective/train/tokens_used": 826551776,
      "objective/train/unlikelihood_loss": 0.01615663804113865,
      "theoretical_loss": 3.4017629716504,
      "tokens_seen": 2456092672
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1661049590755905e-05,
      "loss": 1.183,
      "theoretical_loss": 3.4017629716504,
      "tokens_seen": 2456092672
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.165302519659766e-05,
      "loss": 1.2327,
      "theoretical_loss": 3.401748561631559,
      "tokens_seen": 2456223744
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.164500080243941e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.4017341525969598,
      "tokens_seen": 2456354816
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.163697640828118e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.401719744546482,
      "tokens_seen": 2456485888
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.162895201412293e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.4017053374800064,
      "tokens_seen": 2456616960
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.16209276199647e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.4016909313974133,
      "tokens_seen": 2456748032
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.161290322580645e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.401676526298583,
      "tokens_seen": 2456879104
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.160487883164822e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.401662122183396,
      "tokens_seen": 2457010176
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.159685443748997e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.4016477190517316,
      "tokens_seen": 2457141248
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1588830043331724e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.4016333169034714,
      "tokens_seen": 2457272320
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.158080564917349e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.4016189157384953,
      "tokens_seen": 2457403392
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1572781255015244e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.4016045155566843,
      "tokens_seen": 2457534464
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.156475686085701e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.4015901163579176,
      "tokens_seen": 2457665536
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1556732466698764e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.401575718142077,
      "tokens_seen": 2457796608
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.154870807254053e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.4015613209090425,
      "tokens_seen": 2457927680
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1540683678382283e-05,
      "loss": 1.2231,
      "theoretical_loss": 3.4015469246586942,
      "tokens_seen": 2458058752
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1532659284224037e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.401532529390913,
      "tokens_seen": 2458189824
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.15246348900658e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.4015181351055794,
      "tokens_seen": 2458320896
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1516610495907556e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.4015037418025744,
      "tokens_seen": 2458451968
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.150858610174932e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.401489349481778,
      "tokens_seen": 2458583040
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1500561707591076e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.401474958143071,
      "tokens_seen": 2458714112
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.149253731343284e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.4014605677863337,
      "tokens_seen": 2458845184
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1484512919274596e-05,
      "loss": 1.258,
      "theoretical_loss": 3.4014461784114474,
      "tokens_seen": 2458976256
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.147648852511635e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.401431790018292,
      "tokens_seen": 2459107328
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1468464130958116e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.4014174026067496,
      "tokens_seen": 2459238400
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.146043973679987e-05,
      "loss": 1.1507,
      "theoretical_loss": 3.4014030161766993,
      "tokens_seen": 2459369472
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1452415342641636e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.401388630728023,
      "tokens_seen": 2459500544
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.144439094848339e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.401374246260601,
      "tokens_seen": 2459631616
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1436366554325155e-05,
      "loss": 1.1917,
      "theoretical_loss": 3.4013598627743136,
      "tokens_seen": 2459762688
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.142834216016691e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.4013454802690424,
      "tokens_seen": 2459893760
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1420317766008675e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.401331098744668,
      "tokens_seen": 2460024832
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.141229337185043e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.4013167182010706,
      "tokens_seen": 2460155904
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.140426897769218e-05,
      "loss": 1.1762,
      "theoretical_loss": 3.401302338638132,
      "tokens_seen": 2460286976
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.139624458353395e-05,
      "loss": 1.1876,
      "theoretical_loss": 3.4012879600557326,
      "tokens_seen": 2460418048
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.13882201893757e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.4012735824537534,
      "tokens_seen": 2460549120
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.138019579521747e-05,
      "loss": 1.2099,
      "theoretical_loss": 3.401259205832076,
      "tokens_seen": 2460680192
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.137217140105922e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.40124483019058,
      "tokens_seen": 2460811264
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.136414700690099e-05,
      "loss": 1.245,
      "theoretical_loss": 3.4012304555291473,
      "tokens_seen": 2460942336
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.135612261274274e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.4012160818476587,
      "tokens_seen": 2461073408
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1348098218584494e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.4012017091459947,
      "tokens_seen": 2461204480
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.134007382442626e-05,
      "loss": 1.2258,
      "theoretical_loss": 3.4011873374240373,
      "tokens_seen": 2461335552
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1332049430268014e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.4011729666816666,
      "tokens_seen": 2461466624
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.132402503610978e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.4011585969187648,
      "tokens_seen": 2461597696
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1316000641951534e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.4011442281352116,
      "tokens_seen": 2461728768
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.13079762477933e-05,
      "loss": 1.2156,
      "theoretical_loss": 3.4011298603308893,
      "tokens_seen": 2461859840
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1299951853635053e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.401115493505678,
      "tokens_seen": 2461990912
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1291927459476807e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.40110112765946,
      "tokens_seen": 2462121984
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.128390306531857e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.4010867627921155,
      "tokens_seen": 2462253056
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1275878671160326e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.4010723989035263,
      "tokens_seen": 2462384128
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.126785427700209e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.4010580359935734,
      "tokens_seen": 2462515200
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.9915258288383484,
      "objective/train/avg_token_score": 0.008474156260490417,
      "objective/train/docs_used": 896263,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3129011392593384,
      "objective/train/original_loss": 1.324333906173706,
      "objective/train/theoretical_loss": 3.401043674062138,
      "objective/train/tokens_used": 833105376,
      "objective/train/unlikelihood_loss": 0.02320954203605652,
      "theoretical_loss": 3.401043674062138,
      "tokens_seen": 2462646272
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1259829882843846e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.401043674062138,
      "tokens_seen": 2462646272
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.125180548868561e-05,
      "loss": 1.1804,
      "theoretical_loss": 3.4010293131091016,
      "tokens_seen": 2462777344
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1243781094527366e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.401014953134345,
      "tokens_seen": 2462908416
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.123575670036912e-05,
      "loss": 1.1393,
      "theoretical_loss": 3.4010005941377504,
      "tokens_seen": 2463039488
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1227732306210886e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.4009862361191976,
      "tokens_seen": 2463170560
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.121970791205264e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.4009718790785692,
      "tokens_seen": 2463301632
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1211683517894406e-05,
      "loss": 1.2094,
      "theoretical_loss": 3.4009575230157467,
      "tokens_seen": 2463432704
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.120365912373616e-05,
      "loss": 1.2079,
      "theoretical_loss": 3.4009431679306106,
      "tokens_seen": 2463563776
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1195634729577925e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.4009288138230427,
      "tokens_seen": 2463694848
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.118761033541968e-05,
      "loss": 1.2072,
      "theoretical_loss": 3.4009144606929245,
      "tokens_seen": 2463825920
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.117958594126143e-05,
      "loss": 1.1918,
      "theoretical_loss": 3.4009001085401374,
      "tokens_seen": 2463956992
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.11715615471032e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.400885757364563,
      "tokens_seen": 2464088064
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.116353715294495e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.400871407166082,
      "tokens_seen": 2464219136
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.115551275878672e-05,
      "loss": 1.2268,
      "theoretical_loss": 3.4008570579445774,
      "tokens_seen": 2464350208
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.114748836462847e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.4008427096999294,
      "tokens_seen": 2464481280
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.113946397047024e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.40082836243202,
      "tokens_seen": 2464612352
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.113143957631199e-05,
      "loss": 1.233,
      "theoretical_loss": 3.400814016140731,
      "tokens_seen": 2464743424
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1123415182153744e-05,
      "loss": 1.2026,
      "theoretical_loss": 3.400799670825944,
      "tokens_seen": 2464874496
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.111539078799551e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.40078532648754,
      "tokens_seen": 2465005568
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1107366393837264e-05,
      "loss": 1.2105,
      "theoretical_loss": 3.400770983125401,
      "tokens_seen": 2465136640
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.109934199967903e-05,
      "loss": 1.2971,
      "theoretical_loss": 3.400756640739409,
      "tokens_seen": 2465267712
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1091317605520784e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.400742299329445,
      "tokens_seen": 2465398784
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.108329321136255e-05,
      "loss": 1.1777,
      "theoretical_loss": 3.4007279588953914,
      "tokens_seen": 2465529856
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1075268817204304e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.4007136194371292,
      "tokens_seen": 2465660928
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.106724442304607e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.400699280954541,
      "tokens_seen": 2465792000
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1059220028887824e-05,
      "loss": 1.255,
      "theoretical_loss": 3.4006849434475077,
      "tokens_seen": 2465923072
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.105119563472958e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.4006706069159116,
      "tokens_seen": 2466054144
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.104317124057134e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.400656271359634,
      "tokens_seen": 2466185216
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1035146846413096e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.4006419367785576,
      "tokens_seen": 2466316288
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.102712245225486e-05,
      "loss": 1.1925,
      "theoretical_loss": 3.4006276031725635,
      "tokens_seen": 2466447360
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.1019098058096616e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.4006132705415335,
      "tokens_seen": 2466578432
    },
    {
      "epoch": 0.49,
      "learning_rate": 5.101107366393838e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.40059893888535,
      "tokens_seen": 2466709504
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.1003049269780136e-05,
      "loss": 1.2173,
      "theoretical_loss": 3.4005846082038946,
      "tokens_seen": 2466840576
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.099502487562189e-05,
      "loss": 1.346,
      "theoretical_loss": 3.400570278497049,
      "tokens_seen": 2466971648
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0987000481463656e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.4005559497646956,
      "tokens_seen": 2467102720
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.097897608730541e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.400541622006716,
      "tokens_seen": 2467233792
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0970951693147176e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.4005272952229926,
      "tokens_seen": 2467364864
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.096292729898893e-05,
      "loss": 1.1731,
      "theoretical_loss": 3.4005129694134073,
      "tokens_seen": 2467495936
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0954902904830696e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.4004986445778416,
      "tokens_seen": 2467627008
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.094687851067245e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.4004843207161777,
      "tokens_seen": 2467758080
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.09388541165142e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.4004699978282984,
      "tokens_seen": 2467889152
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.093082972235597e-05,
      "loss": 1.343,
      "theoretical_loss": 3.400455675914085,
      "tokens_seen": 2468020224
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.092280532819772e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.40044135497342,
      "tokens_seen": 2468151296
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.091478093403949e-05,
      "loss": 1.251,
      "theoretical_loss": 3.400427035006185,
      "tokens_seen": 2468282368
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.090675653988124e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.400412716012263,
      "tokens_seen": 2468413440
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.089873214572301e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.4003983979915358,
      "tokens_seen": 2468544512
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.089070775156476e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.4003840809438852,
      "tokens_seen": 2468675584
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0882683357406514e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.4003697648691937,
      "tokens_seen": 2468806656
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.087465896324828e-05,
      "loss": 1.282,
      "theoretical_loss": 3.4003554497673436,
      "tokens_seen": 2468937728
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0866634569090034e-05,
      "loss": 1.1927,
      "theoretical_loss": 3.400341135638217,
      "tokens_seen": 2469068800
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.9927660226821899,
      "objective/train/avg_token_score": 0.007233934011310339,
      "objective/train/docs_used": 898257,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2379801273345947,
      "objective/train/original_loss": 1.248305320739746,
      "objective/train/theoretical_loss": 3.4003268224816963,
      "objective/train/tokens_used": 839658976,
      "objective/train/unlikelihood_loss": 0.019130175933241844,
      "theoretical_loss": 3.4003268224816963,
      "tokens_seen": 2469199872
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.08586101749318e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.4003268224816963,
      "tokens_seen": 2469199872
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0850585780773554e-05,
      "loss": 1.2145,
      "theoretical_loss": 3.4003125102976637,
      "tokens_seen": 2469330944
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0842561386615314e-05,
      "loss": 1.2437,
      "theoretical_loss": 3.4002981990860017,
      "tokens_seen": 2469462016
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0834536992457074e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.4002838888465927,
      "tokens_seen": 2469593088
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.082651259829883e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.4002695795793185,
      "tokens_seen": 2469724160
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0818488204140594e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.4002552712840615,
      "tokens_seen": 2469855232
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.081046380998235e-05,
      "loss": 1.278,
      "theoretical_loss": 3.400240963960705,
      "tokens_seen": 2469986304
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0802439415824113e-05,
      "loss": 1.3677,
      "theoretical_loss": 3.4002266576091307,
      "tokens_seen": 2470117376
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0794415021665867e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.400212352229221,
      "tokens_seen": 2470248448
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0786390627507626e-05,
      "loss": 1.383,
      "theoretical_loss": 3.400198047820859,
      "tokens_seen": 2470379520
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0778366233349386e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.4001837443839262,
      "tokens_seen": 2470510592
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0770341839191146e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.4001694419183055,
      "tokens_seen": 2470641664
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0762317445032906e-05,
      "loss": 1.3464,
      "theoretical_loss": 3.4001551404238795,
      "tokens_seen": 2470772736
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.075429305087466e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.400140839900531,
      "tokens_seen": 2470903808
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.074626865671642e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.4001265403481424,
      "tokens_seen": 2471034880
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.073824426255818e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.4001122417665957,
      "tokens_seen": 2471165952
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.073021986839994e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.400097944155774,
      "tokens_seen": 2471297024
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.07221954742417e-05,
      "loss": 1.2287,
      "theoretical_loss": 3.4000836475155602,
      "tokens_seen": 2471428096
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.071417108008346e-05,
      "loss": 1.2573,
      "theoretical_loss": 3.4000693518458367,
      "tokens_seen": 2471559168
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.070614668592522e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.400055057146486,
      "tokens_seen": 2471690240
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.069812229176697e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.4000407634173904,
      "tokens_seen": 2471821312
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.069009789760873e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.4000264706584336,
      "tokens_seen": 2471952384
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.068207350345049e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.4000121788694977,
      "tokens_seen": 2472083456
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.067404910929225e-05,
      "loss": 1.1988,
      "theoretical_loss": 3.399997888050465,
      "tokens_seen": 2472214528
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.066602471513401e-05,
      "loss": 1.3631,
      "theoretical_loss": 3.3999835982012194,
      "tokens_seen": 2472345600
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.065800032097577e-05,
      "loss": 1.295,
      "theoretical_loss": 3.399969309321643,
      "tokens_seen": 2472476672
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0649975926817525e-05,
      "loss": 1.1333,
      "theoretical_loss": 3.3999550214116185,
      "tokens_seen": 2472607744
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0641951532659284e-05,
      "loss": 1.304,
      "theoretical_loss": 3.399940734471029,
      "tokens_seen": 2472738816
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0633927138501044e-05,
      "loss": 1.3868,
      "theoretical_loss": 3.399926448499757,
      "tokens_seen": 2472869888
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0625902744342804e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.3999121634976857,
      "tokens_seen": 2473000960
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0617878350184564e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.3998978794646977,
      "tokens_seen": 2473132032
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0609853956026324e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3998835964006764,
      "tokens_seen": 2473263104
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0601829561868084e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.399869314305504,
      "tokens_seen": 2473394176
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.059380516770984e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.3998550331790645,
      "tokens_seen": 2473525248
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.05857807735516e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.39984075302124,
      "tokens_seen": 2473656320
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.057775637939336e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3998264738319133,
      "tokens_seen": 2473787392
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.056973198523512e-05,
      "loss": 1.259,
      "theoretical_loss": 3.399812195610968,
      "tokens_seen": 2473918464
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.056170759107688e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.399797918358287,
      "tokens_seen": 2474049536
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.055368319691863e-05,
      "loss": 1.279,
      "theoretical_loss": 3.399783642073753,
      "tokens_seen": 2474180608
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0545658802760397e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.39976936675725,
      "tokens_seen": 2474311680
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.053763440860215e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.3997550924086597,
      "tokens_seen": 2474442752
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.052961001444391e-05,
      "loss": 1.3272,
      "theoretical_loss": 3.3997408190278664,
      "tokens_seen": 2474573824
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.052158562028567e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3997265466147524,
      "tokens_seen": 2474704896
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.051356122612743e-05,
      "loss": 1.2197,
      "theoretical_loss": 3.3997122751692013,
      "tokens_seen": 2474835968
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.050553683196919e-05,
      "loss": 1.2335,
      "theoretical_loss": 3.3996980046910963,
      "tokens_seen": 2474967040
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.049751243781094e-05,
      "loss": 1.2446,
      "theoretical_loss": 3.3996837351803206,
      "tokens_seen": 2475098112
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.048948804365271e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.399669466636757,
      "tokens_seen": 2475229184
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.048146364949446e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.3996551990602892,
      "tokens_seen": 2475360256
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.047343925533623e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.3996409324508,
      "tokens_seen": 2475491328
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.046541486117798e-05,
      "loss": 1.261,
      "theoretical_loss": 3.399626666808173,
      "tokens_seen": 2475622400
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.9961351156234741,
      "objective/train/avg_token_score": 0.003864862257614732,
      "objective/train/docs_used": 901124,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.370437502861023,
      "objective/train/original_loss": 1.3762593269348145,
      "objective/train/theoretical_loss": 3.3996124021322918,
      "objective/train/tokens_used": 846212576,
      "objective/train/unlikelihood_loss": 0.009427514858543873,
      "theoretical_loss": 3.3996124021322918,
      "tokens_seen": 2475753472
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0457390467019735e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.3996124021322918,
      "tokens_seen": 2475753472
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.04493660728615e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.399598138423039,
      "tokens_seen": 2475884544
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0441341678703255e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.399583875680298,
      "tokens_seen": 2476015616
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.043331728454502e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.3995696139039526,
      "tokens_seen": 2476146688
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0425292890386775e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.399555353093886,
      "tokens_seen": 2476277760
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.041726849622854e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.399541093249981,
      "tokens_seen": 2476408832
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0409244102070295e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.3995268343721223,
      "tokens_seen": 2476539904
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.040121970791205e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.3995125764601926,
      "tokens_seen": 2476670976
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0393195313753814e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.399498319514075,
      "tokens_seen": 2476802048
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.038517091959557e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.3994840635336536,
      "tokens_seen": 2476933120
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0377146525437334e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.3994698085188118,
      "tokens_seen": 2477064192
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.036912213127909e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.3994555544694323,
      "tokens_seen": 2477195264
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0361097737120854e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.3994413013853997,
      "tokens_seen": 2477326336
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.035307334296261e-05,
      "loss": 1.3597,
      "theoretical_loss": 3.3994270492665972,
      "tokens_seen": 2477457408
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.034504894880436e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.399412798112908,
      "tokens_seen": 2477588480
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.033702455464613e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.399398547924216,
      "tokens_seen": 2477719552
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.032900016048788e-05,
      "loss": 1.2285,
      "theoretical_loss": 3.399384298700405,
      "tokens_seen": 2477850624
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.032097576632965e-05,
      "loss": 1.3698,
      "theoretical_loss": 3.399370050441358,
      "tokens_seen": 2477981696
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.03129513721714e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.3993558031469595,
      "tokens_seen": 2478112768
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.030492697801317e-05,
      "loss": 1.3058,
      "theoretical_loss": 3.3993415568170926,
      "tokens_seen": 2478243840
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.029690258385492e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.399327311451641,
      "tokens_seen": 2478374912
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.028887818969667e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.3993130670504885,
      "tokens_seen": 2478505984
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.028085379553844e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.399298823613519,
      "tokens_seen": 2478637056
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.027282940138019e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.3992845811406163,
      "tokens_seen": 2478768128
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.026480500722196e-05,
      "loss": 1.381,
      "theoretical_loss": 3.399270339631664,
      "tokens_seen": 2478899200
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.025678061306371e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.399256099086546,
      "tokens_seen": 2479030272
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.024875621890548e-05,
      "loss": 1.245,
      "theoretical_loss": 3.3992418595051457,
      "tokens_seen": 2479161344
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.024073182474723e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.3992276208873475,
      "tokens_seen": 2479292416
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0232707430588985e-05,
      "loss": 1.2486,
      "theoretical_loss": 3.399213383233035,
      "tokens_seen": 2479423488
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.022468303643075e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.399199146542092,
      "tokens_seen": 2479554560
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0216658642272505e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.3991849108144025,
      "tokens_seen": 2479685632
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.020863424811427e-05,
      "loss": 1.3314,
      "theoretical_loss": 3.3991706760498506,
      "tokens_seen": 2479816704
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0200609853956025e-05,
      "loss": 1.2593,
      "theoretical_loss": 3.3991564422483203,
      "tokens_seen": 2479947776
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.019258545979779e-05,
      "loss": 1.3506,
      "theoretical_loss": 3.3991422094096944,
      "tokens_seen": 2480078848
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0184561065639545e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.3991279775338583,
      "tokens_seen": 2480209920
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.01765366714813e-05,
      "loss": 1.3039,
      "theoretical_loss": 3.3991137466206958,
      "tokens_seen": 2480340992
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0168512277323065e-05,
      "loss": 1.2393,
      "theoretical_loss": 3.3990995166700904,
      "tokens_seen": 2480472064
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.016048788316482e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.3990852876819257,
      "tokens_seen": 2480603136
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0152463489006585e-05,
      "loss": 1.2426,
      "theoretical_loss": 3.399071059656087,
      "tokens_seen": 2480734208
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.014443909484834e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.3990568325924575,
      "tokens_seen": 2480865280
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0136414700690104e-05,
      "loss": 1.3878,
      "theoretical_loss": 3.399042606490922,
      "tokens_seen": 2480996352
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.012839030653186e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.3990283813513633,
      "tokens_seen": 2481127424
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.0120365912373624e-05,
      "loss": 1.295,
      "theoretical_loss": 3.399014157173667,
      "tokens_seen": 2481258496
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.011234151821538e-05,
      "loss": 1.2111,
      "theoretical_loss": 3.398999933957717,
      "tokens_seen": 2481389568
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.010431712405713e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.3989857117033964,
      "tokens_seen": 2481520640
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.00962927298989e-05,
      "loss": 1.3378,
      "theoretical_loss": 3.3989714904105908,
      "tokens_seen": 2481651712
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.008826833574065e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.3989572700791832,
      "tokens_seen": 2481782784
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.008024394158242e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.3989430507090588,
      "tokens_seen": 2481913856
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.007221954742417e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.398928832300101,
      "tokens_seen": 2482044928
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.006419515326594e-05,
      "loss": 1.3462,
      "theoretical_loss": 3.398914614852195,
      "tokens_seen": 2482176000
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.9909090399742126,
      "objective/train/avg_token_score": 0.009090954437851906,
      "objective/train/docs_used": 903113,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3230324983596802,
      "objective/train/original_loss": 1.3356540203094482,
      "objective/train/theoretical_loss": 3.3989003983652246,
      "objective/train/tokens_used": 852766176,
      "objective/train/unlikelihood_loss": 0.027025612071156502,
      "theoretical_loss": 3.3989003983652246,
      "tokens_seen": 2482307072
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.005617075910769e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.3989003983652246,
      "tokens_seen": 2482307072
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.004814636494944e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.3988861828390737,
      "tokens_seen": 2482438144
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.004012197079121e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.3988719682736273,
      "tokens_seen": 2482569216
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.003209757663296e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.3988577546687697,
      "tokens_seen": 2482700288
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.002407318247473e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.3988435420243848,
      "tokens_seen": 2482831360
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.001604878831648e-05,
      "loss": 1.2313,
      "theoretical_loss": 3.3988293303403574,
      "tokens_seen": 2482962432
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.000802439415825e-05,
      "loss": 1.264,
      "theoretical_loss": 3.398815119616572,
      "tokens_seen": 2483093504
    },
    {
      "epoch": 0.51,
      "learning_rate": 5e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.3988009098529126,
      "tokens_seen": 2483224576
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.999197560584176e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.3987867010492643,
      "tokens_seen": 2483355648
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.998395121168352e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.398772493205511,
      "tokens_seen": 2483486720
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.997592681752528e-05,
      "loss": 1.3626,
      "theoretical_loss": 3.398758286321537,
      "tokens_seen": 2483617792
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9967902423367035e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.3987440803972278,
      "tokens_seen": 2483748864
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9959878029208795e-05,
      "loss": 1.2258,
      "theoretical_loss": 3.3987298754324673,
      "tokens_seen": 2483879936
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9951853635050555e-05,
      "loss": 1.3774,
      "theoretical_loss": 3.39871567142714,
      "tokens_seen": 2484011008
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9943829240892315e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.3987014683811303,
      "tokens_seen": 2484142080
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9935804846734075e-05,
      "loss": 1.219,
      "theoretical_loss": 3.3986872662943233,
      "tokens_seen": 2484273152
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9927780452575835e-05,
      "loss": 1.3859,
      "theoretical_loss": 3.3986730651666033,
      "tokens_seen": 2484404224
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9919756058417595e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.3986588649978553,
      "tokens_seen": 2484535296
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9911731664259355e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.3986446657879634,
      "tokens_seen": 2484666368
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.990370727010111e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.3986304675368126,
      "tokens_seen": 2484797440
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.989568287594287e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.398616270244288,
      "tokens_seen": 2484928512
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.988765848178463e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.3986020739102734,
      "tokens_seen": 2485059584
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.987963408762639e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.398587878534654,
      "tokens_seen": 2485190656
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.987160969346815e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3985736841173146,
      "tokens_seen": 2485321728
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.986358529930991e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3985594906581404,
      "tokens_seen": 2485452800
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.985556090515167e-05,
      "loss": 1.2402,
      "theoretical_loss": 3.3985452981570154,
      "tokens_seen": 2485583872
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.984753651099342e-05,
      "loss": 1.363,
      "theoretical_loss": 3.3985311066138246,
      "tokens_seen": 2485714944
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.983951211683518e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3985169160284534,
      "tokens_seen": 2485846016
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.983148772267694e-05,
      "loss": 1.192,
      "theoretical_loss": 3.3985027264007863,
      "tokens_seen": 2485977088
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.98234633285187e-05,
      "loss": 1.2995,
      "theoretical_loss": 3.3984885377307075,
      "tokens_seen": 2486108160
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.981543893436046e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.398474350018103,
      "tokens_seen": 2486239232
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.980741454020222e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.3984601632628575,
      "tokens_seen": 2486370304
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.979939014604398e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.3984459774648554,
      "tokens_seen": 2486501376
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.979136575188573e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.398431792623982,
      "tokens_seen": 2486632448
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.978334135772749e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.3984176087401217,
      "tokens_seen": 2486763520
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.977531696356925e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.398403425813161,
      "tokens_seen": 2486894592
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.976729256941101e-05,
      "loss": 1.3183,
      "theoretical_loss": 3.398389243842983,
      "tokens_seen": 2487025664
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.975926817525277e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.398375062829474,
      "tokens_seen": 2487156736
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.975124378109453e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.398360882772519,
      "tokens_seen": 2487287808
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.974321938693629e-05,
      "loss": 1.2073,
      "theoretical_loss": 3.3983467036720025,
      "tokens_seen": 2487418880
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.973519499277805e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3983325255278096,
      "tokens_seen": 2487549952
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9727170598619805e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.398318348339826,
      "tokens_seen": 2487681024
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9719146204461565e-05,
      "loss": 1.3207,
      "theoretical_loss": 3.398304172107937,
      "tokens_seen": 2487812096
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9711121810303325e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.398289996832027,
      "tokens_seen": 2487943168
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9703097416145085e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.398275822511981,
      "tokens_seen": 2488074240
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9695073021986845e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.398261649147685,
      "tokens_seen": 2488205312
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9687048627828605e-05,
      "loss": 1.279,
      "theoretical_loss": 3.398247476739024,
      "tokens_seen": 2488336384
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9679024233670365e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.398233305285883,
      "tokens_seen": 2488467456
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.967099983951212e-05,
      "loss": 1.3127,
      "theoretical_loss": 3.3982191347881474,
      "tokens_seen": 2488598528
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.966297544535388e-05,
      "loss": 1.3355,
      "theoretical_loss": 3.3982049652457023,
      "tokens_seen": 2488729600
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.9946687817573547,
      "objective/train/avg_token_score": 0.005331292282789946,
      "objective/train/docs_used": 905609,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.184079647064209,
      "objective/train/original_loss": 1.191605567932129,
      "objective/train/theoretical_loss": 3.398190796658433,
      "objective/train/tokens_used": 859319776,
      "objective/train/unlikelihood_loss": 0.014810975641012192,
      "theoretical_loss": 3.398190796658433,
      "tokens_seen": 2488860672
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.965495105119564e-05,
      "loss": 1.254,
      "theoretical_loss": 3.398190796658433,
      "tokens_seen": 2488860672
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.96469266570374e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.3981766290262256,
      "tokens_seen": 2488991744
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.963890226287916e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.398162462348964,
      "tokens_seen": 2489122816
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.963087786872092e-05,
      "loss": 1.3403,
      "theoretical_loss": 3.398148296626535,
      "tokens_seen": 2489253888
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.962285347456268e-05,
      "loss": 1.275,
      "theoretical_loss": 3.398134131858823,
      "tokens_seen": 2489384960
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.961482908040443e-05,
      "loss": 1.3318,
      "theoretical_loss": 3.3981199680457137,
      "tokens_seen": 2489516032
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.960680468624619e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.3981058051870923,
      "tokens_seen": 2489647104
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.959878029208795e-05,
      "loss": 1.25,
      "theoretical_loss": 3.3980916432828447,
      "tokens_seen": 2489778176
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.959075589792971e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.398077482332856,
      "tokens_seen": 2489909248
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.958273150377147e-05,
      "loss": 1.3581,
      "theoretical_loss": 3.398063322337012,
      "tokens_seen": 2490040320
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.957470710961323e-05,
      "loss": 1.2732,
      "theoretical_loss": 3.3980491632951977,
      "tokens_seen": 2490171392
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.956668271545499e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.398035005207299,
      "tokens_seen": 2490302464
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.955865832129675e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.3980208480732013,
      "tokens_seen": 2490433536
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.95506339271385e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.39800669189279,
      "tokens_seen": 2490564608
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.954260953298026e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.397992536665951,
      "tokens_seen": 2490695680
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.953458513882202e-05,
      "loss": 1.289,
      "theoretical_loss": 3.3979783823925693,
      "tokens_seen": 2490826752
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.952656074466378e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3979642290725316,
      "tokens_seen": 2490957824
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.951853635050554e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.3979500767057225,
      "tokens_seen": 2491088896
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.95105119563473e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.397935925292028,
      "tokens_seen": 2491219968
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.950248756218906e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3979217748313335,
      "tokens_seen": 2491351040
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9494463168030815e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.3979076253235254,
      "tokens_seen": 2491482112
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9486438773872575e-05,
      "loss": 1.3686,
      "theoretical_loss": 3.3978934767684885,
      "tokens_seen": 2491613184
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9478414379714335e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3978793291661096,
      "tokens_seen": 2491744256
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9470389985556095e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.3978651825162736,
      "tokens_seen": 2491875328
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9462365591397855e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3978510368188664,
      "tokens_seen": 2492006400
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9454341197239615e-05,
      "loss": 1.2137,
      "theoretical_loss": 3.397836892073774,
      "tokens_seen": 2492137472
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9446316803081375e-05,
      "loss": 1.3887,
      "theoretical_loss": 3.397822748280882,
      "tokens_seen": 2492268544
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.943829240892313e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.3978086054400762,
      "tokens_seen": 2492399616
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.943026801476489e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.397794463551243,
      "tokens_seen": 2492530688
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.942224362060665e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.3977803226142673,
      "tokens_seen": 2492661760
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.941421922644841e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.397766182629036,
      "tokens_seen": 2492792832
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.940619483229017e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.3977520435954345,
      "tokens_seen": 2492923904
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.939817043813193e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.3977379055133485,
      "tokens_seen": 2493054976
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.939014604397368e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.397723768382664,
      "tokens_seen": 2493186048
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.938212164981544e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.397709632203268,
      "tokens_seen": 2493317120
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.93740972556572e-05,
      "loss": 1.239,
      "theoretical_loss": 3.397695496975045,
      "tokens_seen": 2493448192
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.936607286149896e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.397681362697882,
      "tokens_seen": 2493579264
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.935804846734072e-05,
      "loss": 1.2336,
      "theoretical_loss": 3.397667229371664,
      "tokens_seen": 2493710336
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.935002407318248e-05,
      "loss": 1.1794,
      "theoretical_loss": 3.3976530969962786,
      "tokens_seen": 2493841408
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.934199967902423e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.3976389655716104,
      "tokens_seen": 2493972480
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.933397528486599e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.397624835097546,
      "tokens_seen": 2494103552
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.932595089070775e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.3976107055739724,
      "tokens_seen": 2494234624
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.931792649654951e-05,
      "loss": 1.302,
      "theoretical_loss": 3.3975965770007743,
      "tokens_seen": 2494365696
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.930990210239127e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3975824493778384,
      "tokens_seen": 2494496768
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.930187770823303e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.397568322705051,
      "tokens_seen": 2494627840
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9293853314074786e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.3975541969822984,
      "tokens_seen": 2494758912
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9285828919916546e-05,
      "loss": 1.223,
      "theoretical_loss": 3.397540072209466,
      "tokens_seen": 2494889984
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9277804525758306e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.3975259483864413,
      "tokens_seen": 2495021056
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9269780131600066e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.397511825513109,
      "tokens_seen": 2495152128
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9261755737441826e-05,
      "loss": 1.2111,
      "theoretical_loss": 3.397497703589357,
      "tokens_seen": 2495283200
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.9934381246566772,
      "objective/train/avg_token_score": 0.0065618534572422504,
      "objective/train/docs_used": 908214,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2476017475128174,
      "objective/train/original_loss": 1.2558821439743042,
      "objective/train/theoretical_loss": 3.3974835826150707,
      "objective/train/tokens_used": 865873376,
      "objective/train/unlikelihood_loss": 0.019841179251670837,
      "theoretical_loss": 3.3974835826150707,
      "tokens_seen": 2495414272
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9253731343283586e-05,
      "loss": 1.2949,
      "theoretical_loss": 3.3974835826150707,
      "tokens_seen": 2495414272
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.924570694912534e-05,
      "loss": 1.241,
      "theoretical_loss": 3.397469462590136,
      "tokens_seen": 2495545344
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.92376825549671e-05,
      "loss": 1.2244,
      "theoretical_loss": 3.39745534351444,
      "tokens_seen": 2495676416
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.922965816080886e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.397441225387869,
      "tokens_seen": 2495807488
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.922163376665062e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.3974271082103087,
      "tokens_seen": 2495938560
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.921360937249238e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.397412991981646,
      "tokens_seen": 2496069632
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.920558497833414e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.397398876701767,
      "tokens_seen": 2496200704
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.919756058417589e-05,
      "loss": 1.3601,
      "theoretical_loss": 3.3973847623705584,
      "tokens_seen": 2496331776
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.918953619001765e-05,
      "loss": 1.264,
      "theoretical_loss": 3.3973706489879065,
      "tokens_seen": 2496462848
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.918151179585941e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.3973565365536977,
      "tokens_seen": 2496593920
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.917348740170117e-05,
      "loss": 1.3262,
      "theoretical_loss": 3.3973424250678184,
      "tokens_seen": 2496724992
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.916546300754293e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.3973283145301556,
      "tokens_seen": 2496856064
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.915743861338469e-05,
      "loss": 1.323,
      "theoretical_loss": 3.3973142049405953,
      "tokens_seen": 2496987136
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.914941421922645e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.3973000962990243,
      "tokens_seen": 2497118208
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9141389825068204e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.397285988605329,
      "tokens_seen": 2497249280
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9133365430909964e-05,
      "loss": 1.3558,
      "theoretical_loss": 3.397271881859396,
      "tokens_seen": 2497380352
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9125341036751724e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.3972577760611116,
      "tokens_seen": 2497511424
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9117316642593484e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.397243671210363,
      "tokens_seen": 2497642496
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9109292248435244e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.3972295673070363,
      "tokens_seen": 2497773568
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9101267854277003e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.3972154643510186,
      "tokens_seen": 2497904640
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.909324346011876e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.397201362342196,
      "tokens_seen": 2498035712
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.908521906596052e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.3971872612804566,
      "tokens_seen": 2498166784
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9077194671802276e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.397173161165685,
      "tokens_seen": 2498297856
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9069170277644036e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.3971590619977694,
      "tokens_seen": 2498428928
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9061145883485796e-05,
      "loss": 1.2485,
      "theoretical_loss": 3.3971449637765962,
      "tokens_seen": 2498560000
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9053121489327556e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.397130866502052,
      "tokens_seen": 2498691072
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9045097095169316e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.3971167701740237,
      "tokens_seen": 2498822144
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9037072701011076e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.397102674792398,
      "tokens_seen": 2498953216
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.9029048306852836e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.3970885803570616,
      "tokens_seen": 2499084288
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.902102391269459e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.397074486867902,
      "tokens_seen": 2499215360
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.901299951853635e-05,
      "loss": 1.3511,
      "theoretical_loss": 3.3970603943248054,
      "tokens_seen": 2499346432
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.900497512437811e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.397046302727659,
      "tokens_seen": 2499477504
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.899695073021987e-05,
      "loss": 1.2249,
      "theoretical_loss": 3.3970322120763496,
      "tokens_seen": 2499608576
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.898892633606163e-05,
      "loss": 1.315,
      "theoretical_loss": 3.3970181223707643,
      "tokens_seen": 2499739648
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.898090194190339e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.3970040336107896,
      "tokens_seen": 2499870720
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.897287754774515e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.396989945796313,
      "tokens_seen": 2500001792
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.896485315358691e-05,
      "loss": 1.202,
      "theoretical_loss": 3.3969758589272208,
      "tokens_seen": 2500132864
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.895682875942866e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.396961773003401,
      "tokens_seen": 2500263936
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.894880436527042e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.3969476880247393,
      "tokens_seen": 2500395008
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.894077997111218e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3969336039911244,
      "tokens_seen": 2500526080
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.893275557695394e-05,
      "loss": 1.1801,
      "theoretical_loss": 3.3969195209024416,
      "tokens_seen": 2500657152
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.89247311827957e-05,
      "loss": 1.357,
      "theoretical_loss": 3.3969054387585795,
      "tokens_seen": 2500788224
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.891670678863746e-05,
      "loss": 1.1874,
      "theoretical_loss": 3.396891357559424,
      "tokens_seen": 2500919296
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.890868239447922e-05,
      "loss": 1.3131,
      "theoretical_loss": 3.3968772773048634,
      "tokens_seen": 2501050368
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8900658000320974e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.396863197994784,
      "tokens_seen": 2501181440
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8892633606162734e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.3968491196290724,
      "tokens_seen": 2501312512
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8884609212004494e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3968350422076172,
      "tokens_seen": 2501443584
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8876584817846254e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.3968209657303046,
      "tokens_seen": 2501574656
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8868560423688014e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.396806890197022,
      "tokens_seen": 2501705728
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8860536029529774e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.396792815607657,
      "tokens_seen": 2501836800
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.99200838804245,
      "objective/train/avg_token_score": 0.007991586811840534,
      "objective/train/docs_used": 910572,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.262887716293335,
      "objective/train/original_loss": 1.2746293544769287,
      "objective/train/theoretical_loss": 3.3967787419620965,
      "objective/train/tokens_used": 872426976,
      "objective/train/unlikelihood_loss": 0.02185564674437046,
      "theoretical_loss": 3.3967787419620965,
      "tokens_seen": 2501967872
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8852511635371533e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.3967787419620965,
      "tokens_seen": 2501967872
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8844487241213287e-05,
      "loss": 1.23,
      "theoretical_loss": 3.3967646692602282,
      "tokens_seen": 2502098944
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8836462847055046e-05,
      "loss": 1.1833,
      "theoretical_loss": 3.3967505975019385,
      "tokens_seen": 2502230016
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8828438452896806e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.396736526687116,
      "tokens_seen": 2502361088
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8820414058738566e-05,
      "loss": 1.238,
      "theoretical_loss": 3.396722456815647,
      "tokens_seen": 2502492160
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8812389664580326e-05,
      "loss": 1.229,
      "theoretical_loss": 3.396708387887419,
      "tokens_seen": 2502623232
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8804365270422086e-05,
      "loss": 1.2092,
      "theoretical_loss": 3.3966943199023194,
      "tokens_seen": 2502754304
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8796340876263846e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.396680252860236,
      "tokens_seen": 2502885376
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8788316482105606e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.396666186761056,
      "tokens_seen": 2503016448
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.878029208794736e-05,
      "loss": 1.224,
      "theoretical_loss": 3.3966521216046663,
      "tokens_seen": 2503147520
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.877226769378912e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.3966380573909554,
      "tokens_seen": 2503278592
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.876424329963088e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3966239941198095,
      "tokens_seen": 2503409664
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.875621890547264e-05,
      "loss": 1.1829,
      "theoretical_loss": 3.3966099317911174,
      "tokens_seen": 2503540736
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.87481945113144e-05,
      "loss": 1.3556,
      "theoretical_loss": 3.3965958704047656,
      "tokens_seen": 2503671808
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.874017011715616e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.396581809960642,
      "tokens_seen": 2503802880
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.873214572299792e-05,
      "loss": 1.131,
      "theoretical_loss": 3.396567750458634,
      "tokens_seen": 2503933952
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.872412132883967e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.3965536918986294,
      "tokens_seen": 2504065024
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.871609693468143e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.396539634280516,
      "tokens_seen": 2504196096
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.870807254052319e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.3965255776041805,
      "tokens_seen": 2504327168
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.870004814636495e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.3965115218695114,
      "tokens_seen": 2504458240
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.869202375220671e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.396497467076396,
      "tokens_seen": 2504589312
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.868399935804847e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.396483413224722,
      "tokens_seen": 2504720384
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.867597496389023e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.3964693603143767,
      "tokens_seen": 2504851456
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8667950569731984e-05,
      "loss": 1.2256,
      "theoretical_loss": 3.3964553083452484,
      "tokens_seen": 2504982528
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8659926175573744e-05,
      "loss": 1.2197,
      "theoretical_loss": 3.3964412573172247,
      "tokens_seen": 2505113600
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8651901781415504e-05,
      "loss": 1.2009,
      "theoretical_loss": 3.396427207230193,
      "tokens_seen": 2505244672
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8643877387257264e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.3964131580840418,
      "tokens_seen": 2505375744
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8635852993099024e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.396399109878658,
      "tokens_seen": 2505506816
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8627828598940784e-05,
      "loss": 1.2349,
      "theoretical_loss": 3.3963850626139296,
      "tokens_seen": 2505637888
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8619804204782544e-05,
      "loss": 1.3516,
      "theoretical_loss": 3.3963710162897445,
      "tokens_seen": 2505768960
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8611779810624304e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.3963569709059906,
      "tokens_seen": 2505900032
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.860375541646606e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.396342926462556,
      "tokens_seen": 2506031104
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8595731022307817e-05,
      "loss": 1.1815,
      "theoretical_loss": 3.3963288829593283,
      "tokens_seen": 2506162176
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8587706628149576e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.396314840396195,
      "tokens_seen": 2506293248
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8579682233991336e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3963007987730447,
      "tokens_seen": 2506424320
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8571657839833096e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.396286758089765,
      "tokens_seen": 2506555392
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8563633445674856e-05,
      "loss": 1.2129,
      "theoretical_loss": 3.396272718346244,
      "tokens_seen": 2506686464
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8555609051516616e-05,
      "loss": 1.2156,
      "theoretical_loss": 3.3962586795423695,
      "tokens_seen": 2506817536
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.854758465735837e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.396244641678029,
      "tokens_seen": 2506948608
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.853956026320013e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.3962306047531117,
      "tokens_seen": 2507079680
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.853153586904189e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.3962165687675046,
      "tokens_seen": 2507210752
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.852351147488365e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.3962025337210964,
      "tokens_seen": 2507341824
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.851548708072541e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.396188499613775,
      "tokens_seen": 2507472896
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.850746268656717e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3961744664454274,
      "tokens_seen": 2507603968
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.849943829240893e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.3961604342159437,
      "tokens_seen": 2507735040
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.849141389825069e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.3961464029252104,
      "tokens_seen": 2507866112
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.848338950409244e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.3961323725731165,
      "tokens_seen": 2507997184
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.84753651099342e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.3961183431595496,
      "tokens_seen": 2508128256
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.846734071577596e-05,
      "loss": 1.2268,
      "theoretical_loss": 3.3961043146843983,
      "tokens_seen": 2508259328
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.845931632161772e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.3960902871475507,
      "tokens_seen": 2508390400
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.9928663372993469,
      "objective/train/avg_token_score": 0.007133623585104942,
      "objective/train/docs_used": 912960,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2874207496643066,
      "objective/train/original_loss": 1.2978150844573975,
      "objective/train/theoretical_loss": 3.3960762605488948,
      "objective/train/tokens_used": 878980576,
      "objective/train/unlikelihood_loss": 0.018837247043848038,
      "theoretical_loss": 3.3960762605488948,
      "tokens_seen": 2508521472
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.845129192745948e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.3960762605488948,
      "tokens_seen": 2508521472
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.844326753330124e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.396062234888319,
      "tokens_seen": 2508652544
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8435243139143e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.396048210165712,
      "tokens_seen": 2508783616
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8427218744984754e-05,
      "loss": 1.1922,
      "theoretical_loss": 3.396034186380961,
      "tokens_seen": 2508914688
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8419194350826514e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.3960201635339553,
      "tokens_seen": 2509045760
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8411169956668274e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.3960061416245826,
      "tokens_seen": 2509176832
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8403145562510034e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.3959921206527315,
      "tokens_seen": 2509307904
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8395121168351794e-05,
      "loss": 1.3573,
      "theoretical_loss": 3.3959781006182905,
      "tokens_seen": 2509438976
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8387096774193554e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.395964081521148,
      "tokens_seen": 2509570048
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8379072380035314e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.3959500633611914,
      "tokens_seen": 2509701120
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.837104798587707e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.3959360461383103,
      "tokens_seen": 2509832192
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.836302359171883e-05,
      "loss": 1.1284,
      "theoretical_loss": 3.395922029852393,
      "tokens_seen": 2509963264
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.835499919756059e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.3959080145033274,
      "tokens_seen": 2510094336
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8346974803402347e-05,
      "loss": 1.1922,
      "theoretical_loss": 3.395894000091002,
      "tokens_seen": 2510225408
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8338950409244106e-05,
      "loss": 1.301,
      "theoretical_loss": 3.3958799866153058,
      "tokens_seen": 2510356480
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8330926015085866e-05,
      "loss": 1.2066,
      "theoretical_loss": 3.3958659740761266,
      "tokens_seen": 2510487552
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8322901620927626e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.395851962473354,
      "tokens_seen": 2510618624
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8314877226769386e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.3958379518068753,
      "tokens_seen": 2510749696
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.830685283261114e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.3958239420765794,
      "tokens_seen": 2510880768
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.82988284384529e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.3958099332823557,
      "tokens_seen": 2511011840
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.829080404429466e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.395795925424092,
      "tokens_seen": 2511142912
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.828277965013642e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.3957819185016773,
      "tokens_seen": 2511273984
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.827475525597818e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.3957679125149998,
      "tokens_seen": 2511405056
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.826673086181994e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.3957539074639485,
      "tokens_seen": 2511536128
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.82587064676617e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.395739903348412,
      "tokens_seen": 2511667200
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.825068207350345e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.395725900168279,
      "tokens_seen": 2511798272
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.824265767934521e-05,
      "loss": 1.3524,
      "theoretical_loss": 3.395711897923438,
      "tokens_seen": 2511929344
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.823463328518697e-05,
      "loss": 1.3358,
      "theoretical_loss": 3.395697896613778,
      "tokens_seen": 2512060416
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.822660889102873e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.3956838962391878,
      "tokens_seen": 2512191488
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.821858449687049e-05,
      "loss": 1.1999,
      "theoretical_loss": 3.395669896799556,
      "tokens_seen": 2512322560
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.821056010271225e-05,
      "loss": 1.2258,
      "theoretical_loss": 3.3956558982947715,
      "tokens_seen": 2512453632
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.820253570855401e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.395641900724723,
      "tokens_seen": 2512584704
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8194511314395764e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.3956279040892996,
      "tokens_seen": 2512715776
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8186486920237524e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.3956139083883894,
      "tokens_seen": 2512846848
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8178462526079284e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.395599913621882,
      "tokens_seen": 2512977920
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8170438131921044e-05,
      "loss": 1.2496,
      "theoretical_loss": 3.3955859197896663,
      "tokens_seen": 2513108992
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8162413737762804e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.395571926891631,
      "tokens_seen": 2513240064
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8154389343604564e-05,
      "loss": 1.328,
      "theoretical_loss": 3.395557934927665,
      "tokens_seen": 2513371136
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8146364949446324e-05,
      "loss": 1.2287,
      "theoretical_loss": 3.395543943897657,
      "tokens_seen": 2513502208
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8138340555288084e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.395529953801496,
      "tokens_seen": 2513633280
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.813031616112984e-05,
      "loss": 1.2071,
      "theoretical_loss": 3.3955159646390713,
      "tokens_seen": 2513764352
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.81222917669716e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.395501976410272,
      "tokens_seen": 2513895424
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.811426737281336e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.3954879891149865,
      "tokens_seen": 2514026496
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8106242978655117e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3954740027531045,
      "tokens_seen": 2514157568
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8098218584496877e-05,
      "loss": 1.2106,
      "theoretical_loss": 3.395460017324515,
      "tokens_seen": 2514288640
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8090194190338636e-05,
      "loss": 1.287,
      "theoretical_loss": 3.3954460328291067,
      "tokens_seen": 2514419712
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.8082169796180396e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.395432049266769,
      "tokens_seen": 2514550784
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.807414540202215e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.3954180666373905,
      "tokens_seen": 2514681856
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.806612100786391e-05,
      "loss": 1.2313,
      "theoretical_loss": 3.3954040849408607,
      "tokens_seen": 2514812928
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.805809661370567e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.395390104177069,
      "tokens_seen": 2514944000
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.9932106137275696,
      "objective/train/avg_token_score": 0.006789429113268852,
      "objective/train/docs_used": 915361,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2340677976608276,
      "objective/train/original_loss": 1.2423590421676636,
      "objective/train/theoretical_loss": 3.395376124345904,
      "objective/train/tokens_used": 885534176,
      "objective/train/unlikelihood_loss": 0.02600879967212677,
      "theoretical_loss": 3.395376124345904,
      "tokens_seen": 2515075072
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.805007221954743e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.395376124345904,
      "tokens_seen": 2515075072
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.804204782538919e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.3953621454472556,
      "tokens_seen": 2515206144
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.803402343123095e-05,
      "loss": 1.223,
      "theoretical_loss": 3.3953481674810124,
      "tokens_seen": 2515337216
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.80259990370727e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.3953341904470635,
      "tokens_seen": 2515468288
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.801797464291446e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.395320214345299,
      "tokens_seen": 2515599360
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.800995024875622e-05,
      "loss": 1.3257,
      "theoretical_loss": 3.3953062391756075,
      "tokens_seen": 2515730432
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.800192585459798e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.3952922649378783,
      "tokens_seen": 2515861504
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.799390146043974e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.395278291632001,
      "tokens_seen": 2515992576
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.79858770662815e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.3952643192578646,
      "tokens_seen": 2516123648
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7977852672123255e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.3952503478153586,
      "tokens_seen": 2516254720
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7969828277965015e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.3952363773043723,
      "tokens_seen": 2516385792
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7961803883806775e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.3952224077247957,
      "tokens_seen": 2516516864
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7953779489648534e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.3952084390765167,
      "tokens_seen": 2516647936
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7945755095490294e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.395194471359426,
      "tokens_seen": 2516779008
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7937730701332054e-05,
      "loss": 1.2268,
      "theoretical_loss": 3.395180504573413,
      "tokens_seen": 2516910080
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.792970630717381e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.3951665387183665,
      "tokens_seen": 2517041152
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.792168191301557e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.395152573794176,
      "tokens_seen": 2517172224
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.791365751885733e-05,
      "loss": 1.262,
      "theoretical_loss": 3.395138609800732,
      "tokens_seen": 2517303296
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.790563312469909e-05,
      "loss": 1.284,
      "theoretical_loss": 3.395124646737923,
      "tokens_seen": 2517434368
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.789760873054085e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.3951106846056383,
      "tokens_seen": 2517565440
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.78895843363826e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.395096723403768,
      "tokens_seen": 2517696512
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.788155994222436e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.3950827631322023,
      "tokens_seen": 2517827584
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.787353554806612e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.3950688037908296,
      "tokens_seen": 2517958656
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.786551115390788e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.3950548453795406,
      "tokens_seen": 2518089728
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.785748675974964e-05,
      "loss": 1.2238,
      "theoretical_loss": 3.3950408878982237,
      "tokens_seen": 2518220800
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.78494623655914e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.395026931346769,
      "tokens_seen": 2518351872
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.784143797143316e-05,
      "loss": 1.1602,
      "theoretical_loss": 3.3950129757250673,
      "tokens_seen": 2518482944
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.783341357727491e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.3949990210330068,
      "tokens_seen": 2518614016
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.782538918311667e-05,
      "loss": 1.2161,
      "theoretical_loss": 3.3949850672704773,
      "tokens_seen": 2518745088
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.781736478895843e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.3949711144373698,
      "tokens_seen": 2518876160
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.780934039480019e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.3949571625335726,
      "tokens_seen": 2519007232
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.780131600064195e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.394943211558976,
      "tokens_seen": 2519138304
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.779329160648371e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.39492926151347,
      "tokens_seen": 2519269376
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.778526721232547e-05,
      "loss": 1.3308,
      "theoretical_loss": 3.394915312396944,
      "tokens_seen": 2519400448
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7777242818167225e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.394901364209288,
      "tokens_seen": 2519531520
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7769218424008985e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.394887416950392,
      "tokens_seen": 2519662592
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7761194029850745e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.3948734706201456,
      "tokens_seen": 2519793664
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7753169635692505e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.394859525218439,
      "tokens_seen": 2519924736
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7745145241534265e-05,
      "loss": 1.2003,
      "theoretical_loss": 3.3948455807451614,
      "tokens_seen": 2520055808
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7737120847376025e-05,
      "loss": 1.2709,
      "theoretical_loss": 3.3948316372002036,
      "tokens_seen": 2520186880
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7729096453217785e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.3948176945834545,
      "tokens_seen": 2520317952
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.772107205905954e-05,
      "loss": 1.257,
      "theoretical_loss": 3.3948037528948047,
      "tokens_seen": 2520449024
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.77130476649013e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.3947898121341447,
      "tokens_seen": 2520580096
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.770502327074306e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.3947758723013632,
      "tokens_seen": 2520711168
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.769699887658482e-05,
      "loss": 1.26,
      "theoretical_loss": 3.394761933396351,
      "tokens_seen": 2520842240
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.768897448242658e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3947479954189976,
      "tokens_seen": 2520973312
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.768095008826834e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.394734058369194,
      "tokens_seen": 2521104384
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.76729256941101e-05,
      "loss": 1.3031,
      "theoretical_loss": 3.3947201222468295,
      "tokens_seen": 2521235456
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.766490129995186e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.3947061870517943,
      "tokens_seen": 2521366528
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.765687690579361e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.3946922527839782,
      "tokens_seen": 2521497600
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.9940690994262695,
      "objective/train/avg_token_score": 0.005930920131504536,
      "objective/train/docs_used": 917863,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2279638051986694,
      "objective/train/original_loss": 1.2349457740783691,
      "objective/train/theoretical_loss": 3.394678319443272,
      "objective/train/tokens_used": 892087776,
      "objective/train/unlikelihood_loss": 0.021120529621839523,
      "theoretical_loss": 3.394678319443272,
      "tokens_seen": 2521628672
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.764885251163537e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.394678319443272,
      "tokens_seen": 2521628672
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.764082811747713e-05,
      "loss": 1.2927,
      "theoretical_loss": 3.3946643870295654,
      "tokens_seen": 2521759744
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.763280372331889e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.3946504555427484,
      "tokens_seen": 2521890816
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.762477932916065e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.3946365249827117,
      "tokens_seen": 2522021888
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.761675493500241e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.394622595349345,
      "tokens_seen": 2522152960
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.760873054084417e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.3946086666425384,
      "tokens_seen": 2522284032
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.760070614668592e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.3945947388621827,
      "tokens_seen": 2522415104
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.759268175252768e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.3945808120081677,
      "tokens_seen": 2522546176
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.758465735836944e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.3945668860803835,
      "tokens_seen": 2522677248
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.75766329642112e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.3945529610787215,
      "tokens_seen": 2522808320
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.756860857005296e-05,
      "loss": 1.2402,
      "theoretical_loss": 3.3945390370030704,
      "tokens_seen": 2522939392
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.756058417589472e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3945251138533212,
      "tokens_seen": 2523070464
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.755255978173648e-05,
      "loss": 1.236,
      "theoretical_loss": 3.394511191629365,
      "tokens_seen": 2523201536
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.754453538757824e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.394497270331091,
      "tokens_seen": 2523332608
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7536510993419995e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.39448334995839,
      "tokens_seen": 2523463680
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7528486599261755e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.394469430511153,
      "tokens_seen": 2523594752
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7520462205103515e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.394455511989269,
      "tokens_seen": 2523725824
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7512437810945275e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.39444159439263,
      "tokens_seen": 2523856896
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7504413416787035e-05,
      "loss": 1.3573,
      "theoretical_loss": 3.3944276777211253,
      "tokens_seen": 2523987968
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7496389022628795e-05,
      "loss": 1.2941,
      "theoretical_loss": 3.3944137619746457,
      "tokens_seen": 2524119040
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7488364628470555e-05,
      "loss": 1.26,
      "theoretical_loss": 3.394399847153082,
      "tokens_seen": 2524250112
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.748034023431231e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.3943859332563244,
      "tokens_seen": 2524381184
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.747231584015407e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.3943720202842638,
      "tokens_seen": 2524512256
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.746429144599583e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.3943581082367897,
      "tokens_seen": 2524643328
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.745626705183759e-05,
      "loss": 1.4163,
      "theoretical_loss": 3.394344197113794,
      "tokens_seen": 2524774400
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.744824265767935e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.394330286915167,
      "tokens_seen": 2524905472
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.744021826352111e-05,
      "loss": 1.3397,
      "theoretical_loss": 3.3943163776407985,
      "tokens_seen": 2525036544
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.743219386936287e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.3943024692905794,
      "tokens_seen": 2525167616
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.742416947520462e-05,
      "loss": 1.3803,
      "theoretical_loss": 3.3942885618644008,
      "tokens_seen": 2525298688
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.741614508104638e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.394274655362153,
      "tokens_seen": 2525429760
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.740812068688814e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.3942607497837267,
      "tokens_seen": 2525560832
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.74000962927299e-05,
      "loss": 1.2297,
      "theoretical_loss": 3.394246845129013,
      "tokens_seen": 2525691904
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.739207189857166e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.3942329413979015,
      "tokens_seen": 2525822976
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.738404750441342e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.3942190385902844,
      "tokens_seen": 2525954048
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.737602311025518e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.3942051367060513,
      "tokens_seen": 2526085120
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.736799871609694e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.3941912357450934,
      "tokens_seen": 2526216192
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.735997432193869e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.3941773357073015,
      "tokens_seen": 2526347264
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.735194992778045e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.3941634365925664,
      "tokens_seen": 2526478336
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.734392553362221e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.394149538400779,
      "tokens_seen": 2526609408
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.733590113946397e-05,
      "loss": 1.276,
      "theoretical_loss": 3.3941356411318298,
      "tokens_seen": 2526740480
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.732787674530573e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.3941217447856102,
      "tokens_seen": 2526871552
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.731985235114749e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.3941078493620105,
      "tokens_seen": 2527002624
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.731182795698925e-05,
      "loss": 1.276,
      "theoretical_loss": 3.394093954860922,
      "tokens_seen": 2527133696
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7303803562831006e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.3940800612822355,
      "tokens_seen": 2527264768
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7295779168672765e-05,
      "loss": 1.227,
      "theoretical_loss": 3.3940661686258418,
      "tokens_seen": 2527395840
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7287754774514525e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.3940522768916317,
      "tokens_seen": 2527526912
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7279730380356285e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.394038386079497,
      "tokens_seen": 2527657984
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7271705986198045e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.3940244961893278,
      "tokens_seen": 2527789056
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7263681592039805e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.3940106072210154,
      "tokens_seen": 2527920128
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7255657197881565e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.3939967191744507,
      "tokens_seen": 2528051200
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.9934902787208557,
      "objective/train/avg_token_score": 0.00650973804295063,
      "objective/train/docs_used": 920165,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1782459020614624,
      "objective/train/original_loss": 1.1857519149780273,
      "objective/train/theoretical_loss": 3.3939828320495247,
      "objective/train/tokens_used": 898641376,
      "objective/train/unlikelihood_loss": 0.022171668708324432,
      "theoretical_loss": 3.3939828320495247,
      "tokens_seen": 2528182272
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.724763280372332e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.3939828320495247,
      "tokens_seen": 2528182272
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.723960840956508e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.393968945846129,
      "tokens_seen": 2528313344
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.723158401540684e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.393955060564154,
      "tokens_seen": 2528444416
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.72235596212486e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.3939411762034917,
      "tokens_seen": 2528575488
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.721553522709036e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.3939272927640323,
      "tokens_seen": 2528706560
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.720751083293212e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.393913410245667,
      "tokens_seen": 2528837632
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.719948643877388e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.393899528648287,
      "tokens_seen": 2528968704
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.719146204461564e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.3938856479717847,
      "tokens_seen": 2529099776
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.718343765045739e-05,
      "loss": 1.256,
      "theoretical_loss": 3.3938717682160497,
      "tokens_seen": 2529230848
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.717541325629915e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.393857889380974,
      "tokens_seen": 2529361920
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.716738886214091e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.393844011466448,
      "tokens_seen": 2529492992
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.715936446798267e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3938301344723643,
      "tokens_seen": 2529624064
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.715134007382443e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.393816258398613,
      "tokens_seen": 2529755136
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.714331567966619e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3938023832450854,
      "tokens_seen": 2529886208
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.713529128550795e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.393788509011674,
      "tokens_seen": 2530017280
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.71272668913497e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.393774635698269,
      "tokens_seen": 2530148352
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.711924249719146e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3937607633047624,
      "tokens_seen": 2530279424
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.711121810303322e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3937468918310447,
      "tokens_seen": 2530410496
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.710319370887498e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.3937330212770083,
      "tokens_seen": 2530541568
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.709516931471674e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.393719151642544,
      "tokens_seen": 2530672640
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.70871449205585e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.393705282927543,
      "tokens_seen": 2530803712
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.707912052640026e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3936914151318973,
      "tokens_seen": 2530934784
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.707109613224202e-05,
      "loss": 1.3524,
      "theoretical_loss": 3.3936775482554977,
      "tokens_seen": 2531065856
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7063071738083776e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.3936636822982367,
      "tokens_seen": 2531196928
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7055047343925536e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.3936498172600045,
      "tokens_seen": 2531328000
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7047022949767295e-05,
      "loss": 1.2485,
      "theoretical_loss": 3.3936359531406937,
      "tokens_seen": 2531459072
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7038998555609055e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.3936220899401954,
      "tokens_seen": 2531590144
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7030974161450815e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.393608227658401,
      "tokens_seen": 2531721216
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7022949767292575e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.393594366295202,
      "tokens_seen": 2531852288
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.7014925373134335e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.39358050585049,
      "tokens_seen": 2531983360
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.700690097897609e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.393566646324157,
      "tokens_seen": 2532114432
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.699887658481785e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.3935527877160947,
      "tokens_seen": 2532245504
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.699085219065961e-05,
      "loss": 1.3388,
      "theoretical_loss": 3.3935389300261942,
      "tokens_seen": 2532376576
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.698282779650137e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.393525073254347,
      "tokens_seen": 2532507648
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.697480340234313e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.3935112174004454,
      "tokens_seen": 2532638720
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.696677900818489e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.3934973624643807,
      "tokens_seen": 2532769792
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.695875461402665e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.3934835084460446,
      "tokens_seen": 2532900864
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.69507302198684e-05,
      "loss": 1.3663,
      "theoretical_loss": 3.393469655345329,
      "tokens_seen": 2533031936
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.694270582571016e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.393455803162126,
      "tokens_seen": 2533163008
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.693468143155192e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.393441951896327,
      "tokens_seen": 2533294080
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.692665703739368e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3934281015478227,
      "tokens_seen": 2533425152
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.691863264323544e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.3934142521165067,
      "tokens_seen": 2533556224
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.69106082490772e-05,
      "loss": 1.307,
      "theoretical_loss": 3.39340040360227,
      "tokens_seen": 2533687296
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.690258385491896e-05,
      "loss": 1.3283,
      "theoretical_loss": 3.3933865560050043,
      "tokens_seen": 2533818368
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.689455946076072e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.393372709324602,
      "tokens_seen": 2533949440
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.688653506660247e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.3933588635609544,
      "tokens_seen": 2534080512
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.687851067244423e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.3933450187139536,
      "tokens_seen": 2534211584
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.687048627828599e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.3933311747834916,
      "tokens_seen": 2534342656
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.686246188412775e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.3933173317694596,
      "tokens_seen": 2534473728
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.685443748996951e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3933034896717507,
      "tokens_seen": 2534604800
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.9930554628372192,
      "objective/train/avg_token_score": 0.006944500375539064,
      "objective/train/docs_used": 922493,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.308261752128601,
      "objective/train/original_loss": 1.3183566331863403,
      "objective/train/theoretical_loss": 3.3932896484902564,
      "objective/train/tokens_used": 905194976,
      "objective/train/unlikelihood_loss": 0.017484791576862335,
      "theoretical_loss": 3.3932896484902564,
      "tokens_seen": 2534735872
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.684641309581127e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.3932896484902564,
      "tokens_seen": 2534735872
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.683838870165303e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.3932758082248684,
      "tokens_seen": 2534866944
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6830364307494786e-05,
      "loss": 1.2423,
      "theoretical_loss": 3.3932619688754793,
      "tokens_seen": 2534998016
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6822339913336546e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.3932481304419806,
      "tokens_seen": 2535129088
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6814315519178306e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.393234292924264,
      "tokens_seen": 2535260160
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6806291125020066e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3932204563222226,
      "tokens_seen": 2535391232
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6798266730861825e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.393206620635748,
      "tokens_seen": 2535522304
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6790242336703585e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.3931927858647315,
      "tokens_seen": 2535653376
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6782217942545345e-05,
      "loss": 1.1961,
      "theoretical_loss": 3.3931789520090665,
      "tokens_seen": 2535784448
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.67741935483871e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.3931651190686445,
      "tokens_seen": 2535915520
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.676616915422886e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.3931512870433576,
      "tokens_seen": 2536046592
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.675814476007062e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.393137455933098,
      "tokens_seen": 2536177664
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.675012036591238e-05,
      "loss": 1.3529,
      "theoretical_loss": 3.3931236257377577,
      "tokens_seen": 2536308736
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.674209597175414e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.3931097964572294,
      "tokens_seen": 2536439808
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.67340715775959e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.393095968091405,
      "tokens_seen": 2536570880
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.672604718343766e-05,
      "loss": 1.1901,
      "theoretical_loss": 3.393082140640177,
      "tokens_seen": 2536701952
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.671802278927941e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3930683141034375,
      "tokens_seen": 2536833024
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.670999839512117e-05,
      "loss": 1.296,
      "theoretical_loss": 3.3930544884810785,
      "tokens_seen": 2536964096
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.670197400096293e-05,
      "loss": 1.1804,
      "theoretical_loss": 3.393040663772992,
      "tokens_seen": 2537095168
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.669394960680469e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.393026839979072,
      "tokens_seen": 2537226240
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.668592521264645e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.393013017099209,
      "tokens_seen": 2537357312
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.667790081848821e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.392999195133296,
      "tokens_seen": 2537488384
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6669876424329964e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.3929853740812255,
      "tokens_seen": 2537619456
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6661852030171723e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.39297155394289,
      "tokens_seen": 2537750528
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6653827636013483e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.392957734718181,
      "tokens_seen": 2537881600
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.664580324185524e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.392943916406992,
      "tokens_seen": 2538012672
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6637778847697e-05,
      "loss": 1.2148,
      "theoretical_loss": 3.392930099009215,
      "tokens_seen": 2538143744
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.662975445353876e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3929162825247423,
      "tokens_seen": 2538274816
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6621730059380516e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.392902466953467,
      "tokens_seen": 2538405888
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6613705665222276e-05,
      "loss": 1.208,
      "theoretical_loss": 3.392888652295281,
      "tokens_seen": 2538536960
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6605681271064036e-05,
      "loss": 1.1875,
      "theoretical_loss": 3.3928748385500764,
      "tokens_seen": 2538668032
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6597656876905796e-05,
      "loss": 1.2332,
      "theoretical_loss": 3.392861025717747,
      "tokens_seen": 2538799104
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6589632482747556e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.392847213798184,
      "tokens_seen": 2538930176
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6581608088589316e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.3928334027912816,
      "tokens_seen": 2539061248
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.657358369443107e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.3928195926969305,
      "tokens_seen": 2539192320
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.656555930027283e-05,
      "loss": 1.351,
      "theoretical_loss": 3.3928057835150245,
      "tokens_seen": 2539323392
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.655753490611459e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.392791975245456,
      "tokens_seen": 2539454464
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.654951051195635e-05,
      "loss": 1.27,
      "theoretical_loss": 3.3927781678881175,
      "tokens_seen": 2539585536
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.654148611779811e-05,
      "loss": 1.1698,
      "theoretical_loss": 3.3927643614429015,
      "tokens_seen": 2539716608
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.653346172363987e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.392750555909701,
      "tokens_seen": 2539847680
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.652543732948162e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.3927367512884086,
      "tokens_seen": 2539978752
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.651741293532338e-05,
      "loss": 1.2159,
      "theoretical_loss": 3.392722947578917,
      "tokens_seen": 2540109824
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.650938854116514e-05,
      "loss": 1.363,
      "theoretical_loss": 3.392709144781119,
      "tokens_seen": 2540240896
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.65013641470069e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.3926953428949074,
      "tokens_seen": 2540371968
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.649333975284866e-05,
      "loss": 1.3209,
      "theoretical_loss": 3.3926815419201746,
      "tokens_seen": 2540503040
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.648531535869042e-05,
      "loss": 1.255,
      "theoretical_loss": 3.3926677418568136,
      "tokens_seen": 2540634112
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6477290964532174e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.3926539427047175,
      "tokens_seen": 2540765184
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6469266570373934e-05,
      "loss": 1.2293,
      "theoretical_loss": 3.3926401444637784,
      "tokens_seen": 2540896256
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6461242176215694e-05,
      "loss": 1.2494,
      "theoretical_loss": 3.39262634713389,
      "tokens_seen": 2541027328
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6453217782057454e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.3926125507149445,
      "tokens_seen": 2541158400
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.9934854507446289,
      "objective/train/avg_token_score": 0.006514542270451784,
      "objective/train/docs_used": 924855,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2264972925186157,
      "objective/train/original_loss": 1.2349741458892822,
      "objective/train/theoretical_loss": 3.3925987552068353,
      "objective/train/tokens_used": 911748576,
      "objective/train/unlikelihood_loss": 0.020402656868100166,
      "theoretical_loss": 3.3925987552068353,
      "tokens_seen": 2541289472
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6445193387899214e-05,
      "loss": 1.297,
      "theoretical_loss": 3.3925987552068353,
      "tokens_seen": 2541289472
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6437168993740974e-05,
      "loss": 1.29,
      "theoretical_loss": 3.392584960609455,
      "tokens_seen": 2541420544
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6429144599582734e-05,
      "loss": 1.1988,
      "theoretical_loss": 3.3925711669226963,
      "tokens_seen": 2541551616
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6421120205424494e-05,
      "loss": 1.3407,
      "theoretical_loss": 3.392557374146452,
      "tokens_seen": 2541682688
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.641309581126625e-05,
      "loss": 1.283,
      "theoretical_loss": 3.392543582280616,
      "tokens_seen": 2541813760
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6405071417108007e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.3925297913250807,
      "tokens_seen": 2541944832
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6397047022949767e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.3925160012797386,
      "tokens_seen": 2542075904
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6389022628791526e-05,
      "loss": 1.2127,
      "theoretical_loss": 3.392502212144483,
      "tokens_seen": 2542206976
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6380998234633286e-05,
      "loss": 1.323,
      "theoretical_loss": 3.392488423919208,
      "tokens_seen": 2542338048
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6372973840475046e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.392474636603805,
      "tokens_seen": 2542469120
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6364949446316806e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.392460850198168,
      "tokens_seen": 2542600192
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.635692505215856e-05,
      "loss": 1.2971,
      "theoretical_loss": 3.3924470647021896,
      "tokens_seen": 2542731264
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.634890065800032e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.3924332801157635,
      "tokens_seen": 2542862336
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.634087626384208e-05,
      "loss": 1.3324,
      "theoretical_loss": 3.392419496438782,
      "tokens_seen": 2542993408
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.633285186968384e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.392405713671139,
      "tokens_seen": 2543124480
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.63248274755256e-05,
      "loss": 1.2016,
      "theoretical_loss": 3.3923919318127274,
      "tokens_seen": 2543255552
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.631680308136736e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.3923781508634403,
      "tokens_seen": 2543386624
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.630877868720912e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.3923643708231705,
      "tokens_seen": 2543517696
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.630075429305087e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.3923505916918115,
      "tokens_seen": 2543648768
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.629272989889263e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.392336813469257,
      "tokens_seen": 2543779840
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.628470550473439e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.3923230361553998,
      "tokens_seen": 2543910912
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.627668111057615e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.392309259750133,
      "tokens_seen": 2544041984
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.626865671641791e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.39229548425335,
      "tokens_seen": 2544173056
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.626063232225967e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.3922817096649442,
      "tokens_seen": 2544304128
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.625260792810143e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.392267935984809,
      "tokens_seen": 2544435200
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.624458353394319e-05,
      "loss": 1.301,
      "theoretical_loss": 3.3922541632128373,
      "tokens_seen": 2544566272
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6236559139784944e-05,
      "loss": 1.241,
      "theoretical_loss": 3.392240391348923,
      "tokens_seen": 2544697344
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6228534745626704e-05,
      "loss": 1.2441,
      "theoretical_loss": 3.392226620392959,
      "tokens_seen": 2544828416
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6220510351468464e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.392212850344839,
      "tokens_seen": 2544959488
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6212485957310224e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.3921990812044562,
      "tokens_seen": 2545090560
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6204461563151984e-05,
      "loss": 1.1736,
      "theoretical_loss": 3.392185312971704,
      "tokens_seen": 2545221632
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6196437168993744e-05,
      "loss": 1.1621,
      "theoretical_loss": 3.3921715456464763,
      "tokens_seen": 2545352704
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6188412774835504e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.392157779228666,
      "tokens_seen": 2545483776
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.618038838067726e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.3921440137181667,
      "tokens_seen": 2545614848
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.617236398651902e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3921302491148717,
      "tokens_seen": 2545745920
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.616433959236078e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.392116485418675,
      "tokens_seen": 2545876992
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6156315198202537e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.3921027226294695,
      "tokens_seen": 2546008064
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6148290804044296e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.3920889607471496,
      "tokens_seen": 2546139136
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6140266409886056e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.3920751997716083,
      "tokens_seen": 2546270208
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6132242015727816e-05,
      "loss": 1.218,
      "theoretical_loss": 3.392061439702739,
      "tokens_seen": 2546401280
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.612421762156957e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.392047680540436,
      "tokens_seen": 2546532352
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.611619322741133e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.392033922284592,
      "tokens_seen": 2546663424
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.610816883325309e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.392020164935101,
      "tokens_seen": 2546794496
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.610014443909485e-05,
      "loss": 1.1684,
      "theoretical_loss": 3.392006408491857,
      "tokens_seen": 2546925568
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.609212004493661e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.3919926529547535,
      "tokens_seen": 2547056640
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.608409565077837e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.391978898323684,
      "tokens_seen": 2547187712
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.607607125662013e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.391965144598542,
      "tokens_seen": 2547318784
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.606804686246189e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.391951391779222,
      "tokens_seen": 2547449856
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.606002246830364e-05,
      "loss": 1.297,
      "theoretical_loss": 3.391937639865617,
      "tokens_seen": 2547580928
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.60519980741454e-05,
      "loss": 1.1974,
      "theoretical_loss": 3.391923888857621,
      "tokens_seen": 2547712000
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.9907525777816772,
      "objective/train/avg_token_score": 0.009247391484677792,
      "objective/train/docs_used": 927107,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.187857985496521,
      "objective/train/original_loss": 1.1991902589797974,
      "objective/train/theoretical_loss": 3.391910138755128,
      "objective/train/tokens_used": 918302176,
      "objective/train/unlikelihood_loss": 0.029310837388038635,
      "theoretical_loss": 3.391910138755128,
      "tokens_seen": 2547843072
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.604397367998716e-05,
      "loss": 1.2123,
      "theoretical_loss": 3.391910138755128,
      "tokens_seen": 2547843072
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.603594928582892e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.391896389558031,
      "tokens_seen": 2547974144
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.602792489167068e-05,
      "loss": 1.1745,
      "theoretical_loss": 3.3918826412662253,
      "tokens_seen": 2548105216
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.601990049751244e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3918688938796033,
      "tokens_seen": 2548236288
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.60118761033542e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.3918551473980596,
      "tokens_seen": 2548367360
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.6003851709195954e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.391841401821488,
      "tokens_seen": 2548498432
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.5995827315037714e-05,
      "loss": 1.323,
      "theoretical_loss": 3.391827657149782,
      "tokens_seen": 2548629504
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.5987802920879474e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.3918139133828356,
      "tokens_seen": 2548760576
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.5979778526721234e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.391800170520543,
      "tokens_seen": 2548891648
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.5971754132562994e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.391786428562798,
      "tokens_seen": 2549022720
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.5963729738404754e-05,
      "loss": 1.2023,
      "theoretical_loss": 3.3917726875094947,
      "tokens_seen": 2549153792
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5955705344246514e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.3917589473605267,
      "tokens_seen": 2549284864
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5947680950088274e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.3917452081157884,
      "tokens_seen": 2549415936
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.593965655593003e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.3917314697751735,
      "tokens_seen": 2549547008
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.593163216177179e-05,
      "loss": 1.229,
      "theoretical_loss": 3.391717732338576,
      "tokens_seen": 2549678080
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.592360776761355e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.3917039958058903,
      "tokens_seen": 2549809152
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.591558337345531e-05,
      "loss": 1.254,
      "theoretical_loss": 3.3916902601770103,
      "tokens_seen": 2549940224
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5907558979297067e-05,
      "loss": 1.238,
      "theoretical_loss": 3.39167652545183,
      "tokens_seen": 2550071296
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5899534585138826e-05,
      "loss": 1.289,
      "theoretical_loss": 3.3916627916302433,
      "tokens_seen": 2550202368
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5891510190980586e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.391649058712145,
      "tokens_seen": 2550333440
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.588348579682234e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3916353266974286,
      "tokens_seen": 2550464512
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.58754614026641e-05,
      "loss": 1.2176,
      "theoretical_loss": 3.3916215955859883,
      "tokens_seen": 2550595584
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.586743700850586e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.3916078653777184,
      "tokens_seen": 2550726656
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.585941261434762e-05,
      "loss": 1.176,
      "theoretical_loss": 3.3915941360725133,
      "tokens_seen": 2550857728
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.585138822018938e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.3915804076702667,
      "tokens_seen": 2550988800
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.584336382603114e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3915666801708735,
      "tokens_seen": 2551119872
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.58353394318729e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3915529535742275,
      "tokens_seen": 2551250944
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.582731503771465e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.391539227880223,
      "tokens_seen": 2551382016
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.581929064355641e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.3915255030887543,
      "tokens_seen": 2551513088
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.581126624939817e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3915117791997154,
      "tokens_seen": 2551644160
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.580324185523993e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.3914980562130013,
      "tokens_seen": 2551775232
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.579521746108169e-05,
      "loss": 1.2296,
      "theoretical_loss": 3.3914843341285055,
      "tokens_seen": 2551906304
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.578719306692345e-05,
      "loss": 1.1697,
      "theoretical_loss": 3.391470612946123,
      "tokens_seen": 2552037376
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.577916867276521e-05,
      "loss": 1.1614,
      "theoretical_loss": 3.3914568926657482,
      "tokens_seen": 2552168448
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.577114427860697e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3914431732872754,
      "tokens_seen": 2552299520
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5763119884448725e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.3914294548105985,
      "tokens_seen": 2552430592
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5755095490290484e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.3914157372356124,
      "tokens_seen": 2552561664
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5747071096132244e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.391402020562211,
      "tokens_seen": 2552692736
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5739046701974004e-05,
      "loss": 1.2403,
      "theoretical_loss": 3.3913883047902895,
      "tokens_seen": 2552823808
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5731022307815764e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.3913745899197423,
      "tokens_seen": 2552954880
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5722997913657524e-05,
      "loss": 1.1823,
      "theoretical_loss": 3.391360875950463,
      "tokens_seen": 2553085952
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5714973519499284e-05,
      "loss": 1.2056,
      "theoretical_loss": 3.391347162882347,
      "tokens_seen": 2553217024
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.570694912534104e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.3913334507152886,
      "tokens_seen": 2553348096
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.56989247311828e-05,
      "loss": 1.2121,
      "theoretical_loss": 3.391319739449182,
      "tokens_seen": 2553479168
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.569090033702456e-05,
      "loss": 1.1803,
      "theoretical_loss": 3.3913060290839225,
      "tokens_seen": 2553610240
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.568287594286632e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.391292319619404,
      "tokens_seen": 2553741312
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.567485154870808e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.391278611055521,
      "tokens_seen": 2553872384
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.566682715454984e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.391264903392169,
      "tokens_seen": 2554003456
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5658802760391597e-05,
      "loss": 1.29,
      "theoretical_loss": 3.3912511966292422,
      "tokens_seen": 2554134528
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.565077836623335e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.3912374907666347,
      "tokens_seen": 2554265600
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.9925299882888794,
      "objective/train/avg_token_score": 0.007469974923878908,
      "objective/train/docs_used": 929369,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2683384418487549,
      "objective/train/original_loss": 1.2790807485580444,
      "objective/train/theoretical_loss": 3.3912237858042413,
      "objective/train/tokens_used": 924855776,
      "objective/train/unlikelihood_loss": 0.020853674039244652,
      "theoretical_loss": 3.3912237858042413,
      "tokens_seen": 2554396672
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.564275397207511e-05,
      "loss": 1.2173,
      "theoretical_loss": 3.3912237858042413,
      "tokens_seen": 2554396672
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.563472957791687e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.3912100817419577,
      "tokens_seen": 2554527744
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.562670518375863e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.3911963785796777,
      "tokens_seen": 2554658816
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.561868078960039e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.391182676317296,
      "tokens_seen": 2554789888
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.561065639544215e-05,
      "loss": 1.1987,
      "theoretical_loss": 3.3911689749547076,
      "tokens_seen": 2554920960
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.560263200128391e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.391155274491807,
      "tokens_seen": 2555052032
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.559460760712567e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.3911415749284894,
      "tokens_seen": 2555183104
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.558658321296742e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.3911278762646493,
      "tokens_seen": 2555314176
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.557855881880918e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3911141785001817,
      "tokens_seen": 2555445248
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.557053442465094e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.391100481634981,
      "tokens_seen": 2555576320
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.55625100304927e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.391086785668943,
      "tokens_seen": 2555707392
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.555448563633446e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.391073090601961,
      "tokens_seen": 2555838464
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.554646124217622e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.3910593964339313,
      "tokens_seen": 2555969536
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.553843684801798e-05,
      "loss": 1.2375,
      "theoretical_loss": 3.3910457031647483,
      "tokens_seen": 2556100608
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5530412453859735e-05,
      "loss": 1.2349,
      "theoretical_loss": 3.391032010794307,
      "tokens_seen": 2556231680
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5522388059701495e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.391018319322502,
      "tokens_seen": 2556362752
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5514363665543255e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.391004628749229,
      "tokens_seen": 2556493824
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5506339271385014e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3909909390743818,
      "tokens_seen": 2556624896
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5498314877226774e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.390977250297856,
      "tokens_seen": 2556755968
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5490290483068534e-05,
      "loss": 1.262,
      "theoretical_loss": 3.390963562419547,
      "tokens_seen": 2556887040
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5482266088910294e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.3909498754393494,
      "tokens_seen": 2557018112
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5474241694752054e-05,
      "loss": 1.2609,
      "theoretical_loss": 3.3909361893571583,
      "tokens_seen": 2557149184
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.546621730059381e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3909225041728686,
      "tokens_seen": 2557280256
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.545819290643557e-05,
      "loss": 1.3614,
      "theoretical_loss": 3.3909088198863757,
      "tokens_seen": 2557411328
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.545016851227733e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.3908951364975746,
      "tokens_seen": 2557542400
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.544214411811909e-05,
      "loss": 1.1932,
      "theoretical_loss": 3.3908814540063603,
      "tokens_seen": 2557673472
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.543411972396085e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.390867772412628,
      "tokens_seen": 2557804544
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.542609532980261e-05,
      "loss": 1.1903,
      "theoretical_loss": 3.390854091716273,
      "tokens_seen": 2557935616
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.541807093564437e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.3908404119171895,
      "tokens_seen": 2558066688
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.541004654148612e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.3908267330152744,
      "tokens_seen": 2558197760
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.540202214732788e-05,
      "loss": 1.175,
      "theoretical_loss": 3.390813055010421,
      "tokens_seen": 2558328832
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.539399775316964e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.3907993779025265,
      "tokens_seen": 2558459904
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.53859733590114e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.3907857016914846,
      "tokens_seen": 2558590976
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.537794896485316e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.390772026377191,
      "tokens_seen": 2558722048
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.536992457069492e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.390758351959541,
      "tokens_seen": 2558853120
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.536190017653667e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.39074467843843,
      "tokens_seen": 2558984192
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.535387578237843e-05,
      "loss": 1.1908,
      "theoretical_loss": 3.390731005813753,
      "tokens_seen": 2559115264
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.534585138822019e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.390717334085406,
      "tokens_seen": 2559246336
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.533782699406195e-05,
      "loss": 1.2673,
      "theoretical_loss": 3.3907036632532837,
      "tokens_seen": 2559377408
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.532980259990371e-05,
      "loss": 1.2262,
      "theoretical_loss": 3.3906899933172814,
      "tokens_seen": 2559508480
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.532177820574547e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.390676324277295,
      "tokens_seen": 2559639552
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5313753811587225e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.390662656133219,
      "tokens_seen": 2559770624
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5305729417428985e-05,
      "loss": 1.2269,
      "theoretical_loss": 3.3906489888849496,
      "tokens_seen": 2559901696
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5297705023270745e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.3906353225323826,
      "tokens_seen": 2560032768
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5289680629112505e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.390621657075412,
      "tokens_seen": 2560163840
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5281656234954265e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.390607992513935,
      "tokens_seen": 2560294912
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5273631840796025e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.390594328847846,
      "tokens_seen": 2560425984
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.526560744663778e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.3905806660770406,
      "tokens_seen": 2560557056
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.525758305247954e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.390567004201414,
      "tokens_seen": 2560688128
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.52495586583213e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.390553343220863,
      "tokens_seen": 2560819200
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.9932734966278076,
      "objective/train/avg_token_score": 0.0067265271209180355,
      "objective/train/docs_used": 931622,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.0733290910720825,
      "objective/train/original_loss": 1.0811107158660889,
      "objective/train/theoretical_loss": 3.3905396831352816,
      "objective/train/tokens_used": 931409376,
      "objective/train/unlikelihood_loss": 0.02481810748577118,
      "theoretical_loss": 3.3905396831352816,
      "tokens_seen": 2560950272
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.524153426416306e-05,
      "loss": 1.1652,
      "theoretical_loss": 3.3905396831352816,
      "tokens_seen": 2560950272
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.523350987000482e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.3905260239445663,
      "tokens_seen": 2561081344
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.522548547584658e-05,
      "loss": 1.1742,
      "theoretical_loss": 3.3905123656486125,
      "tokens_seen": 2561212416
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.521746108168833e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.390498708247316,
      "tokens_seen": 2561343488
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.520943668753009e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.390485051740572,
      "tokens_seen": 2561474560
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.520141229337185e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3904713961282757,
      "tokens_seen": 2561605632
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.519338789921361e-05,
      "loss": 1.1744,
      "theoretical_loss": 3.390457741410324,
      "tokens_seen": 2561736704
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.518536350505537e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.390444087586612,
      "tokens_seen": 2561867776
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.517733911089713e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.390430434657035,
      "tokens_seen": 2561998848
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.516931471673888e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.3904167826214895,
      "tokens_seen": 2562129920
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.516129032258064e-05,
      "loss": 1.1886,
      "theoretical_loss": 3.3904031314798706,
      "tokens_seen": 2562260992
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.51532659284224e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.3903894812320736,
      "tokens_seen": 2562392064
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.514524153426416e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.3903758318779955,
      "tokens_seen": 2562523136
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.513721714010592e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.3903621834175315,
      "tokens_seen": 2562654208
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.512919274594768e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.3903485358505767,
      "tokens_seen": 2562785280
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.512116835178944e-05,
      "loss": 1.1794,
      "theoretical_loss": 3.390334889177028,
      "tokens_seen": 2562916352
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5113143957631196e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.390321243396781,
      "tokens_seen": 2563047424
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5105119563472956e-05,
      "loss": 1.351,
      "theoretical_loss": 3.390307598509731,
      "tokens_seen": 2563178496
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5097095169314715e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.390293954515774,
      "tokens_seen": 2563309568
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5089070775156475e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3902803114148066,
      "tokens_seen": 2563440640
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5081046380998235e-05,
      "loss": 1.2931,
      "theoretical_loss": 3.390266669206724,
      "tokens_seen": 2563571712
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5073021986839995e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.390253027891422,
      "tokens_seen": 2563702784
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5064997592681755e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.390239387468797,
      "tokens_seen": 2563833856
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.505697319852351e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.3902257479387448,
      "tokens_seen": 2563964928
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.504894880436527e-05,
      "loss": 1.2187,
      "theoretical_loss": 3.3902121093011615,
      "tokens_seen": 2564096000
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.504092441020703e-05,
      "loss": 1.27,
      "theoretical_loss": 3.3901984715559426,
      "tokens_seen": 2564227072
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.503290001604879e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.3901848347029846,
      "tokens_seen": 2564358144
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.502487562189055e-05,
      "loss": 1.3686,
      "theoretical_loss": 3.390171198742183,
      "tokens_seen": 2564489216
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.501685122773231e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.390157563673435,
      "tokens_seen": 2564620288
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.500882683357407e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.3901439294966353,
      "tokens_seen": 2564751360
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.500080243941583e-05,
      "loss": 1.2007,
      "theoretical_loss": 3.3901302962116806,
      "tokens_seen": 2564882432
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.499277804525758e-05,
      "loss": 1.2941,
      "theoretical_loss": 3.3901166638184668,
      "tokens_seen": 2565013504
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.498475365109934e-05,
      "loss": 1.3672,
      "theoretical_loss": 3.3901030323168904,
      "tokens_seen": 2565144576
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.49767292569411e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.390089401706847,
      "tokens_seen": 2565275648
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.496870486278286e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.390075771988233,
      "tokens_seen": 2565406720
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.496068046862462e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.390062143160945,
      "tokens_seen": 2565537792
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.495265607446638e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.390048515224878,
      "tokens_seen": 2565668864
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.494463168030814e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.390034888179929,
      "tokens_seen": 2565799936
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.493660728614989e-05,
      "loss": 1.1741,
      "theoretical_loss": 3.3900212620259946,
      "tokens_seen": 2565931008
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.492858289199165e-05,
      "loss": 1.2732,
      "theoretical_loss": 3.3900076367629706,
      "tokens_seen": 2566062080
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.492055849783341e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.389994012390753,
      "tokens_seen": 2566193152
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.491253410367517e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.389980388909238,
      "tokens_seen": 2566324224
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.490450970951693e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.389966766318323,
      "tokens_seen": 2566455296
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.489648531535869e-05,
      "loss": 1.2233,
      "theoretical_loss": 3.3899531446179028,
      "tokens_seen": 2566586368
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.488846092120045e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3899395238078744,
      "tokens_seen": 2566717440
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4880436527042206e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.389925903888134,
      "tokens_seen": 2566848512
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4872412132883966e-05,
      "loss": 1.3419,
      "theoretical_loss": 3.3899122848585783,
      "tokens_seen": 2566979584
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4864387738725726e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.3898986667191036,
      "tokens_seen": 2567110656
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4856363344567486e-05,
      "loss": 1.3043,
      "theoretical_loss": 3.389885049469606,
      "tokens_seen": 2567241728
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4848338950409245e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.389871433109982,
      "tokens_seen": 2567372800
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.990434467792511,
      "objective/train/avg_token_score": 0.009565500542521477,
      "objective/train/docs_used": 933971,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2258433103561401,
      "objective/train/original_loss": 1.2393816709518433,
      "objective/train/theoretical_loss": 3.389857817640128,
      "objective/train/tokens_used": 937962976,
      "objective/train/unlikelihood_loss": 0.025451118126511574,
      "theoretical_loss": 3.389857817640128,
      "tokens_seen": 2567503872
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4840314556251005e-05,
      "loss": 1.2286,
      "theoretical_loss": 3.389857817640128,
      "tokens_seen": 2567503872
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4832290162092765e-05,
      "loss": 1.3781,
      "theoretical_loss": 3.3898442030599405,
      "tokens_seen": 2567634944
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4824265767934525e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.389830589369316,
      "tokens_seen": 2567766016
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.481624137377628e-05,
      "loss": 1.3478,
      "theoretical_loss": 3.389816976568151,
      "tokens_seen": 2567897088
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.480821697961804e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.3898033646563417,
      "tokens_seen": 2568028160
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.48001925854598e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.389789753633785,
      "tokens_seen": 2568159232
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.479216819130156e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.3897761435003773,
      "tokens_seen": 2568290304
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.478414379714332e-05,
      "loss": 1.3435,
      "theoretical_loss": 3.3897625342560147,
      "tokens_seen": 2568421376
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.477611940298508e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.3897489259005944,
      "tokens_seen": 2568552448
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.476809500882684e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.389735318434013,
      "tokens_seen": 2568683520
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.476007061466859e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.389721711856166,
      "tokens_seen": 2568814592
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.475204622051035e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.3897081061669514,
      "tokens_seen": 2568945664
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.474402182635211e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.3896945013662654,
      "tokens_seen": 2569076736
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.473599743219387e-05,
      "loss": 1.3119,
      "theoretical_loss": 3.3896808974540042,
      "tokens_seen": 2569207808
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.472797303803563e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.389667294430065,
      "tokens_seen": 2569338880
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.471994864387739e-05,
      "loss": 1.2227,
      "theoretical_loss": 3.389653692294344,
      "tokens_seen": 2569469952
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.471192424971915e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.3896400910467386,
      "tokens_seen": 2569601024
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.47038998555609e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.389626490687145,
      "tokens_seen": 2569732096
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.469587546140266e-05,
      "loss": 1.3938,
      "theoretical_loss": 3.3896128912154593,
      "tokens_seen": 2569863168
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.468785106724442e-05,
      "loss": 1.303,
      "theoretical_loss": 3.38959929263158,
      "tokens_seen": 2569994240
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.467982667308618e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.389585694935402,
      "tokens_seen": 2570125312
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.467180227892794e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.389572098126823,
      "tokens_seen": 2570256384
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.46637778847697e-05,
      "loss": 1.3323,
      "theoretical_loss": 3.3895585022057397,
      "tokens_seen": 2570387456
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.465575349061146e-05,
      "loss": 1.3367,
      "theoretical_loss": 3.3895449071720494,
      "tokens_seen": 2570518528
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.464772909645322e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.389531313025648,
      "tokens_seen": 2570649600
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4639704702294976e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.389517719766433,
      "tokens_seen": 2570780672
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4631680308136736e-05,
      "loss": 1.4279,
      "theoretical_loss": 3.389504127394301,
      "tokens_seen": 2570911744
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4623655913978496e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.389490535909149,
      "tokens_seen": 2571042816
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4615631519820256e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.3894769453108737,
      "tokens_seen": 2571173888
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4607607125662015e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.3894633555993723,
      "tokens_seen": 2571304960
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4599582731503775e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.3894497667745416,
      "tokens_seen": 2571436032
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4591558337345535e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.389436178836279,
      "tokens_seen": 2571567104
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.458353394318729e-05,
      "loss": 1.3508,
      "theoretical_loss": 3.3894225917844807,
      "tokens_seen": 2571698176
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.457550954902905e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.389409005619044,
      "tokens_seen": 2571829248
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.456748515487081e-05,
      "loss": 1.3458,
      "theoretical_loss": 3.3893954203398664,
      "tokens_seen": 2571960320
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.455946076071257e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.3893818359468444,
      "tokens_seen": 2572091392
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.455143636655433e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.3893682524398745,
      "tokens_seen": 2572222464
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.454341197239609e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.389354669818855,
      "tokens_seen": 2572353536
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.453538757823785e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.3893410880836825,
      "tokens_seen": 2572484608
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.452736318407961e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3893275072342535,
      "tokens_seen": 2572615680
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.451933878992136e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.3893139272704658,
      "tokens_seen": 2572746752
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.451131439576312e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.3893003481922164,
      "tokens_seen": 2572877824
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.450329000160488e-05,
      "loss": 1.2322,
      "theoretical_loss": 3.3892867699994023,
      "tokens_seen": 2573008896
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.449526560744664e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.3892731926919204,
      "tokens_seen": 2573139968
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.44872412132884e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.3892596162696687,
      "tokens_seen": 2573271040
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.447921681913016e-05,
      "loss": 1.306,
      "theoretical_loss": 3.3892460407325435,
      "tokens_seen": 2573402112
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.447119242497192e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.389232466080443,
      "tokens_seen": 2573533184
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4463168030813673e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.3892188923132633,
      "tokens_seen": 2573664256
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.445514363665543e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.3892053194309026,
      "tokens_seen": 2573795328
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.444711924249719e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.389191747433258,
      "tokens_seen": 2573926400
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.9935245513916016,
      "objective/train/avg_token_score": 0.006475494243204594,
      "objective/train/docs_used": 936262,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2518751621246338,
      "objective/train/original_loss": 1.2601573467254639,
      "objective/train/theoretical_loss": 3.3891781763202258,
      "objective/train/tokens_used": 944516576,
      "objective/train/unlikelihood_loss": 0.018883198499679565,
      "theoretical_loss": 3.3891781763202258,
      "tokens_seen": 2574057472
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.443909484833895e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3891781763202258,
      "tokens_seen": 2574057472
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.443107045418071e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.3891646060917044,
      "tokens_seen": 2574188544
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.442304606002247e-05,
      "loss": 1.3617,
      "theoretical_loss": 3.3891510367475908,
      "tokens_seen": 2574319616
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.441502166586423e-05,
      "loss": 1.259,
      "theoretical_loss": 3.389137468287782,
      "tokens_seen": 2574450688
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4406997271705986e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.389123900712176,
      "tokens_seen": 2574581760
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4398972877547746e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.38911033402067,
      "tokens_seen": 2574712832
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4390948483389506e-05,
      "loss": 1.3644,
      "theoretical_loss": 3.389096768213161,
      "tokens_seen": 2574843904
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4382924089231266e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3890832032895464,
      "tokens_seen": 2574974976
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4374899695073026e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3890696392497244,
      "tokens_seen": 2575106048
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4366875300914786e-05,
      "loss": 1.2365,
      "theoretical_loss": 3.3890560760935915,
      "tokens_seen": 2575237120
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4358850906756545e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.3890425138210456,
      "tokens_seen": 2575368192
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4350826512598305e-05,
      "loss": 1.3402,
      "theoretical_loss": 3.3890289524319837,
      "tokens_seen": 2575499264
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.434280211844006e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.3890153919263044,
      "tokens_seen": 2575630336
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.433477772428182e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.389001832303904,
      "tokens_seen": 2575761408
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.432675333012358e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.3889882735646806,
      "tokens_seen": 2575892480
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.431872893596534e-05,
      "loss": 1.2388,
      "theoretical_loss": 3.388974715708532,
      "tokens_seen": 2576023552
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.43107045418071e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.3889611587353556,
      "tokens_seen": 2576154624
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.430268014764886e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.388947602645048,
      "tokens_seen": 2576285696
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.429465575349062e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.388934047437508,
      "tokens_seen": 2576416768
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.428663135933237e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.388920493112633,
      "tokens_seen": 2576547840
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.427860696517413e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.3889069396703206,
      "tokens_seen": 2576678912
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.427058257101589e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.3888933871104676,
      "tokens_seen": 2576809984
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.426255817685765e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.3888798354329728,
      "tokens_seen": 2576941056
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.425453378269941e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.388866284637733,
      "tokens_seen": 2577072128
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.424650938854117e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.388852734724647,
      "tokens_seen": 2577203200
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.423848499438293e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.3888391856936115,
      "tokens_seen": 2577334272
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4230460600224684e-05,
      "loss": 1.1601,
      "theoretical_loss": 3.3888256375445245,
      "tokens_seen": 2577465344
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4222436206066444e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.3888120902772836,
      "tokens_seen": 2577596416
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4214411811908203e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.388798543891787,
      "tokens_seen": 2577727488
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.420638741774996e-05,
      "loss": 1.3445,
      "theoretical_loss": 3.388784998387932,
      "tokens_seen": 2577858560
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.419836302359172e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.388771453765617,
      "tokens_seen": 2577989632
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.419033862943348e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.388757910024739,
      "tokens_seen": 2578120704
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.418231423527524e-05,
      "loss": 1.3728,
      "theoretical_loss": 3.388744367165197,
      "tokens_seen": 2578251776
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4174289841117e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.388730825186887,
      "tokens_seen": 2578382848
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4166265446958756e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.388717284089709,
      "tokens_seen": 2578513920
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4158241052800516e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.3887037438735597,
      "tokens_seen": 2578644992
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4150216658642276e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.388690204538337,
      "tokens_seen": 2578776064
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4142192264484036e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.3886766660839385,
      "tokens_seen": 2578907136
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4134167870325796e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.388663128510263,
      "tokens_seen": 2579038208
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4126143476167556e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3886495918172086,
      "tokens_seen": 2579169280
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.4118119082009316e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.3886360560046724,
      "tokens_seen": 2579300352
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.411009468785107e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.3886225210725525,
      "tokens_seen": 2579431424
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.410207029369283e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.3886089870207474,
      "tokens_seen": 2579562496
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.409404589953459e-05,
      "loss": 1.2673,
      "theoretical_loss": 3.3885954538491543,
      "tokens_seen": 2579693568
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.408602150537635e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.3885819215576722,
      "tokens_seen": 2579824640
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.407799711121811e-05,
      "loss": 1.2127,
      "theoretical_loss": 3.388568390146199,
      "tokens_seen": 2579955712
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.406997271705987e-05,
      "loss": 1.1773,
      "theoretical_loss": 3.388554859614632,
      "tokens_seen": 2580086784
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.406194832290163e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.3885413299628704,
      "tokens_seen": 2580217856
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.405392392874339e-05,
      "loss": 1.2233,
      "theoretical_loss": 3.3885278011908113,
      "tokens_seen": 2580348928
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.404589953458514e-05,
      "loss": 1.3808,
      "theoretical_loss": 3.388514273298353,
      "tokens_seen": 2580480000
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.9910797476768494,
      "objective/train/avg_token_score": 0.008920260705053806,
      "objective/train/docs_used": 938550,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2044166326522827,
      "objective/train/original_loss": 1.216614842414856,
      "objective/train/theoretical_loss": 3.388500746285394,
      "objective/train/tokens_used": 951070176,
      "objective/train/unlikelihood_loss": 0.025076264515519142,
      "theoretical_loss": 3.388500746285394,
      "tokens_seen": 2580611072
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.40378751404269e-05,
      "loss": 1.2118,
      "theoretical_loss": 3.388500746285394,
      "tokens_seen": 2580611072
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.402985074626866e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.3884872201518323,
      "tokens_seen": 2580742144
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.402182635211042e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3884736948975664,
      "tokens_seen": 2580873216
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.401380195795218e-05,
      "loss": 1.286,
      "theoretical_loss": 3.3884601705224937,
      "tokens_seen": 2581004288
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.400577756379394e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3884466470265133,
      "tokens_seen": 2581135360
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.3997753169635694e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3884331244095227,
      "tokens_seen": 2581266432
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.3989728775477454e-05,
      "loss": 1.3328,
      "theoretical_loss": 3.38841960267142,
      "tokens_seen": 2581397504
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.3981704381319214e-05,
      "loss": 1.1939,
      "theoretical_loss": 3.388406081812105,
      "tokens_seen": 2581528576
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.3973679987160974e-05,
      "loss": 1.3242,
      "theoretical_loss": 3.3883925618314743,
      "tokens_seen": 2581659648
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.3965655593002733e-05,
      "loss": 1.3874,
      "theoretical_loss": 3.3883790427294267,
      "tokens_seen": 2581790720
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.395763119884449e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.3883655245058604,
      "tokens_seen": 2581921792
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.3949606804686246e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.388352007160674,
      "tokens_seen": 2582052864
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.3941582410528006e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.3883384906937657,
      "tokens_seen": 2582183936
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3933558016369766e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.388324975105034,
      "tokens_seen": 2582315008
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3925533622211526e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.3883114603943767,
      "tokens_seen": 2582446080
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3917509228053286e-05,
      "loss": 1.325,
      "theoretical_loss": 3.388297946561693,
      "tokens_seen": 2582577152
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3909484833895046e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.388284433606881,
      "tokens_seen": 2582708224
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.39014604397368e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.388270921529839,
      "tokens_seen": 2582839296
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.389343604557856e-05,
      "loss": 1.196,
      "theoretical_loss": 3.3882574103304655,
      "tokens_seen": 2582970368
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.388541165142032e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.388243900008659,
      "tokens_seen": 2583101440
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.387738725726208e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.388230390564318,
      "tokens_seen": 2583232512
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.386936286310384e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.388216881997341,
      "tokens_seen": 2583363584
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.38613384689456e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.388203374307626,
      "tokens_seen": 2583494656
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.385331407478735e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.3881898674950723,
      "tokens_seen": 2583625728
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.384528968062911e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.3881763615595784,
      "tokens_seen": 2583756800
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.383726528647087e-05,
      "loss": 1.2256,
      "theoretical_loss": 3.388162856501042,
      "tokens_seen": 2583887872
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.382924089231263e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.3881493523193624,
      "tokens_seen": 2584018944
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.382121649815439e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.3881358490144384,
      "tokens_seen": 2584150016
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.381319210399615e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.3881223465861683,
      "tokens_seen": 2584281088
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3805167709837904e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3881088450344503,
      "tokens_seen": 2584412160
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3797143315679664e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.3880953443591837,
      "tokens_seen": 2584543232
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3789118921521424e-05,
      "loss": 1.221,
      "theoretical_loss": 3.3880818445602667,
      "tokens_seen": 2584674304
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3781094527363184e-05,
      "loss": 1.2058,
      "theoretical_loss": 3.388068345637598,
      "tokens_seen": 2584805376
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3773070133204944e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.388054847591077,
      "tokens_seen": 2584936448
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3765045739046704e-05,
      "loss": 1.1997,
      "theoretical_loss": 3.3880413504206017,
      "tokens_seen": 2585067520
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.375702134488846e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.388027854126071,
      "tokens_seen": 2585198592
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.374899695073022e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3880143587073834,
      "tokens_seen": 2585329664
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.374097255657198e-05,
      "loss": 1.3464,
      "theoretical_loss": 3.3880008641644386,
      "tokens_seen": 2585460736
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.373294816241374e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.3879873704971337,
      "tokens_seen": 2585591808
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.37249237682555e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.3879738777053694,
      "tokens_seen": 2585722880
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.371689937409726e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.387960385789043,
      "tokens_seen": 2585853952
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3708874979939017e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.387946894748054,
      "tokens_seen": 2585985024
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3700850585780776e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3879334045823017,
      "tokens_seen": 2586116096
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.369282619162253e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.387919915291684,
      "tokens_seen": 2586247168
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.368480179746429e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.3879064268761003,
      "tokens_seen": 2586378240
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.367677740330605e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3878929393354493,
      "tokens_seen": 2586509312
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.366875300914781e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.3878794526696305,
      "tokens_seen": 2586640384
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.366072861498957e-05,
      "loss": 1.4086,
      "theoretical_loss": 3.387865966878542,
      "tokens_seen": 2586771456
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.365270422083133e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.387852481962083,
      "tokens_seen": 2586902528
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.364467982667309e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.387838997920153,
      "tokens_seen": 2587033600
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.9919148683547974,
      "objective/train/avg_token_score": 0.008085177280008793,
      "objective/train/docs_used": 940829,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2944910526275635,
      "objective/train/original_loss": 1.306321382522583,
      "objective/train/theoretical_loss": 3.38782551475265,
      "objective/train/tokens_used": 957623776,
      "objective/train/unlikelihood_loss": 0.020549198612570763,
      "theoretical_loss": 3.38782551475265,
      "tokens_seen": 2587164672
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.363665543251484e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.38782551475265,
      "tokens_seen": 2587164672
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.36286310383566e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.387812032459474,
      "tokens_seen": 2587295744
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.362060664419836e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.3877985510405235,
      "tokens_seen": 2587426816
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.361258225004012e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.3877850704956978,
      "tokens_seen": 2587557888
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.360455785588188e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.3877715908248955,
      "tokens_seen": 2587688960
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.359653346172364e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.387758112028016,
      "tokens_seen": 2587820032
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.35885090675654e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.3877446341049584,
      "tokens_seen": 2587951104
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.358048467340716e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.3877311570556214,
      "tokens_seen": 2588082176
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3572460279248915e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.3877176808799043,
      "tokens_seen": 2588213248
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3564435885090675e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.3877042055777067,
      "tokens_seen": 2588344320
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3556411490932434e-05,
      "loss": 1.3198,
      "theoretical_loss": 3.3876907311489273,
      "tokens_seen": 2588475392
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3548387096774194e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3876772575934657,
      "tokens_seen": 2588606464
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3540362702615954e-05,
      "loss": 1.1997,
      "theoretical_loss": 3.3876637849112203,
      "tokens_seen": 2588737536
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3532338308457714e-05,
      "loss": 1.3641,
      "theoretical_loss": 3.3876503131020907,
      "tokens_seen": 2588868608
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3524313914299474e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.3876368421659766,
      "tokens_seen": 2588999680
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.351628952014123e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.3876233721027766,
      "tokens_seen": 2589130752
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.350826512598299e-05,
      "loss": 1.257,
      "theoretical_loss": 3.3876099029123896,
      "tokens_seen": 2589261824
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.350024073182475e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.387596434594716,
      "tokens_seen": 2589392896
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.349221633766651e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.3875829671496547,
      "tokens_seen": 2589523968
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.348419194350827e-05,
      "loss": 1.3358,
      "theoretical_loss": 3.3875695005771043,
      "tokens_seen": 2589655040
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.347616754935003e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.387556034876965,
      "tokens_seen": 2589786112
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3468143155191787e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.3875425700491353,
      "tokens_seen": 2589917184
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.346011876103354e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.387529106093515,
      "tokens_seen": 2590048256
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.34520943668753e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.387515643010004,
      "tokens_seen": 2590179328
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.344406997271706e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.3875021807985006,
      "tokens_seen": 2590310400
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.343604557855882e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.387488719458905,
      "tokens_seen": 2590441472
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.342802118440058e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.387475258991116,
      "tokens_seen": 2590572544
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.341999679024234e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.387461799395034,
      "tokens_seen": 2590703616
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.34119723960841e-05,
      "loss": 1.2357,
      "theoretical_loss": 3.3874483406705576,
      "tokens_seen": 2590834688
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.340394800192586e-05,
      "loss": 1.2725,
      "theoretical_loss": 3.387434882817586,
      "tokens_seen": 2590965760
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.339592360776761e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.3874214258360196,
      "tokens_seen": 2591096832
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.338789921360937e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.3874079697257575,
      "tokens_seen": 2591227904
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.337987481945113e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.387394514486699,
      "tokens_seen": 2591358976
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.337185042529289e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.387381060118744,
      "tokens_seen": 2591490048
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.336382603113465e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.387367606621792,
      "tokens_seen": 2591621120
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.335580163697641e-05,
      "loss": 1.2414,
      "theoretical_loss": 3.3873541539957426,
      "tokens_seen": 2591752192
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.334777724281817e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.387340702240495,
      "tokens_seen": 2591883264
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3339752848659925e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.387327251355949,
      "tokens_seen": 2592014336
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3331728454501685e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.387313801342004,
      "tokens_seen": 2592145408
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3323704060343445e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.3873003521985603,
      "tokens_seen": 2592276480
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3315679666185204e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.387286903925517,
      "tokens_seen": 2592407552
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3307655272026964e-05,
      "loss": 1.1962,
      "theoretical_loss": 3.387273456522774,
      "tokens_seen": 2592538624
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3299630877868724e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.387260009990231,
      "tokens_seen": 2592669696
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3291606483710484e-05,
      "loss": 1.253,
      "theoretical_loss": 3.387246564327787,
      "tokens_seen": 2592800768
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.328358208955224e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3872331195353427,
      "tokens_seen": 2592931840
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3275557695394e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.3872196756127972,
      "tokens_seen": 2593062912
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.326753330123576e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.387206232560051,
      "tokens_seen": 2593193984
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.325950890707752e-05,
      "loss": 1.2648,
      "theoretical_loss": 3.3871927903770027,
      "tokens_seen": 2593325056
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.325148451291928e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.3871793490635533,
      "tokens_seen": 2593456128
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.324346011876104e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.3871659086196018,
      "tokens_seen": 2593587200
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.9943293929100037,
      "objective/train/avg_token_score": 0.005670545622706413,
      "objective/train/docs_used": 943233,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.254280924797058,
      "objective/train/original_loss": 1.2618498802185059,
      "objective/train/theoretical_loss": 3.387152469045048,
      "objective/train/tokens_used": 964177376,
      "objective/train/unlikelihood_loss": 0.0170451533049345,
      "theoretical_loss": 3.387152469045048,
      "tokens_seen": 2593718272
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.32354357246028e-05,
      "loss": 1.2227,
      "theoretical_loss": 3.387152469045048,
      "tokens_seen": 2593718272
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.322741133044456e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.3871390303397924,
      "tokens_seen": 2593849344
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.321938693628631e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.3871255925037342,
      "tokens_seen": 2593980416
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.321136254212807e-05,
      "loss": 1.3583,
      "theoretical_loss": 3.3871121555367734,
      "tokens_seen": 2594111488
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.320333814796983e-05,
      "loss": 1.2182,
      "theoretical_loss": 3.3870987194388102,
      "tokens_seen": 2594242560
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.319531375381159e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.387085284209744,
      "tokens_seen": 2594373632
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.318728935965335e-05,
      "loss": 1.3039,
      "theoretical_loss": 3.3870718498494754,
      "tokens_seen": 2594504704
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.317926496549511e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.3870584163579034,
      "tokens_seen": 2594635776
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.317124057133687e-05,
      "loss": 1.1818,
      "theoretical_loss": 3.387044983734929,
      "tokens_seen": 2594766848
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.316321617717862e-05,
      "loss": 1.2179,
      "theoretical_loss": 3.3870315519804515,
      "tokens_seen": 2594897920
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.315519178302038e-05,
      "loss": 1.219,
      "theoretical_loss": 3.387018121094371,
      "tokens_seen": 2595028992
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.314716738886214e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.387004691076587,
      "tokens_seen": 2595160064
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.31391429947039e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.3869912619270006,
      "tokens_seen": 2595291136
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.313111860054566e-05,
      "loss": 1.3443,
      "theoretical_loss": 3.386977833645511,
      "tokens_seen": 2595422208
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.312309420638742e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.386964406232019,
      "tokens_seen": 2595553280
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.311506981222918e-05,
      "loss": 1.2117,
      "theoretical_loss": 3.3869509796864237,
      "tokens_seen": 2595684352
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.310704541807094e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.3869375540086257,
      "tokens_seen": 2595815424
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3099021023912695e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.3869241291985253,
      "tokens_seen": 2595946496
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3090996629754455e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.386910705256022,
      "tokens_seen": 2596077568
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3082972235596215e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.3868972821810166,
      "tokens_seen": 2596208640
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3074947841437975e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.3868838599734086,
      "tokens_seen": 2596339712
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3066923447279734e-05,
      "loss": 1.1969,
      "theoretical_loss": 3.386870438633099,
      "tokens_seen": 2596470784
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3058899053121494e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3868570181599873,
      "tokens_seen": 2596601856
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.3050874658963254e-05,
      "loss": 1.3597,
      "theoretical_loss": 3.3868435985539738,
      "tokens_seen": 2596732928
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.304285026480501e-05,
      "loss": 1.3496,
      "theoretical_loss": 3.3868301798149587,
      "tokens_seen": 2596864000
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.303482587064677e-05,
      "loss": 1.3341,
      "theoretical_loss": 3.3868167619428426,
      "tokens_seen": 2596995072
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.302680147648853e-05,
      "loss": 1.2473,
      "theoretical_loss": 3.386803344937525,
      "tokens_seen": 2597126144
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.301877708233029e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.3867899287989074,
      "tokens_seen": 2597257216
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.301075268817205e-05,
      "loss": 1.2573,
      "theoretical_loss": 3.386776513526889,
      "tokens_seen": 2597388288
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.300272829401381e-05,
      "loss": 1.3774,
      "theoretical_loss": 3.38676309912137,
      "tokens_seen": 2597519360
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.299470389985557e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.3867496855822514,
      "tokens_seen": 2597650432
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.298667950569732e-05,
      "loss": 1.34,
      "theoretical_loss": 3.3867362729094332,
      "tokens_seen": 2597781504
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.297865511153908e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.386722861102816,
      "tokens_seen": 2597912576
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.297063071738084e-05,
      "loss": 1.358,
      "theoretical_loss": 3.3867094501623,
      "tokens_seen": 2598043648
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.29626063232226e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.3866960400877852,
      "tokens_seen": 2598174720
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.295458192906436e-05,
      "loss": 1.3592,
      "theoretical_loss": 3.3866826308791724,
      "tokens_seen": 2598305792
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.294655753490612e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.386669222536362,
      "tokens_seen": 2598436864
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.293853314074788e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.386655815059255,
      "tokens_seen": 2598567936
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.293050874658964e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.3866424084477504,
      "tokens_seen": 2598699008
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.292248435243139e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.38662900270175,
      "tokens_seen": 2598830080
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.291445995827315e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.3866155978211534,
      "tokens_seen": 2598961152
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.290643556411491e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.386602193805862,
      "tokens_seen": 2599092224
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.289841116995667e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.3865887906557752,
      "tokens_seen": 2599223296
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.289038677579843e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.3865753883707943,
      "tokens_seen": 2599354368
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.288236238164019e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.3865619869508197,
      "tokens_seen": 2599485440
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.287433798748195e-05,
      "loss": 1.248,
      "theoretical_loss": 3.386548586395752,
      "tokens_seen": 2599616512
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2866313593323705e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.3865351867054914,
      "tokens_seen": 2599747584
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2858289199165465e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.3865217878799387,
      "tokens_seen": 2599878656
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2850264805007225e-05,
      "loss": 1.3333,
      "theoretical_loss": 3.386508389918995,
      "tokens_seen": 2600009728
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2842240410848985e-05,
      "loss": 1.3476,
      "theoretical_loss": 3.38649499282256,
      "tokens_seen": 2600140800
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.9934045076370239,
      "objective/train/avg_token_score": 0.006595509126782417,
      "objective/train/docs_used": 945762,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2631226778030396,
      "objective/train/original_loss": 1.271596908569336,
      "objective/train/theoretical_loss": 3.386481596590535,
      "objective/train/tokens_used": 970730976,
      "objective/train/unlikelihood_loss": 0.01985674351453781,
      "theoretical_loss": 3.386481596590535,
      "tokens_seen": 2600271872
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2834216016690745e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.386481596590535,
      "tokens_seen": 2600271872
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2826191622532505e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.3864682012228204,
      "tokens_seen": 2600402944
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2818167228374264e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.386454806719317,
      "tokens_seen": 2600534016
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.281014283421602e-05,
      "loss": 1.3866,
      "theoretical_loss": 3.386441413079926,
      "tokens_seen": 2600665088
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.280211844005778e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.386428020304547,
      "tokens_seen": 2600796160
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.279409404589954e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.3864146283930814,
      "tokens_seen": 2600927232
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.27860696517413e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.38640123734543,
      "tokens_seen": 2601058304
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.277804525758306e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3863878471614934,
      "tokens_seen": 2601189376
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.277002086342482e-05,
      "loss": 1.3576,
      "theoretical_loss": 3.386374457841172,
      "tokens_seen": 2601320448
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.276199646926658e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.3863610693843675,
      "tokens_seen": 2601451520
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.275397207510834e-05,
      "loss": 1.226,
      "theoretical_loss": 3.38634768179098,
      "tokens_seen": 2601582592
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.274594768095009e-05,
      "loss": 1.3694,
      "theoretical_loss": 3.38633429506091,
      "tokens_seen": 2601713664
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.273792328679185e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.386320909194059,
      "tokens_seen": 2601844736
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.272989889263361e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.386307524190328,
      "tokens_seen": 2601975808
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.272187449847537e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.3862941400496176,
      "tokens_seen": 2602106880
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.271385010431713e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.3862807567718285,
      "tokens_seen": 2602237952
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.270582571015889e-05,
      "loss": 1.2104,
      "theoretical_loss": 3.3862673743568616,
      "tokens_seen": 2602369024
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.269780131600065e-05,
      "loss": 1.269,
      "theoretical_loss": 3.386253992804618,
      "tokens_seen": 2602500096
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.26897769218424e-05,
      "loss": 1.3256,
      "theoretical_loss": 3.3862406121149986,
      "tokens_seen": 2602631168
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.268175252768416e-05,
      "loss": 1.331,
      "theoretical_loss": 3.386227232287905,
      "tokens_seen": 2602762240
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.267372813352592e-05,
      "loss": 1.281,
      "theoretical_loss": 3.3862138533232367,
      "tokens_seen": 2602893312
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.266570373936768e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.3862004752208956,
      "tokens_seen": 2603024384
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.265767934520944e-05,
      "loss": 1.298,
      "theoretical_loss": 3.386187097980783,
      "tokens_seen": 2603155456
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.26496549510512e-05,
      "loss": 1.3595,
      "theoretical_loss": 3.386173721602799,
      "tokens_seen": 2603286528
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2641630556892955e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3861603460868457,
      "tokens_seen": 2603417600
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2633606162734715e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.386146971432823,
      "tokens_seen": 2603548672
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2625581768576475e-05,
      "loss": 1.3996,
      "theoretical_loss": 3.386133597640633,
      "tokens_seen": 2603679744
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2617557374418235e-05,
      "loss": 1.3544,
      "theoretical_loss": 3.3861202247101767,
      "tokens_seen": 2603810816
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2609532980259995e-05,
      "loss": 1.2334,
      "theoretical_loss": 3.3861068526413542,
      "tokens_seen": 2603941888
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2601508586101755e-05,
      "loss": 1.4899,
      "theoretical_loss": 3.3860934814340675,
      "tokens_seen": 2604072960
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.259348419194351e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.386080111088218,
      "tokens_seen": 2604204032
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.258545979778527e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.386066741603706,
      "tokens_seen": 2604335104
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.257743540362703e-05,
      "loss": 1.223,
      "theoretical_loss": 3.3860533729804327,
      "tokens_seen": 2604466176
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.256941100946879e-05,
      "loss": 1.3447,
      "theoretical_loss": 3.3860400052183,
      "tokens_seen": 2604597248
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.256138661531055e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.3860266383172086,
      "tokens_seen": 2604728320
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.255336222115231e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.3860132722770597,
      "tokens_seen": 2604859392
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.254533782699406e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.385999907097755,
      "tokens_seen": 2604990464
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.253731343283582e-05,
      "loss": 1.3929,
      "theoretical_loss": 3.3859865427791953,
      "tokens_seen": 2605121536
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.252928903867758e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.385973179321282,
      "tokens_seen": 2605252608
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.252126464451934e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.385959816723916,
      "tokens_seen": 2605383680
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.25132402503611e-05,
      "loss": 1.291,
      "theoretical_loss": 3.3859464549869993,
      "tokens_seen": 2605514752
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.250521585620286e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.385933094110433,
      "tokens_seen": 2605645824
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.249719146204461e-05,
      "loss": 1.2223,
      "theoretical_loss": 3.385919734094118,
      "tokens_seen": 2605776896
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.248916706788637e-05,
      "loss": 1.2065,
      "theoretical_loss": 3.3859063749379557,
      "tokens_seen": 2605907968
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.248114267372813e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.385893016641848,
      "tokens_seen": 2606039040
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.247311827956989e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.385879659205696,
      "tokens_seen": 2606170112
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.246509388541165e-05,
      "loss": 1.3624,
      "theoretical_loss": 3.385866302629401,
      "tokens_seen": 2606301184
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.245706949125341e-05,
      "loss": 1.3547,
      "theoretical_loss": 3.3858529469128644,
      "tokens_seen": 2606432256
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2449045097095166e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.385839592055988,
      "tokens_seen": 2606563328
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2441020702936926e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.3858262380586726,
      "tokens_seen": 2606694400
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.9930676221847534,
      "objective/train/avg_token_score": 0.006932341493666172,
      "objective/train/docs_used": 948007,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3535586595535278,
      "objective/train/original_loss": 1.3637501001358032,
      "objective/train/theoretical_loss": 3.38581288492082,
      "objective/train/tokens_used": 977284576,
      "objective/train/unlikelihood_loss": 0.017819629982113838,
      "theoretical_loss": 3.38581288492082,
      "tokens_seen": 2606825472
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2432996308778686e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.38581288492082,
      "tokens_seen": 2606825472
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2424971914620446e-05,
      "loss": 1.2446,
      "theoretical_loss": 3.385799532642332,
      "tokens_seen": 2606956544
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2416947520462206e-05,
      "loss": 1.2311,
      "theoretical_loss": 3.38578618122311,
      "tokens_seen": 2607087616
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2408923126303965e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.3857728306630546,
      "tokens_seen": 2607218688
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2400898732145725e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.385759480962068,
      "tokens_seen": 2607349760
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.239287433798748e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.385746132120052,
      "tokens_seen": 2607480832
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.238484994382924e-05,
      "loss": 1.3423,
      "theoretical_loss": 3.3857327841369083,
      "tokens_seen": 2607611904
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2376825549671e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.3857194370125376,
      "tokens_seen": 2607742976
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.236880115551276e-05,
      "loss": 1.3801,
      "theoretical_loss": 3.385706090746842,
      "tokens_seen": 2607874048
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.236077676135452e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.385692745339723,
      "tokens_seen": 2608005120
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.235275236719628e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3856794007910826,
      "tokens_seen": 2608136192
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.234472797303804e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.385666057100822,
      "tokens_seen": 2608267264
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.233670357887979e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.3856527142688426,
      "tokens_seen": 2608398336
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.232867918472155e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.3856393722950466,
      "tokens_seen": 2608529408
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.232065479056331e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.385626031179336,
      "tokens_seen": 2608660480
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.231263039640507e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3856126909216115,
      "tokens_seen": 2608791552
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.230460600224683e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.3855993515217757,
      "tokens_seen": 2608922624
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.229658160808859e-05,
      "loss": 1.2078,
      "theoretical_loss": 3.3855860129797297,
      "tokens_seen": 2609053696
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.228855721393035e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.385572675295376,
      "tokens_seen": 2609184768
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.228053281977211e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.3855593384686156,
      "tokens_seen": 2609315840
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2272508425613864e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.3855460024993502,
      "tokens_seen": 2609446912
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2264484031455623e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.3855326673874826,
      "tokens_seen": 2609577984
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.225645963729738e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.3855193331329136,
      "tokens_seen": 2609709056
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.224843524313914e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.3855059997355457,
      "tokens_seen": 2609840128
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.22404108489809e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.3854926671952805,
      "tokens_seen": 2609971200
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.223238645482266e-05,
      "loss": 1.3691,
      "theoretical_loss": 3.3854793355120196,
      "tokens_seen": 2610102272
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.222436206066442e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.3854660046856653,
      "tokens_seen": 2610233344
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2216337666506176e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.3854526747161193,
      "tokens_seen": 2610364416
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2208313272347936e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3854393456032836,
      "tokens_seen": 2610495488
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2200288878189696e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.38542601734706,
      "tokens_seen": 2610626560
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2192264484031456e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.3854126899473505,
      "tokens_seen": 2610757632
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2184240089873216e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.385399363404057,
      "tokens_seen": 2610888704
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2176215695714976e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.3853860377170815,
      "tokens_seen": 2611019776
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2168191301556736e-05,
      "loss": 1.22,
      "theoretical_loss": 3.3853727128863262,
      "tokens_seen": 2611150848
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2160166907398495e-05,
      "loss": 1.337,
      "theoretical_loss": 3.3853593889116933,
      "tokens_seen": 2611281920
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.215214251324025e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.385346065793084,
      "tokens_seen": 2611412992
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.214411811908201e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.385332743530401,
      "tokens_seen": 2611544064
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.213609372492377e-05,
      "loss": 1.1873,
      "theoretical_loss": 3.385319422123546,
      "tokens_seen": 2611675136
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.212806933076553e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.385306101572421,
      "tokens_seen": 2611806208
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.212004493660729e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.385292781876929,
      "tokens_seen": 2611937280
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.211202054244905e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.385279463036971,
      "tokens_seen": 2612068352
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.210399614829081e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.3852661450524497,
      "tokens_seen": 2612199424
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.209597175413256e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.385252827923267,
      "tokens_seen": 2612330496
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.208794735997432e-05,
      "loss": 1.2254,
      "theoretical_loss": 3.3852395116493255,
      "tokens_seen": 2612461568
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.207992296581608e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.385226196230527,
      "tokens_seen": 2612592640
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.207189857165784e-05,
      "loss": 1.3613,
      "theoretical_loss": 3.3852128816667735,
      "tokens_seen": 2612723712
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.20638741774996e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.3851995679579674,
      "tokens_seen": 2612854784
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.205584978334136e-05,
      "loss": 1.3382,
      "theoretical_loss": 3.385186255104011,
      "tokens_seen": 2612985856
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.204782538918312e-05,
      "loss": 1.337,
      "theoretical_loss": 3.3851729431048065,
      "tokens_seen": 2613116928
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2039800995024874e-05,
      "loss": 1.2233,
      "theoretical_loss": 3.385159631960256,
      "tokens_seen": 2613248000
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.9929156303405762,
      "objective/train/avg_token_score": 0.007084388751536608,
      "objective/train/docs_used": 950332,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.224981665611267,
      "objective/train/original_loss": 1.234660267829895,
      "objective/train/theoretical_loss": 3.3851463216702617,
      "objective/train/tokens_used": 983838176,
      "objective/train/unlikelihood_loss": 0.01988373138010502,
      "theoretical_loss": 3.3851463216702617,
      "tokens_seen": 2613379072
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2031776600866634e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.3851463216702617,
      "tokens_seen": 2613379072
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2023752206708394e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.3851330122347267,
      "tokens_seen": 2613510144
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.2015727812550153e-05,
      "loss": 1.311,
      "theoretical_loss": 3.3851197036535523,
      "tokens_seen": 2613641216
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.200770341839191e-05,
      "loss": 1.2955,
      "theoretical_loss": 3.385106395926641,
      "tokens_seen": 2613772288
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.199967902423367e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.3850930890538957,
      "tokens_seen": 2613903360
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.199165463007543e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.385079783035218,
      "tokens_seen": 2614034432
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.198363023591719e-05,
      "loss": 1.3682,
      "theoretical_loss": 3.3850664778705113,
      "tokens_seen": 2614165504
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.1975605841758946e-05,
      "loss": 1.2369,
      "theoretical_loss": 3.385053173559677,
      "tokens_seen": 2614296576
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.1967581447600706e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.3850398701026174,
      "tokens_seen": 2614427648
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.1959557053442466e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.385026567499236,
      "tokens_seen": 2614558720
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.1951532659284226e-05,
      "loss": 1.276,
      "theoretical_loss": 3.385013265749434,
      "tokens_seen": 2614689792
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.1943508265125986e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.3849999648531144,
      "tokens_seen": 2614820864
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.1935483870967746e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.38498666481018,
      "tokens_seen": 2614951936
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.1927459476809506e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.3849733656205325,
      "tokens_seen": 2615083008
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.191943508265126e-05,
      "loss": 1.275,
      "theoretical_loss": 3.384960067284075,
      "tokens_seen": 2615214080
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.191141068849302e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.38494676980071,
      "tokens_seen": 2615345152
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.190338629433478e-05,
      "loss": 1.1826,
      "theoretical_loss": 3.3849334731703395,
      "tokens_seen": 2615476224
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.189536190017654e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.3849201773928668,
      "tokens_seen": 2615607296
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.18873375060183e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.384906882468194,
      "tokens_seen": 2615738368
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.187931311186006e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.3848935883962237,
      "tokens_seen": 2615869440
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.187128871770182e-05,
      "loss": 1.262,
      "theoretical_loss": 3.3848802951768584,
      "tokens_seen": 2616000512
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.186326432354357e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.384867002810001,
      "tokens_seen": 2616131584
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.185523992938533e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.3848537112955537,
      "tokens_seen": 2616262656
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.184721553522709e-05,
      "loss": 1.314,
      "theoretical_loss": 3.3848404206334193,
      "tokens_seen": 2616393728
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.183919114106885e-05,
      "loss": 1.3227,
      "theoretical_loss": 3.3848271308235005,
      "tokens_seen": 2616524800
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.183116674691061e-05,
      "loss": 1.207,
      "theoretical_loss": 3.3848138418657,
      "tokens_seen": 2616655872
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.182314235275237e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.384800553759921,
      "tokens_seen": 2616786944
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.181511795859413e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.384787266506065,
      "tokens_seen": 2616918016
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.180709356443589e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.3847739801040357,
      "tokens_seen": 2617049088
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1799069170277644e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.3847606945537354,
      "tokens_seen": 2617180160
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1791044776119404e-05,
      "loss": 1.191,
      "theoretical_loss": 3.384747409855067,
      "tokens_seen": 2617311232
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1783020381961164e-05,
      "loss": 1.355,
      "theoretical_loss": 3.3847341260079338,
      "tokens_seen": 2617442304
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1774995987802923e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.3847208430122375,
      "tokens_seen": 2617573376
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1766971593644683e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.384707560867881,
      "tokens_seen": 2617704448
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.175894719948644e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.3846942795747683,
      "tokens_seen": 2617835520
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.17509228053282e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.384680999132801,
      "tokens_seen": 2617966592
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1742898411169956e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3846677195418824,
      "tokens_seen": 2618097664
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1734874017011716e-05,
      "loss": 1.2211,
      "theoretical_loss": 3.3846544408019152,
      "tokens_seen": 2618228736
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1726849622853476e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.384641162912803,
      "tokens_seen": 2618359808
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1718825228695236e-05,
      "loss": 1.2111,
      "theoretical_loss": 3.3846278858744476,
      "tokens_seen": 2618490880
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1710800834536996e-05,
      "loss": 1.3481,
      "theoretical_loss": 3.3846146096867527,
      "tokens_seen": 2618621952
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1702776440378756e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.384601334349621,
      "tokens_seen": 2618753024
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1694752046220516e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.384588059862955,
      "tokens_seen": 2618884096
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1686727652062276e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.384574786226658,
      "tokens_seen": 2619015168
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.167870325790403e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.3845615134406337,
      "tokens_seen": 2619146240
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.167067886374579e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.384548241504784,
      "tokens_seen": 2619277312
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.166265446958755e-05,
      "loss": 1.3509,
      "theoretical_loss": 3.384534970419012,
      "tokens_seen": 2619408384
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.165463007542931e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.3845217001832215,
      "tokens_seen": 2619539456
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.164660568127107e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.384508430797315,
      "tokens_seen": 2619670528
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.163858128711283e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.3844951622611954,
      "tokens_seen": 2619801600
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.99135422706604,
      "objective/train/avg_token_score": 0.008645761758089066,
      "objective/train/docs_used": 952663,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2298343181610107,
      "objective/train/original_loss": 1.2415592670440674,
      "objective/train/theoretical_loss": 3.3844818945747663,
      "objective/train/tokens_used": 990391776,
      "objective/train/unlikelihood_loss": 0.0225478895008564,
      "theoretical_loss": 3.3844818945747663,
      "tokens_seen": 2619932672
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.163055689295459e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3844818945747663,
      "tokens_seen": 2619932672
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.162253249879634e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.38446862773793,
      "tokens_seen": 2620063744
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.16145081046381e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.3844553617505904,
      "tokens_seen": 2620194816
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.160648371047986e-05,
      "loss": 1.3325,
      "theoretical_loss": 3.3844420966126503,
      "tokens_seen": 2620325888
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.159845931632162e-05,
      "loss": 1.2145,
      "theoretical_loss": 3.3844288323240126,
      "tokens_seen": 2620456960
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.159043492216338e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.384415568884581,
      "tokens_seen": 2620588032
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.158241052800514e-05,
      "loss": 1.3774,
      "theoretical_loss": 3.384402306294258,
      "tokens_seen": 2620719104
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.15743861338469e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.384389044552947,
      "tokens_seen": 2620850176
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1566361739688654e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.3843757836605515,
      "tokens_seen": 2620981248
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1558337345530414e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.384362523616974,
      "tokens_seen": 2621112320
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1550312951372174e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.384349264422119,
      "tokens_seen": 2621243392
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1542288557213934e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.3843360060758885,
      "tokens_seen": 2621374464
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1534264163055694e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3843227485781866,
      "tokens_seen": 2621505536
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1526239768897453e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.384309491928916,
      "tokens_seen": 2621636608
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.151821537473921e-05,
      "loss": 1.3339,
      "theoretical_loss": 3.38429623612798,
      "tokens_seen": 2621767680
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.151019098058097e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.384282981175282,
      "tokens_seen": 2621898752
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1502166586422726e-05,
      "loss": 1.2269,
      "theoretical_loss": 3.3842697270707256,
      "tokens_seen": 2622029824
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1494142192264486e-05,
      "loss": 1.3064,
      "theoretical_loss": 3.3842564738142142,
      "tokens_seen": 2622160896
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1486117798106246e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.3842432214056504,
      "tokens_seen": 2622291968
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1478093403948006e-05,
      "loss": 1.3123,
      "theoretical_loss": 3.3842299698449385,
      "tokens_seen": 2622423040
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1470069009789766e-05,
      "loss": 1.2441,
      "theoretical_loss": 3.384216719131981,
      "tokens_seen": 2622554112
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1462044615631526e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.3842034692666823,
      "tokens_seen": 2622685184
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1454020221473286e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.3841902202489447,
      "tokens_seen": 2622816256
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.144599582731504e-05,
      "loss": 1.2485,
      "theoretical_loss": 3.3841769720786727,
      "tokens_seen": 2622947328
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.14379714331568e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.3841637247557688,
      "tokens_seen": 2623078400
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.142994703899856e-05,
      "loss": 1.1824,
      "theoretical_loss": 3.3841504782801373,
      "tokens_seen": 2623209472
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.142192264484032e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.384137232651681,
      "tokens_seen": 2623340544
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.141389825068208e-05,
      "loss": 1.3039,
      "theoretical_loss": 3.3841239878703036,
      "tokens_seen": 2623471616
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.140587385652384e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.384110743935909,
      "tokens_seen": 2623602688
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.13978494623656e-05,
      "loss": 1.3478,
      "theoretical_loss": 3.3840975008484,
      "tokens_seen": 2623733760
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.138982506820735e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.384084258607681,
      "tokens_seen": 2623864832
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.138180067404911e-05,
      "loss": 1.3886,
      "theoretical_loss": 3.384071017213655,
      "tokens_seen": 2623995904
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.137377627989087e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.384057776666226,
      "tokens_seen": 2624126976
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.136575188573263e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.384044536965297,
      "tokens_seen": 2624258048
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.135772749157439e-05,
      "loss": 1.265,
      "theoretical_loss": 3.384031298110772,
      "tokens_seen": 2624389120
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.134970309741615e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.3840180601025542,
      "tokens_seen": 2624520192
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.134167870325791e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.3840048229405477,
      "tokens_seen": 2624651264
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.133365430909967e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.383991586624656,
      "tokens_seen": 2624782336
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1325629914941424e-05,
      "loss": 1.3394,
      "theoretical_loss": 3.3839783511547825,
      "tokens_seen": 2624913408
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1317605520783184e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.3839651165308315,
      "tokens_seen": 2625044480
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1309581126624944e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.383951882752706,
      "tokens_seen": 2625175552
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1301556732466704e-05,
      "loss": 1.3738,
      "theoretical_loss": 3.383938649820311,
      "tokens_seen": 2625306624
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1293532338308464e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.3839254177335487,
      "tokens_seen": 2625437696
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1285507944150224e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.3839121864923234,
      "tokens_seen": 2625568768
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.127748354999198e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.383898956096539,
      "tokens_seen": 2625699840
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1269459155833737e-05,
      "loss": 1.271,
      "theoretical_loss": 3.383885726546099,
      "tokens_seen": 2625830912
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1261434761675496e-05,
      "loss": 1.3676,
      "theoretical_loss": 3.3838724978409074,
      "tokens_seen": 2625961984
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1253410367517256e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.383859269980868,
      "tokens_seen": 2626093056
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1245385973359016e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.383846042965885,
      "tokens_seen": 2626224128
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.123736157920077e-05,
      "loss": 1.3007,
      "theoretical_loss": 3.3838328167958616,
      "tokens_seen": 2626355200
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.9925761818885803,
      "objective/train/avg_token_score": 0.007423733361065388,
      "objective/train/docs_used": 954702,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.144209623336792,
      "objective/train/original_loss": 1.1544349193572998,
      "objective/train/theoretical_loss": 3.383819591470702,
      "objective/train/tokens_used": 996945376,
      "objective/train/unlikelihood_loss": 0.02000565081834793,
      "theoretical_loss": 3.383819591470702,
      "tokens_seen": 2626486272
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.122933718504253e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.383819591470702,
      "tokens_seen": 2626486272
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.122131279088429e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.38380636699031,
      "tokens_seen": 2626617344
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.121328839672605e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3837931433545894,
      "tokens_seen": 2626748416
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.120526400256781e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.3837799205634442,
      "tokens_seen": 2626879488
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.119723960840957e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.3837666986167783,
      "tokens_seen": 2627010560
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.118921521425132e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.383753477514496,
      "tokens_seen": 2627141632
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.118119082009308e-05,
      "loss": 1.3314,
      "theoretical_loss": 3.3837402572565005,
      "tokens_seen": 2627272704
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.117316642593484e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.383727037842696,
      "tokens_seen": 2627403776
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.11651420317766e-05,
      "loss": 1.2079,
      "theoretical_loss": 3.383713819272987,
      "tokens_seen": 2627534848
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.115711763761836e-05,
      "loss": 1.329,
      "theoretical_loss": 3.383700601547277,
      "tokens_seen": 2627665920
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.114909324346012e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.3836873846654703,
      "tokens_seen": 2627796992
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1141068849301875e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.383674168627471,
      "tokens_seen": 2627928064
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1133044455143635e-05,
      "loss": 1.3821,
      "theoretical_loss": 3.3836609534331825,
      "tokens_seen": 2628059136
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1125020060985395e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3836477390825097,
      "tokens_seen": 2628190208
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1116995666827154e-05,
      "loss": 1.3398,
      "theoretical_loss": 3.383634525575356,
      "tokens_seen": 2628321280
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1108971272668914e-05,
      "loss": 1.286,
      "theoretical_loss": 3.383621312911626,
      "tokens_seen": 2628452352
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.1100946878510674e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.3836081010912236,
      "tokens_seen": 2628583424
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.109292248435243e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.3835948901140527,
      "tokens_seen": 2628714496
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.108489809019419e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.3835816799800176,
      "tokens_seen": 2628845568
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.107687369603595e-05,
      "loss": 1.322,
      "theoretical_loss": 3.383568470689023,
      "tokens_seen": 2628976640
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.106884930187771e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.383555262240972,
      "tokens_seen": 2629107712
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.106082490771947e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.38354205463577,
      "tokens_seen": 2629238784
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.105280051356123e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.38352884787332,
      "tokens_seen": 2629369856
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.104477611940299e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.383515641953527,
      "tokens_seen": 2629500928
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.103675172524475e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.383502436876295,
      "tokens_seen": 2629632000
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.10287273310865e-05,
      "loss": 1.2334,
      "theoretical_loss": 3.3834892326415282,
      "tokens_seen": 2629763072
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.102070293692826e-05,
      "loss": 1.203,
      "theoretical_loss": 3.383476029249131,
      "tokens_seen": 2629894144
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.101267854277002e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3834628266990077,
      "tokens_seen": 2630025216
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.100465414861178e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.3834496249910626,
      "tokens_seen": 2630156288
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.099662975445354e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.3834364241252,
      "tokens_seen": 2630287360
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.09886053602953e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.383423224101324,
      "tokens_seen": 2630418432
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.098058096613706e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.383410024919339,
      "tokens_seen": 2630549504
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.097255657197881e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.3833968265791494,
      "tokens_seen": 2630680576
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.096453217782057e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.38338362908066,
      "tokens_seen": 2630811648
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.095650778366233e-05,
      "loss": 1.302,
      "theoretical_loss": 3.3833704324237748,
      "tokens_seen": 2630942720
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.094848338950409e-05,
      "loss": 1.316,
      "theoretical_loss": 3.383357236608398,
      "tokens_seen": 2631073792
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.094045899534585e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.3833440416344347,
      "tokens_seen": 2631204864
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.093243460118761e-05,
      "loss": 1.2134,
      "theoretical_loss": 3.3833308475017883,
      "tokens_seen": 2631335936
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.092441020702937e-05,
      "loss": 1.1942,
      "theoretical_loss": 3.383317654210364,
      "tokens_seen": 2631467008
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.0916385812871125e-05,
      "loss": 1.2247,
      "theoretical_loss": 3.3833044617600665,
      "tokens_seen": 2631598080
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0908361418712885e-05,
      "loss": 1.2453,
      "theoretical_loss": 3.3832912701508,
      "tokens_seen": 2631729152
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0900337024554645e-05,
      "loss": 1.2375,
      "theoretical_loss": 3.3832780793824684,
      "tokens_seen": 2631860224
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0892312630396405e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.383264889454977,
      "tokens_seen": 2631991296
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0884288236238165e-05,
      "loss": 1.2172,
      "theoretical_loss": 3.3832517003682296,
      "tokens_seen": 2632122368
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0876263842079925e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.383238512122132,
      "tokens_seen": 2632253440
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0868239447921684e-05,
      "loss": 1.3582,
      "theoretical_loss": 3.3832253247165873,
      "tokens_seen": 2632384512
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0860215053763444e-05,
      "loss": 1.261,
      "theoretical_loss": 3.3832121381515012,
      "tokens_seen": 2632515584
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.08521906596052e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.3831989524267776,
      "tokens_seen": 2632646656
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.084416626544696e-05,
      "loss": 1.2023,
      "theoretical_loss": 3.3831857675423214,
      "tokens_seen": 2632777728
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.083614187128872e-05,
      "loss": 1.2395,
      "theoretical_loss": 3.383172583498037,
      "tokens_seen": 2632908800
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.9916519522666931,
      "objective/train/avg_token_score": 0.008348029106855392,
      "objective/train/docs_used": 956895,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2607543468475342,
      "objective/train/original_loss": 1.2739973068237305,
      "objective/train/theoretical_loss": 3.3831594002938297,
      "objective/train/tokens_used": 1003498976,
      "objective/train/unlikelihood_loss": 0.02044493891298771,
      "theoretical_loss": 3.3831594002938297,
      "tokens_seen": 2633039872
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.082811747713048e-05,
      "loss": 1.2586,
      "theoretical_loss": 3.3831594002938297,
      "tokens_seen": 2633039872
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.082009308297224e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.3831462179296032,
      "tokens_seen": 2633170944
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0812068688814e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.3831330364052628,
      "tokens_seen": 2633302016
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.080404429465576e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.383119855720713,
      "tokens_seen": 2633433088
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.079601990049751e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.3831066758758586,
      "tokens_seen": 2633564160
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.078799550633927e-05,
      "loss": 1.2346,
      "theoretical_loss": 3.3830934968706043,
      "tokens_seen": 2633695232
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.077997111218103e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.383080318704855,
      "tokens_seen": 2633826304
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.077194671802279e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.3830671413785147,
      "tokens_seen": 2633957376
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.076392232386455e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.383053964891489,
      "tokens_seen": 2634088448
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.075589792970631e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.3830407892436822,
      "tokens_seen": 2634219520
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.074787353554807e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.3830276144349996,
      "tokens_seen": 2634350592
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.073984914138983e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.3830144404653453,
      "tokens_seen": 2634481664
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.073182474723158e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.383001267334625,
      "tokens_seen": 2634612736
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.072380035307334e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.3829880950427427,
      "tokens_seen": 2634743808
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.07157759589151e-05,
      "loss": 1.2465,
      "theoretical_loss": 3.382974923589604,
      "tokens_seen": 2634874880
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.070775156475686e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.382961752975113,
      "tokens_seen": 2635005952
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.069972717059862e-05,
      "loss": 1.334,
      "theoretical_loss": 3.382948583199175,
      "tokens_seen": 2635137024
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.069170277644038e-05,
      "loss": 1.3337,
      "theoretical_loss": 3.382935414261695,
      "tokens_seen": 2635268096
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.068367838228214e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3829222461625776,
      "tokens_seen": 2635399168
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0675653988123895e-05,
      "loss": 1.234,
      "theoretical_loss": 3.3829090789017284,
      "tokens_seen": 2635530240
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0667629593965655e-05,
      "loss": 1.239,
      "theoretical_loss": 3.3828959124790514,
      "tokens_seen": 2635661312
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0659605199807415e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.3828827468944525,
      "tokens_seen": 2635792384
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0651580805649175e-05,
      "loss": 1.2441,
      "theoretical_loss": 3.382869582147836,
      "tokens_seen": 2635923456
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0643556411490935e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.382856418239107,
      "tokens_seen": 2636054528
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0635532017332695e-05,
      "loss": 1.1703,
      "theoretical_loss": 3.382843255168171,
      "tokens_seen": 2636185600
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0627507623174455e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.3828300929349324,
      "tokens_seen": 2636316672
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.061948322901621e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.3828169315392964,
      "tokens_seen": 2636447744
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.061145883485797e-05,
      "loss": 1.281,
      "theoretical_loss": 3.3828037709811687,
      "tokens_seen": 2636578816
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.060343444069973e-05,
      "loss": 1.266,
      "theoretical_loss": 3.3827906112604538,
      "tokens_seen": 2636709888
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.059541004654149e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.3827774523770566,
      "tokens_seen": 2636840960
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.058738565238325e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.382764294330882,
      "tokens_seen": 2636972032
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.057936125822501e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.3827511371218364,
      "tokens_seen": 2637103104
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.057133686406677e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.382737980749824,
      "tokens_seen": 2637234176
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.056331246990853e-05,
      "loss": 1.316,
      "theoretical_loss": 3.38272482521475,
      "tokens_seen": 2637365248
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.055528807575028e-05,
      "loss": 1.2393,
      "theoretical_loss": 3.382711670516519,
      "tokens_seen": 2637496320
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.054726368159204e-05,
      "loss": 1.218,
      "theoretical_loss": 3.3826985166550374,
      "tokens_seen": 2637627392
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.05392392874338e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.38268536363021,
      "tokens_seen": 2637758464
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.053121489327556e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3826722114419416,
      "tokens_seen": 2637889536
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.052319049911732e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.3826590600901376,
      "tokens_seen": 2638020608
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.051516610495908e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.3826459095747037,
      "tokens_seen": 2638151680
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.050714171080084e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.3826327598955443,
      "tokens_seen": 2638282752
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.049911731664259e-05,
      "loss": 1.2393,
      "theoretical_loss": 3.382619611052565,
      "tokens_seen": 2638413824
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.049109292248435e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.382606463045672,
      "tokens_seen": 2638544896
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.048306852832611e-05,
      "loss": 1.2803,
      "theoretical_loss": 3.382593315874769,
      "tokens_seen": 2638675968
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.047504413416787e-05,
      "loss": 1.239,
      "theoretical_loss": 3.3825801695397626,
      "tokens_seen": 2638807040
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.046701974000963e-05,
      "loss": 1.2407,
      "theoretical_loss": 3.382567024040558,
      "tokens_seen": 2638938112
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.045899534585139e-05,
      "loss": 1.2241,
      "theoretical_loss": 3.3825538793770598,
      "tokens_seen": 2639069184
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.045097095169315e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.382540735549174,
      "tokens_seen": 2639200256
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0442946557534905e-05,
      "loss": 1.2254,
      "theoretical_loss": 3.382527592556806,
      "tokens_seen": 2639331328
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0434922163376665e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.382514450399861,
      "tokens_seen": 2639462400
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.9917089343070984,
      "objective/train/avg_token_score": 0.008291028439998627,
      "objective/train/docs_used": 959162,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2331702709197998,
      "objective/train/original_loss": 1.244254231452942,
      "objective/train/theoretical_loss": 3.382501309078244,
      "objective/train/tokens_used": 1010052576,
      "objective/train/unlikelihood_loss": 0.02237582951784134,
      "theoretical_loss": 3.382501309078244,
      "tokens_seen": 2639593472
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0426897769218425e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.382501309078244,
      "tokens_seen": 2639593472
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0418873375060185e-05,
      "loss": 1.2181,
      "theoretical_loss": 3.382488168591861,
      "tokens_seen": 2639724544
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0410848980901945e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.3824750289406174,
      "tokens_seen": 2639855616
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0402824586743705e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.382461890124419,
      "tokens_seen": 2639986688
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0394800192585465e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.38244875214317,
      "tokens_seen": 2640117760
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0386775798427225e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.382435614996777,
      "tokens_seen": 2640248832
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.037875140426898e-05,
      "loss": 1.2282,
      "theoretical_loss": 3.382422478685146,
      "tokens_seen": 2640379904
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.037072701011074e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.382409343208181,
      "tokens_seen": 2640510976
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.03627026159525e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.3823962085657886,
      "tokens_seen": 2640642048
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.035467822179426e-05,
      "loss": 1.1906,
      "theoretical_loss": 3.3823830747578736,
      "tokens_seen": 2640773120
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.034665382763602e-05,
      "loss": 1.2229,
      "theoretical_loss": 3.3823699417843427,
      "tokens_seen": 2640904192
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.033862943347778e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.3823568096451004,
      "tokens_seen": 2641035264
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.033060503931954e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.3823436783400527,
      "tokens_seen": 2641166336
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.032258064516129e-05,
      "loss": 1.2101,
      "theoretical_loss": 3.3823305478691053,
      "tokens_seen": 2641297408
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.031455625100305e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.382317418232164,
      "tokens_seen": 2641428480
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.030653185684481e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.382304289429134,
      "tokens_seen": 2641559552
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.029850746268657e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.382291161459921,
      "tokens_seen": 2641690624
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.029048306852833e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.382278034324431,
      "tokens_seen": 2641821696
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.028245867437009e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.3822649080225693,
      "tokens_seen": 2641952768
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.027443428021185e-05,
      "loss": 1.1827,
      "theoretical_loss": 3.3822517825542424,
      "tokens_seen": 2642083840
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.026640988605361e-05,
      "loss": 1.1672,
      "theoretical_loss": 3.382238657919355,
      "tokens_seen": 2642214912
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.025838549189536e-05,
      "loss": 1.2226,
      "theoretical_loss": 3.3822255341178136,
      "tokens_seen": 2642345984
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.025036109773712e-05,
      "loss": 1.2715,
      "theoretical_loss": 3.3822124111495233,
      "tokens_seen": 2642477056
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.024233670357888e-05,
      "loss": 1.214,
      "theoretical_loss": 3.3821992890143906,
      "tokens_seen": 2642608128
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.023431230942064e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.3821861677123204,
      "tokens_seen": 2642739200
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.02262879152624e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.38217304724322,
      "tokens_seen": 2642870272
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.021826352110416e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.382159927606993,
      "tokens_seen": 2643001344
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.021023912694592e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.3821468088035473,
      "tokens_seen": 2643132416
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0202214732787675e-05,
      "loss": 1.2178,
      "theoretical_loss": 3.3821336908327875,
      "tokens_seen": 2643263488
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0194190338629435e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.38212057369462,
      "tokens_seen": 2643394560
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0186165944471195e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.3821074573889502,
      "tokens_seen": 2643525632
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0178141550312955e-05,
      "loss": 1.225,
      "theoretical_loss": 3.382094341915685,
      "tokens_seen": 2643656704
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0170117156154715e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.382081227274729,
      "tokens_seen": 2643787776
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0162092761996475e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.3820681134659893,
      "tokens_seen": 2643918848
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.0154068367838235e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.382055000489371,
      "tokens_seen": 2644049920
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.014604397367999e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.38204188834478,
      "tokens_seen": 2644180992
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.013801957952175e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.382028777032123,
      "tokens_seen": 2644312064
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.012999518536351e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.3820156665513057,
      "tokens_seen": 2644443136
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.012197079120527e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.3820025569022336,
      "tokens_seen": 2644574208
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.011394639704703e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.3819894480848136,
      "tokens_seen": 2644705280
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.010592200288879e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.3819763400989507,
      "tokens_seen": 2644836352
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.009789760873055e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.381963232944552,
      "tokens_seen": 2644967424
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.008987321457231e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.381950126621523,
      "tokens_seen": 2645098496
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.008184882041406e-05,
      "loss": 1.2911,
      "theoretical_loss": 3.3819370211297692,
      "tokens_seen": 2645229568
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.007382442625582e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.381923916469198,
      "tokens_seen": 2645360640
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.006580003209758e-05,
      "loss": 1.2298,
      "theoretical_loss": 3.381910812639714,
      "tokens_seen": 2645491712
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.005777563793934e-05,
      "loss": 1.187,
      "theoretical_loss": 3.3818977096412244,
      "tokens_seen": 2645622784
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.00497512437811e-05,
      "loss": 1.2353,
      "theoretical_loss": 3.3818846074736357,
      "tokens_seen": 2645753856
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.004172684962286e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.3818715061368527,
      "tokens_seen": 2645884928
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.003370245546462e-05,
      "loss": 1.1752,
      "theoretical_loss": 3.3818584056307825,
      "tokens_seen": 2646016000
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.9936308860778809,
      "objective/train/avg_token_score": 0.006369124632328749,
      "objective/train/docs_used": 961281,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2024009227752686,
      "objective/train/original_loss": 1.2112183570861816,
      "objective/train/theoretical_loss": 3.3818453059553315,
      "objective/train/tokens_used": 1016606176,
      "objective/train/unlikelihood_loss": 0.0187978595495224,
      "theoretical_loss": 3.3818453059553315,
      "tokens_seen": 2646147072
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.002567806130637e-05,
      "loss": 1.2145,
      "theoretical_loss": 3.3818453059553315,
      "tokens_seen": 2646147072
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.001765366714813e-05,
      "loss": 1.213,
      "theoretical_loss": 3.3818322071104046,
      "tokens_seen": 2646278144
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.000962927298989e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.3818191090959093,
      "tokens_seen": 2646409216
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.000160487883165e-05,
      "loss": 1.2298,
      "theoretical_loss": 3.3818060119117517,
      "tokens_seen": 2646540288
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.999358048467341e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.3817929155578375,
      "tokens_seen": 2646671360
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.998555609051517e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.381779820034073,
      "tokens_seen": 2646802432
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.997753169635693e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.381766725340365,
      "tokens_seen": 2646933504
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.9969507302198685e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.3817536314766192,
      "tokens_seen": 2647064576
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.9961482908040445e-05,
      "loss": 1.2709,
      "theoretical_loss": 3.381740538442742,
      "tokens_seen": 2647195648
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.9953458513882205e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.3817274462386404,
      "tokens_seen": 2647326720
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.9945434119723965e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.3817143548642195,
      "tokens_seen": 2647457792
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.9937409725565725e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.381701264319387,
      "tokens_seen": 2647588864
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.9929385331407485e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.381688174604048,
      "tokens_seen": 2647719936
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.992136093724924e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3816750857181104,
      "tokens_seen": 2647851008
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.9913336543091e-05,
      "loss": 1.2578,
      "theoretical_loss": 3.381661997661479,
      "tokens_seen": 2647982080
    },
    {
      "epoch": 0.6,
      "learning_rate": 3.990531214893276e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.3816489104340612,
      "tokens_seen": 2648113152
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.989728775477452e-05,
      "loss": 1.2264,
      "theoretical_loss": 3.381635824035763,
      "tokens_seen": 2648244224
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.988926336061628e-05,
      "loss": 1.2218,
      "theoretical_loss": 3.3816227384664908,
      "tokens_seen": 2648375296
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.988123896645804e-05,
      "loss": 1.1659,
      "theoretical_loss": 3.3816096537261515,
      "tokens_seen": 2648506368
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.987321457229979e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.381596569814651,
      "tokens_seen": 2648637440
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.986519017814155e-05,
      "loss": 1.1497,
      "theoretical_loss": 3.3815834867318966,
      "tokens_seen": 2648768512
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.985716578398331e-05,
      "loss": 1.199,
      "theoretical_loss": 3.381570404477794,
      "tokens_seen": 2648899584
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.984914138982507e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.3815573230522498,
      "tokens_seen": 2649030656
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.984111699566683e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.381544242455171,
      "tokens_seen": 2649161728
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.983309260150859e-05,
      "loss": 1.1861,
      "theoretical_loss": 3.3815311626864637,
      "tokens_seen": 2649292800
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9825068207350343e-05,
      "loss": 1.219,
      "theoretical_loss": 3.381518083746035,
      "tokens_seen": 2649423872
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.98170438131921e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.3815050056337905,
      "tokens_seen": 2649554944
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.980901941903386e-05,
      "loss": 1.1781,
      "theoretical_loss": 3.3814919283496376,
      "tokens_seen": 2649686016
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.980099502487562e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.3814788518934833,
      "tokens_seen": 2649817088
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.979297063071738e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.381465776265233,
      "tokens_seen": 2649948160
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.978494623655914e-05,
      "loss": 1.258,
      "theoretical_loss": 3.381452701464794,
      "tokens_seen": 2650079232
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9776921842400896e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.381439627492073,
      "tokens_seen": 2650210304
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9768897448242656e-05,
      "loss": 1.2137,
      "theoretical_loss": 3.3814265543469766,
      "tokens_seen": 2650341376
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9760873054084416e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.3814134820294117,
      "tokens_seen": 2650472448
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9752848659926176e-05,
      "loss": 1.2157,
      "theoretical_loss": 3.3814004105392845,
      "tokens_seen": 2650603520
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9744824265767936e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.381387339876502,
      "tokens_seen": 2650734592
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9736799871609696e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.381374270040971,
      "tokens_seen": 2650865664
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.972877547745145e-05,
      "loss": 1.2103,
      "theoretical_loss": 3.381361201032598,
      "tokens_seen": 2650996736
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.972075108329321e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.3813481328512904,
      "tokens_seen": 2651127808
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.971272668913497e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.381335065496954,
      "tokens_seen": 2651258880
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.970470229497673e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.3813219989694963,
      "tokens_seen": 2651389952
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.969667790081849e-05,
      "loss": 1.1922,
      "theoretical_loss": 3.3813089332688238,
      "tokens_seen": 2651521024
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.968865350666025e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.3812958683948438,
      "tokens_seen": 2651652096
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.968062911250201e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.381282804347462,
      "tokens_seen": 2651783168
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.967260471834376e-05,
      "loss": 1.286,
      "theoretical_loss": 3.3812697411265864,
      "tokens_seen": 2651914240
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.966458032418552e-05,
      "loss": 1.2181,
      "theoretical_loss": 3.3812566787321234,
      "tokens_seen": 2652045312
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.965655593002728e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.38124361716398,
      "tokens_seen": 2652176384
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.964853153586904e-05,
      "loss": 1.2125,
      "theoretical_loss": 3.3812305564220626,
      "tokens_seen": 2652307456
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.96405071417108e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3812174965062787,
      "tokens_seen": 2652438528
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.963248274755256e-05,
      "loss": 1.2009,
      "theoretical_loss": 3.3812044374165353,
      "tokens_seen": 2652569600
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.9926072955131531,
      "objective/train/avg_token_score": 0.007392706349492073,
      "objective/train/docs_used": 963694,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.305254340171814,
      "objective/train/original_loss": 1.3146860599517822,
      "objective/train/theoretical_loss": 3.381191379152739,
      "objective/train/tokens_used": 1023159776,
      "objective/train/unlikelihood_loss": 0.02184179611504078,
      "theoretical_loss": 3.381191379152739,
      "tokens_seen": 2652700672
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.962445835339432e-05,
      "loss": 1.263,
      "theoretical_loss": 3.381191379152739,
      "tokens_seen": 2652700672
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.961643395923608e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.381178321714797,
      "tokens_seen": 2652831744
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9608409565077834e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.3811652651026156,
      "tokens_seen": 2652962816
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9600385170919594e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.381152209316103,
      "tokens_seen": 2653093888
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9592360776761354e-05,
      "loss": 1.2129,
      "theoretical_loss": 3.3811391543551648,
      "tokens_seen": 2653224960
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9584336382603114e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.381126100219709,
      "tokens_seen": 2653356032
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9576311988444873e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.3811130469096424,
      "tokens_seen": 2653487104
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.956828759428663e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.381099994424872,
      "tokens_seen": 2653618176
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.956026320012839e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.3810869427653047,
      "tokens_seen": 2653749248
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9552238805970146e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.381073891930848,
      "tokens_seen": 2653880320
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9544214411811906e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.3810608419214088,
      "tokens_seen": 2654011392
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9536190017653666e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.3810477927368936,
      "tokens_seen": 2654142464
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9528165623495426e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.38103474437721,
      "tokens_seen": 2654273536
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9520141229337186e-05,
      "loss": 1.236,
      "theoretical_loss": 3.381021696842266,
      "tokens_seen": 2654404608
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9512116835178946e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.3810086501319674,
      "tokens_seen": 2654535680
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9504092441020706e-05,
      "loss": 1.35,
      "theoretical_loss": 3.380995604246222,
      "tokens_seen": 2654666752
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.949606804686246e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.3809825591849365,
      "tokens_seen": 2654797824
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.948804365270422e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.380969514948019,
      "tokens_seen": 2654928896
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.948001925854598e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.380956471535376,
      "tokens_seen": 2655059968
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.947199486438774e-05,
      "loss": 1.2201,
      "theoretical_loss": 3.3809434289469147,
      "tokens_seen": 2655191040
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.94639704702295e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.3809303871825427,
      "tokens_seen": 2655322112
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.945594607607126e-05,
      "loss": 1.2336,
      "theoretical_loss": 3.3809173462421667,
      "tokens_seen": 2655453184
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.944792168191302e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.380904306125695,
      "tokens_seen": 2655584256
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.943989728775478e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.380891266833034,
      "tokens_seen": 2655715328
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.943187289359653e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.3808782283640912,
      "tokens_seen": 2655846400
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.942384849943829e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.380865190718774,
      "tokens_seen": 2655977472
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.941582410528005e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.38085215389699,
      "tokens_seen": 2656108544
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.940779971112181e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.380839117898646,
      "tokens_seen": 2656239616
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.939977531696357e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.380826082723649,
      "tokens_seen": 2656370688
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.939175092280533e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.3808130483719077,
      "tokens_seen": 2656501760
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.938372652864709e-05,
      "loss": 1.3087,
      "theoretical_loss": 3.380800014843329,
      "tokens_seen": 2656632832
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9375702134488844e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.3807869821378196,
      "tokens_seen": 2656763904
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9367677740330604e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.3807739502552874,
      "tokens_seen": 2656894976
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9359653346172364e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.38076091919564,
      "tokens_seen": 2657026048
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9351628952014124e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.3807478889587848,
      "tokens_seen": 2657157120
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9343604557855884e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.3807348595446287,
      "tokens_seen": 2657288192
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9335580163697644e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.38072183095308,
      "tokens_seen": 2657419264
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9327555769539403e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.3807088031840458,
      "tokens_seen": 2657550336
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.931953137538116e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.3806957762374337,
      "tokens_seen": 2657681408
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9311506981222916e-05,
      "loss": 1.3516,
      "theoretical_loss": 3.380682750113151,
      "tokens_seen": 2657812480
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9303482587064676e-05,
      "loss": 1.2068,
      "theoretical_loss": 3.3806697248111055,
      "tokens_seen": 2657943552
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9295458192906436e-05,
      "loss": 1.1972,
      "theoretical_loss": 3.3806567003312047,
      "tokens_seen": 2658074624
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9287433798748196e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.3806436766733556,
      "tokens_seen": 2658205696
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9279409404589956e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.380630653837467,
      "tokens_seen": 2658336768
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9271385010431716e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.3806176318234455,
      "tokens_seen": 2658467840
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9263360616273476e-05,
      "loss": 1.2536,
      "theoretical_loss": 3.380604610631199,
      "tokens_seen": 2658598912
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.925533622211523e-05,
      "loss": 1.3134,
      "theoretical_loss": 3.3805915902606354,
      "tokens_seen": 2658729984
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.924731182795699e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.380578570711662,
      "tokens_seen": 2658861056
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.923928743379875e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.380565551984186,
      "tokens_seen": 2658992128
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.923126303964051e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.3805525340781166,
      "tokens_seen": 2659123200
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.9953900575637817,
      "objective/train/avg_token_score": 0.004609938710927963,
      "objective/train/docs_used": 966064,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3459446430206299,
      "objective/train/original_loss": 1.3525340557098389,
      "objective/train/theoretical_loss": 3.38053951699336,
      "objective/train/tokens_used": 1029713376,
      "objective/train/unlikelihood_loss": 0.012804620899260044,
      "theoretical_loss": 3.38053951699336,
      "tokens_seen": 2659254272
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.922323864548227e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.38053951699336,
      "tokens_seen": 2659254272
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.921521425132403e-05,
      "loss": 1.1813,
      "theoretical_loss": 3.3805265007298244,
      "tokens_seen": 2659385344
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.920718985716579e-05,
      "loss": 1.3597,
      "theoretical_loss": 3.380513485287418,
      "tokens_seen": 2659516416
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.919916546300754e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.3805004706660475,
      "tokens_seen": 2659647488
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.91911410688493e-05,
      "loss": 1.1868,
      "theoretical_loss": 3.3804874568656214,
      "tokens_seen": 2659778560
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.918311667469106e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.3804744438860475,
      "tokens_seen": 2659909632
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.917509228053282e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.3804614317272335,
      "tokens_seen": 2660040704
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.916706788637458e-05,
      "loss": 1.1667,
      "theoretical_loss": 3.3804484203890866,
      "tokens_seen": 2660171776
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.915904349221634e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.3804354098715153,
      "tokens_seen": 2660302848
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.91510190980581e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.380422400174427,
      "tokens_seen": 2660433920
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.914299470389986e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.3804093912977304,
      "tokens_seen": 2660564992
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9134970309741614e-05,
      "loss": 1.2387,
      "theoretical_loss": 3.380396383241332,
      "tokens_seen": 2660696064
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9126945915583374e-05,
      "loss": 1.2143,
      "theoretical_loss": 3.380383376005141,
      "tokens_seen": 2660827136
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9118921521425134e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.380370369589064,
      "tokens_seen": 2660958208
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9110897127266894e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.38035736399301,
      "tokens_seen": 2661089280
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9102872733108654e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.380344359216886,
      "tokens_seen": 2661220352
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9094848338950414e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.380331355260601,
      "tokens_seen": 2661351424
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9086823944792174e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.380318352124062,
      "tokens_seen": 2661482496
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.907879955063393e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.3803053498071773,
      "tokens_seen": 2661613568
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9070775156475687e-05,
      "loss": 1.2492,
      "theoretical_loss": 3.3802923483098546,
      "tokens_seen": 2661744640
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9062750762317446e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.3802793476320026,
      "tokens_seen": 2661875712
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9054726368159206e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.380266347773529,
      "tokens_seen": 2662006784
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9046701974000966e-05,
      "loss": 1.1958,
      "theoretical_loss": 3.3802533487343407,
      "tokens_seen": 2662137856
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9038677579842726e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.380240350514348,
      "tokens_seen": 2662268928
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9030653185684486e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.3802273531134563,
      "tokens_seen": 2662400000
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.902262879152624e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.3802143565315763,
      "tokens_seen": 2662531072
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.9014604397368e-05,
      "loss": 1.2258,
      "theoretical_loss": 3.380201360768614,
      "tokens_seen": 2662662144
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.900658000320976e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.3801883658244782,
      "tokens_seen": 2662793216
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.899855560905152e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.3801753716990777,
      "tokens_seen": 2662924288
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.899053121489328e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.3801623783923196,
      "tokens_seen": 2663055360
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.898250682073504e-05,
      "loss": 1.31,
      "theoretical_loss": 3.3801493859041125,
      "tokens_seen": 2663186432
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.89744824265768e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.3801363942343645,
      "tokens_seen": 2663317504
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.896645803241856e-05,
      "loss": 1.1825,
      "theoretical_loss": 3.3801234033829837,
      "tokens_seen": 2663448576
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.895843363826031e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.3801104133498785,
      "tokens_seen": 2663579648
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.895040924410207e-05,
      "loss": 1.176,
      "theoretical_loss": 3.3800974241349566,
      "tokens_seen": 2663710720
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.894238484994383e-05,
      "loss": 1.3563,
      "theoretical_loss": 3.380084435738127,
      "tokens_seen": 2663841792
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.893436045578559e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.3800714481592973,
      "tokens_seen": 2663972864
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.892633606162735e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.380058461398376,
      "tokens_seen": 2664103936
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.891831166746911e-05,
      "loss": 1.2218,
      "theoretical_loss": 3.380045475455271,
      "tokens_seen": 2664235008
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.891028727331087e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.380032490329891,
      "tokens_seen": 2664366080
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.8902262879152624e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.380019506022144,
      "tokens_seen": 2664497152
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.8894238484994384e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.3800065225319385,
      "tokens_seen": 2664628224
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8886214090836144e-05,
      "loss": 1.2431,
      "theoretical_loss": 3.379993539859183,
      "tokens_seen": 2664759296
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8878189696677904e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.379980558003785,
      "tokens_seen": 2664890368
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8870165302519664e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.3799675769656536,
      "tokens_seen": 2665021440
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8862140908361424e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.379954596744697,
      "tokens_seen": 2665152512
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8854116514203184e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3799416173408234,
      "tokens_seen": 2665283584
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.884609212004494e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.379928638753942,
      "tokens_seen": 2665414656
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.88380677258867e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.37991566098396,
      "tokens_seen": 2665545728
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.883004333172846e-05,
      "loss": 1.2927,
      "theoretical_loss": 3.379902684030786,
      "tokens_seen": 2665676800
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.992882251739502,
      "objective/train/avg_token_score": 0.007117701694369316,
      "objective/train/docs_used": 968517,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1695009469985962,
      "objective/train/original_loss": 1.1788755655288696,
      "objective/train/theoretical_loss": 3.3798897078943293,
      "objective/train/tokens_used": 1036266976,
      "objective/train/unlikelihood_loss": 0.02379632368683815,
      "theoretical_loss": 3.3798897078943293,
      "tokens_seen": 2665807872
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8822018937570217e-05,
      "loss": 1.1886,
      "theoretical_loss": 3.3798897078943293,
      "tokens_seen": 2665807872
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8813994543411976e-05,
      "loss": 1.269,
      "theoretical_loss": 3.379876732574498,
      "tokens_seen": 2665938944
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8805970149253736e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.3798637580712,
      "tokens_seen": 2666070016
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8797945755095496e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.379850784384344,
      "tokens_seen": 2666201088
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8789921360937256e-05,
      "loss": 1.3495,
      "theoretical_loss": 3.379837811513839,
      "tokens_seen": 2666332160
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.878189696677901e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.3798248394595927,
      "tokens_seen": 2666463232
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.877387257262077e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.3798118682215144,
      "tokens_seen": 2666594304
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.876584817846253e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.3797988977995126,
      "tokens_seen": 2666725376
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.875782378430429e-05,
      "loss": 1.212,
      "theoretical_loss": 3.379785928193495,
      "tokens_seen": 2666856448
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.874979939014605e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.379772959403371,
      "tokens_seen": 2666987520
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.874177499598781e-05,
      "loss": 1.2149,
      "theoretical_loss": 3.379759991429049,
      "tokens_seen": 2667118592
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.873375060182957e-05,
      "loss": 1.3668,
      "theoretical_loss": 3.3797470242704373,
      "tokens_seen": 2667249664
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.872572620767132e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.379734057927445,
      "tokens_seen": 2667380736
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.871770181351308e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.37972109239998,
      "tokens_seen": 2667511808
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.870967741935484e-05,
      "loss": 1.2965,
      "theoretical_loss": 3.3797081276879517,
      "tokens_seen": 2667642880
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.87016530251966e-05,
      "loss": 1.313,
      "theoretical_loss": 3.3796951637912684,
      "tokens_seen": 2667773952
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.869362863103836e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.379682200709839,
      "tokens_seen": 2667905024
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.868560423688012e-05,
      "loss": 1.1669,
      "theoretical_loss": 3.3796692384435714,
      "tokens_seen": 2668036096
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.867757984272188e-05,
      "loss": 1.2955,
      "theoretical_loss": 3.379656276992375,
      "tokens_seen": 2668167168
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.866955544856364e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.3796433163561588,
      "tokens_seen": 2668298240
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8661531054405394e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.3796303565348307,
      "tokens_seen": 2668429312
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8653506660247154e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.3796173975282997,
      "tokens_seen": 2668560384
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8645482266088914e-05,
      "loss": 1.2931,
      "theoretical_loss": 3.3796044393364753,
      "tokens_seen": 2668691456
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8637457871930674e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.3795914819592654,
      "tokens_seen": 2668822528
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8629433477772434e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.379578525396579,
      "tokens_seen": 2668953600
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8621409083614194e-05,
      "loss": 1.1797,
      "theoretical_loss": 3.379565569648325,
      "tokens_seen": 2669084672
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.861338468945595e-05,
      "loss": 1.2232,
      "theoretical_loss": 3.379552614714412,
      "tokens_seen": 2669215744
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.860536029529771e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.3795396605947494,
      "tokens_seen": 2669346816
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.859733590113947e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.379526707289245,
      "tokens_seen": 2669477888
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.858931150698123e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.3795137547978094,
      "tokens_seen": 2669608960
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8581287112822987e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.3795008031203495,
      "tokens_seen": 2669740032
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8573262718664747e-05,
      "loss": 1.1786,
      "theoretical_loss": 3.3794878522567755,
      "tokens_seen": 2669871104
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.85652383245065e-05,
      "loss": 1.3861,
      "theoretical_loss": 3.3794749022069954,
      "tokens_seen": 2670002176
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.855721393034826e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.379461952970919,
      "tokens_seen": 2670133248
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.854918953619002e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.3794490045484546,
      "tokens_seen": 2670264320
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.854116514203178e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.3794360569395114,
      "tokens_seen": 2670395392
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.853314074787354e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.3794231101439984,
      "tokens_seen": 2670526464
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.85251163537153e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.3794101641618246,
      "tokens_seen": 2670657536
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.851709195955705e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.379397218992899,
      "tokens_seen": 2670788608
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.850906756539881e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3793842746371303,
      "tokens_seen": 2670919680
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.850104317124057e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.3793713310944278,
      "tokens_seen": 2671050752
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.849301877708233e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.3793583883647003,
      "tokens_seen": 2671181824
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.848499438292409e-05,
      "loss": 1.297,
      "theoretical_loss": 3.379345446447857,
      "tokens_seen": 2671312896
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.847696998876585e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.3793325053438075,
      "tokens_seen": 2671443968
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8468945594607605e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.3793195650524597,
      "tokens_seen": 2671575040
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8460921200449365e-05,
      "loss": 1.2294,
      "theoretical_loss": 3.3793066255737236,
      "tokens_seen": 2671706112
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8452896806291125e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.3792936869075083,
      "tokens_seen": 2671837184
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8444872412132885e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.379280749053722,
      "tokens_seen": 2671968256
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8436848017974645e-05,
      "loss": 1.2193,
      "theoretical_loss": 3.379267812012275,
      "tokens_seen": 2672099328
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8428823623816404e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.3792548757830763,
      "tokens_seen": 2672230400
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.9920362830162048,
      "objective/train/avg_token_score": 0.007963757030665874,
      "objective/train/docs_used": 970931,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2503935098648071,
      "objective/train/original_loss": 1.2620632648468018,
      "objective/train/theoretical_loss": 3.3792419403660343,
      "objective/train/tokens_used": 1042820576,
      "objective/train/unlikelihood_loss": 0.019985873252153397,
      "theoretical_loss": 3.3792419403660343,
      "tokens_seen": 2672361472
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.842079922965816e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.3792419403660343,
      "tokens_seen": 2672361472
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.841277483549992e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.379229005761059,
      "tokens_seen": 2672492544
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.840475044134168e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.3792160719680586,
      "tokens_seen": 2672623616
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.839672604718344e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3792031389869432,
      "tokens_seen": 2672754688
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.83887016530252e-05,
      "loss": 1.2498,
      "theoretical_loss": 3.379190206817622,
      "tokens_seen": 2672885760
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.838067725886696e-05,
      "loss": 1.2289,
      "theoretical_loss": 3.3791772754600036,
      "tokens_seen": 2673016832
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.837265286470871e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.379164344913998,
      "tokens_seen": 2673147904
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.836462847055047e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.3791514151795137,
      "tokens_seen": 2673278976
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.835660407639223e-05,
      "loss": 1.3519,
      "theoretical_loss": 3.3791384862564606,
      "tokens_seen": 2673410048
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.834857968223399e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.379125558144748,
      "tokens_seen": 2673541120
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.834055528807575e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3791126308442845,
      "tokens_seen": 2673672192
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.833253089391751e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.3790997043549806,
      "tokens_seen": 2673803264
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.832450649975927e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.3790867786767445,
      "tokens_seen": 2673934336
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.831648210560103e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.379073853809486,
      "tokens_seen": 2674065408
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.830845771144278e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.379060929753115,
      "tokens_seen": 2674196480
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.830043331728454e-05,
      "loss": 1.319,
      "theoretical_loss": 3.37904800650754,
      "tokens_seen": 2674327552
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.82924089231263e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.3790350840726706,
      "tokens_seen": 2674458624
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.828438452896806e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.3790221624484165,
      "tokens_seen": 2674589696
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.827636013480982e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.3790092416346873,
      "tokens_seen": 2674720768
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.826833574065158e-05,
      "loss": 1.1991,
      "theoretical_loss": 3.378996321631392,
      "tokens_seen": 2674851840
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.826031134649334e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.37898340243844,
      "tokens_seen": 2674982912
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8252286952335095e-05,
      "loss": 1.1677,
      "theoretical_loss": 3.3789704840557415,
      "tokens_seen": 2675113984
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8244262558176855e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.378957566483205,
      "tokens_seen": 2675245056
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8236238164018615e-05,
      "loss": 1.2625,
      "theoretical_loss": 3.378944649720741,
      "tokens_seen": 2675376128
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8228213769860375e-05,
      "loss": 1.2403,
      "theoretical_loss": 3.3789317337682583,
      "tokens_seen": 2675507200
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8220189375702135e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.3789188186256665,
      "tokens_seen": 2675638272
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8212164981543895e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.3789059042928753,
      "tokens_seen": 2675769344
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8204140587385655e-05,
      "loss": 1.284,
      "theoretical_loss": 3.378892990769794,
      "tokens_seen": 2675900416
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8196116193227415e-05,
      "loss": 1.3242,
      "theoretical_loss": 3.378880078056333,
      "tokens_seen": 2676031488
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.818809179906917e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.3788671661524012,
      "tokens_seen": 2676162560
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.818006740491093e-05,
      "loss": 1.3332,
      "theoretical_loss": 3.378854255057908,
      "tokens_seen": 2676293632
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.817204301075269e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.3788413447727637,
      "tokens_seen": 2676424704
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.816401861659445e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.378828435296877,
      "tokens_seen": 2676555776
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.815599422243621e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.3788155266301585,
      "tokens_seen": 2676686848
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.814796982827797e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.3788026187725175,
      "tokens_seen": 2676817920
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.813994543411973e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.3787897117238637,
      "tokens_seen": 2676948992
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.813192103996148e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3787768054841067,
      "tokens_seen": 2677080064
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.812389664580324e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.378763900053156,
      "tokens_seen": 2677211136
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8115872251645e-05,
      "loss": 1.1876,
      "theoretical_loss": 3.378750995430922,
      "tokens_seen": 2677342208
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.810784785748676e-05,
      "loss": 1.2182,
      "theoretical_loss": 3.3787380916173135,
      "tokens_seen": 2677473280
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.809982346332852e-05,
      "loss": 1.3031,
      "theoretical_loss": 3.3787251886122407,
      "tokens_seen": 2677604352
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.809179906917028e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.378712286415614,
      "tokens_seen": 2677735424
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.808377467501204e-05,
      "loss": 1.1919,
      "theoretical_loss": 3.3786993850273426,
      "tokens_seen": 2677866496
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.807575028085379e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.3786864844473357,
      "tokens_seen": 2677997568
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.806772588669555e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3786735846755036,
      "tokens_seen": 2678128640
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.805970149253731e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.3786606857117567,
      "tokens_seen": 2678259712
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.805167709837907e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.3786477875560044,
      "tokens_seen": 2678390784
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.804365270422083e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.378634890208156,
      "tokens_seen": 2678521856
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.803562831006259e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.3786219936681223,
      "tokens_seen": 2678652928
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.802760391590435e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.3786090979358123,
      "tokens_seen": 2678784000
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.9920907020568848,
      "objective/train/avg_token_score": 0.007909329608082771,
      "objective/train/docs_used": 973380,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2764972448349,
      "objective/train/original_loss": 1.2869486808776855,
      "objective/train/theoretical_loss": 3.3785962030111367,
      "objective/train/tokens_used": 1049374176,
      "objective/train/unlikelihood_loss": 0.021762266755104065,
      "theoretical_loss": 3.3785962030111367,
      "tokens_seen": 2678915072
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.801957952174611e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.3785962030111367,
      "tokens_seen": 2678915072
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8011555127587865e-05,
      "loss": 1.2202,
      "theoretical_loss": 3.378583308894005,
      "tokens_seen": 2679046144
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.8003530733429625e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.3785704155843272,
      "tokens_seen": 2679177216
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.7995506339271385e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.378557523082013,
      "tokens_seen": 2679308288
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.7987481945113145e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.3785446313869727,
      "tokens_seen": 2679439360
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.7979457550954905e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.378531740499116,
      "tokens_seen": 2679570432
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.7971433156796665e-05,
      "loss": 1.2509,
      "theoretical_loss": 3.378518850418353,
      "tokens_seen": 2679701504
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.7963408762638425e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.378505961144594,
      "tokens_seen": 2679832576
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.795538436848018e-05,
      "loss": 1.1729,
      "theoretical_loss": 3.378493072677748,
      "tokens_seen": 2679963648
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.794735997432194e-05,
      "loss": 1.299,
      "theoretical_loss": 3.378480185017726,
      "tokens_seen": 2680094720
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.79393355801637e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.3784672981644386,
      "tokens_seen": 2680225792
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.793131118600546e-05,
      "loss": 1.29,
      "theoretical_loss": 3.378454412117794,
      "tokens_seen": 2680356864
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.792328679184722e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3784415268777037,
      "tokens_seen": 2680487936
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.791526239768898e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.378428642444077,
      "tokens_seen": 2680619008
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.790723800353074e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.378415758816825,
      "tokens_seen": 2680750080
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.789921360937249e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.3784028759958566,
      "tokens_seen": 2680881152
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.789118921521425e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.3783899939810826,
      "tokens_seen": 2681012224
    },
    {
      "epoch": 0.62,
      "learning_rate": 3.788316482105601e-05,
      "loss": 1.2431,
      "theoretical_loss": 3.378377112772413,
      "tokens_seen": 2681143296
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.787514042689777e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.3783642323697585,
      "tokens_seen": 2681274368
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.786711603273953e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.378351352773028,
      "tokens_seen": 2681405440
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.785909163858129e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.378338473982133,
      "tokens_seen": 2681536512
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.785106724442305e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.3783255959969827,
      "tokens_seen": 2681667584
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.784304285026481e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.378312718817488,
      "tokens_seen": 2681798656
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.783501845610656e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.3782998424435586,
      "tokens_seen": 2681929728
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.782699406194832e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3782869668751054,
      "tokens_seen": 2682060800
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.781896966779008e-05,
      "loss": 1.3211,
      "theoretical_loss": 3.3782740921120378,
      "tokens_seen": 2682191872
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.781094527363184e-05,
      "loss": 1.294,
      "theoretical_loss": 3.378261218154267,
      "tokens_seen": 2682322944
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.78029208794736e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.3782483450017025,
      "tokens_seen": 2682454016
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.779489648531536e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.378235472654255,
      "tokens_seen": 2682585088
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.778687209115712e-05,
      "loss": 1.2026,
      "theoretical_loss": 3.3782226011118346,
      "tokens_seen": 2682716160
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7778847696998876e-05,
      "loss": 1.193,
      "theoretical_loss": 3.3782097303743517,
      "tokens_seen": 2682847232
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7770823302840635e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.378196860441717,
      "tokens_seen": 2682978304
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7762798908682395e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.3781839913138403,
      "tokens_seen": 2683109376
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7754774514524155e-05,
      "loss": 1.3227,
      "theoretical_loss": 3.378171122990632,
      "tokens_seen": 2683240448
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7746750120365915e-05,
      "loss": 1.257,
      "theoretical_loss": 3.378158255472003,
      "tokens_seen": 2683371520
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7738725726207675e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.378145388757863,
      "tokens_seen": 2683502592
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7730701332049435e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.378132522848124,
      "tokens_seen": 2683633664
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7722676937891195e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.378119657742694,
      "tokens_seen": 2683764736
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.771465254373295e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.378106793441485,
      "tokens_seen": 2683895808
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.770662814957471e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.378093929944407,
      "tokens_seen": 2684026880
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.769860375541647e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.378081067251371,
      "tokens_seen": 2684157952
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.769057936125823e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.3780682053622866,
      "tokens_seen": 2684289024
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.768255496709999e-05,
      "loss": 1.293,
      "theoretical_loss": 3.378055344277065,
      "tokens_seen": 2684420096
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.767453057294175e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.3780424839956167,
      "tokens_seen": 2684551168
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.766650617878351e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.3780296245178514,
      "tokens_seen": 2684682240
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.765848178462526e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.3780167658436806,
      "tokens_seen": 2684813312
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.765045739046702e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3780039079730146,
      "tokens_seen": 2684944384
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.764243299630878e-05,
      "loss": 1.1959,
      "theoretical_loss": 3.377991050905764,
      "tokens_seen": 2685075456
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.763440860215054e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.3779781946418392,
      "tokens_seen": 2685206528
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.76263842079923e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.3779653391811504,
      "tokens_seen": 2685337600
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.9927713871002197,
      "objective/train/avg_token_score": 0.0072286115027964115,
      "objective/train/docs_used": 975809,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1838074922561646,
      "objective/train/original_loss": 1.191867709159851,
      "objective/train/theoretical_loss": 3.3779524845236093,
      "objective/train/tokens_used": 1055927776,
      "objective/train/unlikelihood_loss": 0.02560616470873356,
      "theoretical_loss": 3.3779524845236093,
      "tokens_seen": 2685468672
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.761835981383406e-05,
      "loss": 1.1882,
      "theoretical_loss": 3.3779524845236093,
      "tokens_seen": 2685468672
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.761033541967582e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3779396306691254,
      "tokens_seen": 2685599744
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.760231102551757e-05,
      "loss": 1.3629,
      "theoretical_loss": 3.3779267776176103,
      "tokens_seen": 2685730816
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.759428663135933e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.377913925368974,
      "tokens_seen": 2685861888
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.758626223720109e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.377901073923127,
      "tokens_seen": 2685992960
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.757823784304285e-05,
      "loss": 1.2329,
      "theoretical_loss": 3.3778882232799807,
      "tokens_seen": 2686124032
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.757021344888461e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.3778753734394455,
      "tokens_seen": 2686255104
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.756218905472637e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.377862524401432,
      "tokens_seen": 2686386176
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.755416466056813e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.377849676165851,
      "tokens_seen": 2686517248
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.754614026640989e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.3778368287326135,
      "tokens_seen": 2686648320
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7538115872251646e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.3778239821016296,
      "tokens_seen": 2686779392
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7530091478093406e-05,
      "loss": 1.3475,
      "theoretical_loss": 3.3778111362728103,
      "tokens_seen": 2686910464
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7522067083935165e-05,
      "loss": 1.23,
      "theoretical_loss": 3.377798291246067,
      "tokens_seen": 2687041536
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7514042689776925e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.37778544702131,
      "tokens_seen": 2687172608
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7506018295618685e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.3777726035984497,
      "tokens_seen": 2687303680
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7497993901460445e-05,
      "loss": 1.1935,
      "theoretical_loss": 3.3777597609773977,
      "tokens_seen": 2687434752
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7489969507302205e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.3777469191580645,
      "tokens_seen": 2687565824
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.748194511314396e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.377734078140361,
      "tokens_seen": 2687696896
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.747392071898572e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.377721237924198,
      "tokens_seen": 2687827968
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.746589632482748e-05,
      "loss": 1.1818,
      "theoretical_loss": 3.3777083985094865,
      "tokens_seen": 2687959040
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.745787193066924e-05,
      "loss": 1.254,
      "theoretical_loss": 3.3776955598961376,
      "tokens_seen": 2688090112
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7449847536511e-05,
      "loss": 1.26,
      "theoretical_loss": 3.377682722084061,
      "tokens_seen": 2688221184
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.744182314235276e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.3776698850731695,
      "tokens_seen": 2688352256
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.743379874819452e-05,
      "loss": 1.3481,
      "theoretical_loss": 3.377657048863373,
      "tokens_seen": 2688483328
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.742577435403627e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.377644213454582,
      "tokens_seen": 2688614400
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.741774995987803e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.3776313788467087,
      "tokens_seen": 2688745472
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.740972556571979e-05,
      "loss": 1.2035,
      "theoretical_loss": 3.377618545039663,
      "tokens_seen": 2688876544
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.740170117156155e-05,
      "loss": 1.3582,
      "theoretical_loss": 3.3776057120333567,
      "tokens_seen": 2689007616
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.739367677740331e-05,
      "loss": 1.2038,
      "theoretical_loss": 3.3775928798277,
      "tokens_seen": 2689138688
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.738565238324507e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.377580048422604,
      "tokens_seen": 2689269760
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.737762798908683e-05,
      "loss": 1.1966,
      "theoretical_loss": 3.377567217817981,
      "tokens_seen": 2689400832
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.736960359492859e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.3775543880137406,
      "tokens_seen": 2689531904
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.736157920077034e-05,
      "loss": 1.2137,
      "theoretical_loss": 3.377541559009795,
      "tokens_seen": 2689662976
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.73535548066121e-05,
      "loss": 1.205,
      "theoretical_loss": 3.3775287308060538,
      "tokens_seen": 2689794048
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.734553041245386e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.3775159034024296,
      "tokens_seen": 2689925120
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.733750601829562e-05,
      "loss": 1.28,
      "theoretical_loss": 3.377503076798833,
      "tokens_seen": 2690056192
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.732948162413738e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.3774902509951747,
      "tokens_seen": 2690187264
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.732145722997914e-05,
      "loss": 1.2817,
      "theoretical_loss": 3.3774774259913665,
      "tokens_seen": 2690318336
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.73134328358209e-05,
      "loss": 1.3583,
      "theoretical_loss": 3.377464601787319,
      "tokens_seen": 2690449408
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7305408441662656e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.3774517783829436,
      "tokens_seen": 2690580480
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7297384047504416e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.3774389557781515,
      "tokens_seen": 2690711552
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7289359653346176e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.377426133972854,
      "tokens_seen": 2690842624
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7281335259187936e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.377413312966962,
      "tokens_seen": 2690973696
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7273310865029695e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.377400492760387,
      "tokens_seen": 2691104768
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7265286470871455e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.37738767335304,
      "tokens_seen": 2691235840
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7257262076713215e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.377374854744833,
      "tokens_seen": 2691366912
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.724923768255497e-05,
      "loss": 1.3472,
      "theoretical_loss": 3.377362036935676,
      "tokens_seen": 2691497984
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.724121328839673e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.377349219925481,
      "tokens_seen": 2691629056
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.723318889423849e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.3773364037141596,
      "tokens_seen": 2691760128
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.722516450008025e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.377323588301623,
      "tokens_seen": 2691891200
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.9927344918251038,
      "objective/train/avg_token_score": 0.007265455089509487,
      "objective/train/docs_used": 978239,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2303799390792847,
      "objective/train/original_loss": 1.2410392761230469,
      "objective/train/theoretical_loss": 3.377310773687782,
      "objective/train/tokens_used": 1062481376,
      "objective/train/unlikelihood_loss": 0.017754778265953064,
      "theoretical_loss": 3.377310773687782,
      "tokens_seen": 2692022272
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.721714010592201e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.377310773687782,
      "tokens_seen": 2692022272
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.720911571176377e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.377297959872548,
      "tokens_seen": 2692153344
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.720109131760552e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.377285146855833,
      "tokens_seen": 2692284416
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.719306692344728e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.3772723346375475,
      "tokens_seen": 2692415488
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.718504252928904e-05,
      "loss": 1.3216,
      "theoretical_loss": 3.3772595232176035,
      "tokens_seen": 2692546560
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.71770181351308e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.3772467125959125,
      "tokens_seen": 2692677632
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.716899374097256e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.377233902772385,
      "tokens_seen": 2692808704
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.716096934681432e-05,
      "loss": 1.3903,
      "theoretical_loss": 3.377221093746934,
      "tokens_seen": 2692939776
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7152944952656074e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.377208285519469,
      "tokens_seen": 2693070848
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7144920558497834e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.377195478089903,
      "tokens_seen": 2693201920
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7136896164339593e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.377182671458147,
      "tokens_seen": 2693332992
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7128871770181353e-05,
      "loss": 1.3061,
      "theoretical_loss": 3.3771698656241123,
      "tokens_seen": 2693464064
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.712084737602311e-05,
      "loss": 1.242,
      "theoretical_loss": 3.3771570605877104,
      "tokens_seen": 2693595136
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.711282298186487e-05,
      "loss": 1.3907,
      "theoretical_loss": 3.3771442563488527,
      "tokens_seen": 2693726208
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7104798587706626e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.3771314529074514,
      "tokens_seen": 2693857280
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7096774193548386e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.377118650263417,
      "tokens_seen": 2693988352
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7088749799390146e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.377105848416662,
      "tokens_seen": 2694119424
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7080725405231906e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.3770930473670977,
      "tokens_seen": 2694250496
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.7072701011073666e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.3770802471146353,
      "tokens_seen": 2694381568
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.706467661691542e-05,
      "loss": 1.3313,
      "theoretical_loss": 3.3770674476591864,
      "tokens_seen": 2694512640
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.705665222275718e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.377054649000663,
      "tokens_seen": 2694643712
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.704862782859894e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.3770418511389773,
      "tokens_seen": 2694774784
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.70406034344407e-05,
      "loss": 1.2492,
      "theoretical_loss": 3.3770290540740393,
      "tokens_seen": 2694905856
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.703257904028246e-05,
      "loss": 1.2437,
      "theoretical_loss": 3.377016257805762,
      "tokens_seen": 2695036928
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.702455464612422e-05,
      "loss": 1.3298,
      "theoretical_loss": 3.3770034623340566,
      "tokens_seen": 2695168000
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.701653025196598e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.3769906676588346,
      "tokens_seen": 2695299072
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.700850585780773e-05,
      "loss": 1.2319,
      "theoretical_loss": 3.376977873780008,
      "tokens_seen": 2695430144
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.700048146364949e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3769650806974885,
      "tokens_seen": 2695561216
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.699245706949125e-05,
      "loss": 1.1985,
      "theoretical_loss": 3.3769522884111876,
      "tokens_seen": 2695692288
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.698443267533301e-05,
      "loss": 1.2709,
      "theoretical_loss": 3.3769394969210174,
      "tokens_seen": 2695823360
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.697640828117477e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.3769267062268886,
      "tokens_seen": 2695954432
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.696838388701653e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.3769139163287147,
      "tokens_seen": 2696085504
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.696035949285829e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.376901127226406,
      "tokens_seen": 2696216576
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6952335098700044e-05,
      "loss": 1.3609,
      "theoretical_loss": 3.376888338919875,
      "tokens_seen": 2696347648
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6944310704541804e-05,
      "loss": 1.294,
      "theoretical_loss": 3.376875551409033,
      "tokens_seen": 2696478720
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6936286310383564e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3768627646937928,
      "tokens_seen": 2696609792
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6928261916225324e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3768499787740653,
      "tokens_seen": 2696740864
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6920237522067084e-05,
      "loss": 1.256,
      "theoretical_loss": 3.3768371936497625,
      "tokens_seen": 2696871936
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6912213127908844e-05,
      "loss": 1.2927,
      "theoretical_loss": 3.376824409320797,
      "tokens_seen": 2697003008
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6904188733750604e-05,
      "loss": 1.2434,
      "theoretical_loss": 3.376811625787079,
      "tokens_seen": 2697134080
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6896164339592364e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.3767988430485225,
      "tokens_seen": 2697265152
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.688813994543412e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.3767860611050375,
      "tokens_seen": 2697396224
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6880115551275877e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.376773279956537,
      "tokens_seen": 2697527296
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.6872091157117637e-05,
      "loss": 1.2278,
      "theoretical_loss": 3.3767604996029332,
      "tokens_seen": 2697658368
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6864066762959396e-05,
      "loss": 1.2971,
      "theoretical_loss": 3.376747720044137,
      "tokens_seen": 2697789440
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6856042368801156e-05,
      "loss": 1.238,
      "theoretical_loss": 3.3767349412800614,
      "tokens_seen": 2697920512
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6848017974642916e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3767221633106175,
      "tokens_seen": 2698051584
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6839993580484676e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3767093861357176,
      "tokens_seen": 2698182656
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.683196918632643e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.3766966097552737,
      "tokens_seen": 2698313728
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.682394479216819e-05,
      "loss": 1.1883,
      "theoretical_loss": 3.3766838341691985,
      "tokens_seen": 2698444800
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.9915187358856201,
      "objective/train/avg_token_score": 0.008481224998831749,
      "objective/train/docs_used": 980656,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4161356687545776,
      "objective/train/original_loss": 1.4288616180419922,
      "objective/train/theoretical_loss": 3.376671059377403,
      "objective/train/tokens_used": 1069034976,
      "objective/train/unlikelihood_loss": 0.019608987495303154,
      "theoretical_loss": 3.376671059377403,
      "tokens_seen": 2698575872
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.681592039800995e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.376671059377403,
      "tokens_seen": 2698575872
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.680789600385171e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.3766582853797997,
      "tokens_seen": 2698706944
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.679987160969347e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.3766455121763004,
      "tokens_seen": 2698838016
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.679184721553523e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.376632739766818,
      "tokens_seen": 2698969088
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.678382282137699e-05,
      "loss": 1.3507,
      "theoretical_loss": 3.376619968151263,
      "tokens_seen": 2699100160
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.677579842721875e-05,
      "loss": 1.2495,
      "theoretical_loss": 3.376607197329549,
      "tokens_seen": 2699231232
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.67677740330605e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.3765944273015878,
      "tokens_seen": 2699362304
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.675974963890226e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.376581658067291,
      "tokens_seen": 2699493376
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.675172524474402e-05,
      "loss": 1.255,
      "theoretical_loss": 3.3765688896265713,
      "tokens_seen": 2699624448
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.674370085058578e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.3765561219793403,
      "tokens_seen": 2699755520
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.673567645642754e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.376543355125511,
      "tokens_seen": 2699886592
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.67276520622693e-05,
      "loss": 1.261,
      "theoretical_loss": 3.376530589064995,
      "tokens_seen": 2700017664
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.671962766811106e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.376517823797704,
      "tokens_seen": 2700148736
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6711603273952814e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.376505059323551,
      "tokens_seen": 2700279808
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6703578879794574e-05,
      "loss": 1.2662,
      "theoretical_loss": 3.3764922956424477,
      "tokens_seen": 2700410880
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6695554485636334e-05,
      "loss": 1.304,
      "theoretical_loss": 3.376479532754307,
      "tokens_seen": 2700541952
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6687530091478094e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.3764667706590408,
      "tokens_seen": 2700673024
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6679505697319854e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.376454009356561,
      "tokens_seen": 2700804096
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6671481303161614e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.3764412488467808,
      "tokens_seen": 2700935168
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6663456909003374e-05,
      "loss": 1.323,
      "theoretical_loss": 3.3764284891296117,
      "tokens_seen": 2701066240
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.665543251484513e-05,
      "loss": 1.205,
      "theoretical_loss": 3.376415730204966,
      "tokens_seen": 2701197312
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.664740812068689e-05,
      "loss": 1.3754,
      "theoretical_loss": 3.3764029720727566,
      "tokens_seen": 2701328384
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.663938372652865e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.3763902147328952,
      "tokens_seen": 2701459456
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6631359332370407e-05,
      "loss": 1.3031,
      "theoretical_loss": 3.3763774581852948,
      "tokens_seen": 2701590528
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6623334938212166e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.376364702429867,
      "tokens_seen": 2701721600
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6615310544053926e-05,
      "loss": 1.329,
      "theoretical_loss": 3.3763519474665245,
      "tokens_seen": 2701852672
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6607286149895686e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.37633919329518,
      "tokens_seen": 2701983744
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6599261755737446e-05,
      "loss": 1.2881,
      "theoretical_loss": 3.3763264399157458,
      "tokens_seen": 2702114816
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.65912373615792e-05,
      "loss": 1.2407,
      "theoretical_loss": 3.376313687328134,
      "tokens_seen": 2702245888
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.658321296742096e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.3763009355322575,
      "tokens_seen": 2702376960
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.657518857326272e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.376288184528028,
      "tokens_seen": 2702508032
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.656716417910448e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3762754343153594,
      "tokens_seen": 2702639104
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.655913978494624e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.3762626848941624,
      "tokens_seen": 2702770176
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6551115390788e-05,
      "loss": 1.3394,
      "theoretical_loss": 3.3762499362643505,
      "tokens_seen": 2702901248
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.654309099662976e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.376237188425836,
      "tokens_seen": 2703032320
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.653506660247151e-05,
      "loss": 1.251,
      "theoretical_loss": 3.3762244413785316,
      "tokens_seen": 2703163392
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.652704220831327e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3762116951223495,
      "tokens_seen": 2703294464
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.651901781415503e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.3761989496572022,
      "tokens_seen": 2703425536
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.651099341999679e-05,
      "loss": 1.3149,
      "theoretical_loss": 3.376186204983003,
      "tokens_seen": 2703556608
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.650296902583855e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.376173461099664,
      "tokens_seen": 2703687680
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.649494463168031e-05,
      "loss": 1.261,
      "theoretical_loss": 3.376160718007097,
      "tokens_seen": 2703818752
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.648692023752207e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.3761479757052157,
      "tokens_seen": 2703949824
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6478895843363824e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.3761352341939324,
      "tokens_seen": 2704080896
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6470871449205584e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.3761224934731597,
      "tokens_seen": 2704211968
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6462847055047344e-05,
      "loss": 1.3339,
      "theoretical_loss": 3.3761097535428104,
      "tokens_seen": 2704343040
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6454822660889104e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.376097014402797,
      "tokens_seen": 2704474112
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6446798266730864e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.3760842760530316,
      "tokens_seen": 2704605184
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6438773872572624e-05,
      "loss": 1.331,
      "theoretical_loss": 3.376071538493428,
      "tokens_seen": 2704736256
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6430749478414384e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.376058801723898,
      "tokens_seen": 2704867328
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6422725084256144e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.3760460657443545,
      "tokens_seen": 2704998400
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.9928800463676453,
      "objective/train/avg_token_score": 0.007119883317500353,
      "objective/train/docs_used": 983087,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4221529960632324,
      "objective/train/original_loss": 1.4327404499053955,
      "objective/train/theoretical_loss": 3.3760333305547108,
      "objective/train/tokens_used": 1075588576,
      "objective/train/unlikelihood_loss": 0.017439723014831543,
      "theoretical_loss": 3.3760333305547108,
      "tokens_seen": 2705129472
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.64147006900979e-05,
      "loss": 1.4111,
      "theoretical_loss": 3.3760333305547108,
      "tokens_seen": 2705129472
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.640667629593966e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.3760205961548793,
      "tokens_seen": 2705260544
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.639865190178142e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.376007862544772,
      "tokens_seen": 2705391616
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.639062750762318e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.375995129724303,
      "tokens_seen": 2705522688
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6382603113464937e-05,
      "loss": 1.243,
      "theoretical_loss": 3.375982397693384,
      "tokens_seen": 2705653760
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6374578719306696e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.3759696664519288,
      "tokens_seen": 2705784832
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6366554325148456e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.3759569359998487,
      "tokens_seen": 2705915904
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.635852993099021e-05,
      "loss": 1.314,
      "theoretical_loss": 3.375944206337058,
      "tokens_seen": 2706046976
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.635050553683197e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.375931477463469,
      "tokens_seen": 2706178048
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.634248114267373e-05,
      "loss": 1.268,
      "theoretical_loss": 3.3759187493789944,
      "tokens_seen": 2706309120
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.633445674851549e-05,
      "loss": 1.25,
      "theoretical_loss": 3.375906022083547,
      "tokens_seen": 2706440192
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.632643235435725e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.37589329557704,
      "tokens_seen": 2706571264
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.631840796019901e-05,
      "loss": 1.3564,
      "theoretical_loss": 3.375880569859387,
      "tokens_seen": 2706702336
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.631038356604077e-05,
      "loss": 1.3527,
      "theoretical_loss": 3.3758678449304993,
      "tokens_seen": 2706833408
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.630235917188253e-05,
      "loss": 1.3676,
      "theoretical_loss": 3.3758551207902903,
      "tokens_seen": 2706964480
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.629433477772428e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.3758423974386735,
      "tokens_seen": 2707095552
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.628631038356604e-05,
      "loss": 1.3992,
      "theoretical_loss": 3.3758296748755616,
      "tokens_seen": 2707226624
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.62782859894078e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.375816953100868,
      "tokens_seen": 2707357696
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.627026159524956e-05,
      "loss": 1.3961,
      "theoretical_loss": 3.3758042321145045,
      "tokens_seen": 2707488768
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.626223720109132e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.3757915119163853,
      "tokens_seen": 2707619840
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.625421280693308e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.3757787925064227,
      "tokens_seen": 2707750912
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.624618841277484e-05,
      "loss": 1.284,
      "theoretical_loss": 3.3757660738845296,
      "tokens_seen": 2707881984
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6238164018616595e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.3757533560506197,
      "tokens_seen": 2708013056
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6230139624458354e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.375740639004606,
      "tokens_seen": 2708144128
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6222115230300114e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.3757279227464005,
      "tokens_seen": 2708275200
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6214090836141874e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.375715207275918,
      "tokens_seen": 2708406272
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6206066441983634e-05,
      "loss": 1.3579,
      "theoretical_loss": 3.37570249259307,
      "tokens_seen": 2708537344
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6198042047825394e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.3756897786977698,
      "tokens_seen": 2708668416
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6190017653667154e-05,
      "loss": 1.2153,
      "theoretical_loss": 3.3756770655899313,
      "tokens_seen": 2708799488
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.618199325950891e-05,
      "loss": 1.4005,
      "theoretical_loss": 3.3756643532694675,
      "tokens_seen": 2708930560
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.617396886535067e-05,
      "loss": 1.376,
      "theoretical_loss": 3.375651641736291,
      "tokens_seen": 2709061632
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.616594447119243e-05,
      "loss": 1.3678,
      "theoretical_loss": 3.3756389309903154,
      "tokens_seen": 2709192704
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.615792007703419e-05,
      "loss": 1.3464,
      "theoretical_loss": 3.375626221031454,
      "tokens_seen": 2709323776
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.614989568287595e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.375613511859619,
      "tokens_seen": 2709454848
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.614187128871771e-05,
      "loss": 1.3419,
      "theoretical_loss": 3.375600803474725,
      "tokens_seen": 2709585920
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6133846894559467e-05,
      "loss": 1.3773,
      "theoretical_loss": 3.3755880958766844,
      "tokens_seen": 2709716992
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6125822500401226e-05,
      "loss": 1.3252,
      "theoretical_loss": 3.37557538906541,
      "tokens_seen": 2709848064
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.611779810624298e-05,
      "loss": 1.26,
      "theoretical_loss": 3.375562683040816,
      "tokens_seen": 2709979136
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.610977371208474e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.375549977802815,
      "tokens_seen": 2710110208
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.61017493179265e-05,
      "loss": 1.343,
      "theoretical_loss": 3.3755372733513207,
      "tokens_seen": 2710241280
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.609372492376826e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.375524569686246,
      "tokens_seen": 2710372352
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.608570052961002e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.3755118668075044,
      "tokens_seen": 2710503424
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.607767613545178e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.375499164715009,
      "tokens_seen": 2710634496
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.606965174129354e-05,
      "loss": 1.3946,
      "theoretical_loss": 3.3754864634086736,
      "tokens_seen": 2710765568
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.606162734713529e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.3754737628884115,
      "tokens_seen": 2710896640
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.605360295297705e-05,
      "loss": 1.3379,
      "theoretical_loss": 3.375461063154135,
      "tokens_seen": 2711027712
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.604557855881881e-05,
      "loss": 1.308,
      "theoretical_loss": 3.3754483642057584,
      "tokens_seen": 2711158784
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.603755416466057e-05,
      "loss": 1.2014,
      "theoretical_loss": 3.3754356660431952,
      "tokens_seen": 2711289856
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.602952977050233e-05,
      "loss": 1.3556,
      "theoretical_loss": 3.375422968666358,
      "tokens_seen": 2711420928
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.602150537634409e-05,
      "loss": 1.4083,
      "theoretical_loss": 3.3754102720751615,
      "tokens_seen": 2711552000
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.9941312074661255,
      "objective/train/avg_token_score": 0.005868761800229549,
      "objective/train/docs_used": 985372,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3065677881240845,
      "objective/train/original_loss": 1.3138186931610107,
      "objective/train/theoretical_loss": 3.375397576269518,
      "objective/train/tokens_used": 1082142176,
      "objective/train/unlikelihood_loss": 0.01792612299323082,
      "theoretical_loss": 3.375397576269518,
      "tokens_seen": 2711683072
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.601348098218585e-05,
      "loss": 1.3217,
      "theoretical_loss": 3.375397576269518,
      "tokens_seen": 2711683072
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.6005456588027605e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.3753848812493406,
      "tokens_seen": 2711814144
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5997432193869365e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.375372187014544,
      "tokens_seen": 2711945216
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5989407799711125e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.375359493565041,
      "tokens_seen": 2712076288
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5981383405552884e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.375346800900745,
      "tokens_seen": 2712207360
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5973359011394644e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.375334109021569,
      "tokens_seen": 2712338432
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5965334617236404e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.3753214179274282,
      "tokens_seen": 2712469504
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5957310223078164e-05,
      "loss": 1.33,
      "theoretical_loss": 3.3753087276182345,
      "tokens_seen": 2712600576
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.5949285828919924e-05,
      "loss": 1.4672,
      "theoretical_loss": 3.375296038093902,
      "tokens_seen": 2712731648
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.594126143476168e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.3752833493543437,
      "tokens_seen": 2712862720
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.593323704060344e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.3752706613994743,
      "tokens_seen": 2712993792
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.59252126464452e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.375257974229206,
      "tokens_seen": 2713124864
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.591718825228696e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3752452878434536,
      "tokens_seen": 2713255936
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.590916385812872e-05,
      "loss": 1.4021,
      "theoretical_loss": 3.37523260224213,
      "tokens_seen": 2713387008
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.590113946397048e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.3752199174251487,
      "tokens_seen": 2713518080
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.589311506981223e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.375207233392424,
      "tokens_seen": 2713649152
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.588509067565399e-05,
      "loss": 1.4439,
      "theoretical_loss": 3.375194550143869,
      "tokens_seen": 2713780224
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.587706628149575e-05,
      "loss": 1.3764,
      "theoretical_loss": 3.375181867679397,
      "tokens_seen": 2713911296
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.586904188733751e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.375169185998923,
      "tokens_seen": 2714042368
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.586101749317927e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.3751565051023595,
      "tokens_seen": 2714173440
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.585299309902103e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.3751438249896206,
      "tokens_seen": 2714304512
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.584496870486278e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.3751311456606197,
      "tokens_seen": 2714435584
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.583694431070454e-05,
      "loss": 1.4214,
      "theoretical_loss": 3.3751184671152705,
      "tokens_seen": 2714566656
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.58289199165463e-05,
      "loss": 1.318,
      "theoretical_loss": 3.3751057893534875,
      "tokens_seen": 2714697728
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.582089552238806e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.3750931123751835,
      "tokens_seen": 2714828800
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.581287112822982e-05,
      "loss": 1.2866,
      "theoretical_loss": 3.375080436180273,
      "tokens_seen": 2714959872
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.580484673407158e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.375067760768669,
      "tokens_seen": 2715090944
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5796822339913335e-05,
      "loss": 1.3272,
      "theoretical_loss": 3.3750550861402857,
      "tokens_seen": 2715222016
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5788797945755095e-05,
      "loss": 1.1797,
      "theoretical_loss": 3.3750424122950373,
      "tokens_seen": 2715353088
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5780773551596855e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.3750297392328372,
      "tokens_seen": 2715484160
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5772749157438615e-05,
      "loss": 1.3191,
      "theoretical_loss": 3.3750170669535993,
      "tokens_seen": 2715615232
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5764724763280375e-05,
      "loss": 1.1867,
      "theoretical_loss": 3.375004395457237,
      "tokens_seen": 2715746304
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5756700369122135e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3749917247436647,
      "tokens_seen": 2715877376
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.574867597496389e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.3749790548127963,
      "tokens_seen": 2716008448
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.574065158080565e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.3749663856645453,
      "tokens_seen": 2716139520
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.573262718664741e-05,
      "loss": 1.299,
      "theoretical_loss": 3.374953717298826,
      "tokens_seen": 2716270592
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.572460279248917e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.3749410497155523,
      "tokens_seen": 2716401664
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.571657839833093e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3749283829146375,
      "tokens_seen": 2716532736
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.570855400417269e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.374915716895996,
      "tokens_seen": 2716663808
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.570052961001444e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.374903051659542,
      "tokens_seen": 2716794880
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.56925052158562e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.374890387205189,
      "tokens_seen": 2716925952
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.568448082169796e-05,
      "loss": 1.358,
      "theoretical_loss": 3.374877723532851,
      "tokens_seen": 2717057024
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.567645642753972e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.3748650606424424,
      "tokens_seen": 2717188096
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.566843203338148e-05,
      "loss": 1.2949,
      "theoretical_loss": 3.3748523985338768,
      "tokens_seen": 2717319168
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.566040763922324e-05,
      "loss": 1.3296,
      "theoretical_loss": 3.3748397372070684,
      "tokens_seen": 2717450240
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5652383245065e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.3748270766619313,
      "tokens_seen": 2717581312
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.564435885090675e-05,
      "loss": 1.2885,
      "theoretical_loss": 3.374814416898379,
      "tokens_seen": 2717712384
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.563633445674851e-05,
      "loss": 1.206,
      "theoretical_loss": 3.3748017579163267,
      "tokens_seen": 2717843456
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.562831006259027e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.374789099715687,
      "tokens_seen": 2717974528
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.562028566843203e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.3747764422963753,
      "tokens_seen": 2718105600
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.9934532642364502,
      "objective/train/avg_token_score": 0.0065467203967273235,
      "objective/train/docs_used": 987833,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3913750648498535,
      "objective/train/original_loss": 1.3997931480407715,
      "objective/train/theoretical_loss": 3.3747637856583053,
      "objective/train/tokens_used": 1088695776,
      "objective/train/unlikelihood_loss": 0.01991441287100315,
      "theoretical_loss": 3.3747637856583053,
      "tokens_seen": 2718236672
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.561226127427379e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.3747637856583053,
      "tokens_seen": 2718236672
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.560423688011555e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.3747511298013904,
      "tokens_seen": 2718367744
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.559621248595731e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.374738474725546,
      "tokens_seen": 2718498816
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5588188091799066e-05,
      "loss": 1.1952,
      "theoretical_loss": 3.374725820430685,
      "tokens_seen": 2718629888
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5580163697640826e-05,
      "loss": 1.261,
      "theoretical_loss": 3.374713166916722,
      "tokens_seen": 2718760960
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5572139303482585e-05,
      "loss": 1.2117,
      "theoretical_loss": 3.3747005141835715,
      "tokens_seen": 2718892032
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5564114909324345e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.3746878622311476,
      "tokens_seen": 2719023104
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5556090515166105e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.3746752110593636,
      "tokens_seen": 2719154176
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5548066121007865e-05,
      "loss": 1.2451,
      "theoretical_loss": 3.3746625606681353,
      "tokens_seen": 2719285248
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5540041726849625e-05,
      "loss": 1.2819,
      "theoretical_loss": 3.3746499110573756,
      "tokens_seen": 2719416320
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.553201733269138e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.374637262226999,
      "tokens_seen": 2719547392
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.552399293853314e-05,
      "loss": 1.3538,
      "theoretical_loss": 3.3746246141769207,
      "tokens_seen": 2719678464
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.55159685443749e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.374611966907054,
      "tokens_seen": 2719809536
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.550794415021666e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.3745993204173126,
      "tokens_seen": 2719940608
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.549991975605842e-05,
      "loss": 1.3647,
      "theoretical_loss": 3.3745866747076123,
      "tokens_seen": 2720071680
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.549189536190018e-05,
      "loss": 1.403,
      "theoretical_loss": 3.3745740297778664,
      "tokens_seen": 2720202752
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.548387096774194e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.3745613856279895,
      "tokens_seen": 2720333824
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.54758465735837e-05,
      "loss": 1.239,
      "theoretical_loss": 3.374548742257896,
      "tokens_seen": 2720464896
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.546782217942545e-05,
      "loss": 1.3601,
      "theoretical_loss": 3.3745360996675,
      "tokens_seen": 2720595968
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.545979778526721e-05,
      "loss": 1.25,
      "theoretical_loss": 3.374523457856716,
      "tokens_seen": 2720727040
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.545177339110897e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.3745108168254587,
      "tokens_seen": 2720858112
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.544374899695073e-05,
      "loss": 1.3558,
      "theoretical_loss": 3.374498176573642,
      "tokens_seen": 2720989184
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.543572460279249e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.3744855371011804,
      "tokens_seen": 2721120256
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.542770020863425e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.3744728984079884,
      "tokens_seen": 2721251328
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.541967581447601e-05,
      "loss": 1.3785,
      "theoretical_loss": 3.37446026049398,
      "tokens_seen": 2721382400
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.541165142031776e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.3744476233590706,
      "tokens_seen": 2721513472
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.540362702615952e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.3744349870031742,
      "tokens_seen": 2721644544
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.539560263200128e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.3744223514262046,
      "tokens_seen": 2721775616
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.538757823784304e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.3744097166280773,
      "tokens_seen": 2721906688
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.53795538436848e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.374397082608706,
      "tokens_seen": 2722037760
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.537152944952656e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.374384449368006,
      "tokens_seen": 2722168832
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.536350505536832e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.374371816905891,
      "tokens_seen": 2722299904
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.535548066121008e-05,
      "loss": 1.237,
      "theoretical_loss": 3.374359185222276,
      "tokens_seen": 2722430976
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5347456267051836e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.374346554317075,
      "tokens_seen": 2722562048
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5339431872893596e-05,
      "loss": 1.2115,
      "theoretical_loss": 3.3743339241902035,
      "tokens_seen": 2722693120
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5331407478735356e-05,
      "loss": 1.3949,
      "theoretical_loss": 3.3743212948415753,
      "tokens_seen": 2722824192
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5323383084577115e-05,
      "loss": 1.252,
      "theoretical_loss": 3.374308666271105,
      "tokens_seen": 2722955264
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5315358690418875e-05,
      "loss": 1.257,
      "theoretical_loss": 3.374296038478708,
      "tokens_seen": 2723086336
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5307334296260635e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.3742834114642983,
      "tokens_seen": 2723217408
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5299309902102395e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.3742707852277904,
      "tokens_seen": 2723348480
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.529128550794415e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.3742581597690986,
      "tokens_seen": 2723479552
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.528326111378591e-05,
      "loss": 1.2468,
      "theoretical_loss": 3.374245535088139,
      "tokens_seen": 2723610624
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.527523671962767e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.3742329111848246,
      "tokens_seen": 2723741696
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.526721232546943e-05,
      "loss": 1.268,
      "theoretical_loss": 3.3742202880590715,
      "tokens_seen": 2723872768
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.525918793131119e-05,
      "loss": 1.304,
      "theoretical_loss": 3.3742076657107933,
      "tokens_seen": 2724003840
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.525116353715295e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.3741950441399053,
      "tokens_seen": 2724134912
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.524313914299471e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.374182423346322,
      "tokens_seen": 2724265984
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.523511474883646e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.3741698033299583,
      "tokens_seen": 2724397056
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.522709035467822e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.3741571840907287,
      "tokens_seen": 2724528128
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.521906596051998e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.374144565628548,
      "tokens_seen": 2724659200
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.9909141063690186,
      "objective/train/avg_token_score": 0.009085885249078274,
      "objective/train/docs_used": 990115,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3129764795303345,
      "objective/train/original_loss": 1.3259563446044922,
      "objective/train/theoretical_loss": 3.374131947943331,
      "objective/train/tokens_used": 1095249376,
      "objective/train/unlikelihood_loss": 0.021410929039120674,
      "theoretical_loss": 3.374131947943331,
      "tokens_seen": 2724790272
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.521104156636174e-05,
      "loss": 1.303,
      "theoretical_loss": 3.374131947943331,
      "tokens_seen": 2724790272
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.52030171722035e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.3741193310349926,
      "tokens_seen": 2724921344
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.519499277804526e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.374106714903448,
      "tokens_seen": 2725052416
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.518696838388702e-05,
      "loss": 1.315,
      "theoretical_loss": 3.374094099548611,
      "tokens_seen": 2725183488
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.517894398972878e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.3740814849703975,
      "tokens_seen": 2725314560
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.517091959557053e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.3740688711687215,
      "tokens_seen": 2725445632
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.516289520141229e-05,
      "loss": 1.3215,
      "theoretical_loss": 3.374056258143498,
      "tokens_seen": 2725576704
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.515487080725405e-05,
      "loss": 1.274,
      "theoretical_loss": 3.3740436458946426,
      "tokens_seen": 2725707776
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.514684641309581e-05,
      "loss": 1.308,
      "theoretical_loss": 3.3740310344220696,
      "tokens_seen": 2725838848
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.513882201893757e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.3740184237256936,
      "tokens_seen": 2725969920
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.513079762477933e-05,
      "loss": 1.3204,
      "theoretical_loss": 3.37400581380543,
      "tokens_seen": 2726100992
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.512277323062109e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.373993204661194,
      "tokens_seen": 2726232064
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5114748836462846e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.3739805962928995,
      "tokens_seen": 2726363136
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5106724442304606e-05,
      "loss": 1.3525,
      "theoretical_loss": 3.3739679887004623,
      "tokens_seen": 2726494208
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5098700048146366e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.373955381883797,
      "tokens_seen": 2726625280
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5090675653988126e-05,
      "loss": 1.3413,
      "theoretical_loss": 3.373942775842819,
      "tokens_seen": 2726756352
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5082651259829885e-05,
      "loss": 1.253,
      "theoretical_loss": 3.3739301705774425,
      "tokens_seen": 2726887424
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5074626865671645e-05,
      "loss": 1.2739,
      "theoretical_loss": 3.3739175660875835,
      "tokens_seen": 2727018496
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.5066602471513405e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.3739049623731563,
      "tokens_seen": 2727149568
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.505857807735516e-05,
      "loss": 1.1811,
      "theoretical_loss": 3.373892359434076,
      "tokens_seen": 2727280640
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.505055368319692e-05,
      "loss": 1.28,
      "theoretical_loss": 3.373879757270258,
      "tokens_seen": 2727411712
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.504252928903868e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.373867155881617,
      "tokens_seen": 2727542784
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.503450489488044e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.3738545552680685,
      "tokens_seen": 2727673856
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.50264805007222e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.373841955429527,
      "tokens_seen": 2727804928
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.501845610656396e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.373829356365908,
      "tokens_seen": 2727936000
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.501043171240572e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.3738167580771266,
      "tokens_seen": 2728067072
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.500240731824748e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.3738041605630977,
      "tokens_seen": 2728198144
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.499438292408923e-05,
      "loss": 1.257,
      "theoretical_loss": 3.3737915638237364,
      "tokens_seen": 2728329216
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.498635852993099e-05,
      "loss": 1.3264,
      "theoretical_loss": 3.373778967858958,
      "tokens_seen": 2728460288
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.497833413577275e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.373766372668678,
      "tokens_seen": 2728591360
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.497030974161451e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3737537782528113,
      "tokens_seen": 2728722432
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.496228534745627e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.373741184611273,
      "tokens_seen": 2728853504
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.495426095329803e-05,
      "loss": 1.29,
      "theoretical_loss": 3.373728591743978,
      "tokens_seen": 2728984576
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.494623655913979e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.373715999650842,
      "tokens_seen": 2729115648
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.4938212164981543e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.3737034083317803,
      "tokens_seen": 2729246720
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.49301877708233e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.373690817786708,
      "tokens_seen": 2729377792
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.492216337666506e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.37367822801554,
      "tokens_seen": 2729508864
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.491413898250682e-05,
      "loss": 1.3256,
      "theoretical_loss": 3.373665639018192,
      "tokens_seen": 2729639936
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.490611458834858e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3736530507945792,
      "tokens_seen": 2729771008
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.489809019419034e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.373640463344617,
      "tokens_seen": 2729902080
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.48900658000321e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.37362787666822,
      "tokens_seen": 2730033152
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.488204140587386e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.3736152907653043,
      "tokens_seen": 2730164224
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.4874017011715616e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.3736027056357853,
      "tokens_seen": 2730295296
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.4865992617557376e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.373590121279578,
      "tokens_seen": 2730426368
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.4857968223399136e-05,
      "loss": 1.3855,
      "theoretical_loss": 3.373577537696598,
      "tokens_seen": 2730557440
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.4849943829240896e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.37356495488676,
      "tokens_seen": 2730688512
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4841919435082656e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.3735523728499803,
      "tokens_seen": 2730819584
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4833895040924415e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.373539791586174,
      "tokens_seen": 2730950656
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4825870646766175e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.373527211095256,
      "tokens_seen": 2731081728
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.481784625260793e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.3735146313771427,
      "tokens_seen": 2731212800
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.9928444623947144,
      "objective/train/avg_token_score": 0.007155559025704861,
      "objective/train/docs_used": 992345,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3333886861801147,
      "objective/train/original_loss": 1.343772530555725,
      "objective/train/theoretical_loss": 3.3735020524317485,
      "objective/train/tokens_used": 1101802976,
      "objective/train/unlikelihood_loss": 0.017148640006780624,
      "theoretical_loss": 3.3735020524317485,
      "tokens_seen": 2731343872
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.480982185844969e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.3735020524317485,
      "tokens_seen": 2731343872
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.480179746429145e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.3734894742589896,
      "tokens_seen": 2731474944
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.479377307013321e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.373476896858781,
      "tokens_seen": 2731606016
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.478574867597497e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.3734643202310384,
      "tokens_seen": 2731737088
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.477772428181673e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.3734517443756777,
      "tokens_seen": 2731868160
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.476969988765849e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.373439169292614,
      "tokens_seen": 2731999232
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.476167549350024e-05,
      "loss": 1.3536,
      "theoretical_loss": 3.3734265949817623,
      "tokens_seen": 2732130304
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4753651099342e-05,
      "loss": 1.2786,
      "theoretical_loss": 3.3734140214430393,
      "tokens_seen": 2732261376
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.474562670518376e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.3734014486763595,
      "tokens_seen": 2732392448
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.473760231102552e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.373388876681639,
      "tokens_seen": 2732523520
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.472957791686728e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.3733763054587933,
      "tokens_seen": 2732654592
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.472155352270904e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.373363735007738,
      "tokens_seen": 2732785664
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.47135291285508e-05,
      "loss": 1.2701,
      "theoretical_loss": 3.3733511653283887,
      "tokens_seen": 2732916736
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.470550473439256e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.373338596420661,
      "tokens_seen": 2733047808
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4697480340234314e-05,
      "loss": 1.284,
      "theoretical_loss": 3.3733260282844704,
      "tokens_seen": 2733178880
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4689455946076073e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.3733134609197326,
      "tokens_seen": 2733309952
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.468143155191783e-05,
      "loss": 1.1706,
      "theoretical_loss": 3.3733008943263636,
      "tokens_seen": 2733441024
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.467340715775959e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.3732883285042785,
      "tokens_seen": 2733572096
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.466538276360135e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.3732757634533934,
      "tokens_seen": 2733703168
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.465735836944311e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.3732631991736235,
      "tokens_seen": 2733834240
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.464933397528487e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.3732506356648853,
      "tokens_seen": 2733965312
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4641309581126626e-05,
      "loss": 1.3518,
      "theoretical_loss": 3.373238072927094,
      "tokens_seen": 2734096384
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4633285186968386e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.3732255109601654,
      "tokens_seen": 2734227456
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4625260792810146e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.373212949764015,
      "tokens_seen": 2734358528
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4617236398651906e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.373200389338559,
      "tokens_seen": 2734489600
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4609212004493666e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.373187829683713,
      "tokens_seen": 2734620672
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4601187610335426e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.373175270799393,
      "tokens_seen": 2734751744
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4593163216177186e-05,
      "loss": 1.1977,
      "theoretical_loss": 3.3731627126855144,
      "tokens_seen": 2734882816
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.458513882201894e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.3731501553419934,
      "tokens_seen": 2735013888
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.45771144278607e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.3731375987687455,
      "tokens_seen": 2735144960
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.456909003370246e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.373125042965687,
      "tokens_seen": 2735276032
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.456106563954422e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.373112487932733,
      "tokens_seen": 2735407104
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.455304124538598e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.3730999336698,
      "tokens_seen": 2735538176
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.454501685122774e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.3730873801768038,
      "tokens_seen": 2735669248
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.453699245706949e-05,
      "loss": 1.2259,
      "theoretical_loss": 3.37307482745366,
      "tokens_seen": 2735800320
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.452896806291125e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.373062275500285,
      "tokens_seen": 2735931392
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.452094366875301e-05,
      "loss": 1.1992,
      "theoretical_loss": 3.3730497243165942,
      "tokens_seen": 2736062464
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.451291927459477e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.3730371739025036,
      "tokens_seen": 2736193536
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.450489488043653e-05,
      "loss": 1.2403,
      "theoretical_loss": 3.3730246242579294,
      "tokens_seen": 2736324608
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.449687048627829e-05,
      "loss": 1.2227,
      "theoretical_loss": 3.3730120753827877,
      "tokens_seen": 2736455680
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4488846092120044e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.3729995272769937,
      "tokens_seen": 2736586752
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4480821697961804e-05,
      "loss": 1.2094,
      "theoretical_loss": 3.3729869799404644,
      "tokens_seen": 2736717824
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4472797303803564e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.3729744333731153,
      "tokens_seen": 2736848896
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4464772909645324e-05,
      "loss": 1.3647,
      "theoretical_loss": 3.3729618875748626,
      "tokens_seen": 2736979968
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4456748515487084e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.372949342545622,
      "tokens_seen": 2737111040
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4448724121328844e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3729367982853096,
      "tokens_seen": 2737242112
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.44406997271706e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.3729242547938414,
      "tokens_seen": 2737373184
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4432675333012357e-05,
      "loss": 1.4011,
      "theoretical_loss": 3.372911712071134,
      "tokens_seen": 2737504256
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4424650938854116e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.3728991701171025,
      "tokens_seen": 2737635328
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4416626544695876e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.3728866289316644,
      "tokens_seen": 2737766400
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.9949920177459717,
      "objective/train/avg_token_score": 0.00500798923894763,
      "objective/train/docs_used": 994721,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1675324440002441,
      "objective/train/original_loss": 1.1745431423187256,
      "objective/train/theoretical_loss": 3.3728740885147346,
      "objective/train/tokens_used": 1108356576,
      "objective/train/unlikelihood_loss": 0.015171666629612446,
      "theoretical_loss": 3.3728740885147346,
      "tokens_seen": 2737897472
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4408602150537636e-05,
      "loss": 1.2132,
      "theoretical_loss": 3.3728740885147346,
      "tokens_seen": 2737897472
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4400577756379396e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.3728615488662292,
      "tokens_seen": 2738028544
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.439255336222115e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.3728490099860653,
      "tokens_seen": 2738159616
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.438452896806291e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.3728364718741584,
      "tokens_seen": 2738290688
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.437650457390467e-05,
      "loss": 1.2067,
      "theoretical_loss": 3.372823934530425,
      "tokens_seen": 2738421760
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.436848017974643e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.3728113979547807,
      "tokens_seen": 2738552832
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.436045578558819e-05,
      "loss": 1.2056,
      "theoretical_loss": 3.3727988621471416,
      "tokens_seen": 2738683904
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.435243139142995e-05,
      "loss": 1.3453,
      "theoretical_loss": 3.372786327107425,
      "tokens_seen": 2738814976
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.43444069972717e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.372773792835546,
      "tokens_seen": 2738946048
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.433638260311346e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.372761259331422,
      "tokens_seen": 2739077120
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.432835820895522e-05,
      "loss": 1.341,
      "theoretical_loss": 3.372748726594968,
      "tokens_seen": 2739208192
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.432033381479698e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.3727361946261003,
      "tokens_seen": 2739339264
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.431230942063874e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.372723663424736,
      "tokens_seen": 2739470336
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.43042850264805e-05,
      "loss": 1.2157,
      "theoretical_loss": 3.372711132990791,
      "tokens_seen": 2739601408
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.429626063232226e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.3726986033241815,
      "tokens_seen": 2739732480
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4288236238164015e-05,
      "loss": 1.3332,
      "theoretical_loss": 3.372686074424824,
      "tokens_seen": 2739863552
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4280211844005774e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.3726735462926345,
      "tokens_seen": 2739994624
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4272187449847534e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.3726610189275297,
      "tokens_seen": 2740125696
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4264163055689294e-05,
      "loss": 1.3601,
      "theoretical_loss": 3.3726484923294255,
      "tokens_seen": 2740256768
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4256138661531054e-05,
      "loss": 1.2303,
      "theoretical_loss": 3.3726359664982386,
      "tokens_seen": 2740387840
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4248114267372814e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.372623441433886,
      "tokens_seen": 2740518912
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4240089873214574e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3726109171362824,
      "tokens_seen": 2740649984
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4232065479056334e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.3725983936053456,
      "tokens_seen": 2740781056
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.422404108489809e-05,
      "loss": 1.1812,
      "theoretical_loss": 3.3725858708409913,
      "tokens_seen": 2740912128
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.421601669073985e-05,
      "loss": 1.1929,
      "theoretical_loss": 3.3725733488431366,
      "tokens_seen": 2741043200
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.420799229658161e-05,
      "loss": 1.3719,
      "theoretical_loss": 3.372560827611697,
      "tokens_seen": 2741174272
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.419996790242337e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.37254830714659,
      "tokens_seen": 2741305344
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.419194350826513e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.372535787447731,
      "tokens_seen": 2741436416
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4183919114106887e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.372523268515037,
      "tokens_seen": 2741567488
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4175894719948646e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.3725107503484253,
      "tokens_seen": 2741698560
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.41678703257904e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.372498232947811,
      "tokens_seen": 2741829632
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.415984593163216e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.3724857163131112,
      "tokens_seen": 2741960704
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.415182153747392e-05,
      "loss": 1.2085,
      "theoretical_loss": 3.3724732004442424,
      "tokens_seen": 2742091776
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.414379714331568e-05,
      "loss": 1.2556,
      "theoretical_loss": 3.3724606853411214,
      "tokens_seen": 2742222848
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.413577274915744e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.3724481710036645,
      "tokens_seen": 2742353920
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.41277483549992e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.372435657431788,
      "tokens_seen": 2742484992
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.411972396084096e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.372423144625409,
      "tokens_seen": 2742616064
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.411169956668271e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.372410632584444,
      "tokens_seen": 2742747136
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.410367517252447e-05,
      "loss": 1.2917,
      "theoretical_loss": 3.3723981213088092,
      "tokens_seen": 2742878208
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.409565077836623e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.3723856107984216,
      "tokens_seen": 2743009280
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.408762638420799e-05,
      "loss": 1.1831,
      "theoretical_loss": 3.3723731010531974,
      "tokens_seen": 2743140352
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.407960199004975e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.372360592073054,
      "tokens_seen": 2743271424
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.407157759589151e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3723480838579074,
      "tokens_seen": 2743402496
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.406355320173327e-05,
      "loss": 1.3112,
      "theoretical_loss": 3.3723355764076746,
      "tokens_seen": 2743533568
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.405552880757503e-05,
      "loss": 1.1544,
      "theoretical_loss": 3.372323069722272,
      "tokens_seen": 2743664640
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4047504413416785e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.3723105638016166,
      "tokens_seen": 2743795712
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4039480019258545e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.372298058645625,
      "tokens_seen": 2743926784
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4031455625100304e-05,
      "loss": 1.1939,
      "theoretical_loss": 3.372285554254214,
      "tokens_seen": 2744057856
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4023431230942064e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.3722730506273,
      "tokens_seen": 2744188928
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4015406836783824e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.3722605477648,
      "tokens_seen": 2744320000
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.9928019046783447,
      "objective/train/avg_token_score": 0.007198074366897345,
      "objective/train/docs_used": 997363,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3310866355895996,
      "objective/train/original_loss": 1.3418378829956055,
      "objective/train/theoretical_loss": 3.372248045666631,
      "objective/train/tokens_used": 1114910176,
      "objective/train/unlikelihood_loss": 0.018284082412719727,
      "theoretical_loss": 3.372248045666631,
      "tokens_seen": 2744451072
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.4007382442625584e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.372248045666631,
      "tokens_seen": 2744451072
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.3999358048467344e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.372235544332709,
      "tokens_seen": 2744582144
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.39913336543091e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.3722230437629515,
      "tokens_seen": 2744713216
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.398330926015086e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.3722105439572756,
      "tokens_seen": 2744844288
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.397528486599262e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.3721980449155966,
      "tokens_seen": 2744975360
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.396726047183438e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.372185546637833,
      "tokens_seen": 2745106432
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.395923607767614e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.372173049123901,
      "tokens_seen": 2745237504
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.39512116835179e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.3721605523737175,
      "tokens_seen": 2745368576
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.394318728935966e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.372148056387199,
      "tokens_seen": 2745499648
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.3935162895201417e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.3721355611642627,
      "tokens_seen": 2745630720
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.392713850104317e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.372123066704826,
      "tokens_seen": 2745761792
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.391911410688493e-05,
      "loss": 1.2224,
      "theoretical_loss": 3.372110573008805,
      "tokens_seen": 2745892864
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.391108971272669e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3720980800761167,
      "tokens_seen": 2746023936
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.390306531856845e-05,
      "loss": 1.2123,
      "theoretical_loss": 3.3720855879066782,
      "tokens_seen": 2746155008
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.389504092441021e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.372073096500406,
      "tokens_seen": 2746286080
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.388701653025197e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.3720606058572185,
      "tokens_seen": 2746417152
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.387899213609373e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.372048115977031,
      "tokens_seen": 2746548224
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.387096774193548e-05,
      "loss": 1.3,
      "theoretical_loss": 3.3720356268597618,
      "tokens_seen": 2746679296
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.386294334777724e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.3720231385053268,
      "tokens_seen": 2746810368
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.3854918953619e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3720106509136434,
      "tokens_seen": 2746941440
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.384689455946076e-05,
      "loss": 1.2327,
      "theoretical_loss": 3.371998164084629,
      "tokens_seen": 2747072512
    },
    {
      "epoch": 0.66,
      "learning_rate": 3.383887016530252e-05,
      "loss": 1.307,
      "theoretical_loss": 3.3719856780182003,
      "tokens_seen": 2747203584
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.383084577114428e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.3719731927142744,
      "tokens_seen": 2747334656
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.382282137698604e-05,
      "loss": 1.1849,
      "theoretical_loss": 3.371960708172768,
      "tokens_seen": 2747465728
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3814796982827795e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3719482243935985,
      "tokens_seen": 2747596800
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3806772588669555e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.3719357413766833,
      "tokens_seen": 2747727872
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3798748194511315e-05,
      "loss": 1.2002,
      "theoretical_loss": 3.3719232591219392,
      "tokens_seen": 2747858944
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3790723800353074e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.3719107776292834,
      "tokens_seen": 2747990016
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3782699406194834e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.3718982968986326,
      "tokens_seen": 2748121088
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3774675012036594e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.371885816929905,
      "tokens_seen": 2748252160
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3766650617878354e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.371873337723016,
      "tokens_seen": 2748383232
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3758626223720114e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.3718608592778847,
      "tokens_seen": 2748514304
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.375060182956187e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.3718483815944267,
      "tokens_seen": 2748645376
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.374257743540363e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.3718359046725603,
      "tokens_seen": 2748776448
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.373455304124539e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.3718234285122017,
      "tokens_seen": 2748907520
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.372652864708715e-05,
      "loss": 1.2547,
      "theoretical_loss": 3.371810953113269,
      "tokens_seen": 2749038592
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.371850425292891e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.3717984784756796,
      "tokens_seen": 2749169664
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.371047985877067e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.3717860045993495,
      "tokens_seen": 2749300736
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.370245546461243e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.3717735314841972,
      "tokens_seen": 2749431808
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.369443107045418e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.3717610591301392,
      "tokens_seen": 2749562880
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.368640667629594e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.3717485875370934,
      "tokens_seen": 2749693952
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.36783822821377e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.371736116704976,
      "tokens_seen": 2749825024
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.367035788797946e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.371723646633706,
      "tokens_seen": 2749956096
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.366233349382122e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.371711177323199,
      "tokens_seen": 2750087168
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.365430909966298e-05,
      "loss": 1.3382,
      "theoretical_loss": 3.3716987087733736,
      "tokens_seen": 2750218240
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.364628470550474e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.371686240984147,
      "tokens_seen": 2750349312
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.363826031134649e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.371673773955435,
      "tokens_seen": 2750480384
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.363023591718825e-05,
      "loss": 1.3617,
      "theoretical_loss": 3.371661307687157,
      "tokens_seen": 2750611456
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.362221152303001e-05,
      "loss": 1.4057,
      "theoretical_loss": 3.3716488421792294,
      "tokens_seen": 2750742528
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.361418712887177e-05,
      "loss": 1.1903,
      "theoretical_loss": 3.37163637743157,
      "tokens_seen": 2750873600
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.9917548298835754,
      "objective/train/avg_token_score": 0.008245176635682583,
      "objective/train/docs_used": 999260,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2184526920318604,
      "objective/train/original_loss": 1.2290620803833008,
      "objective/train/theoretical_loss": 3.3716239134440955,
      "objective/train/tokens_used": 1121463776,
      "objective/train/unlikelihood_loss": 0.02336023934185505,
      "theoretical_loss": 3.3716239134440955,
      "tokens_seen": 2751004672
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.360616273471353e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.3716239134440955,
      "tokens_seen": 2751004672
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.359813834055529e-05,
      "loss": 1.2124,
      "theoretical_loss": 3.3716114502167245,
      "tokens_seen": 2751135744
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.359011394639705e-05,
      "loss": 1.2147,
      "theoretical_loss": 3.371598987749373,
      "tokens_seen": 2751266816
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.358208955223881e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.3715865260419595,
      "tokens_seen": 2751397888
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3574065158080565e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.3715740650944013,
      "tokens_seen": 2751528960
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3566040763922325e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.3715616049066153,
      "tokens_seen": 2751660032
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3558016369764085e-05,
      "loss": 1.1907,
      "theoretical_loss": 3.37154914547852,
      "tokens_seen": 2751791104
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3549991975605845e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.371536686810032,
      "tokens_seen": 2751922176
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3541967581447604e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.3715242289010687,
      "tokens_seen": 2752053248
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3533943187289364e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.3715117717515484,
      "tokens_seen": 2752184320
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3525918793131124e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.371499315361388,
      "tokens_seen": 2752315392
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.351789439897288e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.371486859730506,
      "tokens_seen": 2752446464
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.350987000481464e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.3714744048588186,
      "tokens_seen": 2752577536
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.35018456106564e-05,
      "loss": 1.3906,
      "theoretical_loss": 3.3714619507462444,
      "tokens_seen": 2752708608
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.349382121649816e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.371449497392701,
      "tokens_seen": 2752839680
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.348579682233992e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.371437044798105,
      "tokens_seen": 2752970752
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.347777242818168e-05,
      "loss": 1.2908,
      "theoretical_loss": 3.371424592962375,
      "tokens_seen": 2753101824
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.346974803402344e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.371412141885428,
      "tokens_seen": 2753232896
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.34617236398652e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.3713996915671824,
      "tokens_seen": 2753363968
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.345369924570695e-05,
      "loss": 1.2077,
      "theoretical_loss": 3.3713872420075557,
      "tokens_seen": 2753495040
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.344567485154871e-05,
      "loss": 1.369,
      "theoretical_loss": 3.371374793206465,
      "tokens_seen": 2753626112
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.343765045739047e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.3713623451638277,
      "tokens_seen": 2753757184
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.342962606323223e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.3713498978795626,
      "tokens_seen": 2753888256
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.342160166907399e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.371337451353587,
      "tokens_seen": 2754019328
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.341357727491575e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.3713250055858177,
      "tokens_seen": 2754150400
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.340555288075751e-05,
      "loss": 1.3163,
      "theoretical_loss": 3.371312560576174,
      "tokens_seen": 2754281472
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.339752848659926e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3713001163245724,
      "tokens_seen": 2754412544
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.338950409244102e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.3712876728309316,
      "tokens_seen": 2754543616
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.338147969828278e-05,
      "loss": 1.1965,
      "theoretical_loss": 3.3712752300951685,
      "tokens_seen": 2754674688
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.337345530412454e-05,
      "loss": 1.304,
      "theoretical_loss": 3.371262788117201,
      "tokens_seen": 2754805760
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.33654309099663e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.3712503468969475,
      "tokens_seen": 2754936832
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.335740651580806e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.3712379064343256,
      "tokens_seen": 2755067904
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.334938212164982e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.371225466729253,
      "tokens_seen": 2755198976
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3341357727491575e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.371213027781647,
      "tokens_seen": 2755330048
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 1.351,
      "theoretical_loss": 3.3712005895914263,
      "tokens_seen": 2755461120
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3325308939175095e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.371188152158509,
      "tokens_seen": 2755592192
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3317284545016855e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.3711757154828117,
      "tokens_seen": 2755723264
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3309260150858615e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.371163279564253,
      "tokens_seen": 2755854336
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3301235756700375e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.371150844402751,
      "tokens_seen": 2755985408
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3293211362542134e-05,
      "loss": 1.2285,
      "theoretical_loss": 3.3711384099982236,
      "tokens_seen": 2756116480
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3285186968383894e-05,
      "loss": 1.3202,
      "theoretical_loss": 3.371125976350588,
      "tokens_seen": 2756247552
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.327716257422565e-05,
      "loss": 1.2353,
      "theoretical_loss": 3.3711135434597628,
      "tokens_seen": 2756378624
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.326913818006741e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.3711011113256664,
      "tokens_seen": 2756509696
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.326111378590917e-05,
      "loss": 1.288,
      "theoretical_loss": 3.3710886799482154,
      "tokens_seen": 2756640768
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.325308939175093e-05,
      "loss": 1.295,
      "theoretical_loss": 3.371076249327329,
      "tokens_seen": 2756771840
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.324506499759269e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.3710638194629245,
      "tokens_seen": 2756902912
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.323704060343445e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.3710513903549204,
      "tokens_seen": 2757033984
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.322901620927621e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.3710389620032344,
      "tokens_seen": 2757165056
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.322099181511796e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.3710265344077843,
      "tokens_seen": 2757296128
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.321296742095972e-05,
      "loss": 1.3537,
      "theoretical_loss": 3.371014107568489,
      "tokens_seen": 2757427200
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.9923155307769775,
      "objective/train/avg_token_score": 0.007684460841119289,
      "objective/train/docs_used": 1001599,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4274193048477173,
      "objective/train/original_loss": 1.4408724308013916,
      "objective/train/theoretical_loss": 3.3710016814852652,
      "objective/train/tokens_used": 1128017376,
      "objective/train/unlikelihood_loss": 0.019767532125115395,
      "theoretical_loss": 3.3710016814852652,
      "tokens_seen": 2757558272
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.320494302680148e-05,
      "loss": 1.3502,
      "theoretical_loss": 3.3710016814852652,
      "tokens_seen": 2757558272
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.319691863264324e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.3709892561580324,
      "tokens_seen": 2757689344
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3188894238485e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.3709768315867077,
      "tokens_seen": 2757820416
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.318086984432676e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.3709644077712095,
      "tokens_seen": 2757951488
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.317284545016851e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.370951984711456,
      "tokens_seen": 2758082560
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.316482105601027e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.370939562407365,
      "tokens_seen": 2758213632
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.315679666185203e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.3709271408588553,
      "tokens_seen": 2758344704
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.314877226769379e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.3709147200658443,
      "tokens_seen": 2758475776
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.314074787353555e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.3709023000282503,
      "tokens_seen": 2758606848
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.313272347937731e-05,
      "loss": 1.32,
      "theoretical_loss": 3.370889880745992,
      "tokens_seen": 2758737920
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3124699085219065e-05,
      "loss": 1.3194,
      "theoretical_loss": 3.3708774622189868,
      "tokens_seen": 2758868992
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3116674691060825e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.3708650444471537,
      "tokens_seen": 2759000064
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3108650296902585e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.37085262743041,
      "tokens_seen": 2759131136
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3100625902744345e-05,
      "loss": 1.2248,
      "theoretical_loss": 3.370840211168675,
      "tokens_seen": 2759262208
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3092601508586105e-05,
      "loss": 1.2441,
      "theoretical_loss": 3.370827795661866,
      "tokens_seen": 2759393280
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.3084577114427865e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.3708153809099013,
      "tokens_seen": 2759524352
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.307655272026962e-05,
      "loss": 1.188,
      "theoretical_loss": 3.3708029669127,
      "tokens_seen": 2759655424
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.306852832611138e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.3707905536701794,
      "tokens_seen": 2759786496
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.306050393195314e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3707781411822584,
      "tokens_seen": 2759917568
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.30524795377949e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.370765729448855,
      "tokens_seen": 2760048640
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.304445514363666e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.370753318469888,
      "tokens_seen": 2760179712
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.303643074947842e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.3707409082452746,
      "tokens_seen": 2760310784
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.302840635532017e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3707284987749344,
      "tokens_seen": 2760441856
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.302038196116193e-05,
      "loss": 1.2535,
      "theoretical_loss": 3.370716090058785,
      "tokens_seen": 2760572928
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.301235756700369e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.370703682096745,
      "tokens_seen": 2760704000
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.300433317284545e-05,
      "loss": 1.3509,
      "theoretical_loss": 3.3706912748887325,
      "tokens_seen": 2760835072
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.299630877868721e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.3706788684346662,
      "tokens_seen": 2760966144
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.298828438452897e-05,
      "loss": 1.2221,
      "theoretical_loss": 3.3706664627344645,
      "tokens_seen": 2761097216
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.298025999037072e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.3706540577880455,
      "tokens_seen": 2761228288
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.297223559621248e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.370641653595328,
      "tokens_seen": 2761359360
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.296421120205424e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.3706292501562305,
      "tokens_seen": 2761490432
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2956186807896e-05,
      "loss": 1.332,
      "theoretical_loss": 3.3706168474706706,
      "tokens_seen": 2761621504
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.294816241373776e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.370604445538568,
      "tokens_seen": 2761752576
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.294013801957952e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.3705920443598405,
      "tokens_seen": 2761883648
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.293211362542128e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.370579643934406,
      "tokens_seen": 2762014720
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2924089231263036e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.370567244262184,
      "tokens_seen": 2762145792
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2916064837104796e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.3705548453430927,
      "tokens_seen": 2762276864
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2908040442946556e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3705424471770504,
      "tokens_seen": 2762407936
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2900016048788316e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.370530049763976,
      "tokens_seen": 2762539008
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2891991654630076e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.370517653103788,
      "tokens_seen": 2762670080
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2883967260471835e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.3705052571964043,
      "tokens_seen": 2762801152
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.2875942866313595e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.3704928620417443,
      "tokens_seen": 2762932224
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.286791847215535e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.370480467639726,
      "tokens_seen": 2763063296
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.285989407799711e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.370468073990269,
      "tokens_seen": 2763194368
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.285186968383887e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.3704556810932904,
      "tokens_seen": 2763325440
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.284384528968063e-05,
      "loss": 1.2276,
      "theoretical_loss": 3.3704432889487097,
      "tokens_seen": 2763456512
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.283582089552239e-05,
      "loss": 1.216,
      "theoretical_loss": 3.370430897556446,
      "tokens_seen": 2763587584
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.282779650136415e-05,
      "loss": 1.2272,
      "theoretical_loss": 3.3704185069164168,
      "tokens_seen": 2763718656
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.281977210720591e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.3704061170285415,
      "tokens_seen": 2763849728
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.281174771304767e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.3703937278927385,
      "tokens_seen": 2763980800
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.9913116693496704,
      "objective/train/avg_token_score": 0.008688334375619888,
      "objective/train/docs_used": 1003856,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2408310174942017,
      "objective/train/original_loss": 1.2516766786575317,
      "objective/train/theoretical_loss": 3.370381339508927,
      "objective/train/tokens_used": 1134570976,
      "objective/train/unlikelihood_loss": 0.028941579163074493,
      "theoretical_loss": 3.370381339508927,
      "tokens_seen": 2764111872
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.280372331888942e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.370381339508927,
      "tokens_seen": 2764111872
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.279569892473118e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.370368951877025,
      "tokens_seen": 2764242944
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.278767453057294e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.3703565649969516,
      "tokens_seen": 2764374016
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.27796501364147e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.370344178868625,
      "tokens_seen": 2764505088
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.277162574225646e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.3703317934919648,
      "tokens_seen": 2764636160
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.276360134809822e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.3703194088668895,
      "tokens_seen": 2764767232
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.275557695393998e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.3703070249933176,
      "tokens_seen": 2764898304
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2747552559781734e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.370294641871168,
      "tokens_seen": 2765029376
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2739528165623493e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3702822595003594,
      "tokens_seen": 2765160448
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.273150377146525e-05,
      "loss": 1.341,
      "theoretical_loss": 3.3702698778808107,
      "tokens_seen": 2765291520
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.272347937730701e-05,
      "loss": 1.3432,
      "theoretical_loss": 3.3702574970124406,
      "tokens_seen": 2765422592
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.271545498314877e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.370245116895168,
      "tokens_seen": 2765553664
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.270743058899053e-05,
      "loss": 1.3106,
      "theoretical_loss": 3.370232737528912,
      "tokens_seen": 2765684736
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.269940619483229e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.370220358913591,
      "tokens_seen": 2765815808
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2691381800674046e-05,
      "loss": 1.268,
      "theoretical_loss": 3.3702079810491243,
      "tokens_seen": 2765946880
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2683357406515806e-05,
      "loss": 1.3371,
      "theoretical_loss": 3.3701956039354304,
      "tokens_seen": 2766077952
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2675333012357566e-05,
      "loss": 1.3714,
      "theoretical_loss": 3.370183227572428,
      "tokens_seen": 2766209024
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2667308618199326e-05,
      "loss": 1.2115,
      "theoretical_loss": 3.3701708519600366,
      "tokens_seen": 2766340096
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2659284224041086e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.3701584770981747,
      "tokens_seen": 2766471168
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2651259829882846e-05,
      "loss": 1.2327,
      "theoretical_loss": 3.3701461029867614,
      "tokens_seen": 2766602240
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2643235435724606e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.370133729625716,
      "tokens_seen": 2766733312
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2635211041566365e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.370121357014957,
      "tokens_seen": 2766864384
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.262718664740812e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.370108985154403,
      "tokens_seen": 2766995456
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.261916225324988e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3700966140439736,
      "tokens_seen": 2767126528
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.261113785909164e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.370084243683588,
      "tokens_seen": 2767257600
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.26031134649334e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.3700718740731643,
      "tokens_seen": 2767388672
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.259508907077516e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.3700595052126223,
      "tokens_seen": 2767519744
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.258706467661692e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.3700471371018805,
      "tokens_seen": 2767650816
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.257904028245868e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.370034769740858,
      "tokens_seen": 2767781888
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.257101588830043e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.3700224031294748,
      "tokens_seen": 2767912960
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.256299149414219e-05,
      "loss": 1.302,
      "theoretical_loss": 3.3700100372676487,
      "tokens_seen": 2768044032
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.255496709998395e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.369997672155299,
      "tokens_seen": 2768175104
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.254694270582571e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.369985307792346,
      "tokens_seen": 2768306176
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.253891831166747e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.3699729441787074,
      "tokens_seen": 2768437248
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.253089391750923e-05,
      "loss": 1.307,
      "theoretical_loss": 3.3699605813143023,
      "tokens_seen": 2768568320
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.252286952335099e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.369948219199051,
      "tokens_seen": 2768699392
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.251484512919275e-05,
      "loss": 1.159,
      "theoretical_loss": 3.3699358578328718,
      "tokens_seen": 2768830464
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2506820735034504e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.369923497215684,
      "tokens_seen": 2768961536
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2498796340876264e-05,
      "loss": 1.3164,
      "theoretical_loss": 3.369911137347407,
      "tokens_seen": 2769092608
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2490771946718023e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.3698987782279595,
      "tokens_seen": 2769223680
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.248274755255978e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.3698864198572607,
      "tokens_seen": 2769354752
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.247472315840154e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3698740622352306,
      "tokens_seen": 2769485824
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.24666987642433e-05,
      "loss": 1.3453,
      "theoretical_loss": 3.3698617053617874,
      "tokens_seen": 2769616896
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.245867437008506e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.3698493492368513,
      "tokens_seen": 2769747968
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2450649975926816e-05,
      "loss": 1.28,
      "theoretical_loss": 3.369836993860341,
      "tokens_seen": 2769879040
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2442625581768576e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.369824639232175,
      "tokens_seen": 2770010112
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2434601187610336e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.3698122853522743,
      "tokens_seen": 2770141184
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2426576793452096e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.369799932220557,
      "tokens_seen": 2770272256
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2418552399293856e-05,
      "loss": 1.2499,
      "theoretical_loss": 3.3697875798369425,
      "tokens_seen": 2770403328
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2410528005135616e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.3697752282013504,
      "tokens_seen": 2770534400
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.9930559396743774,
      "objective/train/avg_token_score": 0.006943980697542429,
      "objective/train/docs_used": 1006223,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1844812631607056,
      "objective/train/original_loss": 1.1936628818511963,
      "objective/train/theoretical_loss": 3.3697628773136996,
      "objective/train/tokens_used": 1141124576,
      "objective/train/unlikelihood_loss": 0.02009393274784088,
      "theoretical_loss": 3.3697628773136996,
      "tokens_seen": 2770665472
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2402503610977376e-05,
      "loss": 1.246,
      "theoretical_loss": 3.3697628773136996,
      "tokens_seen": 2770665472
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.239447921681913e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.36975052717391,
      "tokens_seen": 2770796544
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.238645482266089e-05,
      "loss": 1.3649,
      "theoretical_loss": 3.3697381777819007,
      "tokens_seen": 2770927616
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.237843042850265e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.369725829137591,
      "tokens_seen": 2771058688
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.237040603434441e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3697134812409,
      "tokens_seen": 2771189760
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.236238164018617e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.3697011340917475,
      "tokens_seen": 2771320832
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.235435724602793e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.3696887876900528,
      "tokens_seen": 2771451904
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.234633285186969e-05,
      "loss": 1.3127,
      "theoretical_loss": 3.369676442035735,
      "tokens_seen": 2771582976
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.233830845771145e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3696640971287137,
      "tokens_seen": 2771714048
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.23302840635532e-05,
      "loss": 1.2446,
      "theoretical_loss": 3.3696517529689087,
      "tokens_seen": 2771845120
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.232225966939496e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.369639409556239,
      "tokens_seen": 2771976192
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.231423527523672e-05,
      "loss": 1.284,
      "theoretical_loss": 3.369627066890624,
      "tokens_seen": 2772107264
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.230621088107848e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.3696147249719837,
      "tokens_seen": 2772238336
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.229818648692024e-05,
      "loss": 1.2178,
      "theoretical_loss": 3.369602383800237,
      "tokens_seen": 2772369408
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2290162092762e-05,
      "loss": 1.2173,
      "theoretical_loss": 3.3695900433753034,
      "tokens_seen": 2772500480
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.228213769860376e-05,
      "loss": 1.2433,
      "theoretical_loss": 3.369577703697103,
      "tokens_seen": 2772631552
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2274113304445514e-05,
      "loss": 1.2124,
      "theoretical_loss": 3.369565364765555,
      "tokens_seen": 2772762624
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2266088910287274e-05,
      "loss": 1.2148,
      "theoretical_loss": 3.3695530265805784,
      "tokens_seen": 2772893696
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2258064516129034e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.3695406891420934,
      "tokens_seen": 2773024768
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2250040121970793e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.3695283524500197,
      "tokens_seen": 2773155840
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2242015727812553e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.369516016504276,
      "tokens_seen": 2773286912
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.223399133365431e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.3695036813047827,
      "tokens_seen": 2773417984
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.222596693949607e-05,
      "loss": 1.2673,
      "theoretical_loss": 3.3694913468514596,
      "tokens_seen": 2773549056
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2217942545337826e-05,
      "loss": 1.256,
      "theoretical_loss": 3.369479013144225,
      "tokens_seen": 2773680128
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2209918151179586e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3694666801829998,
      "tokens_seen": 2773811200
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2201893757021346e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.369454347967703,
      "tokens_seen": 2773942272
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2193869362863106e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.369442016498254,
      "tokens_seen": 2774073344
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2185844968704866e-05,
      "loss": 1.302,
      "theoretical_loss": 3.3694296857745734,
      "tokens_seen": 2774204416
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2177820574546626e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.36941735579658,
      "tokens_seen": 2774335488
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2169796180388386e-05,
      "loss": 1.289,
      "theoretical_loss": 3.369405026564194,
      "tokens_seen": 2774466560
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2161771786230146e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.369392698077335,
      "tokens_seen": 2774597632
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.21537473920719e-05,
      "loss": 1.2079,
      "theoretical_loss": 3.3693803703359224,
      "tokens_seen": 2774728704
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.214572299791366e-05,
      "loss": 1.275,
      "theoretical_loss": 3.369368043339876,
      "tokens_seen": 2774859776
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.213769860375542e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.3693557170891157,
      "tokens_seen": 2774990848
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.212967420959718e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.369343391583561,
      "tokens_seen": 2775121920
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.212164981543894e-05,
      "loss": 1.1995,
      "theoretical_loss": 3.369331066823132,
      "tokens_seen": 2775252992
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.21136254212807e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.3693187428077485,
      "tokens_seen": 2775384064
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.210560102712246e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.3693064195373297,
      "tokens_seen": 2775515136
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.209757663296421e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.369294097011796,
      "tokens_seen": 2775646208
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.208955223880597e-05,
      "loss": 1.1947,
      "theoretical_loss": 3.369281775231067,
      "tokens_seen": 2775777280
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.208152784464773e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.3692694541950625,
      "tokens_seen": 2775908352
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.207350345048949e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.369257133903702,
      "tokens_seen": 2776039424
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.206547905633125e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.369244814356906,
      "tokens_seen": 2776170496
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.205745466217301e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.369232495554594,
      "tokens_seen": 2776301568
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.204943026801477e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.3692201774966857,
      "tokens_seen": 2776432640
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2041405873856524e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.369207860183101,
      "tokens_seen": 2776563712
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2033381479698284e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.36919554361376,
      "tokens_seen": 2776694784
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2025357085540044e-05,
      "loss": 1.2287,
      "theoretical_loss": 3.369183227788583,
      "tokens_seen": 2776825856
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2017332691381804e-05,
      "loss": 1.2527,
      "theoretical_loss": 3.369170912707489,
      "tokens_seen": 2776956928
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2009308297223564e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.369158598370398,
      "tokens_seen": 2777088000
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.9946650862693787,
      "objective/train/avg_token_score": 0.005334955640137196,
      "objective/train/docs_used": 1008708,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2242995500564575,
      "objective/train/original_loss": 1.2315924167633057,
      "objective/train/theoretical_loss": 3.3691462847772313,
      "objective/train/tokens_used": 1147678176,
      "objective/train/unlikelihood_loss": 0.014776166528463364,
      "theoretical_loss": 3.3691462847772313,
      "tokens_seen": 2777219072
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.2001283903065323e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.3691462847772313,
      "tokens_seen": 2777219072
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.199325950890708e-05,
      "loss": 1.236,
      "theoretical_loss": 3.3691339719279076,
      "tokens_seen": 2777350144
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.198523511474884e-05,
      "loss": 1.2495,
      "theoretical_loss": 3.369121659822347,
      "tokens_seen": 2777481216
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.1977210720590596e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.3691093484604693,
      "tokens_seen": 2777612288
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.1969186326432356e-05,
      "loss": 1.3519,
      "theoretical_loss": 3.369097037842195,
      "tokens_seen": 2777743360
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.1961161932274116e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.369084727967444,
      "tokens_seen": 2777874432
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.1953137538115876e-05,
      "loss": 1.2131,
      "theoretical_loss": 3.369072418836136,
      "tokens_seen": 2778005504
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.1945113143957636e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.3690601104481916,
      "tokens_seen": 2778136576
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.1937088749799396e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.3690478028035304,
      "tokens_seen": 2778267648
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.1929064355641156e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.369035495902072,
      "tokens_seen": 2778398720
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.192103996148291e-05,
      "loss": 1.2134,
      "theoretical_loss": 3.3690231897437375,
      "tokens_seen": 2778529792
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.191301556732467e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3690108843284468,
      "tokens_seen": 2778660864
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.190499117316643e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.368998579656119,
      "tokens_seen": 2778791936
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.189696677900819e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.3689862757266753,
      "tokens_seen": 2778923008
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.188894238484995e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.368973972540035,
      "tokens_seen": 2779054080
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.188091799069171e-05,
      "loss": 1.2263,
      "theoretical_loss": 3.368961670096119,
      "tokens_seen": 2779185152
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.187289359653347e-05,
      "loss": 1.2803,
      "theoretical_loss": 3.368949368394847,
      "tokens_seen": 2779316224
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.186486920237522e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.368937067436139,
      "tokens_seen": 2779447296
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.185684480821698e-05,
      "loss": 1.3064,
      "theoretical_loss": 3.3689247672199154,
      "tokens_seen": 2779578368
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.184882041405874e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3689124677460964,
      "tokens_seen": 2779709440
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.18407960199005e-05,
      "loss": 1.233,
      "theoretical_loss": 3.368900169014602,
      "tokens_seen": 2779840512
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.183277162574226e-05,
      "loss": 1.23,
      "theoretical_loss": 3.3688878710253523,
      "tokens_seen": 2779971584
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.182474723158402e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.3688755737782685,
      "tokens_seen": 2780102656
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1816722837425774e-05,
      "loss": 1.1864,
      "theoretical_loss": 3.3688632772732694,
      "tokens_seen": 2780233728
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1808698443267534e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.368850981510276,
      "tokens_seen": 2780364800
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1800674049109294e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3688386864892084,
      "tokens_seen": 2780495872
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1792649654951054e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.368826392209987,
      "tokens_seen": 2780626944
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1784625260792814e-05,
      "loss": 1.226,
      "theoretical_loss": 3.368814098672532,
      "tokens_seen": 2780758016
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1776600866634574e-05,
      "loss": 1.25,
      "theoretical_loss": 3.3688018058767635,
      "tokens_seen": 2780889088
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.176857647247633e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.368789513822602,
      "tokens_seen": 2781020160
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.176055207831809e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.368777222509968,
      "tokens_seen": 2781151232
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.175252768415985e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3687649319387813,
      "tokens_seen": 2781282304
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1744503290001607e-05,
      "loss": 1.2548,
      "theoretical_loss": 3.3687526421089626,
      "tokens_seen": 2781413376
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1736478895843366e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.3687403530204323,
      "tokens_seen": 2781544448
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1728454501685126e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3687280646731104,
      "tokens_seen": 2781675520
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.172043010752688e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.368715777066918,
      "tokens_seen": 2781806592
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.171240571336864e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.3687034902017743,
      "tokens_seen": 2781937664
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.17043813192104e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3686912040776007,
      "tokens_seen": 2782068736
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.169635692505216e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.368678918694317,
      "tokens_seen": 2782199808
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.168833253089392e-05,
      "loss": 1.272,
      "theoretical_loss": 3.3686666340518445,
      "tokens_seen": 2782330880
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.168030813673568e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.368654350150103,
      "tokens_seen": 2782461952
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.167228374257743e-05,
      "loss": 1.295,
      "theoretical_loss": 3.3686420669890125,
      "tokens_seen": 2782593024
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.166425934841919e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.368629784568494,
      "tokens_seen": 2782724096
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.165623495426095e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.368617502888468,
      "tokens_seen": 2782855168
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.164821056010271e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.3686052219488554,
      "tokens_seen": 2782986240
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.164018616594447e-05,
      "loss": 1.289,
      "theoretical_loss": 3.3685929417495757,
      "tokens_seen": 2783117312
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.163216177178623e-05,
      "loss": 1.2329,
      "theoretical_loss": 3.36858066229055,
      "tokens_seen": 2783248384
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1624137377627985e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3685683835716986,
      "tokens_seen": 2783379456
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1616112983469745e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.3685561055929427,
      "tokens_seen": 2783510528
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1608088589311505e-05,
      "loss": 1.4112,
      "theoretical_loss": 3.368543828354202,
      "tokens_seen": 2783641600
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.9910603165626526,
      "objective/train/avg_token_score": 0.008939657360315323,
      "objective/train/docs_used": 1011173,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2079942226409912,
      "objective/train/original_loss": 1.2205007076263428,
      "objective/train/theoretical_loss": 3.3685315518553973,
      "objective/train/tokens_used": 1154231776,
      "objective/train/unlikelihood_loss": 0.021104658022522926,
      "theoretical_loss": 3.3685315518553973,
      "tokens_seen": 2783772672
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1600064195153265e-05,
      "loss": 1.2337,
      "theoretical_loss": 3.3685315518553973,
      "tokens_seen": 2783772672
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1592039800995024e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.368519276096449,
      "tokens_seen": 2783903744
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1584015406836784e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.368507001077278,
      "tokens_seen": 2784034816
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1575991012678544e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.368494726797805,
      "tokens_seen": 2784165888
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.15679666185203e-05,
      "loss": 1.2178,
      "theoretical_loss": 3.3684824532579505,
      "tokens_seen": 2784296960
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.155994222436206e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.368470180457635,
      "tokens_seen": 2784428032
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.155191783020382e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.3684579083967785,
      "tokens_seen": 2784559104
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.154389343604558e-05,
      "loss": 1.257,
      "theoretical_loss": 3.368445637075303,
      "tokens_seen": 2784690176
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.153586904188734e-05,
      "loss": 1.2593,
      "theoretical_loss": 3.3684333664931283,
      "tokens_seen": 2784821248
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.15278446477291e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.368421096650175,
      "tokens_seen": 2784952320
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.151982025357086e-05,
      "loss": 1.3323,
      "theoretical_loss": 3.3684088275463644,
      "tokens_seen": 2785083392
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.151179585941262e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.3683965591816163,
      "tokens_seen": 2785214464
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.150377146525437e-05,
      "loss": 1.1966,
      "theoretical_loss": 3.368384291555852,
      "tokens_seen": 2785345536
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.149574707109613e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.3683720246689925,
      "tokens_seen": 2785476608
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.148772267693789e-05,
      "loss": 1.3064,
      "theoretical_loss": 3.368359758520958,
      "tokens_seen": 2785607680
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.147969828277965e-05,
      "loss": 1.3211,
      "theoretical_loss": 3.368347493111669,
      "tokens_seen": 2785738752
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.147167388862141e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.368335228441047,
      "tokens_seen": 2785869824
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.146364949446317e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.3683229645090123,
      "tokens_seen": 2786000896
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.145562510030493e-05,
      "loss": 1.292,
      "theoretical_loss": 3.368310701315486,
      "tokens_seen": 2786131968
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.144760070614668e-05,
      "loss": 1.3452,
      "theoretical_loss": 3.3682984388603883,
      "tokens_seen": 2786263040
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.143957631198844e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.36828617714364,
      "tokens_seen": 2786394112
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.14315519178302e-05,
      "loss": 1.2723,
      "theoretical_loss": 3.3682739161651627,
      "tokens_seen": 2786525184
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.142352752367196e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.368261655924877,
      "tokens_seen": 2786656256
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.141550312951372e-05,
      "loss": 1.265,
      "theoretical_loss": 3.368249396422703,
      "tokens_seen": 2786787328
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.140747873535548e-05,
      "loss": 1.3702,
      "theoretical_loss": 3.368237137658563,
      "tokens_seen": 2786918400
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.139945434119724e-05,
      "loss": 1.3314,
      "theoretical_loss": 3.3682248796323764,
      "tokens_seen": 2787049472
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1391429947039e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.3682126223440645,
      "tokens_seen": 2787180544
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1383405552880755e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.368200365793548,
      "tokens_seen": 2787311616
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1375381158722515e-05,
      "loss": 1.313,
      "theoretical_loss": 3.368188109980749,
      "tokens_seen": 2787442688
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1367356764564275e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.368175854905587,
      "tokens_seen": 2787573760
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1359332370406035e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.3681636005679834,
      "tokens_seen": 2787704832
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1351307976247795e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.368151346967859,
      "tokens_seen": 2787835904
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1343283582089554e-05,
      "loss": 1.2241,
      "theoretical_loss": 3.3681390941051355,
      "tokens_seen": 2787966976
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1335259187931314e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.3681268419797328,
      "tokens_seen": 2788098048
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.132723479377307e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.368114590591573,
      "tokens_seen": 2788229120
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.131921039961483e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.3681023399405756,
      "tokens_seen": 2788360192
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.131118600545659e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.368090090026663,
      "tokens_seen": 2788491264
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.130316161129835e-05,
      "loss": 1.33,
      "theoretical_loss": 3.368077840849755,
      "tokens_seen": 2788622336
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.129513721714011e-05,
      "loss": 1.271,
      "theoretical_loss": 3.368065592409774,
      "tokens_seen": 2788753408
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.128711282298187e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.36805334470664,
      "tokens_seen": 2788884480
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.127908842882363e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.3680410977402744,
      "tokens_seen": 2789015552
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.127106403466538e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.3680288515105983,
      "tokens_seen": 2789146624
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.126303964050714e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.3680166060175325,
      "tokens_seen": 2789277696
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.12550152463489e-05,
      "loss": 1.277,
      "theoretical_loss": 3.368004361260998,
      "tokens_seen": 2789408768
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.124699085219066e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.3679921172409166,
      "tokens_seen": 2789539840
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.123896645803242e-05,
      "loss": 1.3523,
      "theoretical_loss": 3.367979873957209,
      "tokens_seen": 2789670912
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.123094206387418e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.3679676314097957,
      "tokens_seen": 2789801984
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.122291766971594e-05,
      "loss": 1.2463,
      "theoretical_loss": 3.367955389598599,
      "tokens_seen": 2789933056
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.12148932755577e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.3679431485235396,
      "tokens_seen": 2790064128
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.120686888139945e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.3679309081845377,
      "tokens_seen": 2790195200
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.9916509985923767,
      "objective/train/avg_token_score": 0.00834900327026844,
      "objective/train/docs_used": 1013591,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3987348079681396,
      "objective/train/original_loss": 1.4108185768127441,
      "objective/train/theoretical_loss": 3.367918668581516,
      "objective/train/tokens_used": 1160785376,
      "objective/train/unlikelihood_loss": 0.020987041294574738,
      "theoretical_loss": 3.367918668581516,
      "tokens_seen": 2790326272
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.119884448724121e-05,
      "loss": 1.3577,
      "theoretical_loss": 3.367918668581516,
      "tokens_seen": 2790326272
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.119082009308297e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.3679064297143944,
      "tokens_seen": 2790457344
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.118279569892473e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.367894191583095,
      "tokens_seen": 2790588416
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.117477130476649e-05,
      "loss": 1.2405,
      "theoretical_loss": 3.3678819541875384,
      "tokens_seen": 2790719488
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.116674691060825e-05,
      "loss": 1.3456,
      "theoretical_loss": 3.3678697175276464,
      "tokens_seen": 2790850560
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.115872251645001e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3678574816033393,
      "tokens_seen": 2790981632
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1150698122291765e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3678452464145394,
      "tokens_seen": 2791112704
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1142673728133525e-05,
      "loss": 1.3072,
      "theoretical_loss": 3.3678330119611672,
      "tokens_seen": 2791243776
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1134649333975285e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.3678207782431446,
      "tokens_seen": 2791374848
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1126624939817045e-05,
      "loss": 1.227,
      "theoretical_loss": 3.3678085452603925,
      "tokens_seen": 2791505920
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1118600545658805e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.367796313012832,
      "tokens_seen": 2791636992
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1110576151500565e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3677840815003846,
      "tokens_seen": 2791768064
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.1102551757342325e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.3677718507229715,
      "tokens_seen": 2791899136
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.109452736318408e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.3677596206805145,
      "tokens_seen": 2792030208
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.108650296902584e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.367747391372934,
      "tokens_seen": 2792161280
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.10784785748676e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.3677351628001526,
      "tokens_seen": 2792292352
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.107045418070936e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.3677229349620905,
      "tokens_seen": 2792423424
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.106242978655112e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.3677107078586697,
      "tokens_seen": 2792554496
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.105440539239288e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.3676984814898114,
      "tokens_seen": 2792685568
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.104638099823464e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.3676862558554372,
      "tokens_seen": 2792816640
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.10383566040764e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.367674030955468,
      "tokens_seen": 2792947712
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.103033220991815e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.367661806789826,
      "tokens_seen": 2793078784
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.102230781575991e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.367649583358432,
      "tokens_seen": 2793209856
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.101428342160167e-05,
      "loss": 1.2336,
      "theoretical_loss": 3.367637360661208,
      "tokens_seen": 2793340928
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.100625902744343e-05,
      "loss": 1.266,
      "theoretical_loss": 3.3676251386980747,
      "tokens_seen": 2793472000
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.099823463328519e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.3676129174689535,
      "tokens_seen": 2793603072
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.099021023912695e-05,
      "loss": 1.2313,
      "theoretical_loss": 3.367600696973767,
      "tokens_seen": 2793734144
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.098218584496871e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.3675884772124363,
      "tokens_seen": 2793865216
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.097416145081046e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.367576258184882,
      "tokens_seen": 2793996288
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.096613705665222e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3675640398910267,
      "tokens_seen": 2794127360
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.095811266249398e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.3675518223307916,
      "tokens_seen": 2794258432
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.095008826833574e-05,
      "loss": 1.2061,
      "theoretical_loss": 3.367539605504098,
      "tokens_seen": 2794389504
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.09420638741775e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.3675273894108675,
      "tokens_seen": 2794520576
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.093403948001926e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.3675151740510216,
      "tokens_seen": 2794651648
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.092601508586102e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.3675029594244825,
      "tokens_seen": 2794782720
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.091799069170278e-05,
      "loss": 1.3261,
      "theoretical_loss": 3.367490745531171,
      "tokens_seen": 2794913792
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.0909966297544535e-05,
      "loss": 1.3247,
      "theoretical_loss": 3.367478532371009,
      "tokens_seen": 2795044864
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.0901941903386295e-05,
      "loss": 1.254,
      "theoretical_loss": 3.3674663199439183,
      "tokens_seen": 2795175936
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.0893917509228055e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.3674541082498206,
      "tokens_seen": 2795307008
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.0885893115069815e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.367441897288637,
      "tokens_seen": 2795438080
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.0877868720911575e-05,
      "loss": 1.3575,
      "theoretical_loss": 3.3674296870602896,
      "tokens_seen": 2795569152
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.0869844326753335e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.3674174775647,
      "tokens_seen": 2795700224
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.0861819932595095e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.3674052688017895,
      "tokens_seen": 2795831296
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.085379553843685e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.36739306077148,
      "tokens_seen": 2795962368
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.084577114427861e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.367380853473694,
      "tokens_seen": 2796093440
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.083774675012037e-05,
      "loss": 1.275,
      "theoretical_loss": 3.367368646908352,
      "tokens_seen": 2796224512
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.082972235596213e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.3673564410753767,
      "tokens_seen": 2796355584
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.082169796180389e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.367344235974689,
      "tokens_seen": 2796486656
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.081367356764565e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.3673320316062108,
      "tokens_seen": 2796617728
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.080564917348741e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.3673198279698644,
      "tokens_seen": 2796748800
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.9920945167541504,
      "objective/train/avg_token_score": 0.007905470207333565,
      "objective/train/docs_used": 1015946,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2241928577423096,
      "objective/train/original_loss": 1.2360543012619019,
      "objective/train/theoretical_loss": 3.367307625065571,
      "objective/train/tokens_used": 1167338976,
      "objective/train/unlikelihood_loss": 0.020899029448628426,
      "theoretical_loss": 3.367307625065571,
      "tokens_seen": 2796879872
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.079762477932916e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.367307625065571,
      "tokens_seen": 2796879872
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.078960038517092e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.367295422893253,
      "tokens_seen": 2797010944
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.078157599101268e-05,
      "loss": 1.3766,
      "theoretical_loss": 3.3672832214528317,
      "tokens_seen": 2797142016
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.077355159685444e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.3672710207442287,
      "tokens_seen": 2797273088
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.07655272026962e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.367258820767366,
      "tokens_seen": 2797404160
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.075750280853796e-05,
      "loss": 1.3694,
      "theoretical_loss": 3.367246621522166,
      "tokens_seen": 2797535232
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.074947841437972e-05,
      "loss": 1.3025,
      "theoretical_loss": 3.3672344230085502,
      "tokens_seen": 2797666304
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.074145402022148e-05,
      "loss": 1.2975,
      "theoretical_loss": 3.36722222522644,
      "tokens_seen": 2797797376
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.073342962606323e-05,
      "loss": 1.2398,
      "theoretical_loss": 3.3672100281757578,
      "tokens_seen": 2797928448
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.072540523190499e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.3671978318564255,
      "tokens_seen": 2798059520
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.071738083774675e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.3671856362683648,
      "tokens_seen": 2798190592
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.070935644358851e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.3671734414114978,
      "tokens_seen": 2798321664
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.070133204943027e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.3671612472857455,
      "tokens_seen": 2798452736
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.069330765527203e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.367149053891031,
      "tokens_seen": 2798583808
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.068528326111379e-05,
      "loss": 1.2391,
      "theoretical_loss": 3.367136861227276,
      "tokens_seen": 2798714880
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0677258866955545e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.3671246692944017,
      "tokens_seen": 2798845952
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0669234472797305e-05,
      "loss": 1.2496,
      "theoretical_loss": 3.367112478092331,
      "tokens_seen": 2798977024
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0661210078639065e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3671002876209855,
      "tokens_seen": 2799108096
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0653185684480825e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.3670880978802873,
      "tokens_seen": 2799239168
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0645161290322585e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.367075908870158,
      "tokens_seen": 2799370240
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0637136896164345e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.36706372059052,
      "tokens_seen": 2799501312
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0629112502006105e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.367051533041295,
      "tokens_seen": 2799632384
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.062108810784786e-05,
      "loss": 1.266,
      "theoretical_loss": 3.3670393462224055,
      "tokens_seen": 2799763456
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.061306371368962e-05,
      "loss": 1.258,
      "theoretical_loss": 3.367027160133773,
      "tokens_seen": 2799894528
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.060503931953138e-05,
      "loss": 1.2072,
      "theoretical_loss": 3.36701497477532,
      "tokens_seen": 2800025600
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.059701492537314e-05,
      "loss": 1.2826,
      "theoretical_loss": 3.3670027901469686,
      "tokens_seen": 2800156672
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.05889905312149e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.3669906062486405,
      "tokens_seen": 2800287744
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.058096613705666e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.3669784230802584,
      "tokens_seen": 2800418816
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.057294174289842e-05,
      "loss": 1.2456,
      "theoretical_loss": 3.3669662406417435,
      "tokens_seen": 2800549888
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.056491734874018e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3669540589330187,
      "tokens_seen": 2800680960
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.055689295458193e-05,
      "loss": 1.2224,
      "theoretical_loss": 3.3669418779540057,
      "tokens_seen": 2800812032
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.054886856042369e-05,
      "loss": 1.2303,
      "theoretical_loss": 3.366929697704627,
      "tokens_seen": 2800943104
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.054084416626545e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.3669175181848043,
      "tokens_seen": 2801074176
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.053281977210721e-05,
      "loss": 1.3,
      "theoretical_loss": 3.36690533939446,
      "tokens_seen": 2801205248
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.052479537794897e-05,
      "loss": 1.2574,
      "theoretical_loss": 3.3668931613335165,
      "tokens_seen": 2801336320
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0516770983790727e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3668809840018956,
      "tokens_seen": 2801467392
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0508746589632486e-05,
      "loss": 1.3572,
      "theoretical_loss": 3.36686880739952,
      "tokens_seen": 2801598464
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.050072219547424e-05,
      "loss": 1.2625,
      "theoretical_loss": 3.366856631526311,
      "tokens_seen": 2801729536
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0492697801316e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.366844456382192,
      "tokens_seen": 2801860608
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.048467340715776e-05,
      "loss": 1.2549,
      "theoretical_loss": 3.3668322819670844,
      "tokens_seen": 2801991680
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.047664901299952e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.3668201082809106,
      "tokens_seen": 2802122752
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.046862461884128e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.3668079353235933,
      "tokens_seen": 2802253824
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.046060022468304e-05,
      "loss": 1.255,
      "theoretical_loss": 3.3667957630950545,
      "tokens_seen": 2802384896
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.04525758305248e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.3667835915952162,
      "tokens_seen": 2802515968
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.044455143636656e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.366771420824001,
      "tokens_seen": 2802647040
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0436527042208312e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.366759250781331,
      "tokens_seen": 2802778112
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0428502648050072e-05,
      "loss": 1.3335,
      "theoretical_loss": 3.3667470814671288,
      "tokens_seen": 2802909184
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0420478253891832e-05,
      "loss": 1.231,
      "theoretical_loss": 3.3667349128813164,
      "tokens_seen": 2803040256
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0412453859733592e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.3667227450238166,
      "tokens_seen": 2803171328
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.040442946557535e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.3667105778945516,
      "tokens_seen": 2803302400
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.9903537034988403,
      "objective/train/avg_token_score": 0.009646275080740452,
      "objective/train/docs_used": 1018418,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2624685764312744,
      "objective/train/original_loss": 1.2745890617370605,
      "objective/train/theoretical_loss": 3.3666984114934437,
      "objective/train/tokens_used": 1173892576,
      "objective/train/unlikelihood_loss": 0.02859090268611908,
      "theoretical_loss": 3.3666984114934437,
      "tokens_seen": 2803433472
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.039640507141711e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3666984114934437,
      "tokens_seen": 2803433472
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.038838067725887e-05,
      "loss": 1.2289,
      "theoretical_loss": 3.366686245820415,
      "tokens_seen": 2803564544
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0380356283100625e-05,
      "loss": 1.21,
      "theoretical_loss": 3.3666740808753883,
      "tokens_seen": 2803695616
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0372331888942384e-05,
      "loss": 1.311,
      "theoretical_loss": 3.366661916658286,
      "tokens_seen": 2803826688
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0364307494784144e-05,
      "loss": 1.3432,
      "theoretical_loss": 3.36664975316903,
      "tokens_seen": 2803957760
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0356283100625904e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.366637590407543,
      "tokens_seen": 2804088832
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0348258706467664e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.366625428373748,
      "tokens_seen": 2804219904
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0340234312309424e-05,
      "loss": 1.272,
      "theoretical_loss": 3.366613267067567,
      "tokens_seen": 2804350976
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0332209918151184e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.3666011064889223,
      "tokens_seen": 2804482048
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0324185523992937e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.3665889466377363,
      "tokens_seen": 2804613120
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0316161129834697e-05,
      "loss": 1.284,
      "theoretical_loss": 3.366576787513932,
      "tokens_seen": 2804744192
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0308136735676457e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.366564629117432,
      "tokens_seen": 2804875264
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0300112341518217e-05,
      "loss": 1.285,
      "theoretical_loss": 3.3665524714481574,
      "tokens_seen": 2805006336
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0292087947359977e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.3665403145060324,
      "tokens_seen": 2805137408
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0284063553201737e-05,
      "loss": 1.303,
      "theoretical_loss": 3.366528158290979,
      "tokens_seen": 2805268480
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0276039159043497e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.3665160028029195,
      "tokens_seen": 2805399552
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0268014764885256e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.3665038480417766,
      "tokens_seen": 2805530624
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.025999037072701e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3664916940074727,
      "tokens_seen": 2805661696
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.025196597656877e-05,
      "loss": 1.262,
      "theoretical_loss": 3.366479540699931,
      "tokens_seen": 2805792768
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.024394158241053e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.3664673881190734,
      "tokens_seen": 2805923840
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.023591718825229e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.3664552362648226,
      "tokens_seen": 2806054912
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.022789279409405e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.366443085137101,
      "tokens_seen": 2806185984
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.021986839993581e-05,
      "loss": 1.317,
      "theoretical_loss": 3.3664309347358325,
      "tokens_seen": 2806317056
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.021184400577757e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.3664187850609384,
      "tokens_seen": 2806448128
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0203819611619322e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.366406636112342,
      "tokens_seen": 2806579200
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0195795217461082e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.366394487889966,
      "tokens_seen": 2806710272
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0187770823302842e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.366382340393732,
      "tokens_seen": 2806841344
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0179746429144602e-05,
      "loss": 1.2332,
      "theoretical_loss": 3.366370193623564,
      "tokens_seen": 2806972416
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0171722034986362e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3663580475793844,
      "tokens_seen": 2807103488
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0163697640828122e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.3663459022611155,
      "tokens_seen": 2807234560
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0155673246669878e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.3663337576686807,
      "tokens_seen": 2807365632
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0147648852511635e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.3663216138020022,
      "tokens_seen": 2807496704
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0139624458353395e-05,
      "loss": 1.355,
      "theoretical_loss": 3.3663094706610024,
      "tokens_seen": 2807627776
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0131600064195155e-05,
      "loss": 1.3221,
      "theoretical_loss": 3.366297328245605,
      "tokens_seen": 2807758848
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0123575670036914e-05,
      "loss": 1.257,
      "theoretical_loss": 3.366285186555732,
      "tokens_seen": 2807889920
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0115551275878674e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.366273045591307,
      "tokens_seen": 2808020992
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.010752688172043e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3662609053522514,
      "tokens_seen": 2808152064
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.009950248756219e-05,
      "loss": 1.295,
      "theoretical_loss": 3.3662487658384896,
      "tokens_seen": 2808283136
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.009147809340395e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.3662366270499433,
      "tokens_seen": 2808414208
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0083453699245707e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.366224488986536,
      "tokens_seen": 2808545280
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0075429305087467e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3662123516481905,
      "tokens_seen": 2808676352
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0067404910929227e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.366200215034829,
      "tokens_seen": 2808807424
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0059380516770984e-05,
      "loss": 1.298,
      "theoretical_loss": 3.366188079146375,
      "tokens_seen": 2808938496
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0051356122612743e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.3661759439827508,
      "tokens_seen": 2809069568
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0043331728454503e-05,
      "loss": 1.1965,
      "theoretical_loss": 3.36616380954388,
      "tokens_seen": 2809200640
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0035307334296263e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3661516758296854,
      "tokens_seen": 2809331712
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.002728294013802e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.3661395428400898,
      "tokens_seen": 2809462784
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.001925854597978e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3661274105750154,
      "tokens_seen": 2809593856
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0011234151821536e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3661152790343865,
      "tokens_seen": 2809724928
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.0003209757663296e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3661031482181247,
      "tokens_seen": 2809856000
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.9916039109230042,
      "objective/train/avg_token_score": 0.008396035060286522,
      "objective/train/docs_used": 1020904,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3400930166244507,
      "objective/train/original_loss": 1.3534297943115234,
      "objective/train/theoretical_loss": 3.366091018126154,
      "objective/train/tokens_used": 1180446176,
      "objective/train/unlikelihood_loss": 0.01811951957643032,
      "theoretical_loss": 3.366091018126154,
      "tokens_seen": 2809987072
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9995185363505056e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.366091018126154,
      "tokens_seen": 2809987072
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9987160969346816e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.3660788887583966,
      "tokens_seen": 2810118144
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9979136575188576e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.3660667601147756,
      "tokens_seen": 2810249216
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9971112181030336e-05,
      "loss": 1.2435,
      "theoretical_loss": 3.366054632195215,
      "tokens_seen": 2810380288
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.996308778687209e-05,
      "loss": 1.2426,
      "theoretical_loss": 3.3660425049996365,
      "tokens_seen": 2810511360
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.995506339271385e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.3660303785279635,
      "tokens_seen": 2810642432
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.994703899855561e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3660182527801195,
      "tokens_seen": 2810773504
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.993901460439737e-05,
      "loss": 1.279,
      "theoretical_loss": 3.3660061277560276,
      "tokens_seen": 2810904576
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.993099021023913e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.3659940034556097,
      "tokens_seen": 2811035648
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.992296581608089e-05,
      "loss": 1.344,
      "theoretical_loss": 3.36598187987879,
      "tokens_seen": 2811166720
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.991494142192265e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.3659697570254914,
      "tokens_seen": 2811297792
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.99069170277644e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.365957634895637,
      "tokens_seen": 2811428864
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.989889263360616e-05,
      "loss": 1.3842,
      "theoretical_loss": 3.365945513489149,
      "tokens_seen": 2811559936
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.989086823944792e-05,
      "loss": 1.257,
      "theoretical_loss": 3.365933392805952,
      "tokens_seen": 2811691008
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.988284384528968e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.3659212728459678,
      "tokens_seen": 2811822080
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.987481945113144e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.3659091536091204,
      "tokens_seen": 2811953152
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.98667950569732e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.365897035095333,
      "tokens_seen": 2812084224
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.985877066281496e-05,
      "loss": 1.2744,
      "theoretical_loss": 3.3658849173045278,
      "tokens_seen": 2812215296
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9850746268656714e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.3658728002366285,
      "tokens_seen": 2812346368
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9842721874498474e-05,
      "loss": 1.3683,
      "theoretical_loss": 3.365860683891559,
      "tokens_seen": 2812477440
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9834697480340234e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3658485682692416,
      "tokens_seen": 2812608512
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9826673086181994e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.3658364533696,
      "tokens_seen": 2812739584
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9818648692023754e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.3658243391925566,
      "tokens_seen": 2812870656
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9810624297865514e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.365812225738036,
      "tokens_seen": 2813001728
    },
    {
      "epoch": 0.7,
      "learning_rate": 2.9802599903707273e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.3658001130059603,
      "tokens_seen": 2813132800
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9794575509549033e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.365788000996253,
      "tokens_seen": 2813263872
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9786551115390786e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.3657758897088375,
      "tokens_seen": 2813394944
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9778526721232546e-05,
      "loss": 1.3338,
      "theoretical_loss": 3.3657637791436374,
      "tokens_seen": 2813526016
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9770502327074306e-05,
      "loss": 1.278,
      "theoretical_loss": 3.3657516693005753,
      "tokens_seen": 2813657088
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9762477932916066e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.365739560179575,
      "tokens_seen": 2813788160
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9754453538757826e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.36572745178056,
      "tokens_seen": 2813919232
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9746429144599586e-05,
      "loss": 1.2117,
      "theoretical_loss": 3.365715344103453,
      "tokens_seen": 2814050304
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9738404750441346e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.365703237148178,
      "tokens_seen": 2814181376
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.97303803562831e-05,
      "loss": 1.3097,
      "theoretical_loss": 3.3656911309146578,
      "tokens_seen": 2814312448
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.972235596212486e-05,
      "loss": 1.2494,
      "theoretical_loss": 3.365679025402816,
      "tokens_seen": 2814443520
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.971433156796662e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.3656669206125756,
      "tokens_seen": 2814574592
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.970630717380838e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.3656548165438607,
      "tokens_seen": 2814705664
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.969828277965014e-05,
      "loss": 1.332,
      "theoretical_loss": 3.3656427131965945,
      "tokens_seen": 2814836736
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.96902583854919e-05,
      "loss": 1.257,
      "theoretical_loss": 3.3656306105706997,
      "tokens_seen": 2814967808
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.968223399133366e-05,
      "loss": 1.2114,
      "theoretical_loss": 3.3656185086661004,
      "tokens_seen": 2815098880
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.967420959717541e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.36560640748272,
      "tokens_seen": 2815229952
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.966618520301717e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.3655943070204817,
      "tokens_seen": 2815361024
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.965816080885893e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.365582207279309,
      "tokens_seen": 2815492096
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.965013641470069e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.365570108259126,
      "tokens_seen": 2815623168
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.964211202054245e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.365558009959855,
      "tokens_seen": 2815754240
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.963408762638421e-05,
      "loss": 1.374,
      "theoretical_loss": 3.36554591238142,
      "tokens_seen": 2815885312
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.962606323222597e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.3655338155237455,
      "tokens_seen": 2816016384
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.961803883806773e-05,
      "loss": 1.1937,
      "theoretical_loss": 3.3655217193867535,
      "tokens_seen": 2816147456
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9610014443909484e-05,
      "loss": 1.2646,
      "theoretical_loss": 3.365509623970368,
      "tokens_seen": 2816278528
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9601990049751244e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.365497529274513,
      "tokens_seen": 2816409600
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.9903793931007385,
      "objective/train/avg_token_score": 0.009620551019906998,
      "objective/train/docs_used": 1023297,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3528820276260376,
      "objective/train/original_loss": 1.3689510822296143,
      "objective/train/theoretical_loss": 3.3654854352991115,
      "objective/train/tokens_used": 1186999776,
      "objective/train/unlikelihood_loss": 0.020886769518256187,
      "theoretical_loss": 3.3654854352991115,
      "tokens_seen": 2816540672
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9593965655593004e-05,
      "loss": 1.379,
      "theoretical_loss": 3.3654854352991115,
      "tokens_seen": 2816540672
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9585941261434764e-05,
      "loss": 1.3377,
      "theoretical_loss": 3.3654733420440874,
      "tokens_seen": 2816671744
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9577916867276524e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.365461249509364,
      "tokens_seen": 2816802816
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9569892473118284e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.3654491576948655,
      "tokens_seen": 2816933888
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9561868078960044e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.365437066600515,
      "tokens_seen": 2817064960
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9553843684801797e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.365424976226236,
      "tokens_seen": 2817196032
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9545819290643557e-05,
      "loss": 1.298,
      "theoretical_loss": 3.365412886571952,
      "tokens_seen": 2817327104
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9537794896485316e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3654007976375873,
      "tokens_seen": 2817458176
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9529770502327076e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.3653887094230646,
      "tokens_seen": 2817589248
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9521746108168836e-05,
      "loss": 1.2456,
      "theoretical_loss": 3.3653766219283083,
      "tokens_seen": 2817720320
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9513721714010596e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.365364535153242,
      "tokens_seen": 2817851392
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9505697319852356e-05,
      "loss": 1.343,
      "theoretical_loss": 3.3653524490977893,
      "tokens_seen": 2817982464
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9497672925694113e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.3653403637618737,
      "tokens_seen": 2818113536
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.948964853153587e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.365328279145419,
      "tokens_seen": 2818244608
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.948162413737763e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.365316195248349,
      "tokens_seen": 2818375680
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.947359974321939e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.3653041120705875,
      "tokens_seen": 2818506752
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.946557534906115e-05,
      "loss": 1.2077,
      "theoretical_loss": 3.365292029612058,
      "tokens_seen": 2818637824
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.945755095490291e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.3652799478726845,
      "tokens_seen": 2818768896
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9449526560744665e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.36526786685239,
      "tokens_seen": 2818899968
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9441502166586425e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.3652557865510992,
      "tokens_seen": 2819031040
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.943347777242818e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.3652437069687355,
      "tokens_seen": 2819162112
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.942545337826994e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.365231628105223,
      "tokens_seen": 2819293184
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.94174289841117e-05,
      "loss": 1.356,
      "theoretical_loss": 3.365219549960485,
      "tokens_seen": 2819424256
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.940940458995346e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.3652074725344456,
      "tokens_seen": 2819555328
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9401380195795218e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.365195395827029,
      "tokens_seen": 2819686400
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9393355801636978e-05,
      "loss": 1.2722,
      "theoretical_loss": 3.3651833198381578,
      "tokens_seen": 2819817472
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9385331407478738e-05,
      "loss": 1.3031,
      "theoretical_loss": 3.3651712445677573,
      "tokens_seen": 2819948544
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9377307013320494e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.3651591700157506,
      "tokens_seen": 2820079616
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9369282619162254e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.3651470961820618,
      "tokens_seen": 2820210688
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9361258225004014e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.365135023066614,
      "tokens_seen": 2820341760
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.935323383084577e-05,
      "loss": 1.2296,
      "theoretical_loss": 3.3651229506693325,
      "tokens_seen": 2820472832
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.934520943668753e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.3651108789901403,
      "tokens_seen": 2820603904
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.933718504252929e-05,
      "loss": 1.319,
      "theoretical_loss": 3.3650988080289617,
      "tokens_seen": 2820734976
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.932916064837105e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.3650867377857203,
      "tokens_seen": 2820866048
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.932113625421281e-05,
      "loss": 1.2478,
      "theoretical_loss": 3.3650746682603403,
      "tokens_seen": 2820997120
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9313111860054567e-05,
      "loss": 1.266,
      "theoretical_loss": 3.365062599452745,
      "tokens_seen": 2821128192
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9305087465896323e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.3650505313628596,
      "tokens_seen": 2821259264
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9297063071738083e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.365038463990607,
      "tokens_seen": 2821390336
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9289038677579843e-05,
      "loss": 1.286,
      "theoretical_loss": 3.365026397335912,
      "tokens_seen": 2821521408
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9281014283421603e-05,
      "loss": 1.32,
      "theoretical_loss": 3.3650143313986978,
      "tokens_seen": 2821652480
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9272989889263363e-05,
      "loss": 1.219,
      "theoretical_loss": 3.365002266178889,
      "tokens_seen": 2821783552
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9264965495105123e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3649902016764095,
      "tokens_seen": 2821914624
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9256941100946876e-05,
      "loss": 1.342,
      "theoretical_loss": 3.364978137891183,
      "tokens_seen": 2822045696
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9248916706788636e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.3649660748231343,
      "tokens_seen": 2822176768
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9240892312630396e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.3649540124721864,
      "tokens_seen": 2822307840
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9232867918472156e-05,
      "loss": 1.311,
      "theoretical_loss": 3.364941950838264,
      "tokens_seen": 2822438912
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9224843524313916e-05,
      "loss": 1.289,
      "theoretical_loss": 3.3649298899212914,
      "tokens_seen": 2822569984
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9216819130155675e-05,
      "loss": 1.3209,
      "theoretical_loss": 3.3649178297211924,
      "tokens_seen": 2822701056
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9208794735997435e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.364905770237891,
      "tokens_seen": 2822832128
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.920077034183919e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.3648937114713116,
      "tokens_seen": 2822963200
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.9929764866828918,
      "objective/train/avg_token_score": 0.007023543585091829,
      "objective/train/docs_used": 1025771,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.299753189086914,
      "objective/train/original_loss": 1.309305191040039,
      "objective/train/theoretical_loss": 3.364881653421378,
      "objective/train/tokens_used": 1193553376,
      "objective/train/unlikelihood_loss": 0.020278848707675934,
      "theoretical_loss": 3.364881653421378,
      "tokens_seen": 2823094272
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.919274594768095e-05,
      "loss": 1.3519,
      "theoretical_loss": 3.364881653421378,
      "tokens_seen": 2823094272
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9184721553522708e-05,
      "loss": 1.347,
      "theoretical_loss": 3.364869596088014,
      "tokens_seen": 2823225344
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9176697159364468e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.3648575394711453,
      "tokens_seen": 2823356416
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9168672765206228e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.3648454835706945,
      "tokens_seen": 2823487488
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9160648371047988e-05,
      "loss": 1.3392,
      "theoretical_loss": 3.3648334283865866,
      "tokens_seen": 2823618560
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9152623976889748e-05,
      "loss": 1.316,
      "theoretical_loss": 3.364821373918745,
      "tokens_seen": 2823749632
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9144599582731508e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3648093201670948,
      "tokens_seen": 2823880704
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.913657518857326e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.3647972671315594,
      "tokens_seen": 2824011776
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.912855079441502e-05,
      "loss": 1.3503,
      "theoretical_loss": 3.3647852148120636,
      "tokens_seen": 2824142848
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.912052640025678e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.3647731632085316,
      "tokens_seen": 2824273920
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.911250200609854e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.3647611123208874,
      "tokens_seen": 2824404992
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.91044776119403e-05,
      "loss": 1.2212,
      "theoretical_loss": 3.364749062149055,
      "tokens_seen": 2824536064
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.909645321778206e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.3647370126929594,
      "tokens_seen": 2824667136
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.908842882362382e-05,
      "loss": 1.2024,
      "theoretical_loss": 3.3647249639525243,
      "tokens_seen": 2824798208
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9080404429465574e-05,
      "loss": 1.3172,
      "theoretical_loss": 3.3647129159276745,
      "tokens_seen": 2824929280
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9072380035307333e-05,
      "loss": 1.2819,
      "theoretical_loss": 3.3647008686183337,
      "tokens_seen": 2825060352
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9064355641149093e-05,
      "loss": 1.2362,
      "theoretical_loss": 3.364688822024427,
      "tokens_seen": 2825191424
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9056331246990853e-05,
      "loss": 1.2052,
      "theoretical_loss": 3.3646767761458776,
      "tokens_seen": 2825322496
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9048306852832613e-05,
      "loss": 1.3207,
      "theoretical_loss": 3.3646647309826108,
      "tokens_seen": 2825453568
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9040282458674373e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3646526865345505,
      "tokens_seen": 2825584640
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9032258064516133e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.364640642801621,
      "tokens_seen": 2825715712
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9024233670357893e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.364628599783747,
      "tokens_seen": 2825846784
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9016209276199646e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.3646165574808524,
      "tokens_seen": 2825977856
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9008184882041406e-05,
      "loss": 1.2578,
      "theoretical_loss": 3.3646045158928626,
      "tokens_seen": 2826108928
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.9000160487883166e-05,
      "loss": 1.2261,
      "theoretical_loss": 3.364592475019701,
      "tokens_seen": 2826240000
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8992136093724926e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.364580434861292,
      "tokens_seen": 2826371072
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8984111699566686e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.3645683954175607,
      "tokens_seen": 2826502144
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8976087305408446e-05,
      "loss": 1.283,
      "theoretical_loss": 3.364556356688431,
      "tokens_seen": 2826633216
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8968062911250205e-05,
      "loss": 1.3302,
      "theoretical_loss": 3.3645443186738277,
      "tokens_seen": 2826764288
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.896003851709196e-05,
      "loss": 1.2433,
      "theoretical_loss": 3.364532281373675,
      "tokens_seen": 2826895360
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.895201412293372e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.364520244787898,
      "tokens_seen": 2827026432
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.894398972877548e-05,
      "loss": 1.265,
      "theoretical_loss": 3.36450820891642,
      "tokens_seen": 2827157504
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8935965334617238e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.364496173759167,
      "tokens_seen": 2827288576
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8927940940458998e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.364484139316062,
      "tokens_seen": 2827419648
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8919916546300758e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.3644721055870304,
      "tokens_seen": 2827550720
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8911892152142518e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.364460072571996,
      "tokens_seen": 2827681792
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.890386775798427e-05,
      "loss": 1.2378,
      "theoretical_loss": 3.3644480402708847,
      "tokens_seen": 2827812864
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.889584336382603e-05,
      "loss": 1.3529,
      "theoretical_loss": 3.36443600868362,
      "tokens_seen": 2827943936
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.888781896966779e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.3644239778101266,
      "tokens_seen": 2828075008
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.887979457550955e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.3644119476503294,
      "tokens_seen": 2828206080
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.887177018135131e-05,
      "loss": 1.3263,
      "theoretical_loss": 3.3643999182041524,
      "tokens_seen": 2828337152
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.886374578719307e-05,
      "loss": 1.241,
      "theoretical_loss": 3.3643878894715207,
      "tokens_seen": 2828468224
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.885572139303483e-05,
      "loss": 1.2306,
      "theoretical_loss": 3.364375861452359,
      "tokens_seen": 2828599296
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.884769699887659e-05,
      "loss": 1.2248,
      "theoretical_loss": 3.3643638341465913,
      "tokens_seen": 2828730368
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8839672604718344e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.3643518075541428,
      "tokens_seen": 2828861440
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8831648210560103e-05,
      "loss": 1.3021,
      "theoretical_loss": 3.3643397816749383,
      "tokens_seen": 2828992512
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8823623816401863e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.364327756508902,
      "tokens_seen": 2829123584
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8815599422243623e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.3643157320559585,
      "tokens_seen": 2829254656
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8807575028085383e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.364303708316033,
      "tokens_seen": 2829385728
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.8799550633927143e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.3642916852890496,
      "tokens_seen": 2829516800
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.9901697039604187,
      "objective/train/avg_token_score": 0.009830369614064693,
      "objective/train/docs_used": 1028281,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.252549648284912,
      "objective/train/original_loss": 1.2670451402664185,
      "objective/train/theoretical_loss": 3.3642796629749334,
      "objective/train/tokens_used": 1200106976,
      "objective/train/unlikelihood_loss": 0.02714966982603073,
      "theoretical_loss": 3.3642796629749334,
      "tokens_seen": 2829647872
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.87915262397689e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.3642796629749334,
      "tokens_seen": 2829647872
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8783501845610656e-05,
      "loss": 1.2238,
      "theoretical_loss": 3.364267641373609,
      "tokens_seen": 2829778944
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8775477451452416e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.364255620485001,
      "tokens_seen": 2829910016
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8767453057294176e-05,
      "loss": 1.31,
      "theoretical_loss": 3.3642436003090346,
      "tokens_seen": 2830041088
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8759428663135936e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.364231580845634,
      "tokens_seen": 2830172160
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8751404268977696e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.364219562094724,
      "tokens_seen": 2830303232
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8743379874819452e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.36420754405623,
      "tokens_seen": 2830434304
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8735355480661212e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.364195526730076,
      "tokens_seen": 2830565376
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.872733108650297e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.364183510116187,
      "tokens_seen": 2830696448
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.871930669234473e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3641714942144882,
      "tokens_seen": 2830827520
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.871128229818649e-05,
      "loss": 1.2445,
      "theoretical_loss": 3.364159479024904,
      "tokens_seen": 2830958592
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.870325790402825e-05,
      "loss": 1.2167,
      "theoretical_loss": 3.3641474645473597,
      "tokens_seen": 2831089664
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8695233509870005e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.3641354507817796,
      "tokens_seen": 2831220736
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8687209115711765e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.3641234377280886,
      "tokens_seen": 2831351808
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8679184721553525e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.364111425386212,
      "tokens_seen": 2831482880
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8671160327395285e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.3640994137560742,
      "tokens_seen": 2831613952
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.866313593323704e-05,
      "loss": 1.3456,
      "theoretical_loss": 3.3640874028376007,
      "tokens_seen": 2831745024
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.86551115390788e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.3640753926307156,
      "tokens_seen": 2831876096
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8647087144920558e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.3640633831353446,
      "tokens_seen": 2832007168
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8639062750762318e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.364051374351412,
      "tokens_seen": 2832138240
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8631038356604077e-05,
      "loss": 1.3244,
      "theoretical_loss": 3.364039366278843,
      "tokens_seen": 2832269312
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8623013962445837e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.3640273589175624,
      "tokens_seen": 2832400384
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8614989568287597e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3640153522674954,
      "tokens_seen": 2832531456
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8606965174129354e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.3640033463285666,
      "tokens_seen": 2832662528
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.859894077997111e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.3639913411007014,
      "tokens_seen": 2832793600
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.859091638581287e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.3639793365838244,
      "tokens_seen": 2832924672
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.858289199165463e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.363967332777861,
      "tokens_seen": 2833055744
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.857486759749639e-05,
      "loss": 1.3333,
      "theoretical_loss": 3.363955329682736,
      "tokens_seen": 2833186816
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.856684320333815e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.363943327298374,
      "tokens_seen": 2833317888
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.855881880917991e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.363931325624701,
      "tokens_seen": 2833448960
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.855079441502167e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3639193246616417,
      "tokens_seen": 2833580032
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8542770020863423e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.36390732440912,
      "tokens_seen": 2833711104
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8534745626705183e-05,
      "loss": 1.3396,
      "theoretical_loss": 3.363895324867063,
      "tokens_seen": 2833842176
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8526721232546943e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.363883326035394,
      "tokens_seen": 2833973248
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8518696838388703e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.3638713279140395,
      "tokens_seen": 2834104320
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8510672444230462e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.3638593305029234,
      "tokens_seen": 2834235392
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8502648050072222e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.363847333801971,
      "tokens_seen": 2834366464
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8494623655913982e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.3638353378111083,
      "tokens_seen": 2834497536
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8486599261755735e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.3638233425302597,
      "tokens_seen": 2834628608
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8478574867597495e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.3638113479593508,
      "tokens_seen": 2834759680
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8470550473439255e-05,
      "loss": 1.3439,
      "theoretical_loss": 3.363799354098306,
      "tokens_seen": 2834890752
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8462526079281015e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.363787360947051,
      "tokens_seen": 2835021824
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8454501685122775e-05,
      "loss": 1.3436,
      "theoretical_loss": 3.3637753685055114,
      "tokens_seen": 2835152896
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8446477290964535e-05,
      "loss": 1.29,
      "theoretical_loss": 3.363763376773612,
      "tokens_seen": 2835283968
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8438452896806295e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.3637513857512773,
      "tokens_seen": 2835415040
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8430428502648048e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.363739395438434,
      "tokens_seen": 2835546112
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8422404108489808e-05,
      "loss": 1.3493,
      "theoretical_loss": 3.3637274058350055,
      "tokens_seen": 2835677184
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8414379714331568e-05,
      "loss": 1.2102,
      "theoretical_loss": 3.3637154169409187,
      "tokens_seen": 2835808256
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8406355320173328e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.363703428756098,
      "tokens_seen": 2835939328
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8398330926015088e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.3636914412804684,
      "tokens_seen": 2836070400
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.9904579520225525,
      "objective/train/avg_token_score": 0.009542101062834263,
      "objective/train/docs_used": 1030811,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4244540929794312,
      "objective/train/original_loss": 1.4394457340240479,
      "objective/train/theoretical_loss": 3.3636794545139566,
      "objective/train/tokens_used": 1206660576,
      "objective/train/unlikelihood_loss": 0.024126995354890823,
      "theoretical_loss": 3.3636794545139566,
      "tokens_seen": 2836201472
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8390306531856847e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.3636794545139566,
      "tokens_seen": 2836201472
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8382282137698607e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.363667468456486,
      "tokens_seen": 2836332544
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8374257743540367e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.363655483107983,
      "tokens_seen": 2836463616
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.836623334938212e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.363643498468373,
      "tokens_seen": 2836594688
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.835820895522388e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.363631514537581,
      "tokens_seen": 2836725760
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.835018456106564e-05,
      "loss": 1.3685,
      "theoretical_loss": 3.363619531315532,
      "tokens_seen": 2836856832
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.83421601669074e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.3636075488021517,
      "tokens_seen": 2836987904
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.833413577274916e-05,
      "loss": 1.331,
      "theoretical_loss": 3.3635955669973656,
      "tokens_seen": 2837118976
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.832611137859092e-05,
      "loss": 1.246,
      "theoretical_loss": 3.363583585901099,
      "tokens_seen": 2837250048
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.831808698443268e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.363571605513277,
      "tokens_seen": 2837381120
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8310062590274433e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.363559625833825,
      "tokens_seen": 2837512192
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8302038196116193e-05,
      "loss": 1.3045,
      "theoretical_loss": 3.363547646862669,
      "tokens_seen": 2837643264
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8294013801957953e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.3635356685997335,
      "tokens_seen": 2837774336
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8285989407799713e-05,
      "loss": 1.3611,
      "theoretical_loss": 3.3635236910449446,
      "tokens_seen": 2837905408
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8277965013641473e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3635117141982276,
      "tokens_seen": 2838036480
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8269940619483233e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.363499738059508,
      "tokens_seen": 2838167552
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8261916225324992e-05,
      "loss": 1.2015,
      "theoretical_loss": 3.363487762628711,
      "tokens_seen": 2838298624
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8253891831166746e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.363475787905762,
      "tokens_seen": 2838429696
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8245867437008505e-05,
      "loss": 1.278,
      "theoretical_loss": 3.363463813890587,
      "tokens_seen": 2838560768
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8237843042850265e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.3634518405831106,
      "tokens_seen": 2838691840
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8229818648692025e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.363439867983259,
      "tokens_seen": 2838822912
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8221794254533785e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.363427896090958,
      "tokens_seen": 2838953984
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8213769860375545e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3634159249061324,
      "tokens_seen": 2839085056
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8205745466217305e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.363403954428708,
      "tokens_seen": 2839216128
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8197721072059065e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.3633919846586102,
      "tokens_seen": 2839347200
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8189696677900818e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.363380015595765,
      "tokens_seen": 2839478272
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8181672283742578e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3633680472400975,
      "tokens_seen": 2839609344
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8173647889584338e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.3633560795915334,
      "tokens_seen": 2839740416
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8165623495426098e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.3633441126499983,
      "tokens_seen": 2839871488
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8157599101267858e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.3633321464154182,
      "tokens_seen": 2840002560
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8149574707109618e-05,
      "loss": 1.3682,
      "theoretical_loss": 3.363320180887718,
      "tokens_seen": 2840133632
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8141550312951377e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.363308216066824,
      "tokens_seen": 2840264704
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.813352591879313e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.363296251952661,
      "tokens_seen": 2840395776
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.812550152463489e-05,
      "loss": 1.2245,
      "theoretical_loss": 3.3632842885451555,
      "tokens_seen": 2840526848
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.811747713047665e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.3632723258442327,
      "tokens_seen": 2840657920
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.810945273631841e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.363260363849818,
      "tokens_seen": 2840788992
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.810142834216017e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.3632484025618385,
      "tokens_seen": 2840920064
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8093403948001927e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.363236441980218,
      "tokens_seen": 2841051136
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8085379553843687e-05,
      "loss": 1.3372,
      "theoretical_loss": 3.3632244821048833,
      "tokens_seen": 2841182208
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8077355159685447e-05,
      "loss": 1.3388,
      "theoretical_loss": 3.36321252293576,
      "tokens_seen": 2841313280
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8069330765527203e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.3632005644727734,
      "tokens_seen": 2841444352
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8061306371368963e-05,
      "loss": 1.3426,
      "theoretical_loss": 3.3631886067158496,
      "tokens_seen": 2841575424
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8053281977210723e-05,
      "loss": 1.211,
      "theoretical_loss": 3.3631766496649145,
      "tokens_seen": 2841706496
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.804525758305248e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3631646933198933,
      "tokens_seen": 2841837568
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.803723318889424e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3631527376807124,
      "tokens_seen": 2841968640
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8029208794736e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.3631407827472968,
      "tokens_seen": 2842099712
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.802118440057776e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.3631288285195735,
      "tokens_seen": 2842230784
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8013160006419516e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.363116874997467,
      "tokens_seen": 2842361856
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.8005135612261276e-05,
      "loss": 1.3617,
      "theoretical_loss": 3.363104922180904,
      "tokens_seen": 2842492928
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7997111218103032e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.36309297006981,
      "tokens_seen": 2842624000
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.9920662641525269,
      "objective/train/avg_token_score": 0.007933743298053741,
      "objective/train/docs_used": 1033097,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2653154134750366,
      "objective/train/original_loss": 1.2757244110107422,
      "objective/train/theoretical_loss": 3.3630810186641105,
      "objective/train/tokens_used": 1213214176,
      "objective/train/unlikelihood_loss": 0.02240237407386303,
      "theoretical_loss": 3.3630810186641105,
      "tokens_seen": 2842755072
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7989086823944792e-05,
      "loss": 1.263,
      "theoretical_loss": 3.3630810186641105,
      "tokens_seen": 2842755072
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7981062429786552e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.363069067963732,
      "tokens_seen": 2842886144
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7973038035628312e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.3630571179685997,
      "tokens_seen": 2843017216
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.796501364147007e-05,
      "loss": 1.2486,
      "theoretical_loss": 3.3630451686786405,
      "tokens_seen": 2843148288
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7956989247311828e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.363033220093779,
      "tokens_seen": 2843279360
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7948964853153585e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.363021272213942,
      "tokens_seen": 2843410432
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7940940458995345e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.3630093250390547,
      "tokens_seen": 2843541504
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7932916064837105e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.3629973785690437,
      "tokens_seen": 2843672576
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7924891670678864e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.3629854328038347,
      "tokens_seen": 2843803648
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7916867276520624e-05,
      "loss": 1.2263,
      "theoretical_loss": 3.3629734877433535,
      "tokens_seen": 2843934720
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7908842882362384e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.3629615433875264,
      "tokens_seen": 2844065792
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7900818488204144e-05,
      "loss": 1.286,
      "theoretical_loss": 3.362949599736279,
      "tokens_seen": 2844196864
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7892794094045897e-05,
      "loss": 1.297,
      "theoretical_loss": 3.362937656789537,
      "tokens_seen": 2844327936
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7884769699887657e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.362925714547227,
      "tokens_seen": 2844459008
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7876745305729417e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.3629137730092746,
      "tokens_seen": 2844590080
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7868720911571177e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.362901832175606,
      "tokens_seen": 2844721152
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7860696517412937e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.362889892046147,
      "tokens_seen": 2844852224
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7852672123254697e-05,
      "loss": 1.27,
      "theoretical_loss": 3.362877952620824,
      "tokens_seen": 2844983296
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7844647729096457e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.3628660138995627,
      "tokens_seen": 2845114368
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.783662333493821e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.3628540758822894,
      "tokens_seen": 2845245440
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.782859894077997e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.36284213856893,
      "tokens_seen": 2845376512
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.782057454662173e-05,
      "loss": 1.2745,
      "theoretical_loss": 3.3628302019594107,
      "tokens_seen": 2845507584
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.781255015246349e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.362818266053657,
      "tokens_seen": 2845638656
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.780452575830525e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.3628063308515963,
      "tokens_seen": 2845769728
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.779650136414701e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.3627943963531535,
      "tokens_seen": 2845900800
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.778847696998877e-05,
      "loss": 1.324,
      "theoretical_loss": 3.362782462558255,
      "tokens_seen": 2846031872
    },
    {
      "epoch": 0.72,
      "learning_rate": 2.7780452575830522e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.362770529466827,
      "tokens_seen": 2846162944
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7772428181672282e-05,
      "loss": 1.2148,
      "theoretical_loss": 3.3627585970787957,
      "tokens_seen": 2846294016
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7764403787514042e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.362746665394087,
      "tokens_seen": 2846425088
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7756379393355802e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.3627347344126277,
      "tokens_seen": 2846556160
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7748354999197562e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.3627228041343433,
      "tokens_seen": 2846687232
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7740330605039322e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.36271087455916,
      "tokens_seen": 2846818304
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7732306210881082e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.3626989456870047,
      "tokens_seen": 2846949376
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7724281816722842e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.362687017517803,
      "tokens_seen": 2847080448
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7716257422564595e-05,
      "loss": 1.227,
      "theoretical_loss": 3.3626750900514812,
      "tokens_seen": 2847211520
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7708233028406355e-05,
      "loss": 1.3001,
      "theoretical_loss": 3.3626631632879653,
      "tokens_seen": 2847342592
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7700208634248115e-05,
      "loss": 1.322,
      "theoretical_loss": 3.3626512372271824,
      "tokens_seen": 2847473664
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7692184240089875e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.3626393118690574,
      "tokens_seen": 2847604736
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7684159845931635e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.362627387213518,
      "tokens_seen": 2847735808
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7676135451773394e-05,
      "loss": 1.242,
      "theoretical_loss": 3.3626154632604894,
      "tokens_seen": 2847866880
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7668111057615154e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.3626035400098986,
      "tokens_seen": 2847997952
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7660086663456907e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.3625916174616712,
      "tokens_seen": 2848129024
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7652062269298667e-05,
      "loss": 1.1927,
      "theoretical_loss": 3.362579695615734,
      "tokens_seen": 2848260096
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7644037875140427e-05,
      "loss": 1.2398,
      "theoretical_loss": 3.3625677744720135,
      "tokens_seen": 2848391168
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7636013480982187e-05,
      "loss": 1.2971,
      "theoretical_loss": 3.362555854030435,
      "tokens_seen": 2848522240
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7627989086823947e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.3625439342909265,
      "tokens_seen": 2848653312
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7619964692665707e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.3625320152534126,
      "tokens_seen": 2848784384
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7611940298507467e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.3625200969178213,
      "tokens_seen": 2848915456
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7603915904349227e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3625081792840774,
      "tokens_seen": 2849046528
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.759589151019098e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.362496262352108,
      "tokens_seen": 2849177600
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.9938709139823914,
      "objective/train/avg_token_score": 0.006129080429673195,
      "objective/train/docs_used": 1035566,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2274131774902344,
      "objective/train/original_loss": 1.2361645698547363,
      "objective/train/theoretical_loss": 3.3624843461218403,
      "objective/train/tokens_used": 1219767776,
      "objective/train/unlikelihood_loss": 0.015049198642373085,
      "theoretical_loss": 3.3624843461218403,
      "tokens_seen": 2849308672
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.758786711603274e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.3624843461218403,
      "tokens_seen": 2849308672
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.75798427218745e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.3624724305931992,
      "tokens_seen": 2849439744
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.757181832771626e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.3624605157661125,
      "tokens_seen": 2849570816
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.756379393355802e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3624486016405055,
      "tokens_seen": 2849701888
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.755576953939978e-05,
      "loss": 1.2436,
      "theoretical_loss": 3.362436688216305,
      "tokens_seen": 2849832960
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.754774514524154e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.3624247754934378,
      "tokens_seen": 2849964032
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7539720751083292e-05,
      "loss": 1.253,
      "theoretical_loss": 3.36241286347183,
      "tokens_seen": 2850095104
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7531696356925052e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.3624009521514084,
      "tokens_seen": 2850226176
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7523671962766812e-05,
      "loss": 1.2393,
      "theoretical_loss": 3.362389041532099,
      "tokens_seen": 2850357248
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7515647568608572e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.3623771316138287,
      "tokens_seen": 2850488320
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7507623174450332e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.362365222396524,
      "tokens_seen": 2850619392
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7499598780292092e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.3623533138801114,
      "tokens_seen": 2850750464
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7491574386133852e-05,
      "loss": 1.3264,
      "theoretical_loss": 3.362341406064517,
      "tokens_seen": 2850881536
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7483549991975605e-05,
      "loss": 1.269,
      "theoretical_loss": 3.3623294989496677,
      "tokens_seen": 2851012608
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7475525597817365e-05,
      "loss": 1.2187,
      "theoretical_loss": 3.36231759253549,
      "tokens_seen": 2851143680
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7467501203659125e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.362305686821911,
      "tokens_seen": 2851274752
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7459476809500885e-05,
      "loss": 1.3419,
      "theoretical_loss": 3.362293781808856,
      "tokens_seen": 2851405824
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7451452415342645e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.362281877496253,
      "tokens_seen": 2851536896
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7443428021184405e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.362269973884027,
      "tokens_seen": 2851667968
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.743540362702616e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.3622580709721066,
      "tokens_seen": 2851799040
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.742737923286792e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.3622461687604166,
      "tokens_seen": 2851930112
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7419354838709678e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.3622342672488847,
      "tokens_seen": 2852061184
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7411330444551437e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.3622223664374373,
      "tokens_seen": 2852192256
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7403306050393197e-05,
      "loss": 1.2246,
      "theoretical_loss": 3.3622104663260006,
      "tokens_seen": 2852323328
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7395281656234957e-05,
      "loss": 1.235,
      "theoretical_loss": 3.362198566914502,
      "tokens_seen": 2852454400
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7387257262076714e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3621866682028676,
      "tokens_seen": 2852585472
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7379232867918474e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3621747701910243,
      "tokens_seen": 2852716544
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7371208473760234e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.3621628728788986,
      "tokens_seen": 2852847616
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.736318407960199e-05,
      "loss": 1.3287,
      "theoretical_loss": 3.3621509762664177,
      "tokens_seen": 2852978688
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.735515968544375e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.362139080353508,
      "tokens_seen": 2853109760
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.734713529128551e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.362127185140096,
      "tokens_seen": 2853240832
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7339110897127266e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.3621152906261083,
      "tokens_seen": 2853371904
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7331086502969026e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.3621033968114724,
      "tokens_seen": 2853502976
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7323062108810786e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.3620915036961145,
      "tokens_seen": 2853634048
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7315037714652546e-05,
      "loss": 1.2404,
      "theoretical_loss": 3.362079611279962,
      "tokens_seen": 2853765120
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7307013320494303e-05,
      "loss": 1.2473,
      "theoretical_loss": 3.3620677195629405,
      "tokens_seen": 2853896192
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7298988926336063e-05,
      "loss": 1.1671,
      "theoretical_loss": 3.362055828544978,
      "tokens_seen": 2854027264
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.729096453217782e-05,
      "loss": 1.3067,
      "theoretical_loss": 3.362043938226001,
      "tokens_seen": 2854158336
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.728294013801958e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.3620320486059354,
      "tokens_seen": 2854289408
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.727491574386134e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.3620201596847092,
      "tokens_seen": 2854420480
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.72668913497031e-05,
      "loss": 1.334,
      "theoretical_loss": 3.362008271462249,
      "tokens_seen": 2854551552
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.725886695554486e-05,
      "loss": 1.3494,
      "theoretical_loss": 3.361996383938481,
      "tokens_seen": 2854682624
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.725084256138662e-05,
      "loss": 1.2082,
      "theoretical_loss": 3.361984497113333,
      "tokens_seen": 2854813696
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7242818167228372e-05,
      "loss": 1.2527,
      "theoretical_loss": 3.3619726109867307,
      "tokens_seen": 2854944768
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.723479377307013e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.3619607255586024,
      "tokens_seen": 2855075840
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.722676937891189e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.3619488408288736,
      "tokens_seen": 2855206912
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.721874498475365e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.361936956797472,
      "tokens_seen": 2855337984
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.721072059059541e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.3619250734643247,
      "tokens_seen": 2855469056
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.720269619643717e-05,
      "loss": 1.3614,
      "theoretical_loss": 3.361913190829358,
      "tokens_seen": 2855600128
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.719467180227893e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.361901308892499,
      "tokens_seen": 2855731200
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.9928230047225952,
      "objective/train/avg_token_score": 0.007176929619163275,
      "objective/train/docs_used": 1037860,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3564598560333252,
      "objective/train/original_loss": 1.3679580688476562,
      "objective/train/theoretical_loss": 3.361889427653675,
      "objective/train/tokens_used": 1226321376,
      "objective/train/unlikelihood_loss": 0.018936797976493835,
      "theoretical_loss": 3.361889427653675,
      "tokens_seen": 2855862272
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7186647408120684e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.361889427653675,
      "tokens_seen": 2855862272
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7178623013962444e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.3618775471128126,
      "tokens_seen": 2855993344
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7170598619804204e-05,
      "loss": 1.3523,
      "theoretical_loss": 3.361865667269839,
      "tokens_seen": 2856124416
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7162574225645964e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.361853788124681,
      "tokens_seen": 2856255488
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7154549831487724e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.3618419096772656,
      "tokens_seen": 2856386560
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7146525437329484e-05,
      "loss": 1.301,
      "theoretical_loss": 3.3618300319275196,
      "tokens_seen": 2856517632
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7138501043171244e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.3618181548753707,
      "tokens_seen": 2856648704
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7130476649013004e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.3618062785207456,
      "tokens_seen": 2856779776
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7122452254854757e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.361794402863571,
      "tokens_seen": 2856910848
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7114427860696517e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.3617825279037743,
      "tokens_seen": 2857041920
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7106403466538277e-05,
      "loss": 1.3515,
      "theoretical_loss": 3.3617706536412824,
      "tokens_seen": 2857172992
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7098379072380037e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.3617587800760225,
      "tokens_seen": 2857304064
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7090354678221796e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.3617469072079214,
      "tokens_seen": 2857435136
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7082330284063556e-05,
      "loss": 1.2251,
      "theoretical_loss": 3.361735035036906,
      "tokens_seen": 2857566208
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7074305889905316e-05,
      "loss": 1.226,
      "theoretical_loss": 3.3617231635629046,
      "tokens_seen": 2857697280
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.706628149574707e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.3617112927858432,
      "tokens_seen": 2857828352
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.705825710158883e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.361699422705649,
      "tokens_seen": 2857959424
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.705023270743059e-05,
      "loss": 1.3038,
      "theoretical_loss": 3.3616875533222492,
      "tokens_seen": 2858090496
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.704220831327235e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.3616756846355718,
      "tokens_seen": 2858221568
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.703418391911411e-05,
      "loss": 1.2276,
      "theoretical_loss": 3.3616638166455424,
      "tokens_seen": 2858352640
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.702615952495587e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3616519493520896,
      "tokens_seen": 2858483712
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.701813513079763e-05,
      "loss": 1.317,
      "theoretical_loss": 3.36164008275514,
      "tokens_seen": 2858614784
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7010110736639382e-05,
      "loss": 1.2662,
      "theoretical_loss": 3.3616282168546205,
      "tokens_seen": 2858745856
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.7002086342481142e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.3616163516504587,
      "tokens_seen": 2858876928
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6994061948322902e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.3616044871425816,
      "tokens_seen": 2859008000
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.698603755416466e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.3615926233309166,
      "tokens_seen": 2859139072
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.697801316000642e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.361580760215391,
      "tokens_seen": 2859270144
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.696998876584818e-05,
      "loss": 1.3086,
      "theoretical_loss": 3.3615688977959315,
      "tokens_seen": 2859401216
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.696196437168994e-05,
      "loss": 1.2155,
      "theoretical_loss": 3.361557036072466,
      "tokens_seen": 2859532288
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.69539399775317e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.3615451750449212,
      "tokens_seen": 2859663360
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6945915583373454e-05,
      "loss": 1.3572,
      "theoretical_loss": 3.361533314713225,
      "tokens_seen": 2859794432
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6937891189215214e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.3615214550773045,
      "tokens_seen": 2859925504
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6929866795056974e-05,
      "loss": 1.3207,
      "theoretical_loss": 3.3615095961370867,
      "tokens_seen": 2860056576
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6921842400898734e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.361497737892499,
      "tokens_seen": 2860187648
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6913818006740494e-05,
      "loss": 1.3359,
      "theoretical_loss": 3.361485880343469,
      "tokens_seen": 2860318720
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6905793612582254e-05,
      "loss": 1.259,
      "theoretical_loss": 3.3614740234899236,
      "tokens_seen": 2860449792
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6897769218424014e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.3614621673317906,
      "tokens_seen": 2860580864
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6889744824265767e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.361450311868997,
      "tokens_seen": 2860711936
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6881720430107527e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.36143845710147,
      "tokens_seen": 2860843008
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6873696035949287e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.3614266030291375,
      "tokens_seen": 2860974080
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6865671641791047e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.361414749651927,
      "tokens_seen": 2861105152
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6857647247632807e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.361402896969765,
      "tokens_seen": 2861236224
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6849622853474566e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.3613910449825797,
      "tokens_seen": 2861367296
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6841598459316326e-05,
      "loss": 1.3466,
      "theoretical_loss": 3.361379193690298,
      "tokens_seen": 2861498368
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.683357406515808e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.361367343092848,
      "tokens_seen": 2861629440
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.682554967099984e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.361355493190157,
      "tokens_seen": 2861760512
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.68175252768416e-05,
      "loss": 1.255,
      "theoretical_loss": 3.3613436439821514,
      "tokens_seen": 2861891584
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.680950088268336e-05,
      "loss": 1.2024,
      "theoretical_loss": 3.3613317954687596,
      "tokens_seen": 2862022656
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.680147648852512e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.361319947649909,
      "tokens_seen": 2862153728
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.679345209436688e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.361308100525527,
      "tokens_seen": 2862284800
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.9913198947906494,
      "objective/train/avg_token_score": 0.00868011824786663,
      "objective/train/docs_used": 1040389,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.330885410308838,
      "objective/train/original_loss": 1.343265414237976,
      "objective/train/theoretical_loss": 3.3612962540955413,
      "objective/train/tokens_used": 1232874976,
      "objective/train/unlikelihood_loss": 0.022511886432766914,
      "theoretical_loss": 3.3612962540955413,
      "tokens_seen": 2862415872
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.678542770020864e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.3612962540955413,
      "tokens_seen": 2862415872
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6777403306050395e-05,
      "loss": 1.2336,
      "theoretical_loss": 3.3612844083598796,
      "tokens_seen": 2862546944
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.6769378911892152e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.3612725633184684,
      "tokens_seen": 2862678016
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6761354517733912e-05,
      "loss": 1.274,
      "theoretical_loss": 3.361260718971236,
      "tokens_seen": 2862809088
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6753330123575672e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.36124887531811,
      "tokens_seen": 2862940160
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6745305729417432e-05,
      "loss": 1.3473,
      "theoretical_loss": 3.3612370323590173,
      "tokens_seen": 2863071232
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.673728133525919e-05,
      "loss": 1.3155,
      "theoretical_loss": 3.3612251900938857,
      "tokens_seen": 2863202304
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6729256941100948e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.3612133485226434,
      "tokens_seen": 2863333376
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6721232546942708e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.3612015076452177,
      "tokens_seen": 2863464448
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6713208152784465e-05,
      "loss": 1.2417,
      "theoretical_loss": 3.361189667461536,
      "tokens_seen": 2863595520
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6705183758626224e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.3611778279715256,
      "tokens_seen": 2863726592
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6697159364467984e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.361165989175115,
      "tokens_seen": 2863857664
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6689134970309744e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.3611541510722307,
      "tokens_seen": 2863988736
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.66811105761515e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.361142313662801,
      "tokens_seen": 2864119808
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.667308618199326e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.3611304769467534,
      "tokens_seen": 2864250880
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.666506178783502e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.361118640924016,
      "tokens_seen": 2864381952
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.665703739367678e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.361106805594516,
      "tokens_seen": 2864513024
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6649012999518537e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.361094970958181,
      "tokens_seen": 2864644096
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6640988605360297e-05,
      "loss": 1.3684,
      "theoretical_loss": 3.3610831370149388,
      "tokens_seen": 2864775168
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6632964211202053e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.3610713037647173,
      "tokens_seen": 2864906240
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6624939817043813e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.361059471207444,
      "tokens_seen": 2865037312
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6616915422885573e-05,
      "loss": 1.2732,
      "theoretical_loss": 3.361047639343047,
      "tokens_seen": 2865168384
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6608891028727333e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.3610358081714535,
      "tokens_seen": 2865299456
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6600866634569093e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.361023977692591,
      "tokens_seen": 2865430528
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.659284224041085e-05,
      "loss": 1.2128,
      "theoretical_loss": 3.3610121479063886,
      "tokens_seen": 2865561600
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6584817846252606e-05,
      "loss": 1.3692,
      "theoretical_loss": 3.3610003188127724,
      "tokens_seen": 2865692672
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6576793452094366e-05,
      "loss": 1.3034,
      "theoretical_loss": 3.3609884904116716,
      "tokens_seen": 2865823744
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6568769057936126e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.3609766627030133,
      "tokens_seen": 2865954816
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6560744663777886e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.3609648356867248,
      "tokens_seen": 2866085888
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6552720269619646e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.360953009362735,
      "tokens_seen": 2866216960
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6544695875461406e-05,
      "loss": 1.2276,
      "theoretical_loss": 3.360941183730971,
      "tokens_seen": 2866348032
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.653667148130316e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.3609293587913607,
      "tokens_seen": 2866479104
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.652864708714492e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.360917534543832,
      "tokens_seen": 2866610176
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.652062269298668e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.3609057109883134,
      "tokens_seen": 2866741248
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.651259829882844e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.3608938881247314,
      "tokens_seen": 2866872320
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.65045739046702e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.3608820659530148,
      "tokens_seen": 2867003392
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6496549510511958e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.360870244473092,
      "tokens_seen": 2867134464
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6488525116353718e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.3608584236848893,
      "tokens_seen": 2867265536
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6480500722195478e-05,
      "loss": 1.3497,
      "theoretical_loss": 3.360846603588336,
      "tokens_seen": 2867396608
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.647247632803723e-05,
      "loss": 1.2284,
      "theoretical_loss": 3.3608347841833597,
      "tokens_seen": 2867527680
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.646445193387899e-05,
      "loss": 1.2288,
      "theoretical_loss": 3.360822965469888,
      "tokens_seen": 2867658752
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.645642753972075e-05,
      "loss": 1.2685,
      "theoretical_loss": 3.3608111474478486,
      "tokens_seen": 2867789824
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.644840314556251e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.3607993301171706,
      "tokens_seen": 2867920896
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.644037875140427e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.360787513477781,
      "tokens_seen": 2868051968
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.643235435724603e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.3607756975296077,
      "tokens_seen": 2868183040
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.642432996308779e-05,
      "loss": 1.3403,
      "theoretical_loss": 3.3607638822725794,
      "tokens_seen": 2868314112
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6416305568929544e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.360752067706623,
      "tokens_seen": 2868445184
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6408281174771304e-05,
      "loss": 1.2321,
      "theoretical_loss": 3.360740253831668,
      "tokens_seen": 2868576256
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6400256780613064e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.360728440647641,
      "tokens_seen": 2868707328
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6392232386454824e-05,
      "loss": 1.3363,
      "theoretical_loss": 3.360716628154471,
      "tokens_seen": 2868838400
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.9911837577819824,
      "objective/train/avg_token_score": 0.008816240355372429,
      "objective/train/docs_used": 1042847,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3210104703903198,
      "objective/train/original_loss": 1.3339134454727173,
      "objective/train/theoretical_loss": 3.3607048163520856,
      "objective/train/tokens_used": 1239428576,
      "objective/train/unlikelihood_loss": 0.025549929589033127,
      "theoretical_loss": 3.3607048163520856,
      "tokens_seen": 2868969472
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6384207992296583e-05,
      "loss": 1.341,
      "theoretical_loss": 3.3607048163520856,
      "tokens_seen": 2868969472
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6376183598138343e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.360693005240413,
      "tokens_seen": 2869100544
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6368159203980103e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3606811948193807,
      "tokens_seen": 2869231616
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6360134809821856e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.3606693850889178,
      "tokens_seen": 2869362688
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6352110415663616e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.3606575760489514,
      "tokens_seen": 2869493760
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6344086021505376e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.36064576769941,
      "tokens_seen": 2869624832
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6336061627347136e-05,
      "loss": 1.2765,
      "theoretical_loss": 3.360633960040222,
      "tokens_seen": 2869755904
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6328037233188896e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.360622153071315,
      "tokens_seen": 2869886976
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6320012839030656e-05,
      "loss": 1.2431,
      "theoretical_loss": 3.360610346792617,
      "tokens_seen": 2870018048
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6311988444872416e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.360598541204057,
      "tokens_seen": 2870149120
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6303964050714176e-05,
      "loss": 1.2848,
      "theoretical_loss": 3.360586736305563,
      "tokens_seen": 2870280192
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.629593965655593e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.360574932097062,
      "tokens_seen": 2870411264
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.628791526239769e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.360563128578483,
      "tokens_seen": 2870542336
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.627989086823945e-05,
      "loss": 1.3325,
      "theoretical_loss": 3.3605513257497543,
      "tokens_seen": 2870673408
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.627186647408121e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3605395236108038,
      "tokens_seen": 2870804480
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.626384207992297e-05,
      "loss": 1.2389,
      "theoretical_loss": 3.3605277221615597,
      "tokens_seen": 2870935552
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.625581768576473e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.3605159214019507,
      "tokens_seen": 2871066624
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6247793291606488e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.3605041213319042,
      "tokens_seen": 2871197696
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.623976889744824e-05,
      "loss": 1.3409,
      "theoretical_loss": 3.360492321951349,
      "tokens_seen": 2871328768
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.623174450329e-05,
      "loss": 1.249,
      "theoretical_loss": 3.3604805232602133,
      "tokens_seen": 2871459840
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.622372010913176e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.360468725258425,
      "tokens_seen": 2871590912
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.621569571497352e-05,
      "loss": 1.2263,
      "theoretical_loss": 3.360456927945913,
      "tokens_seen": 2871721984
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.620767132081528e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.3604451313226047,
      "tokens_seen": 2871853056
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.619964692665704e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.360433335388429,
      "tokens_seen": 2871984128
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.61916225324988e-05,
      "loss": 1.2907,
      "theoretical_loss": 3.360421540143314,
      "tokens_seen": 2872115200
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.618359813834056e-05,
      "loss": 1.29,
      "theoretical_loss": 3.360409745587188,
      "tokens_seen": 2872246272
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6175573744182314e-05,
      "loss": 1.2249,
      "theoretical_loss": 3.36039795171998,
      "tokens_seen": 2872377344
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6167549350024074e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.3603861585416173,
      "tokens_seen": 2872508416
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6159524955865834e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.360374366052029,
      "tokens_seen": 2872639488
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6151500561707594e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.3603625742511425,
      "tokens_seen": 2872770560
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6143476167549354e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.3603507831388866,
      "tokens_seen": 2872901632
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6135451773391113e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.3603389927151905,
      "tokens_seen": 2873032704
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6127427379232873e-05,
      "loss": 1.276,
      "theoretical_loss": 3.3603272029799816,
      "tokens_seen": 2873163776
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6119402985074626e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.3603154139331886,
      "tokens_seen": 2873294848
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6111378590916386e-05,
      "loss": 1.2027,
      "theoretical_loss": 3.3603036255747396,
      "tokens_seen": 2873425920
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6103354196758146e-05,
      "loss": 1.3445,
      "theoretical_loss": 3.3602918379045636,
      "tokens_seen": 2873556992
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6095329802599906e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.3602800509225887,
      "tokens_seen": 2873688064
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6087305408441666e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.3602682646287434,
      "tokens_seen": 2873819136
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6079281014283426e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.360256479022956,
      "tokens_seen": 2873950208
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6071256620125182e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.3602446941051554,
      "tokens_seen": 2874081280
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.606323222596694e-05,
      "loss": 1.2223,
      "theoretical_loss": 3.3602329098752692,
      "tokens_seen": 2874212352
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.60552078318087e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3602211263332267,
      "tokens_seen": 2874343424
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.604718343765046e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.3602093434789557,
      "tokens_seen": 2874474496
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.603915904349222e-05,
      "loss": 1.1889,
      "theoretical_loss": 3.3601975613123853,
      "tokens_seen": 2874605568
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6031134649333975e-05,
      "loss": 1.2645,
      "theoretical_loss": 3.3601857798334445,
      "tokens_seen": 2874736640
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6023110255175735e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.3601739990420603,
      "tokens_seen": 2874867712
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6015085861017495e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.360162218938162,
      "tokens_seen": 2874998784
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.6007061466859255e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.3601504395216786,
      "tokens_seen": 2875129856
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.599903707270101e-05,
      "loss": 1.3545,
      "theoretical_loss": 3.360138660792538,
      "tokens_seen": 2875260928
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.599101267854277e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.360126882750669,
      "tokens_seen": 2875392000
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.9937359094619751,
      "objective/train/avg_token_score": 0.0062640272080898285,
      "objective/train/docs_used": 1045259,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1488901376724243,
      "objective/train/original_loss": 1.1573392152786255,
      "objective/train/theoretical_loss": 3.360115105396001,
      "objective/train/tokens_used": 1245982176,
      "objective/train/unlikelihood_loss": 0.01799384132027626,
      "theoretical_loss": 3.360115105396001,
      "tokens_seen": 2875523072
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5982988284384528e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.360115105396001,
      "tokens_seen": 2875523072
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5974963890226288e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.3601033287284605,
      "tokens_seen": 2875654144
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5966939496068048e-05,
      "loss": 1.306,
      "theoretical_loss": 3.3600915527479778,
      "tokens_seen": 2875785216
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5958915101909808e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.360079777454481,
      "tokens_seen": 2875916288
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5950890707751568e-05,
      "loss": 1.2955,
      "theoretical_loss": 3.3600680028478993,
      "tokens_seen": 2876047360
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5942866313593324e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.36005622892816,
      "tokens_seen": 2876178432
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.593484191943508e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.360044455695193,
      "tokens_seen": 2876309504
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.592681752527684e-05,
      "loss": 1.2941,
      "theoretical_loss": 3.3600326831489267,
      "tokens_seen": 2876440576
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.59187931311186e-05,
      "loss": 1.334,
      "theoretical_loss": 3.360020911289289,
      "tokens_seen": 2876571648
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.591076873696036e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3600091401162095,
      "tokens_seen": 2876702720
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.590274434280212e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.3599973696296166,
      "tokens_seen": 2876833792
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.589471994864388e-05,
      "loss": 1.293,
      "theoretical_loss": 3.3599855998294386,
      "tokens_seen": 2876964864
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5886695554485633e-05,
      "loss": 1.3615,
      "theoretical_loss": 3.3599738307156044,
      "tokens_seen": 2877095936
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5878671160327393e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.3599620622880426,
      "tokens_seen": 2877227008
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5870646766169153e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.3599502945466826,
      "tokens_seen": 2877358080
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5862622372010913e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3599385274914524,
      "tokens_seen": 2877489152
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5854597977852673e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.359926761122281,
      "tokens_seen": 2877620224
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5846573583694433e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.359914995439097,
      "tokens_seen": 2877751296
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5838549189536193e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.3599032304418297,
      "tokens_seen": 2877882368
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5830524795377953e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.3598914661304073,
      "tokens_seen": 2878013440
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5822500401219706e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.3598797025047586,
      "tokens_seen": 2878144512
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5814476007061466e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.359867939564813,
      "tokens_seen": 2878275584
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5806451612903226e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3598561773104985,
      "tokens_seen": 2878406656
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5798427218744985e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.3598444157417444,
      "tokens_seen": 2878537728
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5790402824586745e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.3598326548584794,
      "tokens_seen": 2878668800
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5782378430428505e-05,
      "loss": 1.3129,
      "theoretical_loss": 3.3598208946606323,
      "tokens_seen": 2878799872
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5774354036270265e-05,
      "loss": 1.2103,
      "theoretical_loss": 3.359809135148132,
      "tokens_seen": 2878930944
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5766329642112018e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.3597973763209072,
      "tokens_seen": 2879062016
    },
    {
      "epoch": 0.74,
      "learning_rate": 2.5758305247953778e-05,
      "loss": 1.3127,
      "theoretical_loss": 3.359785618178887,
      "tokens_seen": 2879193088
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5750280853795538e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.359773860722001,
      "tokens_seen": 2879324160
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5742256459637298e-05,
      "loss": 1.3325,
      "theoretical_loss": 3.3597621039501764,
      "tokens_seen": 2879455232
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5734232065479058e-05,
      "loss": 1.276,
      "theoretical_loss": 3.3597503478633435,
      "tokens_seen": 2879586304
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5726207671320818e-05,
      "loss": 1.218,
      "theoretical_loss": 3.3597385924614303,
      "tokens_seen": 2879717376
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5718183277162578e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.3597268377443665,
      "tokens_seen": 2879848448
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5710158883004338e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.3597150837120804,
      "tokens_seen": 2879979520
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.570213448884609e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3597033303645016,
      "tokens_seen": 2880110592
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.569411009468785e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.3596915777015584,
      "tokens_seen": 2880241664
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.568608570052961e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.3596798257231804,
      "tokens_seen": 2880372736
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.567806130637137e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.3596680744292957,
      "tokens_seen": 2880503808
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.567003691221313e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.3596563238198343,
      "tokens_seen": 2880634880
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.566201251805489e-05,
      "loss": 1.2309,
      "theoretical_loss": 3.359644573894725,
      "tokens_seen": 2880765952
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.565398812389665e-05,
      "loss": 1.2567,
      "theoretical_loss": 3.359632824653896,
      "tokens_seen": 2880897024
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5645963729738403e-05,
      "loss": 1.224,
      "theoretical_loss": 3.3596210760972767,
      "tokens_seen": 2881028096
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5637939335580163e-05,
      "loss": 1.2869,
      "theoretical_loss": 3.3596093282247965,
      "tokens_seen": 2881159168
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5629914941421923e-05,
      "loss": 1.325,
      "theoretical_loss": 3.3595975810363843,
      "tokens_seen": 2881290240
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5621890547263683e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.359585834531969,
      "tokens_seen": 2881421312
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5613866153105443e-05,
      "loss": 1.2842,
      "theoretical_loss": 3.3595740887114798,
      "tokens_seen": 2881552384
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5605841758947203e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.3595623435748454,
      "tokens_seen": 2881683456
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5597817364788963e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.3595505991219956,
      "tokens_seen": 2881814528
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5589792970630716e-05,
      "loss": 1.288,
      "theoretical_loss": 3.359538855352859,
      "tokens_seen": 2881945600
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.9919629693031311,
      "objective/train/avg_token_score": 0.008036998100578785,
      "objective/train/docs_used": 1047629,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2653743028640747,
      "objective/train/original_loss": 1.2768603563308716,
      "objective/train/theoretical_loss": 3.3595271122673642,
      "objective/train/tokens_used": 1252535776,
      "objective/train/unlikelihood_loss": 0.021995466202497482,
      "theoretical_loss": 3.3595271122673642,
      "tokens_seen": 2882076672
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5581768576472476e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.3595271122673642,
      "tokens_seen": 2882076672
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5573744182314236e-05,
      "loss": 1.2495,
      "theoretical_loss": 3.359515369865442,
      "tokens_seen": 2882207744
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5565719788155996e-05,
      "loss": 1.3434,
      "theoretical_loss": 3.3595036281470194,
      "tokens_seen": 2882338816
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5557695393997755e-05,
      "loss": 1.1864,
      "theoretical_loss": 3.359491887112027,
      "tokens_seen": 2882469888
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5549670999839515e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.3594801467603936,
      "tokens_seen": 2882600960
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5541646605681275e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.359468407092048,
      "tokens_seen": 2882732032
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5533622211523035e-05,
      "loss": 1.3396,
      "theoretical_loss": 3.35945666810692,
      "tokens_seen": 2882863104
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.552559781736479e-05,
      "loss": 1.2995,
      "theoretical_loss": 3.359444929804938,
      "tokens_seen": 2882994176
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5517573423206548e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.359433192186032,
      "tokens_seen": 2883125248
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5509549029048308e-05,
      "loss": 1.305,
      "theoretical_loss": 3.3594214552501307,
      "tokens_seen": 2883256320
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5501524634890068e-05,
      "loss": 1.2221,
      "theoretical_loss": 3.359409718997163,
      "tokens_seen": 2883387392
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5493500240731828e-05,
      "loss": 1.3621,
      "theoretical_loss": 3.359397983427059,
      "tokens_seen": 2883518464
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5485475846573588e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.359386248539747,
      "tokens_seen": 2883649536
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5477451452415348e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.3593745143351574,
      "tokens_seen": 2883780608
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.54694270582571e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.359362780813218,
      "tokens_seen": 2883911680
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.546140266409886e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.3593510479738597,
      "tokens_seen": 2884042752
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.545337826994062e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.35933931581701,
      "tokens_seen": 2884173824
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.544535387578238e-05,
      "loss": 1.311,
      "theoretical_loss": 3.3593275843425996,
      "tokens_seen": 2884304896
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.543732948162414e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.359315853550557,
      "tokens_seen": 2884435968
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.54293050874659e-05,
      "loss": 1.2447,
      "theoretical_loss": 3.359304123440812,
      "tokens_seen": 2884567040
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5421280693307657e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.3592923940132935,
      "tokens_seen": 2884698112
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5413256299149413e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.359280665267931,
      "tokens_seen": 2884829184
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5405231904991173e-05,
      "loss": 1.2547,
      "theoretical_loss": 3.359268937204654,
      "tokens_seen": 2884960256
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5397207510832933e-05,
      "loss": 1.1757,
      "theoretical_loss": 3.359257209823392,
      "tokens_seen": 2885091328
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5389183116674693e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3592454831240732,
      "tokens_seen": 2885222400
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5381158722516453e-05,
      "loss": 1.3566,
      "theoretical_loss": 3.3592337571066286,
      "tokens_seen": 2885353472
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.537313432835821e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.3592220317709867,
      "tokens_seen": 2885484544
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.536510993419997e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.3592103071170767,
      "tokens_seen": 2885615616
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.535708554004173e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.359198583144828,
      "tokens_seen": 2885746688
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5349061145883486e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.359186859854171,
      "tokens_seen": 2885877760
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5341036751725246e-05,
      "loss": 1.2077,
      "theoretical_loss": 3.359175137245034,
      "tokens_seen": 2886008832
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5333012357567006e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.359163415317347,
      "tokens_seen": 2886139904
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5324987963408762e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.3591516940710395,
      "tokens_seen": 2886270976
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5316963569250522e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.35913997350604,
      "tokens_seen": 2886402048
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5308939175092282e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.3591282536222797,
      "tokens_seen": 2886533120
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5300914780934042e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3591165344196865,
      "tokens_seen": 2886664192
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.52928903867758e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.3591048158981907,
      "tokens_seen": 2886795264
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.528486599261756e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.359093098057721,
      "tokens_seen": 2886926336
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5276841598459315e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.359081380898208,
      "tokens_seen": 2887057408
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5268817204301075e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.3590696644195805,
      "tokens_seen": 2887188480
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5260792810142835e-05,
      "loss": 1.2465,
      "theoretical_loss": 3.3590579486217687,
      "tokens_seen": 2887319552
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5252768415984595e-05,
      "loss": 1.294,
      "theoretical_loss": 3.359046233504701,
      "tokens_seen": 2887450624
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5244744021826355e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.3590345190683077,
      "tokens_seen": 2887581696
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5236719627668114e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.3590228053125184,
      "tokens_seen": 2887712768
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5228695233509868e-05,
      "loss": 1.2425,
      "theoretical_loss": 3.3590110922372625,
      "tokens_seen": 2887843840
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5220670839351628e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.3589993798424693,
      "tokens_seen": 2887974912
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5212646445193387e-05,
      "loss": 1.2365,
      "theoretical_loss": 3.3589876681280693,
      "tokens_seen": 2888105984
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5204622051035147e-05,
      "loss": 1.3851,
      "theoretical_loss": 3.358975957093991,
      "tokens_seen": 2888237056
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5196597656876907e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.3589642467401646,
      "tokens_seen": 2888368128
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5188573262718667e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.3589525370665196,
      "tokens_seen": 2888499200
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.9933603405952454,
      "objective/train/avg_token_score": 0.006639623083174229,
      "objective/train/docs_used": 1050131,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3532625436782837,
      "objective/train/original_loss": 1.3631024360656738,
      "objective/train/theoretical_loss": 3.3589408280729858,
      "objective/train/tokens_used": 1259089376,
      "objective/train/unlikelihood_loss": 0.016430485993623734,
      "theoretical_loss": 3.3589408280729858,
      "tokens_seen": 2888630272
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5180548868560427e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.3589408280729858,
      "tokens_seen": 2888630272
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.517252447440218e-05,
      "loss": 1.3429,
      "theoretical_loss": 3.3589291197594924,
      "tokens_seen": 2888761344
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.516450008024394e-05,
      "loss": 1.3467,
      "theoretical_loss": 3.3589174121259693,
      "tokens_seen": 2888892416
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.51564756860857e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.3589057051723463,
      "tokens_seen": 2889023488
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.514845129192746e-05,
      "loss": 1.3075,
      "theoretical_loss": 3.3588939988985533,
      "tokens_seen": 2889154560
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.514042689776922e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.3588822933045193,
      "tokens_seen": 2889285632
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.513240250361098e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.3588705883901744,
      "tokens_seen": 2889416704
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.512437810945274e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.358858884155448,
      "tokens_seen": 2889547776
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5116353715294493e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.3588471806002707,
      "tokens_seen": 2889678848
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5108329321136253e-05,
      "loss": 1.27,
      "theoretical_loss": 3.358835477724571,
      "tokens_seen": 2889809920
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5100304926978013e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.3588237755282795,
      "tokens_seen": 2889940992
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5092280532819772e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.3588120740113254,
      "tokens_seen": 2890072064
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5084256138661532e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.358800373173639,
      "tokens_seen": 2890203136
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5076231744503292e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.35878867301515,
      "tokens_seen": 2890334208
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5068207350345052e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.3587769735357873,
      "tokens_seen": 2890465280
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5060182956186812e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.3587652747354815,
      "tokens_seen": 2890596352
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5052158562028565e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.3587535766141627,
      "tokens_seen": 2890727424
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5044134167870325e-05,
      "loss": 1.1978,
      "theoretical_loss": 3.35874187917176,
      "tokens_seen": 2890858496
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5036109773712085e-05,
      "loss": 1.339,
      "theoretical_loss": 3.3587301824082036,
      "tokens_seen": 2890989568
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5028085379553845e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.358718486323423,
      "tokens_seen": 2891120640
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5020060985395605e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.3587067909173483,
      "tokens_seen": 2891251712
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5012036591237365e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.3586950961899094,
      "tokens_seen": 2891382784
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.5004012197079125e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.3586834021410357,
      "tokens_seen": 2891513856
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.499598780292088e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.3586717087706575,
      "tokens_seen": 2891644928
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.498796340876264e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.358660016078705,
      "tokens_seen": 2891776000
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4979939014604398e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.3586483240651077,
      "tokens_seen": 2891907072
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4971914620446157e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.3586366327297954,
      "tokens_seen": 2892038144
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4963890226287917e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.358624942072698,
      "tokens_seen": 2892169216
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4955865832129677e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.358613252093745,
      "tokens_seen": 2892300288
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4947841437971434e-05,
      "loss": 1.2829,
      "theoretical_loss": 3.3586015627928676,
      "tokens_seen": 2892431360
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4939817043813194e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.358589874169995,
      "tokens_seen": 2892562432
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4931792649654954e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.3585781862250568,
      "tokens_seen": 2892693504
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.492376825549671e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.3585664989579835,
      "tokens_seen": 2892824576
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.491574386133847e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.358554812368705,
      "tokens_seen": 2892955648
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.490771946718023e-05,
      "loss": 1.2014,
      "theoretical_loss": 3.358543126457151,
      "tokens_seen": 2893086720
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.489969507302199e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.358531441223252,
      "tokens_seen": 2893217792
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4891670678863746e-05,
      "loss": 1.3357,
      "theoretical_loss": 3.3585197566669374,
      "tokens_seen": 2893348864
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4883646284705506e-05,
      "loss": 1.3278,
      "theoretical_loss": 3.358508072788138,
      "tokens_seen": 2893479936
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4875621890547266e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3584963895867825,
      "tokens_seen": 2893611008
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4867597496389026e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.3584847070628023,
      "tokens_seen": 2893742080
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4859573102230783e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.358473025216127,
      "tokens_seen": 2893873152
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4851548708072543e-05,
      "loss": 1.2234,
      "theoretical_loss": 3.3584613440466864,
      "tokens_seen": 2894004224
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4843524313914302e-05,
      "loss": 1.2376,
      "theoretical_loss": 3.3584496635544108,
      "tokens_seen": 2894135296
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.483549991975606e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.35843798373923,
      "tokens_seen": 2894266368
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.482747552559782e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.3584263046010747,
      "tokens_seen": 2894397440
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.481945113143958e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.3584146261398744,
      "tokens_seen": 2894528512
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.481142673728134e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.3584029483555593,
      "tokens_seen": 2894659584
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4803402343123095e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.3583912712480597,
      "tokens_seen": 2894790656
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4795377948964855e-05,
      "loss": 1.3289,
      "theoretical_loss": 3.358379594817306,
      "tokens_seen": 2894921728
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4787353554806615e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.3583679190632276,
      "tokens_seen": 2895052800
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.9946760535240173,
      "objective/train/avg_token_score": 0.005323933903127909,
      "objective/train/docs_used": 1052611,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2621833086013794,
      "objective/train/original_loss": 1.2697293758392334,
      "objective/train/theoretical_loss": 3.3583562439857557,
      "objective/train/tokens_used": 1265642976,
      "objective/train/unlikelihood_loss": 0.014400391839444637,
      "theoretical_loss": 3.3583562439857557,
      "tokens_seen": 2895183872
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.4779329160648375e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.3583562439857557,
      "tokens_seen": 2895183872
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.477130476649013e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.358344569584819,
      "tokens_seen": 2895314944
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.476328037233189e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.3583328958603493,
      "tokens_seen": 2895446016
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.475525597817365e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.3583212228122754,
      "tokens_seen": 2895577088
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4747231584015408e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.3583095504405285,
      "tokens_seen": 2895708160
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4739207189857168e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.358297878745038,
      "tokens_seen": 2895839232
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4731182795698928e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.358286207725735,
      "tokens_seen": 2895970304
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4723158401540687e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.358274537382549,
      "tokens_seen": 2896101376
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4715134007382444e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.35826286771541,
      "tokens_seen": 2896232448
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4707109613224204e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.358251198724249,
      "tokens_seen": 2896363520
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4699085219065964e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.358239530408996,
      "tokens_seen": 2896494592
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.469106082490772e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.3582278627695814,
      "tokens_seen": 2896625664
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.468303643074948e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.358216195805935,
      "tokens_seen": 2896756736
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.467501203659124e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.358204529517988,
      "tokens_seen": 2896887808
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4666987642432997e-05,
      "loss": 1.211,
      "theoretical_loss": 3.3581928639056695,
      "tokens_seen": 2897018880
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4658963248274757e-05,
      "loss": 1.325,
      "theoretical_loss": 3.3581811989689103,
      "tokens_seen": 2897149952
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4650938854116516e-05,
      "loss": 1.2113,
      "theoretical_loss": 3.358169534707641,
      "tokens_seen": 2897281024
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4642914459958273e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.358157871121792,
      "tokens_seen": 2897412096
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4634890065800033e-05,
      "loss": 1.2308,
      "theoretical_loss": 3.3581462082112927,
      "tokens_seen": 2897543168
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4626865671641793e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.3581345459760747,
      "tokens_seen": 2897674240
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.461884127748355e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.3581228844160678,
      "tokens_seen": 2897805312
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.461081688332531e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.358111223531202,
      "tokens_seen": 2897936384
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.460279248916707e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.3580995633214084,
      "tokens_seen": 2898067456
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4594768095008826e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.3580879037866165,
      "tokens_seen": 2898198528
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4586743700850586e-05,
      "loss": 1.3358,
      "theoretical_loss": 3.3580762449267576,
      "tokens_seen": 2898329600
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4578719306692345e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.3580645867417616,
      "tokens_seen": 2898460672
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4570694912534102e-05,
      "loss": 1.2407,
      "theoretical_loss": 3.358052929231559,
      "tokens_seen": 2898591744
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4562670518375862e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.358041272396081,
      "tokens_seen": 2898722816
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4554646124217622e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.3580296162352563,
      "tokens_seen": 2898853888
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.454662173005938e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.3580179607490166,
      "tokens_seen": 2898984960
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4538597335901138e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.358006305937293,
      "tokens_seen": 2899116032
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4530572941742898e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.357994651800014,
      "tokens_seen": 2899247104
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4522548547584658e-05,
      "loss": 1.1931,
      "theoretical_loss": 3.3579829983371114,
      "tokens_seen": 2899378176
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4514524153426418e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.357971345548516,
      "tokens_seen": 2899509248
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4506499759268174e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.3579596934341573,
      "tokens_seen": 2899640320
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4498475365109934e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.357948041993967,
      "tokens_seen": 2899771392
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4490450970951694e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.357936391227874,
      "tokens_seen": 2899902464
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4482426576793454e-05,
      "loss": 1.3518,
      "theoretical_loss": 3.35792474113581,
      "tokens_seen": 2900033536
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.447440218263521e-05,
      "loss": 1.263,
      "theoretical_loss": 3.357913091717706,
      "tokens_seen": 2900164608
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.446637778847697e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.357901442973491,
      "tokens_seen": 2900295680
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.445835339431873e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.357889794903097,
      "tokens_seen": 2900426752
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4450329000160487e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.3578781475064536,
      "tokens_seen": 2900557824
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4442304606002247e-05,
      "loss": 1.309,
      "theoretical_loss": 3.3578665007834916,
      "tokens_seen": 2900688896
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4434280211844007e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.3578548547341422,
      "tokens_seen": 2900819968
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4426255817685767e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3578432093583355,
      "tokens_seen": 2900951040
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4418231423527523e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3578315646560024,
      "tokens_seen": 2901082112
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4410207029369283e-05,
      "loss": 1.1948,
      "theoretical_loss": 3.357819920627073,
      "tokens_seen": 2901213184
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4402182635211043e-05,
      "loss": 1.3411,
      "theoretical_loss": 3.357808277271478,
      "tokens_seen": 2901344256
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4394158241052803e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.357796634589149,
      "tokens_seen": 2901475328
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.438613384689456e-05,
      "loss": 1.2534,
      "theoretical_loss": 3.3577849925800156,
      "tokens_seen": 2901606400
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.9908344745635986,
      "objective/train/avg_token_score": 0.00916551798582077,
      "objective/train/docs_used": 1055165,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3563495874404907,
      "objective/train/original_loss": 1.3700411319732666,
      "objective/train/theoretical_loss": 3.357773351244009,
      "objective/train/tokens_used": 1272196576,
      "objective/train/unlikelihood_loss": 0.024363506585359573,
      "theoretical_loss": 3.357773351244009,
      "tokens_seen": 2901737472
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.437810945273632e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.357773351244009,
      "tokens_seen": 2901737472
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.437008505857808e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.3577617105810598,
      "tokens_seen": 2901868544
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4362060664419836e-05,
      "loss": 1.3861,
      "theoretical_loss": 3.357750070591098,
      "tokens_seen": 2901999616
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4354036270261596e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.3577384312740555,
      "tokens_seen": 2902130688
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4346011876103356e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.357726792629862,
      "tokens_seen": 2902261760
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4337987481945116e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.357715154658449,
      "tokens_seen": 2902392832
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4329963087786872e-05,
      "loss": 1.3573,
      "theoretical_loss": 3.357703517359747,
      "tokens_seen": 2902523904
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4321938693628632e-05,
      "loss": 1.287,
      "theoretical_loss": 3.357691880733686,
      "tokens_seen": 2902654976
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4313914299470392e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.3576802447801977,
      "tokens_seen": 2902786048
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4305889905312152e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.3576686094992123,
      "tokens_seen": 2902917120
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4297865511153908e-05,
      "loss": 1.3183,
      "theoretical_loss": 3.357656974890661,
      "tokens_seen": 2903048192
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4289841116995668e-05,
      "loss": 1.3066,
      "theoretical_loss": 3.357645340954474,
      "tokens_seen": 2903179264
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4281816722837428e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.357633707690583,
      "tokens_seen": 2903310336
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4273792328679185e-05,
      "loss": 1.2002,
      "theoretical_loss": 3.357622075098918,
      "tokens_seen": 2903441408
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4265767934520945e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.35761044317941,
      "tokens_seen": 2903572480
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4257743540362704e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.35759881193199,
      "tokens_seen": 2903703552
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4249719146204464e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.3575871813565885,
      "tokens_seen": 2903834624
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.424169475204622e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3575755514531367,
      "tokens_seen": 2903965696
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.423367035788798e-05,
      "loss": 1.315,
      "theoretical_loss": 3.3575639222215656,
      "tokens_seen": 2904096768
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.422564596372974e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.357552293661805,
      "tokens_seen": 2904227840
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.42176215695715e-05,
      "loss": 1.2566,
      "theoretical_loss": 3.3575406657737874,
      "tokens_seen": 2904358912
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4209597175413257e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.357529038557442,
      "tokens_seen": 2904489984
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4201572781255017e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.3575174120127014,
      "tokens_seen": 2904621056
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4193548387096777e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.357505786139495,
      "tokens_seen": 2904752128
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4185523992938533e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.3574941609377547,
      "tokens_seen": 2904883200
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4177499598780293e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.357482536407411,
      "tokens_seen": 2905014272
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4169475204622053e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.357470912548395,
      "tokens_seen": 2905145344
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4161450810463813e-05,
      "loss": 1.28,
      "theoretical_loss": 3.357459289360637,
      "tokens_seen": 2905276416
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.415342641630557e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.357447666844069,
      "tokens_seen": 2905407488
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.414540202214733e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.357436044998621,
      "tokens_seen": 2905538560
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.413737762798909e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.357424423824225,
      "tokens_seen": 2905669632
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.412935323383085e-05,
      "loss": 1.2988,
      "theoretical_loss": 3.357412803320811,
      "tokens_seen": 2905800704
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4121328839672606e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.3574011834883106,
      "tokens_seen": 2905931776
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4113304445514366e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.3573895643266547,
      "tokens_seen": 2906062848
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4105280051356126e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.357377945835774,
      "tokens_seen": 2906193920
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4097255657197882e-05,
      "loss": 1.2538,
      "theoretical_loss": 3.3573663280156,
      "tokens_seen": 2906324992
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4089231263039642e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.3573547108660633,
      "tokens_seen": 2906456064
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4081206868881402e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.357343094387095,
      "tokens_seen": 2906587136
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4073182474723162e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.357331478578627,
      "tokens_seen": 2906718208
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.406515808056492e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.3573198634405887,
      "tokens_seen": 2906849280
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.405713368640668e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.357308248972913,
      "tokens_seen": 2906980352
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4049109292248438e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.35729663517553,
      "tokens_seen": 2907111424
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4041084898090198e-05,
      "loss": 1.2577,
      "theoretical_loss": 3.3572850220483703,
      "tokens_seen": 2907242496
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4033060503931955e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.3572734095913663,
      "tokens_seen": 2907373568
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4025036109773715e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.357261797804448,
      "tokens_seen": 2907504640
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.4017011715615474e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.3572501866875477,
      "tokens_seen": 2907635712
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.400898732145723e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.357238576240595,
      "tokens_seen": 2907766784
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.400096292729899e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.3572269664635224,
      "tokens_seen": 2907897856
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.399293853314075e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.35721535735626,
      "tokens_seen": 2908028928
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3984914138982507e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.35720374891874,
      "tokens_seen": 2908160000
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.9926728010177612,
      "objective/train/avg_token_score": 0.007327239494770765,
      "objective/train/docs_used": 1057433,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3090240955352783,
      "objective/train/original_loss": 1.3184435367584229,
      "objective/train/theoretical_loss": 3.357192141150893,
      "objective/train/tokens_used": 1278750176,
      "objective/train/unlikelihood_loss": 0.02221887931227684,
      "theoretical_loss": 3.357192141150893,
      "tokens_seen": 2908291072
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3976889744824267e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.357192141150893,
      "tokens_seen": 2908291072
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3968865350666027e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.35718053405265,
      "tokens_seen": 2908422144
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3960840956507784e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.3571689276239427,
      "tokens_seen": 2908553216
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3952816562349544e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.357157321864702,
      "tokens_seen": 2908684288
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.39447921681913e-05,
      "loss": 1.2282,
      "theoretical_loss": 3.3571457167748595,
      "tokens_seen": 2908815360
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.393676777403306e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.3571341123543457,
      "tokens_seen": 2908946432
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.392874337987482e-05,
      "loss": 1.1968,
      "theoretical_loss": 3.3571225086030925,
      "tokens_seen": 2909077504
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.392071898571658e-05,
      "loss": 1.331,
      "theoretical_loss": 3.357110905521031,
      "tokens_seen": 2909208576
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3912694591558336e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.3570993031080922,
      "tokens_seen": 2909339648
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3904670197400096e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.3570877013642075,
      "tokens_seen": 2909470720
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3896645803241856e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.357076100289308,
      "tokens_seen": 2909601792
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3888621409083613e-05,
      "loss": 1.3283,
      "theoretical_loss": 3.3570644998833257,
      "tokens_seen": 2909732864
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3880597014925373e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.3570529001461913,
      "tokens_seen": 2909863936
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3872572620767132e-05,
      "loss": 1.2285,
      "theoretical_loss": 3.357041301077836,
      "tokens_seen": 2909995008
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3864548226608892e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.357029702678192,
      "tokens_seen": 2910126080
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.385652383245065e-05,
      "loss": 1.2023,
      "theoretical_loss": 3.357018104947189,
      "tokens_seen": 2910257152
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.384849943829241e-05,
      "loss": 1.3861,
      "theoretical_loss": 3.35700650788476,
      "tokens_seen": 2910388224
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.384047504413417e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.3569949114908355,
      "tokens_seen": 2910519296
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.383245064997593e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.356983315765347,
      "tokens_seen": 2910650368
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3824426255817685e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.356971720708226,
      "tokens_seen": 2910781440
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3816401861659445e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.3569601263194038,
      "tokens_seen": 2910912512
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3808377467501205e-05,
      "loss": 1.3438,
      "theoretical_loss": 3.3569485325988113,
      "tokens_seen": 2911043584
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.380035307334296e-05,
      "loss": 1.33,
      "theoretical_loss": 3.3569369395463813,
      "tokens_seen": 2911174656
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.379232867918472e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.3569253471620435,
      "tokens_seen": 2911305728
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.378430428502648e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.3569137554457305,
      "tokens_seen": 2911436800
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.377627989086824e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.3569021643973733,
      "tokens_seen": 2911567872
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3768255496709998e-05,
      "loss": 1.2209,
      "theoretical_loss": 3.3568905740169033,
      "tokens_seen": 2911698944
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3760231102551758e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3568789843042524,
      "tokens_seen": 2911830016
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3752206708393518e-05,
      "loss": 1.2611,
      "theoretical_loss": 3.3568673952593513,
      "tokens_seen": 2911961088
    },
    {
      "epoch": 0.76,
      "learning_rate": 2.3744182314235277e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.356855806882132,
      "tokens_seen": 2912092160
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3736157920077034e-05,
      "loss": 1.3822,
      "theoretical_loss": 3.3568442191725256,
      "tokens_seen": 2912223232
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3728133525918794e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.3568326321304642,
      "tokens_seen": 2912354304
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3720109131760554e-05,
      "loss": 1.297,
      "theoretical_loss": 3.356821045755879,
      "tokens_seen": 2912485376
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.371208473760231e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3568094600487015,
      "tokens_seen": 2912616448
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.370406034344407e-05,
      "loss": 1.2434,
      "theoretical_loss": 3.356797875008863,
      "tokens_seen": 2912747520
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.369603594928583e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.3567862906362955,
      "tokens_seen": 2912878592
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.368801155512759e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.35677470693093,
      "tokens_seen": 2913009664
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3679987160969346e-05,
      "loss": 1.195,
      "theoretical_loss": 3.3567631238926983,
      "tokens_seen": 2913140736
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3671962766811106e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.3567515415215325,
      "tokens_seen": 2913271808
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3663938372652866e-05,
      "loss": 1.3092,
      "theoretical_loss": 3.356739959817363,
      "tokens_seen": 2913402880
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3655913978494626e-05,
      "loss": 1.2473,
      "theoretical_loss": 3.3567283787801223,
      "tokens_seen": 2913533952
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3647889584336383e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.356716798409742,
      "tokens_seen": 2913665024
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3639865190178143e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3567052187061535,
      "tokens_seen": 2913796096
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3631840796019903e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.356693639669288,
      "tokens_seen": 2913927168
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.362381640186166e-05,
      "loss": 1.2349,
      "theoretical_loss": 3.3566820612990775,
      "tokens_seen": 2914058240
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.361579200770342e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.3566704835954537,
      "tokens_seen": 2914189312
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.360776761354518e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3566589065583483,
      "tokens_seen": 2914320384
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.359974321938694e-05,
      "loss": 1.279,
      "theoretical_loss": 3.356647330187693,
      "tokens_seen": 2914451456
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3591718825228695e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.356635754483419,
      "tokens_seen": 2914582528
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3583694431070455e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.3566241794454585,
      "tokens_seen": 2914713600
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.9948214888572693,
      "objective/train/avg_token_score": 0.005178466439247131,
      "objective/train/docs_used": 1059748,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3445072174072266,
      "objective/train/original_loss": 1.3519647121429443,
      "objective/train/theoretical_loss": 3.3566126050737424,
      "objective/train/tokens_used": 1285303776,
      "objective/train/unlikelihood_loss": 0.01440105214715004,
      "theoretical_loss": 3.3566126050737424,
      "tokens_seen": 2914844672
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3575670036912215e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.3566126050737424,
      "tokens_seen": 2914844672
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3567645642753975e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.3566010313682035,
      "tokens_seen": 2914975744
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.355962124859573e-05,
      "loss": 1.2532,
      "theoretical_loss": 3.3565894583287728,
      "tokens_seen": 2915106816
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.355159685443749e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.356577885955382,
      "tokens_seen": 2915237888
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.354357246027925e-05,
      "loss": 1.3992,
      "theoretical_loss": 3.3565663142479636,
      "tokens_seen": 2915368960
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.353554806612101e-05,
      "loss": 1.308,
      "theoretical_loss": 3.3565547432064484,
      "tokens_seen": 2915500032
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3527523671962768e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.356543172830768,
      "tokens_seen": 2915631104
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3519499277804528e-05,
      "loss": 1.279,
      "theoretical_loss": 3.356531603120855,
      "tokens_seen": 2915762176
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3511474883646288e-05,
      "loss": 1.3352,
      "theoretical_loss": 3.356520034076641,
      "tokens_seen": 2915893248
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3503450489488044e-05,
      "loss": 1.2106,
      "theoretical_loss": 3.3565084656980573,
      "tokens_seen": 2916024320
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3495426095329804e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.356496897985036,
      "tokens_seen": 2916155392
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3487401701171564e-05,
      "loss": 1.2828,
      "theoretical_loss": 3.356485330937509,
      "tokens_seen": 2916286464
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3479377307013324e-05,
      "loss": 1.2219,
      "theoretical_loss": 3.356473764555408,
      "tokens_seen": 2916417536
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.347135291285508e-05,
      "loss": 1.2365,
      "theoretical_loss": 3.3564621988386647,
      "tokens_seen": 2916548608
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.346332851869684e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.356450633787211,
      "tokens_seen": 2916679680
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.34553041245386e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.3564390694009787,
      "tokens_seen": 2916810752
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.344727973038036e-05,
      "loss": 1.235,
      "theoretical_loss": 3.3564275056799,
      "tokens_seen": 2916941824
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3439255336222117e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.3564159426239057,
      "tokens_seen": 2917072896
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3431230942063876e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.356404380232929,
      "tokens_seen": 2917203968
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3423206547905636e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.356392818506901,
      "tokens_seen": 2917335040
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3415182153747393e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.3563812574457534,
      "tokens_seen": 2917466112
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3407157759589153e-05,
      "loss": 1.315,
      "theoretical_loss": 3.356369697049419,
      "tokens_seen": 2917597184
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3399133365430913e-05,
      "loss": 1.2542,
      "theoretical_loss": 3.3563581373178293,
      "tokens_seen": 2917728256
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3391108971272673e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.356346578250916,
      "tokens_seen": 2917859328
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.338308457711443e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.356335019848611,
      "tokens_seen": 2917990400
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.337506018295619e-05,
      "loss": 1.3295,
      "theoretical_loss": 3.356323462110846,
      "tokens_seen": 2918121472
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.336703578879795e-05,
      "loss": 1.3257,
      "theoretical_loss": 3.3563119050375536,
      "tokens_seen": 2918252544
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3359011394639705e-05,
      "loss": 1.2486,
      "theoretical_loss": 3.3563003486286656,
      "tokens_seen": 2918383616
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3350987000481465e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.3562887928841136,
      "tokens_seen": 2918514688
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3342962606323225e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.35627723780383,
      "tokens_seen": 2918645760
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3334938212164982e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3562656833877464,
      "tokens_seen": 2918776832
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3326913818006742e-05,
      "loss": 1.3422,
      "theoretical_loss": 3.3562541296357953,
      "tokens_seen": 2918907904
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.33188894238485e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.356242576547908,
      "tokens_seen": 2919038976
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3310865029690258e-05,
      "loss": 1.289,
      "theoretical_loss": 3.3562310241240167,
      "tokens_seen": 2919170048
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3302840635532018e-05,
      "loss": 1.3462,
      "theoretical_loss": 3.3562194723640544,
      "tokens_seen": 2919301120
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3294816241373778e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.356207921267952,
      "tokens_seen": 2919432192
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3286791847215534e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.356196370835642,
      "tokens_seen": 2919563264
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3278767453057294e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.356184821067056,
      "tokens_seen": 2919694336
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3270743058899054e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.3561732719621267,
      "tokens_seen": 2919825408
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.326271866474081e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.356161723520786,
      "tokens_seen": 2919956480
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.325469427058257e-05,
      "loss": 1.2751,
      "theoretical_loss": 3.356150175742966,
      "tokens_seen": 2920087552
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.324666987642433e-05,
      "loss": 1.318,
      "theoretical_loss": 3.3561386286285986,
      "tokens_seen": 2920218624
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3238645482266087e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.3561270821776157,
      "tokens_seen": 2920349696
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3230621088107847e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.35611553638995,
      "tokens_seen": 2920480768
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3222596693949607e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.3561039912655333,
      "tokens_seen": 2920611840
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3214572299791367e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.356092446804298,
      "tokens_seen": 2920742912
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3206547905633123e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.3560809030061756,
      "tokens_seen": 2920873984
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3198523511474883e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.356069359871099,
      "tokens_seen": 2921005056
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3190499117316643e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.356057817399,
      "tokens_seen": 2921136128
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3182474723158403e-05,
      "loss": 1.2319,
      "theoretical_loss": 3.3560462755898106,
      "tokens_seen": 2921267200
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.9905661940574646,
      "objective/train/avg_token_score": 0.009433820843696594,
      "objective/train/docs_used": 1062024,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.41095769405365,
      "objective/train/original_loss": 1.4247554540634155,
      "objective/train/theoretical_loss": 3.3560347344434636,
      "objective/train/tokens_used": 1291857376,
      "objective/train/unlikelihood_loss": 0.024458011612296104,
      "theoretical_loss": 3.3560347344434636,
      "tokens_seen": 2921398272
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.317445032900016e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.3560347344434636,
      "tokens_seen": 2921398272
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.316642593484192e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.356023193959891,
      "tokens_seen": 2921529344
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.315840154068368e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.3560116541390244,
      "tokens_seen": 2921660416
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3150377146525436e-05,
      "loss": 1.2593,
      "theoretical_loss": 3.3560001149807963,
      "tokens_seen": 2921791488
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3142352752367196e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.3559885764851396,
      "tokens_seen": 2921922560
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3134328358208956e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.3559770386519854,
      "tokens_seen": 2922053632
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3126303964050716e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.3559655014812666,
      "tokens_seen": 2922184704
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3118279569892472e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.355953964972916,
      "tokens_seen": 2922315776
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3110255175734232e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.355942429126865,
      "tokens_seen": 2922446848
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3102230781575992e-05,
      "loss": 1.3493,
      "theoretical_loss": 3.355930893943046,
      "tokens_seen": 2922577920
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3094206387417752e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.3559193594213914,
      "tokens_seen": 2922708992
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.308618199325951e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.3559078255618338,
      "tokens_seen": 2922840064
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3078157599101268e-05,
      "loss": 1.289,
      "theoretical_loss": 3.3558962923643048,
      "tokens_seen": 2922971136
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3070133204943028e-05,
      "loss": 1.2633,
      "theoretical_loss": 3.3558847598287374,
      "tokens_seen": 2923102208
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3062108810784785e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.355873227955064,
      "tokens_seen": 2923233280
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3054084416626545e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.355861696743216,
      "tokens_seen": 2923364352
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3046060022468305e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.355850166193127,
      "tokens_seen": 2923495424
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.3038035628310064e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.3558386363047283,
      "tokens_seen": 2923626496
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.303001123415182e-05,
      "loss": 1.3512,
      "theoretical_loss": 3.355827107077953,
      "tokens_seen": 2923757568
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.302198683999358e-05,
      "loss": 1.2912,
      "theoretical_loss": 3.3558155785127326,
      "tokens_seen": 2923888640
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.301396244583534e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.3558040506090006,
      "tokens_seen": 2924019712
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.30059380516771e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.3557925233666888,
      "tokens_seen": 2924150784
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2997913657518857e-05,
      "loss": 1.3564,
      "theoretical_loss": 3.3557809967857293,
      "tokens_seen": 2924281856
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2989889263360617e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.355769470866055,
      "tokens_seen": 2924412928
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2981864869202377e-05,
      "loss": 1.4283,
      "theoretical_loss": 3.355757945607598,
      "tokens_seen": 2924544000
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2973840475044137e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.3557464210102914,
      "tokens_seen": 2924675072
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2965816080885893e-05,
      "loss": 1.3229,
      "theoretical_loss": 3.355734897074067,
      "tokens_seen": 2924806144
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2957791686727653e-05,
      "loss": 1.3582,
      "theoretical_loss": 3.3557233737988574,
      "tokens_seen": 2924937216
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2949767292569413e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.355711851184595,
      "tokens_seen": 2925068288
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.294174289841117e-05,
      "loss": 1.3443,
      "theoretical_loss": 3.3557003292312126,
      "tokens_seen": 2925199360
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.293371850425293e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.355688807938642,
      "tokens_seen": 2925330432
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.292569411009469e-05,
      "loss": 1.2486,
      "theoretical_loss": 3.3556772873068166,
      "tokens_seen": 2925461504
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.291766971593645e-05,
      "loss": 1.3439,
      "theoretical_loss": 3.3556657673356685,
      "tokens_seen": 2925592576
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2909645321778206e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.3556542480251297,
      "tokens_seen": 2925723648
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2901620927619966e-05,
      "loss": 1.364,
      "theoretical_loss": 3.3556427293751336,
      "tokens_seen": 2925854720
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2893596533461726e-05,
      "loss": 1.3022,
      "theoretical_loss": 3.355631211385612,
      "tokens_seen": 2925985792
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2885572139303486e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.355619694056498,
      "tokens_seen": 2926116864
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2877547745145242e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.3556081773877238,
      "tokens_seen": 2926247936
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2869523350987002e-05,
      "loss": 1.3285,
      "theoretical_loss": 3.3555966613792223,
      "tokens_seen": 2926379008
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2861498956828762e-05,
      "loss": 1.2185,
      "theoretical_loss": 3.3555851460309256,
      "tokens_seen": 2926510080
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.285347456267052e-05,
      "loss": 1.3503,
      "theoretical_loss": 3.3555736313427666,
      "tokens_seen": 2926641152
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.284545016851228e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.355562117314678,
      "tokens_seen": 2926772224
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.283742577435404e-05,
      "loss": 1.279,
      "theoretical_loss": 3.3555506039465923,
      "tokens_seen": 2926903296
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2829401380195798e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.355539091238442,
      "tokens_seen": 2927034368
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2821376986037555e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.3555275791901593,
      "tokens_seen": 2927165440
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2813352591879315e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.3555160678016778,
      "tokens_seen": 2927296512
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2805328197721075e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.35550455707293,
      "tokens_seen": 2927427584
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2797303803562835e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.3554930470038475,
      "tokens_seen": 2927558656
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.278927940940459e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.355481537594364,
      "tokens_seen": 2927689728
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.278125501524635e-05,
      "loss": 1.2337,
      "theoretical_loss": 3.3554700288444117,
      "tokens_seen": 2927820800
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.9944932460784912,
      "objective/train/avg_token_score": 0.005506803747266531,
      "objective/train/docs_used": 1064246,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3141322135925293,
      "objective/train/original_loss": 1.3220288753509521,
      "objective/train/theoretical_loss": 3.3554585207539236,
      "objective/train/tokens_used": 1298410976,
      "objective/train/unlikelihood_loss": 0.014275084249675274,
      "theoretical_loss": 3.3554585207539236,
      "tokens_seen": 2927951872
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.277323062108811e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.3554585207539236,
      "tokens_seen": 2927951872
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2765206226929867e-05,
      "loss": 1.2778,
      "theoretical_loss": 3.355447013322832,
      "tokens_seen": 2928082944
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2757181832771627e-05,
      "loss": 1.3076,
      "theoretical_loss": 3.3554355065510704,
      "tokens_seen": 2928214016
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2749157438613387e-05,
      "loss": 1.2012,
      "theoretical_loss": 3.3554240004385703,
      "tokens_seen": 2928345088
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2741133044455147e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.3554124949852655,
      "tokens_seen": 2928476160
    },
    {
      "epoch": 0.77,
      "learning_rate": 2.2733108650296904e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.3554009901910877,
      "tokens_seen": 2928607232
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2725084256138663e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.3553894860559708,
      "tokens_seen": 2928738304
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2717059861980423e-05,
      "loss": 1.3073,
      "theoretical_loss": 3.355377982579847,
      "tokens_seen": 2928869376
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2709035467822183e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.355366479762649,
      "tokens_seen": 2929000448
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.270101107366394e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.3553549776043097,
      "tokens_seen": 2929131520
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.26929866795057e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.3553434761047614,
      "tokens_seen": 2929262592
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.268496228534746e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.3553319752639377,
      "tokens_seen": 2929393664
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2676937891189216e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.3553204750817707,
      "tokens_seen": 2929524736
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2668913497030976e-05,
      "loss": 1.2192,
      "theoretical_loss": 3.3553089755581937,
      "tokens_seen": 2929655808
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2660889102872736e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.3552974766931394,
      "tokens_seen": 2929786880
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2652864708714492e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.35528597848654,
      "tokens_seen": 2929917952
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2644840314556252e-05,
      "loss": 1.2519,
      "theoretical_loss": 3.3552744809383297,
      "tokens_seen": 2930049024
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2636815920398012e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3552629840484403,
      "tokens_seen": 2930180096
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.262879152623977e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.3552514878168047,
      "tokens_seen": 2930311168
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.262076713208153e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.355239992243356,
      "tokens_seen": 2930442240
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.261274273792329e-05,
      "loss": 1.3193,
      "theoretical_loss": 3.3552284973280275,
      "tokens_seen": 2930573312
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2604718343765045e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.3552170030707513,
      "tokens_seen": 2930704384
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2596693949606805e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.355205509471461,
      "tokens_seen": 2930835456
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2588669555448565e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.355194016530089,
      "tokens_seen": 2930966528
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.258064516129032e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.355182524246568,
      "tokens_seen": 2931097600
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.257262076713208e-05,
      "loss": 1.3801,
      "theoretical_loss": 3.3551710326208317,
      "tokens_seen": 2931228672
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.256459637297384e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.355159541652813,
      "tokens_seen": 2931359744
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2556571978815598e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.355148051342444,
      "tokens_seen": 2931490816
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2548547584657358e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.3551365616896582,
      "tokens_seen": 2931621888
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2540523190499118e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.3551250726943884,
      "tokens_seen": 2931752960
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2532498796340878e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.3551135843565683,
      "tokens_seen": 2931884032
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2524474402182634e-05,
      "loss": 1.3611,
      "theoretical_loss": 3.35510209667613,
      "tokens_seen": 2932015104
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2516450008024394e-05,
      "loss": 1.2101,
      "theoretical_loss": 3.355090609653007,
      "tokens_seen": 2932146176
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2508425613866154e-05,
      "loss": 1.22,
      "theoretical_loss": 3.3550791232871315,
      "tokens_seen": 2932277248
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2500401219707914e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.3550676375784376,
      "tokens_seen": 2932408320
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.249237682554967e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.3550561525268576,
      "tokens_seen": 2932539392
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.248435243139143e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.355044668132325,
      "tokens_seen": 2932670464
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.247632803723319e-05,
      "loss": 1.3423,
      "theoretical_loss": 3.355033184394773,
      "tokens_seen": 2932801536
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2468303643074947e-05,
      "loss": 1.236,
      "theoretical_loss": 3.3550217013141337,
      "tokens_seen": 2932932608
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2460279248916707e-05,
      "loss": 1.2337,
      "theoretical_loss": 3.355010218890341,
      "tokens_seen": 2933063680
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2452254854758466e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.354998737123328,
      "tokens_seen": 2933194752
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2444230460600226e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.354987256013027,
      "tokens_seen": 2933325824
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2436206066441983e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.354975775559372,
      "tokens_seen": 2933456896
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2428181672283743e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.3549642957622954,
      "tokens_seen": 2933587968
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2420157278125503e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.354952816621731,
      "tokens_seen": 2933719040
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2412132883967263e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.3549413381376114,
      "tokens_seen": 2933850112
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.240410848980902e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3549298603098703,
      "tokens_seen": 2933981184
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.239608409565078e-05,
      "loss": 1.348,
      "theoretical_loss": 3.35491838313844,
      "tokens_seen": 2934112256
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.238805970149254e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.3549069066232544,
      "tokens_seen": 2934243328
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2380035307334295e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.3548954307642465,
      "tokens_seen": 2934374400
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.9888792037963867,
      "objective/train/avg_token_score": 0.011120806448161602,
      "objective/train/docs_used": 1066769,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.327968716621399,
      "objective/train/original_loss": 1.3447877168655396,
      "objective/train/theoretical_loss": 3.354883955561349,
      "objective/train/tokens_used": 1304964576,
      "objective/train/unlikelihood_loss": 0.025336189195513725,
      "theoretical_loss": 3.354883955561349,
      "tokens_seen": 2934505472
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2372010913176055e-05,
      "loss": 1.2857,
      "theoretical_loss": 3.354883955561349,
      "tokens_seen": 2934505472
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2363986519017815e-05,
      "loss": 1.2441,
      "theoretical_loss": 3.3548724810144956,
      "tokens_seen": 2934636544
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2355962124859575e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.3548610071236196,
      "tokens_seen": 2934767616
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.234793773070133e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.3548495338886535,
      "tokens_seen": 2934898688
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.233991333654309e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.354838061309531,
      "tokens_seen": 2935029760
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.233188894238485e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3548265893861853,
      "tokens_seen": 2935160832
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.232386454822661e-05,
      "loss": 1.2596,
      "theoretical_loss": 3.35481511811855,
      "tokens_seen": 2935291904
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2315840154068368e-05,
      "loss": 1.2191,
      "theoretical_loss": 3.3548036475065577,
      "tokens_seen": 2935422976
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2307815759910128e-05,
      "loss": 1.1993,
      "theoretical_loss": 3.354792177550142,
      "tokens_seen": 2935554048
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2299791365751888e-05,
      "loss": 1.3692,
      "theoretical_loss": 3.354780708249236,
      "tokens_seen": 2935685120
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2291766971593644e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.354769239603773,
      "tokens_seen": 2935816192
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2283742577435404e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.354757771613686,
      "tokens_seen": 2935947264
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2275718183277164e-05,
      "loss": 1.1952,
      "theoretical_loss": 3.354746304278909,
      "tokens_seen": 2936078336
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2267693789118924e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.354734837599375,
      "tokens_seen": 2936209408
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.225966939496068e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.3547233715750173,
      "tokens_seen": 2936340480
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.225164500080244e-05,
      "loss": 1.2096,
      "theoretical_loss": 3.354711906205769,
      "tokens_seen": 2936471552
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.22436206066442e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.3547004414915635,
      "tokens_seen": 2936602624
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.223559621248596e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.3546889774323345,
      "tokens_seen": 2936733696
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2227571818327717e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.3546775140280145,
      "tokens_seen": 2936864768
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2219547424169477e-05,
      "loss": 1.2194,
      "theoretical_loss": 3.354666051278538,
      "tokens_seen": 2936995840
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2211523030011236e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.3546545891838377,
      "tokens_seen": 2937126912
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2203498635852993e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.354643127743847,
      "tokens_seen": 2937257984
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2195474241694753e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3546316669584995,
      "tokens_seen": 2937389056
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2187449847536513e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.3546202068277284,
      "tokens_seen": 2937520128
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2179425453378273e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.3546087473514676,
      "tokens_seen": 2937651200
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.217140105922003e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.3545972885296496,
      "tokens_seen": 2937782272
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.216337666506179e-05,
      "loss": 1.2468,
      "theoretical_loss": 3.3545858303622085,
      "tokens_seen": 2937913344
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.215535227090355e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.3545743728490773,
      "tokens_seen": 2938044416
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.214732787674531e-05,
      "loss": 1.3,
      "theoretical_loss": 3.35456291599019,
      "tokens_seen": 2938175488
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2139303482587065e-05,
      "loss": 1.3425,
      "theoretical_loss": 3.3545514597854798,
      "tokens_seen": 2938306560
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2131279088428825e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.35454000423488,
      "tokens_seen": 2938437632
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2123254694270585e-05,
      "loss": 1.2702,
      "theoretical_loss": 3.354528549338325,
      "tokens_seen": 2938568704
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2115230300112342e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.354517095095747,
      "tokens_seen": 2938699776
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2107205905954102e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.3545056415070795,
      "tokens_seen": 2938830848
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.209918151179586e-05,
      "loss": 1.2465,
      "theoretical_loss": 3.354494188572257,
      "tokens_seen": 2938961920
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.209115711763762e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.3544827362912124,
      "tokens_seen": 2939092992
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2083132723479378e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.3544712846638793,
      "tokens_seen": 2939224064
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2075108329321138e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.354459833690192,
      "tokens_seen": 2939355136
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2067083935162898e-05,
      "loss": 1.3125,
      "theoretical_loss": 3.3544483833700824,
      "tokens_seen": 2939486208
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2059059541004658e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.3544369337034854,
      "tokens_seen": 2939617280
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2051035146846414e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.354425484690334,
      "tokens_seen": 2939748352
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2043010752688174e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.3544140363305623,
      "tokens_seen": 2939879424
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2034986358529934e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.3544025886241036,
      "tokens_seen": 2940010496
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.2026961964371694e-05,
      "loss": 1.271,
      "theoretical_loss": 3.354391141570891,
      "tokens_seen": 2940141568
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.201893757021345e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.354379695170859,
      "tokens_seen": 2940272640
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.201091317605521e-05,
      "loss": 1.2538,
      "theoretical_loss": 3.35436824942394,
      "tokens_seen": 2940403712
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.200288878189697e-05,
      "loss": 1.2285,
      "theoretical_loss": 3.3543568043300693,
      "tokens_seen": 2940534784
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1994864387738727e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.354345359889179,
      "tokens_seen": 2940665856
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1986839993580487e-05,
      "loss": 1.274,
      "theoretical_loss": 3.3543339161012034,
      "tokens_seen": 2940796928
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1978815599422247e-05,
      "loss": 1.272,
      "theoretical_loss": 3.354322472966076,
      "tokens_seen": 2940928000
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.9926429986953735,
      "objective/train/avg_token_score": 0.0073569766245782375,
      "objective/train/docs_used": 1069310,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3058805465698242,
      "objective/train/original_loss": 1.3160061836242676,
      "objective/train/theoretical_loss": 3.354311030483731,
      "objective/train/tokens_used": 1311518176,
      "objective/train/unlikelihood_loss": 0.01961762271821499,
      "theoretical_loss": 3.354311030483731,
      "tokens_seen": 2941059072
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1970791205264003e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.354311030483731,
      "tokens_seen": 2941059072
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1962766811105763e-05,
      "loss": 1.2251,
      "theoretical_loss": 3.354299588654101,
      "tokens_seen": 2941190144
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1954742416947523e-05,
      "loss": 1.305,
      "theoretical_loss": 3.3542881474771207,
      "tokens_seen": 2941321216
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.194671802278928e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.3542767069527235,
      "tokens_seen": 2941452288
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.193869362863104e-05,
      "loss": 1.3351,
      "theoretical_loss": 3.354265267080843,
      "tokens_seen": 2941583360
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.19306692344728e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.3542538278614127,
      "tokens_seen": 2941714432
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1922644840314556e-05,
      "loss": 1.2066,
      "theoretical_loss": 3.3542423892943667,
      "tokens_seen": 2941845504
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1914620446156316e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.3542309513796384,
      "tokens_seen": 2941976576
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1906596051998076e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.354219514117162,
      "tokens_seen": 2942107648
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1898571657839832e-05,
      "loss": 1.346,
      "theoretical_loss": 3.3542080775068706,
      "tokens_seen": 2942238720
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1890547263681592e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.3541966415486986,
      "tokens_seen": 2942369792
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1882522869523352e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.354185206242579,
      "tokens_seen": 2942500864
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.187449847536511e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3541737715884468,
      "tokens_seen": 2942631936
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.186647408120687e-05,
      "loss": 1.29,
      "theoretical_loss": 3.3541623375862346,
      "tokens_seen": 2942763008
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.185844968704863e-05,
      "loss": 1.266,
      "theoretical_loss": 3.3541509042358766,
      "tokens_seen": 2942894080
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1850425292890388e-05,
      "loss": 1.2259,
      "theoretical_loss": 3.354139471537307,
      "tokens_seen": 2943025152
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1842400898732145e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.354128039490459,
      "tokens_seen": 2943156224
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1834376504573905e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.354116608095267,
      "tokens_seen": 2943287296
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1826352110415665e-05,
      "loss": 1.3334,
      "theoretical_loss": 3.354105177351664,
      "tokens_seen": 2943418368
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.181832771625742e-05,
      "loss": 1.271,
      "theoretical_loss": 3.3540937472595846,
      "tokens_seen": 2943549440
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.181030332209918e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.3540823178189623,
      "tokens_seen": 2943680512
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.180227892794094e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.3540708890297317,
      "tokens_seen": 2943811584
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.17942545337827e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.3540594608918255,
      "tokens_seen": 2943942656
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1786230139624457e-05,
      "loss": 1.3826,
      "theoretical_loss": 3.3540480334051783,
      "tokens_seen": 2944073728
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1778205745466217e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.354036606569724,
      "tokens_seen": 2944204800
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1770181351307977e-05,
      "loss": 1.3624,
      "theoretical_loss": 3.3540251803853964,
      "tokens_seen": 2944335872
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1762156957149737e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.354013754852129,
      "tokens_seen": 2944466944
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1754132562991494e-05,
      "loss": 1.359,
      "theoretical_loss": 3.3540023299698563,
      "tokens_seen": 2944598016
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1746108168833253e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.353990905738512,
      "tokens_seen": 2944729088
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.1738083774675013e-05,
      "loss": 1.252,
      "theoretical_loss": 3.3539794821580307,
      "tokens_seen": 2944860160
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.173005938051677e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.353968059228345,
      "tokens_seen": 2944991232
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.172203498635853e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3539566369493894,
      "tokens_seen": 2945122304
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.171401059220029e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.3539452153210987,
      "tokens_seen": 2945253376
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.170598619804205e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.3539337943434058,
      "tokens_seen": 2945384448
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1697961803883806e-05,
      "loss": 1.2653,
      "theoretical_loss": 3.3539223740162454,
      "tokens_seen": 2945515520
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1689937409725566e-05,
      "loss": 1.2551,
      "theoretical_loss": 3.3539109543395513,
      "tokens_seen": 2945646592
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1681913015567326e-05,
      "loss": 1.3122,
      "theoretical_loss": 3.353899535313257,
      "tokens_seen": 2945777664
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1673888621409086e-05,
      "loss": 1.1929,
      "theoretical_loss": 3.3538881169372976,
      "tokens_seen": 2945908736
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1665864227250842e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.353876699211606,
      "tokens_seen": 2946039808
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1657839833092602e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.353865282136117,
      "tokens_seen": 2946170880
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1649815438934362e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.3538538657107644,
      "tokens_seen": 2946301952
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.164179104477612e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.353842449935482,
      "tokens_seen": 2946433024
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.163376665061788e-05,
      "loss": 1.234,
      "theoretical_loss": 3.353831034810204,
      "tokens_seen": 2946564096
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.162574225645964e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.353819620334865,
      "tokens_seen": 2946695168
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.16177178623014e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.3538082065093984,
      "tokens_seen": 2946826240
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1609693468143155e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.3537967933337387,
      "tokens_seen": 2946957312
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1601669073984915e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.3537853808078197,
      "tokens_seen": 2947088384
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1593644679826675e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.353773968931576,
      "tokens_seen": 2947219456
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1585620285668435e-05,
      "loss": 1.2054,
      "theoretical_loss": 3.3537625577049415,
      "tokens_seen": 2947350528
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.157759589151019e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.35375114712785,
      "tokens_seen": 2947481600
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.9914300441741943,
      "objective/train/avg_token_score": 0.008569997735321522,
      "objective/train/docs_used": 1071710,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.377148151397705,
      "objective/train/original_loss": 1.3885929584503174,
      "objective/train/theoretical_loss": 3.3537397372002355,
      "objective/train/tokens_used": 1318071776,
      "objective/train/unlikelihood_loss": 0.023678481578826904,
      "theoretical_loss": 3.3537397372002355,
      "tokens_seen": 2947612672
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.156957149735195e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.3537397372002355,
      "tokens_seen": 2947612672
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.156154710319371e-05,
      "loss": 1.1978,
      "theoretical_loss": 3.353728327922033,
      "tokens_seen": 2947743744
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.155352270903547e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.353716919293176,
      "tokens_seen": 2947874816
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1545498314877227e-05,
      "loss": 1.2253,
      "theoretical_loss": 3.353705511313599,
      "tokens_seen": 2948005888
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1537473920718987e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3536941039832366,
      "tokens_seen": 2948136960
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1529449526560747e-05,
      "loss": 1.3362,
      "theoretical_loss": 3.353682697302022,
      "tokens_seen": 2948268032
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1521425132402504e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.35367129126989,
      "tokens_seen": 2948399104
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1513400738244264e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.353659885886774,
      "tokens_seen": 2948530176
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1505376344086024e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.35364848115261,
      "tokens_seen": 2948661248
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1497351949927783e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.3536370770673303,
      "tokens_seen": 2948792320
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.148932755576954e-05,
      "loss": 1.1896,
      "theoretical_loss": 3.3536256736308703,
      "tokens_seen": 2948923392
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.14813031616113e-05,
      "loss": 1.2791,
      "theoretical_loss": 3.3536142708431638,
      "tokens_seen": 2949054464
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.147327876745306e-05,
      "loss": 1.4149,
      "theoretical_loss": 3.3536028687041455,
      "tokens_seen": 2949185536
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.146525437329482e-05,
      "loss": 1.1967,
      "theoretical_loss": 3.3535914672137492,
      "tokens_seen": 2949316608
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1457229979136576e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.353580066371909,
      "tokens_seen": 2949447680
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1449205584978336e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.35356866617856,
      "tokens_seen": 2949578752
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1441181190820096e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.3535572666336355,
      "tokens_seen": 2949709824
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1433156796661853e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.353545867737071,
      "tokens_seen": 2949840896
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1425132402503612e-05,
      "loss": 1.2145,
      "theoretical_loss": 3.3535344694887996,
      "tokens_seen": 2949971968
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1417108008345372e-05,
      "loss": 1.2248,
      "theoretical_loss": 3.3535230718887563,
      "tokens_seen": 2950103040
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1409083614187132e-05,
      "loss": 1.2168,
      "theoretical_loss": 3.3535116749368754,
      "tokens_seen": 2950234112
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.140105922002889e-05,
      "loss": 1.309,
      "theoretical_loss": 3.353500278633091,
      "tokens_seen": 2950365184
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.139303482587065e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.353488882977338,
      "tokens_seen": 2950496256
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.138501043171241e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.3534774879695504,
      "tokens_seen": 2950627328
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.137698603755417e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.3534660936096623,
      "tokens_seen": 2950758400
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1368961643395925e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.3534546998976085,
      "tokens_seen": 2950889472
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1360937249237685e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.3534433068333227,
      "tokens_seen": 2951020544
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1352912855079445e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.35343191441674,
      "tokens_seen": 2951151616
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.13448884609212e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.353420522647795,
      "tokens_seen": 2951282688
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.133686406676296e-05,
      "loss": 1.334,
      "theoretical_loss": 3.353409131526422,
      "tokens_seen": 2951413760
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.132883967260472e-05,
      "loss": 1.263,
      "theoretical_loss": 3.3533977410525546,
      "tokens_seen": 2951544832
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1320815278446478e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.353386351226128,
      "tokens_seen": 2951675904
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1312790884288238e-05,
      "loss": 1.2524,
      "theoretical_loss": 3.3533749620470767,
      "tokens_seen": 2951806976
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1304766490129997e-05,
      "loss": 1.2823,
      "theoretical_loss": 3.3533635735153347,
      "tokens_seen": 2951938048
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1296742095971754e-05,
      "loss": 1.317,
      "theoretical_loss": 3.353352185630837,
      "tokens_seen": 2952069120
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1288717701813514e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.3533407983935177,
      "tokens_seen": 2952200192
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1280693307655274e-05,
      "loss": 1.3432,
      "theoretical_loss": 3.3533294118033115,
      "tokens_seen": 2952331264
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.127266891349703e-05,
      "loss": 1.2119,
      "theoretical_loss": 3.3533180258601525,
      "tokens_seen": 2952462336
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.126464451933879e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.3533066405639755,
      "tokens_seen": 2952593408
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.125662012518055e-05,
      "loss": 1.234,
      "theoretical_loss": 3.3532952559147153,
      "tokens_seen": 2952724480
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1248595731022307e-05,
      "loss": 1.3502,
      "theoretical_loss": 3.353283871912306,
      "tokens_seen": 2952855552
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1240571336864067e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.3532724885566827,
      "tokens_seen": 2952986624
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1232546942705826e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.353261105847779,
      "tokens_seen": 2953117696
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1224522548547583e-05,
      "loss": 1.2,
      "theoretical_loss": 3.35324972378553,
      "tokens_seen": 2953248768
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1216498154389343e-05,
      "loss": 1.3322,
      "theoretical_loss": 3.3532383423698704,
      "tokens_seen": 2953379840
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1208473760231103e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.353226961600735,
      "tokens_seen": 2953510912
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1200449366072863e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.3532155814780573,
      "tokens_seen": 2953641984
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.119242497191462e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.353204202001773,
      "tokens_seen": 2953773056
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.118440057775638e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.3531928231718164,
      "tokens_seen": 2953904128
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.117637618359814e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.353181444988122,
      "tokens_seen": 2954035200
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.9946377277374268,
      "objective/train/avg_token_score": 0.0053623141720891,
      "objective/train/docs_used": 1074062,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.275276780128479,
      "objective/train/original_loss": 1.2826308012008667,
      "objective/train/theoretical_loss": 3.353170067450624,
      "objective/train/tokens_used": 1324625376,
      "objective/train/unlikelihood_loss": 0.015402885153889656,
      "theoretical_loss": 3.353170067450624,
      "tokens_seen": 2954166272
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1168351789439896e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.353170067450624,
      "tokens_seen": 2954166272
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1160327395281655e-05,
      "loss": 1.2637,
      "theoretical_loss": 3.3531586905592583,
      "tokens_seen": 2954297344
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1152303001123415e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.3531473143139583,
      "tokens_seen": 2954428416
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1144278606965175e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.353135938714659,
      "tokens_seen": 2954559488
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1136254212806932e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.353124563761295,
      "tokens_seen": 2954690560
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.112822981864869e-05,
      "loss": 1.3127,
      "theoretical_loss": 3.3531131894538015,
      "tokens_seen": 2954821632
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.112020542449045e-05,
      "loss": 1.3183,
      "theoretical_loss": 3.3531018157921126,
      "tokens_seen": 2954952704
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.111218103033221e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.353090442776163,
      "tokens_seen": 2955083776
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1104156636173968e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.3530790704058884,
      "tokens_seen": 2955214848
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1096132242015728e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.3530676986812216,
      "tokens_seen": 2955345920
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1088107847857488e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.353056327602099,
      "tokens_seen": 2955476992
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1080083453699248e-05,
      "loss": 1.2176,
      "theoretical_loss": 3.3530449571684544,
      "tokens_seen": 2955608064
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1072059059541004e-05,
      "loss": 1.2253,
      "theoretical_loss": 3.3530335873802235,
      "tokens_seen": 2955739136
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1064034665382764e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.3530222182373395,
      "tokens_seen": 2955870208
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1056010271224524e-05,
      "loss": 1.2514,
      "theoretical_loss": 3.353010849739739,
      "tokens_seen": 2956001280
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.104798587706628e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.3529994818873554,
      "tokens_seen": 2956132352
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.103996148290804e-05,
      "loss": 1.2275,
      "theoretical_loss": 3.352988114680124,
      "tokens_seen": 2956263424
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.10319370887498e-05,
      "loss": 1.3595,
      "theoretical_loss": 3.352976748117979,
      "tokens_seen": 2956394496
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.102391269459156e-05,
      "loss": 1.339,
      "theoretical_loss": 3.352965382200856,
      "tokens_seen": 2956525568
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1015888300433317e-05,
      "loss": 1.3151,
      "theoretical_loss": 3.35295401692869,
      "tokens_seen": 2956656640
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.1007863906275077e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.3529426523014147,
      "tokens_seen": 2956787712
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0999839512116837e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.3529312883189655,
      "tokens_seen": 2956918784
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0991815117958597e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.3529199249812778,
      "tokens_seen": 2957049856
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0983790723800353e-05,
      "loss": 1.3074,
      "theoretical_loss": 3.3529085622882855,
      "tokens_seen": 2957180928
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0975766329642113e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.3528972002399238,
      "tokens_seen": 2957312000
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0967741935483873e-05,
      "loss": 1.313,
      "theoretical_loss": 3.352885838836128,
      "tokens_seen": 2957443072
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.095971754132563e-05,
      "loss": 1.2703,
      "theoretical_loss": 3.3528744780768323,
      "tokens_seen": 2957574144
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.095169314716739e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.352863117961972,
      "tokens_seen": 2957705216
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.094366875300915e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.3528517584914814,
      "tokens_seen": 2957836288
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.093564435885091e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.352840399665296,
      "tokens_seen": 2957967360
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0927619964692666e-05,
      "loss": 1.215,
      "theoretical_loss": 3.352829041483351,
      "tokens_seen": 2958098432
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0919595570534426e-05,
      "loss": 1.2293,
      "theoretical_loss": 3.3528176839455806,
      "tokens_seen": 2958229504
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0911571176376185e-05,
      "loss": 1.1934,
      "theoretical_loss": 3.35280632705192,
      "tokens_seen": 2958360576
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0903546782217945e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.352794970802304,
      "tokens_seen": 2958491648
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0895522388059702e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.352783615196668,
      "tokens_seen": 2958622720
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0887497993901462e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.3527722602349463,
      "tokens_seen": 2958753792
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.087947359974322e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.3527609059170747,
      "tokens_seen": 2958884864
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0871449205584978e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.3527495522429875,
      "tokens_seen": 2959015936
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0863424811426738e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.3527381992126193,
      "tokens_seen": 2959147008
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0855400417268498e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.3527268468259064,
      "tokens_seen": 2959278080
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0847376023110258e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3527154950827827,
      "tokens_seen": 2959409152
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0839351628952014e-05,
      "loss": 1.3444,
      "theoretical_loss": 3.3527041439831833,
      "tokens_seen": 2959540224
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0831327234793774e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.352692793527044,
      "tokens_seen": 2959671296
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0823302840635534e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.352681443714299,
      "tokens_seen": 2959802368
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0815278446477294e-05,
      "loss": 1.3943,
      "theoretical_loss": 3.3526700945448837,
      "tokens_seen": 2959933440
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.080725405231905e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.352658746018733,
      "tokens_seen": 2960064512
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.079922965816081e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3526473981357823,
      "tokens_seen": 2960195584
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.079120526400257e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.352636050895966,
      "tokens_seen": 2960326656
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0783180869844327e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.3526247042992203,
      "tokens_seen": 2960457728
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0775156475686087e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.3526133583454794,
      "tokens_seen": 2960588800
    },
    {
      "debugging/Compilability": 1.0,
      "debugging/distinct-1-grams": 0.7459160326454821,
      "debugging/entropy-1-grams": 5.28426573657414,
      "debugging/length": 471.5,
      "debugging/num_segments": 12,
      "debugging/score": 0.006897032404886403,
      "debugging/score_std": 0.00540563131039379,
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.9946767687797546,
      "objective/train/avg_token_score": 0.0053232889622449875,
      "objective/train/docs_used": 1076444,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.244368553161621,
      "objective/train/original_loss": 1.2512257099151611,
      "objective/train/theoretical_loss": 3.3526020130346783,
      "objective/train/tokens_used": 1331178976,
      "objective/train/unlikelihood_loss": 0.01698208414018154,
      "theoretical_loss": 3.3526020130346783,
      "tokens_seen": 2960719872
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0767132081527847e-05,
      "loss": 1.2633,
      "theoretical_loss": 3.3526020130346783,
      "tokens_seen": 2960719872
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0759107687369607e-05,
      "loss": 1.2762,
      "theoretical_loss": 3.3525906683667523,
      "tokens_seen": 2960850944
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0751083293211363e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.3525793243416366,
      "tokens_seen": 2960982016
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0743058899053123e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.352567980959267,
      "tokens_seen": 2961113088
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0735034504894883e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.3525566382195775,
      "tokens_seen": 2961244160
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.0727010110736643e-05,
      "loss": 1.1799,
      "theoretical_loss": 3.3525452961225035,
      "tokens_seen": 2961375232
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.07189857165784e-05,
      "loss": 1.3472,
      "theoretical_loss": 3.352533954667981,
      "tokens_seen": 2961506304
    },
    {
      "epoch": 0.79,
      "learning_rate": 2.071096132242016e-05,
      "loss": 1.237,
      "theoretical_loss": 3.352522613855944,
      "tokens_seen": 2961637376
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.070293692826192e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.352511273686328,
      "tokens_seen": 2961768448
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0694912534103676e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.352499934159069,
      "tokens_seen": 2961899520
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0686888139945436e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.3524885952741013,
      "tokens_seen": 2962030592
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0678863745787196e-05,
      "loss": 1.3334,
      "theoretical_loss": 3.35247725703136,
      "tokens_seen": 2962161664
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0670839351628955e-05,
      "loss": 1.194,
      "theoretical_loss": 3.3524659194307813,
      "tokens_seen": 2962292736
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0662814957470712e-05,
      "loss": 1.3532,
      "theoretical_loss": 3.3524545824722995,
      "tokens_seen": 2962423808
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0654790563312472e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.3524432461558504,
      "tokens_seen": 2962554880
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0646766169154232e-05,
      "loss": 1.2776,
      "theoretical_loss": 3.3524319104813687,
      "tokens_seen": 2962685952
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.063874177499599e-05,
      "loss": 1.1902,
      "theoretical_loss": 3.35242057544879,
      "tokens_seen": 2962817024
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0630717380837748e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3524092410580493,
      "tokens_seen": 2962948096
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0622692986679508e-05,
      "loss": 1.1833,
      "theoretical_loss": 3.352397907309082,
      "tokens_seen": 2963079168
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0614668592521265e-05,
      "loss": 1.2057,
      "theoretical_loss": 3.352386574201824,
      "tokens_seen": 2963210240
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0606644198363025e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.352375241736209,
      "tokens_seen": 2963341312
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0598619804204784e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.3523639099121745,
      "tokens_seen": 2963472384
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.059059541004654e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.3523525787296538,
      "tokens_seen": 2963603456
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.05825710158883e-05,
      "loss": 1.261,
      "theoretical_loss": 3.352341248188583,
      "tokens_seen": 2963734528
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.057454662173006e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.3523299182888975,
      "tokens_seen": 2963865600
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0566522227571817e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.3523185890305327,
      "tokens_seen": 2963996672
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0558497833413577e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.352307260413424,
      "tokens_seen": 2964127744
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0550473439255337e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.352295932437506,
      "tokens_seen": 2964258816
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0542449045097094e-05,
      "loss": 1.3629,
      "theoretical_loss": 3.352284605102715,
      "tokens_seen": 2964389888
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0534424650938854e-05,
      "loss": 1.2332,
      "theoretical_loss": 3.352273278408986,
      "tokens_seen": 2964520960
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0526400256780613e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3522619523562542,
      "tokens_seen": 2964652032
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0518375862622373e-05,
      "loss": 1.2539,
      "theoretical_loss": 3.352250626944455,
      "tokens_seen": 2964783104
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.051035146846413e-05,
      "loss": 1.219,
      "theoretical_loss": 3.3522393021735244,
      "tokens_seen": 2964914176
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.050232707430589e-05,
      "loss": 1.2403,
      "theoretical_loss": 3.352227978043397,
      "tokens_seen": 2965045248
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.049430268014765e-05,
      "loss": 1.3459,
      "theoretical_loss": 3.3522166545540086,
      "tokens_seen": 2965176320
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0486278285989406e-05,
      "loss": 1.2207,
      "theoretical_loss": 3.352205331705295,
      "tokens_seen": 2965307392
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0478253891831166e-05,
      "loss": 1.1669,
      "theoretical_loss": 3.3521940094971905,
      "tokens_seen": 2965438464
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0470229497672926e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3521826879296315,
      "tokens_seen": 2965569536
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0462205103514686e-05,
      "loss": 1.2006,
      "theoretical_loss": 3.3521713670025535,
      "tokens_seen": 2965700608
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0454180709356442e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.3521600467158916,
      "tokens_seen": 2965831680
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0446156315198202e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3521487270695816,
      "tokens_seen": 2965962752
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0438131921039962e-05,
      "loss": 1.2006,
      "theoretical_loss": 3.3521374080635584,
      "tokens_seen": 2966093824
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0430107526881722e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.352126089697758,
      "tokens_seen": 2966224896
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.042208313272348e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.3521147719721154,
      "tokens_seen": 2966355968
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.041405873856524e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.3521034548865667,
      "tokens_seen": 2966487040
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0406034344407e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.352092138441047,
      "tokens_seen": 2966618112
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0398009950248755e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.352080822635492,
      "tokens_seen": 2966749184
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0389985556090515e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.3520695074698375,
      "tokens_seen": 2966880256
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0381961161932275e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.3520581929440185,
      "tokens_seen": 2967011328
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0373936767774035e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.352046879057971,
      "tokens_seen": 2967142400
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.9933345317840576,
      "objective/train/avg_token_score": 0.006665495224297047,
      "objective/train/docs_used": 1078969,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2845011949539185,
      "objective/train/original_loss": 1.293672800064087,
      "objective/train/theoretical_loss": 3.3520355658116303,
      "objective/train/tokens_used": 1337732576,
      "objective/train/unlikelihood_loss": 0.0200672447681427,
      "theoretical_loss": 3.3520355658116303,
      "tokens_seen": 2967273472
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.036591237361579e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.3520355658116303,
      "tokens_seen": 2967273472
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.035788797945755e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.352024253204932,
      "tokens_seen": 2967404544
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.034986358529931e-05,
      "loss": 1.245,
      "theoretical_loss": 3.3520129412378115,
      "tokens_seen": 2967535616
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.034183919114107e-05,
      "loss": 1.2462,
      "theoretical_loss": 3.352001629910205,
      "tokens_seen": 2967666688
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0333814796982827e-05,
      "loss": 1.2,
      "theoretical_loss": 3.3519903192220477,
      "tokens_seen": 2967797760
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0325790402824587e-05,
      "loss": 1.3209,
      "theoretical_loss": 3.3519790091732755,
      "tokens_seen": 2967928832
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0317766008666347e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.351967699763823,
      "tokens_seen": 2968059904
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0309741614508104e-05,
      "loss": 1.2093,
      "theoretical_loss": 3.3519563909936276,
      "tokens_seen": 2968190976
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0301717220349864e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.351945082862623,
      "tokens_seen": 2968322048
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0293692826191624e-05,
      "loss": 1.3089,
      "theoretical_loss": 3.3519337753707465,
      "tokens_seen": 2968453120
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0285668432033384e-05,
      "loss": 1.3078,
      "theoretical_loss": 3.351922468517933,
      "tokens_seen": 2968584192
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.027764403787514e-05,
      "loss": 1.2578,
      "theoretical_loss": 3.351911162304118,
      "tokens_seen": 2968715264
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.02696196437169e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3518998567292373,
      "tokens_seen": 2968846336
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.026159524955866e-05,
      "loss": 1.2194,
      "theoretical_loss": 3.351888551793227,
      "tokens_seen": 2968977408
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.025357085540042e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.3518772474960223,
      "tokens_seen": 2969108480
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0245546461242176e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.3518659438375593,
      "tokens_seen": 2969239552
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0237522067083936e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.3518546408177734,
      "tokens_seen": 2969370624
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0229497672925696e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3518433384366,
      "tokens_seen": 2969501696
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0221473278767453e-05,
      "loss": 1.27,
      "theoretical_loss": 3.351832036693976,
      "tokens_seen": 2969632768
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0213448884609213e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.351820735589836,
      "tokens_seen": 2969763840
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0205424490450972e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.351809435124116,
      "tokens_seen": 2969894912
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0197400096292732e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.351798135296752,
      "tokens_seen": 2970025984
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.018937570213449e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.35178683610768,
      "tokens_seen": 2970157056
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.018135130797625e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.3517755375568354,
      "tokens_seen": 2970288128
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.017332691381801e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.351764239644154,
      "tokens_seen": 2970419200
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.016530251965977e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.3517529423695716,
      "tokens_seen": 2970550272
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0157278125501525e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.351741645733024,
      "tokens_seen": 2970681344
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0149253731343285e-05,
      "loss": 1.2252,
      "theoretical_loss": 3.351730349734447,
      "tokens_seen": 2970812416
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0141229337185045e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.3517190543737767,
      "tokens_seen": 2970943488
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0133204943026805e-05,
      "loss": 1.1804,
      "theoretical_loss": 3.3517077596509486,
      "tokens_seen": 2971074560
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.012518054886856e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.3516964655658987,
      "tokens_seen": 2971205632
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.011715615471032e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.351685172118563,
      "tokens_seen": 2971336704
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.010913176055208e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.351673879308877,
      "tokens_seen": 2971467776
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0101107366393838e-05,
      "loss": 1.3271,
      "theoretical_loss": 3.351662587136777,
      "tokens_seen": 2971598848
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0093082972235598e-05,
      "loss": 1.2444,
      "theoretical_loss": 3.3516512956021987,
      "tokens_seen": 2971729920
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0085058578077357e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.351640004705078,
      "tokens_seen": 2971860992
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0077034183919117e-05,
      "loss": 1.253,
      "theoretical_loss": 3.3516287144453503,
      "tokens_seen": 2971992064
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0069009789760874e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.351617424822952,
      "tokens_seen": 2972123136
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0060985395602634e-05,
      "loss": 1.2201,
      "theoretical_loss": 3.3516061358378195,
      "tokens_seen": 2972254208
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0052961001444394e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.3515948474898876,
      "tokens_seen": 2972385280
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0044936607286154e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.351583559779093,
      "tokens_seen": 2972516352
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.003691221312791e-05,
      "loss": 1.309,
      "theoretical_loss": 3.351572272705372,
      "tokens_seen": 2972647424
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.002888781896967e-05,
      "loss": 1.285,
      "theoretical_loss": 3.3515609862686597,
      "tokens_seen": 2972778496
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.002086342481143e-05,
      "loss": 1.321,
      "theoretical_loss": 3.351549700468892,
      "tokens_seen": 2972909568
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0012839030653186e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.351538415306006,
      "tokens_seen": 2973040640
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.0004814636494946e-05,
      "loss": 1.2175,
      "theoretical_loss": 3.3515271307799366,
      "tokens_seen": 2973171712
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9996790242336706e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.3515158468906203,
      "tokens_seen": 2973302784
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9988765848178466e-05,
      "loss": 1.3592,
      "theoretical_loss": 3.3515045636379925,
      "tokens_seen": 2973433856
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9980741454020223e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.35149328102199,
      "tokens_seen": 2973564928
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9972717059861983e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.351481999042549,
      "tokens_seen": 2973696000
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.9927148818969727,
      "objective/train/avg_token_score": 0.00728512741625309,
      "objective/train/docs_used": 1081215,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2408437728881836,
      "objective/train/original_loss": 1.2506749629974365,
      "objective/train/theoretical_loss": 3.3514707176996046,
      "objective/train/tokens_used": 1344286176,
      "objective/train/unlikelihood_loss": 0.019975585862994194,
      "theoretical_loss": 3.3514707176996046,
      "tokens_seen": 2973827072
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9964692665703743e-05,
      "loss": 1.3277,
      "theoretical_loss": 3.3514707176996046,
      "tokens_seen": 2973827072
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.99566682715455e-05,
      "loss": 1.209,
      "theoretical_loss": 3.351459436993093,
      "tokens_seen": 2973958144
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.994864387738726e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.3514481569229506,
      "tokens_seen": 2974089216
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.994061948322902e-05,
      "loss": 1.2916,
      "theoretical_loss": 3.351436877489114,
      "tokens_seen": 2974220288
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9932595089070775e-05,
      "loss": 1.2909,
      "theoretical_loss": 3.351425598691518,
      "tokens_seen": 2974351360
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9924570694912535e-05,
      "loss": 1.2266,
      "theoretical_loss": 3.3514143205300995,
      "tokens_seen": 2974482432
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9916546300754295e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.3514030430047943,
      "tokens_seen": 2974613504
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.990852190659605e-05,
      "loss": 1.2774,
      "theoretical_loss": 3.3513917661155386,
      "tokens_seen": 2974744576
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.990049751243781e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.3513804898622688,
      "tokens_seen": 2974875648
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.989247311827957e-05,
      "loss": 1.2369,
      "theoretical_loss": 3.3513692142449205,
      "tokens_seen": 2975006720
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9884448724121328e-05,
      "loss": 1.3481,
      "theoretical_loss": 3.35135793926343,
      "tokens_seen": 2975137792
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9876424329963088e-05,
      "loss": 1.3481,
      "theoretical_loss": 3.3513466649177337,
      "tokens_seen": 2975268864
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9868399935804848e-05,
      "loss": 1.1769,
      "theoretical_loss": 3.3513353912077672,
      "tokens_seen": 2975399936
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9860375541646604e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.351324118133467,
      "tokens_seen": 2975531008
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9852351147488364e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3513128456947694,
      "tokens_seen": 2975662080
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9844326753330124e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.35130157389161,
      "tokens_seen": 2975793152
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.983630235917188e-05,
      "loss": 1.3237,
      "theoretical_loss": 3.351290302723926,
      "tokens_seen": 2975924224
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.982827796501364e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3512790321916524,
      "tokens_seen": 2976055296
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.98202535708554e-05,
      "loss": 1.2996,
      "theoretical_loss": 3.3512677622947264,
      "tokens_seen": 2976186368
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.981222917669716e-05,
      "loss": 1.2159,
      "theoretical_loss": 3.3512564930330835,
      "tokens_seen": 2976317440
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9804204782538917e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.35124522440666,
      "tokens_seen": 2976448512
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9796180388380677e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.3512339564153923,
      "tokens_seen": 2976579584
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9788155994222437e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.3512226890592167,
      "tokens_seen": 2976710656
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9780131600064197e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.3512114223380696,
      "tokens_seen": 2976841728
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9772107205905953e-05,
      "loss": 1.228,
      "theoretical_loss": 3.3512001562518874,
      "tokens_seen": 2976972800
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9764082811747713e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.3511888908006053,
      "tokens_seen": 2977103872
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9756058417589473e-05,
      "loss": 1.1916,
      "theoretical_loss": 3.3511776259841604,
      "tokens_seen": 2977234944
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.974803402343123e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.3511663618024885,
      "tokens_seen": 2977366016
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.974000962927299e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.3511550982555267,
      "tokens_seen": 2977497088
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.973198523511475e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.3511438353432106,
      "tokens_seen": 2977628160
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.972396084095651e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.351132573065477,
      "tokens_seen": 2977759232
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9715936446798266e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.3511213114222618,
      "tokens_seen": 2977890304
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9707912052640026e-05,
      "loss": 1.2211,
      "theoretical_loss": 3.351110050413501,
      "tokens_seen": 2978021376
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.9699887658481786e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.3510987900391322,
      "tokens_seen": 2978152448
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9691863264323545e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.3510875302990906,
      "tokens_seen": 2978283520
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9683838870165302e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.3510762711933126,
      "tokens_seen": 2978414592
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9675814476007062e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.3510650127217354,
      "tokens_seen": 2978545664
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9667790081848822e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.3510537548842945,
      "tokens_seen": 2978676736
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.965976568769058e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.3510424976809263,
      "tokens_seen": 2978807808
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9651741293532338e-05,
      "loss": 1.2026,
      "theoretical_loss": 3.351031241111568,
      "tokens_seen": 2978938880
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9643716899374098e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.3510199851761553,
      "tokens_seen": 2979069952
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9635692505215858e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.351008729874625,
      "tokens_seen": 2979201024
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9627668111057615e-05,
      "loss": 1.3156,
      "theoretical_loss": 3.350997475206913,
      "tokens_seen": 2979332096
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9619643716899374e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.350986221172956,
      "tokens_seen": 2979463168
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9611619322741134e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.3509749677726908,
      "tokens_seen": 2979594240
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9603594928582894e-05,
      "loss": 1.284,
      "theoretical_loss": 3.3509637150060536,
      "tokens_seen": 2979725312
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.959557053442465e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.3509524628729803,
      "tokens_seen": 2979856384
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.958754614026641e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.3509412113734083,
      "tokens_seen": 2979987456
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.957952174610817e-05,
      "loss": 1.311,
      "theoretical_loss": 3.3509299605072735,
      "tokens_seen": 2980118528
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.957149735194993e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.350918710274512,
      "tokens_seen": 2980249600
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.9895803332328796,
      "objective/train/avg_token_score": 0.010419688187539577,
      "objective/train/docs_used": 1083727,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3000433444976807,
      "objective/train/original_loss": 1.3151015043258667,
      "objective/train/theoretical_loss": 3.3509074606750615,
      "objective/train/tokens_used": 1350839776,
      "objective/train/unlikelihood_loss": 0.02682066708803177,
      "theoretical_loss": 3.3509074606750615,
      "tokens_seen": 2980380672
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9563472957791687e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.3509074606750615,
      "tokens_seen": 2980380672
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9555448563633447e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.3508962117088577,
      "tokens_seen": 2980511744
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9547424169475207e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.350884963375837,
      "tokens_seen": 2980642816
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9539399775316963e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.3508737156759363,
      "tokens_seen": 2980773888
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9531375381158723e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.3508624686090913,
      "tokens_seen": 2980904960
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9523350987000483e-05,
      "loss": 1.2222,
      "theoretical_loss": 3.3508512221752396,
      "tokens_seen": 2981036032
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9515326592842243e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.3508399763743175,
      "tokens_seen": 2981167104
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9507302198684e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.3508287312062612,
      "tokens_seen": 2981298176
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.949927780452576e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.3508174866710077,
      "tokens_seen": 2981429248
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.949125341036752e-05,
      "loss": 1.313,
      "theoretical_loss": 3.350806242768493,
      "tokens_seen": 2981560320
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.948322901620928e-05,
      "loss": 1.2536,
      "theoretical_loss": 3.350794999498654,
      "tokens_seen": 2981691392
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9475204622051036e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.350783756861427,
      "tokens_seen": 2981822464
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9467180227892796e-05,
      "loss": 1.2445,
      "theoretical_loss": 3.3507725148567493,
      "tokens_seen": 2981953536
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9459155833734556e-05,
      "loss": 1.3524,
      "theoretical_loss": 3.350761273484557,
      "tokens_seen": 2982084608
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9451131439576312e-05,
      "loss": 1.3203,
      "theoretical_loss": 3.350750032744787,
      "tokens_seen": 2982215680
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9443107045418072e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.3507387926373755,
      "tokens_seen": 2982346752
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9435082651259832e-05,
      "loss": 1.202,
      "theoretical_loss": 3.3507275531622596,
      "tokens_seen": 2982477824
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9427058257101592e-05,
      "loss": 1.2615,
      "theoretical_loss": 3.350716314319375,
      "tokens_seen": 2982608896
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.941903386294335e-05,
      "loss": 1.3395,
      "theoretical_loss": 3.3507050761086603,
      "tokens_seen": 2982739968
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9411009468785108e-05,
      "loss": 1.2527,
      "theoretical_loss": 3.35069383853005,
      "tokens_seen": 2982871040
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9402985074626868e-05,
      "loss": 1.2467,
      "theoretical_loss": 3.350682601583482,
      "tokens_seen": 2983002112
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9394960680468628e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.3506713652688926,
      "tokens_seen": 2983133184
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9386936286310385e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.3506601295862186,
      "tokens_seen": 2983264256
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9378911892152144e-05,
      "loss": 1.2399,
      "theoretical_loss": 3.350648894535397,
      "tokens_seen": 2983395328
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9370887497993904e-05,
      "loss": 1.1985,
      "theoretical_loss": 3.350637660116364,
      "tokens_seen": 2983526400
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.936286310383566e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.3506264263290566,
      "tokens_seen": 2983657472
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.935483870967742e-05,
      "loss": 1.234,
      "theoretical_loss": 3.350615193173411,
      "tokens_seen": 2983788544
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.934681431551918e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.350603960649365,
      "tokens_seen": 2983919616
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.933878992136094e-05,
      "loss": 1.2191,
      "theoretical_loss": 3.350592728756854,
      "tokens_seen": 2984050688
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9330765527202697e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.350581497495816,
      "tokens_seen": 2984181760
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9322741133044457e-05,
      "loss": 1.2485,
      "theoretical_loss": 3.350570266866187,
      "tokens_seen": 2984312832
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9314716738886217e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.3505590368679044,
      "tokens_seen": 2984443904
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9306692344727973e-05,
      "loss": 1.2698,
      "theoretical_loss": 3.3505478075009045,
      "tokens_seen": 2984574976
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9298667950569733e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.350536578765124,
      "tokens_seen": 2984706048
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9290643556411493e-05,
      "loss": 1.2169,
      "theoretical_loss": 3.3505253506605,
      "tokens_seen": 2984837120
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.928261916225325e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.350514123186969,
      "tokens_seen": 2984968192
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.927459476809501e-05,
      "loss": 1.165,
      "theoretical_loss": 3.3505028963444685,
      "tokens_seen": 2985099264
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.926657037393677e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.3504916701329344,
      "tokens_seen": 2985230336
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9258545979778526e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.350480444552304,
      "tokens_seen": 2985361408
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9250521585620286e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.3504692196025143,
      "tokens_seen": 2985492480
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9242497191462046e-05,
      "loss": 1.2498,
      "theoretical_loss": 3.3504579952835023,
      "tokens_seen": 2985623552
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9234472797303802e-05,
      "loss": 1.1689,
      "theoretical_loss": 3.350446771595204,
      "tokens_seen": 2985754624
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9226448403145562e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.350435548537557,
      "tokens_seen": 2985885696
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9218424008987322e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.350424326110498,
      "tokens_seen": 2986016768
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.921039961482908e-05,
      "loss": 1.2179,
      "theoretical_loss": 3.350413104313964,
      "tokens_seen": 2986147840
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.920237522067084e-05,
      "loss": 1.293,
      "theoretical_loss": 3.350401883147892,
      "tokens_seen": 2986278912
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.91943508265126e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3503906626122184,
      "tokens_seen": 2986409984
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9186326432354355e-05,
      "loss": 1.2049,
      "theoretical_loss": 3.3503794427068803,
      "tokens_seen": 2986541056
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9178302038196115e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.3503682234318153,
      "tokens_seen": 2986672128
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9170277644037875e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3503570047869595,
      "tokens_seen": 2986803200
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.9937630891799927,
      "objective/train/avg_token_score": 0.00623689778149128,
      "objective/train/docs_used": 1086010,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.337925910949707,
      "objective/train/original_loss": 1.3462193012237549,
      "objective/train/theoretical_loss": 3.35034578677225,
      "objective/train/tokens_used": 1357393376,
      "objective/train/unlikelihood_loss": 0.01813862845301628,
      "theoretical_loss": 3.35034578677225,
      "tokens_seen": 2986934272
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9162253249879635e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.35034578677225,
      "tokens_seen": 2986934272
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.915422885572139e-05,
      "loss": 1.249,
      "theoretical_loss": 3.3503345693876243,
      "tokens_seen": 2987065344
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.914620446156315e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.3503233526330183,
      "tokens_seen": 2987196416
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.913818006740491e-05,
      "loss": 1.269,
      "theoretical_loss": 3.35031213650837,
      "tokens_seen": 2987327488
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.913015567324667e-05,
      "loss": 1.2265,
      "theoretical_loss": 3.3503009210136163,
      "tokens_seen": 2987458560
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9122131279088428e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.3502897061486934,
      "tokens_seen": 2987589632
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9114106884930188e-05,
      "loss": 1.2698,
      "theoretical_loss": 3.3502784919135395,
      "tokens_seen": 2987720704
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9106082490771947e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3502672783080905,
      "tokens_seen": 2987851776
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9098058096613707e-05,
      "loss": 1.2522,
      "theoretical_loss": 3.3502560653322835,
      "tokens_seen": 2987982848
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9090033702455464e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.3502448529860565,
      "tokens_seen": 2988113920
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9082009308297224e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.3502336412693454,
      "tokens_seen": 2988244992
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9073984914138984e-05,
      "loss": 1.3255,
      "theoretical_loss": 3.350222430182088,
      "tokens_seen": 2988376064
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.906596051998074e-05,
      "loss": 1.1727,
      "theoretical_loss": 3.3502112197242213,
      "tokens_seen": 2988507136
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.90579361258225e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.3502000098956817,
      "tokens_seen": 2988638208
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.904991173166426e-05,
      "loss": 1.2477,
      "theoretical_loss": 3.350188800696407,
      "tokens_seen": 2988769280
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.904188733750602e-05,
      "loss": 1.2879,
      "theoretical_loss": 3.350177592126334,
      "tokens_seen": 2988900352
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9033862943347776e-05,
      "loss": 1.2128,
      "theoretical_loss": 3.3501663841854,
      "tokens_seen": 2989031424
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9025838549189536e-05,
      "loss": 1.235,
      "theoretical_loss": 3.3501551768735416,
      "tokens_seen": 2989162496
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9017814155031296e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.3501439701906968,
      "tokens_seen": 2989293568
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9009789760873056e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.3501327641368013,
      "tokens_seen": 2989424640
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.9001765366714813e-05,
      "loss": 1.1954,
      "theoretical_loss": 3.350121558711794,
      "tokens_seen": 2989555712
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8993740972556573e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.3501103539156105,
      "tokens_seen": 2989686784
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8985716578398332e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.3500991497481887,
      "tokens_seen": 2989817856
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.897769218424009e-05,
      "loss": 1.2321,
      "theoretical_loss": 3.3500879462094657,
      "tokens_seen": 2989948928
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.896966779008185e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.350076743299378,
      "tokens_seen": 2990080000
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.896164339592361e-05,
      "loss": 1.3319,
      "theoretical_loss": 3.350065541017864,
      "tokens_seen": 2990211072
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.895361900176537e-05,
      "loss": 1.2546,
      "theoretical_loss": 3.35005433936486,
      "tokens_seen": 2990342144
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8945594607607125e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.3500431383403035,
      "tokens_seen": 2990473216
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8937570213448885e-05,
      "loss": 1.1814,
      "theoretical_loss": 3.3500319379441317,
      "tokens_seen": 2990604288
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8929545819290645e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.3500207381762817,
      "tokens_seen": 2990735360
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8921521425132405e-05,
      "loss": 1.3731,
      "theoretical_loss": 3.3500095390366904,
      "tokens_seen": 2990866432
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.891349703097416e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.3499983405252958,
      "tokens_seen": 2990997504
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.890547263681592e-05,
      "loss": 1.2561,
      "theoretical_loss": 3.3499871426420347,
      "tokens_seen": 2991128576
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.889744824265768e-05,
      "loss": 1.1406,
      "theoretical_loss": 3.349975945386844,
      "tokens_seen": 2991259648
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8889423848499438e-05,
      "loss": 1.238,
      "theoretical_loss": 3.3499647487596613,
      "tokens_seen": 2991390720
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8881399454341198e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.3499535527604243,
      "tokens_seen": 2991521792
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8873375060182958e-05,
      "loss": 1.2042,
      "theoretical_loss": 3.3499423573890694,
      "tokens_seen": 2991652864
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8865350666024717e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.3499311626455346,
      "tokens_seen": 2991783936
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8857326271866474e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.3499199685297567,
      "tokens_seen": 2991915008
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8849301877708234e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.3499087750416736,
      "tokens_seen": 2992046080
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8841277483549994e-05,
      "loss": 1.1977,
      "theoretical_loss": 3.349897582181222,
      "tokens_seen": 2992177152
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8833253089391754e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.349886389948339,
      "tokens_seen": 2992308224
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.882522869523351e-05,
      "loss": 1.2287,
      "theoretical_loss": 3.3498751983429633,
      "tokens_seen": 2992439296
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.881720430107527e-05,
      "loss": 1.1813,
      "theoretical_loss": 3.3498640073650305,
      "tokens_seen": 2992570368
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.880917990691703e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.349852817014479,
      "tokens_seen": 2992701440
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8801155512758787e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.349841627291246,
      "tokens_seen": 2992832512
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8793131118600546e-05,
      "loss": 1.1734,
      "theoretical_loss": 3.3498304381952684,
      "tokens_seen": 2992963584
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8785106724442306e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.3498192497264845,
      "tokens_seen": 2993094656
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8777082330284066e-05,
      "loss": 1.2245,
      "theoretical_loss": 3.3498080618848305,
      "tokens_seen": 2993225728
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8769057936125823e-05,
      "loss": 1.2187,
      "theoretical_loss": 3.349796874670245,
      "tokens_seen": 2993356800
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.9904163479804993,
      "objective/train/avg_token_score": 0.00958362314850092,
      "objective/train/docs_used": 1088408,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2993043661117554,
      "objective/train/original_loss": 1.3121954202651978,
      "objective/train/theoretical_loss": 3.3497856880826644,
      "objective/train/tokens_used": 1363946976,
      "objective/train/unlikelihood_loss": 0.029201652854681015,
      "theoretical_loss": 3.3497856880826644,
      "tokens_seen": 2993487872
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8761033541967583e-05,
      "loss": 1.3761,
      "theoretical_loss": 3.3497856880826644,
      "tokens_seen": 2993487872
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8753009147809343e-05,
      "loss": 1.3182,
      "theoretical_loss": 3.3497745021220267,
      "tokens_seen": 2993618944
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8744984753651103e-05,
      "loss": 1.3317,
      "theoretical_loss": 3.3497633167882688,
      "tokens_seen": 2993750016
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.873696035949286e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.3497521320813286,
      "tokens_seen": 2993881088
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.872893596533462e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.3497409480011435,
      "tokens_seen": 2994012160
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.872091157117638e-05,
      "loss": 1.2583,
      "theoretical_loss": 3.3497297645476505,
      "tokens_seen": 2994143232
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8712887177018135e-05,
      "loss": 1.348,
      "theoretical_loss": 3.3497185817207877,
      "tokens_seen": 2994274304
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8704862782859895e-05,
      "loss": 1.1662,
      "theoretical_loss": 3.3497073995204922,
      "tokens_seen": 2994405376
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8696838388701655e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.3496962179467014,
      "tokens_seen": 2994536448
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.8688813994543415e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.349685036999353,
      "tokens_seen": 2994667520
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.868078960038517e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.3496738566783844,
      "tokens_seen": 2994798592
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.867276520622693e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.3496626769837325,
      "tokens_seen": 2994929664
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.866474081206869e-05,
      "loss": 1.3694,
      "theoretical_loss": 3.3496514979153362,
      "tokens_seen": 2995060736
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.865671641791045e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.3496403194731315,
      "tokens_seen": 2995191808
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8648692023752208e-05,
      "loss": 1.336,
      "theoretical_loss": 3.349629141657057,
      "tokens_seen": 2995322880
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8640667629593968e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.3496179644670496,
      "tokens_seen": 2995453952
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8632643235435728e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.3496067879030473,
      "tokens_seen": 2995585024
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8624618841277484e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.3495956119649875,
      "tokens_seen": 2995716096
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8616594447119244e-05,
      "loss": 1.3018,
      "theoretical_loss": 3.3495844366528074,
      "tokens_seen": 2995847168
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8608570052961004e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.3495732619664453,
      "tokens_seen": 2995978240
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.860054565880276e-05,
      "loss": 1.2677,
      "theoretical_loss": 3.349562087905838,
      "tokens_seen": 2996109312
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.859252126464452e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.3495509144709232,
      "tokens_seen": 2996240384
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.858449687048628e-05,
      "loss": 1.2127,
      "theoretical_loss": 3.3495397416616393,
      "tokens_seen": 2996371456
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8576472476328037e-05,
      "loss": 1.2705,
      "theoretical_loss": 3.349528569477923,
      "tokens_seen": 2996502528
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8568448082169797e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.349517397919712,
      "tokens_seen": 2996633600
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8560423688011557e-05,
      "loss": 1.2619,
      "theoretical_loss": 3.3495062269869447,
      "tokens_seen": 2996764672
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8552399293853313e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.3494950566795576,
      "tokens_seen": 2996895744
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8544374899695073e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.3494838869974894,
      "tokens_seen": 2997026816
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8536350505536833e-05,
      "loss": 1.2414,
      "theoretical_loss": 3.349472717940677,
      "tokens_seen": 2997157888
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.852832611137859e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.3494615495090585,
      "tokens_seen": 2997288960
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.852030171722035e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.3494503817025714,
      "tokens_seen": 2997420032
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.851227732306211e-05,
      "loss": 1.2808,
      "theoretical_loss": 3.349439214521153,
      "tokens_seen": 2997551104
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8504252928903866e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.349428047964742,
      "tokens_seen": 2997682176
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8496228534745626e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.349416882033275,
      "tokens_seen": 2997813248
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8488204140587386e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.3494057167266904,
      "tokens_seen": 2997944320
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8480179746429146e-05,
      "loss": 1.2599,
      "theoretical_loss": 3.3493945520449255,
      "tokens_seen": 2998075392
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8472155352270902e-05,
      "loss": 1.1726,
      "theoretical_loss": 3.3493833879879182,
      "tokens_seen": 2998206464
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8464130958112662e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.3493722245556063,
      "tokens_seen": 2998337536
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8456106563954422e-05,
      "loss": 1.3327,
      "theoretical_loss": 3.3493610617479277,
      "tokens_seen": 2998468608
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8448082169796182e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.3493498995648197,
      "tokens_seen": 2998599680
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8440057775637938e-05,
      "loss": 1.2882,
      "theoretical_loss": 3.34933873800622,
      "tokens_seen": 2998730752
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8432033381479698e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.349327577072067,
      "tokens_seen": 2998861824
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8424008987321458e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.349316416762298,
      "tokens_seen": 2998992896
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8415984593163215e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3493052570768507,
      "tokens_seen": 2999123968
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8407960199004975e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.3492940980156636,
      "tokens_seen": 2999255040
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8399935804846734e-05,
      "loss": 1.2334,
      "theoretical_loss": 3.3492829395786736,
      "tokens_seen": 2999386112
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8391911410688494e-05,
      "loss": 1.218,
      "theoretical_loss": 3.3492717817658186,
      "tokens_seen": 2999517184
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.838388701653025e-05,
      "loss": 1.1868,
      "theoretical_loss": 3.3492606245770373,
      "tokens_seen": 2999648256
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.837586262237201e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.3492494680122666,
      "tokens_seen": 2999779328
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.836783822821377e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.349238312071445,
      "tokens_seen": 2999910400
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.9921075701713562,
      "objective/train/avg_token_score": 0.007892429828643799,
      "objective/train/docs_used": 1090669,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4129098653793335,
      "objective/train/original_loss": 1.4229018688201904,
      "objective/train/theoretical_loss": 3.3492271567545093,
      "objective/train/tokens_used": 1370500576,
      "objective/train/unlikelihood_loss": 0.023720279335975647,
      "theoretical_loss": 3.3492271567545093,
      "tokens_seen": 3000041472
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.835981383405553e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.3492271567545093,
      "tokens_seen": 3000041472
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8351789439897287e-05,
      "loss": 1.1929,
      "theoretical_loss": 3.3492160020613984,
      "tokens_seen": 3000172544
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8343765045739047e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.34920484799205,
      "tokens_seen": 3000303616
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8335740651580807e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.3491936945464023,
      "tokens_seen": 3000434688
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8327716257422563e-05,
      "loss": 1.3173,
      "theoretical_loss": 3.349182541724392,
      "tokens_seen": 3000565760
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8319691863264323e-05,
      "loss": 1.245,
      "theoretical_loss": 3.3491713895259574,
      "tokens_seen": 3000696832
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8311667469106083e-05,
      "loss": 1.1735,
      "theoretical_loss": 3.3491602379510375,
      "tokens_seen": 3000827904
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8303643074947843e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.349149086999569,
      "tokens_seen": 3000958976
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.82956186807896e-05,
      "loss": 1.3859,
      "theoretical_loss": 3.34913793667149,
      "tokens_seen": 3001090048
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.828759428663136e-05,
      "loss": 1.1937,
      "theoretical_loss": 3.349126786966739,
      "tokens_seen": 3001221120
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.827956989247312e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.3491156378852533,
      "tokens_seen": 3001352192
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.827154549831488e-05,
      "loss": 1.1617,
      "theoretical_loss": 3.349104489426971,
      "tokens_seen": 3001483264
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8263521104156636e-05,
      "loss": 1.1943,
      "theoretical_loss": 3.3490933415918303,
      "tokens_seen": 3001614336
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8255496709998396e-05,
      "loss": 1.2,
      "theoretical_loss": 3.3490821943797693,
      "tokens_seen": 3001745408
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8247472315840156e-05,
      "loss": 1.3096,
      "theoretical_loss": 3.3490710477907255,
      "tokens_seen": 3001876480
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8239447921681912e-05,
      "loss": 1.2921,
      "theoretical_loss": 3.3490599018246368,
      "tokens_seen": 3002007552
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8231423527523672e-05,
      "loss": 1.3062,
      "theoretical_loss": 3.3490487564814417,
      "tokens_seen": 3002138624
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8223399133365432e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.3490376117610783,
      "tokens_seen": 3002269696
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8215374739207192e-05,
      "loss": 1.2663,
      "theoretical_loss": 3.3490264676634833,
      "tokens_seen": 3002400768
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.820735034504895e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.3490153241885965,
      "tokens_seen": 3002531840
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.819932595089071e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.3490041813363547,
      "tokens_seen": 3002662912
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8191301556732468e-05,
      "loss": 1.3299,
      "theoretical_loss": 3.3489930391066967,
      "tokens_seen": 3002793984
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8183277162574228e-05,
      "loss": 1.2787,
      "theoretical_loss": 3.34898189749956,
      "tokens_seen": 3002925056
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8175252768415985e-05,
      "loss": 1.215,
      "theoretical_loss": 3.3489707565148827,
      "tokens_seen": 3003056128
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8167228374257745e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.348959616152603,
      "tokens_seen": 3003187200
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8159203980099505e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3489484764126587,
      "tokens_seen": 3003318272
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8151179585941264e-05,
      "loss": 1.3603,
      "theoretical_loss": 3.3489373372949887,
      "tokens_seen": 3003449344
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.814315519178302e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3489261987995302,
      "tokens_seen": 3003580416
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.813513079762478e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.3489150609262217,
      "tokens_seen": 3003711488
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.812710640346654e-05,
      "loss": 1.2135,
      "theoretical_loss": 3.348903923675001,
      "tokens_seen": 3003842560
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8119082009308297e-05,
      "loss": 1.2803,
      "theoretical_loss": 3.3488927870458065,
      "tokens_seen": 3003973632
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8111057615150057e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.3488816510385764,
      "tokens_seen": 3004104704
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8103033220991817e-05,
      "loss": 1.2214,
      "theoretical_loss": 3.3488705156532483,
      "tokens_seen": 3004235776
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8095008826833577e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.348859380889761,
      "tokens_seen": 3004366848
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8086984432675334e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.3488482467480525,
      "tokens_seen": 3004497920
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8078960038517093e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.3488371132280603,
      "tokens_seen": 3004628992
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8070935644358853e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.3488259803297233,
      "tokens_seen": 3004760064
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8062911250200613e-05,
      "loss": 1.2154,
      "theoretical_loss": 3.34881484805298,
      "tokens_seen": 3004891136
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.805488685604237e-05,
      "loss": 1.2082,
      "theoretical_loss": 3.348803716397767,
      "tokens_seen": 3005022208
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.804686246188413e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.3487925853640244,
      "tokens_seen": 3005153280
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.803883806772589e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.348781454951689,
      "tokens_seen": 3005284352
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8030813673567646e-05,
      "loss": 1.2537,
      "theoretical_loss": 3.3487703251606993,
      "tokens_seen": 3005415424
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8022789279409406e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.3487591959909944,
      "tokens_seen": 3005546496
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8014764885251166e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.3487480674425116,
      "tokens_seen": 3005677568
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.8006740491092926e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.3487369395151894,
      "tokens_seen": 3005808640
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7998716096934682e-05,
      "loss": 1.1925,
      "theoretical_loss": 3.348725812208966,
      "tokens_seen": 3005939712
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7990691702776442e-05,
      "loss": 1.1781,
      "theoretical_loss": 3.3487146855237797,
      "tokens_seen": 3006070784
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7982667308618202e-05,
      "loss": 1.1788,
      "theoretical_loss": 3.348703559459569,
      "tokens_seen": 3006201856
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7974642914459962e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.3486924340162716,
      "tokens_seen": 3006332928
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.796661852030172e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.3486813091938266,
      "tokens_seen": 3006464000
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.9927536249160767,
      "objective/train/avg_token_score": 0.007246372289955616,
      "objective/train/docs_used": 1093144,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.273648738861084,
      "objective/train/original_loss": 1.2825168371200562,
      "objective/train/theoretical_loss": 3.348670184992171,
      "objective/train/tokens_used": 1377054176,
      "objective/train/unlikelihood_loss": 0.02656622789800167,
      "theoretical_loss": 3.348670184992171,
      "tokens_seen": 3006595072
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.795859412614348e-05,
      "loss": 1.272,
      "theoretical_loss": 3.348670184992171,
      "tokens_seen": 3006595072
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.795056973198524e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.3486590614112446,
      "tokens_seen": 3006726144
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7942545337826995e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.348647938450985,
      "tokens_seen": 3006857216
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7934520943668755e-05,
      "loss": 1.2124,
      "theoretical_loss": 3.3486368161113305,
      "tokens_seen": 3006988288
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7926496549510515e-05,
      "loss": 1.2131,
      "theoretical_loss": 3.3486256943922195,
      "tokens_seen": 3007119360
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.791847215535227e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.34861457329359,
      "tokens_seen": 3007250432
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.791044776119403e-05,
      "loss": 1.1774,
      "theoretical_loss": 3.3486034528153805,
      "tokens_seen": 3007381504
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.790242336703579e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.34859233295753,
      "tokens_seen": 3007512576
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7894398972877548e-05,
      "loss": 1.3037,
      "theoretical_loss": 3.3485812137199757,
      "tokens_seen": 3007643648
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7886374578719307e-05,
      "loss": 1.2027,
      "theoretical_loss": 3.348570095102657,
      "tokens_seen": 3007774720
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7878350184561067e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.3485589771055118,
      "tokens_seen": 3007905792
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7870325790402824e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.3485478597284786,
      "tokens_seen": 3008036864
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7862301396244584e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.348536742971496,
      "tokens_seen": 3008167936
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7854277002086344e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.348525626834502,
      "tokens_seen": 3008299008
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.78462526079281e-05,
      "loss": 1.1881,
      "theoretical_loss": 3.348514511317435,
      "tokens_seen": 3008430080
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.783822821376986e-05,
      "loss": 1.2163,
      "theoretical_loss": 3.3485033964202335,
      "tokens_seen": 3008561152
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.783020381961162e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.3484922821428365,
      "tokens_seen": 3008692224
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7822179425453377e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.3484811684851814,
      "tokens_seen": 3008823296
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7814155031295136e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.3484700554472075,
      "tokens_seen": 3008954368
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7806130637136896e-05,
      "loss": 1.1953,
      "theoretical_loss": 3.348458943028853,
      "tokens_seen": 3009085440
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7798106242978656e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.3484478312300565,
      "tokens_seen": 3009216512
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7790081848820413e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.3484367200507563,
      "tokens_seen": 3009347584
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7782057454662173e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.3484256094908904,
      "tokens_seen": 3009478656
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7774033060503933e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.3484144995503984,
      "tokens_seen": 3009609728
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.776600866634569e-05,
      "loss": 1.2449,
      "theoretical_loss": 3.3484033902292176,
      "tokens_seen": 3009740800
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.775798427218745e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.3483922815272873,
      "tokens_seen": 3009871872
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.774995987802921e-05,
      "loss": 1.2497,
      "theoretical_loss": 3.3483811734445457,
      "tokens_seen": 3010002944
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.774193548387097e-05,
      "loss": 1.2025,
      "theoretical_loss": 3.348370065980932,
      "tokens_seen": 3010134016
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7733911089712725e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.348358959136384,
      "tokens_seen": 3010265088
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7725886695554485e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3483478529108397,
      "tokens_seen": 3010396160
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7717862301396245e-05,
      "loss": 1.2321,
      "theoretical_loss": 3.348336747304239,
      "tokens_seen": 3010527232
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.7709837907238005e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.3483256423165195,
      "tokens_seen": 3010658304
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.770181351307976e-05,
      "loss": 1.2403,
      "theoretical_loss": 3.3483145379476205,
      "tokens_seen": 3010789376
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.769378911892152e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3483034341974798,
      "tokens_seen": 3010920448
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.768576472476328e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.3482923310660366,
      "tokens_seen": 3011051520
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.767774033060504e-05,
      "loss": 1.2153,
      "theoretical_loss": 3.3482812285532293,
      "tokens_seen": 3011182592
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7669715936446798e-05,
      "loss": 1.2008,
      "theoretical_loss": 3.3482701266589965,
      "tokens_seen": 3011313664
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7661691542288558e-05,
      "loss": 1.2658,
      "theoretical_loss": 3.3482590253832765,
      "tokens_seen": 3011444736
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7653667148130318e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.3482479247260084,
      "tokens_seen": 3011575808
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7645642753972074e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.34823682468713,
      "tokens_seen": 3011706880
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7637618359813834e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.3482257252665817,
      "tokens_seen": 3011837952
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7629593965655594e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.3482146264643005,
      "tokens_seen": 3011969024
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7621569571497354e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.348203528280225,
      "tokens_seen": 3012100096
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.761354517733911e-05,
      "loss": 1.2118,
      "theoretical_loss": 3.3481924307142954,
      "tokens_seen": 3012231168
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.760552078318087e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.348181333766449,
      "tokens_seen": 3012362240
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.759749638902263e-05,
      "loss": 1.1918,
      "theoretical_loss": 3.3481702374366247,
      "tokens_seen": 3012493312
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.758947199486439e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.3481591417247616,
      "tokens_seen": 3012624384
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7581447600706147e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.3481480466307985,
      "tokens_seen": 3012755456
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7573423206547907e-05,
      "loss": 1.2631,
      "theoretical_loss": 3.3481369521546736,
      "tokens_seen": 3012886528
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7565398812389666e-05,
      "loss": 1.3175,
      "theoretical_loss": 3.3481258582963256,
      "tokens_seen": 3013017600
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.9914937615394592,
      "objective/train/avg_token_score": 0.008506190031766891,
      "objective/train/docs_used": 1095529,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.267470121383667,
      "objective/train/original_loss": 1.2790398597717285,
      "objective/train/theoretical_loss": 3.3481147650556937,
      "objective/train/tokens_used": 1383607776,
      "objective/train/unlikelihood_loss": 0.02353137731552124,
      "theoretical_loss": 3.3481147650556937,
      "tokens_seen": 3013148672
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7557374418231423e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.3481147650556937,
      "tokens_seen": 3013148672
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7549350024073183e-05,
      "loss": 1.2779,
      "theoretical_loss": 3.3481036724327167,
      "tokens_seen": 3013279744
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7541325629914943e-05,
      "loss": 1.2228,
      "theoretical_loss": 3.3480925804273323,
      "tokens_seen": 3013410816
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7533301235756703e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.3480814890394806,
      "tokens_seen": 3013541888
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.752527684159846e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.3480703982690994,
      "tokens_seen": 3013672960
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.751725244744022e-05,
      "loss": 1.2483,
      "theoretical_loss": 3.348059308116128,
      "tokens_seen": 3013804032
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.750922805328198e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.348048218580505,
      "tokens_seen": 3013935104
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.750120365912374e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.348037129662169,
      "tokens_seen": 3014066176
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7493179264965495e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.3480260413610594,
      "tokens_seen": 3014197248
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7485154870807255e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.348014953677114,
      "tokens_seen": 3014328320
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7477130476649015e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.3480038666102727,
      "tokens_seen": 3014459392
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7469106082490772e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.347992780160474,
      "tokens_seen": 3014590464
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.746108168833253e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3479816943276566,
      "tokens_seen": 3014721536
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.745305729417429e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.347970609111759,
      "tokens_seen": 3014852608
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.744503290001605e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3479595245127203,
      "tokens_seen": 3014983680
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7437008505857808e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.3479484405304794,
      "tokens_seen": 3015114752
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7428984111699568e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3479373571649753,
      "tokens_seen": 3015245824
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7420959717541328e-05,
      "loss": 1.2406,
      "theoretical_loss": 3.347926274416147,
      "tokens_seen": 3015376896
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7412935323383088e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3479151922839328,
      "tokens_seen": 3015507968
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7404910929224844e-05,
      "loss": 1.3406,
      "theoretical_loss": 3.347904110768272,
      "tokens_seen": 3015639040
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7396886535066604e-05,
      "loss": 1.1709,
      "theoretical_loss": 3.3478930298691036,
      "tokens_seen": 3015770112
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7388862140908364e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.3478819495863665,
      "tokens_seen": 3015901184
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.738083774675012e-05,
      "loss": 1.1974,
      "theoretical_loss": 3.3478708699199995,
      "tokens_seen": 3016032256
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.737281335259188e-05,
      "loss": 1.2005,
      "theoretical_loss": 3.3478597908699412,
      "tokens_seen": 3016163328
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.736478895843364e-05,
      "loss": 1.2156,
      "theoretical_loss": 3.347848712436131,
      "tokens_seen": 3016294400
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.73567645642754e-05,
      "loss": 1.3393,
      "theoretical_loss": 3.3478376346185077,
      "tokens_seen": 3016425472
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7348740170117157e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.3478265574170103,
      "tokens_seen": 3016556544
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7340715775958917e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.3478154808315774,
      "tokens_seen": 3016687616
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7332691381800677e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3478044048621487,
      "tokens_seen": 3016818688
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7324666987642436e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.3477933295086624,
      "tokens_seen": 3016949760
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7316642593484193e-05,
      "loss": 1.2558,
      "theoretical_loss": 3.347782254771058,
      "tokens_seen": 3017080832
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7308618199325953e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.3477711806492745,
      "tokens_seen": 3017211904
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7300593805167713e-05,
      "loss": 1.2243,
      "theoretical_loss": 3.347760107143251,
      "tokens_seen": 3017342976
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.729256941100947e-05,
      "loss": 1.1562,
      "theoretical_loss": 3.3477490342529257,
      "tokens_seen": 3017474048
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.728454501685123e-05,
      "loss": 1.2363,
      "theoretical_loss": 3.3477379619782384,
      "tokens_seen": 3017605120
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.727652062269299e-05,
      "loss": 1.2972,
      "theoretical_loss": 3.347726890319128,
      "tokens_seen": 3017736192
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7268496228534746e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.3477158192755336,
      "tokens_seen": 3017867264
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7260471834376506e-05,
      "loss": 1.1757,
      "theoretical_loss": 3.3477047488473937,
      "tokens_seen": 3017998336
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7252447440218265e-05,
      "loss": 1.2528,
      "theoretical_loss": 3.347693679034648,
      "tokens_seen": 3018129408
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7244423046060022e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3476826098372356,
      "tokens_seen": 3018260480
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7236398651901782e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.347671541255095,
      "tokens_seen": 3018391552
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7228374257743542e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.347660473288166,
      "tokens_seen": 3018522624
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.72203498635853e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.347649405936387,
      "tokens_seen": 3018653696
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7212325469427058e-05,
      "loss": 1.1924,
      "theoretical_loss": 3.347638339199697,
      "tokens_seen": 3018784768
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7204301075268818e-05,
      "loss": 1.2416,
      "theoretical_loss": 3.3476272730780363,
      "tokens_seen": 3018915840
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7196276681110575e-05,
      "loss": 1.3309,
      "theoretical_loss": 3.3476162075713427,
      "tokens_seen": 3019046912
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7188252286952335e-05,
      "loss": 1.2034,
      "theoretical_loss": 3.3476051426795563,
      "tokens_seen": 3019177984
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7180227892794094e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.3475940784026155,
      "tokens_seen": 3019309056
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.717220349863585e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3475830147404597,
      "tokens_seen": 3019440128
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.716417910447761e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.3475719516930287,
      "tokens_seen": 3019571200
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.9923089742660522,
      "objective/train/avg_token_score": 0.007690987084060907,
      "objective/train/docs_used": 1097674,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2523478269577026,
      "objective/train/original_loss": 1.2640082836151123,
      "objective/train/theoretical_loss": 3.34756088926026,
      "objective/train/tokens_used": 1390161376,
      "objective/train/unlikelihood_loss": 0.020240917801856995,
      "theoretical_loss": 3.34756088926026,
      "tokens_seen": 3019702272
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.715615471031937e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.34756088926026,
      "tokens_seen": 3019702272
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.714813031616113e-05,
      "loss": 1.308,
      "theoretical_loss": 3.3475498274420947,
      "tokens_seen": 3019833344
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7140105922002887e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.347538766238471,
      "tokens_seen": 3019964416
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7132081527844647e-05,
      "loss": 1.2276,
      "theoretical_loss": 3.347527705649328,
      "tokens_seen": 3020095488
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7124057133686407e-05,
      "loss": 1.2709,
      "theoretical_loss": 3.347516645674605,
      "tokens_seen": 3020226560
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7116032739528167e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3475055863142416,
      "tokens_seen": 3020357632
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7108008345369923e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.3474945275681764,
      "tokens_seen": 3020488704
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7099983951211683e-05,
      "loss": 1.2612,
      "theoretical_loss": 3.3474834694363493,
      "tokens_seen": 3020619776
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7091959557053443e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.3474724119186994,
      "tokens_seen": 3020750848
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.70839351628952e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.3474613550151653,
      "tokens_seen": 3020881920
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.707591076873696e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.3474502987256867,
      "tokens_seen": 3021012992
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.706788637457872e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.3474392430502027,
      "tokens_seen": 3021144064
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.705986198042048e-05,
      "loss": 1.2748,
      "theoretical_loss": 3.347428187988653,
      "tokens_seen": 3021275136
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7051837586262236e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.3474171335409766,
      "tokens_seen": 3021406208
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7043813192103996e-05,
      "loss": 1.2327,
      "theoretical_loss": 3.3474060797071123,
      "tokens_seen": 3021537280
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7035788797945756e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3473950264870003,
      "tokens_seen": 3021668352
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7027764403787516e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.3473839738805795,
      "tokens_seen": 3021799424
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7019740009629272e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.347372921887789,
      "tokens_seen": 3021930496
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7011715615471032e-05,
      "loss": 1.215,
      "theoretical_loss": 3.347361870508568,
      "tokens_seen": 3022061568
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.7003691221312792e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.3473508197428563,
      "tokens_seen": 3022192640
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.699566682715455e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.3473397695905933,
      "tokens_seen": 3022323712
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.698764243299631e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.347328720051718,
      "tokens_seen": 3022454784
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.697961803883807e-05,
      "loss": 1.2372,
      "theoretical_loss": 3.3473176711261696,
      "tokens_seen": 3022585856
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.697159364467983e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.347306622813888,
      "tokens_seen": 3022716928
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6963569250521585e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.347295575114812,
      "tokens_seen": 3022848000
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6955544856363345e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.347284528028881,
      "tokens_seen": 3022979072
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6947520462205105e-05,
      "loss": 1.2374,
      "theoretical_loss": 3.3472734815560354,
      "tokens_seen": 3023110144
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6939496068046865e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3472624356962135,
      "tokens_seen": 3023241216
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.693147167388862e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.347251390449355,
      "tokens_seen": 3023372288
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.692344727973038e-05,
      "loss": 1.2068,
      "theoretical_loss": 3.347240345815399,
      "tokens_seen": 3023503360
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.691542288557214e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.347229301794286,
      "tokens_seen": 3023634432
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6907398491413897e-05,
      "loss": 1.2487,
      "theoretical_loss": 3.347218258385954,
      "tokens_seen": 3023765504
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6899374097255657e-05,
      "loss": 1.2108,
      "theoretical_loss": 3.347207215590344,
      "tokens_seen": 3023896576
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6891349703097417e-05,
      "loss": 1.2106,
      "theoretical_loss": 3.347196173407394,
      "tokens_seen": 3024027648
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6883325308939177e-05,
      "loss": 1.2624,
      "theoretical_loss": 3.347185131837044,
      "tokens_seen": 3024158720
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6875300914780934e-05,
      "loss": 1.218,
      "theoretical_loss": 3.347174090879234,
      "tokens_seen": 3024289792
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6867276520622694e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.3471630505339025,
      "tokens_seen": 3024420864
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6859252126464453e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.3471520108009893,
      "tokens_seen": 3024551936
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6851227732306213e-05,
      "loss": 1.3347,
      "theoretical_loss": 3.3471409716804343,
      "tokens_seen": 3024683008
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.684320333814797e-05,
      "loss": 1.1941,
      "theoretical_loss": 3.347129933172177,
      "tokens_seen": 3024814080
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.683517894398973e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.347118895276157,
      "tokens_seen": 3024945152
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.682715454983149e-05,
      "loss": 1.2897,
      "theoretical_loss": 3.3471078579923126,
      "tokens_seen": 3025076224
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6819130155673246e-05,
      "loss": 1.2337,
      "theoretical_loss": 3.3470968213205845,
      "tokens_seen": 3025207296
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6811105761515006e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.347085785260912,
      "tokens_seen": 3025338368
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6803081367356766e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.347074749813235,
      "tokens_seen": 3025469440
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6795056973198526e-05,
      "loss": 1.3805,
      "theoretical_loss": 3.347063714977492,
      "tokens_seen": 3025600512
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6787032579040282e-05,
      "loss": 1.1927,
      "theoretical_loss": 3.347052680753624,
      "tokens_seen": 3025731584
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6779008184882042e-05,
      "loss": 1.3143,
      "theoretical_loss": 3.347041647141569,
      "tokens_seen": 3025862656
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6770983790723802e-05,
      "loss": 1.23,
      "theoretical_loss": 3.347030614141268,
      "tokens_seen": 3025993728
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6762959396565562e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.3470195817526593,
      "tokens_seen": 3026124800
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.9922446012496948,
      "objective/train/avg_token_score": 0.007755397818982601,
      "objective/train/docs_used": 1099986,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2939573526382446,
      "objective/train/original_loss": 1.3069510459899902,
      "objective/train/theoretical_loss": 3.3470085499756834,
      "objective/train/tokens_used": 1396714976,
      "objective/train/unlikelihood_loss": 0.018861880525946617,
      "theoretical_loss": 3.3470085499756834,
      "tokens_seen": 3026255872
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.675493500240732e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3470085499756834,
      "tokens_seen": 3026255872
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.674691060824908e-05,
      "loss": 1.3353,
      "theoretical_loss": 3.3469975188102796,
      "tokens_seen": 3026386944
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.673888621409084e-05,
      "loss": 1.331,
      "theoretical_loss": 3.3469864882563876,
      "tokens_seen": 3026518016
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.67308618199326e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.346975458313947,
      "tokens_seen": 3026649088
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6722837425774355e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.346964428982897,
      "tokens_seen": 3026780160
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6714813031616115e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.3469534002631782,
      "tokens_seen": 3026911232
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.6706788637457875e-05,
      "loss": 1.2215,
      "theoretical_loss": 3.3469423721547296,
      "tokens_seen": 3027042304
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.669876424329963e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.3469313446574906,
      "tokens_seen": 3027173376
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.669073984914139e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.346920317771401,
      "tokens_seen": 3027304448
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.668271545498315e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.3469092914964014,
      "tokens_seen": 3027435520
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.667469106082491e-05,
      "loss": 1.2859,
      "theoretical_loss": 3.34689826583243,
      "tokens_seen": 3027566592
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.3468872407794277,
      "tokens_seen": 3027697664
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6658642272508427e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.346876216337334,
      "tokens_seen": 3027828736
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6650617878350187e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.3468651925060877,
      "tokens_seen": 3027959808
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6642593484191947e-05,
      "loss": 1.2491,
      "theoretical_loss": 3.3468541692856295,
      "tokens_seen": 3028090880
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6634569090033704e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.3468431466758983,
      "tokens_seen": 3028221952
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6626544695875464e-05,
      "loss": 1.2025,
      "theoretical_loss": 3.3468321246768347,
      "tokens_seen": 3028353024
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6618520301717224e-05,
      "loss": 1.2581,
      "theoretical_loss": 3.346821103288378,
      "tokens_seen": 3028484096
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.661049590755898e-05,
      "loss": 1.1905,
      "theoretical_loss": 3.346810082510468,
      "tokens_seen": 3028615168
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.660247151340074e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3467990623430444,
      "tokens_seen": 3028746240
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.65944471192425e-05,
      "loss": 1.3536,
      "theoretical_loss": 3.346788042786047,
      "tokens_seen": 3028877312
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6586422725084256e-05,
      "loss": 1.32,
      "theoretical_loss": 3.3467770238394157,
      "tokens_seen": 3029008384
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6578398330926016e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.34676600550309,
      "tokens_seen": 3029139456
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6570373936767776e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.3467549877770097,
      "tokens_seen": 3029270528
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6562349542609533e-05,
      "loss": 1.2554,
      "theoretical_loss": 3.346743970661115,
      "tokens_seen": 3029401600
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6554325148451293e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.3467329541553457,
      "tokens_seen": 3029532672
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6546300754293053e-05,
      "loss": 1.2415,
      "theoretical_loss": 3.346721938259641,
      "tokens_seen": 3029663744
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.653827636013481e-05,
      "loss": 1.1952,
      "theoretical_loss": 3.346710922973941,
      "tokens_seen": 3029794816
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.653025196597657e-05,
      "loss": 1.2928,
      "theoretical_loss": 3.346699908298186,
      "tokens_seen": 3029925888
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.652222757181833e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.346688894232315,
      "tokens_seen": 3030056960
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6514203177660085e-05,
      "loss": 1.2319,
      "theoretical_loss": 3.3466778807762694,
      "tokens_seen": 3030188032
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6506178783501845e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.346666867929987,
      "tokens_seen": 3030319104
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6498154389343605e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.3466558556934087,
      "tokens_seen": 3030450176
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.649012999518536e-05,
      "loss": 1.2101,
      "theoretical_loss": 3.346644844066475,
      "tokens_seen": 3030581248
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.648210560102712e-05,
      "loss": 1.3154,
      "theoretical_loss": 3.3466338330491245,
      "tokens_seen": 3030712320
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.647408120686888e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.3466228226412977,
      "tokens_seen": 3030843392
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.646605681271064e-05,
      "loss": 1.3071,
      "theoretical_loss": 3.346611812842935,
      "tokens_seen": 3030974464
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6458032418552398e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.3466008036539754,
      "tokens_seen": 3031105536
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6450008024394158e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.3465897950743595,
      "tokens_seen": 3031236608
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6441983630235918e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.346578787104027,
      "tokens_seen": 3031367680
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6433959236077674e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.3465677797429176,
      "tokens_seen": 3031498752
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6425934841919434e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.3465567729909713,
      "tokens_seen": 3031629824
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6417910447761194e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.346545766848129,
      "tokens_seen": 3031760896
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6409886053602954e-05,
      "loss": 1.261,
      "theoretical_loss": 3.346534761314329,
      "tokens_seen": 3031891968
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.640186165944471e-05,
      "loss": 1.3553,
      "theoretical_loss": 3.3465237563895123,
      "tokens_seen": 3032023040
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.639383726528647e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.346512752073619,
      "tokens_seen": 3032154112
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.638581287112823e-05,
      "loss": 1.2088,
      "theoretical_loss": 3.3465017483665886,
      "tokens_seen": 3032285184
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.637778847696999e-05,
      "loss": 1.3348,
      "theoretical_loss": 3.3464907452683614,
      "tokens_seen": 3032416256
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6369764082811747e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.3464797427788775,
      "tokens_seen": 3032547328
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6361739688653507e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.3464687408980764,
      "tokens_seen": 3032678400
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.9925715327262878,
      "objective/train/avg_token_score": 0.007428442593663931,
      "objective/train/docs_used": 1102459,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.263049602508545,
      "objective/train/original_loss": 1.2735052108764648,
      "objective/train/theoretical_loss": 3.3464577396258983,
      "objective/train/tokens_used": 1403268576,
      "objective/train/unlikelihood_loss": 0.0216961782425642,
      "theoretical_loss": 3.3464577396258983,
      "tokens_seen": 3032809472
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6353715294495267e-05,
      "loss": 1.2766,
      "theoretical_loss": 3.3464577396258983,
      "tokens_seen": 3032809472
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6345690900337023e-05,
      "loss": 1.2454,
      "theoretical_loss": 3.346446738962284,
      "tokens_seen": 3032940544
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6337666506178783e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.3464357389071724,
      "tokens_seen": 3033071616
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6329642112020543e-05,
      "loss": 1.2324,
      "theoretical_loss": 3.3464247394605042,
      "tokens_seen": 3033202688
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6321617717862303e-05,
      "loss": 1.2344,
      "theoretical_loss": 3.346413740622219,
      "tokens_seen": 3033333760
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.631359332370406e-05,
      "loss": 1.2608,
      "theoretical_loss": 3.3464027423922578,
      "tokens_seen": 3033464832
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.630556892954582e-05,
      "loss": 1.2443,
      "theoretical_loss": 3.3463917447705596,
      "tokens_seen": 3033595904
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.629754453538758e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.346380747757065,
      "tokens_seen": 3033726976
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.628952014122934e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.3463697513517143,
      "tokens_seen": 3033858048
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6281495747071096e-05,
      "loss": 1.2727,
      "theoretical_loss": 3.346358755554447,
      "tokens_seen": 3033989120
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6273471352912855e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.3463477603652034,
      "tokens_seen": 3034120192
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6265446958754615e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.3463367657839242,
      "tokens_seen": 3034251264
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6257422564596375e-05,
      "loss": 1.2426,
      "theoretical_loss": 3.3463257718105486,
      "tokens_seen": 3034382336
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6249398170438132e-05,
      "loss": 1.319,
      "theoretical_loss": 3.3463147784450173,
      "tokens_seen": 3034513408
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.624137377627989e-05,
      "loss": 1.2156,
      "theoretical_loss": 3.3463037856872706,
      "tokens_seen": 3034644480
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.623334938212165e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.3462927935372484,
      "tokens_seen": 3034775552
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6225324987963408e-05,
      "loss": 1.2082,
      "theoretical_loss": 3.3462818019948903,
      "tokens_seen": 3034906624
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6217300593805168e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.3462708110601374,
      "tokens_seen": 3035037696
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6209276199646928e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.3462598207329295,
      "tokens_seen": 3035168768
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6201251805488688e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.346248831013207,
      "tokens_seen": 3035299840
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6193227411330444e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3462378419009093,
      "tokens_seen": 3035430912
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6185203017172204e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.346226853395977,
      "tokens_seen": 3035561984
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6177178623013964e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.346215865498351,
      "tokens_seen": 3035693056
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6169154228855724e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.346204878207971,
      "tokens_seen": 3035824128
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.616112983469748e-05,
      "loss": 1.3628,
      "theoretical_loss": 3.346193891524777,
      "tokens_seen": 3035955200
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.615310544053924e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.346182905448709,
      "tokens_seen": 3036086272
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6145081046381e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.3461719199797084,
      "tokens_seen": 3036217344
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6137056652222757e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.346160935117714,
      "tokens_seen": 3036348416
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6129032258064517e-05,
      "loss": 1.3328,
      "theoretical_loss": 3.3461499508626673,
      "tokens_seen": 3036479488
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6121007863906277e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.3461389672145083,
      "tokens_seen": 3036610560
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6112983469748037e-05,
      "loss": 1.35,
      "theoretical_loss": 3.3461279841731764,
      "tokens_seen": 3036741632
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6104959075589793e-05,
      "loss": 1.2771,
      "theoretical_loss": 3.346117001738613,
      "tokens_seen": 3036872704
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6096934681431553e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.3461060199107573,
      "tokens_seen": 3037003776
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6088910287273313e-05,
      "loss": 1.322,
      "theoretical_loss": 3.3460950386895507,
      "tokens_seen": 3037134848
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6080885893115073e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.3460840580749327,
      "tokens_seen": 3037265920
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.607286149895683e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.346073078066844,
      "tokens_seen": 3037396992
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.606483710479859e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.346062098665225,
      "tokens_seen": 3037528064
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.605681271064035e-05,
      "loss": 1.3042,
      "theoretical_loss": 3.3460511198700162,
      "tokens_seen": 3037659136
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6048788316482106e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.346040141681157,
      "tokens_seen": 3037790208
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6040763922323866e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.3460291640985886,
      "tokens_seen": 3037921280
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6032739528165626e-05,
      "loss": 1.2721,
      "theoretical_loss": 3.3460181871222514,
      "tokens_seen": 3038052352
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6024715134007385e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.3460072107520853,
      "tokens_seen": 3038183424
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6016690739849142e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.345996234988031,
      "tokens_seen": 3038314496
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6008666345690902e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.3459852598300284,
      "tokens_seen": 3038445568
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.6000641951532662e-05,
      "loss": 1.1804,
      "theoretical_loss": 3.3459742852780185,
      "tokens_seen": 3038576640
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.599261755737442e-05,
      "loss": 1.265,
      "theoretical_loss": 3.345963311331942,
      "tokens_seen": 3038707712
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5984593163216178e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.345952337991738,
      "tokens_seen": 3038838784
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5976568769057938e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.345941365257348,
      "tokens_seen": 3038969856
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5968544374899698e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.345930393128712,
      "tokens_seen": 3039100928
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5960519980741454e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.345919421605771,
      "tokens_seen": 3039232000
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.9932354092597961,
      "objective/train/avg_token_score": 0.006764549762010574,
      "objective/train/docs_used": 1104383,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2275776863098145,
      "objective/train/original_loss": 1.2356407642364502,
      "objective/train/theoretical_loss": 3.3459084506884644,
      "objective/train/tokens_used": 1409822176,
      "objective/train/unlikelihood_loss": 0.020724713802337646,
      "theoretical_loss": 3.3459084506884644,
      "tokens_seen": 3039363072
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5952495586583214e-05,
      "loss": 1.2284,
      "theoretical_loss": 3.3459084506884644,
      "tokens_seen": 3039363072
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5944471192424974e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.345897480376734,
      "tokens_seen": 3039494144
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5936446798266734e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.345886510670519,
      "tokens_seen": 3039625216
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.592842240410849e-05,
      "loss": 1.2539,
      "theoretical_loss": 3.34587554156976,
      "tokens_seen": 3039756288
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.592039800995025e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.3458645730743988,
      "tokens_seen": 3039887360
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.591237361579201e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.345853605184374,
      "tokens_seen": 3040018432
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5904349221633767e-05,
      "loss": 1.2205,
      "theoretical_loss": 3.345842637899628,
      "tokens_seen": 3040149504
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5896324827475527e-05,
      "loss": 1.2896,
      "theoretical_loss": 3.3458316712201,
      "tokens_seen": 3040280576
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5888300433317287e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.3458207051457305,
      "tokens_seen": 3040411648
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5880276039159043e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.3458097396764606,
      "tokens_seen": 3040542720
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5872251645000803e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3457987748122306,
      "tokens_seen": 3040673792
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5864227250842563e-05,
      "loss": 1.3245,
      "theoretical_loss": 3.3457878105529812,
      "tokens_seen": 3040804864
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.585620285668432e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.345776846898653,
      "tokens_seen": 3040935936
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.584817846252608e-05,
      "loss": 1.2256,
      "theoretical_loss": 3.345765883849186,
      "tokens_seen": 3041067008
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.584015406836784e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.345754921404521,
      "tokens_seen": 3041198080
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5832129674209596e-05,
      "loss": 1.2993,
      "theoretical_loss": 3.345743959564599,
      "tokens_seen": 3041329152
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5824105280051356e-05,
      "loss": 1.242,
      "theoretical_loss": 3.34573299832936,
      "tokens_seen": 3041460224
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5816080885893116e-05,
      "loss": 1.282,
      "theoretical_loss": 3.345722037698745,
      "tokens_seen": 3041591296
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5808056491734872e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.3457110776726946,
      "tokens_seen": 3041722368
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5800032097576632e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.345700118251149,
      "tokens_seen": 3041853440
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5792007703418392e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.345689159434049,
      "tokens_seen": 3041984512
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.578398330926015e-05,
      "loss": 1.2351,
      "theoretical_loss": 3.345678201221336,
      "tokens_seen": 3042115584
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.577595891510191e-05,
      "loss": 1.438,
      "theoretical_loss": 3.345667243612949,
      "tokens_seen": 3042246656
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.576793452094367e-05,
      "loss": 1.2601,
      "theoretical_loss": 3.34565628660883,
      "tokens_seen": 3042377728
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.575991012678543e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.345645330208919,
      "tokens_seen": 3042508800
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5751885732627185e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.345634374413157,
      "tokens_seen": 3042639872
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5743861338468945e-05,
      "loss": 1.1872,
      "theoretical_loss": 3.3456234192214844,
      "tokens_seen": 3042770944
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5735836944310705e-05,
      "loss": 1.3847,
      "theoretical_loss": 3.345612464633842,
      "tokens_seen": 3042902016
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5727812550152465e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.345601510650171,
      "tokens_seen": 3043033088
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.571978815599422e-05,
      "loss": 1.3483,
      "theoretical_loss": 3.3455905572704108,
      "tokens_seen": 3043164160
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.571176376183598e-05,
      "loss": 1.1818,
      "theoretical_loss": 3.345579604494503,
      "tokens_seen": 3043295232
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.570373936767774e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.345568652322388,
      "tokens_seen": 3043426304
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.56957149735195e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.345557700754007,
      "tokens_seen": 3043557376
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5687690579361257e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.3455467497893006,
      "tokens_seen": 3043688448
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5679666185203017e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.345535799428209,
      "tokens_seen": 3043819520
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5671641791044777e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.3455248496706735,
      "tokens_seen": 3043950592
    },
    {
      "epoch": 0.84,
      "learning_rate": 1.5663617396886534e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.345513900516634,
      "tokens_seen": 3044081664
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5655593002728294e-05,
      "loss": 1.2361,
      "theoretical_loss": 3.345502951966033,
      "tokens_seen": 3044212736
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5647568608570054e-05,
      "loss": 1.2894,
      "theoretical_loss": 3.345492004018809,
      "tokens_seen": 3044343808
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5639544214411813e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.3454810566749043,
      "tokens_seen": 3044474880
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.563151982025357e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.3454701099342596,
      "tokens_seen": 3044605952
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.562349542609533e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.3454591637968147,
      "tokens_seen": 3044737024
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.561547103193709e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3454482182625114,
      "tokens_seen": 3044868096
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.560744663777885e-05,
      "loss": 1.3007,
      "theoretical_loss": 3.3454372733312905,
      "tokens_seen": 3044999168
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5599422243620606e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.345426329003092,
      "tokens_seen": 3045130240
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5591397849462366e-05,
      "loss": 1.291,
      "theoretical_loss": 3.345415385277857,
      "tokens_seen": 3045261312
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5583373455304126e-05,
      "loss": 1.2241,
      "theoretical_loss": 3.345404442155527,
      "tokens_seen": 3045392384
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5575349061145883e-05,
      "loss": 1.167,
      "theoretical_loss": 3.3453934996360424,
      "tokens_seen": 3045523456
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5567324666987642e-05,
      "loss": 1.3256,
      "theoretical_loss": 3.345382557719344,
      "tokens_seen": 3045654528
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5559300272829402e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.345371616405372,
      "tokens_seen": 3045785600
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.9917280673980713,
      "objective/train/avg_token_score": 0.008271921426057816,
      "objective/train/docs_used": 1106703,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2452493906021118,
      "objective/train/original_loss": 1.256644368171692,
      "objective/train/theoretical_loss": 3.345360675694069,
      "objective/train/tokens_used": 1416375776,
      "objective/train/unlikelihood_loss": 0.021693279966711998,
      "theoretical_loss": 3.345360675694069,
      "tokens_seen": 3045916672
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5551275878671162e-05,
      "loss": 1.298,
      "theoretical_loss": 3.345360675694069,
      "tokens_seen": 3045916672
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.554325148451292e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.3453497355853745,
      "tokens_seen": 3046047744
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.553522709035468e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.345338796079229,
      "tokens_seen": 3046178816
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.552720269619644e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.345327857175575,
      "tokens_seen": 3046309888
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.55191783020382e-05,
      "loss": 1.2572,
      "theoretical_loss": 3.345316918874352,
      "tokens_seen": 3046440960
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5511153907879955e-05,
      "loss": 1.2726,
      "theoretical_loss": 3.3453059811755015,
      "tokens_seen": 3046572032
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5503129513721715e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.3452950440789637,
      "tokens_seen": 3046703104
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5495105119563475e-05,
      "loss": 1.2827,
      "theoretical_loss": 3.345284107584681,
      "tokens_seen": 3046834176
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.548708072540523e-05,
      "loss": 1.3141,
      "theoretical_loss": 3.3452731716925928,
      "tokens_seen": 3046965248
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.547905633124699e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.345262236402641,
      "tokens_seen": 3047096320
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.547103193708875e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.3452513017147663,
      "tokens_seen": 3047227392
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.546300754293051e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.3452403676289095,
      "tokens_seen": 3047358464
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5454983148772268e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3452294341450117,
      "tokens_seen": 3047489536
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5446958754614027e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.345218501263014,
      "tokens_seen": 3047620608
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5438934360455787e-05,
      "loss": 1.3036,
      "theoretical_loss": 3.3452075689828566,
      "tokens_seen": 3047751680
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5430909966297547e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.3451966373044817,
      "tokens_seen": 3047882752
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5422885572139304e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.345185706227829,
      "tokens_seen": 3048013824
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5414861177981064e-05,
      "loss": 1.3033,
      "theoretical_loss": 3.3451747757528407,
      "tokens_seen": 3048144896
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5406836783822824e-05,
      "loss": 1.3,
      "theoretical_loss": 3.3451638458794575,
      "tokens_seen": 3048275968
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.539881238966458e-05,
      "loss": 1.2432,
      "theoretical_loss": 3.3451529166076197,
      "tokens_seen": 3048407040
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.539078799550634e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.3451419879372692,
      "tokens_seen": 3048538112
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.53827636013481e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.345131059868347,
      "tokens_seen": 3048669184
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.537473920718986e-05,
      "loss": 1.2955,
      "theoretical_loss": 3.345120132400793,
      "tokens_seen": 3048800256
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5366714813031616e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3451092055345497,
      "tokens_seen": 3048931328
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5358690418873376e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.345098279269558,
      "tokens_seen": 3049062400
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5350666024715136e-05,
      "loss": 1.2085,
      "theoretical_loss": 3.3450873536057575,
      "tokens_seen": 3049193472
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5342641630556896e-05,
      "loss": 1.283,
      "theoretical_loss": 3.345076428543091,
      "tokens_seen": 3049324544
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5334617236398653e-05,
      "loss": 1.2946,
      "theoretical_loss": 3.3450655040814987,
      "tokens_seen": 3049455616
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5326592842240413e-05,
      "loss": 1.2939,
      "theoretical_loss": 3.3450545802209217,
      "tokens_seen": 3049586688
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5318568448082172e-05,
      "loss": 1.2488,
      "theoretical_loss": 3.345043656961302,
      "tokens_seen": 3049717760
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.531054405392393e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.3450327343025794,
      "tokens_seen": 3049848832
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.530251965976569e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.3450218122446955,
      "tokens_seen": 3049979904
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.529449526560745e-05,
      "loss": 1.2812,
      "theoretical_loss": 3.345010890787592,
      "tokens_seen": 3050110976
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.528647087144921e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.3449999699312096,
      "tokens_seen": 3050242048
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5278446477290965e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.3449890496754895,
      "tokens_seen": 3050373120
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5270422083132725e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.3449781300203725,
      "tokens_seen": 3050504192
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5262397688974485e-05,
      "loss": 1.3107,
      "theoretical_loss": 3.3449672109658,
      "tokens_seen": 3050635264
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5254373294816243e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3449562925117133,
      "tokens_seen": 3050766336
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5246348900658e-05,
      "loss": 1.293,
      "theoretical_loss": 3.344945374658054,
      "tokens_seen": 3050897408
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.523832450649976e-05,
      "loss": 1.1802,
      "theoretical_loss": 3.3449344574047624,
      "tokens_seen": 3051028480
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.523030011234152e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.3449235407517803,
      "tokens_seen": 3051159552
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.522227571818328e-05,
      "loss": 1.3327,
      "theoretical_loss": 3.344912624699049,
      "tokens_seen": 3051290624
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5214251324025036e-05,
      "loss": 1.2452,
      "theoretical_loss": 3.3449017092465088,
      "tokens_seen": 3051421696
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5206226929866796e-05,
      "loss": 1.1614,
      "theoretical_loss": 3.344890794394102,
      "tokens_seen": 3051552768
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5198202535708556e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.344879880141769,
      "tokens_seen": 3051683840
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5190178141550312e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.3448689664894515,
      "tokens_seen": 3051814912
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5182153747392072e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.3448580534370906,
      "tokens_seen": 3051945984
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5174129353233832e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.344847140984628,
      "tokens_seen": 3052077056
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5166104959075592e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3448362291320044,
      "tokens_seen": 3052208128
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5158080564917349e-05,
      "loss": 1.2274,
      "theoretical_loss": 3.3448253178791614,
      "tokens_seen": 3052339200
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.9923360347747803,
      "objective/train/avg_token_score": 0.007663935422897339,
      "objective/train/docs_used": 1109048,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.284757375717163,
      "objective/train/original_loss": 1.2970839738845825,
      "objective/train/theoretical_loss": 3.34481440722604,
      "objective/train/tokens_used": 1422929376,
      "objective/train/unlikelihood_loss": 0.020660171285271645,
      "theoretical_loss": 3.34481440722604,
      "tokens_seen": 3052470272
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5150056170759108e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.34481440722604,
      "tokens_seen": 3052470272
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5142031776600868e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.3448034971725815,
      "tokens_seen": 3052601344
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5134007382442628e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.3447925877187275,
      "tokens_seen": 3052732416
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5125982988284385e-05,
      "loss": 1.2793,
      "theoretical_loss": 3.344781678864419,
      "tokens_seen": 3052863488
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5117958594126145e-05,
      "loss": 1.1704,
      "theoretical_loss": 3.3447707706095975,
      "tokens_seen": 3052994560
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5109934199967905e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.344759862954204,
      "tokens_seen": 3053125632
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5101909805809661e-05,
      "loss": 1.3043,
      "theoretical_loss": 3.34474895589818,
      "tokens_seen": 3053256704
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5093885411651421e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.3447380494414674,
      "tokens_seen": 3053387776
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5085861017493181e-05,
      "loss": 1.2926,
      "theoretical_loss": 3.344727143584007,
      "tokens_seen": 3053518848
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5077836623334939e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.3447162383257405,
      "tokens_seen": 3053649920
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5069812229176697e-05,
      "loss": 1.234,
      "theoretical_loss": 3.3447053336666084,
      "tokens_seen": 3053780992
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5061787835018457e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.3446944296065526,
      "tokens_seen": 3053912064
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5053763440860215e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.344683526145515,
      "tokens_seen": 3054043136
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5045739046701975e-05,
      "loss": 1.2237,
      "theoretical_loss": 3.344672623283436,
      "tokens_seen": 3054174208
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5037714652543734e-05,
      "loss": 1.2236,
      "theoretical_loss": 3.344661721020258,
      "tokens_seen": 3054305280
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5029690258385492e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.3446508193559215,
      "tokens_seen": 3054436352
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5021665864227252e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.3446399182903686,
      "tokens_seen": 3054567424
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.501364147006901e-05,
      "loss": 1.249,
      "theoretical_loss": 3.3446290178235403,
      "tokens_seen": 3054698496
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.5005617075910768e-05,
      "loss": 1.1988,
      "theoretical_loss": 3.344618117955378,
      "tokens_seen": 3054829568
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4997592681752528e-05,
      "loss": 1.2089,
      "theoretical_loss": 3.3446072186858236,
      "tokens_seen": 3054960640
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4989568287594288e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.3445963200148183,
      "tokens_seen": 3055091712
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4981543893436044e-05,
      "loss": 1.2171,
      "theoretical_loss": 3.3445854219423032,
      "tokens_seen": 3055222784
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4973519499277804e-05,
      "loss": 1.254,
      "theoretical_loss": 3.34457452446822,
      "tokens_seen": 3055353856
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4965495105119564e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.344563627592511,
      "tokens_seen": 3055484928
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4957470710961324e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.344552731315116,
      "tokens_seen": 3055616000
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.494944631680308e-05,
      "loss": 1.3208,
      "theoretical_loss": 3.3445418356359777,
      "tokens_seen": 3055747072
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.494142192264484e-05,
      "loss": 1.1802,
      "theoretical_loss": 3.344530940555037,
      "tokens_seen": 3055878144
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.49333975284866e-05,
      "loss": 1.253,
      "theoretical_loss": 3.3445200460722364,
      "tokens_seen": 3056009216
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4925373134328357e-05,
      "loss": 1.2798,
      "theoretical_loss": 3.3445091521875163,
      "tokens_seen": 3056140288
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4917348740170117e-05,
      "loss": 1.257,
      "theoretical_loss": 3.3444982589008188,
      "tokens_seen": 3056271360
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4909324346011877e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.344487366212085,
      "tokens_seen": 3056402432
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4901299951853637e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.3444764741212567,
      "tokens_seen": 3056533504
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4893275557695393e-05,
      "loss": 1.2066,
      "theoretical_loss": 3.3444655826282754,
      "tokens_seen": 3056664576
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4885251163537153e-05,
      "loss": 1.2506,
      "theoretical_loss": 3.344454691733083,
      "tokens_seen": 3056795648
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4877226769378913e-05,
      "loss": 1.285,
      "theoretical_loss": 3.3444438014356206,
      "tokens_seen": 3056926720
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4869202375220673e-05,
      "loss": 1.2712,
      "theoretical_loss": 3.3444329117358294,
      "tokens_seen": 3057057792
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.486117798106243e-05,
      "loss": 1.2468,
      "theoretical_loss": 3.344422022633652,
      "tokens_seen": 3057188864
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.485315358690419e-05,
      "loss": 1.1829,
      "theoretical_loss": 3.3444111341290297,
      "tokens_seen": 3057319936
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.484512919274595e-05,
      "loss": 1.2348,
      "theoretical_loss": 3.344400246221903,
      "tokens_seen": 3057451008
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4837104798587706e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.344389358912215,
      "tokens_seen": 3057582080
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4829080404429466e-05,
      "loss": 1.1844,
      "theoretical_loss": 3.3443784721999066,
      "tokens_seen": 3057713152
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4821056010271226e-05,
      "loss": 1.1918,
      "theoretical_loss": 3.3443675860849194,
      "tokens_seen": 3057844224
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4813031616112986e-05,
      "loss": 1.2649,
      "theoretical_loss": 3.344356700567195,
      "tokens_seen": 3057975296
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4805007221954742e-05,
      "loss": 1.2431,
      "theoretical_loss": 3.344345815646675,
      "tokens_seen": 3058106368
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4796982827796502e-05,
      "loss": 1.211,
      "theoretical_loss": 3.344334931323302,
      "tokens_seen": 3058237440
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4788958433638262e-05,
      "loss": 1.235,
      "theoretical_loss": 3.344324047597016,
      "tokens_seen": 3058368512
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4780934039480022e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.34431316446776,
      "tokens_seen": 3058499584
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4772909645321778e-05,
      "loss": 1.3513,
      "theoretical_loss": 3.3443022819354744,
      "tokens_seen": 3058630656
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4764885251163538e-05,
      "loss": 1.1225,
      "theoretical_loss": 3.3442914000001025,
      "tokens_seen": 3058761728
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4756860857005298e-05,
      "loss": 1.264,
      "theoretical_loss": 3.3442805186615843,
      "tokens_seen": 3058892800
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.9947852492332458,
      "objective/train/avg_token_score": 0.0052147796377539635,
      "objective/train/docs_used": 1111330,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1884770393371582,
      "objective/train/original_loss": 1.1943895816802979,
      "objective/train/theoretical_loss": 3.344269637919863,
      "objective/train/tokens_used": 1429482976,
      "objective/train/unlikelihood_loss": 0.016896842047572136,
      "theoretical_loss": 3.344269637919863,
      "tokens_seen": 3059023872
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4748836462847056e-05,
      "loss": 1.2007,
      "theoretical_loss": 3.344269637919863,
      "tokens_seen": 3059023872
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4740812068688815e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.3442587577748792,
      "tokens_seen": 3059154944
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4732787674530574e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.344247878226575,
      "tokens_seen": 3059286016
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4724763280372333e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.3442369992748926,
      "tokens_seen": 3059417088
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.471673888621409e-05,
      "loss": 1.2051,
      "theoretical_loss": 3.344226120919773,
      "tokens_seen": 3059548160
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.470871449205585e-05,
      "loss": 1.2165,
      "theoretical_loss": 3.344215243161158,
      "tokens_seen": 3059679232
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4700690097897609e-05,
      "loss": 1.2317,
      "theoretical_loss": 3.3442043659989893,
      "tokens_seen": 3059810304
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4692665703739369e-05,
      "loss": 1.2379,
      "theoretical_loss": 3.3441934894332093,
      "tokens_seen": 3059941376
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4684641309581127e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.344182613463759,
      "tokens_seen": 3060072448
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4676616915422885e-05,
      "loss": 1.3272,
      "theoretical_loss": 3.344171738090581,
      "tokens_seen": 3060203520
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4668592521264645e-05,
      "loss": 1.2043,
      "theoretical_loss": 3.344160863313616,
      "tokens_seen": 3060334592
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4660568127106405e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.3441499891328066,
      "tokens_seen": 3060465664
    },
    {
      "epoch": 0.85,
      "learning_rate": 1.4652543732948162e-05,
      "loss": 1.2691,
      "theoretical_loss": 3.3441391155480944,
      "tokens_seen": 3060596736
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4644519338789922e-05,
      "loss": 1.2074,
      "theoretical_loss": 3.344128242559421,
      "tokens_seen": 3060727808
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4636494944631681e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.344117370166728,
      "tokens_seen": 3060858880
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4628470550473438e-05,
      "loss": 1.2137,
      "theoretical_loss": 3.344106498369958,
      "tokens_seen": 3060989952
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4620446156315198e-05,
      "loss": 1.236,
      "theoretical_loss": 3.3440956271690525,
      "tokens_seen": 3061121024
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4612421762156958e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.344084756563953,
      "tokens_seen": 3061252096
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4604397367998718e-05,
      "loss": 1.1317,
      "theoretical_loss": 3.3440738865546016,
      "tokens_seen": 3061383168
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4596372973840474e-05,
      "loss": 1.2965,
      "theoretical_loss": 3.34406301714094,
      "tokens_seen": 3061514240
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4588348579682234e-05,
      "loss": 1.2837,
      "theoretical_loss": 3.34405214832291,
      "tokens_seen": 3061645312
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4580324185523994e-05,
      "loss": 1.2594,
      "theoretical_loss": 3.3440412801004538,
      "tokens_seen": 3061776384
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4572299791365754e-05,
      "loss": 1.2594,
      "theoretical_loss": 3.344030412473513,
      "tokens_seen": 3061907456
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.456427539720751e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.3440195454420296,
      "tokens_seen": 3062038528
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.455625100304927e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3440086790059453,
      "tokens_seen": 3062169600
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.454822660889103e-05,
      "loss": 1.229,
      "theoretical_loss": 3.343997813165202,
      "tokens_seen": 3062300672
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4540202214732787e-05,
      "loss": 1.3219,
      "theoretical_loss": 3.343986947919742,
      "tokens_seen": 3062431744
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4532177820574547e-05,
      "loss": 1.2904,
      "theoretical_loss": 3.343976083269507,
      "tokens_seen": 3062562816
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4524153426416307e-05,
      "loss": 1.3839,
      "theoretical_loss": 3.343965219214439,
      "tokens_seen": 3062693888
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4516129032258066e-05,
      "loss": 1.2386,
      "theoretical_loss": 3.3439543557544793,
      "tokens_seen": 3062824960
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4508104638099823e-05,
      "loss": 1.274,
      "theoretical_loss": 3.343943492889571,
      "tokens_seen": 3062956032
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4500080243941583e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.343932630619655,
      "tokens_seen": 3063087104
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4492055849783343e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.343921768944673,
      "tokens_seen": 3063218176
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4484031455625103e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.3439109078645686,
      "tokens_seen": 3063349248
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.447600706146686e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.3439000473792824,
      "tokens_seen": 3063480320
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4467982667308619e-05,
      "loss": 1.274,
      "theoretical_loss": 3.3438891874887564,
      "tokens_seen": 3063611392
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4459958273150379e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.3438783281929334,
      "tokens_seen": 3063742464
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4451933878992136e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.3438674694917543,
      "tokens_seen": 3063873536
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4443909484833895e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.343856611385162,
      "tokens_seen": 3064004608
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4435885090675655e-05,
      "loss": 1.3446,
      "theoretical_loss": 3.3438457538730986,
      "tokens_seen": 3064135680
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4427860696517415e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.3438348969555056,
      "tokens_seen": 3064266752
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4419836302359172e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.3438240406323247,
      "tokens_seen": 3064397824
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4411811908200932e-05,
      "loss": 1.1959,
      "theoretical_loss": 3.3438131849034987,
      "tokens_seen": 3064528896
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4403787514042692e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.3438023297689696,
      "tokens_seen": 3064659968
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.439576311988445e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.343791475228679,
      "tokens_seen": 3064791040
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4387738725726208e-05,
      "loss": 1.2278,
      "theoretical_loss": 3.3437806212825687,
      "tokens_seen": 3064922112
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4379714331567968e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.343769767930582,
      "tokens_seen": 3065053184
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4371689937409726e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3437589151726597,
      "tokens_seen": 3065184256
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4363665543251484e-05,
      "loss": 1.2246,
      "theoretical_loss": 3.343748063008744,
      "tokens_seen": 3065315328
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4355641149093244e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.343737211438778,
      "tokens_seen": 3065446400
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.9946680068969727,
      "objective/train/avg_token_score": 0.005332026164978743,
      "objective/train/docs_used": 1113675,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2437716722488403,
      "objective/train/original_loss": 1.2507622241973877,
      "objective/train/theoretical_loss": 3.343726360462703,
      "objective/train/tokens_used": 1436036576,
      "objective/train/unlikelihood_loss": 0.016266847029328346,
      "theoretical_loss": 3.343726360462703,
      "tokens_seen": 3065577472
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4347616754935002e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.343726360462703,
      "tokens_seen": 3065577472
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4339592360776762e-05,
      "loss": 1.2618,
      "theoretical_loss": 3.3437155100804614,
      "tokens_seen": 3065708544
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.433156796661852e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.343704660291995,
      "tokens_seen": 3065839616
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4323543572460279e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.3436938110972463,
      "tokens_seen": 3065970688
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4315519178302039e-05,
      "loss": 1.3332,
      "theoretical_loss": 3.343682962496157,
      "tokens_seen": 3066101760
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4307494784143799e-05,
      "loss": 1.217,
      "theoretical_loss": 3.3436721144886694,
      "tokens_seen": 3066232832
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4299470389985555e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.343661267074726,
      "tokens_seen": 3066363904
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4291445995827315e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.3436504202542685,
      "tokens_seen": 3066494976
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4283421601669075e-05,
      "loss": 1.1924,
      "theoretical_loss": 3.343639574027239,
      "tokens_seen": 3066626048
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4275397207510835e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.3436287283935804,
      "tokens_seen": 3066757120
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4267372813352591e-05,
      "loss": 1.258,
      "theoretical_loss": 3.343617883353234,
      "tokens_seen": 3066888192
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4259348419194351e-05,
      "loss": 1.253,
      "theoretical_loss": 3.3436070389061427,
      "tokens_seen": 3067019264
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4251324025036111e-05,
      "loss": 1.2944,
      "theoretical_loss": 3.343596195052248,
      "tokens_seen": 3067150336
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4243299630877868e-05,
      "loss": 1.2772,
      "theoretical_loss": 3.343585351791493,
      "tokens_seen": 3067281408
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4235275236719628e-05,
      "loss": 1.275,
      "theoretical_loss": 3.343574509123819,
      "tokens_seen": 3067412480
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4227250842561388e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.3435636670491684,
      "tokens_seen": 3067543552
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4219226448403147e-05,
      "loss": 1.2588,
      "theoretical_loss": 3.3435528255674845,
      "tokens_seen": 3067674624
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4211202054244904e-05,
      "loss": 1.176,
      "theoretical_loss": 3.3435419846787076,
      "tokens_seen": 3067805696
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4203177660086664e-05,
      "loss": 1.2103,
      "theoretical_loss": 3.3435311443827818,
      "tokens_seen": 3067936768
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4195153265928424e-05,
      "loss": 1.2845,
      "theoretical_loss": 3.3435203046796484,
      "tokens_seen": 3068067840
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4187128871770184e-05,
      "loss": 1.2381,
      "theoretical_loss": 3.34350946556925,
      "tokens_seen": 3068198912
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.417910447761194e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.343498627051528,
      "tokens_seen": 3068329984
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.41710800834537e-05,
      "loss": 1.3388,
      "theoretical_loss": 3.343487789126426,
      "tokens_seen": 3068461056
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.416305568929546e-05,
      "loss": 1.2835,
      "theoretical_loss": 3.3434769517938854,
      "tokens_seen": 3068592128
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4155031295137216e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.343466115053849,
      "tokens_seen": 3068723200
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4147006900978976e-05,
      "loss": 1.2233,
      "theoretical_loss": 3.3434552789062586,
      "tokens_seen": 3068854272
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4138982506820736e-05,
      "loss": 1.2105,
      "theoretical_loss": 3.343444443351057,
      "tokens_seen": 3068985344
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4130958112662496e-05,
      "loss": 1.2733,
      "theoretical_loss": 3.343433608388186,
      "tokens_seen": 3069116416
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4122933718504253e-05,
      "loss": 1.2941,
      "theoretical_loss": 3.3434227740175886,
      "tokens_seen": 3069247488
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4114909324346013e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3434119402392066,
      "tokens_seen": 3069378560
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4106884930187773e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.3434011070529825,
      "tokens_seen": 3069509632
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4098860536029532e-05,
      "loss": 1.243,
      "theoretical_loss": 3.3433902744588586,
      "tokens_seen": 3069640704
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4090836141871289e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.343379442456777,
      "tokens_seen": 3069771776
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4082811747713049e-05,
      "loss": 1.3257,
      "theoretical_loss": 3.3433686110466807,
      "tokens_seen": 3069902848
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4074787353554809e-05,
      "loss": 1.321,
      "theoretical_loss": 3.3433577802285117,
      "tokens_seen": 3070033920
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4066762959396565e-05,
      "loss": 1.2962,
      "theoretical_loss": 3.3433469500022124,
      "tokens_seen": 3070164992
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4058738565238325e-05,
      "loss": 1.3082,
      "theoretical_loss": 3.343336120367725,
      "tokens_seen": 3070296064
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4050714171080085e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.3433252913249927,
      "tokens_seen": 3070427136
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4042689776921843e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.3433144628739564,
      "tokens_seen": 3070558208
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4034665382763602e-05,
      "loss": 1.2711,
      "theoretical_loss": 3.34330363501456,
      "tokens_seen": 3070689280
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4026640988605361e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.3432928077467454,
      "tokens_seen": 3070820352
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.401861659444712e-05,
      "loss": 1.3047,
      "theoretical_loss": 3.343281981070455,
      "tokens_seen": 3070951424
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.401059220028888e-05,
      "loss": 1.2616,
      "theoretical_loss": 3.343271154985631,
      "tokens_seen": 3071082496
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4002567806130638e-05,
      "loss": 1.256,
      "theoretical_loss": 3.343260329492216,
      "tokens_seen": 3071213568
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3994543411972396e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.343249504590153,
      "tokens_seen": 3071344640
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3986519017814156e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3432386802793834,
      "tokens_seen": 3071475712
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3978494623655914e-05,
      "loss": 1.2818,
      "theoretical_loss": 3.3432278565598508,
      "tokens_seen": 3071606784
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3970470229497672e-05,
      "loss": 1.3527,
      "theoretical_loss": 3.343217033431497,
      "tokens_seen": 3071737856
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3962445835339432e-05,
      "loss": 1.2844,
      "theoretical_loss": 3.343206210894264,
      "tokens_seen": 3071868928
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3954421441181192e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.3431953889480956,
      "tokens_seen": 3072000000
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.994194507598877,
      "objective/train/avg_token_score": 0.005805479362607002,
      "objective/train/docs_used": 1116127,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2202064990997314,
      "objective/train/original_loss": 1.2279956340789795,
      "objective/train/theoretical_loss": 3.3431845675929335,
      "objective/train/tokens_used": 1442590176,
      "objective/train/unlikelihood_loss": 0.01819433830678463,
      "theoretical_loss": 3.3431845675929335,
      "tokens_seen": 3072131072
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3946397047022949e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.3431845675929335,
      "tokens_seen": 3072131072
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3938372652864709e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.34317374682872,
      "tokens_seen": 3072262144
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3930348258706468e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.343162926655398,
      "tokens_seen": 3072393216
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3922323864548228e-05,
      "loss": 1.2292,
      "theoretical_loss": 3.3431521070729104,
      "tokens_seen": 3072524288
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3914299470389985e-05,
      "loss": 1.2767,
      "theoretical_loss": 3.343141288081199,
      "tokens_seen": 3072655360
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3906275076231745e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.3431304696802067,
      "tokens_seen": 3072786432
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3898250682073505e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.3431196518698756,
      "tokens_seen": 3072917504
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3890226287915261e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.3431088346501494,
      "tokens_seen": 3073048576
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3882201893757021e-05,
      "loss": 1.2118,
      "theoretical_loss": 3.3430980180209695,
      "tokens_seen": 3073179648
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3874177499598781e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.343087201982279,
      "tokens_seen": 3073310720
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3866153105440541e-05,
      "loss": 1.2145,
      "theoretical_loss": 3.3430763865340203,
      "tokens_seen": 3073441792
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3858128711282297e-05,
      "loss": 1.3091,
      "theoretical_loss": 3.343065571676136,
      "tokens_seen": 3073572864
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3850104317124057e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.343054757408569,
      "tokens_seen": 3073703936
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3842079922965817e-05,
      "loss": 1.1926,
      "theoretical_loss": 3.343043943731262,
      "tokens_seen": 3073835008
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3834055528807577e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.3430331306441565,
      "tokens_seen": 3073966080
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3826031134649334e-05,
      "loss": 1.2839,
      "theoretical_loss": 3.343022318147196,
      "tokens_seen": 3074097152
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3818006740491094e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.3430115062403236,
      "tokens_seen": 3074228224
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3809982346332853e-05,
      "loss": 1.2231,
      "theoretical_loss": 3.3430006949234814,
      "tokens_seen": 3074359296
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3801957952174613e-05,
      "loss": 1.3108,
      "theoretical_loss": 3.3429898841966117,
      "tokens_seen": 3074490368
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.379393355801637e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.3429790740596577,
      "tokens_seen": 3074621440
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.378590916385813e-05,
      "loss": 1.2699,
      "theoretical_loss": 3.3429682645125616,
      "tokens_seen": 3074752512
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.377788476969989e-05,
      "loss": 1.1921,
      "theoretical_loss": 3.3429574555552666,
      "tokens_seen": 3074883584
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3769860375541646e-05,
      "loss": 1.2571,
      "theoretical_loss": 3.342946647187715,
      "tokens_seen": 3075014656
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3761835981383406e-05,
      "loss": 1.3188,
      "theoretical_loss": 3.3429358394098494,
      "tokens_seen": 3075145728
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3753811587225166e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.342925032221613,
      "tokens_seen": 3075276800
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3745787193066926e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.3429142256229483,
      "tokens_seen": 3075407872
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3737762798908682e-05,
      "loss": 1.2867,
      "theoretical_loss": 3.3429034196137977,
      "tokens_seen": 3075538944
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3729738404750442e-05,
      "loss": 1.3131,
      "theoretical_loss": 3.342892614194104,
      "tokens_seen": 3075670016
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3721714010592202e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.34288180936381,
      "tokens_seen": 3075801088
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.371368961643396e-05,
      "loss": 1.1927,
      "theoretical_loss": 3.3428710051228587,
      "tokens_seen": 3075932160
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3705665222275719e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.3428602014711926,
      "tokens_seen": 3076063232
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3697640828117479e-05,
      "loss": 1.259,
      "theoretical_loss": 3.3428493984087546,
      "tokens_seen": 3076194304
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3689616433959237e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.342838595935487,
      "tokens_seen": 3076325376
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3681592039800995e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.3428277940513333,
      "tokens_seen": 3076456448
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3673567645642755e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.342816992756236,
      "tokens_seen": 3076587520
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3665543251484513e-05,
      "loss": 1.1918,
      "theoretical_loss": 3.3428061920501375,
      "tokens_seen": 3076718592
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3657518857326273e-05,
      "loss": 1.2153,
      "theoretical_loss": 3.342795391932981,
      "tokens_seen": 3076849664
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.3649494463168031e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.342784592404709,
      "tokens_seen": 3076980736
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.364147006900979e-05,
      "loss": 1.2424,
      "theoretical_loss": 3.3427737934652644,
      "tokens_seen": 3077111808
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.363344567485155e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.3427629951145903,
      "tokens_seen": 3077242880
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.362542128069331e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.3427521973526293,
      "tokens_seen": 3077373952
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3617396886535066e-05,
      "loss": 1.291,
      "theoretical_loss": 3.342741400179324,
      "tokens_seen": 3077505024
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3609372492376826e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.342730603594618,
      "tokens_seen": 3077636096
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3601348098218586e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.342719807598453,
      "tokens_seen": 3077767168
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3593323704060342e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.342709012190773,
      "tokens_seen": 3077898240
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3585299309902102e-05,
      "loss": 1.2983,
      "theoretical_loss": 3.3426982173715203,
      "tokens_seen": 3078029312
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3577274915743862e-05,
      "loss": 1.2472,
      "theoretical_loss": 3.3426874231406374,
      "tokens_seen": 3078160384
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3569250521585622e-05,
      "loss": 1.2692,
      "theoretical_loss": 3.342676629498068,
      "tokens_seen": 3078291456
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3561226127427378e-05,
      "loss": 1.214,
      "theoretical_loss": 3.3426658364437545,
      "tokens_seen": 3078422528
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3553201733269138e-05,
      "loss": 1.3008,
      "theoretical_loss": 3.34265504397764,
      "tokens_seen": 3078553600
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.9918553829193115,
      "objective/train/avg_token_score": 0.008144626393914223,
      "objective/train/docs_used": 1118389,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3120956420898438,
      "objective/train/original_loss": 1.3246299028396606,
      "objective/train/theoretical_loss": 3.342644252099667,
      "objective/train/tokens_used": 1449143776,
      "objective/train/unlikelihood_loss": 0.021634364500641823,
      "theoretical_loss": 3.342644252099667,
      "tokens_seen": 3078684672
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3545177339110898e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.342644252099667,
      "tokens_seen": 3078684672
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3537152944952658e-05,
      "loss": 1.2396,
      "theoretical_loss": 3.3426334608097785,
      "tokens_seen": 3078815744
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3529128550794415e-05,
      "loss": 1.2168,
      "theoretical_loss": 3.342622670107918,
      "tokens_seen": 3078946816
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3521104156636175e-05,
      "loss": 1.1854,
      "theoretical_loss": 3.342611879994028,
      "tokens_seen": 3079077888
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3513079762477934e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.3426010904680514,
      "tokens_seen": 3079208960
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3505055368319691e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.342590301529931,
      "tokens_seen": 3079340032
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3497030974161451e-05,
      "loss": 1.2314,
      "theoretical_loss": 3.3425795131796106,
      "tokens_seen": 3079471104
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.348900658000321e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.342568725417032,
      "tokens_seen": 3079602176
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.348098218584497e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3425579382421393,
      "tokens_seen": 3079733248
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3472957791686727e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.342547151654874,
      "tokens_seen": 3079864320
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3464933397528487e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3425363656551808,
      "tokens_seen": 3079995392
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3456909003370247e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.3425255802430014,
      "tokens_seen": 3080126464
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3448884609212007e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.3425147954182792,
      "tokens_seen": 3080257536
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3440860215053763e-05,
      "loss": 1.2086,
      "theoretical_loss": 3.3425040111809574,
      "tokens_seen": 3080388608
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3432835820895523e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.342493227530979,
      "tokens_seen": 3080519680
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3424811426737283e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.3424824444682866,
      "tokens_seen": 3080650752
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.341678703257904e-05,
      "loss": 1.2303,
      "theoretical_loss": 3.342471661992824,
      "tokens_seen": 3080781824
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.34087626384208e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.3424608801045332,
      "tokens_seen": 3080912896
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.340073824426256e-05,
      "loss": 1.2757,
      "theoretical_loss": 3.342450098803358,
      "tokens_seen": 3081043968
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.339271385010432e-05,
      "loss": 1.3214,
      "theoretical_loss": 3.3424393180892413,
      "tokens_seen": 3081175040
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3384689455946076e-05,
      "loss": 1.2668,
      "theoretical_loss": 3.342428537962126,
      "tokens_seen": 3081306112
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3376665061787836e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.3424177584219557,
      "tokens_seen": 3081437184
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3368640667629596e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.342406979468673,
      "tokens_seen": 3081568256
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3360616273471354e-05,
      "loss": 1.3574,
      "theoretical_loss": 3.3423962011022206,
      "tokens_seen": 3081699328
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3352591879313112e-05,
      "loss": 1.2355,
      "theoretical_loss": 3.342385423322542,
      "tokens_seen": 3081830400
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3344567485154872e-05,
      "loss": 1.2377,
      "theoretical_loss": 3.342374646129581,
      "tokens_seen": 3081961472
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.333654309099663e-05,
      "loss": 1.1717,
      "theoretical_loss": 3.34236386952328,
      "tokens_seen": 3082092544
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.332851869683839e-05,
      "loss": 1.2529,
      "theoretical_loss": 3.3423530935035815,
      "tokens_seen": 3082223616
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3320494302680148e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.34234231807043,
      "tokens_seen": 3082354688
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3312469908521907e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.3423315432237675,
      "tokens_seen": 3082485760
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3304445514363667e-05,
      "loss": 1.3006,
      "theoretical_loss": 3.342320768963538,
      "tokens_seen": 3082616832
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3296421120205425e-05,
      "loss": 1.255,
      "theoretical_loss": 3.3423099952896838,
      "tokens_seen": 3082747904
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3288396726047183e-05,
      "loss": 1.263,
      "theoretical_loss": 3.342299222202149,
      "tokens_seen": 3082878976
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3280372331888943e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.3422884497008756,
      "tokens_seen": 3083010048
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3272347937730703e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.342277677785808,
      "tokens_seen": 3083141120
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.326432354357246e-05,
      "loss": 1.2503,
      "theoretical_loss": 3.3422669064568886,
      "tokens_seen": 3083272192
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.325629914941422e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.3422561357140608,
      "tokens_seen": 3083403264
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3248274755255979e-05,
      "loss": 1.2543,
      "theoretical_loss": 3.342245365557268,
      "tokens_seen": 3083534336
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3240250361097739e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.342234595986453,
      "tokens_seen": 3083665408
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3232225966939496e-05,
      "loss": 1.2456,
      "theoretical_loss": 3.3422238270015594,
      "tokens_seen": 3083796480
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3224201572781255e-05,
      "loss": 1.2121,
      "theoretical_loss": 3.34221305860253,
      "tokens_seen": 3083927552
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3216177178623015e-05,
      "loss": 1.2937,
      "theoretical_loss": 3.3422022907893085,
      "tokens_seen": 3084058624
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3208152784464772e-05,
      "loss": 1.2465,
      "theoretical_loss": 3.3421915235618376,
      "tokens_seen": 3084189696
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3200128390306532e-05,
      "loss": 1.2761,
      "theoretical_loss": 3.3421807569200612,
      "tokens_seen": 3084320768
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3192103996148292e-05,
      "loss": 1.3032,
      "theoretical_loss": 3.342169990863922,
      "tokens_seen": 3084451840
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3184079601990052e-05,
      "loss": 1.2009,
      "theoretical_loss": 3.3421592253933636,
      "tokens_seen": 3084582912
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3176055207831808e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.342148460508329,
      "tokens_seen": 3084713984
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3168030813673568e-05,
      "loss": 1.3101,
      "theoretical_loss": 3.342137696208762,
      "tokens_seen": 3084845056
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3160006419515328e-05,
      "loss": 1.271,
      "theoretical_loss": 3.342126932494605,
      "tokens_seen": 3084976128
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3151982025357088e-05,
      "loss": 1.2919,
      "theoretical_loss": 3.342116169365802,
      "tokens_seen": 3085107200
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.9940102100372314,
      "objective/train/avg_token_score": 0.005989755503833294,
      "objective/train/docs_used": 1120785,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3332281112670898,
      "objective/train/original_loss": 1.342673659324646,
      "objective/train/theoretical_loss": 3.342105406822296,
      "objective/train/tokens_used": 1455697376,
      "objective/train/unlikelihood_loss": 0.014535754919052124,
      "theoretical_loss": 3.342105406822296,
      "tokens_seen": 3085238272
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3143957631198844e-05,
      "loss": 1.2729,
      "theoretical_loss": 3.342105406822296,
      "tokens_seen": 3085238272
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3135933237040604e-05,
      "loss": 1.3084,
      "theoretical_loss": 3.3420946448640305,
      "tokens_seen": 3085369344
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3127908842882364e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.3420838834909485,
      "tokens_seen": 3085500416
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.311988444872412e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.342073122702994,
      "tokens_seen": 3085631488
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.311186005456588e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.3420623625001094,
      "tokens_seen": 3085762560
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.310383566040764e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.342051602882239,
      "tokens_seen": 3085893632
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.30958112662494e-05,
      "loss": 1.2698,
      "theoretical_loss": 3.3420408438493254,
      "tokens_seen": 3086024704
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3087786872091157e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.342030085401312,
      "tokens_seen": 3086155776
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3079762477932917e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.3420193275381425,
      "tokens_seen": 3086286848
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3071738083774677e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.34200857025976,
      "tokens_seen": 3086417920
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3063713689616437e-05,
      "loss": 1.2111,
      "theoretical_loss": 3.3419978135661084,
      "tokens_seen": 3086548992
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3055689295458193e-05,
      "loss": 1.2214,
      "theoretical_loss": 3.3419870574571306,
      "tokens_seen": 3086680064
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3047664901299953e-05,
      "loss": 1.3002,
      "theoretical_loss": 3.34197630193277,
      "tokens_seen": 3086811136
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3039640507141713e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.34196554699297,
      "tokens_seen": 3086942208
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.303161611298347e-05,
      "loss": 1.1905,
      "theoretical_loss": 3.341954792637675,
      "tokens_seen": 3087073280
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.302359171882523e-05,
      "loss": 1.1751,
      "theoretical_loss": 3.3419440388668264,
      "tokens_seen": 3087204352
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3015567324666988e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.341933285680369,
      "tokens_seen": 3087335424
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.3007542930508748e-05,
      "loss": 1.2852,
      "theoretical_loss": 3.3419225330782467,
      "tokens_seen": 3087466496
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2999518536350506e-05,
      "loss": 1.224,
      "theoretical_loss": 3.3419117810604018,
      "tokens_seen": 3087597568
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2991494142192264e-05,
      "loss": 1.2494,
      "theoretical_loss": 3.341901029626778,
      "tokens_seen": 3087728640
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2983469748034024e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.341890278777319,
      "tokens_seen": 3087859712
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2975445353875784e-05,
      "loss": 1.2,
      "theoretical_loss": 3.3418795285119685,
      "tokens_seen": 3087990784
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.296742095971754e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.341868778830669,
      "tokens_seen": 3088121856
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.29593965655593e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.3418580297333653,
      "tokens_seen": 3088252928
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.295137217140106e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.34184728122,
      "tokens_seen": 3088384000
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2943347777242817e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.3418365332905173,
      "tokens_seen": 3088515072
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2935323383084577e-05,
      "loss": 1.2226,
      "theoretical_loss": 3.34182578594486,
      "tokens_seen": 3088646144
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2927298988926336e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.341815039182972,
      "tokens_seen": 3088777216
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2919274594768096e-05,
      "loss": 1.2858,
      "theoretical_loss": 3.3418042930047966,
      "tokens_seen": 3088908288
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2911250200609853e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.341793547410277,
      "tokens_seen": 3089039360
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2903225806451613e-05,
      "loss": 1.2094,
      "theoretical_loss": 3.341782802399358,
      "tokens_seen": 3089170432
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2895201412293373e-05,
      "loss": 1.2516,
      "theoretical_loss": 3.3417720579719816,
      "tokens_seen": 3089301504
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2887177018135133e-05,
      "loss": 1.2316,
      "theoretical_loss": 3.3417613141280924,
      "tokens_seen": 3089432576
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2879152623976889e-05,
      "loss": 1.2877,
      "theoretical_loss": 3.3417505708676334,
      "tokens_seen": 3089563648
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2871128229818649e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.3417398281905486,
      "tokens_seen": 3089694720
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2863103835660409e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.341729086096781,
      "tokens_seen": 3089825792
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2855079441502169e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.3417183445862753,
      "tokens_seen": 3089956864
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2847055047343925e-05,
      "loss": 1.218,
      "theoretical_loss": 3.341707603658974,
      "tokens_seen": 3090087936
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2839030653185685e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.341696863314821,
      "tokens_seen": 3090219008
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2831006259027445e-05,
      "loss": 1.1984,
      "theoretical_loss": 3.34168612355376,
      "tokens_seen": 3090350080
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2822981864869202e-05,
      "loss": 1.2205,
      "theoretical_loss": 3.3416753843757343,
      "tokens_seen": 3090481152
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2814957470710962e-05,
      "loss": 1.2035,
      "theoretical_loss": 3.341664645780688,
      "tokens_seen": 3090612224
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2806933076552721e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.3416539077685643,
      "tokens_seen": 3090743296
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2798908682394481e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.3416431703393075,
      "tokens_seen": 3090874368
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2790884288236238e-05,
      "loss": 1.1598,
      "theoretical_loss": 3.3416324334928604,
      "tokens_seen": 3091005440
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2782859894077998e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.3416216972291672,
      "tokens_seen": 3091136512
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2774835499919758e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.3416109615481715,
      "tokens_seen": 3091267584
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2766811105761518e-05,
      "loss": 1.232,
      "theoretical_loss": 3.341600226449817,
      "tokens_seen": 3091398656
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2758786711603274e-05,
      "loss": 1.199,
      "theoretical_loss": 3.3415894919340468,
      "tokens_seen": 3091529728
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2750762317445034e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.3415787580008054,
      "tokens_seen": 3091660800
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.9936481714248657,
      "objective/train/avg_token_score": 0.006351863499730825,
      "objective/train/docs_used": 1123259,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2616156339645386,
      "objective/train/original_loss": 1.2703931331634521,
      "objective/train/theoretical_loss": 3.341568024650036,
      "objective/train/tokens_used": 1462250976,
      "objective/train/unlikelihood_loss": 0.018029332160949707,
      "theoretical_loss": 3.341568024650036,
      "tokens_seen": 3091791872
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2742737923286794e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.341568024650036,
      "tokens_seen": 3091791872
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.273471352912855e-05,
      "loss": 1.2527,
      "theoretical_loss": 3.3415572918816823,
      "tokens_seen": 3091922944
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.272668913497031e-05,
      "loss": 1.3501,
      "theoretical_loss": 3.3415465596956886,
      "tokens_seen": 3092054016
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.271866474081207e-05,
      "loss": 1.3437,
      "theoretical_loss": 3.341535828091998,
      "tokens_seen": 3092185088
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2710640346653828e-05,
      "loss": 1.2536,
      "theoretical_loss": 3.341525097070554,
      "tokens_seen": 3092316160
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2702615952495587e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.3415143666313014,
      "tokens_seen": 3092447232
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2694591558337347e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.341503636774183,
      "tokens_seen": 3092578304
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2686567164179105e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.3414929074991426,
      "tokens_seen": 3092709376
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2678542770020865e-05,
      "loss": 1.3114,
      "theoretical_loss": 3.341482178806124,
      "tokens_seen": 3092840448
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2670518375862623e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.341471450695072,
      "tokens_seen": 3092971520
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2662493981704381e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3414607231659286,
      "tokens_seen": 3093102592
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2654469587546141e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.341449996218639,
      "tokens_seen": 3093233664
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.26464451933879e-05,
      "loss": 1.2541,
      "theoretical_loss": 3.3414392698531463,
      "tokens_seen": 3093364736
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2638420799229657e-05,
      "loss": 1.3098,
      "theoretical_loss": 3.3414285440693945,
      "tokens_seen": 3093495808
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.2630396405071417e-05,
      "loss": 1.34,
      "theoretical_loss": 3.3414178188673276,
      "tokens_seen": 3093626880
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2622372010913177e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.341407094246889,
      "tokens_seen": 3093757952
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2614347616754934e-05,
      "loss": 1.3575,
      "theoretical_loss": 3.341396370208023,
      "tokens_seen": 3093889024
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2606323222596694e-05,
      "loss": 1.2873,
      "theoretical_loss": 3.3413856467506724,
      "tokens_seen": 3094020096
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2598298828438454e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.3413749238747825,
      "tokens_seen": 3094151168
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2590274434280214e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.341364201580296,
      "tokens_seen": 3094282240
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.258225004012197e-05,
      "loss": 1.1926,
      "theoretical_loss": 3.3413534798671574,
      "tokens_seen": 3094413312
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.257422564596373e-05,
      "loss": 1.1838,
      "theoretical_loss": 3.34134275873531,
      "tokens_seen": 3094544384
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.256620125180549e-05,
      "loss": 1.2929,
      "theoretical_loss": 3.3413320381846985,
      "tokens_seen": 3094675456
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2558176857647246e-05,
      "loss": 1.23,
      "theoretical_loss": 3.341321318215266,
      "tokens_seen": 3094806528
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2550152463489006e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.3413105988269565,
      "tokens_seen": 3094937600
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2542128069330766e-05,
      "loss": 1.2682,
      "theoretical_loss": 3.3412998800197142,
      "tokens_seen": 3095068672
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2534103675172526e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.341289161793483,
      "tokens_seen": 3095199744
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2526079281014283e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.341278444148206,
      "tokens_seen": 3095330816
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2518054886856043e-05,
      "loss": 1.3755,
      "theoretical_loss": 3.341267727083828,
      "tokens_seen": 3095461888
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2510030492697802e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.341257010600293,
      "tokens_seen": 3095592960
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2502006098539562e-05,
      "loss": 1.2517,
      "theoretical_loss": 3.3412462946975445,
      "tokens_seen": 3095724032
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.249398170438132e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.341235579375526,
      "tokens_seen": 3095855104
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2485957310223079e-05,
      "loss": 1.3005,
      "theoretical_loss": 3.3412248646341824,
      "tokens_seen": 3095986176
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2477932916064839e-05,
      "loss": 1.343,
      "theoretical_loss": 3.341214150473457,
      "tokens_seen": 3096117248
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2469908521906597e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.341203436893294,
      "tokens_seen": 3096248320
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2461884127748355e-05,
      "loss": 1.2248,
      "theoretical_loss": 3.341192723893638,
      "tokens_seen": 3096379392
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2453859733590115e-05,
      "loss": 1.2209,
      "theoretical_loss": 3.3411820114744315,
      "tokens_seen": 3096510464
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2445835339431873e-05,
      "loss": 1.266,
      "theoretical_loss": 3.3411712996356195,
      "tokens_seen": 3096641536
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2437810945273633e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.341160588377146,
      "tokens_seen": 3096772608
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2429786551115391e-05,
      "loss": 1.2173,
      "theoretical_loss": 3.341149877698955,
      "tokens_seen": 3096903680
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2421762156957151e-05,
      "loss": 1.2296,
      "theoretical_loss": 3.34113916760099,
      "tokens_seen": 3097034752
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.241373776279891e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.341128458083195,
      "tokens_seen": 3097165824
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.240571336864067e-05,
      "loss": 1.2609,
      "theoretical_loss": 3.3411177491455146,
      "tokens_seen": 3097296896
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2397688974482428e-05,
      "loss": 1.2924,
      "theoretical_loss": 3.3411070407878927,
      "tokens_seen": 3097427968
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2389664580324187e-05,
      "loss": 1.1931,
      "theoretical_loss": 3.341096333010273,
      "tokens_seen": 3097559040
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2381640186165946e-05,
      "loss": 1.3015,
      "theoretical_loss": 3.3410856258125996,
      "tokens_seen": 3097690112
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2373615792007704e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.341074919194817,
      "tokens_seen": 3097821184
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2365591397849464e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.341064213156869,
      "tokens_seen": 3097952256
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2357567003691222e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.3410535076986996,
      "tokens_seen": 3098083328
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2349542609532982e-05,
      "loss": 1.3381,
      "theoretical_loss": 3.3410428028202523,
      "tokens_seen": 3098214400
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.9929050207138062,
      "objective/train/avg_token_score": 0.007094961125403643,
      "objective/train/docs_used": 1125757,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3266174793243408,
      "objective/train/original_loss": 1.3368617296218872,
      "objective/train/theoretical_loss": 3.3410320985214725,
      "objective/train/tokens_used": 1468804576,
      "objective/train/unlikelihood_loss": 0.01649756170809269,
      "theoretical_loss": 3.3410320985214725,
      "tokens_seen": 3098345472
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.234151821537474e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3410320985214725,
      "tokens_seen": 3098345472
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2333493821216498e-05,
      "loss": 1.2716,
      "theoretical_loss": 3.3410213948023033,
      "tokens_seen": 3098476544
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2325469427058258e-05,
      "loss": 1.2439,
      "theoretical_loss": 3.3410106916626887,
      "tokens_seen": 3098607616
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2317445032900016e-05,
      "loss": 1.2796,
      "theoretical_loss": 3.3409999891025737,
      "tokens_seen": 3098738688
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2309420638741775e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.340989287121902,
      "tokens_seen": 3098869760
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2301396244583535e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.340978585720617,
      "tokens_seen": 3099000832
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2293371850425293e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.3409678848986637,
      "tokens_seen": 3099131904
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2285347456267051e-05,
      "loss": 1.3602,
      "theoretical_loss": 3.340957184655986,
      "tokens_seen": 3099262976
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2277323062108811e-05,
      "loss": 1.258,
      "theoretical_loss": 3.340946484992528,
      "tokens_seen": 3099394048
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2269298667950569e-05,
      "loss": 1.319,
      "theoretical_loss": 3.340935785908234,
      "tokens_seen": 3099525120
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2261274273792329e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.340925087403048,
      "tokens_seen": 3099656192
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2253249879634087e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.3409143894769144,
      "tokens_seen": 3099787264
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2245225485475847e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.3409036921297766,
      "tokens_seen": 3099918336
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2237201091317605e-05,
      "loss": 1.32,
      "theoretical_loss": 3.34089299536158,
      "tokens_seen": 3100049408
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2229176697159365e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.340882299172268,
      "tokens_seen": 3100180480
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2221152303001123e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.3408716035617845,
      "tokens_seen": 3100311552
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2213127908842883e-05,
      "loss": 1.2518,
      "theoretical_loss": 3.3408609085300744,
      "tokens_seen": 3100442624
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2205103514684642e-05,
      "loss": 1.2687,
      "theoretical_loss": 3.340850214077082,
      "tokens_seen": 3100573696
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2197079120526401e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.340839520202751,
      "tokens_seen": 3100704768
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.218905472636816e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.3408288269070257,
      "tokens_seen": 3100835840
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2181030332209918e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.340818134189851,
      "tokens_seen": 3100966912
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2173005938051678e-05,
      "loss": 1.246,
      "theoretical_loss": 3.3408074420511698,
      "tokens_seen": 3101097984
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2164981543893436e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.3407967504909277,
      "tokens_seen": 3101229056
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2156957149735196e-05,
      "loss": 1.2227,
      "theoretical_loss": 3.3407860595090684,
      "tokens_seen": 3101360128
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2148932755576954e-05,
      "loss": 1.1995,
      "theoretical_loss": 3.3407753691055357,
      "tokens_seen": 3101491200
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2140908361418714e-05,
      "loss": 1.2179,
      "theoretical_loss": 3.3407646792802748,
      "tokens_seen": 3101622272
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2132883967260472e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.3407539900332295,
      "tokens_seen": 3101753344
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2124859573102232e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.3407433013643444,
      "tokens_seen": 3101884416
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.211683517894399e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.3407326132735635,
      "tokens_seen": 3102015488
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.210881078478575e-05,
      "loss": 1.238,
      "theoretical_loss": 3.340721925760831,
      "tokens_seen": 3102146560
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2100786390627508e-05,
      "loss": 1.2017,
      "theoretical_loss": 3.340711238826091,
      "tokens_seen": 3102277632
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2092761996469267e-05,
      "loss": 1.3017,
      "theoretical_loss": 3.3407005524692885,
      "tokens_seen": 3102408704
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2084737602311027e-05,
      "loss": 1.2805,
      "theoretical_loss": 3.3406898666903677,
      "tokens_seen": 3102539776
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2076713208152785e-05,
      "loss": 1.318,
      "theoretical_loss": 3.3406791814892727,
      "tokens_seen": 3102670848
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2068688813994545e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.340668496865948,
      "tokens_seen": 3102801920
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2060664419836303e-05,
      "loss": 1.1988,
      "theoretical_loss": 3.3406578128203375,
      "tokens_seen": 3102932992
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2052640025678063e-05,
      "loss": 1.1803,
      "theoretical_loss": 3.3406471293523863,
      "tokens_seen": 3103064064
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2044615631519821e-05,
      "loss": 1.3144,
      "theoretical_loss": 3.340636446462038,
      "tokens_seen": 3103195136
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2036591237361581e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.340625764149238,
      "tokens_seen": 3103326208
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.202856684320334e-05,
      "loss": 1.2383,
      "theoretical_loss": 3.3406150824139296,
      "tokens_seen": 3103457280
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2020542449045099e-05,
      "loss": 1.2204,
      "theoretical_loss": 3.3406044012560576,
      "tokens_seen": 3103588352
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2012518054886857e-05,
      "loss": 1.2654,
      "theoretical_loss": 3.3405937206755665,
      "tokens_seen": 3103719424
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.2004493660728616e-05,
      "loss": 1.1816,
      "theoretical_loss": 3.340583040672401,
      "tokens_seen": 3103850496
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1996469266570375e-05,
      "loss": 1.3668,
      "theoretical_loss": 3.340572361246505,
      "tokens_seen": 3103981568
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1988444872412134e-05,
      "loss": 1.2352,
      "theoretical_loss": 3.340561682397823,
      "tokens_seen": 3104112640
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1980420478253892e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.3405510041263,
      "tokens_seen": 3104243712
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.197239608409565e-05,
      "loss": 1.2413,
      "theoretical_loss": 3.340540326431879,
      "tokens_seen": 3104374784
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.196437168993741e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.3405296493145062,
      "tokens_seen": 3104505856
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1956347295779168e-05,
      "loss": 1.2013,
      "theoretical_loss": 3.340518972774125,
      "tokens_seen": 3104636928
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1948322901620928e-05,
      "loss": 1.2219,
      "theoretical_loss": 3.3405082968106803,
      "tokens_seen": 3104768000
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.9935815930366516,
      "objective/train/avg_token_score": 0.006418364122509956,
      "objective/train/docs_used": 1128234,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2474299669265747,
      "objective/train/original_loss": 1.2566063404083252,
      "objective/train/theoretical_loss": 3.340497621424116,
      "objective/train/tokens_used": 1475358176,
      "objective/train/unlikelihood_loss": 0.017145482823252678,
      "theoretical_loss": 3.340497621424116,
      "tokens_seen": 3104899072
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1940298507462686e-05,
      "loss": 1.336,
      "theoretical_loss": 3.340497621424116,
      "tokens_seen": 3104899072
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1932274113304446e-05,
      "loss": 1.1944,
      "theoretical_loss": 3.3404869466143774,
      "tokens_seen": 3105030144
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1924249719146204e-05,
      "loss": 1.2713,
      "theoretical_loss": 3.3404762723814083,
      "tokens_seen": 3105161216
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1916225324987964e-05,
      "loss": 1.306,
      "theoretical_loss": 3.340465598725154,
      "tokens_seen": 3105292288
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1908200930829723e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.340454925645558,
      "tokens_seen": 3105423360
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.190017653667148e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.3404442531425653,
      "tokens_seen": 3105554432
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.189215214251324e-05,
      "loss": 1.2365,
      "theoretical_loss": 3.34043358121612,
      "tokens_seen": 3105685504
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1884127748354999e-05,
      "loss": 1.3698,
      "theoretical_loss": 3.3404229098661675,
      "tokens_seen": 3105816576
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1876103354196759e-05,
      "loss": 1.1974,
      "theoretical_loss": 3.3404122390926516,
      "tokens_seen": 3105947648
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1868078960038517e-05,
      "loss": 1.2967,
      "theoretical_loss": 3.3404015688955173,
      "tokens_seen": 3106078720
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1860054565880277e-05,
      "loss": 1.2564,
      "theoretical_loss": 3.340390899274709,
      "tokens_seen": 3106209792
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1852030171722035e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.340380230230171,
      "tokens_seen": 3106340864
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1844005777563795e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.340369561761848,
      "tokens_seen": 3106471936
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1835981383405553e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3403588938696847,
      "tokens_seen": 3106603008
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1827956989247313e-05,
      "loss": 1.249,
      "theoretical_loss": 3.3403482265536253,
      "tokens_seen": 3106734080
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1819932595089071e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.340337559813615,
      "tokens_seen": 3106865152
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.181190820093083e-05,
      "loss": 1.2976,
      "theoretical_loss": 3.340326893649598,
      "tokens_seen": 3106996224
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.180388380677259e-05,
      "loss": 1.1961,
      "theoretical_loss": 3.3403162280615186,
      "tokens_seen": 3107127296
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1795859412614348e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.340305563049322,
      "tokens_seen": 3107258368
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1787835018456108e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.340294898612953,
      "tokens_seen": 3107389440
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1779810624297866e-05,
      "loss": 1.2365,
      "theoretical_loss": 3.3402842347523554,
      "tokens_seen": 3107520512
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1771786230139626e-05,
      "loss": 1.3359,
      "theoretical_loss": 3.340273571467474,
      "tokens_seen": 3107651584
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1763761835981384e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.3402629087582545,
      "tokens_seen": 3107782656
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1755737441823144e-05,
      "loss": 1.3449,
      "theoretical_loss": 3.34025224662464,
      "tokens_seen": 3107913728
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1747713047664902e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.340241585066576,
      "tokens_seen": 3108044800
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1739688653506662e-05,
      "loss": 1.2544,
      "theoretical_loss": 3.340230924084007,
      "tokens_seen": 3108175872
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.173166425934842e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.3402202636768776,
      "tokens_seen": 3108306944
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.172363986519018e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.3402096038451328,
      "tokens_seen": 3108438016
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1715615471031938e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.3401989445887175,
      "tokens_seen": 3108569088
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1707591076873696e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.3401882859075753,
      "tokens_seen": 3108700160
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1699566682715456e-05,
      "loss": 1.2898,
      "theoretical_loss": 3.3401776278016517,
      "tokens_seen": 3108831232
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1691542288557215e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.340166970270891,
      "tokens_seen": 3108962304
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1683517894398974e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.340156313315239,
      "tokens_seen": 3109093376
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1675493500240733e-05,
      "loss": 1.2606,
      "theoretical_loss": 3.340145656934639,
      "tokens_seen": 3109224448
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1667469106082491e-05,
      "loss": 1.3687,
      "theoretical_loss": 3.3401350011290365,
      "tokens_seen": 3109355520
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.165944471192425e-05,
      "loss": 1.271,
      "theoretical_loss": 3.340124345898376,
      "tokens_seen": 3109486592
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1651420317766009e-05,
      "loss": 1.2856,
      "theoretical_loss": 3.340113691242602,
      "tokens_seen": 3109617664
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1643395923607767e-05,
      "loss": 1.2834,
      "theoretical_loss": 3.34010303716166,
      "tokens_seen": 3109748736
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1635371529449527e-05,
      "loss": 1.3048,
      "theoretical_loss": 3.3400923836554943,
      "tokens_seen": 3109879808
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1627347135291285e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.3400817307240493,
      "tokens_seen": 3110010880
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.1619322741133044e-05,
      "loss": 1.2364,
      "theoretical_loss": 3.3400710783672705,
      "tokens_seen": 3110141952
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1611298346974803e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.3400604265851026,
      "tokens_seen": 3110273024
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1603273952816562e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.34004977537749,
      "tokens_seen": 3110404096
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1595249558658322e-05,
      "loss": 1.3731,
      "theoretical_loss": 3.3400391247443775,
      "tokens_seen": 3110535168
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.158722516450008e-05,
      "loss": 1.2382,
      "theoretical_loss": 3.3400284746857096,
      "tokens_seen": 3110666240
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.157920077034184e-05,
      "loss": 1.2724,
      "theoretical_loss": 3.340017825201432,
      "tokens_seen": 3110797312
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1571176376183598e-05,
      "loss": 1.2433,
      "theoretical_loss": 3.3400071762914894,
      "tokens_seen": 3110928384
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1563151982025358e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.339996527955826,
      "tokens_seen": 3111059456
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1555127587867116e-05,
      "loss": 1.2955,
      "theoretical_loss": 3.339985880194387,
      "tokens_seen": 3111190528
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1547103193708876e-05,
      "loss": 1.2149,
      "theoretical_loss": 3.339975233007117,
      "tokens_seen": 3111321600
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.9921447038650513,
      "objective/train/avg_token_score": 0.007855276577174664,
      "objective/train/docs_used": 1130437,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2592774629592896,
      "objective/train/original_loss": 1.2692452669143677,
      "objective/train/theoretical_loss": 3.339964586393961,
      "objective/train/tokens_used": 1481911776,
      "objective/train/unlikelihood_loss": 0.022835925221443176,
      "theoretical_loss": 3.339964586393961,
      "tokens_seen": 3111452672
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1539078799550634e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.339964586393961,
      "tokens_seen": 3111452672
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1531054405392392e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.3399539403548646,
      "tokens_seen": 3111583744
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1523030011234152e-05,
      "loss": 1.208,
      "theoretical_loss": 3.3399432948897716,
      "tokens_seen": 3111714816
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.151500561707591e-05,
      "loss": 1.2512,
      "theoretical_loss": 3.339932649998627,
      "tokens_seen": 3111845888
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.150698122291767e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.339922005681376,
      "tokens_seen": 3111976960
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1498956828759429e-05,
      "loss": 1.2595,
      "theoretical_loss": 3.339911361937964,
      "tokens_seen": 3112108032
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1490932434601188e-05,
      "loss": 1.2279,
      "theoretical_loss": 3.3399007187683347,
      "tokens_seen": 3112239104
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1482908040442947e-05,
      "loss": 1.3701,
      "theoretical_loss": 3.3398900761724333,
      "tokens_seen": 3112370176
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1474883646284707e-05,
      "loss": 1.3367,
      "theoretical_loss": 3.339879434150206,
      "tokens_seen": 3112501248
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1466859252126465e-05,
      "loss": 1.1895,
      "theoretical_loss": 3.339868792701596,
      "tokens_seen": 3112632320
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1458834857968225e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.339858151826549,
      "tokens_seen": 3112763392
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1450810463809983e-05,
      "loss": 1.2584,
      "theoretical_loss": 3.3398475115250106,
      "tokens_seen": 3112894464
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1442786069651743e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.3398368717969245,
      "tokens_seen": 3113025536
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1434761675493501e-05,
      "loss": 1.3213,
      "theoretical_loss": 3.3398262326422365,
      "tokens_seen": 3113156608
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.142673728133526e-05,
      "loss": 1.3084,
      "theoretical_loss": 3.3398155940608913,
      "tokens_seen": 3113287680
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.141871288717702e-05,
      "loss": 1.3233,
      "theoretical_loss": 3.339804956052834,
      "tokens_seen": 3113418752
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1410688493018777e-05,
      "loss": 1.2234,
      "theoretical_loss": 3.339794318618009,
      "tokens_seen": 3113549824
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1402664098860537e-05,
      "loss": 1.2902,
      "theoretical_loss": 3.3397836817563618,
      "tokens_seen": 3113680896
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1394639704702296e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.339773045467837,
      "tokens_seen": 3113811968
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1386615310544055e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.339762409752381,
      "tokens_seen": 3113943040
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1378590916385814e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.3397517746099363,
      "tokens_seen": 3114074112
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1370566522227574e-05,
      "loss": 1.2734,
      "theoretical_loss": 3.33974114004045,
      "tokens_seen": 3114205184
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1362542128069332e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.3397305060438667,
      "tokens_seen": 3114336256
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1354517733911092e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.3397198726201305,
      "tokens_seen": 3114467328
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.134649333975285e-05,
      "loss": 1.2865,
      "theoretical_loss": 3.3397092397691877,
      "tokens_seen": 3114598400
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1338468945594608e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.3396986074909822,
      "tokens_seen": 3114729472
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1330444551436368e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.33968797578546,
      "tokens_seen": 3114860544
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1322420157278126e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.339677344652565,
      "tokens_seen": 3114991616
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1314395763119884e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.339666714092244,
      "tokens_seen": 3115122688
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1306371368961644e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.3396560841044405,
      "tokens_seen": 3115253760
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1298346974803403e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.3396454546891006,
      "tokens_seen": 3115384832
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.129032258064516e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.3396348258461686,
      "tokens_seen": 3115515904
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.128229818648692e-05,
      "loss": 1.3049,
      "theoretical_loss": 3.33962419757559,
      "tokens_seen": 3115646976
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1274273792328679e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3396135698773097,
      "tokens_seen": 3115778048
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1266249398170439e-05,
      "loss": 1.3157,
      "theoretical_loss": 3.339602942751273,
      "tokens_seen": 3115909120
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1258225004012197e-05,
      "loss": 1.3243,
      "theoretical_loss": 3.339592316197425,
      "tokens_seen": 3116040192
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1250200609853957e-05,
      "loss": 1.2889,
      "theoretical_loss": 3.3395816902157107,
      "tokens_seen": 3116171264
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1242176215695715e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3395710648060755,
      "tokens_seen": 3116302336
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1234151821537473e-05,
      "loss": 1.2855,
      "theoretical_loss": 3.339560439968464,
      "tokens_seen": 3116433408
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1226127427379233e-05,
      "loss": 1.2918,
      "theoretical_loss": 3.3395498157028225,
      "tokens_seen": 3116564480
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1218103033220991e-05,
      "loss": 1.37,
      "theoretical_loss": 3.3395391920090947,
      "tokens_seen": 3116695552
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1210078639062751e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.3395285688872267,
      "tokens_seen": 3116826624
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.120205424490451e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.339517946337163,
      "tokens_seen": 3116957696
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.119402985074627e-05,
      "loss": 1.3481,
      "theoretical_loss": 3.33950732435885,
      "tokens_seen": 3117088768
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1186005456588028e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.339496702952231,
      "tokens_seen": 3117219840
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1177981062429788e-05,
      "loss": 1.2723,
      "theoretical_loss": 3.3394860821172534,
      "tokens_seen": 3117350912
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1169956668271546e-05,
      "loss": 1.2997,
      "theoretical_loss": 3.3394754618538607,
      "tokens_seen": 3117481984
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1161932274113306e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.3394648421619983,
      "tokens_seen": 3117613056
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1153907879955064e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.3394542230416127,
      "tokens_seen": 3117744128
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1145883485796822e-05,
      "loss": 1.3148,
      "theoretical_loss": 3.3394436044926477,
      "tokens_seen": 3117875200
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.9936418533325195,
      "objective/train/avg_token_score": 0.006358125247061253,
      "objective/train/docs_used": 1132901,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.4101524353027344,
      "objective/train/original_loss": 1.4198355674743652,
      "objective/train/theoretical_loss": 3.3394329865150487,
      "objective/train/tokens_used": 1488465376,
      "objective/train/unlikelihood_loss": 0.01523058395832777,
      "theoretical_loss": 3.3394329865150487,
      "tokens_seen": 3118006272
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1137859091638582e-05,
      "loss": 1.4135,
      "theoretical_loss": 3.3394329865150487,
      "tokens_seen": 3118006272
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.112983469748034e-05,
      "loss": 1.3815,
      "theoretical_loss": 3.3394223691087617,
      "tokens_seen": 3118137344
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.11218103033221e-05,
      "loss": 1.305,
      "theoretical_loss": 3.3394117522737314,
      "tokens_seen": 3118268416
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1113785909163858e-05,
      "loss": 1.3181,
      "theoretical_loss": 3.3394011360099034,
      "tokens_seen": 3118399488
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1105761515005618e-05,
      "loss": 1.328,
      "theoretical_loss": 3.339390520317223,
      "tokens_seen": 3118530560
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1097737120847376e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.3393799051956345,
      "tokens_seen": 3118661632
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1089712726689136e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.3393692906450845,
      "tokens_seen": 3118792704
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1081688332530895e-05,
      "loss": 1.3585,
      "theoretical_loss": 3.3393586766655172,
      "tokens_seen": 3118923776
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1073663938372654e-05,
      "loss": 1.3474,
      "theoretical_loss": 3.3393480632568786,
      "tokens_seen": 3119054848
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1065639544214413e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.339337450419114,
      "tokens_seen": 3119185920
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1057615150056171e-05,
      "loss": 1.3511,
      "theoretical_loss": 3.339326838152168,
      "tokens_seen": 3119316992
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.104959075589793e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3393162264559866,
      "tokens_seen": 3119448064
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1041566361739689e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.339305615330515,
      "tokens_seen": 3119579136
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1033541967581449e-05,
      "loss": 1.294,
      "theoretical_loss": 3.3392950047756984,
      "tokens_seen": 3119710208
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1025517573423207e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.3392843947914823,
      "tokens_seen": 3119841280
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1017493179264967e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.3392737853778116,
      "tokens_seen": 3119972352
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1009468785106725e-05,
      "loss": 1.3392,
      "theoretical_loss": 3.339263176534632,
      "tokens_seen": 3120103424
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.1001444390948485e-05,
      "loss": 1.3768,
      "theoretical_loss": 3.339252568261889,
      "tokens_seen": 3120234496
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0993419996790243e-05,
      "loss": 1.2935,
      "theoretical_loss": 3.339241960559528,
      "tokens_seen": 3120365568
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0985395602632002e-05,
      "loss": 1.301,
      "theoretical_loss": 3.339231353427494,
      "tokens_seen": 3120496640
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0977371208473761e-05,
      "loss": 1.3503,
      "theoretical_loss": 3.3392207468657324,
      "tokens_seen": 3120627712
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.096934681431552e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.339210140874189,
      "tokens_seen": 3120758784
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0961322420157278e-05,
      "loss": 1.3531,
      "theoretical_loss": 3.339199535452809,
      "tokens_seen": 3120889856
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0953298025999038e-05,
      "loss": 1.2995,
      "theoretical_loss": 3.3391889306015377,
      "tokens_seen": 3121020928
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0945273631840796e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.3391783263203205,
      "tokens_seen": 3121152000
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0937249237682554e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.339167722609103,
      "tokens_seen": 3121283072
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0929224843524314e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.33915711946783,
      "tokens_seen": 3121414144
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0921200449366072e-05,
      "loss": 1.319,
      "theoretical_loss": 3.339146516896448,
      "tokens_seen": 3121545216
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0913176055207832e-05,
      "loss": 1.3942,
      "theoretical_loss": 3.339135914894902,
      "tokens_seen": 3121676288
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.090515166104959e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.339125313463137,
      "tokens_seen": 3121807360
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.089712726689135e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3391147126010985,
      "tokens_seen": 3121938432
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0889102872733109e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.339104112308733,
      "tokens_seen": 3122069504
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0881078478574869e-05,
      "loss": 1.37,
      "theoretical_loss": 3.3390935125859844,
      "tokens_seen": 3122200576
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0873054084416627e-05,
      "loss": 1.2809,
      "theoretical_loss": 3.3390829134328,
      "tokens_seen": 3122331648
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0865029690258385e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.3390723148491235,
      "tokens_seen": 3122462720
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0857005296100145e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.3390617168349017,
      "tokens_seen": 3122593792
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0848980901941903e-05,
      "loss": 1.3124,
      "theoretical_loss": 3.339051119390079,
      "tokens_seen": 3122724864
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0840956507783663e-05,
      "loss": 1.3044,
      "theoretical_loss": 3.339040522514602,
      "tokens_seen": 3122855936
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0832932113625421e-05,
      "loss": 1.3604,
      "theoretical_loss": 3.3390299262084158,
      "tokens_seen": 3122987008
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0824907719467181e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.3390193304714657,
      "tokens_seen": 3123118080
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.081688332530894e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.3390087353036972,
      "tokens_seen": 3123249152
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.08088589311507e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.338998140705056,
      "tokens_seen": 3123380224
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0800834536992457e-05,
      "loss": 1.2698,
      "theoretical_loss": 3.3389875466754884,
      "tokens_seen": 3123511296
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0792810142834217e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.3389769532149387,
      "tokens_seen": 3123642368
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0784785748675976e-05,
      "loss": 1.3728,
      "theoretical_loss": 3.3389663603233526,
      "tokens_seen": 3123773440
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0776761354517735e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.3389557680006767,
      "tokens_seen": 3123904512
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0768736960359494e-05,
      "loss": 1.3427,
      "theoretical_loss": 3.3389451762468556,
      "tokens_seen": 3124035584
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0760712566201252e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.3389345850618355,
      "tokens_seen": 3124166656
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0752688172043012e-05,
      "loss": 1.1888,
      "theoretical_loss": 3.3389239944455613,
      "tokens_seen": 3124297728
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.074466377788477e-05,
      "loss": 1.3043,
      "theoretical_loss": 3.3389134043979793,
      "tokens_seen": 3124428800
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.9928231239318848,
      "objective/train/avg_token_score": 0.00717685092240572,
      "objective/train/docs_used": 1135223,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.273779273033142,
      "objective/train/original_loss": 1.2841800451278687,
      "objective/train/theoretical_loss": 3.338902814919035,
      "objective/train/tokens_used": 1495018976,
      "objective/train/unlikelihood_loss": 0.019203534349799156,
      "theoretical_loss": 3.338902814919035,
      "tokens_seen": 3124559872
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.073663938372653e-05,
      "loss": 1.2758,
      "theoretical_loss": 3.338902814919035,
      "tokens_seen": 3124559872
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0728614989568288e-05,
      "loss": 1.2914,
      "theoretical_loss": 3.3388922260086735,
      "tokens_seen": 3124690944
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0720590595410048e-05,
      "loss": 1.2171,
      "theoretical_loss": 3.338881637666841,
      "tokens_seen": 3124822016
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0712566201251806e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.3388710498934824,
      "tokens_seen": 3124953088
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0704541807093566e-05,
      "loss": 1.2777,
      "theoretical_loss": 3.3388604626885448,
      "tokens_seen": 3125084160
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0696517412935324e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.338849876051972,
      "tokens_seen": 3125215232
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0688493018777084e-05,
      "loss": 1.3679,
      "theoretical_loss": 3.338839289983711,
      "tokens_seen": 3125346304
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0680468624618842e-05,
      "loss": 1.2504,
      "theoretical_loss": 3.3388287044837064,
      "tokens_seen": 3125477376
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.06724442304606e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.338818119551905,
      "tokens_seen": 3125608448
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.066441983630236e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.338807535188252,
      "tokens_seen": 3125739520
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0656395442144119e-05,
      "loss": 1.2773,
      "theoretical_loss": 3.3387969513926925,
      "tokens_seen": 3125870592
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0648371047985877e-05,
      "loss": 1.2814,
      "theoretical_loss": 3.338786368165173,
      "tokens_seen": 3126001664
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0640346653827637e-05,
      "loss": 1.2285,
      "theoretical_loss": 3.338775785505639,
      "tokens_seen": 3126132736
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0632322259669395e-05,
      "loss": 1.315,
      "theoretical_loss": 3.338765203414036,
      "tokens_seen": 3126263808
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0624297865511153e-05,
      "loss": 1.3117,
      "theoretical_loss": 3.3387546218903097,
      "tokens_seen": 3126394880
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0616273471352913e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.338744040934406,
      "tokens_seen": 3126525952
    },
    {
      "epoch": 0.89,
      "learning_rate": 1.0608249077194671e-05,
      "loss": 1.323,
      "theoretical_loss": 3.3387334605462704,
      "tokens_seen": 3126657024
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0600224683036431e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.3387228807258493,
      "tokens_seen": 3126788096
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.059220028887819e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.3387123014730875,
      "tokens_seen": 3126919168
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0584175894719948e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.338701722787931,
      "tokens_seen": 3127050240
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0576151500561708e-05,
      "loss": 1.282,
      "theoretical_loss": 3.3386911446703262,
      "tokens_seen": 3127181312
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0568127106403466e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.3386805671202184,
      "tokens_seen": 3127312384
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0560102712245226e-05,
      "loss": 1.3119,
      "theoretical_loss": 3.338669990137553,
      "tokens_seen": 3127443456
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0552078318086984e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3386594137222767,
      "tokens_seen": 3127574528
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0544053923928744e-05,
      "loss": 1.3288,
      "theoretical_loss": 3.338648837874334,
      "tokens_seen": 3127705600
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0536029529770502e-05,
      "loss": 1.3517,
      "theoretical_loss": 3.338638262593672,
      "tokens_seen": 3127836672
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0528005135612262e-05,
      "loss": 1.3109,
      "theoretical_loss": 3.338627687880236,
      "tokens_seen": 3127967744
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.051998074145402e-05,
      "loss": 1.2235,
      "theoretical_loss": 3.3386171137339713,
      "tokens_seen": 3128098816
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.051195634729578e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3386065401548244,
      "tokens_seen": 3128229888
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0503931953137538e-05,
      "loss": 1.253,
      "theoretical_loss": 3.338595967142741,
      "tokens_seen": 3128360960
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0495907558979298e-05,
      "loss": 1.3636,
      "theoretical_loss": 3.3385853946976667,
      "tokens_seen": 3128492032
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0487883164821056e-05,
      "loss": 1.2763,
      "theoretical_loss": 3.3385748228195475,
      "tokens_seen": 3128623104
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0479858770662815e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.338564251508329,
      "tokens_seen": 3128754176
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0471834376504575e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.3385536807639578,
      "tokens_seen": 3128885248
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0463809982346333e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.3385431105863788,
      "tokens_seen": 3129016320
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0455785588188093e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.3385325409755384,
      "tokens_seen": 3129147392
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0447761194029851e-05,
      "loss": 1.2539,
      "theoretical_loss": 3.338521971931382,
      "tokens_seen": 3129278464
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.043973679987161e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.3385114034538566,
      "tokens_seen": 3129409536
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0431712405713369e-05,
      "loss": 1.272,
      "theoretical_loss": 3.3385008355429067,
      "tokens_seen": 3129540608
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0423688011555129e-05,
      "loss": 1.2923,
      "theoretical_loss": 3.338490268198479,
      "tokens_seen": 3129671680
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0415663617396887e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.3384797014205194,
      "tokens_seen": 3129802752
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0407639223238647e-05,
      "loss": 1.3851,
      "theoretical_loss": 3.3384691352089737,
      "tokens_seen": 3129933824
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0399614829080405e-05,
      "loss": 1.3838,
      "theoretical_loss": 3.3384585695637874,
      "tokens_seen": 3130064896
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0391590434922163e-05,
      "loss": 1.2338,
      "theoretical_loss": 3.338448004484907,
      "tokens_seen": 3130195968
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0383566040763923e-05,
      "loss": 1.3346,
      "theoretical_loss": 3.3384374399722785,
      "tokens_seen": 3130327040
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0375541646605682e-05,
      "loss": 1.3551,
      "theoretical_loss": 3.3384268760258475,
      "tokens_seen": 3130458112
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0367517252447442e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3384163126455597,
      "tokens_seen": 3130589184
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.03594928582892e-05,
      "loss": 1.2636,
      "theoretical_loss": 3.3384057498313617,
      "tokens_seen": 3130720256
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.035146846413096e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.338395187583199,
      "tokens_seen": 3130851328
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0343444069972718e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.3383846259010177,
      "tokens_seen": 3130982400
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.9953626990318298,
      "objective/train/avg_token_score": 0.004637277219444513,
      "objective/train/docs_used": 1137606,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1959806680679321,
      "objective/train/original_loss": 1.2014527320861816,
      "objective/train/theoretical_loss": 3.338374064784764,
      "objective/train/tokens_used": 1501572576,
      "objective/train/unlikelihood_loss": 0.015449739061295986,
      "theoretical_loss": 3.338374064784764,
      "tokens_seen": 3131113472
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0335419675814478e-05,
      "loss": 1.254,
      "theoretical_loss": 3.338374064784764,
      "tokens_seen": 3131113472
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0327395281656236e-05,
      "loss": 1.2495,
      "theoretical_loss": 3.3383635042343833,
      "tokens_seen": 3131244544
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0319370887497994e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.3383529442498228,
      "tokens_seen": 3131375616
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0311346493339754e-05,
      "loss": 1.2128,
      "theoretical_loss": 3.338342384831027,
      "tokens_seen": 3131506688
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0303322099181512e-05,
      "loss": 1.3126,
      "theoretical_loss": 3.3383318259779426,
      "tokens_seen": 3131637760
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.029529770502327e-05,
      "loss": 1.2888,
      "theoretical_loss": 3.3383212676905156,
      "tokens_seen": 3131768832
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.028727331086503e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.3383107099686926,
      "tokens_seen": 3131899904
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0279248916706789e-05,
      "loss": 1.218,
      "theoretical_loss": 3.3383001528124185,
      "tokens_seen": 3132030976
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0271224522548547e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.3382895962216406,
      "tokens_seen": 3132162048
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0263200128390307e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.3382790401963036,
      "tokens_seen": 3132293120
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0255175734232065e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.3382684847363544,
      "tokens_seen": 3132424192
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0247151340073825e-05,
      "loss": 1.2945,
      "theoretical_loss": 3.3382579298417387,
      "tokens_seen": 3132555264
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0239126945915583e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.338247375512403,
      "tokens_seen": 3132686336
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0231102551757343e-05,
      "loss": 1.3211,
      "theoretical_loss": 3.338236821748293,
      "tokens_seen": 3132817408
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0223078157599101e-05,
      "loss": 1.2769,
      "theoretical_loss": 3.338226268549355,
      "tokens_seen": 3132948480
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0215053763440861e-05,
      "loss": 1.3406,
      "theoretical_loss": 3.3382157159155352,
      "tokens_seen": 3133079552
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.020702936928262e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.338205163846779,
      "tokens_seen": 3133210624
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0199004975124378e-05,
      "loss": 1.3911,
      "theoretical_loss": 3.3381946123430337,
      "tokens_seen": 3133341696
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0190980580966137e-05,
      "loss": 1.302,
      "theoretical_loss": 3.338184061404244,
      "tokens_seen": 3133472768
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0182956186807896e-05,
      "loss": 1.2303,
      "theoretical_loss": 3.338173511030357,
      "tokens_seen": 3133603840
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0174931792649656e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.3381629612213186,
      "tokens_seen": 3133734912
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0166907398491414e-05,
      "loss": 1.4036,
      "theoretical_loss": 3.338152411977075,
      "tokens_seen": 3133865984
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0158883004333174e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.3381418632975723,
      "tokens_seen": 3133997056
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0150858610174932e-05,
      "loss": 1.2841,
      "theoretical_loss": 3.338131315182756,
      "tokens_seen": 3134128128
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0142834216016692e-05,
      "loss": 1.2992,
      "theoretical_loss": 3.3381207676325735,
      "tokens_seen": 3134259200
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.013480982185845e-05,
      "loss": 1.365,
      "theoretical_loss": 3.3381102206469704,
      "tokens_seen": 3134390272
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.012678542770021e-05,
      "loss": 1.3186,
      "theoretical_loss": 3.338099674225892,
      "tokens_seen": 3134521344
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0118761033541968e-05,
      "loss": 1.2411,
      "theoretical_loss": 3.338089128369286,
      "tokens_seen": 3134652416
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0110736639383726e-05,
      "loss": 1.2813,
      "theoretical_loss": 3.3380785830770976,
      "tokens_seen": 3134783488
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0102712245225486e-05,
      "loss": 1.2337,
      "theoretical_loss": 3.3380680383492733,
      "tokens_seen": 3134914560
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0094687851067244e-05,
      "loss": 1.3604,
      "theoretical_loss": 3.338057494185759,
      "tokens_seen": 3135045632
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0086663456909004e-05,
      "loss": 1.3915,
      "theoretical_loss": 3.338046950586502,
      "tokens_seen": 3135176704
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0078639062750763e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.3380364075514466,
      "tokens_seen": 3135307776
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0070614668592522e-05,
      "loss": 1.2501,
      "theoretical_loss": 3.3380258650805406,
      "tokens_seen": 3135438848
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.006259027443428e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.33801532317373,
      "tokens_seen": 3135569920
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.005456588027604e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.33800478183096,
      "tokens_seen": 3135700992
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0046541486117799e-05,
      "loss": 1.2742,
      "theoretical_loss": 3.3379942410521783,
      "tokens_seen": 3135832064
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0038517091959559e-05,
      "loss": 1.2901,
      "theoretical_loss": 3.3379837008373303,
      "tokens_seen": 3135963136
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0030492697801317e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.3379731611863623,
      "tokens_seen": 3136094208
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0022468303643077e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.337962622099221,
      "tokens_seen": 3136225280
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0014443909484835e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.337952083575852,
      "tokens_seen": 3136356352
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0006419515326593e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.3379415456162023,
      "tokens_seen": 3136487424
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.998395121168353e-06,
      "loss": 1.2802,
      "theoretical_loss": 3.3379310082202176,
      "tokens_seen": 3136618496
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.990370727010111e-06,
      "loss": 1.2475,
      "theoretical_loss": 3.3379204713878448,
      "tokens_seen": 3136749568
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.982346332851871e-06,
      "loss": 1.1973,
      "theoretical_loss": 3.337909935119029,
      "tokens_seen": 3136880640
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.97432193869363e-06,
      "loss": 1.2386,
      "theoretical_loss": 3.3378993994137183,
      "tokens_seen": 3137011712
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.966297544535388e-06,
      "loss": 1.2328,
      "theoretical_loss": 3.3378888642718576,
      "tokens_seen": 3137142784
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.958273150377148e-06,
      "loss": 1.2927,
      "theoretical_loss": 3.337878329693394,
      "tokens_seen": 3137273856
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.950248756218906e-06,
      "loss": 1.2209,
      "theoretical_loss": 3.337867795678273,
      "tokens_seen": 3137404928
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.942224362060664e-06,
      "loss": 1.28,
      "theoretical_loss": 3.337857262226442,
      "tokens_seen": 3137536000
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.9934064149856567,
      "objective/train/avg_token_score": 0.006593545898795128,
      "objective/train/docs_used": 1140026,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3243494033813477,
      "objective/train/original_loss": 1.3334019184112549,
      "objective/train/theoretical_loss": 3.337846729337847,
      "objective/train/tokens_used": 1508126176,
      "objective/train/unlikelihood_loss": 0.017957713454961777,
      "theoretical_loss": 3.337846729337847,
      "tokens_seen": 3137667072
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.934199967902424e-06,
      "loss": 1.2826,
      "theoretical_loss": 3.337846729337847,
      "tokens_seen": 3137667072
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.926175573744182e-06,
      "loss": 1.2892,
      "theoretical_loss": 3.3378361970124337,
      "tokens_seen": 3137798144
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.91815117958594e-06,
      "loss": 1.2834,
      "theoretical_loss": 3.3378256652501492,
      "tokens_seen": 3137929216
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.9101267854277e-06,
      "loss": 1.2971,
      "theoretical_loss": 3.3378151340509397,
      "tokens_seen": 3138060288
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.902102391269458e-06,
      "loss": 1.2622,
      "theoretical_loss": 3.3378046034147513,
      "tokens_seen": 3138191360
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.894077997111218e-06,
      "loss": 1.324,
      "theoretical_loss": 3.3377940733415308,
      "tokens_seen": 3138322432
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.886053602952977e-06,
      "loss": 1.2579,
      "theoretical_loss": 3.3377835438312244,
      "tokens_seen": 3138453504
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.878029208794736e-06,
      "loss": 1.3102,
      "theoretical_loss": 3.337773014883778,
      "tokens_seen": 3138584576
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.870004814636495e-06,
      "loss": 1.2632,
      "theoretical_loss": 3.3377624864991393,
      "tokens_seen": 3138715648
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.861980420478255e-06,
      "loss": 1.2958,
      "theoretical_loss": 3.3377519586772535,
      "tokens_seen": 3138846720
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.853956026320013e-06,
      "loss": 1.2927,
      "theoretical_loss": 3.3377414314180673,
      "tokens_seen": 3138977792
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.845931632161773e-06,
      "loss": 1.2889,
      "theoretical_loss": 3.3377309047215276,
      "tokens_seen": 3139108864
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.837907238003531e-06,
      "loss": 1.2591,
      "theoretical_loss": 3.33772037858758,
      "tokens_seen": 3139239936
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.82988284384529e-06,
      "loss": 1.314,
      "theoretical_loss": 3.3377098530161717,
      "tokens_seen": 3139371008
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.821858449687049e-06,
      "loss": 1.3196,
      "theoretical_loss": 3.337699328007249,
      "tokens_seen": 3139502080
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.813834055528807e-06,
      "loss": 1.2961,
      "theoretical_loss": 3.3376888035607584,
      "tokens_seen": 3139633152
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.805809661370567e-06,
      "loss": 1.3192,
      "theoretical_loss": 3.3376782796766458,
      "tokens_seen": 3139764224
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.797785267212325e-06,
      "loss": 1.2781,
      "theoretical_loss": 3.3376677563548585,
      "tokens_seen": 3139895296
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.789760873054085e-06,
      "loss": 1.3289,
      "theoretical_loss": 3.3376572335953423,
      "tokens_seen": 3140026368
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.781736478895843e-06,
      "loss": 1.2108,
      "theoretical_loss": 3.337646711398044,
      "tokens_seen": 3140157440
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.773712084737603e-06,
      "loss": 1.3529,
      "theoretical_loss": 3.33763618976291,
      "tokens_seen": 3140288512
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.765687690579362e-06,
      "loss": 1.2195,
      "theoretical_loss": 3.337625668689887,
      "tokens_seen": 3140419584
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.757663296421122e-06,
      "loss": 1.2757,
      "theoretical_loss": 3.3376151481789216,
      "tokens_seen": 3140550656
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.74963890226288e-06,
      "loss": 1.2305,
      "theoretical_loss": 3.33760462822996,
      "tokens_seen": 3140681728
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.74161450810464e-06,
      "loss": 1.2543,
      "theoretical_loss": 3.3375941088429486,
      "tokens_seen": 3140812800
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.733590113946398e-06,
      "loss": 1.3473,
      "theoretical_loss": 3.337583590017834,
      "tokens_seen": 3140943872
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.725565719788156e-06,
      "loss": 1.32,
      "theoretical_loss": 3.337573071754563,
      "tokens_seen": 3141074944
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.717541325629916e-06,
      "loss": 1.2744,
      "theoretical_loss": 3.3375625540530827,
      "tokens_seen": 3141206016
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.709516931471674e-06,
      "loss": 1.3099,
      "theoretical_loss": 3.3375520369133382,
      "tokens_seen": 3141337088
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.701492537313434e-06,
      "loss": 1.3176,
      "theoretical_loss": 3.337541520335277,
      "tokens_seen": 3141468160
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.693468143155192e-06,
      "loss": 1.2948,
      "theoretical_loss": 3.337531004318846,
      "tokens_seen": 3141599232
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.685443748996952e-06,
      "loss": 1.3533,
      "theoretical_loss": 3.3375204888639907,
      "tokens_seen": 3141730304
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.67741935483871e-06,
      "loss": 1.2869,
      "theoretical_loss": 3.3375099739706586,
      "tokens_seen": 3141861376
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.66939496068047e-06,
      "loss": 1.2905,
      "theoretical_loss": 3.337499459638796,
      "tokens_seen": 3141992448
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.661370566522229e-06,
      "loss": 1.2745,
      "theoretical_loss": 3.3374889458683494,
      "tokens_seen": 3142123520
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.653346172363987e-06,
      "loss": 1.248,
      "theoretical_loss": 3.3374784326592657,
      "tokens_seen": 3142254592
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.645321778205747e-06,
      "loss": 1.3211,
      "theoretical_loss": 3.337467920011491,
      "tokens_seen": 3142385664
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.637297384047505e-06,
      "loss": 1.2195,
      "theoretical_loss": 3.337457407924972,
      "tokens_seen": 3142516736
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.629272989889263e-06,
      "loss": 1.2968,
      "theoretical_loss": 3.337446896399656,
      "tokens_seen": 3142647808
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.621248595731023e-06,
      "loss": 1.2597,
      "theoretical_loss": 3.337436385435489,
      "tokens_seen": 3142778880
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.613224201572781e-06,
      "loss": 1.2539,
      "theoretical_loss": 3.337425875032418,
      "tokens_seen": 3142909952
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.60519980741454e-06,
      "loss": 1.2616,
      "theoretical_loss": 3.337415365190389,
      "tokens_seen": 3143041024
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.5971754132563e-06,
      "loss": 1.4019,
      "theoretical_loss": 3.33740485590935,
      "tokens_seen": 3143172096
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.589151019098058e-06,
      "loss": 1.2668,
      "theoretical_loss": 3.337394347189246,
      "tokens_seen": 3143303168
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.581126624939817e-06,
      "loss": 1.2643,
      "theoretical_loss": 3.3373838390300246,
      "tokens_seen": 3143434240
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.573102230781576e-06,
      "loss": 1.2312,
      "theoretical_loss": 3.3373733314316323,
      "tokens_seen": 3143565312
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.565077836623336e-06,
      "loss": 1.2486,
      "theoretical_loss": 3.3373628243940163,
      "tokens_seen": 3143696384
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.557053442465094e-06,
      "loss": 1.3267,
      "theoretical_loss": 3.3373523179171225,
      "tokens_seen": 3143827456
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.549029048306854e-06,
      "loss": 1.327,
      "theoretical_loss": 3.3373418120008975,
      "tokens_seen": 3143958528
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.541004654148612e-06,
      "loss": 1.3149,
      "theoretical_loss": 3.337331306645289,
      "tokens_seen": 3144089600
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.9939411878585815,
      "objective/train/avg_token_score": 0.006058818195015192,
      "objective/train/docs_used": 1142279,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.255284309387207,
      "objective/train/original_loss": 1.2632532119750977,
      "objective/train/theoretical_loss": 3.337320801850243,
      "objective/train/tokens_used": 1514679776,
      "objective/train/unlikelihood_loss": 0.02021862007677555,
      "theoretical_loss": 3.337320801850243,
      "tokens_seen": 3144220672
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.53298025999037e-06,
      "loss": 1.2837,
      "theoretical_loss": 3.337320801850243,
      "tokens_seen": 3144220672
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.52495586583213e-06,
      "loss": 1.1882,
      "theoretical_loss": 3.3373102976157063,
      "tokens_seen": 3144351744
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.516931471673888e-06,
      "loss": 1.3184,
      "theoretical_loss": 3.337299793941626,
      "tokens_seen": 3144482816
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.508907077515648e-06,
      "loss": 1.2426,
      "theoretical_loss": 3.337289290827948,
      "tokens_seen": 3144613888
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.500882683357406e-06,
      "loss": 1.3146,
      "theoretical_loss": 3.33727878827462,
      "tokens_seen": 3144744960
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.492858289199166e-06,
      "loss": 1.3045,
      "theoretical_loss": 3.337268286281588,
      "tokens_seen": 3144876032
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.484833895040924e-06,
      "loss": 1.2764,
      "theoretical_loss": 3.3372577848487994,
      "tokens_seen": 3145007104
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.476809500882684e-06,
      "loss": 1.334,
      "theoretical_loss": 3.3372472839762004,
      "tokens_seen": 3145138176
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.468785106724443e-06,
      "loss": 1.2796,
      "theoretical_loss": 3.337236783663738,
      "tokens_seen": 3145269248
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.460760712566202e-06,
      "loss": 1.2473,
      "theoretical_loss": 3.3372262839113596,
      "tokens_seen": 3145400320
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.45273631840796e-06,
      "loss": 1.2318,
      "theoretical_loss": 3.3372157847190107,
      "tokens_seen": 3145531392
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.444711924249719e-06,
      "loss": 1.2637,
      "theoretical_loss": 3.3372052860866392,
      "tokens_seen": 3145662464
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.436687530091479e-06,
      "loss": 1.2708,
      "theoretical_loss": 3.337194788014191,
      "tokens_seen": 3145793536
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.428663135933237e-06,
      "loss": 1.3263,
      "theoretical_loss": 3.337184290501614,
      "tokens_seen": 3145924608
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.420638741774997e-06,
      "loss": 1.2828,
      "theoretical_loss": 3.337173793548854,
      "tokens_seen": 3146055680
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.412614347616755e-06,
      "loss": 1.2891,
      "theoretical_loss": 3.3371632971558585,
      "tokens_seen": 3146186752
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.404589953458515e-06,
      "loss": 1.26,
      "theoretical_loss": 3.337152801322574,
      "tokens_seen": 3146317824
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.396565559300273e-06,
      "loss": 1.2305,
      "theoretical_loss": 3.3371423060489476,
      "tokens_seen": 3146448896
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.388541165142033e-06,
      "loss": 1.3035,
      "theoretical_loss": 3.3371318113349258,
      "tokens_seen": 3146579968
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.380516770983791e-06,
      "loss": 1.236,
      "theoretical_loss": 3.3371213171804555,
      "tokens_seen": 3146711040
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.372492376825551e-06,
      "loss": 1.2398,
      "theoretical_loss": 3.3371108235854834,
      "tokens_seen": 3146842112
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.36446798266731e-06,
      "loss": 1.2447,
      "theoretical_loss": 3.337100330549957,
      "tokens_seen": 3146973184
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.356443588509068e-06,
      "loss": 1.2674,
      "theoretical_loss": 3.337089838073823,
      "tokens_seen": 3147104256
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.348419194350828e-06,
      "loss": 1.2867,
      "theoretical_loss": 3.337079346157028,
      "tokens_seen": 3147235328
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.340394800192586e-06,
      "loss": 1.301,
      "theoretical_loss": 3.3370688547995186,
      "tokens_seen": 3147366400
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.332370406034346e-06,
      "loss": 1.2979,
      "theoretical_loss": 3.337058364001243,
      "tokens_seen": 3147497472
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.324346011876104e-06,
      "loss": 1.3486,
      "theoretical_loss": 3.3370478737621463,
      "tokens_seen": 3147628544
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.316321617717864e-06,
      "loss": 1.2687,
      "theoretical_loss": 3.3370373840821763,
      "tokens_seen": 3147759616
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.308297223559622e-06,
      "loss": 1.2408,
      "theoretical_loss": 3.3370268949612805,
      "tokens_seen": 3147890688
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.30027282940138e-06,
      "loss": 1.263,
      "theoretical_loss": 3.337016406399405,
      "tokens_seen": 3148021760
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.29224843524314e-06,
      "loss": 1.2447,
      "theoretical_loss": 3.337005918396497,
      "tokens_seen": 3148152832
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.284224041084898e-06,
      "loss": 1.2245,
      "theoretical_loss": 3.336995430952503,
      "tokens_seen": 3148283904
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.276199646926657e-06,
      "loss": 1.2577,
      "theoretical_loss": 3.336984944067371,
      "tokens_seen": 3148414976
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.268175252768416e-06,
      "loss": 1.2757,
      "theoretical_loss": 3.336974457741047,
      "tokens_seen": 3148546048
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.260150858610175e-06,
      "loss": 1.2402,
      "theoretical_loss": 3.336963971973478,
      "tokens_seen": 3148677120
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.252126464451933e-06,
      "loss": 1.2465,
      "theoretical_loss": 3.336953486764612,
      "tokens_seen": 3148808192
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.244102070293693e-06,
      "loss": 1.324,
      "theoretical_loss": 3.3369430021143947,
      "tokens_seen": 3148939264
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.236077676135451e-06,
      "loss": 1.3452,
      "theoretical_loss": 3.3369325180227736,
      "tokens_seen": 3149070336
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.228053281977211e-06,
      "loss": 1.2945,
      "theoretical_loss": 3.336922034489696,
      "tokens_seen": 3149201408
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.220028887818969e-06,
      "loss": 1.3171,
      "theoretical_loss": 3.3369115515151084,
      "tokens_seen": 3149332480
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.212004493660729e-06,
      "loss": 1.3502,
      "theoretical_loss": 3.3369010690989582,
      "tokens_seen": 3149463552
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.203980099502487e-06,
      "loss": 1.2917,
      "theoretical_loss": 3.336890587241192,
      "tokens_seen": 3149594624
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.195955705344247e-06,
      "loss": 1.2545,
      "theoretical_loss": 3.336880105941757,
      "tokens_seen": 3149725696
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.187931311186005e-06,
      "loss": 1.3416,
      "theoretical_loss": 3.3368696252006003,
      "tokens_seen": 3149856768
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.179906917027765e-06,
      "loss": 1.2616,
      "theoretical_loss": 3.336859145017669,
      "tokens_seen": 3149987840
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.171882522869524e-06,
      "loss": 1.2606,
      "theoretical_loss": 3.3368486653929104,
      "tokens_seen": 3150118912
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.163858128711282e-06,
      "loss": 1.3489,
      "theoretical_loss": 3.3368381863262706,
      "tokens_seen": 3150249984
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.155833734553042e-06,
      "loss": 1.2293,
      "theoretical_loss": 3.3368277078176973,
      "tokens_seen": 3150381056
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.1478093403948e-06,
      "loss": 1.2682,
      "theoretical_loss": 3.336817229867138,
      "tokens_seen": 3150512128
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.13978494623656e-06,
      "loss": 1.3519,
      "theoretical_loss": 3.336806752474539,
      "tokens_seen": 3150643200
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.9914184212684631,
      "objective/train/avg_token_score": 0.008581587113440037,
      "objective/train/docs_used": 1144741,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2213243246078491,
      "objective/train/original_loss": 1.232299566268921,
      "objective/train/theoretical_loss": 3.3367962756398475,
      "objective/train/tokens_used": 1521233376,
      "objective/train/unlikelihood_loss": 0.02324308268725872,
      "theoretical_loss": 3.3367962756398475,
      "tokens_seen": 3150774272
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.131760552078318e-06,
      "loss": 1.2549,
      "theoretical_loss": 3.3367962756398475,
      "tokens_seen": 3150774272
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.123736157920078e-06,
      "loss": 1.3225,
      "theoretical_loss": 3.3367857993630112,
      "tokens_seen": 3150905344
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.115711763761836e-06,
      "loss": 1.1665,
      "theoretical_loss": 3.3367753236439763,
      "tokens_seen": 3151036416
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.107687369603596e-06,
      "loss": 1.3247,
      "theoretical_loss": 3.3367648484826904,
      "tokens_seen": 3151167488
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.099662975445354e-06,
      "loss": 1.2847,
      "theoretical_loss": 3.3367543738791006,
      "tokens_seen": 3151298560
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.091638581287114e-06,
      "loss": 1.2231,
      "theoretical_loss": 3.336743899833154,
      "tokens_seen": 3151429632
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.083614187128872e-06,
      "loss": 1.35,
      "theoretical_loss": 3.336733426344798,
      "tokens_seen": 3151560704
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.075589792970632e-06,
      "loss": 1.3115,
      "theoretical_loss": 3.336722953413979,
      "tokens_seen": 3151691776
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.06756539881239e-06,
      "loss": 1.2738,
      "theoretical_loss": 3.3367124810406446,
      "tokens_seen": 3151822848
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.059541004654149e-06,
      "loss": 1.2525,
      "theoretical_loss": 3.3367020092247417,
      "tokens_seen": 3151953920
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.051516610495909e-06,
      "loss": 1.3011,
      "theoretical_loss": 3.336691537966218,
      "tokens_seen": 3152084992
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.043492216337667e-06,
      "loss": 1.2416,
      "theoretical_loss": 3.3366810672650202,
      "tokens_seen": 3152216064
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.035467822179427e-06,
      "loss": 1.2866,
      "theoretical_loss": 3.3366705971210955,
      "tokens_seen": 3152347136
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.027443428021185e-06,
      "loss": 1.2536,
      "theoretical_loss": 3.3366601275343912,
      "tokens_seen": 3152478208
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.019419033862945e-06,
      "loss": 1.2328,
      "theoretical_loss": 3.336649658504854,
      "tokens_seen": 3152609280
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.011394639704703e-06,
      "loss": 1.3532,
      "theoretical_loss": 3.3366391900324324,
      "tokens_seen": 3152740352
    },
    {
      "epoch": 0.91,
      "learning_rate": 9.003370245546463e-06,
      "loss": 1.2753,
      "theoretical_loss": 3.336628722117072,
      "tokens_seen": 3152871424
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.995345851388221e-06,
      "loss": 1.3917,
      "theoretical_loss": 3.336618254758721,
      "tokens_seen": 3153002496
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.987321457229981e-06,
      "loss": 1.2087,
      "theoretical_loss": 3.336607787957326,
      "tokens_seen": 3153133568
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.97929706307174e-06,
      "loss": 1.267,
      "theoretical_loss": 3.3365973217128344,
      "tokens_seen": 3153264640
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.971272668913497e-06,
      "loss": 1.2803,
      "theoretical_loss": 3.336586856025194,
      "tokens_seen": 3153395712
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.963248274755257e-06,
      "loss": 1.3028,
      "theoretical_loss": 3.3365763908943515,
      "tokens_seen": 3153526784
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.955223880597016e-06,
      "loss": 1.2999,
      "theoretical_loss": 3.3365659263202536,
      "tokens_seen": 3153657856
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.947199486438774e-06,
      "loss": 1.278,
      "theoretical_loss": 3.336555462302849,
      "tokens_seen": 3153788928
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.939175092280534e-06,
      "loss": 1.3331,
      "theoretical_loss": 3.3365449988420837,
      "tokens_seen": 3153920000
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.931150698122292e-06,
      "loss": 1.2516,
      "theoretical_loss": 3.336534535937905,
      "tokens_seen": 3154051072
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.92312630396405e-06,
      "loss": 1.3133,
      "theoretical_loss": 3.3365240735902613,
      "tokens_seen": 3154182144
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.91510190980581e-06,
      "loss": 1.2029,
      "theoretical_loss": 3.3365136117990986,
      "tokens_seen": 3154313216
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.907077515647568e-06,
      "loss": 1.2635,
      "theoretical_loss": 3.3365031505643645,
      "tokens_seen": 3154444288
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.899053121489328e-06,
      "loss": 1.266,
      "theoretical_loss": 3.336492689886007,
      "tokens_seen": 3154575360
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.891028727331086e-06,
      "loss": 1.2736,
      "theoretical_loss": 3.3364822297639725,
      "tokens_seen": 3154706432
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.883004333172845e-06,
      "loss": 1.31,
      "theoretical_loss": 3.3364717701982083,
      "tokens_seen": 3154837504
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.874979939014604e-06,
      "loss": 1.3266,
      "theoretical_loss": 3.3364613111886627,
      "tokens_seen": 3154968576
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.866955544856363e-06,
      "loss": 1.3456,
      "theoretical_loss": 3.336450852735282,
      "tokens_seen": 3155099648
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.858931150698123e-06,
      "loss": 1.2322,
      "theoretical_loss": 3.3364403948380144,
      "tokens_seen": 3155230720
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.85090675653988e-06,
      "loss": 1.2951,
      "theoretical_loss": 3.3364299374968063,
      "tokens_seen": 3155361792
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.84288236238164e-06,
      "loss": 1.2549,
      "theoretical_loss": 3.336419480711605,
      "tokens_seen": 3155492864
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.834857968223399e-06,
      "loss": 1.3542,
      "theoretical_loss": 3.336409024482359,
      "tokens_seen": 3155623936
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.826833574065159e-06,
      "loss": 1.2439,
      "theoretical_loss": 3.336398568809015,
      "tokens_seen": 3155755008
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.818809179906917e-06,
      "loss": 1.2975,
      "theoretical_loss": 3.33638811369152,
      "tokens_seen": 3155886080
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.810784785748677e-06,
      "loss": 1.2759,
      "theoretical_loss": 3.336377659129822,
      "tokens_seen": 3156017152
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.802760391590435e-06,
      "loss": 1.2417,
      "theoretical_loss": 3.3363672051238678,
      "tokens_seen": 3156148224
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.794735997432195e-06,
      "loss": 1.2173,
      "theoretical_loss": 3.336356751673605,
      "tokens_seen": 3156279296
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.786711603273953e-06,
      "loss": 1.2707,
      "theoretical_loss": 3.336346298778981,
      "tokens_seen": 3156410368
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.778687209115711e-06,
      "loss": 1.257,
      "theoretical_loss": 3.3363358464399435,
      "tokens_seen": 3156541440
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.770662814957471e-06,
      "loss": 1.2575,
      "theoretical_loss": 3.3363253946564395,
      "tokens_seen": 3156672512
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.76263842079923e-06,
      "loss": 1.3726,
      "theoretical_loss": 3.3363149434284165,
      "tokens_seen": 3156803584
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.75461402664099e-06,
      "loss": 1.226,
      "theoretical_loss": 3.336304492755822,
      "tokens_seen": 3156934656
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.746589632482748e-06,
      "loss": 1.3184,
      "theoretical_loss": 3.3362940426386034,
      "tokens_seen": 3157065728
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.738565238324508e-06,
      "loss": 1.3238,
      "theoretical_loss": 3.3362835930767076,
      "tokens_seen": 3157196800
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.991966724395752,
      "objective/train/avg_token_score": 0.008033297955989838,
      "objective/train/docs_used": 1147172,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.362828016281128,
      "objective/train/original_loss": 1.3748083114624023,
      "objective/train/theoretical_loss": 3.3362731440700832,
      "objective/train/tokens_used": 1527786976,
      "objective/train/unlikelihood_loss": 0.01951923966407776,
      "theoretical_loss": 3.3362731440700832,
      "tokens_seen": 3157327872
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.730540844166266e-06,
      "loss": 1.3365,
      "theoretical_loss": 3.3362731440700832,
      "tokens_seen": 3157327872
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.722516450008026e-06,
      "loss": 1.3173,
      "theoretical_loss": 3.3362626956186765,
      "tokens_seen": 3157458944
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.714492055849784e-06,
      "loss": 1.2933,
      "theoretical_loss": 3.336252247722436,
      "tokens_seen": 3157590016
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.706467661691544e-06,
      "loss": 1.2976,
      "theoretical_loss": 3.336241800381308,
      "tokens_seen": 3157721088
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.698443267533302e-06,
      "loss": 1.2717,
      "theoretical_loss": 3.336231353595241,
      "tokens_seen": 3157852160
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.69041887337506e-06,
      "loss": 1.2416,
      "theoretical_loss": 3.3362209073641815,
      "tokens_seen": 3157983232
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.68239447921682e-06,
      "loss": 1.2793,
      "theoretical_loss": 3.336210461688078,
      "tokens_seen": 3158114304
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.674370085058578e-06,
      "loss": 1.3455,
      "theoretical_loss": 3.3362000165668775,
      "tokens_seen": 3158245376
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.666345690900338e-06,
      "loss": 1.2404,
      "theoretical_loss": 3.336189572000527,
      "tokens_seen": 3158376448
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.658321296742097e-06,
      "loss": 1.2391,
      "theoretical_loss": 3.336179127988975,
      "tokens_seen": 3158507520
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.650296902583856e-06,
      "loss": 1.2947,
      "theoretical_loss": 3.3361686845321685,
      "tokens_seen": 3158638592
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.642272508425615e-06,
      "loss": 1.286,
      "theoretical_loss": 3.336158241630055,
      "tokens_seen": 3158769664
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.634248114267373e-06,
      "loss": 1.2418,
      "theoretical_loss": 3.3361477992825814,
      "tokens_seen": 3158900736
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.626223720109133e-06,
      "loss": 1.2735,
      "theoretical_loss": 3.3361373574896964,
      "tokens_seen": 3159031808
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.618199325950891e-06,
      "loss": 1.2833,
      "theoretical_loss": 3.336126916251347,
      "tokens_seen": 3159162880
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.61017493179265e-06,
      "loss": 1.2863,
      "theoretical_loss": 3.336116475567481,
      "tokens_seen": 3159293952
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.602150537634409e-06,
      "loss": 1.2894,
      "theoretical_loss": 3.3361060354380454,
      "tokens_seen": 3159425024
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.594126143476167e-06,
      "loss": 1.2925,
      "theoretical_loss": 3.3360955958629877,
      "tokens_seen": 3159556096
    },
    {
      "epoch": 0.91,
      "learning_rate": 8.586101749317925e-06,
      "loss": 1.2415,
      "theoretical_loss": 3.3360851568422563,
      "tokens_seen": 3159687168
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.578077355159685e-06,
      "loss": 1.3166,
      "theoretical_loss": 3.3360747183757984,
      "tokens_seen": 3159818240
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.570052961001444e-06,
      "loss": 1.304,
      "theoretical_loss": 3.336064280463561,
      "tokens_seen": 3159949312
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.562028566843204e-06,
      "loss": 1.2935,
      "theoretical_loss": 3.3360538431054927,
      "tokens_seen": 3160080384
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.554004172684962e-06,
      "loss": 1.2349,
      "theoretical_loss": 3.3360434063015405,
      "tokens_seen": 3160211456
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.545979778526722e-06,
      "loss": 1.3989,
      "theoretical_loss": 3.336032970051652,
      "tokens_seen": 3160342528
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.53795538436848e-06,
      "loss": 1.1749,
      "theoretical_loss": 3.3360225343557746,
      "tokens_seen": 3160473600
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.52993099021024e-06,
      "loss": 1.2958,
      "theoretical_loss": 3.3360120992138564,
      "tokens_seen": 3160604672
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.521906596051998e-06,
      "loss": 1.2542,
      "theoretical_loss": 3.3360016646258446,
      "tokens_seen": 3160735744
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.513882201893758e-06,
      "loss": 1.2541,
      "theoretical_loss": 3.3359912305916875,
      "tokens_seen": 3160866816
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.505857807735516e-06,
      "loss": 1.2758,
      "theoretical_loss": 3.3359807971113318,
      "tokens_seen": 3160997888
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.497833413577274e-06,
      "loss": 1.2517,
      "theoretical_loss": 3.335970364184726,
      "tokens_seen": 3161128960
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.489809019419034e-06,
      "loss": 1.2449,
      "theoretical_loss": 3.3359599318118174,
      "tokens_seen": 3161260032
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.481784625260792e-06,
      "loss": 1.2343,
      "theoretical_loss": 3.3359494999925534,
      "tokens_seen": 3161391104
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.473760231102552e-06,
      "loss": 1.2522,
      "theoretical_loss": 3.3359390687268817,
      "tokens_seen": 3161522176
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.46573583694431e-06,
      "loss": 1.3099,
      "theoretical_loss": 3.3359286380147504,
      "tokens_seen": 3161653248
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.45771144278607e-06,
      "loss": 1.343,
      "theoretical_loss": 3.335918207856107,
      "tokens_seen": 3161784320
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.449687048627829e-06,
      "loss": 1.2973,
      "theoretical_loss": 3.3359077782508995,
      "tokens_seen": 3161915392
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.441662654469589e-06,
      "loss": 1.3424,
      "theoretical_loss": 3.3358973491990747,
      "tokens_seen": 3162046464
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.433638260311347e-06,
      "loss": 1.3663,
      "theoretical_loss": 3.335886920700581,
      "tokens_seen": 3162177536
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.425613866153107e-06,
      "loss": 1.2548,
      "theoretical_loss": 3.335876492755366,
      "tokens_seen": 3162308608
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.417589471994865e-06,
      "loss": 1.271,
      "theoretical_loss": 3.335866065363377,
      "tokens_seen": 3162439680
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.409565077836623e-06,
      "loss": 1.2545,
      "theoretical_loss": 3.3358556385245626,
      "tokens_seen": 3162570752
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.401540683678383e-06,
      "loss": 1.3046,
      "theoretical_loss": 3.33584521223887,
      "tokens_seen": 3162701824
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.393516289520141e-06,
      "loss": 1.3012,
      "theoretical_loss": 3.335834786506247,
      "tokens_seen": 3162832896
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.385491895361901e-06,
      "loss": 1.2329,
      "theoretical_loss": 3.3358243613266407,
      "tokens_seen": 3162963968
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.37746750120366e-06,
      "loss": 1.2862,
      "theoretical_loss": 3.3358139366999997,
      "tokens_seen": 3163095040
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.36944310704542e-06,
      "loss": 1.347,
      "theoretical_loss": 3.3358035126262715,
      "tokens_seen": 3163226112
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.361418712887177e-06,
      "loss": 1.2024,
      "theoretical_loss": 3.335793089105404,
      "tokens_seen": 3163357184
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.353394318728937e-06,
      "loss": 1.2593,
      "theoretical_loss": 3.3357826661373453,
      "tokens_seen": 3163488256
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.345369924570696e-06,
      "loss": 1.2801,
      "theoretical_loss": 3.335772243722042,
      "tokens_seen": 3163619328
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.337345530412455e-06,
      "loss": 1.2943,
      "theoretical_loss": 3.335761821859443,
      "tokens_seen": 3163750400
    },
    {
      "epoch": 0.92,
      "objective/train/aligned_token_count": 0.99249267578125,
      "objective/train/avg_token_score": 0.0075073144398629665,
      "objective/train/docs_used": 1149760,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.353590965270996,
      "objective/train/original_loss": 1.3648314476013184,
      "objective/train/theoretical_loss": 3.3357514005494955,
      "objective/train/tokens_used": 1534340576,
      "objective/train/unlikelihood_loss": 0.017578840255737305,
      "theoretical_loss": 3.3357514005494955,
      "tokens_seen": 3163881472
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.329321136254214e-06,
      "loss": 1.3163,
      "theoretical_loss": 3.3357514005494955,
      "tokens_seen": 3163881472
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.321296742095974e-06,
      "loss": 1.2357,
      "theoretical_loss": 3.3357409797921473,
      "tokens_seen": 3164012544
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.313272347937732e-06,
      "loss": 1.2731,
      "theoretical_loss": 3.335730559587347,
      "tokens_seen": 3164143616
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.30524795377949e-06,
      "loss": 1.2265,
      "theoretical_loss": 3.3357201399350416,
      "tokens_seen": 3164274688
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.29722355962125e-06,
      "loss": 1.2382,
      "theoretical_loss": 3.3357097208351787,
      "tokens_seen": 3164405760
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.289199165463008e-06,
      "loss": 1.2808,
      "theoretical_loss": 3.335699302287707,
      "tokens_seen": 3164536832
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.281174771304766e-06,
      "loss": 1.2691,
      "theoretical_loss": 3.335688884292574,
      "tokens_seen": 3164667904
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.273150377146526e-06,
      "loss": 1.2178,
      "theoretical_loss": 3.335678466849727,
      "tokens_seen": 3164798976
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.265125982988284e-06,
      "loss": 1.2504,
      "theoretical_loss": 3.3356680499591147,
      "tokens_seen": 3164930048
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.257101588830043e-06,
      "loss": 1.2963,
      "theoretical_loss": 3.3356576336206842,
      "tokens_seen": 3165061120
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.249077194671803e-06,
      "loss": 1.2401,
      "theoretical_loss": 3.335647217834384,
      "tokens_seen": 3165192192
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.24105280051356e-06,
      "loss": 1.2314,
      "theoretical_loss": 3.335636802600162,
      "tokens_seen": 3165323264
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.23302840635532e-06,
      "loss": 1.2387,
      "theoretical_loss": 3.3356263879179653,
      "tokens_seen": 3165454336
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.225004012197079e-06,
      "loss": 1.2055,
      "theoretical_loss": 3.3356159737877427,
      "tokens_seen": 3165585408
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.216979618038837e-06,
      "loss": 1.2325,
      "theoretical_loss": 3.335605560209441,
      "tokens_seen": 3165716480
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.208955223880597e-06,
      "loss": 1.2594,
      "theoretical_loss": 3.3355951471830094,
      "tokens_seen": 3165847552
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.200930829722355e-06,
      "loss": 1.2522,
      "theoretical_loss": 3.335584734708395,
      "tokens_seen": 3165978624
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.192906435564115e-06,
      "loss": 1.2696,
      "theoretical_loss": 3.3355743227855457,
      "tokens_seen": 3166109696
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.184882041405873e-06,
      "loss": 1.3262,
      "theoretical_loss": 3.3355639114144093,
      "tokens_seen": 3166240768
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.176857647247633e-06,
      "loss": 1.302,
      "theoretical_loss": 3.3355535005949344,
      "tokens_seen": 3166371840
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.168833253089391e-06,
      "loss": 1.2444,
      "theoretical_loss": 3.335543090327069,
      "tokens_seen": 3166502912
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.160808858931151e-06,
      "loss": 1.267,
      "theoretical_loss": 3.33553268061076,
      "tokens_seen": 3166633984
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.15278446477291e-06,
      "loss": 1.3564,
      "theoretical_loss": 3.335522271445956,
      "tokens_seen": 3166765056
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.14476007061467e-06,
      "loss": 1.2534,
      "theoretical_loss": 3.335511862832605,
      "tokens_seen": 3166896128
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.136735676456428e-06,
      "loss": 1.2477,
      "theoretical_loss": 3.3355014547706547,
      "tokens_seen": 3167027200
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.128711282298188e-06,
      "loss": 1.2243,
      "theoretical_loss": 3.335491047260054,
      "tokens_seen": 3167158272
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.120686888139946e-06,
      "loss": 1.3058,
      "theoretical_loss": 3.3354806403007493,
      "tokens_seen": 3167289344
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.112662493981704e-06,
      "loss": 1.2571,
      "theoretical_loss": 3.3354702338926896,
      "tokens_seen": 3167420416
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.104638099823464e-06,
      "loss": 1.3404,
      "theoretical_loss": 3.3354598280358227,
      "tokens_seen": 3167551488
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.096613705665222e-06,
      "loss": 1.289,
      "theoretical_loss": 3.335449422730097,
      "tokens_seen": 3167682560
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.088589311506982e-06,
      "loss": 1.245,
      "theoretical_loss": 3.3354390179754594,
      "tokens_seen": 3167813632
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.08056491734874e-06,
      "loss": 1.2612,
      "theoretical_loss": 3.335428613771859,
      "tokens_seen": 3167944704
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.0725405231905e-06,
      "loss": 1.2692,
      "theoretical_loss": 3.3354182101192436,
      "tokens_seen": 3168075776
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.064516129032258e-06,
      "loss": 1.2835,
      "theoretical_loss": 3.335407807017561,
      "tokens_seen": 3168206848
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.056491734874018e-06,
      "loss": 1.3362,
      "theoretical_loss": 3.335397404466759,
      "tokens_seen": 3168337920
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.048467340715777e-06,
      "loss": 1.2957,
      "theoretical_loss": 3.3353870024667858,
      "tokens_seen": 3168468992
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.040442946557536e-06,
      "loss": 1.3714,
      "theoretical_loss": 3.33537660101759,
      "tokens_seen": 3168600064
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.032418552399295e-06,
      "loss": 1.3313,
      "theoretical_loss": 3.335366200119119,
      "tokens_seen": 3168731136
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.024394158241053e-06,
      "loss": 1.2548,
      "theoretical_loss": 3.3353557997713215,
      "tokens_seen": 3168862208
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.016369764082813e-06,
      "loss": 1.3338,
      "theoretical_loss": 3.3353453999741447,
      "tokens_seen": 3168993280
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.008345369924571e-06,
      "loss": 1.338,
      "theoretical_loss": 3.3353350007275377,
      "tokens_seen": 3169124352
    },
    {
      "epoch": 0.92,
      "learning_rate": 8.000320975766331e-06,
      "loss": 1.3795,
      "theoretical_loss": 3.3353246020314478,
      "tokens_seen": 3169255424
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.992296581608089e-06,
      "loss": 1.273,
      "theoretical_loss": 3.3353142038858232,
      "tokens_seen": 3169386496
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.984272187449849e-06,
      "loss": 1.2554,
      "theoretical_loss": 3.3353038062906117,
      "tokens_seen": 3169517568
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.976247793291607e-06,
      "loss": 1.2951,
      "theoretical_loss": 3.3352934092457627,
      "tokens_seen": 3169648640
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.968223399133367e-06,
      "loss": 1.2646,
      "theoretical_loss": 3.3352830127512227,
      "tokens_seen": 3169779712
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.960199004975125e-06,
      "loss": 1.2959,
      "theoretical_loss": 3.335272616806941,
      "tokens_seen": 3169910784
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.952174610816884e-06,
      "loss": 1.3723,
      "theoretical_loss": 3.335262221412865,
      "tokens_seen": 3170041856
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.944150216658643e-06,
      "loss": 1.2372,
      "theoretical_loss": 3.3352518265689435,
      "tokens_seen": 3170172928
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.936125822500402e-06,
      "loss": 1.2929,
      "theoretical_loss": 3.335241432275124,
      "tokens_seen": 3170304000
    },
    {
      "epoch": 0.92,
      "objective/train/aligned_token_count": 0.9928733706474304,
      "objective/train/avg_token_score": 0.007126656360924244,
      "objective/train/docs_used": 1152259,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2615991830825806,
      "objective/train/original_loss": 1.2725520133972168,
      "objective/train/theoretical_loss": 3.335231038531355,
      "objective/train/tokens_used": 1540894176,
      "objective/train/unlikelihood_loss": 0.017856813967227936,
      "theoretical_loss": 3.335231038531355,
      "tokens_seen": 3170435072
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.92810142834216e-06,
      "loss": 1.2796,
      "theoretical_loss": 3.335231038531355,
      "tokens_seen": 3170435072
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.92007703418392e-06,
      "loss": 1.3955,
      "theoretical_loss": 3.335220645337584,
      "tokens_seen": 3170566144
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.912052640025678e-06,
      "loss": 1.2026,
      "theoretical_loss": 3.33521025269376,
      "tokens_seen": 3170697216
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.904028245867436e-06,
      "loss": 1.2868,
      "theoretical_loss": 3.3351998605998316,
      "tokens_seen": 3170828288
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.896003851709196e-06,
      "loss": 1.2757,
      "theoretical_loss": 3.3351894690557455,
      "tokens_seen": 3170959360
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.887979457550954e-06,
      "loss": 1.3234,
      "theoretical_loss": 3.3351790780614508,
      "tokens_seen": 3171090432
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.879955063392714e-06,
      "loss": 1.2541,
      "theoretical_loss": 3.3351686876168958,
      "tokens_seen": 3171221504
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.871930669234472e-06,
      "loss": 1.3042,
      "theoretical_loss": 3.335158297722028,
      "tokens_seen": 3171352576
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.863906275076232e-06,
      "loss": 1.2933,
      "theoretical_loss": 3.335147908376796,
      "tokens_seen": 3171483648
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.85588188091799e-06,
      "loss": 1.313,
      "theoretical_loss": 3.3351375195811483,
      "tokens_seen": 3171614720
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.84785748675975e-06,
      "loss": 1.2693,
      "theoretical_loss": 3.335127131335033,
      "tokens_seen": 3171745792
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.839833092601509e-06,
      "loss": 1.2581,
      "theoretical_loss": 3.335116743638398,
      "tokens_seen": 3171876864
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.831808698443267e-06,
      "loss": 1.2322,
      "theoretical_loss": 3.335106356491192,
      "tokens_seen": 3172007936
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.823784304285027e-06,
      "loss": 1.2871,
      "theoretical_loss": 3.3350959698933624,
      "tokens_seen": 3172139008
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.815759910126785e-06,
      "loss": 1.3045,
      "theoretical_loss": 3.3350855838448585,
      "tokens_seen": 3172270080
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.807735515968545e-06,
      "loss": 1.2859,
      "theoretical_loss": 3.3350751983456277,
      "tokens_seen": 3172401152
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.799711121810303e-06,
      "loss": 1.2245,
      "theoretical_loss": 3.3350648133956184,
      "tokens_seen": 3172532224
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.791686727652063e-06,
      "loss": 1.2769,
      "theoretical_loss": 3.3350544289947797,
      "tokens_seen": 3172663296
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.783662333493821e-06,
      "loss": 1.3708,
      "theoretical_loss": 3.335044045143059,
      "tokens_seen": 3172794368
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.775637939335581e-06,
      "loss": 1.2637,
      "theoretical_loss": 3.335033661840405,
      "tokens_seen": 3172925440
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.76761354517734e-06,
      "loss": 1.2713,
      "theoretical_loss": 3.3350232790867653,
      "tokens_seen": 3173056512
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.7595891510191e-06,
      "loss": 1.2576,
      "theoretical_loss": 3.335012896882089,
      "tokens_seen": 3173187584
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.751564756860857e-06,
      "loss": 1.2714,
      "theoretical_loss": 3.335002515226324,
      "tokens_seen": 3173318656
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.743540362702616e-06,
      "loss": 1.2983,
      "theoretical_loss": 3.3349921341194184,
      "tokens_seen": 3173449728
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.735515968544376e-06,
      "loss": 1.2715,
      "theoretical_loss": 3.3349817535613218,
      "tokens_seen": 3173580800
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.727491574386134e-06,
      "loss": 1.3055,
      "theoretical_loss": 3.334971373551981,
      "tokens_seen": 3173711872
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.719467180227894e-06,
      "loss": 1.2688,
      "theoretical_loss": 3.3349609940913445,
      "tokens_seen": 3173842944
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.711442786069652e-06,
      "loss": 1.3052,
      "theoretical_loss": 3.3349506151793613,
      "tokens_seen": 3173974016
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.703418391911412e-06,
      "loss": 1.3141,
      "theoretical_loss": 3.3349402368159797,
      "tokens_seen": 3174105088
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.69539399775317e-06,
      "loss": 1.2433,
      "theoretical_loss": 3.3349298590011474,
      "tokens_seen": 3174236160
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.68736960359493e-06,
      "loss": 1.2999,
      "theoretical_loss": 3.3349194817348136,
      "tokens_seen": 3174367232
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.679345209436688e-06,
      "loss": 1.229,
      "theoretical_loss": 3.334909105016926,
      "tokens_seen": 3174498304
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.671320815278448e-06,
      "loss": 1.246,
      "theoretical_loss": 3.334898728847433,
      "tokens_seen": 3174629376
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.663296421120206e-06,
      "loss": 1.2779,
      "theoretical_loss": 3.3348883532262836,
      "tokens_seen": 3174760448
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.655272026961964e-06,
      "loss": 1.3215,
      "theoretical_loss": 3.3348779781534255,
      "tokens_seen": 3174891520
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.647247632803724e-06,
      "loss": 1.2363,
      "theoretical_loss": 3.3348676036288074,
      "tokens_seen": 3175022592
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.639223238645483e-06,
      "loss": 1.2441,
      "theoretical_loss": 3.3348572296523775,
      "tokens_seen": 3175153664
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.631198844487242e-06,
      "loss": 1.2517,
      "theoretical_loss": 3.3348468562240847,
      "tokens_seen": 3175284736
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.623174450329e-06,
      "loss": 1.3014,
      "theoretical_loss": 3.334836483343877,
      "tokens_seen": 3175415808
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.61515005617076e-06,
      "loss": 1.3099,
      "theoretical_loss": 3.3348261110117026,
      "tokens_seen": 3175546880
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.607125662012518e-06,
      "loss": 1.2515,
      "theoretical_loss": 3.3348157392275106,
      "tokens_seen": 3175677952
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.599101267854278e-06,
      "loss": 1.336,
      "theoretical_loss": 3.3348053679912493,
      "tokens_seen": 3175809024
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.591076873696036e-06,
      "loss": 1.3118,
      "theoretical_loss": 3.334794997302866,
      "tokens_seen": 3175940096
    },
    {
      "epoch": 0.92,
      "learning_rate": 7.583052479537796e-06,
      "loss": 1.2855,
      "theoretical_loss": 3.334784627162311,
      "tokens_seen": 3176071168
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.575028085379554e-06,
      "loss": 1.2728,
      "theoretical_loss": 3.3347742575695314,
      "tokens_seen": 3176202240
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.567003691221314e-06,
      "loss": 1.2028,
      "theoretical_loss": 3.334763888524476,
      "tokens_seen": 3176333312
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.558979297063072e-06,
      "loss": 1.257,
      "theoretical_loss": 3.334753520027093,
      "tokens_seen": 3176464384
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.5509549029048305e-06,
      "loss": 1.2543,
      "theoretical_loss": 3.3347431520773316,
      "tokens_seen": 3176595456
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.5429305087465905e-06,
      "loss": 1.2489,
      "theoretical_loss": 3.33473278467514,
      "tokens_seen": 3176726528
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.534906114588349e-06,
      "loss": 1.2979,
      "theoretical_loss": 3.3347224178204664,
      "tokens_seen": 3176857600
    },
    {
      "epoch": 0.93,
      "objective/train/aligned_token_count": 0.9929877519607544,
      "objective/train/avg_token_score": 0.007012245245277882,
      "objective/train/docs_used": 1154349,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2917406558990479,
      "objective/train/original_loss": 1.3010215759277344,
      "objective/train/theoretical_loss": 3.3347120515132596,
      "objective/train/tokens_used": 1547447776,
      "objective/train/unlikelihood_loss": 0.02014959044754505,
      "theoretical_loss": 3.3347120515132596,
      "tokens_seen": 3176988672
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.526881720430108e-06,
      "loss": 1.3102,
      "theoretical_loss": 3.3347120515132596,
      "tokens_seen": 3176988672
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.518857326271867e-06,
      "loss": 1.308,
      "theoretical_loss": 3.3347016857534677,
      "tokens_seen": 3177119744
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.510832932113626e-06,
      "loss": 1.2745,
      "theoretical_loss": 3.33469132054104,
      "tokens_seen": 3177250816
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.502808537955384e-06,
      "loss": 1.2992,
      "theoretical_loss": 3.334680955875924,
      "tokens_seen": 3177381888
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.494784143797144e-06,
      "loss": 1.322,
      "theoretical_loss": 3.334670591758069,
      "tokens_seen": 3177512960
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.486759749638902e-06,
      "loss": 1.2794,
      "theoretical_loss": 3.334660228187423,
      "tokens_seen": 3177644032
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.478735355480662e-06,
      "loss": 1.2652,
      "theoretical_loss": 3.334649865163935,
      "tokens_seen": 3177775104
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.47071096132242e-06,
      "loss": 1.3065,
      "theoretical_loss": 3.3346395026875535,
      "tokens_seen": 3177906176
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.4626865671641785e-06,
      "loss": 1.301,
      "theoretical_loss": 3.334629140758227,
      "tokens_seen": 3178037248
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.454662173005938e-06,
      "loss": 1.2748,
      "theoretical_loss": 3.3346187793759032,
      "tokens_seen": 3178168320
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.446637778847697e-06,
      "loss": 1.2702,
      "theoretical_loss": 3.3346084185405322,
      "tokens_seen": 3178299392
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.4386133846894565e-06,
      "loss": 1.2584,
      "theoretical_loss": 3.3345980582520616,
      "tokens_seen": 3178430464
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.430588990531215e-06,
      "loss": 1.2534,
      "theoretical_loss": 3.33458769851044,
      "tokens_seen": 3178561536
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.422564596372975e-06,
      "loss": 1.2451,
      "theoretical_loss": 3.3345773393156164,
      "tokens_seen": 3178692608
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.414540202214733e-06,
      "loss": 1.3047,
      "theoretical_loss": 3.334566980667539,
      "tokens_seen": 3178823680
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.406515808056493e-06,
      "loss": 1.305,
      "theoretical_loss": 3.334556622566157,
      "tokens_seen": 3178954752
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.398491413898251e-06,
      "loss": 1.244,
      "theoretical_loss": 3.3345462650114177,
      "tokens_seen": 3179085824
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.390467019740011e-06,
      "loss": 1.2957,
      "theoretical_loss": 3.3345359080032715,
      "tokens_seen": 3179216896
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.382442625581769e-06,
      "loss": 1.2554,
      "theoretical_loss": 3.3345255515416654,
      "tokens_seen": 3179347968
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.374418231423528e-06,
      "loss": 1.3538,
      "theoretical_loss": 3.3345151956265493,
      "tokens_seen": 3179479040
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.366393837265287e-06,
      "loss": 1.2513,
      "theoretical_loss": 3.3345048402578708,
      "tokens_seen": 3179610112
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.358369443107045e-06,
      "loss": 1.3184,
      "theoretical_loss": 3.334494485435579,
      "tokens_seen": 3179741184
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.3503450489488045e-06,
      "loss": 1.3302,
      "theoretical_loss": 3.334484131159623,
      "tokens_seen": 3179872256
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.3423206547905635e-06,
      "loss": 1.3153,
      "theoretical_loss": 3.334473777429951,
      "tokens_seen": 3180003328
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.334296260632323e-06,
      "loss": 1.2897,
      "theoretical_loss": 3.3344634242465117,
      "tokens_seen": 3180134400
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.326271866474081e-06,
      "loss": 1.3602,
      "theoretical_loss": 3.3344530716092535,
      "tokens_seen": 3180265472
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.318247472315841e-06,
      "loss": 1.298,
      "theoretical_loss": 3.334442719518125,
      "tokens_seen": 3180396544
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.310223078157599e-06,
      "loss": 1.2691,
      "theoretical_loss": 3.334432367973076,
      "tokens_seen": 3180527616
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.302198683999359e-06,
      "loss": 1.2985,
      "theoretical_loss": 3.334422016974054,
      "tokens_seen": 3180658688
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.294174289841117e-06,
      "loss": 1.2878,
      "theoretical_loss": 3.3344116665210084,
      "tokens_seen": 3180789760
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.286149895682877e-06,
      "loss": 1.19,
      "theoretical_loss": 3.3344013166138873,
      "tokens_seen": 3180920832
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.278125501524635e-06,
      "loss": 1.2366,
      "theoretical_loss": 3.3343909672526397,
      "tokens_seen": 3181051904
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.270101107366393e-06,
      "loss": 1.3033,
      "theoretical_loss": 3.3343806184372147,
      "tokens_seen": 3181182976
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.262076713208153e-06,
      "loss": 1.2809,
      "theoretical_loss": 3.33437027016756,
      "tokens_seen": 3181314048
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.2540523190499115e-06,
      "loss": 1.2989,
      "theoretical_loss": 3.3343599224436256,
      "tokens_seen": 3181445120
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.246027924891671e-06,
      "loss": 1.2628,
      "theoretical_loss": 3.3343495752653594,
      "tokens_seen": 3181576192
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.23800353073343e-06,
      "loss": 1.3092,
      "theoretical_loss": 3.334339228632711,
      "tokens_seen": 3181707264
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.2299791365751895e-06,
      "loss": 1.2638,
      "theoretical_loss": 3.334328882545628,
      "tokens_seen": 3181838336
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.221954742416948e-06,
      "loss": 1.2567,
      "theoretical_loss": 3.3343185370040596,
      "tokens_seen": 3181969408
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.213930348258708e-06,
      "loss": 1.2927,
      "theoretical_loss": 3.3343081920079545,
      "tokens_seen": 3182100480
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.205905954100466e-06,
      "loss": 1.3078,
      "theoretical_loss": 3.334297847557262,
      "tokens_seen": 3182231552
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.197881559942225e-06,
      "loss": 1.2319,
      "theoretical_loss": 3.3342875036519306,
      "tokens_seen": 3182362624
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.189857165783984e-06,
      "loss": 1.3072,
      "theoretical_loss": 3.334277160291909,
      "tokens_seen": 3182493696
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.181832771625742e-06,
      "loss": 1.3546,
      "theoretical_loss": 3.334266817477146,
      "tokens_seen": 3182624768
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.173808377467501e-06,
      "loss": 1.3175,
      "theoretical_loss": 3.3342564752075905,
      "tokens_seen": 3182755840
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.16578398330926e-06,
      "loss": 1.3005,
      "theoretical_loss": 3.334246133483191,
      "tokens_seen": 3182886912
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.157759589151019e-06,
      "loss": 1.2505,
      "theoretical_loss": 3.3342357923038968,
      "tokens_seen": 3183017984
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.1497351949927776e-06,
      "loss": 1.3061,
      "theoretical_loss": 3.3342254516696563,
      "tokens_seen": 3183149056
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.1417108008345375e-06,
      "loss": 1.2517,
      "theoretical_loss": 3.334215111580418,
      "tokens_seen": 3183280128
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.133686406676296e-06,
      "loss": 1.3157,
      "theoretical_loss": 3.334204772036132,
      "tokens_seen": 3183411200
    },
    {
      "epoch": 0.93,
      "objective/train/aligned_token_count": 0.994452953338623,
      "objective/train/avg_token_score": 0.005547053646296263,
      "objective/train/docs_used": 1156704,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3317803144454956,
      "objective/train/original_loss": 1.3399930000305176,
      "objective/train/theoretical_loss": 3.334194433036746,
      "objective/train/tokens_used": 1554001376,
      "objective/train/unlikelihood_loss": 0.013857580721378326,
      "theoretical_loss": 3.334194433036746,
      "tokens_seen": 3183542272
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.125662012518056e-06,
      "loss": 1.3345,
      "theoretical_loss": 3.334194433036746,
      "tokens_seen": 3183542272
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.117637618359814e-06,
      "loss": 1.2722,
      "theoretical_loss": 3.3341840945822097,
      "tokens_seen": 3183673344
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.109613224201574e-06,
      "loss": 1.2805,
      "theoretical_loss": 3.334173756672471,
      "tokens_seen": 3183804416
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.101588830043332e-06,
      "loss": 1.242,
      "theoretical_loss": 3.3341634193074796,
      "tokens_seen": 3183935488
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.093564435885092e-06,
      "loss": 1.2486,
      "theoretical_loss": 3.334153082487184,
      "tokens_seen": 3184066560
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.08554004172685e-06,
      "loss": 1.2424,
      "theoretical_loss": 3.3341427462115325,
      "tokens_seen": 3184197632
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.077515647568608e-06,
      "loss": 1.3254,
      "theoretical_loss": 3.3341324104804753,
      "tokens_seen": 3184328704
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.069491253410368e-06,
      "loss": 1.3283,
      "theoretical_loss": 3.33412207529396,
      "tokens_seen": 3184459776
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.061466859252126e-06,
      "loss": 1.2818,
      "theoretical_loss": 3.3341117406519367,
      "tokens_seen": 3184590848
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.053442465093886e-06,
      "loss": 1.2881,
      "theoretical_loss": 3.3341014065543533,
      "tokens_seen": 3184721920
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.0454180709356445e-06,
      "loss": 1.2571,
      "theoretical_loss": 3.334091073001159,
      "tokens_seen": 3184852992
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.037393676777404e-06,
      "loss": 1.2753,
      "theoretical_loss": 3.334080739992303,
      "tokens_seen": 3184984064
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.029369282619163e-06,
      "loss": 1.273,
      "theoretical_loss": 3.334070407527734,
      "tokens_seen": 3185115136
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.021344888460922e-06,
      "loss": 1.3116,
      "theoretical_loss": 3.3340600756074013,
      "tokens_seen": 3185246208
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.013320494302681e-06,
      "loss": 1.2909,
      "theoretical_loss": 3.334049744231253,
      "tokens_seen": 3185377280
    },
    {
      "epoch": 0.93,
      "learning_rate": 7.00529610014444e-06,
      "loss": 1.3431,
      "theoretical_loss": 3.3340394133992386,
      "tokens_seen": 3185508352
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.997271705986198e-06,
      "loss": 1.3133,
      "theoretical_loss": 3.3340290831113073,
      "tokens_seen": 3185639424
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.989247311827957e-06,
      "loss": 1.2903,
      "theoretical_loss": 3.3340187533674075,
      "tokens_seen": 3185770496
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.981222917669716e-06,
      "loss": 1.2865,
      "theoretical_loss": 3.3340084241674885,
      "tokens_seen": 3185901568
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.973198523511474e-06,
      "loss": 1.3363,
      "theoretical_loss": 3.3339980955114994,
      "tokens_seen": 3186032640
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.965174129353234e-06,
      "loss": 1.2673,
      "theoretical_loss": 3.333987767399389,
      "tokens_seen": 3186163712
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.9571497351949924e-06,
      "loss": 1.3185,
      "theoretical_loss": 3.333977439831106,
      "tokens_seen": 3186294784
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.949125341036752e-06,
      "loss": 1.3382,
      "theoretical_loss": 3.3339671128066004,
      "tokens_seen": 3186425856
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.9411009468785106e-06,
      "loss": 1.321,
      "theoretical_loss": 3.3339567863258197,
      "tokens_seen": 3186556928
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.9330765527202705e-06,
      "loss": 1.2145,
      "theoretical_loss": 3.3339464603887143,
      "tokens_seen": 3186688000
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.925052158562029e-06,
      "loss": 1.3045,
      "theoretical_loss": 3.3339361349952323,
      "tokens_seen": 3186819072
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.917027764403789e-06,
      "loss": 1.3054,
      "theoretical_loss": 3.333925810145323,
      "tokens_seen": 3186950144
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.909003370245547e-06,
      "loss": 1.2895,
      "theoretical_loss": 3.3339154858389355,
      "tokens_seen": 3187081216
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.900978976087307e-06,
      "loss": 1.2934,
      "theoretical_loss": 3.333905162076019,
      "tokens_seen": 3187212288
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.892954581929065e-06,
      "loss": 1.3286,
      "theoretical_loss": 3.333894838856522,
      "tokens_seen": 3187343360
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.884930187770823e-06,
      "loss": 1.258,
      "theoretical_loss": 3.333884516180394,
      "tokens_seen": 3187474432
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.876905793612583e-06,
      "loss": 1.3159,
      "theoretical_loss": 3.333874194047584,
      "tokens_seen": 3187605504
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.868881399454341e-06,
      "loss": 1.3358,
      "theoretical_loss": 3.3338638724580414,
      "tokens_seen": 3187736576
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.860857005296101e-06,
      "loss": 1.3127,
      "theoretical_loss": 3.3338535514117145,
      "tokens_seen": 3187867648
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.852832611137859e-06,
      "loss": 1.3311,
      "theoretical_loss": 3.3338432309085526,
      "tokens_seen": 3187998720
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.844808216979618e-06,
      "loss": 1.313,
      "theoretical_loss": 3.333832910948505,
      "tokens_seen": 3188129792
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.8367838228213775e-06,
      "loss": 1.2715,
      "theoretical_loss": 3.333822591531521,
      "tokens_seen": 3188260864
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.8287594286631365e-06,
      "loss": 1.3026,
      "theoretical_loss": 3.3338122726575494,
      "tokens_seen": 3188391936
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.820735034504895e-06,
      "loss": 1.2882,
      "theoretical_loss": 3.3338019543265394,
      "tokens_seen": 3188523008
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.812710640346655e-06,
      "loss": 1.2407,
      "theoretical_loss": 3.33379163653844,
      "tokens_seen": 3188654080
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.804686246188413e-06,
      "loss": 1.2864,
      "theoretical_loss": 3.3337813192932004,
      "tokens_seen": 3188785152
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.796661852030171e-06,
      "loss": 1.2369,
      "theoretical_loss": 3.33377100259077,
      "tokens_seen": 3188916224
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.788637457871931e-06,
      "loss": 1.2363,
      "theoretical_loss": 3.333760686431097,
      "tokens_seen": 3189047296
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.780613063713689e-06,
      "loss": 1.2486,
      "theoretical_loss": 3.3337503708141316,
      "tokens_seen": 3189178368
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.772588669555449e-06,
      "loss": 1.277,
      "theoretical_loss": 3.3337400557398222,
      "tokens_seen": 3189309440
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.764564275397207e-06,
      "loss": 1.2343,
      "theoretical_loss": 3.333729741208119,
      "tokens_seen": 3189440512
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.756539881238967e-06,
      "loss": 1.3455,
      "theoretical_loss": 3.3337194272189694,
      "tokens_seen": 3189571584
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.7485154870807254e-06,
      "loss": 1.3044,
      "theoretical_loss": 3.3337091137723243,
      "tokens_seen": 3189702656
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.740491092922485e-06,
      "loss": 1.2737,
      "theoretical_loss": 3.333698800868132,
      "tokens_seen": 3189833728
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.7324666987642435e-06,
      "loss": 1.3474,
      "theoretical_loss": 3.3336884885063416,
      "tokens_seen": 3189964800
    },
    {
      "epoch": 0.93,
      "objective/train/aligned_token_count": 0.990990400314331,
      "objective/train/avg_token_score": 0.009009545668959618,
      "objective/train/docs_used": 1159129,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3155527114868164,
      "objective/train/original_loss": 1.3283311128616333,
      "objective/train/theoretical_loss": 3.3336781766869032,
      "objective/train/tokens_used": 1560554976,
      "objective/train/unlikelihood_loss": 0.022624194622039795,
      "theoretical_loss": 3.3336781766869032,
      "tokens_seen": 3190095872
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.7244423046060035e-06,
      "loss": 1.3158,
      "theoretical_loss": 3.3336781766869032,
      "tokens_seen": 3190095872
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.716417910447762e-06,
      "loss": 1.247,
      "theoretical_loss": 3.333667865409765,
      "tokens_seen": 3190226944
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.70839351628952e-06,
      "loss": 1.2972,
      "theoretical_loss": 3.3336575546748763,
      "tokens_seen": 3190358016
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.70036912213128e-06,
      "loss": 1.316,
      "theoretical_loss": 3.333647244482187,
      "tokens_seen": 3190489088
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.692344727973038e-06,
      "loss": 1.2386,
      "theoretical_loss": 3.3336369348316457,
      "tokens_seen": 3190620160
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.684320333814798e-06,
      "loss": 1.2894,
      "theoretical_loss": 3.3336266257232015,
      "tokens_seen": 3190751232
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.676295939656556e-06,
      "loss": 1.2142,
      "theoretical_loss": 3.333616317156804,
      "tokens_seen": 3190882304
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.668271545498315e-06,
      "loss": 1.2822,
      "theoretical_loss": 3.333606009132403,
      "tokens_seen": 3191013376
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.660247151340074e-06,
      "loss": 1.3285,
      "theoretical_loss": 3.3335957016499465,
      "tokens_seen": 3191144448
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.652222757181833e-06,
      "loss": 1.2748,
      "theoretical_loss": 3.3335853947093845,
      "tokens_seen": 3191275520
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.6441983630235915e-06,
      "loss": 1.3039,
      "theoretical_loss": 3.333575088310666,
      "tokens_seen": 3191406592
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.636173968865351e-06,
      "loss": 1.2862,
      "theoretical_loss": 3.3335647824537404,
      "tokens_seen": 3191537664
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.62814957470711e-06,
      "loss": 1.2466,
      "theoretical_loss": 3.333554477138557,
      "tokens_seen": 3191668736
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.6201251805488695e-06,
      "loss": 1.2723,
      "theoretical_loss": 3.3335441723650647,
      "tokens_seen": 3191799808
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.612100786390628e-06,
      "loss": 1.3626,
      "theoretical_loss": 3.3335338681332134,
      "tokens_seen": 3191930880
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.604076392232386e-06,
      "loss": 1.2199,
      "theoretical_loss": 3.3335235644429524,
      "tokens_seen": 3192061952
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.596051998074146e-06,
      "loss": 1.2286,
      "theoretical_loss": 3.33351326129423,
      "tokens_seen": 3192193024
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.588027603915904e-06,
      "loss": 1.2873,
      "theoretical_loss": 3.3335029586869966,
      "tokens_seen": 3192324096
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.580003209757664e-06,
      "loss": 1.2636,
      "theoretical_loss": 3.333492656621201,
      "tokens_seen": 3192455168
    },
    {
      "epoch": 0.93,
      "learning_rate": 6.571978815599422e-06,
      "loss": 1.2643,
      "theoretical_loss": 3.333482355096793,
      "tokens_seen": 3192586240
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.563954421441182e-06,
      "loss": 1.3254,
      "theoretical_loss": 3.333472054113721,
      "tokens_seen": 3192717312
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.55593002728294e-06,
      "loss": 1.3068,
      "theoretical_loss": 3.3334617536719353,
      "tokens_seen": 3192848384
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.5479056331247e-06,
      "loss": 1.2869,
      "theoretical_loss": 3.333451453771384,
      "tokens_seen": 3192979456
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.5398812389664584e-06,
      "loss": 1.2426,
      "theoretical_loss": 3.3334411544120184,
      "tokens_seen": 3193110528
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.531856844808218e-06,
      "loss": 1.2531,
      "theoretical_loss": 3.333430855593786,
      "tokens_seen": 3193241600
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.5238324506499765e-06,
      "loss": 1.2705,
      "theoretical_loss": 3.333420557316637,
      "tokens_seen": 3193372672
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.515808056491735e-06,
      "loss": 1.3325,
      "theoretical_loss": 3.3334102595805204,
      "tokens_seen": 3193503744
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.507783662333494e-06,
      "loss": 1.3267,
      "theoretical_loss": 3.3333999623853865,
      "tokens_seen": 3193634816
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.499759268175253e-06,
      "loss": 1.2678,
      "theoretical_loss": 3.3333896657311834,
      "tokens_seen": 3193765888
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.491734874017012e-06,
      "loss": 1.2757,
      "theoretical_loss": 3.333379369617861,
      "tokens_seen": 3193896960
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.48371047985877e-06,
      "loss": 1.2426,
      "theoretical_loss": 3.333369074045369,
      "tokens_seen": 3194028032
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.47568608570053e-06,
      "loss": 1.2668,
      "theoretical_loss": 3.3333587790136567,
      "tokens_seen": 3194159104
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.467661691542288e-06,
      "loss": 1.2405,
      "theoretical_loss": 3.333348484522673,
      "tokens_seen": 3194290176
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.459637297384048e-06,
      "loss": 1.2622,
      "theoretical_loss": 3.333338190572368,
      "tokens_seen": 3194421248
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.451612903225806e-06,
      "loss": 1.2729,
      "theoretical_loss": 3.3333278971626905,
      "tokens_seen": 3194552320
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.443588509067566e-06,
      "loss": 1.3905,
      "theoretical_loss": 3.3333176042935904,
      "tokens_seen": 3194683392
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.4355641149093245e-06,
      "loss": 1.3046,
      "theoretical_loss": 3.333307311965017,
      "tokens_seen": 3194814464
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.427539720751084e-06,
      "loss": 1.2305,
      "theoretical_loss": 3.3332970201769196,
      "tokens_seen": 3194945536
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.419515326592843e-06,
      "loss": 1.308,
      "theoretical_loss": 3.333286728929248,
      "tokens_seen": 3195076608
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.411490932434601e-06,
      "loss": 1.2789,
      "theoretical_loss": 3.333276438221951,
      "tokens_seen": 3195207680
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.403466538276361e-06,
      "loss": 1.2616,
      "theoretical_loss": 3.3332661480549786,
      "tokens_seen": 3195338752
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.395442144118119e-06,
      "loss": 1.2509,
      "theoretical_loss": 3.3332558584282803,
      "tokens_seen": 3195469824
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.387417749959879e-06,
      "loss": 1.2807,
      "theoretical_loss": 3.333245569341805,
      "tokens_seen": 3195600896
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.379393355801637e-06,
      "loss": 1.2845,
      "theoretical_loss": 3.333235280795503,
      "tokens_seen": 3195731968
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.371368961643397e-06,
      "loss": 1.3252,
      "theoretical_loss": 3.333224992789323,
      "tokens_seen": 3195863040
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.363344567485155e-06,
      "loss": 1.2552,
      "theoretical_loss": 3.3332147053232153,
      "tokens_seen": 3195994112
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.355320173326914e-06,
      "loss": 1.2535,
      "theoretical_loss": 3.3332044183971288,
      "tokens_seen": 3196125184
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.347295779168673e-06,
      "loss": 1.2771,
      "theoretical_loss": 3.333194132011013,
      "tokens_seen": 3196256256
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.339271385010432e-06,
      "loss": 1.3199,
      "theoretical_loss": 3.3331838461648178,
      "tokens_seen": 3196387328
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.3312469908521906e-06,
      "loss": 1.2806,
      "theoretical_loss": 3.3331735608584925,
      "tokens_seen": 3196518400
    },
    {
      "epoch": 0.94,
      "objective/train/aligned_token_count": 0.9943403601646423,
      "objective/train/avg_token_score": 0.005659641232341528,
      "objective/train/docs_used": 1161767,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2491236925125122,
      "objective/train/original_loss": 1.2559165954589844,
      "objective/train/theoretical_loss": 3.3331632760919865,
      "objective/train/tokens_used": 1567108576,
      "objective/train/unlikelihood_loss": 0.01795196533203125,
      "theoretical_loss": 3.3331632760919865,
      "tokens_seen": 3196649472
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.32322259669395e-06,
      "loss": 1.2261,
      "theoretical_loss": 3.3331632760919865,
      "tokens_seen": 3196649472
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.315198202535709e-06,
      "loss": 1.2014,
      "theoretical_loss": 3.3331529918652496,
      "tokens_seen": 3196780544
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.307173808377467e-06,
      "loss": 1.2669,
      "theoretical_loss": 3.333142708178231,
      "tokens_seen": 3196911616
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.299149414219227e-06,
      "loss": 1.2111,
      "theoretical_loss": 3.3331324250308807,
      "tokens_seen": 3197042688
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.291125020060985e-06,
      "loss": 1.2744,
      "theoretical_loss": 3.3331221424231483,
      "tokens_seen": 3197173760
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.283100625902745e-06,
      "loss": 1.287,
      "theoretical_loss": 3.3331118603549825,
      "tokens_seen": 3197304832
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.275076231744503e-06,
      "loss": 1.2276,
      "theoretical_loss": 3.3331015788263336,
      "tokens_seen": 3197435904
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.267051837586263e-06,
      "loss": 1.2609,
      "theoretical_loss": 3.3330912978371514,
      "tokens_seen": 3197566976
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.259027443428021e-06,
      "loss": 1.2826,
      "theoretical_loss": 3.333081017387385,
      "tokens_seen": 3197698048
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.251003049269781e-06,
      "loss": 1.2745,
      "theoretical_loss": 3.3330707374769837,
      "tokens_seen": 3197829120
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.242978655111539e-06,
      "loss": 1.2809,
      "theoretical_loss": 3.333060458105898,
      "tokens_seen": 3197960192
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.2349542609532984e-06,
      "loss": 1.3661,
      "theoretical_loss": 3.3330501792740765,
      "tokens_seen": 3198091264
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.2269298667950575e-06,
      "loss": 1.2399,
      "theoretical_loss": 3.3330399009814693,
      "tokens_seen": 3198222336
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.2189054726368165e-06,
      "loss": 1.269,
      "theoretical_loss": 3.3330296232280263,
      "tokens_seen": 3198353408
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.210881078478576e-06,
      "loss": 1.2786,
      "theoretical_loss": 3.333019346013697,
      "tokens_seen": 3198484480
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.202856684320335e-06,
      "loss": 1.3224,
      "theoretical_loss": 3.3330090693384307,
      "tokens_seen": 3198615552
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.194832290162094e-06,
      "loss": 1.2405,
      "theoretical_loss": 3.3329987932021776,
      "tokens_seen": 3198746624
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.186807896003852e-06,
      "loss": 1.2927,
      "theoretical_loss": 3.3329885176048863,
      "tokens_seen": 3198877696
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.178783501845611e-06,
      "loss": 1.235,
      "theoretical_loss": 3.3329782425465075,
      "tokens_seen": 3199008768
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.17075910768737e-06,
      "loss": 1.2716,
      "theoretical_loss": 3.3329679680269906,
      "tokens_seen": 3199139840
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.162734713529129e-06,
      "loss": 1.253,
      "theoretical_loss": 3.332957694046285,
      "tokens_seen": 3199270912
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.154710319370887e-06,
      "loss": 1.3017,
      "theoretical_loss": 3.33294742060434,
      "tokens_seen": 3199401984
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.146685925212646e-06,
      "loss": 1.221,
      "theoretical_loss": 3.3329371477011067,
      "tokens_seen": 3199533056
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.1386615310544054e-06,
      "loss": 1.2473,
      "theoretical_loss": 3.332926875336533,
      "tokens_seen": 3199664128
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.1306371368961645e-06,
      "loss": 1.3113,
      "theoretical_loss": 3.33291660351057,
      "tokens_seen": 3199795200
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.1226127427379236e-06,
      "loss": 1.266,
      "theoretical_loss": 3.3329063322231667,
      "tokens_seen": 3199926272
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.114588348579683e-06,
      "loss": 1.2257,
      "theoretical_loss": 3.332896061474273,
      "tokens_seen": 3200057344
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.106563954421442e-06,
      "loss": 1.3187,
      "theoretical_loss": 3.3328857912638385,
      "tokens_seen": 3200188416
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.098539560263201e-06,
      "loss": 1.2957,
      "theoretical_loss": 3.3328755215918133,
      "tokens_seen": 3200319488
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.090515166104959e-06,
      "loss": 1.2921,
      "theoretical_loss": 3.3328652524581464,
      "tokens_seen": 3200450560
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.082490771946718e-06,
      "loss": 1.3053,
      "theoretical_loss": 3.3328549838627883,
      "tokens_seen": 3200581632
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.074466377788477e-06,
      "loss": 1.2205,
      "theoretical_loss": 3.332844715805688,
      "tokens_seen": 3200712704
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.066441983630236e-06,
      "loss": 1.2949,
      "theoretical_loss": 3.332834448286796,
      "tokens_seen": 3200843776
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.058417589471995e-06,
      "loss": 1.254,
      "theoretical_loss": 3.3328241813060617,
      "tokens_seen": 3200974848
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.050393195313754e-06,
      "loss": 1.3246,
      "theoretical_loss": 3.3328139148634346,
      "tokens_seen": 3201105920
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.042368801155513e-06,
      "loss": 1.2494,
      "theoretical_loss": 3.3328036489588646,
      "tokens_seen": 3201236992
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.034344406997272e-06,
      "loss": 1.263,
      "theoretical_loss": 3.332793383592302,
      "tokens_seen": 3201368064
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.026320012839031e-06,
      "loss": 1.3114,
      "theoretical_loss": 3.332783118763696,
      "tokens_seen": 3201499136
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.0182956186807905e-06,
      "loss": 1.2615,
      "theoretical_loss": 3.332772854472996,
      "tokens_seen": 3201630208
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.0102712245225495e-06,
      "loss": 1.2939,
      "theoretical_loss": 3.332762590720153,
      "tokens_seen": 3201761280
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.002246830364308e-06,
      "loss": 1.2333,
      "theoretical_loss": 3.332752327505116,
      "tokens_seen": 3201892352
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.994222436206067e-06,
      "loss": 1.2855,
      "theoretical_loss": 3.332742064827835,
      "tokens_seen": 3202023424
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.986198042047825e-06,
      "loss": 1.3263,
      "theoretical_loss": 3.3327318026882597,
      "tokens_seen": 3202154496
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.978173647889584e-06,
      "loss": 1.2531,
      "theoretical_loss": 3.33272154108634,
      "tokens_seen": 3202285568
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.970149253731343e-06,
      "loss": 1.1966,
      "theoretical_loss": 3.3327112800220258,
      "tokens_seen": 3202416640
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.962124859573102e-06,
      "loss": 1.3198,
      "theoretical_loss": 3.3327010194952664,
      "tokens_seen": 3202547712
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.954100465414861e-06,
      "loss": 1.2528,
      "theoretical_loss": 3.3326907595060122,
      "tokens_seen": 3202678784
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.94607607125662e-06,
      "loss": 1.1531,
      "theoretical_loss": 3.3326805000542135,
      "tokens_seen": 3202809856
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.938051677098379e-06,
      "loss": 1.3082,
      "theoretical_loss": 3.332670241139819,
      "tokens_seen": 3202940928
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.9300272829401384e-06,
      "loss": 1.255,
      "theoretical_loss": 3.332659982762779,
      "tokens_seen": 3203072000
    },
    {
      "epoch": 0.94,
      "objective/train/aligned_token_count": 0.9910057187080383,
      "objective/train/avg_token_score": 0.008994284085929394,
      "objective/train/docs_used": 1164087,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3110178709030151,
      "objective/train/original_loss": 1.3249279260635376,
      "objective/train/theoretical_loss": 3.3326497249230442,
      "objective/train/tokens_used": 1573662176,
      "objective/train/unlikelihood_loss": 0.022911811247467995,
      "theoretical_loss": 3.3326497249230442,
      "tokens_seen": 3203203072
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.9220028887818975e-06,
      "loss": 1.2794,
      "theoretical_loss": 3.3326497249230442,
      "tokens_seen": 3203203072
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.9139784946236566e-06,
      "loss": 1.2043,
      "theoretical_loss": 3.332639467620563,
      "tokens_seen": 3203334144
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.905954100465415e-06,
      "loss": 1.2717,
      "theoretical_loss": 3.3326292108552864,
      "tokens_seen": 3203465216
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.897929706307174e-06,
      "loss": 1.2147,
      "theoretical_loss": 3.332618954627164,
      "tokens_seen": 3203596288
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.889905312148933e-06,
      "loss": 1.273,
      "theoretical_loss": 3.332608698936146,
      "tokens_seen": 3203727360
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.881880917990692e-06,
      "loss": 1.2348,
      "theoretical_loss": 3.332598443782181,
      "tokens_seen": 3203858432
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.873856523832451e-06,
      "loss": 1.2633,
      "theoretical_loss": 3.3325881891652207,
      "tokens_seen": 3203989504
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.86583212967421e-06,
      "loss": 1.3044,
      "theoretical_loss": 3.332577935085214,
      "tokens_seen": 3204120576
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.857807735515969e-06,
      "loss": 1.2718,
      "theoretical_loss": 3.332567681542111,
      "tokens_seen": 3204251648
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.849783341357728e-06,
      "loss": 1.3007,
      "theoretical_loss": 3.332557428535861,
      "tokens_seen": 3204382720
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.841758947199487e-06,
      "loss": 1.2796,
      "theoretical_loss": 3.3325471760664147,
      "tokens_seen": 3204513792
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.8337345530412454e-06,
      "loss": 1.2683,
      "theoretical_loss": 3.332536924133722,
      "tokens_seen": 3204644864
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.8257101588830045e-06,
      "loss": 1.2826,
      "theoretical_loss": 3.332526672737733,
      "tokens_seen": 3204775936
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.8176857647247636e-06,
      "loss": 1.3071,
      "theoretical_loss": 3.332516421878397,
      "tokens_seen": 3204907008
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.809661370566522e-06,
      "loss": 1.3001,
      "theoretical_loss": 3.3325061715556648,
      "tokens_seen": 3205038080
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.801636976408281e-06,
      "loss": 1.2836,
      "theoretical_loss": 3.3324959217694854,
      "tokens_seen": 3205169152
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.79361258225004e-06,
      "loss": 1.2335,
      "theoretical_loss": 3.3324856725198098,
      "tokens_seen": 3205300224
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.785588188091799e-06,
      "loss": 1.2934,
      "theoretical_loss": 3.3324754238065872,
      "tokens_seen": 3205431296
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.777563793933558e-06,
      "loss": 1.2477,
      "theoretical_loss": 3.3324651756297676,
      "tokens_seen": 3205562368
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.769539399775317e-06,
      "loss": 1.3245,
      "theoretical_loss": 3.3324549279893017,
      "tokens_seen": 3205693440
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.761515005617076e-06,
      "loss": 1.2625,
      "theoretical_loss": 3.3324446808851387,
      "tokens_seen": 3205824512
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.753490611458835e-06,
      "loss": 1.1879,
      "theoretical_loss": 3.3324344343172294,
      "tokens_seen": 3205955584
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.745466217300594e-06,
      "loss": 1.2519,
      "theoretical_loss": 3.332424188285523,
      "tokens_seen": 3206086656
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.737441823142353e-06,
      "loss": 1.2589,
      "theoretical_loss": 3.33241394278997,
      "tokens_seen": 3206217728
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.729417428984112e-06,
      "loss": 1.2614,
      "theoretical_loss": 3.33240369783052,
      "tokens_seen": 3206348800
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.7213930348258714e-06,
      "loss": 1.2697,
      "theoretical_loss": 3.3323934534071236,
      "tokens_seen": 3206479872
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.71336864066763e-06,
      "loss": 1.2396,
      "theoretical_loss": 3.3323832095197305,
      "tokens_seen": 3206610944
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.705344246509389e-06,
      "loss": 1.2459,
      "theoretical_loss": 3.332372966168291,
      "tokens_seen": 3206742016
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.697319852351148e-06,
      "loss": 1.2273,
      "theoretical_loss": 3.332362723352755,
      "tokens_seen": 3206873088
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.689295458192907e-06,
      "loss": 1.2165,
      "theoretical_loss": 3.3323524810730722,
      "tokens_seen": 3207004160
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.681271064034666e-06,
      "loss": 1.2427,
      "theoretical_loss": 3.3323422393291935,
      "tokens_seen": 3207135232
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.673246669876425e-06,
      "loss": 1.2935,
      "theoretical_loss": 3.3323319981210684,
      "tokens_seen": 3207266304
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.665222275718184e-06,
      "loss": 1.2115,
      "theoretical_loss": 3.332321757448647,
      "tokens_seen": 3207397376
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.657197881559942e-06,
      "loss": 1.3001,
      "theoretical_loss": 3.3323115173118794,
      "tokens_seen": 3207528448
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.649173487401701e-06,
      "loss": 1.2961,
      "theoretical_loss": 3.3323012777107155,
      "tokens_seen": 3207659520
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.64114909324346e-06,
      "loss": 1.1979,
      "theoretical_loss": 3.332291038645106,
      "tokens_seen": 3207790592
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.633124699085219e-06,
      "loss": 1.2438,
      "theoretical_loss": 3.332280800115001,
      "tokens_seen": 3207921664
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.6251003049269784e-06,
      "loss": 1.3115,
      "theoretical_loss": 3.3322705621203497,
      "tokens_seen": 3208052736
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.617075910768737e-06,
      "loss": 1.3058,
      "theoretical_loss": 3.3322603246611027,
      "tokens_seen": 3208183808
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.609051516610496e-06,
      "loss": 1.1829,
      "theoretical_loss": 3.3322500877372105,
      "tokens_seen": 3208314880
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.601027122452255e-06,
      "loss": 1.1485,
      "theoretical_loss": 3.332239851348623,
      "tokens_seen": 3208445952
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.593002728294014e-06,
      "loss": 1.2872,
      "theoretical_loss": 3.33222961549529,
      "tokens_seen": 3208577024
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.584978334135773e-06,
      "loss": 1.2858,
      "theoretical_loss": 3.332219380177163,
      "tokens_seen": 3208708096
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.576953939977532e-06,
      "loss": 1.1933,
      "theoretical_loss": 3.3322091453941898,
      "tokens_seen": 3208839168
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.568929545819291e-06,
      "loss": 1.2944,
      "theoretical_loss": 3.332198911146323,
      "tokens_seen": 3208970240
    },
    {
      "epoch": 0.94,
      "learning_rate": 5.56090515166105e-06,
      "loss": 1.2463,
      "theoretical_loss": 3.3321886774335105,
      "tokens_seen": 3209101312
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.552880757502809e-06,
      "loss": 1.2028,
      "theoretical_loss": 3.3321784442557045,
      "tokens_seen": 3209232384
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.544856363344568e-06,
      "loss": 1.263,
      "theoretical_loss": 3.3321682116128537,
      "tokens_seen": 3209363456
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.536831969186327e-06,
      "loss": 1.2823,
      "theoretical_loss": 3.3321579795049097,
      "tokens_seen": 3209494528
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.5288075750280855e-06,
      "loss": 1.2661,
      "theoretical_loss": 3.3321477479318213,
      "tokens_seen": 3209625600
    },
    {
      "epoch": 0.95,
      "objective/train/aligned_token_count": 0.9925563931465149,
      "objective/train/avg_token_score": 0.007443617098033428,
      "objective/train/docs_used": 1166593,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2923239469528198,
      "objective/train/original_loss": 1.3024463653564453,
      "objective/train/theoretical_loss": 3.332137516893539,
      "objective/train/tokens_used": 1580215776,
      "objective/train/unlikelihood_loss": 0.020602405071258545,
      "theoretical_loss": 3.332137516893539,
      "tokens_seen": 3209756672
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.5207831808698445e-06,
      "loss": 1.2867,
      "theoretical_loss": 3.332137516893539,
      "tokens_seen": 3209756672
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.5127587867116036e-06,
      "loss": 1.2611,
      "theoretical_loss": 3.3321272863900138,
      "tokens_seen": 3209887744
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.504734392553363e-06,
      "loss": 1.2246,
      "theoretical_loss": 3.332117056421195,
      "tokens_seen": 3210018816
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.496709998395122e-06,
      "loss": 1.2332,
      "theoretical_loss": 3.3321068269870335,
      "tokens_seen": 3210149888
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.488685604236881e-06,
      "loss": 1.1873,
      "theoretical_loss": 3.3320965980874795,
      "tokens_seen": 3210280960
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.480661210078639e-06,
      "loss": 1.2683,
      "theoretical_loss": 3.3320863697224823,
      "tokens_seen": 3210412032
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.472636815920398e-06,
      "loss": 1.2796,
      "theoretical_loss": 3.3320761418919935,
      "tokens_seen": 3210543104
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.464612421762157e-06,
      "loss": 1.1707,
      "theoretical_loss": 3.3320659145959626,
      "tokens_seen": 3210674176
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.456588027603916e-06,
      "loss": 1.2439,
      "theoretical_loss": 3.3320556878343397,
      "tokens_seen": 3210805248
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.448563633445675e-06,
      "loss": 1.3127,
      "theoretical_loss": 3.332045461607075,
      "tokens_seen": 3210936320
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.440539239287434e-06,
      "loss": 1.2259,
      "theoretical_loss": 3.3320352359141197,
      "tokens_seen": 3211067392
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.4325148451291925e-06,
      "loss": 1.2502,
      "theoretical_loss": 3.332025010755423,
      "tokens_seen": 3211198464
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.4244904509709515e-06,
      "loss": 1.2378,
      "theoretical_loss": 3.332014786130936,
      "tokens_seen": 3211329536
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.416466056812711e-06,
      "loss": 1.3239,
      "theoretical_loss": 3.3320045620406082,
      "tokens_seen": 3211460608
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.40844166265447e-06,
      "loss": 1.3368,
      "theoretical_loss": 3.3319943384843906,
      "tokens_seen": 3211591680
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.400417268496229e-06,
      "loss": 1.2412,
      "theoretical_loss": 3.331984115462233,
      "tokens_seen": 3211722752
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.392392874337988e-06,
      "loss": 1.2523,
      "theoretical_loss": 3.331973892974086,
      "tokens_seen": 3211853824
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.384368480179747e-06,
      "loss": 1.2744,
      "theoretical_loss": 3.3319636710198997,
      "tokens_seen": 3211984896
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.376344086021506e-06,
      "loss": 1.2715,
      "theoretical_loss": 3.3319534495996246,
      "tokens_seen": 3212115968
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.368319691863265e-06,
      "loss": 1.2942,
      "theoretical_loss": 3.3319432287132114,
      "tokens_seen": 3212247040
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.360295297705024e-06,
      "loss": 1.2527,
      "theoretical_loss": 3.3319330083606093,
      "tokens_seen": 3212378112
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.352270903546783e-06,
      "loss": 1.2319,
      "theoretical_loss": 3.33192278854177,
      "tokens_seen": 3212509184
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.344246509388542e-06,
      "loss": 1.3174,
      "theoretical_loss": 3.3319125692566427,
      "tokens_seen": 3212640256
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.3362221152303e-06,
      "loss": 1.2865,
      "theoretical_loss": 3.3319023505051786,
      "tokens_seen": 3212771328
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.328197721072059e-06,
      "loss": 1.3038,
      "theoretical_loss": 3.3318921322873276,
      "tokens_seen": 3212902400
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.3201733269138184e-06,
      "loss": 1.3464,
      "theoretical_loss": 3.33188191460304,
      "tokens_seen": 3213033472
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.312148932755577e-06,
      "loss": 1.3006,
      "theoretical_loss": 3.3318716974522666,
      "tokens_seen": 3213164544
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.304124538597336e-06,
      "loss": 1.3617,
      "theoretical_loss": 3.3318614808349576,
      "tokens_seen": 3213295616
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.296100144439095e-06,
      "loss": 1.2627,
      "theoretical_loss": 3.331851264751063,
      "tokens_seen": 3213426688
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.288075750280854e-06,
      "loss": 1.2774,
      "theoretical_loss": 3.331841049200534,
      "tokens_seen": 3213557760
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.280051356122613e-06,
      "loss": 1.2121,
      "theoretical_loss": 3.3318308341833203,
      "tokens_seen": 3213688832
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.272026961964372e-06,
      "loss": 1.2417,
      "theoretical_loss": 3.3318206196993723,
      "tokens_seen": 3213819904
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.264002567806131e-06,
      "loss": 1.2604,
      "theoretical_loss": 3.331810405748641,
      "tokens_seen": 3213950976
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.25597817364789e-06,
      "loss": 1.2654,
      "theoretical_loss": 3.3318001923310763,
      "tokens_seen": 3214082048
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.247953779489649e-06,
      "loss": 1.2554,
      "theoretical_loss": 3.331789979446629,
      "tokens_seen": 3214213120
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.239929385331407e-06,
      "loss": 1.3469,
      "theoretical_loss": 3.3317797670952487,
      "tokens_seen": 3214344192
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.231904991173166e-06,
      "loss": 1.2477,
      "theoretical_loss": 3.331769555276887,
      "tokens_seen": 3214475264
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.2238805970149255e-06,
      "loss": 1.2865,
      "theoretical_loss": 3.331759343991494,
      "tokens_seen": 3214606336
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.2158562028566845e-06,
      "loss": 1.2605,
      "theoretical_loss": 3.3317491332390192,
      "tokens_seen": 3214737408
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.207831808698444e-06,
      "loss": 1.2859,
      "theoretical_loss": 3.3317389230194143,
      "tokens_seen": 3214868480
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.199807414540203e-06,
      "loss": 1.2127,
      "theoretical_loss": 3.3317287133326294,
      "tokens_seen": 3214999552
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.191783020381962e-06,
      "loss": 1.1514,
      "theoretical_loss": 3.331718504178615,
      "tokens_seen": 3215130624
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.183758626223721e-06,
      "loss": 1.2558,
      "theoretical_loss": 3.331708295557321,
      "tokens_seen": 3215261696
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.17573423206548e-06,
      "loss": 1.2528,
      "theoretical_loss": 3.3316980874686988,
      "tokens_seen": 3215392768
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.167709837907239e-06,
      "loss": 1.2174,
      "theoretical_loss": 3.331687879912698,
      "tokens_seen": 3215523840
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.159685443748997e-06,
      "loss": 1.2738,
      "theoretical_loss": 3.3316776728892696,
      "tokens_seen": 3215654912
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.151661049590756e-06,
      "loss": 1.2601,
      "theoretical_loss": 3.3316674663983648,
      "tokens_seen": 3215785984
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.143636655432515e-06,
      "loss": 1.2617,
      "theoretical_loss": 3.3316572604399326,
      "tokens_seen": 3215917056
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.135612261274273e-06,
      "loss": 1.2595,
      "theoretical_loss": 3.3316470550139243,
      "tokens_seen": 3216048128
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.1275878671160325e-06,
      "loss": 1.2909,
      "theoretical_loss": 3.33163685012029,
      "tokens_seen": 3216179200
    },
    {
      "epoch": 0.95,
      "objective/train/aligned_token_count": 0.9951158165931702,
      "objective/train/avg_token_score": 0.004884195514023304,
      "objective/train/docs_used": 1168814,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1325968503952026,
      "objective/train/original_loss": 1.1392648220062256,
      "objective/train/theoretical_loss": 3.3316266457589814,
      "objective/train/tokens_used": 1586769376,
      "objective/train/unlikelihood_loss": 0.014602080918848515,
      "theoretical_loss": 3.3316266457589814,
      "tokens_seen": 3216310272
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.1195634729577915e-06,
      "loss": 1.1691,
      "theoretical_loss": 3.3316266457589814,
      "tokens_seen": 3216310272
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.111539078799551e-06,
      "loss": 1.3418,
      "theoretical_loss": 3.331616441929948,
      "tokens_seen": 3216441344
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.10351468464131e-06,
      "loss": 1.2314,
      "theoretical_loss": 3.3316062386331406,
      "tokens_seen": 3216572416
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.095490290483069e-06,
      "loss": 1.287,
      "theoretical_loss": 3.33159603586851,
      "tokens_seen": 3216703488
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.087465896324828e-06,
      "loss": 1.3166,
      "theoretical_loss": 3.3315858336360065,
      "tokens_seen": 3216834560
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.079441502166587e-06,
      "loss": 1.2425,
      "theoretical_loss": 3.3315756319355807,
      "tokens_seen": 3216965632
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.071417108008346e-06,
      "loss": 1.2947,
      "theoretical_loss": 3.331565430767183,
      "tokens_seen": 3217096704
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.063392713850105e-06,
      "loss": 1.3201,
      "theoretical_loss": 3.331555230130764,
      "tokens_seen": 3217227776
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.055368319691863e-06,
      "loss": 1.2464,
      "theoretical_loss": 3.331545030026275,
      "tokens_seen": 3217358848
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.047343925533622e-06,
      "loss": 1.2008,
      "theoretical_loss": 3.3315348304536654,
      "tokens_seen": 3217489920
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.039319531375381e-06,
      "loss": 1.2783,
      "theoretical_loss": 3.3315246314128872,
      "tokens_seen": 3217620992
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.03129513721714e-06,
      "loss": 1.2965,
      "theoretical_loss": 3.3315144329038895,
      "tokens_seen": 3217752064
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.023270743058899e-06,
      "loss": 1.2466,
      "theoretical_loss": 3.331504234926624,
      "tokens_seen": 3217883136
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.0152463489006585e-06,
      "loss": 1.3032,
      "theoretical_loss": 3.331494037481041,
      "tokens_seen": 3218014208
    },
    {
      "epoch": 0.95,
      "learning_rate": 5.0072219547424175e-06,
      "loss": 1.2441,
      "theoretical_loss": 3.331483840567091,
      "tokens_seen": 3218145280
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.9991975605841766e-06,
      "loss": 1.2514,
      "theoretical_loss": 3.331473644184725,
      "tokens_seen": 3218276352
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.991173166425936e-06,
      "loss": 1.2642,
      "theoretical_loss": 3.331463448333893,
      "tokens_seen": 3218407424
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.983148772267694e-06,
      "loss": 1.308,
      "theoretical_loss": 3.331453253014546,
      "tokens_seen": 3218538496
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.975124378109453e-06,
      "loss": 1.2228,
      "theoretical_loss": 3.3314430582266352,
      "tokens_seen": 3218669568
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.967099983951212e-06,
      "loss": 1.2012,
      "theoretical_loss": 3.33143286397011,
      "tokens_seen": 3218800640
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.95907558979297e-06,
      "loss": 1.3131,
      "theoretical_loss": 3.331422670244922,
      "tokens_seen": 3218931712
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.951051195634729e-06,
      "loss": 1.2793,
      "theoretical_loss": 3.331412477051022,
      "tokens_seen": 3219062784
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.943026801476488e-06,
      "loss": 1.2225,
      "theoretical_loss": 3.33140228438836,
      "tokens_seen": 3219193856
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.935002407318247e-06,
      "loss": 1.3283,
      "theoretical_loss": 3.3313920922568876,
      "tokens_seen": 3219324928
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.926978013160006e-06,
      "loss": 1.2257,
      "theoretical_loss": 3.3313819006565546,
      "tokens_seen": 3219456000
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.9189536190017655e-06,
      "loss": 1.2496,
      "theoretical_loss": 3.331371709587312,
      "tokens_seen": 3219587072
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.9109292248435245e-06,
      "loss": 1.2413,
      "theoretical_loss": 3.33136151904911,
      "tokens_seen": 3219718144
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.902904830685284e-06,
      "loss": 1.2408,
      "theoretical_loss": 3.3313513290419006,
      "tokens_seen": 3219849216
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.894880436527043e-06,
      "loss": 1.2453,
      "theoretical_loss": 3.3313411395656334,
      "tokens_seen": 3219980288
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.886856042368802e-06,
      "loss": 1.2576,
      "theoretical_loss": 3.3313309506202597,
      "tokens_seen": 3220111360
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.878831648210561e-06,
      "loss": 1.2467,
      "theoretical_loss": 3.33132076220573,
      "tokens_seen": 3220242432
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.87080725405232e-06,
      "loss": 1.2291,
      "theoretical_loss": 3.3313105743219946,
      "tokens_seen": 3220373504
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.862782859894078e-06,
      "loss": 1.209,
      "theoretical_loss": 3.331300386969005,
      "tokens_seen": 3220504576
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.854758465735837e-06,
      "loss": 1.3373,
      "theoretical_loss": 3.3312902001467117,
      "tokens_seen": 3220635648
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.846734071577596e-06,
      "loss": 1.3094,
      "theoretical_loss": 3.3312800138550656,
      "tokens_seen": 3220766720
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.838709677419355e-06,
      "loss": 1.3397,
      "theoretical_loss": 3.331269828094017,
      "tokens_seen": 3220897792
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.830685283261114e-06,
      "loss": 1.2528,
      "theoretical_loss": 3.3312596428635173,
      "tokens_seen": 3221028864
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.822660889102873e-06,
      "loss": 1.204,
      "theoretical_loss": 3.3312494581635166,
      "tokens_seen": 3221159936
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.8146364949446315e-06,
      "loss": 1.3595,
      "theoretical_loss": 3.331239273993966,
      "tokens_seen": 3221291008
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.806612100786391e-06,
      "loss": 1.1886,
      "theoretical_loss": 3.331229090354816,
      "tokens_seen": 3221422080
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.79858770662815e-06,
      "loss": 1.3016,
      "theoretical_loss": 3.331218907246018,
      "tokens_seen": 3221553152
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.790563312469909e-06,
      "loss": 1.2791,
      "theoretical_loss": 3.3312087246675226,
      "tokens_seen": 3221684224
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.782538918311668e-06,
      "loss": 1.2602,
      "theoretical_loss": 3.3311985426192807,
      "tokens_seen": 3221815296
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.774514524153427e-06,
      "loss": 1.2784,
      "theoretical_loss": 3.3311883611012423,
      "tokens_seen": 3221946368
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.766490129995185e-06,
      "loss": 1.2474,
      "theoretical_loss": 3.3311781801133593,
      "tokens_seen": 3222077440
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.758465735836944e-06,
      "loss": 1.3175,
      "theoretical_loss": 3.331167999655582,
      "tokens_seen": 3222208512
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.750441341678703e-06,
      "loss": 1.2436,
      "theoretical_loss": 3.3311578197278613,
      "tokens_seen": 3222339584
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.742416947520462e-06,
      "loss": 1.204,
      "theoretical_loss": 3.331147640330148,
      "tokens_seen": 3222470656
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.734392553362221e-06,
      "loss": 1.2401,
      "theoretical_loss": 3.3311374614623928,
      "tokens_seen": 3222601728
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.72636815920398e-06,
      "loss": 1.2818,
      "theoretical_loss": 3.3311272831245473,
      "tokens_seen": 3222732800
    },
    {
      "epoch": 0.95,
      "objective/train/aligned_token_count": 0.9934417605400085,
      "objective/train/avg_token_score": 0.0065582506358623505,
      "objective/train/docs_used": 1170779,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2796614170074463,
      "objective/train/original_loss": 1.288522481918335,
      "objective/train/theoretical_loss": 3.3311171053165616,
      "objective/train/tokens_used": 1593322976,
      "objective/train/unlikelihood_loss": 0.01919618248939514,
      "theoretical_loss": 3.3311171053165616,
      "tokens_seen": 3222863872
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.718343765045739e-06,
      "loss": 1.2703,
      "theoretical_loss": 3.3311171053165616,
      "tokens_seen": 3222863872
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.7103193708874985e-06,
      "loss": 1.2649,
      "theoretical_loss": 3.3311069280383867,
      "tokens_seen": 3222994944
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.7022949767292575e-06,
      "loss": 1.3126,
      "theoretical_loss": 3.3310967512899734,
      "tokens_seen": 3223126016
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.6942705825710166e-06,
      "loss": 1.3529,
      "theoretical_loss": 3.331086575071273,
      "tokens_seen": 3223257088
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.686246188412776e-06,
      "loss": 1.3056,
      "theoretical_loss": 3.331076399382236,
      "tokens_seen": 3223388160
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.678221794254534e-06,
      "loss": 1.3058,
      "theoretical_loss": 3.3310662242228135,
      "tokens_seen": 3223519232
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.670197400096293e-06,
      "loss": 1.1965,
      "theoretical_loss": 3.331056049592956,
      "tokens_seen": 3223650304
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.662173005938052e-06,
      "loss": 1.3148,
      "theoretical_loss": 3.3310458754926153,
      "tokens_seen": 3223781376
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.654148611779811e-06,
      "loss": 1.1564,
      "theoretical_loss": 3.3310357019217416,
      "tokens_seen": 3223912448
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.64612421762157e-06,
      "loss": 1.2497,
      "theoretical_loss": 3.331025528880286,
      "tokens_seen": 3224043520
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.638099823463328e-06,
      "loss": 1.2703,
      "theoretical_loss": 3.3310153563681992,
      "tokens_seen": 3224174592
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.630075429305087e-06,
      "loss": 1.2481,
      "theoretical_loss": 3.3310051843854325,
      "tokens_seen": 3224305664
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.622051035146846e-06,
      "loss": 1.2929,
      "theoretical_loss": 3.330995012931937,
      "tokens_seen": 3224436736
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.6140266409886055e-06,
      "loss": 1.3141,
      "theoretical_loss": 3.3309848420076626,
      "tokens_seen": 3224567808
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.6060022468303645e-06,
      "loss": 1.3422,
      "theoretical_loss": 3.3309746716125614,
      "tokens_seen": 3224698880
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.597977852672124e-06,
      "loss": 1.2879,
      "theoretical_loss": 3.330964501746584,
      "tokens_seen": 3224829952
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.589953458513883e-06,
      "loss": 1.2896,
      "theoretical_loss": 3.3309543324096813,
      "tokens_seen": 3224961024
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.581929064355641e-06,
      "loss": 1.2544,
      "theoretical_loss": 3.3309441636018047,
      "tokens_seen": 3225092096
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.5739046701974e-06,
      "loss": 1.246,
      "theoretical_loss": 3.330933995322904,
      "tokens_seen": 3225223168
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.565880276039159e-06,
      "loss": 1.3323,
      "theoretical_loss": 3.3309238275729314,
      "tokens_seen": 3225354240
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.557855881880918e-06,
      "loss": 1.2441,
      "theoretical_loss": 3.3309136603518374,
      "tokens_seen": 3225485312
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.549831487722677e-06,
      "loss": 1.274,
      "theoretical_loss": 3.330903493659573,
      "tokens_seen": 3225616384
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.541807093564436e-06,
      "loss": 1.2723,
      "theoretical_loss": 3.3308933274960895,
      "tokens_seen": 3225747456
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.533782699406195e-06,
      "loss": 1.3076,
      "theoretical_loss": 3.3308831618613373,
      "tokens_seen": 3225878528
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.525758305247954e-06,
      "loss": 1.2906,
      "theoretical_loss": 3.330872996755268,
      "tokens_seen": 3226009600
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.517733911089713e-06,
      "loss": 1.2482,
      "theoretical_loss": 3.330862832177832,
      "tokens_seen": 3226140672
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.509709516931472e-06,
      "loss": 1.262,
      "theoretical_loss": 3.330852668128981,
      "tokens_seen": 3226271744
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.5016851227732314e-06,
      "loss": 1.3562,
      "theoretical_loss": 3.330842504608666,
      "tokens_seen": 3226402816
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.4936607286149905e-06,
      "loss": 1.262,
      "theoretical_loss": 3.330832341616838,
      "tokens_seen": 3226533888
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.485636334456749e-06,
      "loss": 1.3281,
      "theoretical_loss": 3.3308221791534476,
      "tokens_seen": 3226664960
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.477611940298508e-06,
      "loss": 1.2446,
      "theoretical_loss": 3.330812017218446,
      "tokens_seen": 3226796032
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.469587546140267e-06,
      "loss": 1.2333,
      "theoretical_loss": 3.3308018558117847,
      "tokens_seen": 3226927104
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.461563151982025e-06,
      "loss": 1.3364,
      "theoretical_loss": 3.3307916949334144,
      "tokens_seen": 3227058176
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.453538757823784e-06,
      "loss": 1.3093,
      "theoretical_loss": 3.330781534583286,
      "tokens_seen": 3227189248
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.445514363665543e-06,
      "loss": 1.31,
      "theoretical_loss": 3.330771374761351,
      "tokens_seen": 3227320320
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.437489969507302e-06,
      "loss": 1.2256,
      "theoretical_loss": 3.330761215467561,
      "tokens_seen": 3227451392
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.429465575349061e-06,
      "loss": 1.2598,
      "theoretical_loss": 3.3307510567018657,
      "tokens_seen": 3227582464
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.42144118119082e-06,
      "loss": 1.2724,
      "theoretical_loss": 3.3307408984642173,
      "tokens_seen": 3227713536
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.413416787032579e-06,
      "loss": 1.2771,
      "theoretical_loss": 3.330730740754566,
      "tokens_seen": 3227844608
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.4053923928743385e-06,
      "loss": 1.3221,
      "theoretical_loss": 3.330720583572864,
      "tokens_seen": 3227975680
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.3973679987160975e-06,
      "loss": 1.344,
      "theoretical_loss": 3.3307104269190617,
      "tokens_seen": 3228106752
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.389343604557856e-06,
      "loss": 1.2469,
      "theoretical_loss": 3.330700270793111,
      "tokens_seen": 3228237824
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.381319210399615e-06,
      "loss": 1.2955,
      "theoretical_loss": 3.3306901151949617,
      "tokens_seen": 3228368896
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.373294816241374e-06,
      "loss": 1.1984,
      "theoretical_loss": 3.330679960124566,
      "tokens_seen": 3228499968
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.365270422083133e-06,
      "loss": 1.287,
      "theoretical_loss": 3.3306698055818744,
      "tokens_seen": 3228631040
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.357246027924892e-06,
      "loss": 1.3215,
      "theoretical_loss": 3.330659651566839,
      "tokens_seen": 3228762112
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.349221633766651e-06,
      "loss": 1.3548,
      "theoretical_loss": 3.33064949807941,
      "tokens_seen": 3228893184
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.34119723960841e-06,
      "loss": 1.2143,
      "theoretical_loss": 3.3306393451195393,
      "tokens_seen": 3229024256
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.333172845450169e-06,
      "loss": 1.2147,
      "theoretical_loss": 3.3306291926871774,
      "tokens_seen": 3229155328
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.325148451291928e-06,
      "loss": 1.2871,
      "theoretical_loss": 3.3306190407822758,
      "tokens_seen": 3229286400
    },
    {
      "epoch": 0.96,
      "objective/train/aligned_token_count": 0.9933333396911621,
      "objective/train/avg_token_score": 0.006666652392596006,
      "objective/train/docs_used": 1173331,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2112382650375366,
      "objective/train/original_loss": 1.2206459045410156,
      "objective/train/theoretical_loss": 3.330608889404786,
      "objective/train/tokens_used": 1599876576,
      "objective/train/unlikelihood_loss": 0.018875544890761375,
      "theoretical_loss": 3.330608889404786,
      "tokens_seen": 3229417472
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.317124057133686e-06,
      "loss": 1.2642,
      "theoretical_loss": 3.330608889404786,
      "tokens_seen": 3229417472
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.3090996629754455e-06,
      "loss": 1.279,
      "theoretical_loss": 3.3305987385546585,
      "tokens_seen": 3229548544
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.3010752688172045e-06,
      "loss": 1.3037,
      "theoretical_loss": 3.330588588231845,
      "tokens_seen": 3229679616
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.293050874658963e-06,
      "loss": 1.2052,
      "theoretical_loss": 3.330578438436297,
      "tokens_seen": 3229810688
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.285026480500722e-06,
      "loss": 1.1643,
      "theoretical_loss": 3.3305682891679647,
      "tokens_seen": 3229941760
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.277002086342481e-06,
      "loss": 1.2738,
      "theoretical_loss": 3.3305581404268003,
      "tokens_seen": 3230072832
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.26897769218424e-06,
      "loss": 1.2458,
      "theoretical_loss": 3.3305479922127548,
      "tokens_seen": 3230203904
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.260953298025999e-06,
      "loss": 1.3135,
      "theoretical_loss": 3.3305378445257787,
      "tokens_seen": 3230334976
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.252928903867758e-06,
      "loss": 1.2393,
      "theoretical_loss": 3.3305276973658247,
      "tokens_seen": 3230466048
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.244904509709517e-06,
      "loss": 1.2688,
      "theoretical_loss": 3.3305175507328424,
      "tokens_seen": 3230597120
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.236880115551276e-06,
      "loss": 1.2781,
      "theoretical_loss": 3.3305074046267844,
      "tokens_seen": 3230728192
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.228855721393035e-06,
      "loss": 1.2162,
      "theoretical_loss": 3.3304972590476014,
      "tokens_seen": 3230859264
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.220831327234794e-06,
      "loss": 1.3294,
      "theoretical_loss": 3.330487113995244,
      "tokens_seen": 3230990336
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.212806933076553e-06,
      "loss": 1.2749,
      "theoretical_loss": 3.330476969469665,
      "tokens_seen": 3231121408
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.2047825389183115e-06,
      "loss": 1.3123,
      "theoretical_loss": 3.3304668254708143,
      "tokens_seen": 3231252480
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.196758144760071e-06,
      "loss": 1.2894,
      "theoretical_loss": 3.3304566819986436,
      "tokens_seen": 3231383552
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.18873375060183e-06,
      "loss": 1.2427,
      "theoretical_loss": 3.3304465390531046,
      "tokens_seen": 3231514624
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.180709356443589e-06,
      "loss": 1.2705,
      "theoretical_loss": 3.3304363966341484,
      "tokens_seen": 3231645696
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.172684962285348e-06,
      "loss": 1.2836,
      "theoretical_loss": 3.330426254741726,
      "tokens_seen": 3231776768
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.164660568127107e-06,
      "loss": 1.3421,
      "theoretical_loss": 3.3304161133757892,
      "tokens_seen": 3231907840
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.156636173968866e-06,
      "loss": 1.2624,
      "theoretical_loss": 3.3304059725362887,
      "tokens_seen": 3232038912
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.148611779810625e-06,
      "loss": 1.2343,
      "theoretical_loss": 3.3303958322231764,
      "tokens_seen": 3232169984
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.140587385652383e-06,
      "loss": 1.2092,
      "theoretical_loss": 3.330385692436403,
      "tokens_seen": 3232301056
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.132562991494142e-06,
      "loss": 1.3114,
      "theoretical_loss": 3.3303755531759207,
      "tokens_seen": 3232432128
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.124538597335901e-06,
      "loss": 1.3017,
      "theoretical_loss": 3.33036541444168,
      "tokens_seen": 3232563200
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.11651420317766e-06,
      "loss": 1.3193,
      "theoretical_loss": 3.3303552762336333,
      "tokens_seen": 3232694272
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.1084898090194186e-06,
      "loss": 1.2658,
      "theoretical_loss": 3.3303451385517304,
      "tokens_seen": 3232825344
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.100465414861178e-06,
      "loss": 1.3347,
      "theoretical_loss": 3.330335001395924,
      "tokens_seen": 3232956416
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.092441020702937e-06,
      "loss": 1.2347,
      "theoretical_loss": 3.330324864766165,
      "tokens_seen": 3233087488
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.084416626544696e-06,
      "loss": 1.298,
      "theoretical_loss": 3.330314728662405,
      "tokens_seen": 3233218560
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.076392232386455e-06,
      "loss": 1.2653,
      "theoretical_loss": 3.3303045930845947,
      "tokens_seen": 3233349632
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.068367838228214e-06,
      "loss": 1.2707,
      "theoretical_loss": 3.3302944580326868,
      "tokens_seen": 3233480704
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.060343444069973e-06,
      "loss": 1.2604,
      "theoretical_loss": 3.330284323506631,
      "tokens_seen": 3233611776
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.052319049911732e-06,
      "loss": 1.2109,
      "theoretical_loss": 3.3302741895063797,
      "tokens_seen": 3233742848
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.044294655753491e-06,
      "loss": 1.2543,
      "theoretical_loss": 3.3302640560318846,
      "tokens_seen": 3233873920
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.03627026159525e-06,
      "loss": 1.2852,
      "theoretical_loss": 3.3302539230830965,
      "tokens_seen": 3234004992
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.028245867437009e-06,
      "loss": 1.2535,
      "theoretical_loss": 3.330243790659967,
      "tokens_seen": 3234136064
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.020221473278768e-06,
      "loss": 1.2753,
      "theoretical_loss": 3.330233658762448,
      "tokens_seen": 3234267136
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.012197079120526e-06,
      "loss": 1.2324,
      "theoretical_loss": 3.3302235273904897,
      "tokens_seen": 3234398208
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.0041726849622855e-06,
      "loss": 1.293,
      "theoretical_loss": 3.330213396544045,
      "tokens_seen": 3234529280
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.9961482908040445e-06,
      "loss": 1.2701,
      "theoretical_loss": 3.330203266223064,
      "tokens_seen": 3234660352
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.988123896645804e-06,
      "loss": 1.3036,
      "theoretical_loss": 3.3301931364274995,
      "tokens_seen": 3234791424
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.980099502487563e-06,
      "loss": 1.2814,
      "theoretical_loss": 3.3301830071573018,
      "tokens_seen": 3234922496
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.972075108329322e-06,
      "loss": 1.3104,
      "theoretical_loss": 3.330172878412423,
      "tokens_seen": 3235053568
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.96405071417108e-06,
      "loss": 1.2789,
      "theoretical_loss": 3.3301627501928146,
      "tokens_seen": 3235184640
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.956026320012839e-06,
      "loss": 1.2311,
      "theoretical_loss": 3.330152622498428,
      "tokens_seen": 3235315712
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.948001925854598e-06,
      "loss": 1.3454,
      "theoretical_loss": 3.3301424953292145,
      "tokens_seen": 3235446784
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.939977531696357e-06,
      "loss": 1.2013,
      "theoretical_loss": 3.3301323686851254,
      "tokens_seen": 3235577856
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.931953137538116e-06,
      "loss": 1.3263,
      "theoretical_loss": 3.3301222425661123,
      "tokens_seen": 3235708928
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.923928743379875e-06,
      "loss": 1.2643,
      "theoretical_loss": 3.3301121169721277,
      "tokens_seen": 3235840000
    },
    {
      "epoch": 0.96,
      "objective/train/aligned_token_count": 0.9917211532592773,
      "objective/train/avg_token_score": 0.008278782479465008,
      "objective/train/docs_used": 1175644,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2828584909439087,
      "objective/train/original_loss": 1.2947909832000732,
      "objective/train/theoretical_loss": 3.330101991903122,
      "objective/train/tokens_used": 1606430176,
      "objective/train/unlikelihood_loss": 0.022535519674420357,
      "theoretical_loss": 3.330101991903122,
      "tokens_seen": 3235971072
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.9159043492216334e-06,
      "loss": 1.2691,
      "theoretical_loss": 3.330101991903122,
      "tokens_seen": 3235971072
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.9078799550633925e-06,
      "loss": 1.259,
      "theoretical_loss": 3.3300918673590467,
      "tokens_seen": 3236102144
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.8998555609051516e-06,
      "loss": 1.2953,
      "theoretical_loss": 3.330081743339854,
      "tokens_seen": 3236233216
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.891831166746911e-06,
      "loss": 1.223,
      "theoretical_loss": 3.3300716198454947,
      "tokens_seen": 3236364288
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.88380677258867e-06,
      "loss": 1.2879,
      "theoretical_loss": 3.3300614968759215,
      "tokens_seen": 3236495360
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.875782378430429e-06,
      "loss": 1.2684,
      "theoretical_loss": 3.3300513744310845,
      "tokens_seen": 3236626432
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.867757984272188e-06,
      "loss": 1.2037,
      "theoretical_loss": 3.330041252510936,
      "tokens_seen": 3236757504
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.859733590113947e-06,
      "loss": 1.2611,
      "theoretical_loss": 3.330031131115428,
      "tokens_seen": 3236888576
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.851709195955706e-06,
      "loss": 1.2859,
      "theoretical_loss": 3.3300210102445105,
      "tokens_seen": 3237019648
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.843684801797465e-06,
      "loss": 1.2506,
      "theoretical_loss": 3.330010889898137,
      "tokens_seen": 3237150720
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.835660407639224e-06,
      "loss": 1.2043,
      "theoretical_loss": 3.330000770076258,
      "tokens_seen": 3237281792
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.827636013480982e-06,
      "loss": 1.253,
      "theoretical_loss": 3.329990650778825,
      "tokens_seen": 3237412864
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.819611619322741e-06,
      "loss": 1.3165,
      "theoretical_loss": 3.3299805320057905,
      "tokens_seen": 3237543936
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.8115872251645e-06,
      "loss": 1.3289,
      "theoretical_loss": 3.329970413757105,
      "tokens_seen": 3237675008
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.803562831006259e-06,
      "loss": 1.3464,
      "theoretical_loss": 3.3299602960327204,
      "tokens_seen": 3237806080
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.795538436848018e-06,
      "loss": 1.2003,
      "theoretical_loss": 3.329950178832589,
      "tokens_seen": 3237937152
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.787514042689777e-06,
      "loss": 1.2878,
      "theoretical_loss": 3.3299400621566613,
      "tokens_seen": 3238068224
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.779489648531536e-06,
      "loss": 1.2501,
      "theoretical_loss": 3.32992994600489,
      "tokens_seen": 3238199296
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.7714652543732952e-06,
      "loss": 1.2533,
      "theoretical_loss": 3.329919830377226,
      "tokens_seen": 3238330368
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.763440860215054e-06,
      "loss": 1.2613,
      "theoretical_loss": 3.329909715273621,
      "tokens_seen": 3238461440
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.755416466056813e-06,
      "loss": 1.2989,
      "theoretical_loss": 3.329899600694027,
      "tokens_seen": 3238592512
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.747392071898572e-06,
      "loss": 1.3156,
      "theoretical_loss": 3.3298894866383955,
      "tokens_seen": 3238723584
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.739367677740331e-06,
      "loss": 1.3383,
      "theoretical_loss": 3.329879373106678,
      "tokens_seen": 3238854656
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.7313432835820893e-06,
      "loss": 1.1862,
      "theoretical_loss": 3.3298692600988264,
      "tokens_seen": 3238985728
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.7233188894238483e-06,
      "loss": 1.2863,
      "theoretical_loss": 3.3298591476147923,
      "tokens_seen": 3239116800
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.7152944952656074e-06,
      "loss": 1.2905,
      "theoretical_loss": 3.3298490356545276,
      "tokens_seen": 3239247872
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.7072701011073664e-06,
      "loss": 1.317,
      "theoretical_loss": 3.329838924217983,
      "tokens_seen": 3239378944
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6992457069491255e-06,
      "loss": 1.3207,
      "theoretical_loss": 3.3298288133051113,
      "tokens_seen": 3239510016
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6912213127908845e-06,
      "loss": 1.2419,
      "theoretical_loss": 3.3298187029158637,
      "tokens_seen": 3239641088
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6831969186326436e-06,
      "loss": 1.2246,
      "theoretical_loss": 3.329808593050192,
      "tokens_seen": 3239772160
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6751725244744022e-06,
      "loss": 1.2402,
      "theoretical_loss": 3.329798483708048,
      "tokens_seen": 3239903232
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6671481303161613e-06,
      "loss": 1.264,
      "theoretical_loss": 3.3297883748893833,
      "tokens_seen": 3240034304
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6591237361579204e-06,
      "loss": 1.3145,
      "theoretical_loss": 3.3297782665941495,
      "tokens_seen": 3240165376
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6510993419996794e-06,
      "loss": 1.311,
      "theoretical_loss": 3.329768158822298,
      "tokens_seen": 3240296448
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6430749478414385e-06,
      "loss": 1.3663,
      "theoretical_loss": 3.3297580515737817,
      "tokens_seen": 3240427520
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6350505536831967e-06,
      "loss": 1.2952,
      "theoretical_loss": 3.329747944848551,
      "tokens_seen": 3240558592
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.6270261595249557e-06,
      "loss": 1.2805,
      "theoretical_loss": 3.3297378386465586,
      "tokens_seen": 3240689664
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.619001765366715e-06,
      "loss": 1.2474,
      "theoretical_loss": 3.329727732967756,
      "tokens_seen": 3240820736
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.610977371208474e-06,
      "loss": 1.2103,
      "theoretical_loss": 3.329717627812095,
      "tokens_seen": 3240951808
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.602952977050233e-06,
      "loss": 1.2939,
      "theoretical_loss": 3.3297075231795263,
      "tokens_seen": 3241082880
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.594928582891992e-06,
      "loss": 1.2923,
      "theoretical_loss": 3.3296974190700035,
      "tokens_seen": 3241213952
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.5869041887337506e-06,
      "loss": 1.2721,
      "theoretical_loss": 3.329687315483477,
      "tokens_seen": 3241345024
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.5788797945755097e-06,
      "loss": 1.185,
      "theoretical_loss": 3.3296772124198992,
      "tokens_seen": 3241476096
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.5708554004172687e-06,
      "loss": 1.1843,
      "theoretical_loss": 3.329667109879222,
      "tokens_seen": 3241607168
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.562831006259028e-06,
      "loss": 1.2928,
      "theoretical_loss": 3.3296570078613965,
      "tokens_seen": 3241738240
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.554806612100787e-06,
      "loss": 1.2775,
      "theoretical_loss": 3.329646906366375,
      "tokens_seen": 3241869312
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.546782217942546e-06,
      "loss": 1.216,
      "theoretical_loss": 3.3296368053941094,
      "tokens_seen": 3242000384
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.538757823784304e-06,
      "loss": 1.2403,
      "theoretical_loss": 3.3296267049445514,
      "tokens_seen": 3242131456
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.530733429626063e-06,
      "loss": 1.2152,
      "theoretical_loss": 3.3296166050176526,
      "tokens_seen": 3242262528
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.5227090354678222e-06,
      "loss": 1.2469,
      "theoretical_loss": 3.329606505613365,
      "tokens_seen": 3242393600
    },
    {
      "epoch": 0.97,
      "objective/train/aligned_token_count": 0.9925158023834229,
      "objective/train/avg_token_score": 0.007484156172722578,
      "objective/train/docs_used": 1177712,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3009167909622192,
      "objective/train/original_loss": 1.3114185333251953,
      "objective/train/theoretical_loss": 3.329596406731641,
      "objective/train/tokens_used": 1612983776,
      "objective/train/unlikelihood_loss": 0.02094210684299469,
      "theoretical_loss": 3.329596406731641,
      "tokens_seen": 3242524672
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.5146846413095813e-06,
      "loss": 1.2629,
      "theoretical_loss": 3.329596406731641,
      "tokens_seen": 3242524672
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.5066602471513404e-06,
      "loss": 1.2934,
      "theoretical_loss": 3.329586308372431,
      "tokens_seen": 3242655744
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.498635852993099e-06,
      "loss": 1.3656,
      "theoretical_loss": 3.329576210535688,
      "tokens_seen": 3242786816
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.490611458834858e-06,
      "loss": 1.258,
      "theoretical_loss": 3.3295661132213636,
      "tokens_seen": 3242917888
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.482587064676617e-06,
      "loss": 1.2404,
      "theoretical_loss": 3.32955601642941,
      "tokens_seen": 3243048960
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.474562670518376e-06,
      "loss": 1.3723,
      "theoretical_loss": 3.329545920159778,
      "tokens_seen": 3243180032
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4665382763601352e-06,
      "loss": 1.2523,
      "theoretical_loss": 3.329535824412421,
      "tokens_seen": 3243311104
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4585138822018943e-06,
      "loss": 1.2603,
      "theoretical_loss": 3.3295257291872895,
      "tokens_seen": 3243442176
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4504894880436533e-06,
      "loss": 1.3463,
      "theoretical_loss": 3.329515634484336,
      "tokens_seen": 3243573248
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4424650938854116e-06,
      "loss": 1.2263,
      "theoretical_loss": 3.3295055403035123,
      "tokens_seen": 3243704320
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4344406997271706e-06,
      "loss": 1.2827,
      "theoretical_loss": 3.3294954466447706,
      "tokens_seen": 3243835392
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4264163055689297e-06,
      "loss": 1.2696,
      "theoretical_loss": 3.329485353508062,
      "tokens_seen": 3243966464
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4183919114106887e-06,
      "loss": 1.2166,
      "theoretical_loss": 3.3294752608933393,
      "tokens_seen": 3244097536
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4103675172524474e-06,
      "loss": 1.2529,
      "theoretical_loss": 3.329465168800554,
      "tokens_seen": 3244228608
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.4023431230942064e-06,
      "loss": 1.2506,
      "theoretical_loss": 3.329455077229658,
      "tokens_seen": 3244359680
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.3943187289359655e-06,
      "loss": 1.2392,
      "theoretical_loss": 3.3294449861806035,
      "tokens_seen": 3244490752
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.3862943347777246e-06,
      "loss": 1.2501,
      "theoretical_loss": 3.329434895653342,
      "tokens_seen": 3244621824
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.3782699406194836e-06,
      "loss": 1.3003,
      "theoretical_loss": 3.3294248056478257,
      "tokens_seen": 3244752896
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.3702455464612427e-06,
      "loss": 1.2612,
      "theoretical_loss": 3.3294147161640066,
      "tokens_seen": 3244883968
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.3622211523030017e-06,
      "loss": 1.3084,
      "theoretical_loss": 3.3294046272018365,
      "tokens_seen": 3245015040
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.35419675814476e-06,
      "loss": 1.3001,
      "theoretical_loss": 3.3293945387612673,
      "tokens_seen": 3245146112
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.346172363986519e-06,
      "loss": 1.3021,
      "theoretical_loss": 3.3293844508422517,
      "tokens_seen": 3245277184
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.338147969828278e-06,
      "loss": 1.3107,
      "theoretical_loss": 3.3293743634447406,
      "tokens_seen": 3245408256
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.330123575670037e-06,
      "loss": 1.2581,
      "theoretical_loss": 3.3293642765686866,
      "tokens_seen": 3245539328
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.3220991815117958e-06,
      "loss": 1.2444,
      "theoretical_loss": 3.3293541902140413,
      "tokens_seen": 3245670400
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.314074787353555e-06,
      "loss": 1.2821,
      "theoretical_loss": 3.3293441043807572,
      "tokens_seen": 3245801472
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.306050393195314e-06,
      "loss": 1.2248,
      "theoretical_loss": 3.329334019068786,
      "tokens_seen": 3245932544
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.298025999037073e-06,
      "loss": 1.32,
      "theoretical_loss": 3.32932393427808,
      "tokens_seen": 3246063616
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.290001604878832e-06,
      "loss": 1.2079,
      "theoretical_loss": 3.3293138500085906,
      "tokens_seen": 3246194688
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.281977210720591e-06,
      "loss": 1.2055,
      "theoretical_loss": 3.3293037662602702,
      "tokens_seen": 3246325760
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.27395281656235e-06,
      "loss": 1.286,
      "theoretical_loss": 3.329293683033071,
      "tokens_seen": 3246456832
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.265928422404109e-06,
      "loss": 1.2581,
      "theoretical_loss": 3.3292836003269444,
      "tokens_seen": 3246587904
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.2579040282458674e-06,
      "loss": 1.2322,
      "theoretical_loss": 3.3292735181418434,
      "tokens_seen": 3246718976
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.2498796340876264e-06,
      "loss": 1.2907,
      "theoretical_loss": 3.329263436477719,
      "tokens_seen": 3246850048
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.241855239929385e-06,
      "loss": 1.3255,
      "theoretical_loss": 3.329253355334524,
      "tokens_seen": 3246981120
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.233830845771144e-06,
      "loss": 1.2574,
      "theoretical_loss": 3.3292432747122103,
      "tokens_seen": 3247112192
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.225806451612903e-06,
      "loss": 1.3715,
      "theoretical_loss": 3.3292331946107296,
      "tokens_seen": 3247243264
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.2177820574546622e-06,
      "loss": 1.3042,
      "theoretical_loss": 3.3292231150300347,
      "tokens_seen": 3247374336
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.2097576632964213e-06,
      "loss": 1.2757,
      "theoretical_loss": 3.329213035970077,
      "tokens_seen": 3247505408
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.2017332691381804e-06,
      "loss": 1.2547,
      "theoretical_loss": 3.329202957430809,
      "tokens_seen": 3247636480
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1937088749799394e-06,
      "loss": 1.3438,
      "theoretical_loss": 3.329192879412182,
      "tokens_seen": 3247767552
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1856844808216985e-06,
      "loss": 1.2465,
      "theoretical_loss": 3.3291828019141496,
      "tokens_seen": 3247898624
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.177660086663457e-06,
      "loss": 1.2904,
      "theoretical_loss": 3.329172724936662,
      "tokens_seen": 3248029696
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.169635692505216e-06,
      "loss": 1.2984,
      "theoretical_loss": 3.329162648479673,
      "tokens_seen": 3248160768
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.161611298346975e-06,
      "loss": 1.3021,
      "theoretical_loss": 3.3291525725431335,
      "tokens_seen": 3248291840
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1535869041887334e-06,
      "loss": 1.3096,
      "theoretical_loss": 3.3291424971269965,
      "tokens_seen": 3248422912
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1455625100304925e-06,
      "loss": 1.2672,
      "theoretical_loss": 3.329132422231214,
      "tokens_seen": 3248553984
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1375381158722516e-06,
      "loss": 1.292,
      "theoretical_loss": 3.3291223478557375,
      "tokens_seen": 3248685056
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1295137217140106e-06,
      "loss": 1.2372,
      "theoretical_loss": 3.3291122740005195,
      "tokens_seen": 3248816128
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1214893275557697e-06,
      "loss": 1.2963,
      "theoretical_loss": 3.329102200665512,
      "tokens_seen": 3248947200
    },
    {
      "epoch": 0.97,
      "objective/train/aligned_token_count": 0.9940483570098877,
      "objective/train/avg_token_score": 0.005951669067144394,
      "objective/train/docs_used": 1180177,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2263606786727905,
      "objective/train/original_loss": 1.234593152999878,
      "objective/train/theoretical_loss": 3.3290921278506675,
      "objective/train/tokens_used": 1619537376,
      "objective/train/unlikelihood_loss": 0.0173080712556839,
      "theoretical_loss": 3.3290921278506675,
      "tokens_seen": 3249078272
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.1134649333975287e-06,
      "loss": 1.2997,
      "theoretical_loss": 3.3290921278506675,
      "tokens_seen": 3249078272
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.105440539239288e-06,
      "loss": 1.3169,
      "theoretical_loss": 3.329082055555938,
      "tokens_seen": 3249209344
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.097416145081047e-06,
      "loss": 1.2805,
      "theoretical_loss": 3.329071983781276,
      "tokens_seen": 3249340416
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.0893917509228055e-06,
      "loss": 1.3153,
      "theoretical_loss": 3.3290619125266327,
      "tokens_seen": 3249471488
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.0813673567645646e-06,
      "loss": 1.2557,
      "theoretical_loss": 3.3290518417919612,
      "tokens_seen": 3249602560
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.073342962606323e-06,
      "loss": 1.3125,
      "theoretical_loss": 3.3290417715772134,
      "tokens_seen": 3249733632
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.0653185684480823e-06,
      "loss": 1.2095,
      "theoretical_loss": 3.329031701882341,
      "tokens_seen": 3249864704
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.0572941742898413e-06,
      "loss": 1.3277,
      "theoretical_loss": 3.3290216327072972,
      "tokens_seen": 3249995776
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.0492697801316004e-06,
      "loss": 1.2744,
      "theoretical_loss": 3.329011564052034,
      "tokens_seen": 3250126848
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.041245385973359e-06,
      "loss": 1.2615,
      "theoretical_loss": 3.3290014959165024,
      "tokens_seen": 3250257920
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.033220991815118e-06,
      "loss": 1.2695,
      "theoretical_loss": 3.3289914283006556,
      "tokens_seen": 3250388992
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.025196597656877e-06,
      "loss": 1.1715,
      "theoretical_loss": 3.3289813612044465,
      "tokens_seen": 3250520064
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.017172203498636e-06,
      "loss": 1.2592,
      "theoretical_loss": 3.328971294627826,
      "tokens_seen": 3250651136
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.0091478093403952e-06,
      "loss": 1.3272,
      "theoretical_loss": 3.328961228570747,
      "tokens_seen": 3250782208
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.001123415182154e-06,
      "loss": 1.3097,
      "theoretical_loss": 3.328951163033161,
      "tokens_seen": 3250913280
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9930990210239125e-06,
      "loss": 1.3208,
      "theoretical_loss": 3.3289410980150214,
      "tokens_seen": 3251044352
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9850746268656716e-06,
      "loss": 1.2799,
      "theoretical_loss": 3.32893103351628,
      "tokens_seen": 3251175424
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9770502327074306e-06,
      "loss": 1.2594,
      "theoretical_loss": 3.3289209695368887,
      "tokens_seen": 3251306496
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9690258385491897e-06,
      "loss": 1.2757,
      "theoretical_loss": 3.3289109060768,
      "tokens_seen": 3251437568
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9610014443909487e-06,
      "loss": 1.2331,
      "theoretical_loss": 3.3289008431359663,
      "tokens_seen": 3251568640
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9529770502327074e-06,
      "loss": 1.2388,
      "theoretical_loss": 3.3288907807143397,
      "tokens_seen": 3251699712
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9449526560744664e-06,
      "loss": 1.2623,
      "theoretical_loss": 3.3288807188118725,
      "tokens_seen": 3251830784
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9369282619162255e-06,
      "loss": 1.3353,
      "theoretical_loss": 3.328870657428517,
      "tokens_seen": 3251961856
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9289038677579846e-06,
      "loss": 1.2503,
      "theoretical_loss": 3.328860596564226,
      "tokens_seen": 3252092928
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9208794735997436e-06,
      "loss": 1.197,
      "theoretical_loss": 3.328850536218951,
      "tokens_seen": 3252224000
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.9128550794415023e-06,
      "loss": 1.2378,
      "theoretical_loss": 3.328840476392645,
      "tokens_seen": 3252355072
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.904830685283261e-06,
      "loss": 1.3471,
      "theoretical_loss": 3.3288304170852596,
      "tokens_seen": 3252486144
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.89680629112502e-06,
      "loss": 1.277,
      "theoretical_loss": 3.3288203582967473,
      "tokens_seen": 3252617216
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.888781896966779e-06,
      "loss": 1.3472,
      "theoretical_loss": 3.328810300027061,
      "tokens_seen": 3252748288
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.880757502808538e-06,
      "loss": 1.2031,
      "theoretical_loss": 3.3288002422761527,
      "tokens_seen": 3252879360
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.872733108650297e-06,
      "loss": 1.2672,
      "theoretical_loss": 3.3287901850439745,
      "tokens_seen": 3253010432
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.864708714492056e-06,
      "loss": 1.3175,
      "theoretical_loss": 3.3287801283304788,
      "tokens_seen": 3253141504
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.856684320333815e-06,
      "loss": 1.2939,
      "theoretical_loss": 3.3287700721356184,
      "tokens_seen": 3253272576
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.848659926175574e-06,
      "loss": 1.2853,
      "theoretical_loss": 3.3287600164593454,
      "tokens_seen": 3253403648
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.840635532017333e-06,
      "loss": 1.2457,
      "theoretical_loss": 3.328749961301612,
      "tokens_seen": 3253534720
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.832611137859092e-06,
      "loss": 1.3385,
      "theoretical_loss": 3.3287399066623706,
      "tokens_seen": 3253665792
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.8245867437008506e-06,
      "loss": 1.2557,
      "theoretical_loss": 3.328729852541574,
      "tokens_seen": 3253796864
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.8165623495426097e-06,
      "loss": 1.2285,
      "theoretical_loss": 3.328719798939174,
      "tokens_seen": 3253927936
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.8085379553843683e-06,
      "loss": 1.2387,
      "theoretical_loss": 3.328709745855123,
      "tokens_seen": 3254059008
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.8005135612261274e-06,
      "loss": 1.3657,
      "theoretical_loss": 3.3286996932893738,
      "tokens_seen": 3254190080
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7924891670678864e-06,
      "loss": 1.3506,
      "theoretical_loss": 3.328689641241879,
      "tokens_seen": 3254321152
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7844647729096455e-06,
      "loss": 1.2875,
      "theoretical_loss": 3.3286795897125905,
      "tokens_seen": 3254452224
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7764403787514046e-06,
      "loss": 1.3047,
      "theoretical_loss": 3.328669538701461,
      "tokens_seen": 3254583296
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7684159845931636e-06,
      "loss": 1.264,
      "theoretical_loss": 3.3286594882084426,
      "tokens_seen": 3254714368
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7603915904349223e-06,
      "loss": 1.2863,
      "theoretical_loss": 3.3286494382334877,
      "tokens_seen": 3254845440
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7523671962766813e-06,
      "loss": 1.3533,
      "theoretical_loss": 3.3286393887765495,
      "tokens_seen": 3254976512
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7443428021184404e-06,
      "loss": 1.2319,
      "theoretical_loss": 3.3286293398375792,
      "tokens_seen": 3255107584
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.736318407960199e-06,
      "loss": 1.3209,
      "theoretical_loss": 3.3286192914165307,
      "tokens_seen": 3255238656
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.728294013801958e-06,
      "loss": 1.244,
      "theoretical_loss": 3.3286092435133554,
      "tokens_seen": 3255369728
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.720269619643717e-06,
      "loss": 1.2455,
      "theoretical_loss": 3.328599196128006,
      "tokens_seen": 3255500800
    },
    {
      "epoch": 0.97,
      "objective/train/aligned_token_count": 0.9918806552886963,
      "objective/train/avg_token_score": 0.00811933632940054,
      "objective/train/docs_used": 1182603,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.3268715143203735,
      "objective/train/original_loss": 1.3396532535552979,
      "objective/train/theoretical_loss": 3.328589149260435,
      "objective/train/tokens_used": 1626090976,
      "objective/train/unlikelihood_loss": 0.018495339900255203,
      "theoretical_loss": 3.328589149260435,
      "tokens_seen": 3255631872
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.7122452254854758e-06,
      "loss": 1.294,
      "theoretical_loss": 3.328589149260435,
      "tokens_seen": 3255631872
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.704220831327235e-06,
      "loss": 1.2714,
      "theoretical_loss": 3.328579102910595,
      "tokens_seen": 3255762944
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.696196437168994e-06,
      "loss": 1.1524,
      "theoretical_loss": 3.3285690570784383,
      "tokens_seen": 3255894016
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.688172043010753e-06,
      "loss": 1.2913,
      "theoretical_loss": 3.3285590117639177,
      "tokens_seen": 3256025088
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.680147648852512e-06,
      "loss": 1.3663,
      "theoretical_loss": 3.328548966966985,
      "tokens_seen": 3256156160
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.672123254694271e-06,
      "loss": 1.233,
      "theoretical_loss": 3.328538922687594,
      "tokens_seen": 3256287232
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.6640988605360297e-06,
      "loss": 1.2759,
      "theoretical_loss": 3.3285288789256957,
      "tokens_seen": 3256418304
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.6560744663777883e-06,
      "loss": 1.31,
      "theoretical_loss": 3.328518835681243,
      "tokens_seen": 3256549376
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.6480500722195474e-06,
      "loss": 1.3293,
      "theoretical_loss": 3.328508792954189,
      "tokens_seen": 3256680448
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.6400256780613064e-06,
      "loss": 1.3696,
      "theoretical_loss": 3.3284987507444863,
      "tokens_seen": 3256811520
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.6320012839030655e-06,
      "loss": 1.3242,
      "theoretical_loss": 3.328488709052087,
      "tokens_seen": 3256942592
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.6239768897448246e-06,
      "loss": 1.2372,
      "theoretical_loss": 3.328478667876943,
      "tokens_seen": 3257073664
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.615952495586583e-06,
      "loss": 1.2489,
      "theoretical_loss": 3.3284686272190083,
      "tokens_seen": 3257204736
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.6079281014283423e-06,
      "loss": 1.3045,
      "theoretical_loss": 3.3284585870782344,
      "tokens_seen": 3257335808
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5999037072701013e-06,
      "loss": 1.2614,
      "theoretical_loss": 3.328448547454574,
      "tokens_seen": 3257466880
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5918793131118604e-06,
      "loss": 1.3008,
      "theoretical_loss": 3.32843850834798,
      "tokens_seen": 3257597952
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5838549189536194e-06,
      "loss": 1.284,
      "theoretical_loss": 3.3284284697584043,
      "tokens_seen": 3257729024
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.575830524795378e-06,
      "loss": 1.2716,
      "theoretical_loss": 3.3284184316858,
      "tokens_seen": 3257860096
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5678061306371367e-06,
      "loss": 1.289,
      "theoretical_loss": 3.32840839413012,
      "tokens_seen": 3257991168
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5597817364788958e-06,
      "loss": 1.3033,
      "theoretical_loss": 3.3283983570913165,
      "tokens_seen": 3258122240
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.551757342320655e-06,
      "loss": 1.2516,
      "theoretical_loss": 3.328388320569342,
      "tokens_seen": 3258253312
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.543732948162414e-06,
      "loss": 1.2649,
      "theoretical_loss": 3.328378284564149,
      "tokens_seen": 3258384384
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.535708554004173e-06,
      "loss": 1.2505,
      "theoretical_loss": 3.32836824907569,
      "tokens_seen": 3258515456
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5276841598459316e-06,
      "loss": 1.2333,
      "theoretical_loss": 3.3283582141039187,
      "tokens_seen": 3258646528
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.5196597656876906e-06,
      "loss": 1.1795,
      "theoretical_loss": 3.3283481796487866,
      "tokens_seen": 3258777600
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.5116353715294497e-06,
      "loss": 1.2224,
      "theoretical_loss": 3.328338145710246,
      "tokens_seen": 3258908672
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.5036109773712088e-06,
      "loss": 1.2339,
      "theoretical_loss": 3.328328112288251,
      "tokens_seen": 3259039744
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.495586583212968e-06,
      "loss": 1.233,
      "theoretical_loss": 3.328318079382753,
      "tokens_seen": 3259170816
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.4875621890547264e-06,
      "loss": 1.2592,
      "theoretical_loss": 3.328308046993705,
      "tokens_seen": 3259301888
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.479537794896485e-06,
      "loss": 1.2766,
      "theoretical_loss": 3.32829801512106,
      "tokens_seen": 3259432960
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.471513400738244e-06,
      "loss": 1.2051,
      "theoretical_loss": 3.32828798376477,
      "tokens_seen": 3259564032
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.463489006580003e-06,
      "loss": 1.3319,
      "theoretical_loss": 3.3282779529247883,
      "tokens_seen": 3259695104
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.4554646124217623e-06,
      "loss": 1.3423,
      "theoretical_loss": 3.328267922601067,
      "tokens_seen": 3259826176
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.4474402182635213e-06,
      "loss": 1.2213,
      "theoretical_loss": 3.3282578927935593,
      "tokens_seen": 3259957248
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.4394158241052804e-06,
      "loss": 1.2553,
      "theoretical_loss": 3.3282478635022175,
      "tokens_seen": 3260088320
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.431391429947039e-06,
      "loss": 1.1988,
      "theoretical_loss": 3.3282378347269943,
      "tokens_seen": 3260219392
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.423367035788798e-06,
      "loss": 1.2429,
      "theoretical_loss": 3.3282278064678428,
      "tokens_seen": 3260350464
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.415342641630557e-06,
      "loss": 1.1554,
      "theoretical_loss": 3.328217778724715,
      "tokens_seen": 3260481536
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.4073182474723158e-06,
      "loss": 1.2153,
      "theoretical_loss": 3.3282077514975645,
      "tokens_seen": 3260612608
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.399293853314075e-06,
      "loss": 1.2688,
      "theoretical_loss": 3.3281977247863432,
      "tokens_seen": 3260743680
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.391269459155834e-06,
      "loss": 1.2132,
      "theoretical_loss": 3.328187698591004,
      "tokens_seen": 3260874752
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3832450649975925e-06,
      "loss": 1.2008,
      "theoretical_loss": 3.3281776729115,
      "tokens_seen": 3261005824
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3752206708393516e-06,
      "loss": 1.2915,
      "theoretical_loss": 3.3281676477477835,
      "tokens_seen": 3261136896
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3671962766811106e-06,
      "loss": 1.2129,
      "theoretical_loss": 3.3281576230998073,
      "tokens_seen": 3261267968
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3591718825228697e-06,
      "loss": 1.206,
      "theoretical_loss": 3.328147598967525,
      "tokens_seen": 3261399040
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3511474883646288e-06,
      "loss": 1.2495,
      "theoretical_loss": 3.3281375753508877,
      "tokens_seen": 3261530112
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.343123094206388e-06,
      "loss": 1.2421,
      "theoretical_loss": 3.3281275522498497,
      "tokens_seen": 3261661184
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3350987000481465e-06,
      "loss": 1.2466,
      "theoretical_loss": 3.328117529664363,
      "tokens_seen": 3261792256
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3270743058899055e-06,
      "loss": 1.2662,
      "theoretical_loss": 3.32810750759438,
      "tokens_seen": 3261923328
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.319049911731664e-06,
      "loss": 1.2368,
      "theoretical_loss": 3.3280974860398542,
      "tokens_seen": 3262054400
    },
    {
      "epoch": 0.98,
      "objective/train/aligned_token_count": 0.992702305316925,
      "objective/train/avg_token_score": 0.007297667674720287,
      "objective/train/docs_used": 1184948,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.346886157989502,
      "objective/train/original_loss": 1.358167052268982,
      "objective/train/theoretical_loss": 3.3280874650007384,
      "objective/train/tokens_used": 1632644576,
      "objective/train/unlikelihood_loss": 0.020045259967446327,
      "theoretical_loss": 3.3280874650007384,
      "tokens_seen": 3262185472
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.311025517573423e-06,
      "loss": 1.3215,
      "theoretical_loss": 3.3280874650007384,
      "tokens_seen": 3262185472
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.3030011234151823e-06,
      "loss": 1.3197,
      "theoretical_loss": 3.328077444476985,
      "tokens_seen": 3262316544
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.2949767292569413e-06,
      "loss": 1.2898,
      "theoretical_loss": 3.328067424468547,
      "tokens_seen": 3262447616
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.2869523350987e-06,
      "loss": 1.2627,
      "theoretical_loss": 3.328057404975377,
      "tokens_seen": 3262578688
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.278927940940459e-06,
      "loss": 1.2317,
      "theoretical_loss": 3.3280473859974276,
      "tokens_seen": 3262709760
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.270903546782218e-06,
      "loss": 1.3559,
      "theoretical_loss": 3.3280373675346526,
      "tokens_seen": 3262840832
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.262879152623977e-06,
      "loss": 1.3233,
      "theoretical_loss": 3.328027349587004,
      "tokens_seen": 3262971904
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.254854758465736e-06,
      "loss": 1.1511,
      "theoretical_loss": 3.328017332154434,
      "tokens_seen": 3263102976
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.2468303643074953e-06,
      "loss": 1.2287,
      "theoretical_loss": 3.328007315236897,
      "tokens_seen": 3263234048
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.238805970149254e-06,
      "loss": 1.2532,
      "theoretical_loss": 3.3279972988343447,
      "tokens_seen": 3263365120
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.2307815759910125e-06,
      "loss": 1.205,
      "theoretical_loss": 3.3279872829467307,
      "tokens_seen": 3263496192
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.2227571818327716e-06,
      "loss": 1.2645,
      "theoretical_loss": 3.327977267574007,
      "tokens_seen": 3263627264
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.2147327876745306e-06,
      "loss": 1.2606,
      "theoretical_loss": 3.327967252716127,
      "tokens_seen": 3263758336
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.2067083935162897e-06,
      "loss": 1.2715,
      "theoretical_loss": 3.3279572383730436,
      "tokens_seen": 3263889408
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.1986839993580488e-06,
      "loss": 1.2785,
      "theoretical_loss": 3.327947224544709,
      "tokens_seen": 3264020480
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.1906596051998074e-06,
      "loss": 1.2803,
      "theoretical_loss": 3.327937211231077,
      "tokens_seen": 3264151552
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.1826352110415665e-06,
      "loss": 1.2691,
      "theoretical_loss": 3.3279271984321,
      "tokens_seen": 3264282624
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.1746108168833255e-06,
      "loss": 1.2902,
      "theoretical_loss": 3.327917186147731,
      "tokens_seen": 3264413696
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.1665864227250846e-06,
      "loss": 1.2925,
      "theoretical_loss": 3.3279071743779225,
      "tokens_seen": 3264544768
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.158562028566843e-06,
      "loss": 1.2732,
      "theoretical_loss": 3.327897163122628,
      "tokens_seen": 3264675840
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.1505376344086023e-06,
      "loss": 1.3162,
      "theoretical_loss": 3.3278871523818,
      "tokens_seen": 3264806912
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.142513240250361e-06,
      "loss": 1.3019,
      "theoretical_loss": 3.3278771421553914,
      "tokens_seen": 3264937984
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.13448884609212e-06,
      "loss": 1.2292,
      "theoretical_loss": 3.3278671324433557,
      "tokens_seen": 3265069056
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.126464451933879e-06,
      "loss": 1.2384,
      "theoretical_loss": 3.327857123245645,
      "tokens_seen": 3265200128
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.118440057775638e-06,
      "loss": 1.2755,
      "theoretical_loss": 3.327847114562213,
      "tokens_seen": 3265331200
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.110415663617397e-06,
      "loss": 1.3188,
      "theoretical_loss": 3.3278371063930114,
      "tokens_seen": 3265462272
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.1023912694591558e-06,
      "loss": 1.3211,
      "theoretical_loss": 3.3278270987379948,
      "tokens_seen": 3265593344
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.094366875300915e-06,
      "loss": 1.2516,
      "theoretical_loss": 3.3278170915971153,
      "tokens_seen": 3265724416
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.086342481142674e-06,
      "loss": 1.1987,
      "theoretical_loss": 3.3278070849703254,
      "tokens_seen": 3265855488
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.078318086984433e-06,
      "loss": 1.2488,
      "theoretical_loss": 3.3277970788575786,
      "tokens_seen": 3265986560
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0702936928261916e-06,
      "loss": 1.3102,
      "theoretical_loss": 3.327787073258828,
      "tokens_seen": 3266117632
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0622692986679506e-06,
      "loss": 1.1898,
      "theoretical_loss": 3.3277770681740266,
      "tokens_seen": 3266248704
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0542449045097093e-06,
      "loss": 1.3102,
      "theoretical_loss": 3.327767063603127,
      "tokens_seen": 3266379776
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0462205103514683e-06,
      "loss": 1.2635,
      "theoretical_loss": 3.327757059546082,
      "tokens_seen": 3266510848
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0381961161932274e-06,
      "loss": 1.2505,
      "theoretical_loss": 3.327747056002845,
      "tokens_seen": 3266641920
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0301717220349865e-06,
      "loss": 1.2993,
      "theoretical_loss": 3.327737052973369,
      "tokens_seen": 3266772992
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0221473278767455e-06,
      "loss": 1.3556,
      "theoretical_loss": 3.3277270504576073,
      "tokens_seen": 3266904064
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0141229337185046e-06,
      "loss": 1.2793,
      "theoretical_loss": 3.3277170484555123,
      "tokens_seen": 3267035136
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.006098539560263e-06,
      "loss": 1.2665,
      "theoretical_loss": 3.327707046967037,
      "tokens_seen": 3267166208
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.9980741454020223e-06,
      "loss": 1.1904,
      "theoretical_loss": 3.327697045992135,
      "tokens_seen": 3267297280
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.9900497512437813e-06,
      "loss": 1.3017,
      "theoretical_loss": 3.327687045530759,
      "tokens_seen": 3267428352
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.98202535708554e-06,
      "loss": 1.2489,
      "theoretical_loss": 3.3276770455828615,
      "tokens_seen": 3267559424
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.974000962927299e-06,
      "loss": 1.2768,
      "theoretical_loss": 3.3276670461483966,
      "tokens_seen": 3267690496
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.965976568769058e-06,
      "loss": 1.2888,
      "theoretical_loss": 3.3276570472273166,
      "tokens_seen": 3267821568
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.9579521746108167e-06,
      "loss": 1.3347,
      "theoretical_loss": 3.327647048819575,
      "tokens_seen": 3267952640
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.9499277804525758e-06,
      "loss": 1.25,
      "theoretical_loss": 3.3276370509251243,
      "tokens_seen": 3268083712
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.941903386294335e-06,
      "loss": 1.2672,
      "theoretical_loss": 3.327627053543918,
      "tokens_seen": 3268214784
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.933878992136094e-06,
      "loss": 1.281,
      "theoretical_loss": 3.3276170566759093,
      "tokens_seen": 3268345856
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.925854597977853e-06,
      "loss": 1.2768,
      "theoretical_loss": 3.327607060321051,
      "tokens_seen": 3268476928
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.917830203819612e-06,
      "loss": 1.1704,
      "theoretical_loss": 3.3275970644792956,
      "tokens_seen": 3268608000
    },
    {
      "epoch": 0.98,
      "objective/train/aligned_token_count": 0.9905927181243896,
      "objective/train/avg_token_score": 0.009407276287674904,
      "objective/train/docs_used": 1187457,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2602373361587524,
      "objective/train/original_loss": 1.2734317779541016,
      "objective/train/theoretical_loss": 3.327587069150597,
      "objective/train/tokens_used": 1639198176,
      "objective/train/unlikelihood_loss": 0.025005128234624863,
      "theoretical_loss": 3.327587069150597,
      "tokens_seen": 3268739072
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.9098058096613706e-06,
      "loss": 1.2838,
      "theoretical_loss": 3.327587069150597,
      "tokens_seen": 3268739072
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.9017814155031295e-06,
      "loss": 1.2723,
      "theoretical_loss": 3.3275770743349087,
      "tokens_seen": 3268870144
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8937570213448886e-06,
      "loss": 1.3132,
      "theoretical_loss": 3.327567080032183,
      "tokens_seen": 3269001216
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8857326271866476e-06,
      "loss": 1.202,
      "theoretical_loss": 3.327557086242373,
      "tokens_seen": 3269132288
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8777082330284065e-06,
      "loss": 1.4224,
      "theoretical_loss": 3.327547092965432,
      "tokens_seen": 3269263360
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8696838388701655e-06,
      "loss": 1.2733,
      "theoretical_loss": 3.327537100201313,
      "tokens_seen": 3269394432
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8616594447119242e-06,
      "loss": 1.2278,
      "theoretical_loss": 3.3275271079499693,
      "tokens_seen": 3269525504
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8536350505536832e-06,
      "loss": 1.2731,
      "theoretical_loss": 3.3275171162113546,
      "tokens_seen": 3269656576
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8456106563954423e-06,
      "loss": 1.3293,
      "theoretical_loss": 3.327507124985421,
      "tokens_seen": 3269787648
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8375862622372011e-06,
      "loss": 1.2301,
      "theoretical_loss": 3.327497134272122,
      "tokens_seen": 3269918720
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8295618680789602e-06,
      "loss": 1.2621,
      "theoretical_loss": 3.327487144071411,
      "tokens_seen": 3270049792
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8215374739207192e-06,
      "loss": 1.2262,
      "theoretical_loss": 3.3274771543832404,
      "tokens_seen": 3270180864
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8135130797624779e-06,
      "loss": 1.2061,
      "theoretical_loss": 3.327467165207565,
      "tokens_seen": 3270311936
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.805488685604237e-06,
      "loss": 1.189,
      "theoretical_loss": 3.3274571765443364,
      "tokens_seen": 3270443008
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.797464291445996e-06,
      "loss": 1.2427,
      "theoretical_loss": 3.327447188393508,
      "tokens_seen": 3270574080
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7894398972877548e-06,
      "loss": 1.2046,
      "theoretical_loss": 3.3274372007550337,
      "tokens_seen": 3270705152
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.781415503129514e-06,
      "loss": 1.266,
      "theoretical_loss": 3.327427213628866,
      "tokens_seen": 3270836224
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.773391108971273e-06,
      "loss": 1.3008,
      "theoretical_loss": 3.3274172270149585,
      "tokens_seen": 3270967296
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7653667148130316e-06,
      "loss": 1.2334,
      "theoretical_loss": 3.3274072409132645,
      "tokens_seen": 3271098368
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7573423206547907e-06,
      "loss": 1.2317,
      "theoretical_loss": 3.3273972553237363,
      "tokens_seen": 3271229440
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7493179264965495e-06,
      "loss": 1.2682,
      "theoretical_loss": 3.3273872702463283,
      "tokens_seen": 3271360512
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7412935323383086e-06,
      "loss": 1.271,
      "theoretical_loss": 3.327377285680993,
      "tokens_seen": 3271491584
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7332691381800676e-06,
      "loss": 1.2495,
      "theoretical_loss": 3.327367301627684,
      "tokens_seen": 3271622656
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7252447440218267e-06,
      "loss": 1.3003,
      "theoretical_loss": 3.3273573180863543,
      "tokens_seen": 3271753728
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7172203498635853e-06,
      "loss": 1.2076,
      "theoretical_loss": 3.327347335056957,
      "tokens_seen": 3271884800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7091959557053444e-06,
      "loss": 1.132,
      "theoretical_loss": 3.3273373525394456,
      "tokens_seen": 3272015872
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7011715615471032e-06,
      "loss": 1.2829,
      "theoretical_loss": 3.327327370533774,
      "tokens_seen": 3272146944
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6931471673888623e-06,
      "loss": 1.2364,
      "theoretical_loss": 3.3273173890398935,
      "tokens_seen": 3272278016
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6851227732306213e-06,
      "loss": 1.3454,
      "theoretical_loss": 3.3273074080577594,
      "tokens_seen": 3272409088
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.67709837907238e-06,
      "loss": 1.322,
      "theoretical_loss": 3.327297427587324,
      "tokens_seen": 3272540160
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.669073984914139e-06,
      "loss": 1.2839,
      "theoretical_loss": 3.3272874476285406,
      "tokens_seen": 3272671232
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6610495907558979e-06,
      "loss": 1.3134,
      "theoretical_loss": 3.3272774681813626,
      "tokens_seen": 3272802304
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.653025196597657e-06,
      "loss": 1.2691,
      "theoretical_loss": 3.3272674892457434,
      "tokens_seen": 3272933376
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.645000802439416e-06,
      "loss": 1.2067,
      "theoretical_loss": 3.3272575108216365,
      "tokens_seen": 3273064448
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.636976408281175e-06,
      "loss": 1.2344,
      "theoretical_loss": 3.3272475329089946,
      "tokens_seen": 3273195520
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6289520141229337e-06,
      "loss": 1.2793,
      "theoretical_loss": 3.327237555507771,
      "tokens_seen": 3273326592
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6209276199646925e-06,
      "loss": 1.2821,
      "theoretical_loss": 3.3272275786179195,
      "tokens_seen": 3273457664
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6129032258064516e-06,
      "loss": 1.3173,
      "theoretical_loss": 3.3272176022393936,
      "tokens_seen": 3273588736
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6048788316482107e-06,
      "loss": 1.2347,
      "theoretical_loss": 3.3272076263721457,
      "tokens_seen": 3273719808
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5968544374899697e-06,
      "loss": 1.3628,
      "theoretical_loss": 3.32719765101613,
      "tokens_seen": 3273850880
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5888300433317286e-06,
      "loss": 1.2998,
      "theoretical_loss": 3.3271876761712993,
      "tokens_seen": 3273981952
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5808056491734874e-06,
      "loss": 1.2394,
      "theoretical_loss": 3.3271777018376074,
      "tokens_seen": 3274113024
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5727812550152463e-06,
      "loss": 1.2021,
      "theoretical_loss": 3.3271677280150076,
      "tokens_seen": 3274244096
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5647568608570053e-06,
      "loss": 1.342,
      "theoretical_loss": 3.3271577547034523,
      "tokens_seen": 3274375168
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5567324666987644e-06,
      "loss": 1.2593,
      "theoretical_loss": 3.3271477819028963,
      "tokens_seen": 3274506240
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5487080725405234e-06,
      "loss": 1.2768,
      "theoretical_loss": 3.327137809613292,
      "tokens_seen": 3274637312
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5406836783822823e-06,
      "loss": 1.2157,
      "theoretical_loss": 3.3271278378345928,
      "tokens_seen": 3274768384
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5326592842240411e-06,
      "loss": 1.2891,
      "theoretical_loss": 3.327117866566753,
      "tokens_seen": 3274899456
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5246348900658002e-06,
      "loss": 1.3012,
      "theoretical_loss": 3.327107895809725,
      "tokens_seen": 3275030528
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.516610495907559e-06,
      "loss": 1.3404,
      "theoretical_loss": 3.327097925563462,
      "tokens_seen": 3275161600
    },
    {
      "epoch": 0.98,
      "objective/train/aligned_token_count": 0.9902477860450745,
      "objective/train/avg_token_score": 0.009752204641699791,
      "objective/train/docs_used": 1189805,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2391982078552246,
      "objective/train/original_loss": 1.2535673379898071,
      "objective/train/theoretical_loss": 3.3270879558279183,
      "objective/train/tokens_used": 1645751776,
      "objective/train/unlikelihood_loss": 0.026408148929476738,
      "theoretical_loss": 3.3270879558279183,
      "tokens_seen": 3275292672
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.508586101749318e-06,
      "loss": 1.2222,
      "theoretical_loss": 3.3270879558279183,
      "tokens_seen": 3275292672
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.500561707591077e-06,
      "loss": 1.2963,
      "theoretical_loss": 3.327077986603047,
      "tokens_seen": 3275423744
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4925373134328358e-06,
      "loss": 1.2634,
      "theoretical_loss": 3.3270680178888012,
      "tokens_seen": 3275554816
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4845129192745948e-06,
      "loss": 1.2514,
      "theoretical_loss": 3.3270580496851347,
      "tokens_seen": 3275685888
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4764885251163537e-06,
      "loss": 1.2723,
      "theoretical_loss": 3.3270480819920003,
      "tokens_seen": 3275816960
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4684641309581128e-06,
      "loss": 1.2925,
      "theoretical_loss": 3.327038114809352,
      "tokens_seen": 3275948032
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4604397367998718e-06,
      "loss": 1.3,
      "theoretical_loss": 3.3270281481371438,
      "tokens_seen": 3276079104
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4524153426416304e-06,
      "loss": 1.2413,
      "theoretical_loss": 3.327018181975328,
      "tokens_seen": 3276210176
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4443909484833895e-06,
      "loss": 1.3327,
      "theoretical_loss": 3.3270082163238586,
      "tokens_seen": 3276341248
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4363665543251486e-06,
      "loss": 1.1364,
      "theoretical_loss": 3.3269982511826885,
      "tokens_seen": 3276472320
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4283421601669074e-06,
      "loss": 1.2824,
      "theoretical_loss": 3.3269882865517717,
      "tokens_seen": 3276603392
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4203177660086665e-06,
      "loss": 1.2415,
      "theoretical_loss": 3.326978322431062,
      "tokens_seen": 3276734464
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4122933718504253e-06,
      "loss": 1.2078,
      "theoretical_loss": 3.326968358820512,
      "tokens_seen": 3276865536
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.4042689776921842e-06,
      "loss": 1.2526,
      "theoretical_loss": 3.3269583957200757,
      "tokens_seen": 3276996608
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3962445835339432e-06,
      "loss": 1.2457,
      "theoretical_loss": 3.3269484331297066,
      "tokens_seen": 3277127680
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3882201893757023e-06,
      "loss": 1.2894,
      "theoretical_loss": 3.326938471049358,
      "tokens_seen": 3277258752
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3801957952174611e-06,
      "loss": 1.2476,
      "theoretical_loss": 3.326928509478983,
      "tokens_seen": 3277389824
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3721714010592202e-06,
      "loss": 1.2582,
      "theoretical_loss": 3.3269185484185364,
      "tokens_seen": 3277520896
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.364147006900979e-06,
      "loss": 1.2013,
      "theoretical_loss": 3.3269085878679703,
      "tokens_seen": 3277651968
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3561226127427379e-06,
      "loss": 1.2507,
      "theoretical_loss": 3.326898627827239,
      "tokens_seen": 3277783040
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.348098218584497e-06,
      "loss": 1.2799,
      "theoretical_loss": 3.326888668296296,
      "tokens_seen": 3277914112
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.340073824426256e-06,
      "loss": 1.3088,
      "theoretical_loss": 3.326878709275094,
      "tokens_seen": 3278045184
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3320494302680148e-06,
      "loss": 1.2525,
      "theoretical_loss": 3.3268687507635875,
      "tokens_seen": 3278176256
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3240250361097737e-06,
      "loss": 1.2887,
      "theoretical_loss": 3.3268587927617297,
      "tokens_seen": 3278307328
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3160006419515328e-06,
      "loss": 1.2682,
      "theoretical_loss": 3.326848835269474,
      "tokens_seen": 3278438400
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3079762477932916e-06,
      "loss": 1.2731,
      "theoretical_loss": 3.3268388782867744,
      "tokens_seen": 3278569472
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2999518536350507e-06,
      "loss": 1.2711,
      "theoretical_loss": 3.326828921813584,
      "tokens_seen": 3278700544
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2919274594768097e-06,
      "loss": 1.2497,
      "theoretical_loss": 3.3268189658498564,
      "tokens_seen": 3278831616
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2839030653185684e-06,
      "loss": 1.2654,
      "theoretical_loss": 3.3268090103955457,
      "tokens_seen": 3278962688
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2758786711603274e-06,
      "loss": 1.2604,
      "theoretical_loss": 3.3267990554506044,
      "tokens_seen": 3279093760
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2678542770020865e-06,
      "loss": 1.3271,
      "theoretical_loss": 3.3267891010149873,
      "tokens_seen": 3279224832
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2598298828438453e-06,
      "loss": 1.3005,
      "theoretical_loss": 3.3267791470886467,
      "tokens_seen": 3279355904
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2518054886856044e-06,
      "loss": 1.2601,
      "theoretical_loss": 3.326769193671537,
      "tokens_seen": 3279486976
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2437810945273632e-06,
      "loss": 1.1954,
      "theoretical_loss": 3.326759240763612,
      "tokens_seen": 3279618048
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.235756700369122e-06,
      "loss": 1.1684,
      "theoretical_loss": 3.3267492883648253,
      "tokens_seen": 3279749120
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2277323062108811e-06,
      "loss": 1.3836,
      "theoretical_loss": 3.32673933647513,
      "tokens_seen": 3279880192
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2197079120526402e-06,
      "loss": 1.2467,
      "theoretical_loss": 3.3267293850944797,
      "tokens_seen": 3280011264
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.211683517894399e-06,
      "loss": 1.3514,
      "theoretical_loss": 3.326719434222828,
      "tokens_seen": 3280142336
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.2036591237361579e-06,
      "loss": 1.3138,
      "theoretical_loss": 3.326709483860129,
      "tokens_seen": 3280273408
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.195634729577917e-06,
      "loss": 1.1859,
      "theoretical_loss": 3.3266995340063366,
      "tokens_seen": 3280404480
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1876103354196758e-06,
      "loss": 1.3376,
      "theoretical_loss": 3.3266895846614033,
      "tokens_seen": 3280535552
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1795859412614348e-06,
      "loss": 1.2334,
      "theoretical_loss": 3.3266796358252835,
      "tokens_seen": 3280666624
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.171561547103194e-06,
      "loss": 1.2703,
      "theoretical_loss": 3.3266696874979313,
      "tokens_seen": 3280797696
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1635371529449528e-06,
      "loss": 1.1957,
      "theoretical_loss": 3.326659739679299,
      "tokens_seen": 3280928768
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1555127587867116e-06,
      "loss": 1.3003,
      "theoretical_loss": 3.3266497923693414,
      "tokens_seen": 3281059840
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1474883646284707e-06,
      "loss": 1.3102,
      "theoretical_loss": 3.326639845568012,
      "tokens_seen": 3281190912
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1394639704702295e-06,
      "loss": 1.2054,
      "theoretical_loss": 3.326629899275264,
      "tokens_seen": 3281321984
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1314395763119886e-06,
      "loss": 1.2803,
      "theoretical_loss": 3.326619953491052,
      "tokens_seen": 3281453056
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1234151821537476e-06,
      "loss": 1.2464,
      "theoretical_loss": 3.3266100082153285,
      "tokens_seen": 3281584128
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1153907879955063e-06,
      "loss": 1.27,
      "theoretical_loss": 3.3266000634480477,
      "tokens_seen": 3281715200
    },
    {
      "epoch": 0.99,
      "objective/train/aligned_token_count": 0.9948239922523499,
      "objective/train/avg_token_score": 0.005176020320504904,
      "objective/train/docs_used": 1192226,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2793099880218506,
      "objective/train/original_loss": 1.2865655422210693,
      "objective/train/theoretical_loss": 3.326590119189164,
      "objective/train/tokens_used": 1652305376,
      "objective/train/unlikelihood_loss": 0.014384533278644085,
      "theoretical_loss": 3.326590119189164,
      "tokens_seen": 3281846272
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.1073663938372653e-06,
      "loss": 1.2561,
      "theoretical_loss": 3.326590119189164,
      "tokens_seen": 3281846272
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0993419996790244e-06,
      "loss": 1.3463,
      "theoretical_loss": 3.3265801754386297,
      "tokens_seen": 3281977344
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0913176055207832e-06,
      "loss": 1.2448,
      "theoretical_loss": 3.3265702321963997,
      "tokens_seen": 3282108416
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0832932113625423e-06,
      "loss": 1.245,
      "theoretical_loss": 3.326560289462427,
      "tokens_seen": 3282239488
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0752688172043011e-06,
      "loss": 1.2622,
      "theoretical_loss": 3.326550347236666,
      "tokens_seen": 3282370560
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.06724442304606e-06,
      "loss": 1.2855,
      "theoretical_loss": 3.3265404055190704,
      "tokens_seen": 3282501632
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.059220028887819e-06,
      "loss": 1.2167,
      "theoretical_loss": 3.326530464309593,
      "tokens_seen": 3282632704
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0511956347295779e-06,
      "loss": 1.3103,
      "theoretical_loss": 3.3265205236081883,
      "tokens_seen": 3282763776
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.043171240571337e-06,
      "loss": 1.2833,
      "theoretical_loss": 3.3265105834148097,
      "tokens_seen": 3282894848
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0351468464130958e-06,
      "loss": 1.2808,
      "theoretical_loss": 3.3265006437294113,
      "tokens_seen": 3283025920
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0271224522548546e-06,
      "loss": 1.267,
      "theoretical_loss": 3.326490704551947,
      "tokens_seen": 3283156992
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0190980580966137e-06,
      "loss": 1.2935,
      "theoretical_loss": 3.32648076588237,
      "tokens_seen": 3283288064
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0110736639383728e-06,
      "loss": 1.2474,
      "theoretical_loss": 3.3264708277206347,
      "tokens_seen": 3283419136
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0030492697801316e-06,
      "loss": 1.2966,
      "theoretical_loss": 3.326460890066694,
      "tokens_seen": 3283550208
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.950248756218907e-07,
      "loss": 1.2243,
      "theoretical_loss": 3.3264509529205024,
      "tokens_seen": 3283681280
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.870004814636495e-07,
      "loss": 1.2607,
      "theoretical_loss": 3.326441016282014,
      "tokens_seen": 3283812352
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.789760873054084e-07,
      "loss": 1.2574,
      "theoretical_loss": 3.3264310801511816,
      "tokens_seen": 3283943424
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.709516931471674e-07,
      "loss": 1.1538,
      "theoretical_loss": 3.3264211445279592,
      "tokens_seen": 3284074496
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.629272989889265e-07,
      "loss": 1.1251,
      "theoretical_loss": 3.3264112094123015,
      "tokens_seen": 3284205568
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.549029048306853e-07,
      "loss": 1.2055,
      "theoretical_loss": 3.326401274804162,
      "tokens_seen": 3284336640
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.468785106724443e-07,
      "loss": 1.2103,
      "theoretical_loss": 3.3263913407034935,
      "tokens_seen": 3284467712
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.388541165142032e-07,
      "loss": 1.278,
      "theoretical_loss": 3.326381407110251,
      "tokens_seen": 3284598784
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.308297223559621e-07,
      "loss": 1.2983,
      "theoretical_loss": 3.3263714740243877,
      "tokens_seen": 3284729856
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.228053281977211e-07,
      "loss": 1.2714,
      "theoretical_loss": 3.326361541445858,
      "tokens_seen": 3284860928
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.147809340394801e-07,
      "loss": 1.1438,
      "theoretical_loss": 3.3263516093746155,
      "tokens_seen": 3284992000
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.067565398812389e-07,
      "loss": 1.3088,
      "theoretical_loss": 3.3263416778106136,
      "tokens_seen": 3285123072
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.98732145722998e-07,
      "loss": 1.2231,
      "theoretical_loss": 3.3263317467538065,
      "tokens_seen": 3285254144
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.90707751564757e-07,
      "loss": 1.2725,
      "theoretical_loss": 3.3263218162041484,
      "tokens_seen": 3285385216
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.826833574065158e-07,
      "loss": 1.2634,
      "theoretical_loss": 3.3263118861615926,
      "tokens_seen": 3285516288
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.746589632482747e-07,
      "loss": 1.3322,
      "theoretical_loss": 3.326301956626093,
      "tokens_seen": 3285647360
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.666345690900338e-07,
      "loss": 1.2759,
      "theoretical_loss": 3.326292027597604,
      "tokens_seen": 3285778432
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.586101749317927e-07,
      "loss": 1.2803,
      "theoretical_loss": 3.3262820990760793,
      "tokens_seen": 3285909504
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.505857807735516e-07,
      "loss": 1.2141,
      "theoretical_loss": 3.3262721710614724,
      "tokens_seen": 3286040576
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.425613866153107e-07,
      "loss": 1.2168,
      "theoretical_loss": 3.326262243553738,
      "tokens_seen": 3286171648
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.345369924570695e-07,
      "loss": 1.2932,
      "theoretical_loss": 3.326252316552829,
      "tokens_seen": 3286302720
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.265125982988285e-07,
      "loss": 1.2509,
      "theoretical_loss": 3.3262423900586997,
      "tokens_seen": 3286433792
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.184882041405875e-07,
      "loss": 1.2396,
      "theoretical_loss": 3.3262324640713046,
      "tokens_seen": 3286564864
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.104638099823463e-07,
      "loss": 1.2557,
      "theoretical_loss": 3.326222538590597,
      "tokens_seen": 3286695936
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.024394158241053e-07,
      "loss": 1.2309,
      "theoretical_loss": 3.326212613616531,
      "tokens_seen": 3286827008
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.944150216658643e-07,
      "loss": 1.3355,
      "theoretical_loss": 3.3262026891490604,
      "tokens_seen": 3286958080
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.863906275076231e-07,
      "loss": 1.2483,
      "theoretical_loss": 3.3261927651881393,
      "tokens_seen": 3287089152
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.783662333493822e-07,
      "loss": 1.2747,
      "theoretical_loss": 3.3261828417337216,
      "tokens_seen": 3287220224
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.703418391911411e-07,
      "loss": 1.2156,
      "theoretical_loss": 3.3261729187857614,
      "tokens_seen": 3287351296
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.623174450329001e-07,
      "loss": 1.1952,
      "theoretical_loss": 3.3261629963442125,
      "tokens_seen": 3287482368
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.54293050874659e-07,
      "loss": 1.2469,
      "theoretical_loss": 3.326153074409029,
      "tokens_seen": 3287613440
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.462686567164179e-07,
      "loss": 1.2098,
      "theoretical_loss": 3.326143152980164,
      "tokens_seen": 3287744512
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.382442625581768e-07,
      "loss": 1.2222,
      "theoretical_loss": 3.326133232057573,
      "tokens_seen": 3287875584
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.302198683999359e-07,
      "loss": 1.1974,
      "theoretical_loss": 3.3261233116412092,
      "tokens_seen": 3288006656
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.221954742416948e-07,
      "loss": 1.2642,
      "theoretical_loss": 3.3261133917310266,
      "tokens_seen": 3288137728
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.141710800834537e-07,
      "loss": 1.1903,
      "theoretical_loss": 3.326103472326979,
      "tokens_seen": 3288268800
    },
    {
      "epoch": 0.99,
      "objective/train/aligned_token_count": 0.9924721121788025,
      "objective/train/avg_token_score": 0.007527893874794245,
      "objective/train/docs_used": 1194491,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.2374454736709595,
      "objective/train/original_loss": 1.2471263408660889,
      "objective/train/theoretical_loss": 3.3260935534290206,
      "objective/train/tokens_used": 1658858976,
      "objective/train/unlikelihood_loss": 0.021435504779219627,
      "theoretical_loss": 3.3260935534290206,
      "tokens_seen": 3288399872
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.061466859252127e-07,
      "loss": 1.2245,
      "theoretical_loss": 3.3260935534290206,
      "tokens_seen": 3288399872
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.981222917669716e-07,
      "loss": 1.2188,
      "theoretical_loss": 3.3260836350371057,
      "tokens_seen": 3288530944
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.900978976087306e-07,
      "loss": 1.3296,
      "theoretical_loss": 3.326073717151188,
      "tokens_seen": 3288662016
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.820735034504895e-07,
      "loss": 1.3692,
      "theoretical_loss": 3.3260637997712212,
      "tokens_seen": 3288793088
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.740491092922485e-07,
      "loss": 1.2695,
      "theoretical_loss": 3.32605388289716,
      "tokens_seen": 3288924160
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.660247151340074e-07,
      "loss": 1.2484,
      "theoretical_loss": 3.326043966528958,
      "tokens_seen": 3289055232
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.580003209757664e-07,
      "loss": 1.3086,
      "theoretical_loss": 3.3260340506665695,
      "tokens_seen": 3289186304
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.499759268175253e-07,
      "loss": 1.2086,
      "theoretical_loss": 3.3260241353099484,
      "tokens_seen": 3289317376
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.419515326592842e-07,
      "loss": 1.2418,
      "theoretical_loss": 3.3260142204590486,
      "tokens_seen": 3289448448
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.339271385010432e-07,
      "loss": 1.2244,
      "theoretical_loss": 3.3260043061138247,
      "tokens_seen": 3289579520
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.259027443428022e-07,
      "loss": 1.302,
      "theoretical_loss": 3.32599439227423,
      "tokens_seen": 3289710592
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.17878350184561e-07,
      "loss": 1.2042,
      "theoretical_loss": 3.325984478940219,
      "tokens_seen": 3289841664
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.098539560263201e-07,
      "loss": 1.2709,
      "theoretical_loss": 3.325974566111746,
      "tokens_seen": 3289972736
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.018295618680789e-07,
      "loss": 1.2391,
      "theoretical_loss": 3.3259646537887644,
      "tokens_seen": 3290103808
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.938051677098379e-07,
      "loss": 1.2842,
      "theoretical_loss": 3.3259547419712288,
      "tokens_seen": 3290234880
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.85780773551597e-07,
      "loss": 1.2267,
      "theoretical_loss": 3.3259448306590933,
      "tokens_seen": 3290365952
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.777563793933558e-07,
      "loss": 1.2596,
      "theoretical_loss": 3.325934919852312,
      "tokens_seen": 3290497024
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.697319852351148e-07,
      "loss": 1.26,
      "theoretical_loss": 3.325925009550838,
      "tokens_seen": 3290628096
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.617075910768738e-07,
      "loss": 1.2922,
      "theoretical_loss": 3.3259150997546274,
      "tokens_seen": 3290759168
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.536831969186327e-07,
      "loss": 1.2322,
      "theoretical_loss": 3.3259051904636325,
      "tokens_seen": 3290890240
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.456588027603916e-07,
      "loss": 1.2446,
      "theoretical_loss": 3.3258952816778087,
      "tokens_seen": 3291021312
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.376344086021506e-07,
      "loss": 1.2345,
      "theoretical_loss": 3.325885373397109,
      "tokens_seen": 3291152384
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.296100144439095e-07,
      "loss": 1.2503,
      "theoretical_loss": 3.325875465621489,
      "tokens_seen": 3291283456
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.215856202856685e-07,
      "loss": 1.2796,
      "theoretical_loss": 3.325865558350901,
      "tokens_seen": 3291414528
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.135612261274273e-07,
      "loss": 1.2455,
      "theoretical_loss": 3.3258556515853,
      "tokens_seen": 3291545600
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.055368319691864e-07,
      "loss": 1.2419,
      "theoretical_loss": 3.325845745324641,
      "tokens_seen": 3291676672
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.975124378109453e-07,
      "loss": 1.2282,
      "theoretical_loss": 3.325835839568877,
      "tokens_seen": 3291807744
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.894880436527042e-07,
      "loss": 1.189,
      "theoretical_loss": 3.3258259343179626,
      "tokens_seen": 3291938816
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.814636494944632e-07,
      "loss": 1.2408,
      "theoretical_loss": 3.3258160295718517,
      "tokens_seen": 3292069888
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.7343925533622214e-07,
      "loss": 1.2358,
      "theoretical_loss": 3.325806125330499,
      "tokens_seen": 3292200960
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.6541486117798104e-07,
      "loss": 1.2983,
      "theoretical_loss": 3.3257962215938583,
      "tokens_seen": 3292332032
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.5739046701974004e-07,
      "loss": 1.2701,
      "theoretical_loss": 3.325786318361884,
      "tokens_seen": 3292463104
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.49366072861499e-07,
      "loss": 1.3388,
      "theoretical_loss": 3.32577641563453,
      "tokens_seen": 3292594176
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.413416787032579e-07,
      "loss": 1.2273,
      "theoretical_loss": 3.3257665134117507,
      "tokens_seen": 3292725248
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.333172845450169e-07,
      "loss": 1.2454,
      "theoretical_loss": 3.3257566116935005,
      "tokens_seen": 3292856320
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.252928903867758e-07,
      "loss": 1.2407,
      "theoretical_loss": 3.325746710479733,
      "tokens_seen": 3292987392
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.1726849622853476e-07,
      "loss": 1.1294,
      "theoretical_loss": 3.325736809770403,
      "tokens_seen": 3293118464
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.0924410207029376e-07,
      "loss": 1.269,
      "theoretical_loss": 3.3257269095654647,
      "tokens_seen": 3293249536
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.0121970791205266e-07,
      "loss": 1.1932,
      "theoretical_loss": 3.325717009864872,
      "tokens_seen": 3293380608
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.9319531375381156e-07,
      "loss": 1.2591,
      "theoretical_loss": 3.325707110668579,
      "tokens_seen": 3293511680
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.8517091959557057e-07,
      "loss": 1.2291,
      "theoretical_loss": 3.325697211976541,
      "tokens_seen": 3293642752
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.771465254373295e-07,
      "loss": 1.3318,
      "theoretical_loss": 3.325687313788711,
      "tokens_seen": 3293773824
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.691221312790884e-07,
      "loss": 1.2638,
      "theoretical_loss": 3.325677416105044,
      "tokens_seen": 3293904896
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.610977371208474e-07,
      "loss": 1.1843,
      "theoretical_loss": 3.325667518925494,
      "tokens_seen": 3294035968
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.5307334296260633e-07,
      "loss": 1.2528,
      "theoretical_loss": 3.3256576222500147,
      "tokens_seen": 3294167040
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.450489488043653e-07,
      "loss": 1.2726,
      "theoretical_loss": 3.3256477260785613,
      "tokens_seen": 3294298112
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.3702455464612424e-07,
      "loss": 1.3067,
      "theoretical_loss": 3.325637830411088,
      "tokens_seen": 3294429184
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.290001604878832e-07,
      "loss": 1.2066,
      "theoretical_loss": 3.3256279352475486,
      "tokens_seen": 3294560256
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.209757663296421e-07,
      "loss": 1.1991,
      "theoretical_loss": 3.3256180405878975,
      "tokens_seen": 3294691328
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.129513721714011e-07,
      "loss": 1.2552,
      "theoretical_loss": 3.325608146432089,
      "tokens_seen": 3294822400
    },
    {
      "epoch": 1.0,
      "objective/train/aligned_token_count": 0.9926229119300842,
      "objective/train/avg_token_score": 0.007377032656222582,
      "objective/train/docs_used": 1196765,
      "objective/train/instantaneous_batch_size": 64,
      "objective/train/instantaneous_microbatch_size": 65536,
      "objective/train/likelihood_loss": 1.1997034549713135,
      "objective/train/original_loss": 1.2074978351593018,
      "objective/train/theoretical_loss": 3.325598252780078,
      "objective/train/tokens_used": 1665412576,
      "objective/train/unlikelihood_loss": 0.025290917605161667,
      "theoretical_loss": 3.325598252780078,
      "tokens_seen": 3294953472
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.0492697801316005e-07,
      "loss": 1.2905,
      "theoretical_loss": 3.325598252780078,
      "tokens_seen": 3294953472
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.9690258385491895e-07,
      "loss": 1.2954,
      "theoretical_loss": 3.325588359631818,
      "tokens_seen": 3295084544
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.888781896966779e-07,
      "loss": 1.1934,
      "theoretical_loss": 3.3255784669872632,
      "tokens_seen": 3295215616
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.808537955384369e-07,
      "loss": 1.2487,
      "theoretical_loss": 3.325568574846369,
      "tokens_seen": 3295346688
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.728294013801958e-07,
      "loss": 1.2548,
      "theoretical_loss": 3.3255586832090884,
      "tokens_seen": 3295477760
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.6480500722195476e-07,
      "loss": 1.243,
      "theoretical_loss": 3.3255487920753772,
      "tokens_seen": 3295608832
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5678061306371366e-07,
      "loss": 1.2173,
      "theoretical_loss": 3.325538901445188,
      "tokens_seen": 3295739904
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4875621890547267e-07,
      "loss": 1.233,
      "theoretical_loss": 3.325529011318477,
      "tokens_seen": 3295870976
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.407318247472316e-07,
      "loss": 1.279,
      "theoretical_loss": 3.325519121695197,
      "tokens_seen": 3296002048
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3270743058899052e-07,
      "loss": 1.2811,
      "theoretical_loss": 3.3255092325753033,
      "tokens_seen": 3296133120
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.246830364307495e-07,
      "loss": 1.2707,
      "theoretical_loss": 3.32549934395875,
      "tokens_seen": 3296264192
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1665864227250845e-07,
      "loss": 1.2729,
      "theoretical_loss": 3.3254894558454913,
      "tokens_seen": 3296395264
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0863424811426738e-07,
      "loss": 1.305,
      "theoretical_loss": 3.3254795682354814,
      "tokens_seen": 3296526336
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0060985395602633e-07,
      "loss": 1.2687,
      "theoretical_loss": 3.3254696811286752,
      "tokens_seen": 3296657408
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.9258545979778528e-07,
      "loss": 1.2899,
      "theoretical_loss": 3.325459794525027,
      "tokens_seen": 3296788480
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.845610656395442e-07,
      "loss": 1.3059,
      "theoretical_loss": 3.325449908424491,
      "tokens_seen": 3296919552
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7653667148130316e-07,
      "loss": 1.2418,
      "theoretical_loss": 3.3254400228270216,
      "tokens_seen": 3297050624
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6851227732306212e-07,
      "loss": 1.2426,
      "theoretical_loss": 3.3254301377325732,
      "tokens_seen": 3297181696
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6048788316482104e-07,
      "loss": 1.2558,
      "theoretical_loss": 3.3254202531411003,
      "tokens_seen": 3297312768
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5246348900658002e-07,
      "loss": 1.2691,
      "theoretical_loss": 3.325410369052557,
      "tokens_seen": 3297443840
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4443909484833895e-07,
      "loss": 1.2384,
      "theoretical_loss": 3.3254004854668984,
      "tokens_seen": 3297574912
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.364147006900979e-07,
      "loss": 1.2175,
      "theoretical_loss": 3.3253906023840782,
      "tokens_seen": 3297705984
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2839030653185683e-07,
      "loss": 1.2067,
      "theoretical_loss": 3.3253807198040515,
      "tokens_seen": 3297837056
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.203659123736158e-07,
      "loss": 1.2628,
      "theoretical_loss": 3.325370837726772,
      "tokens_seen": 3297968128
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1234151821537475e-07,
      "loss": 1.2583,
      "theoretical_loss": 3.3253609561521946,
      "tokens_seen": 3298099200
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0431712405713369e-07,
      "loss": 1.2826,
      "theoretical_loss": 3.325351075080274,
      "tokens_seen": 3298230272
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.629272989889264e-08,
      "loss": 1.1962,
      "theoretical_loss": 3.325341194510964,
      "tokens_seen": 3298361344
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.826833574065158e-08,
      "loss": 1.2356,
      "theoretical_loss": 3.3253313144442194,
      "tokens_seen": 3298492416
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.024394158241052e-08,
      "loss": 1.3311,
      "theoretical_loss": 3.325321434879995,
      "tokens_seen": 3298623488
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.221954742416948e-08,
      "loss": 1.263,
      "theoretical_loss": 3.325311555818245,
      "tokens_seen": 3298754560
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.419515326592842e-08,
      "loss": 1.2581,
      "theoretical_loss": 3.3253016772589232,
      "tokens_seen": 3298885632
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.6170759107687375e-08,
      "loss": 1.2318,
      "theoretical_loss": 3.325291799201985,
      "tokens_seen": 3299016704
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.814636494944632e-08,
      "loss": 1.282,
      "theoretical_loss": 3.3252819216473846,
      "tokens_seen": 3299147776
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.012197079120526e-08,
      "loss": 1.2067,
      "theoretical_loss": 3.3252720445950765,
      "tokens_seen": 3299278848
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.209757663296421e-08,
      "loss": 1.304,
      "theoretical_loss": 3.3252621680450156,
      "tokens_seen": 3299409920
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.407318247472316e-08,
      "loss": 1.2616,
      "theoretical_loss": 3.3252522919971557,
      "tokens_seen": 3299540992
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6048788316482104e-08,
      "loss": 1.2781,
      "theoretical_loss": 3.3252424164514514,
      "tokens_seen": 3299672064
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.024394158241052e-09,
      "loss": 1.251,
      "theoretical_loss": 3.325232541407858,
      "tokens_seen": 3299803136
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0,
      "loss": 1.2211,
      "theoretical_loss": 3.325222666866329,
      "tokens_seen": 3299934208
    }
  ],
  "max_steps": 12588,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 8.42020889296896e+17,
  "trial_name": null,
  "trial_params": null
}