{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 100,
  "global_step": 7120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.121875,
      "epoch": 0.0028089887640449437,
      "grad_norm": 832.14208984375,
      "learning_rate": 1.404494382022472e-07,
      "loss": 10.299293518066406,
      "mean_token_accuracy": 0.3464117765426636,
      "num_tokens": 130814.0,
      "step": 10
    },
    {
      "entropy": 1.12734375,
      "epoch": 0.0056179775280898875,
      "grad_norm": 836.1932373046875,
      "learning_rate": 2.808988764044944e-07,
      "loss": 9.685482788085938,
      "mean_token_accuracy": 0.3462586671113968,
      "num_tokens": 262217.0,
      "step": 20
    },
    {
      "entropy": 1.13671875,
      "epoch": 0.008426966292134831,
      "grad_norm": 682.9151611328125,
      "learning_rate": 4.213483146067416e-07,
      "loss": 6.743853759765625,
      "mean_token_accuracy": 0.34578883945941924,
      "num_tokens": 393111.0,
      "step": 30
    },
    {
      "entropy": 1.14765625,
      "epoch": 0.011235955056179775,
      "grad_norm": 60.9609375,
      "learning_rate": 5.617977528089888e-07,
      "loss": 3.318247604370117,
      "mean_token_accuracy": 0.5958202004432678,
      "num_tokens": 525553.0,
      "step": 40
    },
    {
      "entropy": 1.25,
      "epoch": 0.014044943820224719,
      "grad_norm": 77.60545349121094,
      "learning_rate": 7.022471910112361e-07,
      "loss": 2.5458627700805665,
      "mean_token_accuracy": 0.6743934571743011,
      "num_tokens": 655464.0,
      "step": 50
    },
    {
      "entropy": 1.46171875,
      "epoch": 0.016853932584269662,
      "grad_norm": 13.557156562805176,
      "learning_rate": 8.426966292134832e-07,
      "loss": 1.823989486694336,
      "mean_token_accuracy": 0.6986484050750732,
      "num_tokens": 786189.0,
      "step": 60
    },
    {
      "entropy": 1.8078125,
      "epoch": 0.019662921348314606,
      "grad_norm": 9.053010940551758,
      "learning_rate": 9.831460674157304e-07,
      "loss": 1.3517951011657714,
      "mean_token_accuracy": 0.7538730502128601,
      "num_tokens": 918090.0,
      "step": 70
    },
    {
      "entropy": 2.278125,
      "epoch": 0.02247191011235955,
      "grad_norm": 7.523867607116699,
      "learning_rate": 1.1235955056179777e-06,
      "loss": 1.1615694999694823,
      "mean_token_accuracy": 0.7693759739398957,
      "num_tokens": 1048743.0,
      "step": 80
    },
    {
      "entropy": 2.5359375,
      "epoch": 0.025280898876404494,
      "grad_norm": 8.867629051208496,
      "learning_rate": 1.264044943820225e-06,
      "loss": 1.0422688484191895,
      "mean_token_accuracy": 0.787922203540802,
      "num_tokens": 1180341.0,
      "step": 90
    },
    {
      "entropy": 2.565625,
      "epoch": 0.028089887640449437,
      "grad_norm": 4.355800151824951,
      "learning_rate": 1.4044943820224722e-06,
      "loss": 0.9384889602661133,
      "mean_token_accuracy": 0.8007173240184784,
      "num_tokens": 1311899.0,
      "step": 100
    },
    {
      "epoch": 0.028089887640449437,
      "eval_entropy": 2.6114879893238436,
      "eval_loss": 0.9133086204528809,
      "eval_mean_token_accuracy": 0.8024802939747576,
      "eval_num_tokens": 1311899.0,
      "eval_runtime": 131.5102,
      "eval_samples_per_second": 119.626,
      "eval_steps_per_second": 2.137,
      "step": 100
    },
    {
      "entropy": 2.5546875,
      "epoch": 0.03089887640449438,
      "grad_norm": 5.836198806762695,
      "learning_rate": 1.544943820224719e-06,
      "loss": 0.9030431747436524,
      "mean_token_accuracy": 0.80395547747612,
      "num_tokens": 1442562.0,
      "step": 110
    },
    {
      "entropy": 2.49375,
      "epoch": 0.033707865168539325,
      "grad_norm": 4.069826602935791,
      "learning_rate": 1.6853932584269663e-06,
      "loss": 0.8461490631103515,
      "mean_token_accuracy": 0.8122228145599365,
      "num_tokens": 1573944.0,
      "step": 120
    },
    {
      "entropy": 2.40625,
      "epoch": 0.03651685393258427,
      "grad_norm": 3.7947349548339844,
      "learning_rate": 1.8258426966292136e-06,
      "loss": 0.814113712310791,
      "mean_token_accuracy": 0.8153736054897308,
      "num_tokens": 1704891.0,
      "step": 130
    },
    {
      "entropy": 2.3703125,
      "epoch": 0.03932584269662921,
      "grad_norm": 4.012843132019043,
      "learning_rate": 1.966292134831461e-06,
      "loss": 0.802308464050293,
      "mean_token_accuracy": 0.813105171918869,
      "num_tokens": 1835383.0,
      "step": 140
    },
    {
      "entropy": 2.334375,
      "epoch": 0.042134831460674156,
      "grad_norm": 3.572143793106079,
      "learning_rate": 2.106741573033708e-06,
      "loss": 0.793598985671997,
      "mean_token_accuracy": 0.8176955580711365,
      "num_tokens": 1966763.0,
      "step": 150
    },
    {
      "entropy": 2.353125,
      "epoch": 0.0449438202247191,
      "grad_norm": 4.31054162979126,
      "learning_rate": 2.2471910112359554e-06,
      "loss": 0.7583021640777587,
      "mean_token_accuracy": 0.8313632488250733,
      "num_tokens": 2097971.0,
      "step": 160
    },
    {
      "entropy": 2.33125,
      "epoch": 0.047752808988764044,
      "grad_norm": 4.077391624450684,
      "learning_rate": 2.3876404494382026e-06,
      "loss": 0.7243642807006836,
      "mean_token_accuracy": 0.8355784952640534,
      "num_tokens": 2228856.0,
      "step": 170
    },
    {
      "entropy": 2.3625,
      "epoch": 0.05056179775280899,
      "grad_norm": 13.097944259643555,
      "learning_rate": 2.52808988764045e-06,
      "loss": 0.7547924041748046,
      "mean_token_accuracy": 0.8292012214660645,
      "num_tokens": 2360308.0,
      "step": 180
    },
    {
      "entropy": 2.334375,
      "epoch": 0.05337078651685393,
      "grad_norm": 3.662111282348633,
      "learning_rate": 2.6685393258426968e-06,
      "loss": 0.715250301361084,
      "mean_token_accuracy": 0.8342713713645935,
      "num_tokens": 2491720.0,
      "step": 190
    },
    {
      "entropy": 2.3203125,
      "epoch": 0.056179775280898875,
      "grad_norm": 3.3750202655792236,
      "learning_rate": 2.8089887640449444e-06,
      "loss": 0.7078051567077637,
      "mean_token_accuracy": 0.8342422306537628,
      "num_tokens": 2622853.0,
      "step": 200
    },
    {
      "epoch": 0.056179775280898875,
      "eval_entropy": 2.380449288256228,
      "eval_loss": 0.6887717843055725,
      "eval_mean_token_accuracy": 0.8367773835345095,
      "eval_num_tokens": 2622853.0,
      "eval_runtime": 127.9766,
      "eval_samples_per_second": 122.929,
      "eval_steps_per_second": 2.196,
      "step": 200
    },
    {
      "entropy": 2.3578125,
      "epoch": 0.05898876404494382,
      "grad_norm": 3.572990894317627,
      "learning_rate": 2.9494382022471913e-06,
      "loss": 0.6780576705932617,
      "mean_token_accuracy": 0.8398098230361939,
      "num_tokens": 2754035.0,
      "step": 210
    },
    {
      "entropy": 2.3765625,
      "epoch": 0.06179775280898876,
      "grad_norm": 3.3943042755126953,
      "learning_rate": 3.089887640449438e-06,
      "loss": 0.6490035057067871,
      "mean_token_accuracy": 0.8413715302944184,
      "num_tokens": 2884759.0,
      "step": 220
    },
    {
      "entropy": 2.3421875,
      "epoch": 0.06460674157303371,
      "grad_norm": 2.9690749645233154,
      "learning_rate": 3.230337078651686e-06,
      "loss": 0.676630687713623,
      "mean_token_accuracy": 0.8361791253089905,
      "num_tokens": 3017347.0,
      "step": 230
    },
    {
      "entropy": 2.2515625,
      "epoch": 0.06741573033707865,
      "grad_norm": 10.60129451751709,
      "learning_rate": 3.3707865168539327e-06,
      "loss": 0.6570458889007569,
      "mean_token_accuracy": 0.835248875617981,
      "num_tokens": 3148390.0,
      "step": 240
    },
    {
      "entropy": 2.190625,
      "epoch": 0.0702247191011236,
      "grad_norm": 2.6151316165924072,
      "learning_rate": 3.5112359550561803e-06,
      "loss": 0.6361165046691895,
      "mean_token_accuracy": 0.8389629781246185,
      "num_tokens": 3278719.0,
      "step": 250
    },
    {
      "entropy": 2.178125,
      "epoch": 0.07303370786516854,
      "grad_norm": 2.3367278575897217,
      "learning_rate": 3.651685393258427e-06,
      "loss": 0.6137415409088135,
      "mean_token_accuracy": 0.8461585581302643,
      "num_tokens": 3409983.0,
      "step": 260
    },
    {
      "entropy": 2.1890625,
      "epoch": 0.07584269662921349,
      "grad_norm": 2.472215414047241,
      "learning_rate": 3.7921348314606744e-06,
      "loss": 0.6319524765014648,
      "mean_token_accuracy": 0.8398538887500763,
      "num_tokens": 3541233.0,
      "step": 270
    },
    {
      "entropy": 2.2296875,
      "epoch": 0.07865168539325842,
      "grad_norm": 2.341369867324829,
      "learning_rate": 3.932584269662922e-06,
      "loss": 0.6252285957336425,
      "mean_token_accuracy": 0.8402074754238129,
      "num_tokens": 3671977.0,
      "step": 280
    },
    {
      "entropy": 2.1953125,
      "epoch": 0.08146067415730338,
      "grad_norm": 2.574246406555176,
      "learning_rate": 4.073033707865169e-06,
      "loss": 0.612055492401123,
      "mean_token_accuracy": 0.8448068499565125,
      "num_tokens": 3803486.0,
      "step": 290
    },
    {
      "entropy": 2.159375,
      "epoch": 0.08426966292134831,
      "grad_norm": 1.9848328828811646,
      "learning_rate": 4.213483146067416e-06,
      "loss": 0.6154293537139892,
      "mean_token_accuracy": 0.8447769582271576,
      "num_tokens": 3934832.0,
      "step": 300
    },
    {
      "epoch": 0.08426966292134831,
      "eval_entropy": 2.0921374555160144,
      "eval_loss": 0.6118077039718628,
      "eval_mean_token_accuracy": 0.8438847743743679,
      "eval_num_tokens": 3934832.0,
      "eval_runtime": 127.6727,
      "eval_samples_per_second": 123.221,
      "eval_steps_per_second": 2.201,
      "step": 300
    },
    {
      "entropy": 2.0390625,
      "epoch": 0.08707865168539326,
      "grad_norm": 2.847730875015259,
      "learning_rate": 4.3539325842696635e-06,
      "loss": 0.617510461807251,
      "mean_token_accuracy": 0.8411804437637329,
      "num_tokens": 4066274.0,
      "step": 310
    },
    {
      "entropy": 2.096875,
      "epoch": 0.0898876404494382,
      "grad_norm": 2.367683172225952,
      "learning_rate": 4.494382022471911e-06,
      "loss": 0.5956849098205567,
      "mean_token_accuracy": 0.8486601352691651,
      "num_tokens": 4197501.0,
      "step": 320
    },
    {
      "entropy": 2.159375,
      "epoch": 0.09269662921348315,
      "grad_norm": 2.2467598915100098,
      "learning_rate": 4.634831460674158e-06,
      "loss": 0.6177192211151123,
      "mean_token_accuracy": 0.8402603328227997,
      "num_tokens": 4329231.0,
      "step": 330
    },
    {
      "entropy": 2.11875,
      "epoch": 0.09550561797752809,
      "grad_norm": 2.5109703540802,
      "learning_rate": 4.775280898876405e-06,
      "loss": 0.5997633934020996,
      "mean_token_accuracy": 0.8457850992679596,
      "num_tokens": 4460229.0,
      "step": 340
    },
    {
      "entropy": 2.0390625,
      "epoch": 0.09831460674157304,
      "grad_norm": 2.4322571754455566,
      "learning_rate": 4.915730337078652e-06,
      "loss": 0.608931016921997,
      "mean_token_accuracy": 0.8431453108787537,
      "num_tokens": 4591470.0,
      "step": 350
    },
    {
      "entropy": 1.85,
      "epoch": 0.10112359550561797,
      "grad_norm": 3.578296422958374,
      "learning_rate": 5.0561797752809e-06,
      "loss": 0.5973134994506836,
      "mean_token_accuracy": 0.8443714499473571,
      "num_tokens": 4723264.0,
      "step": 360
    },
    {
      "entropy": 1.91953125,
      "epoch": 0.10393258426966293,
      "grad_norm": 2.3870301246643066,
      "learning_rate": 5.196629213483146e-06,
      "loss": 0.5749616622924805,
      "mean_token_accuracy": 0.8504410922527313,
      "num_tokens": 4853740.0,
      "step": 370
    },
    {
      "entropy": 2.0515625,
      "epoch": 0.10674157303370786,
      "grad_norm": 2.5040910243988037,
      "learning_rate": 5.3370786516853935e-06,
      "loss": 0.6013812065124512,
      "mean_token_accuracy": 0.8464263796806335,
      "num_tokens": 4984627.0,
      "step": 380
    },
    {
      "entropy": 2.1546875,
      "epoch": 0.10955056179775281,
      "grad_norm": 2.3937768936157227,
      "learning_rate": 5.477528089887641e-06,
      "loss": 0.5802605628967286,
      "mean_token_accuracy": 0.8509855687618255,
      "num_tokens": 5116021.0,
      "step": 390
    },
    {
      "entropy": 2.08125,
      "epoch": 0.11235955056179775,
      "grad_norm": 1.9691537618637085,
      "learning_rate": 5.617977528089889e-06,
      "loss": 0.5743422508239746,
      "mean_token_accuracy": 0.8529653191566468,
      "num_tokens": 5247474.0,
      "step": 400
    },
    {
      "epoch": 0.11235955056179775,
      "eval_entropy": 2.06110987544484,
      "eval_loss": 0.5866528749465942,
      "eval_mean_token_accuracy": 0.847210425499071,
      "eval_num_tokens": 5247474.0,
      "eval_runtime": 127.6545,
      "eval_samples_per_second": 123.239,
      "eval_steps_per_second": 2.201,
      "step": 400
    },
    {
      "entropy": 2.03515625,
      "epoch": 0.1151685393258427,
      "grad_norm": 2.3510892391204834,
      "learning_rate": 5.758426966292135e-06,
      "loss": 0.5805181503295899,
      "mean_token_accuracy": 0.848473459482193,
      "num_tokens": 5379956.0,
      "step": 410
    },
    {
      "entropy": 1.99375,
      "epoch": 0.11797752808988764,
      "grad_norm": 2.6800832748413086,
      "learning_rate": 5.8988764044943826e-06,
      "loss": 0.5805004596710205,
      "mean_token_accuracy": 0.848699277639389,
      "num_tokens": 5511211.0,
      "step": 420
    },
    {
      "entropy": 1.88671875,
      "epoch": 0.12078651685393259,
      "grad_norm": 2.39086651802063,
      "learning_rate": 6.03932584269663e-06,
      "loss": 0.5722110271453857,
      "mean_token_accuracy": 0.8508328795433044,
      "num_tokens": 5642237.0,
      "step": 430
    },
    {
      "entropy": 1.88984375,
      "epoch": 0.12359550561797752,
      "grad_norm": 2.485159397125244,
      "learning_rate": 6.179775280898876e-06,
      "loss": 0.5770604133605957,
      "mean_token_accuracy": 0.8501108467578888,
      "num_tokens": 5774030.0,
      "step": 440
    },
    {
      "entropy": 1.99609375,
      "epoch": 0.12640449438202248,
      "grad_norm": 2.1009361743927,
      "learning_rate": 6.320224719101124e-06,
      "loss": 0.5737769126892089,
      "mean_token_accuracy": 0.8525624096393585,
      "num_tokens": 5905167.0,
      "step": 450
    },
    {
      "entropy": 2.0515625,
      "epoch": 0.12921348314606743,
      "grad_norm": 2.3859102725982666,
      "learning_rate": 6.460674157303372e-06,
      "loss": 0.5859375476837159,
      "mean_token_accuracy": 0.8465817093849182,
      "num_tokens": 6037472.0,
      "step": 460
    },
    {
      "entropy": 1.9796875,
      "epoch": 0.13202247191011235,
      "grad_norm": 2.1753172874450684,
      "learning_rate": 6.601123595505619e-06,
      "loss": 0.5598746299743652,
      "mean_token_accuracy": 0.8553961634635925,
      "num_tokens": 6169788.0,
      "step": 470
    },
    {
      "entropy": 1.890625,
      "epoch": 0.1348314606741573,
      "grad_norm": 2.234368085861206,
      "learning_rate": 6.741573033707865e-06,
      "loss": 0.5988113880157471,
      "mean_token_accuracy": 0.8424076497554779,
      "num_tokens": 6301248.0,
      "step": 480
    },
    {
      "entropy": 1.91953125,
      "epoch": 0.13764044943820225,
      "grad_norm": 2.200791120529175,
      "learning_rate": 6.8820224719101126e-06,
      "loss": 0.568276309967041,
      "mean_token_accuracy": 0.8514572620391846,
      "num_tokens": 6433763.0,
      "step": 490
    },
    {
      "entropy": 1.9140625,
      "epoch": 0.1404494382022472,
      "grad_norm": 2.308649778366089,
      "learning_rate": 7.022471910112361e-06,
      "loss": 0.5822394371032715,
      "mean_token_accuracy": 0.8506459176540375,
      "num_tokens": 6565947.0,
      "step": 500
    },
    {
      "epoch": 0.1404494382022472,
      "eval_entropy": 2.079348309608541,
      "eval_loss": 0.5725127458572388,
      "eval_mean_token_accuracy": 0.8500019004336455,
      "eval_num_tokens": 6565947.0,
      "eval_runtime": 129.5228,
      "eval_samples_per_second": 121.461,
      "eval_steps_per_second": 2.17,
      "step": 500
    },
    {
      "entropy": 2.165625,
      "epoch": 0.14325842696629212,
      "grad_norm": 2.0097155570983887,
      "learning_rate": 7.162921348314608e-06,
      "loss": 0.5781090259552002,
      "mean_token_accuracy": 0.8493442416191102,
      "num_tokens": 6697776.0,
      "step": 510
    },
    {
      "entropy": 2.2625,
      "epoch": 0.14606741573033707,
      "grad_norm": 2.2164199352264404,
      "learning_rate": 7.303370786516854e-06,
      "loss": 0.5849899768829345,
      "mean_token_accuracy": 0.8462651312351227,
      "num_tokens": 6829845.0,
      "step": 520
    },
    {
      "entropy": 2.325,
      "epoch": 0.14887640449438203,
      "grad_norm": 2.379678249359131,
      "learning_rate": 7.443820224719102e-06,
      "loss": 0.5576138019561767,
      "mean_token_accuracy": 0.8528270661830902,
      "num_tokens": 6960368.0,
      "step": 530
    },
    {
      "entropy": 2.1390625,
      "epoch": 0.15168539325842698,
      "grad_norm": 1.7464559078216553,
      "learning_rate": 7.584269662921349e-06,
      "loss": 0.5554574489593506,
      "mean_token_accuracy": 0.8532874822616577,
      "num_tokens": 7091153.0,
      "step": 540
    },
    {
      "entropy": 2.18125,
      "epoch": 0.1544943820224719,
      "grad_norm": 2.0346474647521973,
      "learning_rate": 7.724719101123597e-06,
      "loss": 0.5665736675262452,
      "mean_token_accuracy": 0.8527543544769287,
      "num_tokens": 7222459.0,
      "step": 550
    },
    {
      "entropy": 2.1703125,
      "epoch": 0.15730337078651685,
      "grad_norm": 2.339033842086792,
      "learning_rate": 7.865168539325843e-06,
      "loss": 0.5432186603546143,
      "mean_token_accuracy": 0.8581551015377045,
      "num_tokens": 7353747.0,
      "step": 560
    },
    {
      "entropy": 2.1546875,
      "epoch": 0.1601123595505618,
      "grad_norm": 2.061462879180908,
      "learning_rate": 8.00561797752809e-06,
      "loss": 0.5740026950836181,
      "mean_token_accuracy": 0.8499646544456482,
      "num_tokens": 7484915.0,
      "step": 570
    },
    {
      "entropy": 2.246875,
      "epoch": 0.16292134831460675,
      "grad_norm": 1.7877792119979858,
      "learning_rate": 8.146067415730338e-06,
      "loss": 0.5391734600067138,
      "mean_token_accuracy": 0.8582757353782654,
      "num_tokens": 7616484.0,
      "step": 580
    },
    {
      "entropy": 2.396875,
      "epoch": 0.16573033707865167,
      "grad_norm": 2.367311954498291,
      "learning_rate": 8.286516853932584e-06,
      "loss": 0.5691903591156006,
      "mean_token_accuracy": 0.8493059575557709,
      "num_tokens": 7747356.0,
      "step": 590
    },
    {
      "entropy": 2.403125,
      "epoch": 0.16853932584269662,
      "grad_norm": 1.9094247817993164,
      "learning_rate": 8.426966292134832e-06,
      "loss": 0.554432487487793,
      "mean_token_accuracy": 0.8518336236476898,
      "num_tokens": 7878201.0,
      "step": 600
    },
    {
      "epoch": 0.16853932584269662,
      "eval_entropy": 2.287811387900356,
      "eval_loss": 0.5606739521026611,
      "eval_mean_token_accuracy": 0.8524114449253286,
      "eval_num_tokens": 7878201.0,
      "eval_runtime": 125.0306,
      "eval_samples_per_second": 125.825,
      "eval_steps_per_second": 2.247,
      "step": 600
    },
    {
      "entropy": 2.365625,
      "epoch": 0.17134831460674158,
      "grad_norm": 1.7424993515014648,
      "learning_rate": 8.56741573033708e-06,
      "loss": 0.5393033027648926,
      "mean_token_accuracy": 0.8589720845222473,
      "num_tokens": 8010168.0,
      "step": 610
    },
    {
      "entropy": 2.2453125,
      "epoch": 0.17415730337078653,
      "grad_norm": 2.0761020183563232,
      "learning_rate": 8.707865168539327e-06,
      "loss": 0.5451316356658935,
      "mean_token_accuracy": 0.8561301052570343,
      "num_tokens": 8141269.0,
      "step": 620
    },
    {
      "entropy": 2.1609375,
      "epoch": 0.17696629213483145,
      "grad_norm": 1.9591768980026245,
      "learning_rate": 8.848314606741573e-06,
      "loss": 0.5643636226654053,
      "mean_token_accuracy": 0.8521402716636658,
      "num_tokens": 8273144.0,
      "step": 630
    },
    {
      "entropy": 2.35,
      "epoch": 0.1797752808988764,
      "grad_norm": 2.199676036834717,
      "learning_rate": 8.988764044943822e-06,
      "loss": 0.5512661933898926,
      "mean_token_accuracy": 0.853339409828186,
      "num_tokens": 8404074.0,
      "step": 640
    },
    {
      "entropy": 2.6546875,
      "epoch": 0.18258426966292135,
      "grad_norm": 1.5772382020950317,
      "learning_rate": 9.129213483146068e-06,
      "loss": 0.5461832523345947,
      "mean_token_accuracy": 0.8566304683685303,
      "num_tokens": 8535151.0,
      "step": 650
    },
    {
      "entropy": 2.6296875,
      "epoch": 0.1853932584269663,
      "grad_norm": 1.908815622329712,
      "learning_rate": 9.269662921348316e-06,
      "loss": 0.5526309490203858,
      "mean_token_accuracy": 0.8560527503490448,
      "num_tokens": 8666307.0,
      "step": 660
    },
    {
      "entropy": 2.4203125,
      "epoch": 0.18820224719101122,
      "grad_norm": 1.7587661743164062,
      "learning_rate": 9.410112359550562e-06,
      "loss": 0.5451655864715577,
      "mean_token_accuracy": 0.8554248094558716,
      "num_tokens": 8797853.0,
      "step": 670
    },
    {
      "entropy": 2.6328125,
      "epoch": 0.19101123595505617,
      "grad_norm": 2.1652209758758545,
      "learning_rate": 9.55056179775281e-06,
      "loss": 0.5510248661041259,
      "mean_token_accuracy": 0.8524434685707092,
      "num_tokens": 8929332.0,
      "step": 680
    },
    {
      "entropy": 2.4453125,
      "epoch": 0.19382022471910113,
      "grad_norm": 1.934767484664917,
      "learning_rate": 9.691011235955057e-06,
      "loss": 0.5439024448394776,
      "mean_token_accuracy": 0.857047837972641,
      "num_tokens": 9061328.0,
      "step": 690
    },
    {
      "entropy": 2.45625,
      "epoch": 0.19662921348314608,
      "grad_norm": 1.7675893306732178,
      "learning_rate": 9.831460674157303e-06,
      "loss": 0.5488356590270996,
      "mean_token_accuracy": 0.8547171115875244,
      "num_tokens": 9193213.0,
      "step": 700
    },
    {
      "epoch": 0.19662921348314608,
      "eval_entropy": 2.51918371886121,
      "eval_loss": 0.5528153777122498,
      "eval_mean_token_accuracy": 0.8541719907119181,
      "eval_num_tokens": 9193213.0,
      "eval_runtime": 120.9055,
      "eval_samples_per_second": 130.118,
      "eval_steps_per_second": 2.324,
      "step": 700
    },
    {
      "entropy": 2.4234375,
      "epoch": 0.199438202247191,
      "grad_norm": 2.2199792861938477,
      "learning_rate": 9.971910112359552e-06,
      "loss": 0.5681753635406495,
      "mean_token_accuracy": 0.8511766195297241,
      "num_tokens": 9324297.0,
      "step": 710
    },
    {
      "entropy": 2.475,
      "epoch": 0.20224719101123595,
      "grad_norm": 2.0278470516204834,
      "learning_rate": 9.999961543109546e-06,
      "loss": 0.528502082824707,
      "mean_token_accuracy": 0.8589128077030181,
      "num_tokens": 9455706.0,
      "step": 720
    },
    {
      "entropy": 2.521875,
      "epoch": 0.2050561797752809,
      "grad_norm": 2.2629213333129883,
      "learning_rate": 9.999805313005946e-06,
      "loss": 0.5325650215148926,
      "mean_token_accuracy": 0.8591304659843445,
      "num_tokens": 9586251.0,
      "step": 730
    },
    {
      "entropy": 2.4953125,
      "epoch": 0.20786516853932585,
      "grad_norm": 1.9661778211593628,
      "learning_rate": 9.999528909885738e-06,
      "loss": 0.5466290950775147,
      "mean_token_accuracy": 0.8556427180767059,
      "num_tokens": 9717226.0,
      "step": 740
    },
    {
      "entropy": 2.434375,
      "epoch": 0.21067415730337077,
      "grad_norm": 1.7269171476364136,
      "learning_rate": 9.999132340392416e-06,
      "loss": 0.5469918251037598,
      "mean_token_accuracy": 0.8553809940814971,
      "num_tokens": 9848587.0,
      "step": 750
    },
    {
      "entropy": 2.5421875,
      "epoch": 0.21348314606741572,
      "grad_norm": 2.1285669803619385,
      "learning_rate": 9.998615614057743e-06,
      "loss": 0.5462213039398194,
      "mean_token_accuracy": 0.8554510354995728,
      "num_tokens": 9979862.0,
      "step": 760
    },
    {
      "entropy": 2.34375,
      "epoch": 0.21629213483146068,
      "grad_norm": 2.3589627742767334,
      "learning_rate": 9.997978743301516e-06,
      "loss": 0.5489914417266846,
      "mean_token_accuracy": 0.8535460829734802,
      "num_tokens": 10111838.0,
      "step": 770
    },
    {
      "entropy": 2.2921875,
      "epoch": 0.21910112359550563,
      "grad_norm": 1.7968486547470093,
      "learning_rate": 9.997221743431267e-06,
      "loss": 0.5411812782287597,
      "mean_token_accuracy": 0.8545774519443512,
      "num_tokens": 10242186.0,
      "step": 780
    },
    {
      "entropy": 2.4828125,
      "epoch": 0.22191011235955055,
      "grad_norm": 1.8283753395080566,
      "learning_rate": 9.996344632641895e-06,
      "loss": 0.5593603134155274,
      "mean_token_accuracy": 0.8525359392166137,
      "num_tokens": 10373751.0,
      "step": 790
    },
    {
      "entropy": 2.10390625,
      "epoch": 0.2247191011235955,
      "grad_norm": 1.7177592515945435,
      "learning_rate": 9.99534743201523e-06,
      "loss": 0.5581941604614258,
      "mean_token_accuracy": 0.8535070955753327,
      "num_tokens": 10505346.0,
      "step": 800
    },
    {
      "epoch": 0.2247191011235955,
      "eval_entropy": 1.8966025355871887,
      "eval_loss": 0.5411612391471863,
      "eval_mean_token_accuracy": 0.8565805688023143,
      "eval_num_tokens": 10505346.0,
      "eval_runtime": 123.8657,
      "eval_samples_per_second": 127.009,
      "eval_steps_per_second": 2.269,
      "step": 800
    },
    {
      "entropy": 1.8125,
      "epoch": 0.22752808988764045,
      "grad_norm": 1.9482078552246094,
      "learning_rate": 9.994230165519529e-06,
      "loss": 0.5144902229309082,
      "mean_token_accuracy": 0.8626653552055359,
      "num_tokens": 10635148.0,
      "step": 810
    },
    {
      "entropy": 1.84609375,
      "epoch": 0.2303370786516854,
      "grad_norm": 1.6378909349441528,
      "learning_rate": 9.992992860008893e-06,
      "loss": 0.5271161079406739,
      "mean_token_accuracy": 0.8606956720352172,
      "num_tokens": 10766521.0,
      "step": 820
    },
    {
      "entropy": 1.8875,
      "epoch": 0.23314606741573032,
      "grad_norm": 1.9249287843704224,
      "learning_rate": 9.991635545222628e-06,
      "loss": 0.5520608425140381,
      "mean_token_accuracy": 0.8538833498954773,
      "num_tokens": 10897825.0,
      "step": 830
    },
    {
      "entropy": 1.98515625,
      "epoch": 0.23595505617977527,
      "grad_norm": 1.8576961755752563,
      "learning_rate": 9.990158253784525e-06,
      "loss": 0.5513722419738769,
      "mean_token_accuracy": 0.853681880235672,
      "num_tokens": 11028654.0,
      "step": 840
    },
    {
      "entropy": 2.1484375,
      "epoch": 0.23876404494382023,
      "grad_norm": 1.6137062311172485,
      "learning_rate": 9.988561021202083e-06,
      "loss": 0.5378435134887696,
      "mean_token_accuracy": 0.8573130428791046,
      "num_tokens": 11159974.0,
      "step": 850
    },
    {
      "entropy": 2.4359375,
      "epoch": 0.24157303370786518,
      "grad_norm": 1.955034613609314,
      "learning_rate": 9.986843885865649e-06,
      "loss": 0.5395791530609131,
      "mean_token_accuracy": 0.8567401230335235,
      "num_tokens": 11292002.0,
      "step": 860
    },
    {
      "entropy": 2.421875,
      "epoch": 0.2443820224719101,
      "grad_norm": 1.563650131225586,
      "learning_rate": 9.985006889047492e-06,
      "loss": 0.5066062450408936,
      "mean_token_accuracy": 0.8660560488700867,
      "num_tokens": 11422854.0,
      "step": 870
    },
    {
      "entropy": 2.5734375,
      "epoch": 0.24719101123595505,
      "grad_norm": 1.9510332345962524,
      "learning_rate": 9.983050074900824e-06,
      "loss": 0.5264874458312988,
      "mean_token_accuracy": 0.8599664986133575,
      "num_tokens": 11553438.0,
      "step": 880
    },
    {
      "entropy": 2.7,
      "epoch": 0.25,
      "grad_norm": 1.595357894897461,
      "learning_rate": 9.980973490458728e-06,
      "loss": 0.49913692474365234,
      "mean_token_accuracy": 0.8673109352588654,
      "num_tokens": 11685069.0,
      "step": 890
    },
    {
      "entropy": 2.5828125,
      "epoch": 0.25280898876404495,
      "grad_norm": 1.745118498802185,
      "learning_rate": 9.978777185633032e-06,
      "loss": 0.5217289924621582,
      "mean_token_accuracy": 0.859529334306717,
      "num_tokens": 11815835.0,
      "step": 900
    },
    {
      "epoch": 0.25280898876404495,
      "eval_entropy": 2.3949621886120998,
      "eval_loss": 0.5336104035377502,
      "eval_mean_token_accuracy": 0.8577707471372394,
      "eval_num_tokens": 11815835.0,
      "eval_runtime": 124.4757,
      "eval_samples_per_second": 126.386,
      "eval_steps_per_second": 2.257,
      "step": 900
    },
    {
      "entropy": 2.2515625,
      "epoch": 0.2556179775280899,
      "grad_norm": 1.4799717664718628,
      "learning_rate": 9.976461213213104e-06,
      "loss": 0.5158873558044433,
      "mean_token_accuracy": 0.8613489508628845,
      "num_tokens": 11947149.0,
      "step": 910
    },
    {
      "entropy": 2.246875,
      "epoch": 0.25842696629213485,
      "grad_norm": 1.8871139287948608,
      "learning_rate": 9.974025628864592e-06,
      "loss": 0.5089314937591553,
      "mean_token_accuracy": 0.8655749261379242,
      "num_tokens": 12078488.0,
      "step": 920
    },
    {
      "entropy": 2.5171875,
      "epoch": 0.2612359550561798,
      "grad_norm": 1.8753252029418945,
      "learning_rate": 9.971470491128077e-06,
      "loss": 0.5227785587310791,
      "mean_token_accuracy": 0.862045019865036,
      "num_tokens": 12209845.0,
      "step": 930
    },
    {
      "entropy": 2.5078125,
      "epoch": 0.2640449438202247,
      "grad_norm": 1.757841944694519,
      "learning_rate": 9.968795861417676e-06,
      "loss": 0.5381661415100097,
      "mean_token_accuracy": 0.8577485024929047,
      "num_tokens": 12341432.0,
      "step": 940
    },
    {
      "entropy": 2.3859375,
      "epoch": 0.26685393258426965,
      "grad_norm": 1.9045883417129517,
      "learning_rate": 9.966001804019552e-06,
      "loss": 0.5183176517486572,
      "mean_token_accuracy": 0.8622296333312989,
      "num_tokens": 12471897.0,
      "step": 950
    },
    {
      "entropy": 2.403125,
      "epoch": 0.2696629213483146,
      "grad_norm": 1.6883784532546997,
      "learning_rate": 9.963088386090386e-06,
      "loss": 0.5234039306640625,
      "mean_token_accuracy": 0.8607744336128235,
      "num_tokens": 12604287.0,
      "step": 960
    },
    {
      "entropy": 2.45625,
      "epoch": 0.27247191011235955,
      "grad_norm": 1.5262446403503418,
      "learning_rate": 9.960055677655743e-06,
      "loss": 0.5076072692871094,
      "mean_token_accuracy": 0.8646633923053741,
      "num_tokens": 12734991.0,
      "step": 970
    },
    {
      "entropy": 2.5328125,
      "epoch": 0.2752808988764045,
      "grad_norm": 1.690694808959961,
      "learning_rate": 9.956903751608409e-06,
      "loss": 0.5351317405700684,
      "mean_token_accuracy": 0.8547135591506958,
      "num_tokens": 12867494.0,
      "step": 980
    },
    {
      "entropy": 2.571875,
      "epoch": 0.27808988764044945,
      "grad_norm": 1.7971384525299072,
      "learning_rate": 9.953632683706624e-06,
      "loss": 0.5345530986785889,
      "mean_token_accuracy": 0.8580621659755707,
      "num_tokens": 12999312.0,
      "step": 990
    },
    {
      "entropy": 2.79375,
      "epoch": 0.2808988764044944,
      "grad_norm": 1.6464053392410278,
      "learning_rate": 9.950242552572272e-06,
      "loss": 0.5215201854705811,
      "mean_token_accuracy": 0.8577286064624786,
      "num_tokens": 13130633.0,
      "step": 1000
    },
    {
      "epoch": 0.2808988764044944,
      "eval_entropy": 3.1947286476868326,
      "eval_loss": 0.5258083939552307,
      "eval_mean_token_accuracy": 0.8594923481822438,
      "eval_num_tokens": 13130633.0,
      "eval_runtime": 125.9673,
      "eval_samples_per_second": 124.89,
      "eval_steps_per_second": 2.231,
      "step": 1000
    },
    {
      "entropy": 3.696875,
      "epoch": 0.28370786516853935,
      "grad_norm": 1.6463587284088135,
      "learning_rate": 9.946733439688982e-06,
      "loss": 0.5136995792388916,
      "mean_token_accuracy": 0.8632636308670044,
      "num_tokens": 13262304.0,
      "step": 1010
    },
    {
      "entropy": 3.8640625,
      "epoch": 0.28651685393258425,
      "grad_norm": 1.9419281482696533,
      "learning_rate": 9.943105429400171e-06,
      "loss": 0.5002095222473144,
      "mean_token_accuracy": 0.864561277627945,
      "num_tokens": 13392366.0,
      "step": 1020
    },
    {
      "entropy": 4.265625,
      "epoch": 0.2893258426966292,
      "grad_norm": 1.6725404262542725,
      "learning_rate": 9.939358608907026e-06,
      "loss": 0.528085994720459,
      "mean_token_accuracy": 0.8600175321102143,
      "num_tokens": 13524061.0,
      "step": 1030
    },
    {
      "entropy": 4.225,
      "epoch": 0.29213483146067415,
      "grad_norm": 1.522539734840393,
      "learning_rate": 9.935493068266396e-06,
      "loss": 0.5215230941772461,
      "mean_token_accuracy": 0.8587561547756195,
      "num_tokens": 13653978.0,
      "step": 1040
    },
    {
      "entropy": 4.1546875,
      "epoch": 0.2949438202247191,
      "grad_norm": 1.5519685745239258,
      "learning_rate": 9.931508900388635e-06,
      "loss": 0.5285672664642334,
      "mean_token_accuracy": 0.8602302074432373,
      "num_tokens": 13785015.0,
      "step": 1050
    },
    {
      "entropy": 4.05625,
      "epoch": 0.29775280898876405,
      "grad_norm": 1.5624417066574097,
      "learning_rate": 9.927406201035368e-06,
      "loss": 0.4981082916259766,
      "mean_token_accuracy": 0.8646525084972382,
      "num_tokens": 13915466.0,
      "step": 1060
    },
    {
      "entropy": 3.9171875,
      "epoch": 0.300561797752809,
      "grad_norm": 1.5555429458618164,
      "learning_rate": 9.923185068817184e-06,
      "loss": 0.5010139465332031,
      "mean_token_accuracy": 0.8666773021221161,
      "num_tokens": 14047636.0,
      "step": 1070
    },
    {
      "entropy": 3.39375,
      "epoch": 0.30337078651685395,
      "grad_norm": 1.750293493270874,
      "learning_rate": 9.918845605191274e-06,
      "loss": 0.5290132522583008,
      "mean_token_accuracy": 0.8582139611244202,
      "num_tokens": 14178615.0,
      "step": 1080
    },
    {
      "entropy": 3.425,
      "epoch": 0.3061797752808989,
      "grad_norm": 1.5983638763427734,
      "learning_rate": 9.914387914458983e-06,
      "loss": 0.49461803436279295,
      "mean_token_accuracy": 0.8661557137966156,
      "num_tokens": 14309400.0,
      "step": 1090
    },
    {
      "entropy": 3.734375,
      "epoch": 0.3089887640449438,
      "grad_norm": 1.5704827308654785,
      "learning_rate": 9.909812103763312e-06,
      "loss": 0.5299598693847656,
      "mean_token_accuracy": 0.8603205621242523,
      "num_tokens": 14440876.0,
      "step": 1100
    },
    {
      "epoch": 0.3089887640449438,
      "eval_entropy": 4.083907918149467,
      "eval_loss": 0.52057284116745,
      "eval_mean_token_accuracy": 0.8605949876571465,
      "eval_num_tokens": 14440876.0,
      "eval_runtime": 125.1486,
      "eval_samples_per_second": 125.707,
      "eval_steps_per_second": 2.245,
      "step": 1100
    },
    {
      "entropy": 3.859375,
      "epoch": 0.31179775280898875,
      "grad_norm": 1.6799899339675903,
      "learning_rate": 9.905118283086341e-06,
      "loss": 0.51255784034729,
      "mean_token_accuracy": 0.8601176261901855,
      "num_tokens": 14571576.0,
      "step": 1110
    },
    {
      "entropy": 4.1265625,
      "epoch": 0.3146067415730337,
      "grad_norm": 1.6995667219161987,
      "learning_rate": 9.900306565246579e-06,
      "loss": 0.5254480838775635,
      "mean_token_accuracy": 0.8585749983787536,
      "num_tokens": 14703310.0,
      "step": 1120
    },
    {
      "entropy": 4.0171875,
      "epoch": 0.31741573033707865,
      "grad_norm": 1.6145212650299072,
      "learning_rate": 9.895377065896259e-06,
      "loss": 0.5028430461883545,
      "mean_token_accuracy": 0.8649314641952515,
      "num_tokens": 14834265.0,
      "step": 1130
    },
    {
      "entropy": 3.8,
      "epoch": 0.3202247191011236,
      "grad_norm": 1.4251447916030884,
      "learning_rate": 9.890329903518554e-06,
      "loss": 0.5098093032836915,
      "mean_token_accuracy": 0.8632848262786865,
      "num_tokens": 14965478.0,
      "step": 1140
    },
    {
      "entropy": 3.7546875,
      "epoch": 0.32303370786516855,
      "grad_norm": 1.6261520385742188,
      "learning_rate": 9.885165199424738e-06,
      "loss": 0.5035204410552978,
      "mean_token_accuracy": 0.8646722733974457,
      "num_tokens": 15097612.0,
      "step": 1150
    },
    {
      "entropy": 3.3484375,
      "epoch": 0.3258426966292135,
      "grad_norm": 1.6217759847640991,
      "learning_rate": 9.879883077751255e-06,
      "loss": 0.5101470947265625,
      "mean_token_accuracy": 0.8628816902637482,
      "num_tokens": 15228064.0,
      "step": 1160
    },
    {
      "entropy": 3.69375,
      "epoch": 0.32865168539325845,
      "grad_norm": 1.335907220840454,
      "learning_rate": 9.874483665456746e-06,
      "loss": 0.5171346187591552,
      "mean_token_accuracy": 0.861205780506134,
      "num_tokens": 15360788.0,
      "step": 1170
    },
    {
      "entropy": 3.484375,
      "epoch": 0.33146067415730335,
      "grad_norm": 1.496446132659912,
      "learning_rate": 9.868967092319003e-06,
      "loss": 0.5193340301513671,
      "mean_token_accuracy": 0.8619067668914795,
      "num_tokens": 15492328.0,
      "step": 1180
    },
    {
      "entropy": 3.7234375,
      "epoch": 0.3342696629213483,
      "grad_norm": 1.3914567232131958,
      "learning_rate": 9.86333349093183e-06,
      "loss": 0.516145658493042,
      "mean_token_accuracy": 0.8638903021812439,
      "num_tokens": 15624033.0,
      "step": 1190
    },
    {
      "entropy": 3.8015625,
      "epoch": 0.33707865168539325,
      "grad_norm": 1.3881893157958984,
      "learning_rate": 9.857582996701878e-06,
      "loss": 0.5512511253356933,
      "mean_token_accuracy": 0.8553680300712585,
      "num_tokens": 15755794.0,
      "step": 1200
    },
    {
      "epoch": 0.33707865168539325,
      "eval_entropy": 3.676100978647687,
      "eval_loss": 0.511896550655365,
      "eval_mean_token_accuracy": 0.8620759745085367,
      "eval_num_tokens": 15755794.0,
      "eval_runtime": 124.6557,
      "eval_samples_per_second": 126.204,
      "eval_steps_per_second": 2.254,
      "step": 1200
    },
    {
      "entropy": 3.4921875,
      "epoch": 0.3398876404494382,
      "grad_norm": 1.3239006996154785,
      "learning_rate": 9.851715747845372e-06,
      "loss": 0.49883713722229006,
      "mean_token_accuracy": 0.8651662051677704,
      "num_tokens": 15887083.0,
      "step": 1210
    },
    {
      "entropy": 3.5015625,
      "epoch": 0.34269662921348315,
      "grad_norm": 1.9596682786941528,
      "learning_rate": 9.845731885384806e-06,
      "loss": 0.5311936855316162,
      "mean_token_accuracy": 0.8579659998416901,
      "num_tokens": 16018235.0,
      "step": 1220
    },
    {
      "entropy": 3.38125,
      "epoch": 0.3455056179775281,
      "grad_norm": 1.6833865642547607,
      "learning_rate": 9.83963155314554e-06,
      "loss": 0.5027567386627197,
      "mean_token_accuracy": 0.866038566827774,
      "num_tokens": 16149393.0,
      "step": 1230
    },
    {
      "entropy": 3.165625,
      "epoch": 0.34831460674157305,
      "grad_norm": 1.5869909524917603,
      "learning_rate": 9.833414897752346e-06,
      "loss": 0.5135039329528809,
      "mean_token_accuracy": 0.8614062786102294,
      "num_tokens": 16280866.0,
      "step": 1240
    },
    {
      "entropy": 3.26875,
      "epoch": 0.351123595505618,
      "grad_norm": 1.80933678150177,
      "learning_rate": 9.827082068625893e-06,
      "loss": 0.512190580368042,
      "mean_token_accuracy": 0.8624402940273285,
      "num_tokens": 16412868.0,
      "step": 1250
    },
    {
      "entropy": 3.1328125,
      "epoch": 0.3539325842696629,
      "grad_norm": 1.5499094724655151,
      "learning_rate": 9.82063321797914e-06,
      "loss": 0.5112735748291015,
      "mean_token_accuracy": 0.8615074872970581,
      "num_tokens": 16543372.0,
      "step": 1260
    },
    {
      "entropy": 3.171875,
      "epoch": 0.35674157303370785,
      "grad_norm": 1.4488999843597412,
      "learning_rate": 9.814068500813692e-06,
      "loss": 0.5110087871551514,
      "mean_token_accuracy": 0.8628510951995849,
      "num_tokens": 16674900.0,
      "step": 1270
    },
    {
      "entropy": 3.234375,
      "epoch": 0.3595505617977528,
      "grad_norm": 1.338806390762329,
      "learning_rate": 9.807388074916064e-06,
      "loss": 0.5112171173095703,
      "mean_token_accuracy": 0.8626433491706849,
      "num_tokens": 16806663.0,
      "step": 1280
    },
    {
      "entropy": 3.2234375,
      "epoch": 0.36235955056179775,
      "grad_norm": 1.6451371908187866,
      "learning_rate": 9.800592100853894e-06,
      "loss": 0.5251486301422119,
      "mean_token_accuracy": 0.8587349593639374,
      "num_tokens": 16937940.0,
      "step": 1290
    },
    {
      "entropy": 3.3046875,
      "epoch": 0.3651685393258427,
      "grad_norm": 1.816031813621521,
      "learning_rate": 9.793680741972084e-06,
      "loss": 0.5063271999359131,
      "mean_token_accuracy": 0.8647400677204132,
      "num_tokens": 17068818.0,
      "step": 1300
    },
    {
      "epoch": 0.3651685393258427,
      "eval_entropy": 3.4022464412811386,
      "eval_loss": 0.509096622467041,
      "eval_mean_token_accuracy": 0.8629877645774244,
      "eval_num_tokens": 17068818.0,
      "eval_runtime": 114.3982,
      "eval_samples_per_second": 137.52,
      "eval_steps_per_second": 2.456,
      "step": 1300
    },
    {
      "entropy": 3.271875,
      "epoch": 0.36797752808988765,
      "grad_norm": 1.4063043594360352,
      "learning_rate": 9.786654164388873e-06,
      "loss": 0.5131440162658691,
      "mean_token_accuracy": 0.8604735016822815,
      "num_tokens": 17199605.0,
      "step": 1310
    },
    {
      "entropy": 3.0625,
      "epoch": 0.3707865168539326,
      "grad_norm": 1.7652196884155273,
      "learning_rate": 9.779512536991839e-06,
      "loss": 0.5196240425109864,
      "mean_token_accuracy": 0.8597306191921235,
      "num_tokens": 17330477.0,
      "step": 1320
    },
    {
      "entropy": 3.1109375,
      "epoch": 0.37359550561797755,
      "grad_norm": 1.5222713947296143,
      "learning_rate": 9.77225603143385e-06,
      "loss": 0.5136602401733399,
      "mean_token_accuracy": 0.8612237632274627,
      "num_tokens": 17461988.0,
      "step": 1330
    },
    {
      "entropy": 2.9875,
      "epoch": 0.37640449438202245,
      "grad_norm": 1.4366384744644165,
      "learning_rate": 9.764884822128928e-06,
      "loss": 0.5219603538513183,
      "mean_token_accuracy": 0.8586396098136901,
      "num_tokens": 17593734.0,
      "step": 1340
    },
    {
      "entropy": 2.946875,
      "epoch": 0.3792134831460674,
      "grad_norm": 1.4805457592010498,
      "learning_rate": 9.757399086248062e-06,
      "loss": 0.5044600009918213,
      "mean_token_accuracy": 0.8650240898132324,
      "num_tokens": 17725331.0,
      "step": 1350
    },
    {
      "entropy": 3.4125,
      "epoch": 0.38202247191011235,
      "grad_norm": 1.3992596864700317,
      "learning_rate": 9.749799003714954e-06,
      "loss": 0.5088860988616943,
      "mean_token_accuracy": 0.8633273124694825,
      "num_tokens": 17856425.0,
      "step": 1360
    },
    {
      "entropy": 3.0140625,
      "epoch": 0.3848314606741573,
      "grad_norm": 1.6114635467529297,
      "learning_rate": 9.742084757201684e-06,
      "loss": 0.5030006408691406,
      "mean_token_accuracy": 0.8628417015075683,
      "num_tokens": 17987223.0,
      "step": 1370
    },
    {
      "entropy": 3.021875,
      "epoch": 0.38764044943820225,
      "grad_norm": 1.4702385663986206,
      "learning_rate": 9.734256532124326e-06,
      "loss": 0.5308702945709228,
      "mean_token_accuracy": 0.8586233198642731,
      "num_tokens": 18119142.0,
      "step": 1380
    },
    {
      "entropy": 3.0578125,
      "epoch": 0.3904494382022472,
      "grad_norm": 1.5420409440994263,
      "learning_rate": 9.72631451663849e-06,
      "loss": 0.5003273010253906,
      "mean_token_accuracy": 0.8662057340145111,
      "num_tokens": 18250390.0,
      "step": 1390
    },
    {
      "entropy": 3.0046875,
      "epoch": 0.39325842696629215,
      "grad_norm": 1.559325933456421,
      "learning_rate": 9.718258901634802e-06,
      "loss": 0.5051637172698975,
      "mean_token_accuracy": 0.8637124717235565,
      "num_tokens": 18381581.0,
      "step": 1400
    },
    {
      "epoch": 0.39325842696629215,
      "eval_entropy": 3.0663923487544484,
      "eval_loss": 0.5048597455024719,
      "eval_mean_token_accuracy": 0.8639065364501655,
      "eval_num_tokens": 18381581.0,
      "eval_runtime": 119.9524,
      "eval_samples_per_second": 131.152,
      "eval_steps_per_second": 2.343,
      "step": 1400
    },
    {
      "entropy": 3.1390625,
      "epoch": 0.3960674157303371,
      "grad_norm": 1.3189454078674316,
      "learning_rate": 9.71008988073431e-06,
      "loss": 0.5008738994598388,
      "mean_token_accuracy": 0.864854609966278,
      "num_tokens": 18512480.0,
      "step": 1410
    },
    {
      "entropy": 3.184375,
      "epoch": 0.398876404494382,
      "grad_norm": 1.5395656824111938,
      "learning_rate": 9.70180765028384e-06,
      "loss": 0.49518346786499023,
      "mean_token_accuracy": 0.8670604348182678,
      "num_tokens": 18644160.0,
      "step": 1420
    },
    {
      "entropy": 3.215625,
      "epoch": 0.40168539325842695,
      "grad_norm": 1.449451208114624,
      "learning_rate": 9.693412409351264e-06,
      "loss": 0.4907248020172119,
      "mean_token_accuracy": 0.8667110919952392,
      "num_tokens": 18775155.0,
      "step": 1430
    },
    {
      "entropy": 3.1421875,
      "epoch": 0.4044943820224719,
      "grad_norm": 1.3559569120407104,
      "learning_rate": 9.684904359720724e-06,
      "loss": 0.4977103233337402,
      "mean_token_accuracy": 0.8653888046741486,
      "num_tokens": 18905572.0,
      "step": 1440
    },
    {
      "entropy": 2.840625,
      "epoch": 0.40730337078651685,
      "grad_norm": 1.349798560142517,
      "learning_rate": 9.676283705887783e-06,
      "loss": 0.49252843856811523,
      "mean_token_accuracy": 0.8659845530986786,
      "num_tokens": 19036816.0,
      "step": 1450
    },
    {
      "entropy": 3.0,
      "epoch": 0.4101123595505618,
      "grad_norm": 1.8797097206115723,
      "learning_rate": 9.6675506550545e-06,
      "loss": 0.4878744125366211,
      "mean_token_accuracy": 0.8690207064151764,
      "num_tokens": 19168400.0,
      "step": 1460
    },
    {
      "entropy": 3.3953125,
      "epoch": 0.41292134831460675,
      "grad_norm": 2.1975135803222656,
      "learning_rate": 9.658705417124466e-06,
      "loss": 0.4938823699951172,
      "mean_token_accuracy": 0.8655854105949402,
      "num_tokens": 19299202.0,
      "step": 1470
    },
    {
      "entropy": 3.9703125,
      "epoch": 0.4157303370786517,
      "grad_norm": 1.4038432836532593,
      "learning_rate": 9.649748204697741e-06,
      "loss": 0.5063140392303467,
      "mean_token_accuracy": 0.8641268849372864,
      "num_tokens": 19430678.0,
      "step": 1480
    },
    {
      "entropy": 4.34375,
      "epoch": 0.41853932584269665,
      "grad_norm": 1.3492214679718018,
      "learning_rate": 9.640679233065755e-06,
      "loss": 0.480499792098999,
      "mean_token_accuracy": 0.8708866596221924,
      "num_tokens": 19562318.0,
      "step": 1490
    },
    {
      "entropy": 4.309375,
      "epoch": 0.42134831460674155,
      "grad_norm": 1.2927882671356201,
      "learning_rate": 9.631498720206132e-06,
      "loss": 0.5130578517913819,
      "mean_token_accuracy": 0.8614315509796142,
      "num_tokens": 19693523.0,
      "step": 1500
    },
    {
      "epoch": 0.42134831460674155,
      "eval_entropy": 4.088467526690391,
      "eval_loss": 0.4993555247783661,
      "eval_mean_token_accuracy": 0.8653283246471364,
      "eval_num_tokens": 19693523.0,
      "eval_runtime": 128.1393,
      "eval_samples_per_second": 122.773,
      "eval_steps_per_second": 2.193,
      "step": 1500
    },
    {
      "entropy": 4.1609375,
      "epoch": 0.4241573033707865,
      "grad_norm": 1.6482603549957275,
      "learning_rate": 9.622206886777448e-06,
      "loss": 0.5003654003143311,
      "mean_token_accuracy": 0.8649938523769378,
      "num_tokens": 19824098.0,
      "step": 1510
    },
    {
      "entropy": 3.6796875,
      "epoch": 0.42696629213483145,
      "grad_norm": 1.7244060039520264,
      "learning_rate": 9.612803956113932e-06,
      "loss": 0.5100751876831054,
      "mean_token_accuracy": 0.8634325802326203,
      "num_tokens": 19955874.0,
      "step": 1520
    },
    {
      "entropy": 3.6984375,
      "epoch": 0.4297752808988764,
      "grad_norm": 1.5228095054626465,
      "learning_rate": 9.603290154220091e-06,
      "loss": 0.4991715431213379,
      "mean_token_accuracy": 0.8664758741855622,
      "num_tokens": 20087206.0,
      "step": 1530
    },
    {
      "entropy": 3.7421875,
      "epoch": 0.43258426966292135,
      "grad_norm": 1.4505484104156494,
      "learning_rate": 9.59366570976528e-06,
      "loss": 0.4892930030822754,
      "mean_token_accuracy": 0.8676511347293854,
      "num_tokens": 20219174.0,
      "step": 1540
    },
    {
      "entropy": 3.428125,
      "epoch": 0.4353932584269663,
      "grad_norm": 2.2225399017333984,
      "learning_rate": 9.583930854078219e-06,
      "loss": 0.4996492385864258,
      "mean_token_accuracy": 0.8656627058982849,
      "num_tokens": 20350430.0,
      "step": 1550
    },
    {
      "entropy": 3.415625,
      "epoch": 0.43820224719101125,
      "grad_norm": 3.6343190670013428,
      "learning_rate": 9.574085821141406e-06,
      "loss": 0.4848483562469482,
      "mean_token_accuracy": 0.8674822509288788,
      "num_tokens": 20481934.0,
      "step": 1560
    },
    {
      "entropy": 3.5421875,
      "epoch": 0.4410112359550562,
      "grad_norm": 1.6453512907028198,
      "learning_rate": 9.56413084758552e-06,
      "loss": 0.4888915061950684,
      "mean_token_accuracy": 0.8684564471244812,
      "num_tokens": 20613395.0,
      "step": 1570
    },
    {
      "entropy": 3.34375,
      "epoch": 0.4438202247191011,
      "grad_norm": 1.7411277294158936,
      "learning_rate": 9.554066172683715e-06,
      "loss": 0.47826318740844725,
      "mean_token_accuracy": 0.870430201292038,
      "num_tokens": 20745149.0,
      "step": 1580
    },
    {
      "entropy": 3.41875,
      "epoch": 0.44662921348314605,
      "grad_norm": 1.4324166774749756,
      "learning_rate": 9.543892038345885e-06,
      "loss": 0.4814493656158447,
      "mean_token_accuracy": 0.8692177832126617,
      "num_tokens": 20875371.0,
      "step": 1590
    },
    {
      "entropy": 3.6875,
      "epoch": 0.449438202247191,
      "grad_norm": 1.3732210397720337,
      "learning_rate": 9.533608689112827e-06,
      "loss": 0.49297313690185546,
      "mean_token_accuracy": 0.8670674860477448,
      "num_tokens": 21005850.0,
      "step": 1600
    },
    {
      "epoch": 0.449438202247191,
      "eval_entropy": 3.7251445729537367,
      "eval_loss": 0.495996356010437,
      "eval_mean_token_accuracy": 0.8657009179481833,
      "eval_num_tokens": 21005850.0,
      "eval_runtime": 114.9729,
      "eval_samples_per_second": 136.832,
      "eval_steps_per_second": 2.444,
      "step": 1600
    },
    {
      "entropy": 3.6953125,
      "epoch": 0.45224719101123595,
      "grad_norm": 1.517073631286621,
      "learning_rate": 9.523216372150393e-06,
      "loss": 0.5013265609741211,
      "mean_token_accuracy": 0.8644684553146362,
      "num_tokens": 21137058.0,
      "step": 1610
    },
    {
      "entropy": 3.5625,
      "epoch": 0.4550561797752809,
      "grad_norm": 1.5344774723052979,
      "learning_rate": 9.512715337243517e-06,
      "loss": 0.4923240661621094,
      "mean_token_accuracy": 0.8676745176315308,
      "num_tokens": 21267942.0,
      "step": 1620
    },
    {
      "entropy": 3.1921875,
      "epoch": 0.45786516853932585,
      "grad_norm": 1.4356811046600342,
      "learning_rate": 9.50210583679024e-06,
      "loss": 0.4946328639984131,
      "mean_token_accuracy": 0.867253977060318,
      "num_tokens": 21399637.0,
      "step": 1630
    },
    {
      "entropy": 4.1546875,
      "epoch": 0.4606741573033708,
      "grad_norm": 1.4015212059020996,
      "learning_rate": 9.491388125795623e-06,
      "loss": 0.495727014541626,
      "mean_token_accuracy": 0.8641788184642791,
      "num_tokens": 21531156.0,
      "step": 1640
    },
    {
      "entropy": 4.371875,
      "epoch": 0.46348314606741575,
      "grad_norm": 1.3862342834472656,
      "learning_rate": 9.480562461865634e-06,
      "loss": 0.5079226970672608,
      "mean_token_accuracy": 0.8627414643764496,
      "num_tokens": 21662373.0,
      "step": 1650
    },
    {
      "entropy": 4.646875,
      "epoch": 0.46629213483146065,
      "grad_norm": 1.2384926080703735,
      "learning_rate": 9.469629105200937e-06,
      "loss": 0.4975583553314209,
      "mean_token_accuracy": 0.8672258734703064,
      "num_tokens": 21794624.0,
      "step": 1660
    },
    {
      "entropy": 4.728125,
      "epoch": 0.4691011235955056,
      "grad_norm": 1.6070289611816406,
      "learning_rate": 9.458588318590659e-06,
      "loss": 0.4862710475921631,
      "mean_token_accuracy": 0.8686492443084717,
      "num_tokens": 21925953.0,
      "step": 1670
    },
    {
      "entropy": 4.896875,
      "epoch": 0.47191011235955055,
      "grad_norm": 1.3488452434539795,
      "learning_rate": 9.447440367406053e-06,
      "loss": 0.48831663131713865,
      "mean_token_accuracy": 0.8684999406337738,
      "num_tokens": 22056462.0,
      "step": 1680
    },
    {
      "entropy": 4.953125,
      "epoch": 0.4747191011235955,
      "grad_norm": 1.091750979423523,
      "learning_rate": 9.436185519594145e-06,
      "loss": 0.49806861877441405,
      "mean_token_accuracy": 0.8635454416275025,
      "num_tokens": 22188427.0,
      "step": 1690
    },
    {
      "entropy": 5.184375,
      "epoch": 0.47752808988764045,
      "grad_norm": 1.455783724784851,
      "learning_rate": 9.424824045671267e-06,
      "loss": 0.49456963539123533,
      "mean_token_accuracy": 0.8659671008586883,
      "num_tokens": 22319434.0,
      "step": 1700
    },
    {
      "epoch": 0.47752808988764045,
      "eval_entropy": 5.28480871886121,
      "eval_loss": 0.4923085570335388,
      "eval_mean_token_accuracy": 0.8664750089849017,
      "eval_num_tokens": 22319434.0,
      "eval_runtime": 123.6307,
      "eval_samples_per_second": 127.25,
      "eval_steps_per_second": 2.273,
      "step": 1700
    },
    {
      "entropy": 5.04375,
      "epoch": 0.4803370786516854,
      "grad_norm": 1.447575569152832,
      "learning_rate": 9.41335621871657e-06,
      "loss": 0.4862524032592773,
      "mean_token_accuracy": 0.8678418934345246,
      "num_tokens": 22450962.0,
      "step": 1710
    },
    {
      "entropy": 4.83125,
      "epoch": 0.48314606741573035,
      "grad_norm": 1.3019521236419678,
      "learning_rate": 9.401782314365458e-06,
      "loss": 0.487211799621582,
      "mean_token_accuracy": 0.8674911379814148,
      "num_tokens": 22581718.0,
      "step": 1720
    },
    {
      "entropy": 4.871875,
      "epoch": 0.4859550561797753,
      "grad_norm": 1.3167637586593628,
      "learning_rate": 9.390102610802965e-06,
      "loss": 0.49409937858581543,
      "mean_token_accuracy": 0.8652345299720764,
      "num_tokens": 22712049.0,
      "step": 1730
    },
    {
      "entropy": 4.575,
      "epoch": 0.4887640449438202,
      "grad_norm": 1.2791156768798828,
      "learning_rate": 9.378317388757062e-06,
      "loss": 0.4901265621185303,
      "mean_token_accuracy": 0.8667450785636902,
      "num_tokens": 22844964.0,
      "step": 1740
    },
    {
      "entropy": 4.884375,
      "epoch": 0.49157303370786515,
      "grad_norm": 1.2605441808700562,
      "learning_rate": 9.366426931491917e-06,
      "loss": 0.5005804061889648,
      "mean_token_accuracy": 0.8651033520698548,
      "num_tokens": 22976939.0,
      "step": 1750
    },
    {
      "entropy": 4.7,
      "epoch": 0.4943820224719101,
      "grad_norm": 1.3826125860214233,
      "learning_rate": 9.354431524801082e-06,
      "loss": 0.4904569149017334,
      "mean_token_accuracy": 0.8668628811836243,
      "num_tokens": 23108289.0,
      "step": 1760
    },
    {
      "entropy": 4.665625,
      "epoch": 0.49719101123595505,
      "grad_norm": 1.6024537086486816,
      "learning_rate": 9.342331457000621e-06,
      "loss": 0.4679577827453613,
      "mean_token_accuracy": 0.871170723438263,
      "num_tokens": 23238359.0,
      "step": 1770
    },
    {
      "entropy": 4.296875,
      "epoch": 0.5,
      "grad_norm": 1.4190478324890137,
      "learning_rate": 9.330127018922195e-06,
      "loss": 0.486340856552124,
      "mean_token_accuracy": 0.8664220333099365,
      "num_tokens": 23369855.0,
      "step": 1780
    },
    {
      "entropy": 3.9421875,
      "epoch": 0.5028089887640449,
      "grad_norm": 1.3500750064849854,
      "learning_rate": 9.317818503906046e-06,
      "loss": 0.4749727725982666,
      "mean_token_accuracy": 0.8714820444583893,
      "num_tokens": 23500855.0,
      "step": 1790
    },
    {
      "entropy": 4.15625,
      "epoch": 0.5056179775280899,
      "grad_norm": 1.397945761680603,
      "learning_rate": 9.305406207793974e-06,
      "loss": 0.47942180633544923,
      "mean_token_accuracy": 0.8720642805099488,
      "num_tokens": 23631830.0,
      "step": 1800
    },
    {
      "epoch": 0.5056179775280899,
      "eval_entropy": 4.0150133451957295,
      "eval_loss": 0.4904434084892273,
      "eval_mean_token_accuracy": 0.8667532192854694,
      "eval_num_tokens": 23631830.0,
      "eval_runtime": 119.7651,
      "eval_samples_per_second": 131.357,
      "eval_steps_per_second": 2.346,
      "step": 1800
    },
    {
      "entropy": 3.996875,
      "epoch": 0.5084269662921348,
      "grad_norm": 1.471629023551941,
      "learning_rate": 9.29289042892221e-06,
      "loss": 0.4605714797973633,
      "mean_token_accuracy": 0.8735283553600312,
      "num_tokens": 23763023.0,
      "step": 1810
    },
    {
      "entropy": 4.034375,
      "epoch": 0.5112359550561798,
      "grad_norm": 1.1965187788009644,
      "learning_rate": 9.280271468114243e-06,
      "loss": 0.4840362548828125,
      "mean_token_accuracy": 0.8688215911388397,
      "num_tokens": 23893964.0,
      "step": 1820
    },
    {
      "entropy": 4.153125,
      "epoch": 0.5140449438202247,
      "grad_norm": 1.2654021978378296,
      "learning_rate": 9.267549628673603e-06,
      "loss": 0.4879791259765625,
      "mean_token_accuracy": 0.8679976463317871,
      "num_tokens": 24025793.0,
      "step": 1830
    },
    {
      "entropy": 4.2375,
      "epoch": 0.5168539325842697,
      "grad_norm": 1.5571200847625732,
      "learning_rate": 9.254725216376562e-06,
      "loss": 0.49440751075744627,
      "mean_token_accuracy": 0.864652943611145,
      "num_tokens": 24155981.0,
      "step": 1840
    },
    {
      "entropy": 4.146875,
      "epoch": 0.5196629213483146,
      "grad_norm": 1.3446792364120483,
      "learning_rate": 9.24179853946478e-06,
      "loss": 0.48333940505981443,
      "mean_token_accuracy": 0.8702546775341033,
      "num_tokens": 24287592.0,
      "step": 1850
    },
    {
      "entropy": 3.76875,
      "epoch": 0.5224719101123596,
      "grad_norm": 1.2021961212158203,
      "learning_rate": 9.228769908637912e-06,
      "loss": 0.48885130882263184,
      "mean_token_accuracy": 0.866620659828186,
      "num_tokens": 24417844.0,
      "step": 1860
    },
    {
      "entropy": 3.9859375,
      "epoch": 0.5252808988764045,
      "grad_norm": 1.1995372772216797,
      "learning_rate": 9.215639637046121e-06,
      "loss": 0.4787414073944092,
      "mean_token_accuracy": 0.8684794425964355,
      "num_tokens": 24549241.0,
      "step": 1870
    },
    {
      "entropy": 4.065625,
      "epoch": 0.5280898876404494,
      "grad_norm": 1.3320667743682861,
      "learning_rate": 9.202408040282567e-06,
      "loss": 0.4829287528991699,
      "mean_token_accuracy": 0.8685443222522735,
      "num_tokens": 24680460.0,
      "step": 1880
    },
    {
      "entropy": 4.1125,
      "epoch": 0.5308988764044944,
      "grad_norm": 1.408900499343872,
      "learning_rate": 9.189075436375813e-06,
      "loss": 0.47942409515380857,
      "mean_token_accuracy": 0.8694293558597564,
      "num_tokens": 24812344.0,
      "step": 1890
    },
    {
      "entropy": 4.1203125,
      "epoch": 0.5337078651685393,
      "grad_norm": 1.2319111824035645,
      "learning_rate": 9.175642145782179e-06,
      "loss": 0.48389687538146975,
      "mean_token_accuracy": 0.8700027346611023,
      "num_tokens": 24943781.0,
      "step": 1900
    },
    {
      "epoch": 0.5337078651685393,
      "eval_entropy": 4.072842526690391,
      "eval_loss": 0.48570743203163147,
      "eval_mean_token_accuracy": 0.8677168030331568,
      "eval_num_tokens": 24943781.0,
      "eval_runtime": 122.7021,
      "eval_samples_per_second": 128.213,
      "eval_steps_per_second": 2.29,
      "step": 1900
    },
    {
      "entropy": 3.790625,
      "epoch": 0.5365168539325843,
      "grad_norm": 1.3118326663970947,
      "learning_rate": 9.162108491378051e-06,
      "loss": 0.4790935039520264,
      "mean_token_accuracy": 0.8694264650344848,
      "num_tokens": 25075534.0,
      "step": 1910
    },
    {
      "entropy": 3.9984375,
      "epoch": 0.5393258426966292,
      "grad_norm": 1.3163996934890747,
      "learning_rate": 9.14847479845211e-06,
      "loss": 0.49525160789489747,
      "mean_token_accuracy": 0.8655856966972351,
      "num_tokens": 25207621.0,
      "step": 1920
    },
    {
      "entropy": 4.0484375,
      "epoch": 0.5421348314606742,
      "grad_norm": 1.295131802558899,
      "learning_rate": 9.134741394697517e-06,
      "loss": 0.4814631462097168,
      "mean_token_accuracy": 0.8693764925003051,
      "num_tokens": 25339306.0,
      "step": 1930
    },
    {
      "entropy": 3.9953125,
      "epoch": 0.5449438202247191,
      "grad_norm": 1.2692621946334839,
      "learning_rate": 9.120908610204036e-06,
      "loss": 0.47943406105041503,
      "mean_token_accuracy": 0.8698448777198792,
      "num_tokens": 25470668.0,
      "step": 1940
    },
    {
      "entropy": 4.0265625,
      "epoch": 0.547752808988764,
      "grad_norm": 1.2564060688018799,
      "learning_rate": 9.106976777450099e-06,
      "loss": 0.486815357208252,
      "mean_token_accuracy": 0.869417405128479,
      "num_tokens": 25603268.0,
      "step": 1950
    },
    {
      "entropy": 4.15625,
      "epoch": 0.550561797752809,
      "grad_norm": 1.1735256910324097,
      "learning_rate": 9.09294623129482e-06,
      "loss": 0.4783041477203369,
      "mean_token_accuracy": 0.8697896301746368,
      "num_tokens": 25735320.0,
      "step": 1960
    },
    {
      "entropy": 4.009375,
      "epoch": 0.5533707865168539,
      "grad_norm": 1.0734282732009888,
      "learning_rate": 9.07881730896994e-06,
      "loss": 0.4707913398742676,
      "mean_token_accuracy": 0.8718658745288849,
      "num_tokens": 25866035.0,
      "step": 1970
    },
    {
      "entropy": 4.06875,
      "epoch": 0.5561797752808989,
      "grad_norm": 1.325309157371521,
      "learning_rate": 9.06459035007173e-06,
      "loss": 0.4948757171630859,
      "mean_token_accuracy": 0.86621213555336,
      "num_tokens": 25996404.0,
      "step": 1980
    },
    {
      "entropy": 3.971875,
      "epoch": 0.5589887640449438,
      "grad_norm": 1.739433765411377,
      "learning_rate": 9.05026569655281e-06,
      "loss": 0.4780719757080078,
      "mean_token_accuracy": 0.8703970730304718,
      "num_tokens": 26127036.0,
      "step": 1990
    },
    {
      "entropy": 3.8484375,
      "epoch": 0.5617977528089888,
      "grad_norm": 2.187211036682129,
      "learning_rate": 9.035843692713961e-06,
      "loss": 0.4848322868347168,
      "mean_token_accuracy": 0.867662924528122,
      "num_tokens": 26257988.0,
      "step": 2000
    },
    {
      "epoch": 0.5617977528089888,
      "eval_entropy": 3.929715302491103,
      "eval_loss": 0.4836612641811371,
      "eval_mean_token_accuracy": 0.8682058156597232,
      "eval_num_tokens": 26257988.0,
      "eval_runtime": 125.5774,
      "eval_samples_per_second": 125.277,
      "eval_steps_per_second": 2.238,
      "step": 2000
    },
    {
      "entropy": 3.9546875,
      "epoch": 0.5646067415730337,
      "grad_norm": 1.200858473777771,
      "learning_rate": 9.021324685195814e-06,
      "loss": 0.4831418991088867,
      "mean_token_accuracy": 0.8700924217700958,
      "num_tokens": 26388718.0,
      "step": 2010
    },
    {
      "entropy": 3.875,
      "epoch": 0.5674157303370787,
      "grad_norm": 1.1240787506103516,
      "learning_rate": 9.006709022970547e-06,
      "loss": 0.45309786796569823,
      "mean_token_accuracy": 0.8738633096218109,
      "num_tokens": 26519302.0,
      "step": 2020
    },
    {
      "entropy": 4.103125,
      "epoch": 0.5702247191011236,
      "grad_norm": 1.25180184841156,
      "learning_rate": 8.991997057333481e-06,
      "loss": 0.4809992790222168,
      "mean_token_accuracy": 0.868080347776413,
      "num_tokens": 26650848.0,
      "step": 2030
    },
    {
      "entropy": 4.1484375,
      "epoch": 0.5730337078651685,
      "grad_norm": 1.3207292556762695,
      "learning_rate": 8.977189141894645e-06,
      "loss": 0.49120941162109377,
      "mean_token_accuracy": 0.8668237507343293,
      "num_tokens": 26782030.0,
      "step": 2040
    },
    {
      "entropy": 3.9484375,
      "epoch": 0.5758426966292135,
      "grad_norm": 1.320738434791565,
      "learning_rate": 8.962285632570266e-06,
      "loss": 0.4718479156494141,
      "mean_token_accuracy": 0.8704148590564728,
      "num_tokens": 26912783.0,
      "step": 2050
    },
    {
      "entropy": 4.1125,
      "epoch": 0.5786516853932584,
      "grad_norm": 1.4663496017456055,
      "learning_rate": 8.947286887574234e-06,
      "loss": 0.47735843658447263,
      "mean_token_accuracy": 0.8698933899402619,
      "num_tokens": 27043323.0,
      "step": 2060
    },
    {
      "entropy": 4.0984375,
      "epoch": 0.5814606741573034,
      "grad_norm": 1.2799140214920044,
      "learning_rate": 8.932193267409465e-06,
      "loss": 0.4914725303649902,
      "mean_token_accuracy": 0.8674364864826203,
      "num_tokens": 27175479.0,
      "step": 2070
    },
    {
      "entropy": 3.959375,
      "epoch": 0.5842696629213483,
      "grad_norm": 1.2009328603744507,
      "learning_rate": 8.917005134859263e-06,
      "loss": 0.4733433723449707,
      "mean_token_accuracy": 0.8701773166656495,
      "num_tokens": 27307536.0,
      "step": 2080
    },
    {
      "entropy": 4.0296875,
      "epoch": 0.5870786516853933,
      "grad_norm": 1.452924370765686,
      "learning_rate": 8.901722854978582e-06,
      "loss": 0.48818349838256836,
      "mean_token_accuracy": 0.8682938396930695,
      "num_tokens": 27438753.0,
      "step": 2090
    },
    {
      "entropy": 4.35,
      "epoch": 0.5898876404494382,
      "grad_norm": 1.3548771142959595,
      "learning_rate": 8.88634679508526e-06,
      "loss": 0.47675609588623047,
      "mean_token_accuracy": 0.8688257813453675,
      "num_tokens": 27570893.0,
      "step": 2100
    },
    {
      "epoch": 0.5898876404494382,
      "eval_entropy": 4.2899243772241995,
      "eval_loss": 0.47956982254981995,
      "eval_mean_token_accuracy": 0.8688293947868075,
      "eval_num_tokens": 27570893.0,
      "eval_runtime": 121.16,
      "eval_samples_per_second": 129.845,
      "eval_steps_per_second": 2.319,
      "step": 2100
    },
    {
      "entropy": 3.8921875,
      "epoch": 0.5926966292134831,
      "grad_norm": 1.2893325090408325,
      "learning_rate": 8.870877324751186e-06,
      "loss": 0.4787613868713379,
      "mean_token_accuracy": 0.8689257264137268,
      "num_tokens": 27702830.0,
      "step": 2110
    },
    {
      "entropy": 3.734375,
      "epoch": 0.5955056179775281,
      "grad_norm": 1.2767736911773682,
      "learning_rate": 8.85531481579342e-06,
      "loss": 0.4879870891571045,
      "mean_token_accuracy": 0.8691526472568512,
      "num_tokens": 27834484.0,
      "step": 2120
    },
    {
      "entropy": 3.528125,
      "epoch": 0.598314606741573,
      "grad_norm": 1.3836997747421265,
      "learning_rate": 8.839659642265259e-06,
      "loss": 0.47429442405700684,
      "mean_token_accuracy": 0.871291595697403,
      "num_tokens": 27964820.0,
      "step": 2130
    },
    {
      "entropy": 3.671875,
      "epoch": 0.601123595505618,
      "grad_norm": 1.23726224899292,
      "learning_rate": 8.823912180447237e-06,
      "loss": 0.47287673950195314,
      "mean_token_accuracy": 0.8698099255561829,
      "num_tokens": 28095760.0,
      "step": 2140
    },
    {
      "entropy": 3.4890625,
      "epoch": 0.6039325842696629,
      "grad_norm": 1.2016059160232544,
      "learning_rate": 8.80807280883809e-06,
      "loss": 0.4708225250244141,
      "mean_token_accuracy": 0.8709520518779754,
      "num_tokens": 28226822.0,
      "step": 2150
    },
    {
      "entropy": 3.6421875,
      "epoch": 0.6067415730337079,
      "grad_norm": 1.393473744392395,
      "learning_rate": 8.79214190814566e-06,
      "loss": 0.46707677841186523,
      "mean_token_accuracy": 0.8710885047912598,
      "num_tokens": 28357063.0,
      "step": 2160
    },
    {
      "entropy": 3.7171875,
      "epoch": 0.6095505617977528,
      "grad_norm": 1.1839748620986938,
      "learning_rate": 8.77611986127773e-06,
      "loss": 0.47808380126953126,
      "mean_token_accuracy": 0.8690585494041443,
      "num_tokens": 28488340.0,
      "step": 2170
    },
    {
      "entropy": 3.946875,
      "epoch": 0.6123595505617978,
      "grad_norm": 1.4783397912979126,
      "learning_rate": 8.760007053332837e-06,
      "loss": 0.47326364517211916,
      "mean_token_accuracy": 0.8716509461402893,
      "num_tokens": 28620390.0,
      "step": 2180
    },
    {
      "entropy": 4.296875,
      "epoch": 0.6151685393258427,
      "grad_norm": 1.3443355560302734,
      "learning_rate": 8.743803871591008e-06,
      "loss": 0.49116950035095214,
      "mean_token_accuracy": 0.8678078830242157,
      "num_tokens": 28751398.0,
      "step": 2190
    },
    {
      "entropy": 4.234375,
      "epoch": 0.6179775280898876,
      "grad_norm": 1.0788482427597046,
      "learning_rate": 8.727510705504453e-06,
      "loss": 0.47329177856445315,
      "mean_token_accuracy": 0.8697935342788696,
      "num_tokens": 28882200.0,
      "step": 2200
    },
    {
      "epoch": 0.6179775280898876,
      "eval_entropy": 4.024577402135232,
      "eval_loss": 0.47718381881713867,
      "eval_mean_token_accuracy": 0.8694539014979189,
      "eval_num_tokens": 28882200.0,
      "eval_runtime": 127.1742,
      "eval_samples_per_second": 123.704,
      "eval_steps_per_second": 2.21,
      "step": 2200
    },
    {
      "entropy": 4.1734375,
      "epoch": 0.6207865168539326,
      "grad_norm": 1.0492788553237915,
      "learning_rate": 8.711127946688207e-06,
      "loss": 0.47619218826293946,
      "mean_token_accuracy": 0.8692793309688568,
      "num_tokens": 29013246.0,
      "step": 2210
    },
    {
      "entropy": 4.1421875,
      "epoch": 0.6235955056179775,
      "grad_norm": 1.2246413230895996,
      "learning_rate": 8.694655988910707e-06,
      "loss": 0.45670461654663086,
      "mean_token_accuracy": 0.8745538830757141,
      "num_tokens": 29144729.0,
      "step": 2220
    },
    {
      "entropy": 4.48125,
      "epoch": 0.6264044943820225,
      "grad_norm": 1.2382370233535767,
      "learning_rate": 8.678095228084343e-06,
      "loss": 0.466934061050415,
      "mean_token_accuracy": 0.8724118947982789,
      "num_tokens": 29276406.0,
      "step": 2230
    },
    {
      "entropy": 4.971875,
      "epoch": 0.6292134831460674,
      "grad_norm": 4.428104400634766,
      "learning_rate": 8.661446062255931e-06,
      "loss": 0.48734283447265625,
      "mean_token_accuracy": 0.8675418376922608,
      "num_tokens": 29407604.0,
      "step": 2240
    },
    {
      "entropy": 5.30625,
      "epoch": 0.6320224719101124,
      "grad_norm": 1.3471951484680176,
      "learning_rate": 8.644708891597147e-06,
      "loss": 0.48166232109069823,
      "mean_token_accuracy": 0.8676540613174438,
      "num_tokens": 29538003.0,
      "step": 2250
    },
    {
      "entropy": 4.840625,
      "epoch": 0.6348314606741573,
      "grad_norm": 2.8051068782806396,
      "learning_rate": 8.627884118394913e-06,
      "loss": 0.48446054458618165,
      "mean_token_accuracy": 0.8669973611831665,
      "num_tokens": 29669833.0,
      "step": 2260
    },
    {
      "entropy": 4.609375,
      "epoch": 0.6376404494382022,
      "grad_norm": 1.252308964729309,
      "learning_rate": 8.61097214704173e-06,
      "loss": 0.46909432411193847,
      "mean_token_accuracy": 0.8702910363674163,
      "num_tokens": 29801468.0,
      "step": 2270
    },
    {
      "entropy": 4.278125,
      "epoch": 0.6404494382022472,
      "grad_norm": 1.2747924327850342,
      "learning_rate": 8.59397338402594e-06,
      "loss": 0.45398459434509275,
      "mean_token_accuracy": 0.873817777633667,
      "num_tokens": 29932773.0,
      "step": 2280
    },
    {
      "entropy": 4.68125,
      "epoch": 0.6432584269662921,
      "grad_norm": 1.2493188381195068,
      "learning_rate": 8.576888237921983e-06,
      "loss": 0.4726550579071045,
      "mean_token_accuracy": 0.870222544670105,
      "num_tokens": 30063768.0,
      "step": 2290
    },
    {
      "entropy": 4.665625,
      "epoch": 0.6460674157303371,
      "grad_norm": 1.134830117225647,
      "learning_rate": 8.559717119380558e-06,
      "loss": 0.4640795230865479,
      "mean_token_accuracy": 0.8723108887672424,
      "num_tokens": 30195085.0,
      "step": 2300
    },
    {
      "epoch": 0.6460674157303371,
      "eval_entropy": 4.150689501779359,
      "eval_loss": 0.47507718205451965,
      "eval_mean_token_accuracy": 0.8698548914274711,
      "eval_num_tokens": 30195085.0,
      "eval_runtime": 123.8191,
      "eval_samples_per_second": 127.056,
      "eval_steps_per_second": 2.269,
      "step": 2300
    },
    {
      "entropy": 4.19375,
      "epoch": 0.648876404494382,
      "grad_norm": 1.0923227071762085,
      "learning_rate": 8.542460441118756e-06,
      "loss": 0.4910109519958496,
      "mean_token_accuracy": 0.8651893556118011,
      "num_tokens": 30326346.0,
      "step": 2310
    },
    {
      "entropy": 4.265625,
      "epoch": 0.651685393258427,
      "grad_norm": 1.4184261560440063,
      "learning_rate": 8.525118617910144e-06,
      "loss": 0.4746354579925537,
      "mean_token_accuracy": 0.8708074152469635,
      "num_tokens": 30457858.0,
      "step": 2320
    },
    {
      "entropy": 4.2,
      "epoch": 0.6544943820224719,
      "grad_norm": 1.2165019512176514,
      "learning_rate": 8.507692066574795e-06,
      "loss": 0.4689302444458008,
      "mean_token_accuracy": 0.8712755084037781,
      "num_tokens": 30589460.0,
      "step": 2330
    },
    {
      "entropy": 4.228125,
      "epoch": 0.6573033707865169,
      "grad_norm": 1.1056641340255737,
      "learning_rate": 8.490181205969268e-06,
      "loss": 0.47753081321716306,
      "mean_token_accuracy": 0.8684697270393371,
      "num_tokens": 30720816.0,
      "step": 2340
    },
    {
      "entropy": 4.325,
      "epoch": 0.6601123595505618,
      "grad_norm": 1.055174708366394,
      "learning_rate": 8.472586456976534e-06,
      "loss": 0.47216482162475587,
      "mean_token_accuracy": 0.8711139917373657,
      "num_tokens": 30852465.0,
      "step": 2350
    },
    {
      "entropy": 4.178125,
      "epoch": 0.6629213483146067,
      "grad_norm": 1.1944565773010254,
      "learning_rate": 8.45490824249588e-06,
      "loss": 0.48977861404418943,
      "mean_token_accuracy": 0.8679867804050445,
      "num_tokens": 30983836.0,
      "step": 2360
    },
    {
      "entropy": 4.26875,
      "epoch": 0.6657303370786517,
      "grad_norm": 1.291351556777954,
      "learning_rate": 8.437146987432717e-06,
      "loss": 0.46503362655639646,
      "mean_token_accuracy": 0.8736956715583801,
      "num_tokens": 31114856.0,
      "step": 2370
    },
    {
      "entropy": 4.128125,
      "epoch": 0.6685393258426966,
      "grad_norm": 1.3113031387329102,
      "learning_rate": 8.41930311868839e-06,
      "loss": 0.4699450969696045,
      "mean_token_accuracy": 0.8703964829444886,
      "num_tokens": 31245783.0,
      "step": 2380
    },
    {
      "entropy": 4.415625,
      "epoch": 0.6713483146067416,
      "grad_norm": 1.1390258073806763,
      "learning_rate": 8.401377065149904e-06,
      "loss": 0.47075634002685546,
      "mean_token_accuracy": 0.870576149225235,
      "num_tokens": 31377170.0,
      "step": 2390
    },
    {
      "entropy": 4.41875,
      "epoch": 0.6741573033707865,
      "grad_norm": 1.1703011989593506,
      "learning_rate": 8.383369257679625e-06,
      "loss": 0.47453603744506834,
      "mean_token_accuracy": 0.8693708777427673,
      "num_tokens": 31507639.0,
      "step": 2400
    },
    {
      "epoch": 0.6741573033707865,
      "eval_entropy": 4.299599644128114,
      "eval_loss": 0.4732455015182495,
      "eval_mean_token_accuracy": 0.8699907427156524,
      "eval_num_tokens": 31507639.0,
      "eval_runtime": 120.2822,
      "eval_samples_per_second": 130.792,
      "eval_steps_per_second": 2.336,
      "step": 2400
    },
    {
      "entropy": 4.26875,
      "epoch": 0.6769662921348315,
      "grad_norm": 1.1945265531539917,
      "learning_rate": 8.365280129104912e-06,
      "loss": 0.46162776947021483,
      "mean_token_accuracy": 0.8733875215053558,
      "num_tokens": 31639701.0,
      "step": 2410
    },
    {
      "entropy": 4.690625,
      "epoch": 0.6797752808988764,
      "grad_norm": 1.34599769115448,
      "learning_rate": 8.347110114207727e-06,
      "loss": 0.48328146934509275,
      "mean_token_accuracy": 0.8676354289054871,
      "num_tokens": 31769810.0,
      "step": 2420
    },
    {
      "entropy": 4.484375,
      "epoch": 0.6825842696629213,
      "grad_norm": 1.2303920984268188,
      "learning_rate": 8.328859649714171e-06,
      "loss": 0.4698744297027588,
      "mean_token_accuracy": 0.8717637300491333,
      "num_tokens": 31901228.0,
      "step": 2430
    },
    {
      "entropy": 4.5625,
      "epoch": 0.6853932584269663,
      "grad_norm": 1.103874683380127,
      "learning_rate": 8.310529174284004e-06,
      "loss": 0.4678653240203857,
      "mean_token_accuracy": 0.8708121716976166,
      "num_tokens": 32033751.0,
      "step": 2440
    },
    {
      "entropy": 4.44375,
      "epoch": 0.6882022471910112,
      "grad_norm": 1.3066987991333008,
      "learning_rate": 8.292119128500082e-06,
      "loss": 0.47205386161804197,
      "mean_token_accuracy": 0.8693049371242523,
      "num_tokens": 32164885.0,
      "step": 2450
    },
    {
      "entropy": 4.28125,
      "epoch": 0.6910112359550562,
      "grad_norm": 1.2790998220443726,
      "learning_rate": 8.273629954857784e-06,
      "loss": 0.48315629959106443,
      "mean_token_accuracy": 0.86684650182724,
      "num_tokens": 32296139.0,
      "step": 2460
    },
    {
      "entropy": 4.384375,
      "epoch": 0.6938202247191011,
      "grad_norm": 1.3642473220825195,
      "learning_rate": 8.255062097754371e-06,
      "loss": 0.4809610843658447,
      "mean_token_accuracy": 0.8681236147880554,
      "num_tokens": 32427819.0,
      "step": 2470
    },
    {
      "entropy": 4.534375,
      "epoch": 0.6966292134831461,
      "grad_norm": 1.4109039306640625,
      "learning_rate": 8.236416003478295e-06,
      "loss": 0.48529682159423826,
      "mean_token_accuracy": 0.8677134156227112,
      "num_tokens": 32558582.0,
      "step": 2480
    },
    {
      "entropy": 4.659375,
      "epoch": 0.699438202247191,
      "grad_norm": 1.2447803020477295,
      "learning_rate": 8.217692120198492e-06,
      "loss": 0.467620849609375,
      "mean_token_accuracy": 0.8708132028579711,
      "num_tokens": 32688432.0,
      "step": 2490
    },
    {
      "entropy": 4.6125,
      "epoch": 0.702247191011236,
      "grad_norm": 1.5587385892868042,
      "learning_rate": 8.198890897953586e-06,
      "loss": 0.4703376770019531,
      "mean_token_accuracy": 0.8705705344676972,
      "num_tokens": 32819410.0,
      "step": 2500
    },
    {
      "epoch": 0.702247191011236,
      "eval_entropy": 4.677824733096085,
      "eval_loss": 0.4707264006137848,
      "eval_mean_token_accuracy": 0.8707146095211395,
      "eval_num_tokens": 32819410.0,
      "eval_runtime": 123.7734,
      "eval_samples_per_second": 127.103,
      "eval_steps_per_second": 2.27,
      "step": 2500
    },
    {
      "entropy": 4.425,
      "epoch": 0.7050561797752809,
      "grad_norm": 1.2833225727081299,
      "learning_rate": 8.180012788641097e-06,
      "loss": 0.4701727867126465,
      "mean_token_accuracy": 0.8715618431568146,
      "num_tokens": 32951255.0,
      "step": 2510
    },
    {
      "entropy": 4.0359375,
      "epoch": 0.7078651685393258,
      "grad_norm": 1.2633178234100342,
      "learning_rate": 8.161058246006558e-06,
      "loss": 0.46942949295043945,
      "mean_token_accuracy": 0.8702243685722351,
      "num_tokens": 33083373.0,
      "step": 2520
    },
    {
      "entropy": 4.284375,
      "epoch": 0.7106741573033708,
      "grad_norm": 1.189842700958252,
      "learning_rate": 8.142027725632622e-06,
      "loss": 0.4767209529876709,
      "mean_token_accuracy": 0.8694493234157562,
      "num_tokens": 33214565.0,
      "step": 2530
    },
    {
      "entropy": 3.975,
      "epoch": 0.7134831460674157,
      "grad_norm": 1.1373180150985718,
      "learning_rate": 8.122921684928111e-06,
      "loss": 0.4647849082946777,
      "mean_token_accuracy": 0.8710277020931244,
      "num_tokens": 33346006.0,
      "step": 2540
    },
    {
      "entropy": 3.85625,
      "epoch": 0.7162921348314607,
      "grad_norm": 1.0913560390472412,
      "learning_rate": 8.10374058311701e-06,
      "loss": 0.4667331695556641,
      "mean_token_accuracy": 0.8727651000022888,
      "num_tokens": 33477661.0,
      "step": 2550
    },
    {
      "entropy": 3.6203125,
      "epoch": 0.7191011235955056,
      "grad_norm": 1.3130393028259277,
      "learning_rate": 8.084484881227449e-06,
      "loss": 0.44857158660888674,
      "mean_token_accuracy": 0.8778874993324279,
      "num_tokens": 33608757.0,
      "step": 2560
    },
    {
      "entropy": 3.6546875,
      "epoch": 0.7219101123595506,
      "grad_norm": 1.2712119817733765,
      "learning_rate": 8.065155042080599e-06,
      "loss": 0.4690615177154541,
      "mean_token_accuracy": 0.8708488643169403,
      "num_tokens": 33739342.0,
      "step": 2570
    },
    {
      "entropy": 3.965625,
      "epoch": 0.7247191011235955,
      "grad_norm": 1.214709758758545,
      "learning_rate": 8.04575153027957e-06,
      "loss": 0.4657096862792969,
      "mean_token_accuracy": 0.873617309331894,
      "num_tokens": 33869866.0,
      "step": 2580
    },
    {
      "entropy": 4.0171875,
      "epoch": 0.7275280898876404,
      "grad_norm": 1.2564371824264526,
      "learning_rate": 8.026274812198235e-06,
      "loss": 0.45439944267272947,
      "mean_token_accuracy": 0.8727246284484863,
      "num_tokens": 34001271.0,
      "step": 2590
    },
    {
      "entropy": 3.8140625,
      "epoch": 0.7303370786516854,
      "grad_norm": 1.2014402151107788,
      "learning_rate": 8.006725355970008e-06,
      "loss": 0.4674965858459473,
      "mean_token_accuracy": 0.8710847675800324,
      "num_tokens": 34132413.0,
      "step": 2600
    },
    {
      "epoch": 0.7303370786516854,
      "eval_entropy": 3.680771797153025,
      "eval_loss": 0.4692765176296234,
      "eval_mean_token_accuracy": 0.8708580399323188,
      "eval_num_tokens": 34132413.0,
      "eval_runtime": 120.9647,
      "eval_samples_per_second": 130.054,
      "eval_steps_per_second": 2.323,
      "step": 2600
    },
    {
      "entropy": 3.7484375,
      "epoch": 0.7331460674157303,
      "grad_norm": 1.3123708963394165,
      "learning_rate": 7.987103631476615e-06,
      "loss": 0.4625905990600586,
      "mean_token_accuracy": 0.8737405896186828,
      "num_tokens": 34264270.0,
      "step": 2610
    },
    {
      "entropy": 4.1203125,
      "epoch": 0.7359550561797753,
      "grad_norm": 1.1595304012298584,
      "learning_rate": 7.967410110336782e-06,
      "loss": 0.4644637584686279,
      "mean_token_accuracy": 0.8746164739131927,
      "num_tokens": 34395520.0,
      "step": 2620
    },
    {
      "entropy": 3.9859375,
      "epoch": 0.7387640449438202,
      "grad_norm": 1.3156406879425049,
      "learning_rate": 7.94764526589491e-06,
      "loss": 0.4682625770568848,
      "mean_token_accuracy": 0.8708840668201446,
      "num_tokens": 34527712.0,
      "step": 2630
    },
    {
      "entropy": 3.928125,
      "epoch": 0.7415730337078652,
      "grad_norm": 1.1634929180145264,
      "learning_rate": 7.927809573209691e-06,
      "loss": 0.4817975044250488,
      "mean_token_accuracy": 0.8680766820907593,
      "num_tokens": 34659354.0,
      "step": 2640
    },
    {
      "entropy": 3.6625,
      "epoch": 0.7443820224719101,
      "grad_norm": 1.2306469678878784,
      "learning_rate": 7.907903509042696e-06,
      "loss": 0.4635754585266113,
      "mean_token_accuracy": 0.8712450206279755,
      "num_tokens": 34790750.0,
      "step": 2650
    },
    {
      "entropy": 3.5140625,
      "epoch": 0.7471910112359551,
      "grad_norm": 1.2354871034622192,
      "learning_rate": 7.887927551846908e-06,
      "loss": 0.4670205593109131,
      "mean_token_accuracy": 0.8718623101711274,
      "num_tokens": 34922519.0,
      "step": 2660
    },
    {
      "entropy": 3.7125,
      "epoch": 0.75,
      "grad_norm": 1.2895396947860718,
      "learning_rate": 7.86788218175523e-06,
      "loss": 0.4775115966796875,
      "mean_token_accuracy": 0.866829764842987,
      "num_tokens": 35053853.0,
      "step": 2670
    },
    {
      "entropy": 3.7953125,
      "epoch": 0.7528089887640449,
      "grad_norm": 1.293368935585022,
      "learning_rate": 7.847767880568944e-06,
      "loss": 0.4802239894866943,
      "mean_token_accuracy": 0.8680545687675476,
      "num_tokens": 35184906.0,
      "step": 2680
    },
    {
      "entropy": 3.6921875,
      "epoch": 0.7556179775280899,
      "grad_norm": 1.164337158203125,
      "learning_rate": 7.827585131746122e-06,
      "loss": 0.4738199710845947,
      "mean_token_accuracy": 0.8692429184913635,
      "num_tokens": 35315954.0,
      "step": 2690
    },
    {
      "entropy": 4.190625,
      "epoch": 0.7584269662921348,
      "grad_norm": 1.144698143005371,
      "learning_rate": 7.807334420390014e-06,
      "loss": 0.4557361125946045,
      "mean_token_accuracy": 0.8753073394298554,
      "num_tokens": 35447943.0,
      "step": 2700
    },
    {
      "epoch": 0.7584269662921348,
      "eval_entropy": 4.8338523131672595,
      "eval_loss": 0.4654674828052521,
      "eval_mean_token_accuracy": 0.8718371107060714,
      "eval_num_tokens": 35447943.0,
      "eval_runtime": 111.703,
      "eval_samples_per_second": 140.838,
      "eval_steps_per_second": 2.516,
      "step": 2700
    },
    {
      "entropy": 4.896875,
      "epoch": 0.7612359550561798,
      "grad_norm": 1.08592689037323,
      "learning_rate": 7.787016233237387e-06,
      "loss": 0.45949759483337405,
      "mean_token_accuracy": 0.8727444887161255,
      "num_tokens": 35579976.0,
      "step": 2710
    },
    {
      "entropy": 4.8875,
      "epoch": 0.7640449438202247,
      "grad_norm": 1.1390734910964966,
      "learning_rate": 7.766631058646826e-06,
      "loss": 0.45903716087341306,
      "mean_token_accuracy": 0.872628915309906,
      "num_tokens": 35711403.0,
      "step": 2720
    },
    {
      "entropy": 4.74375,
      "epoch": 0.7668539325842697,
      "grad_norm": 1.0319364070892334,
      "learning_rate": 7.746179386586994e-06,
      "loss": 0.45185370445251466,
      "mean_token_accuracy": 0.8728707194328308,
      "num_tokens": 35842170.0,
      "step": 2730
    },
    {
      "entropy": 4.871875,
      "epoch": 0.7696629213483146,
      "grad_norm": 1.1375398635864258,
      "learning_rate": 7.725661708624855e-06,
      "loss": 0.458209228515625,
      "mean_token_accuracy": 0.8718686699867249,
      "num_tokens": 35973847.0,
      "step": 2740
    },
    {
      "entropy": 4.69375,
      "epoch": 0.7724719101123596,
      "grad_norm": 1.075753092765808,
      "learning_rate": 7.705078517913862e-06,
      "loss": 0.47183618545532224,
      "mean_token_accuracy": 0.8694608569145202,
      "num_tokens": 36105220.0,
      "step": 2750
    },
    {
      "entropy": 4.790625,
      "epoch": 0.7752808988764045,
      "grad_norm": 1.3887919187545776,
      "learning_rate": 7.684430309182106e-06,
      "loss": 0.46657438278198243,
      "mean_token_accuracy": 0.8733355402946472,
      "num_tokens": 36237008.0,
      "step": 2760
    },
    {
      "entropy": 4.734375,
      "epoch": 0.7780898876404494,
      "grad_norm": 1.3184854984283447,
      "learning_rate": 7.663717578720412e-06,
      "loss": 0.4694818496704102,
      "mean_token_accuracy": 0.8699236392974854,
      "num_tokens": 36368673.0,
      "step": 2770
    },
    {
      "entropy": 4.571875,
      "epoch": 0.7808988764044944,
      "grad_norm": 1.1244025230407715,
      "learning_rate": 7.642940824370429e-06,
      "loss": 0.45279731750488283,
      "mean_token_accuracy": 0.8733720481395721,
      "num_tokens": 36500193.0,
      "step": 2780
    },
    {
      "entropy": 4.69375,
      "epoch": 0.7837078651685393,
      "grad_norm": 1.3262803554534912,
      "learning_rate": 7.622100545512648e-06,
      "loss": 0.4873329162597656,
      "mean_token_accuracy": 0.8661531925201416,
      "num_tokens": 36631424.0,
      "step": 2790
    },
    {
      "entropy": 4.865625,
      "epoch": 0.7865168539325843,
      "grad_norm": 1.2294501066207886,
      "learning_rate": 7.601197243054411e-06,
      "loss": 0.47661170959472654,
      "mean_token_accuracy": 0.8690319657325745,
      "num_tokens": 36763045.0,
      "step": 2800
    },
    {
      "epoch": 0.7865168539325843,
      "eval_entropy": 4.975200177935943,
      "eval_loss": 0.4631503224372864,
      "eval_mean_token_accuracy": 0.8720892882007721,
      "eval_num_tokens": 36763045.0,
      "eval_runtime": 117.1913,
      "eval_samples_per_second": 134.242,
      "eval_steps_per_second": 2.398,
      "step": 2800
    },
    {
      "entropy": 4.975,
      "epoch": 0.7893258426966292,
      "grad_norm": 1.1538373231887817,
      "learning_rate": 7.580231419417863e-06,
      "loss": 0.4674933433532715,
      "mean_token_accuracy": 0.8721234500408173,
      "num_tokens": 36893659.0,
      "step": 2810
    },
    {
      "entropy": 4.934375,
      "epoch": 0.7921348314606742,
      "grad_norm": 1.4516780376434326,
      "learning_rate": 7.55920357852788e-06,
      "loss": 0.4778285503387451,
      "mean_token_accuracy": 0.8696352362632751,
      "num_tokens": 37025137.0,
      "step": 2820
    },
    {
      "entropy": 4.99375,
      "epoch": 0.7949438202247191,
      "grad_norm": 1.1424376964569092,
      "learning_rate": 7.538114225799955e-06,
      "loss": 0.48140640258789064,
      "mean_token_accuracy": 0.8667447626590729,
      "num_tokens": 37156721.0,
      "step": 2830
    },
    {
      "entropy": 4.88125,
      "epoch": 0.797752808988764,
      "grad_norm": 1.4852889776229858,
      "learning_rate": 7.516963868128054e-06,
      "loss": 0.4619006156921387,
      "mean_token_accuracy": 0.8709936201572418,
      "num_tokens": 37288223.0,
      "step": 2840
    },
    {
      "entropy": 4.6375,
      "epoch": 0.800561797752809,
      "grad_norm": 1.0840098857879639,
      "learning_rate": 7.4957530138724245e-06,
      "loss": 0.45699377059936525,
      "mean_token_accuracy": 0.8749772429466247,
      "num_tokens": 37420217.0,
      "step": 2850
    },
    {
      "entropy": 4.640625,
      "epoch": 0.8033707865168539,
      "grad_norm": 1.2344415187835693,
      "learning_rate": 7.474482172847391e-06,
      "loss": 0.45609078407287595,
      "mean_token_accuracy": 0.8717658758163452,
      "num_tokens": 37551289.0,
      "step": 2860
    },
    {
      "entropy": 4.546875,
      "epoch": 0.8061797752808989,
      "grad_norm": 1.2156758308410645,
      "learning_rate": 7.45315185630908e-06,
      "loss": 0.4577050685882568,
      "mean_token_accuracy": 0.8721380174160004,
      "num_tokens": 37683842.0,
      "step": 2870
    },
    {
      "entropy": 4.625,
      "epoch": 0.8089887640449438,
      "grad_norm": 2.164560317993164,
      "learning_rate": 7.431762576943157e-06,
      "loss": 0.4695699691772461,
      "mean_token_accuracy": 0.8730245769023895,
      "num_tokens": 37816517.0,
      "step": 2880
    },
    {
      "entropy": 4.940625,
      "epoch": 0.8117977528089888,
      "grad_norm": 1.2415045499801636,
      "learning_rate": 7.4103148488524824e-06,
      "loss": 0.46599950790405276,
      "mean_token_accuracy": 0.8713643908500671,
      "num_tokens": 37947399.0,
      "step": 2890
    },
    {
      "entropy": 4.975,
      "epoch": 0.8146067415730337,
      "grad_norm": 1.27390718460083,
      "learning_rate": 7.388809187544764e-06,
      "loss": 0.46910982131958007,
      "mean_token_accuracy": 0.8687225997447967,
      "num_tokens": 38079477.0,
      "step": 2900
    },
    {
      "epoch": 0.8146067415730337,
      "eval_entropy": 4.905026690391459,
      "eval_loss": 0.4616650938987732,
      "eval_mean_token_accuracy": 0.8726058135677487,
      "eval_num_tokens": 38079477.0,
      "eval_runtime": 118.288,
      "eval_samples_per_second": 132.997,
      "eval_steps_per_second": 2.376,
      "step": 2900
    },
    {
      "entropy": 4.659375,
      "epoch": 0.8174157303370787,
      "grad_norm": 1.3528387546539307,
      "learning_rate": 7.367246109920171e-06,
      "loss": 0.4450042724609375,
      "mean_token_accuracy": 0.8755173563957215,
      "num_tokens": 38210375.0,
      "step": 2910
    },
    {
      "entropy": 4.775,
      "epoch": 0.8202247191011236,
      "grad_norm": 1.1432626247406006,
      "learning_rate": 7.345626134258897e-06,
      "loss": 0.4884047508239746,
      "mean_token_accuracy": 0.8652958035469055,
      "num_tokens": 38341942.0,
      "step": 2920
    },
    {
      "entropy": 4.790625,
      "epoch": 0.8230337078651685,
      "grad_norm": 1.2599807977676392,
      "learning_rate": 7.323949780208717e-06,
      "loss": 0.4449580669403076,
      "mean_token_accuracy": 0.8752652704715729,
      "num_tokens": 38473247.0,
      "step": 2930
    },
    {
      "entropy": 4.784375,
      "epoch": 0.8258426966292135,
      "grad_norm": 1.1668078899383545,
      "learning_rate": 7.302217568772488e-06,
      "loss": 0.46388983726501465,
      "mean_token_accuracy": 0.8728438735008239,
      "num_tokens": 38604720.0,
      "step": 2940
    },
    {
      "entropy": 4.8,
      "epoch": 0.8286516853932584,
      "grad_norm": 1.319612741470337,
      "learning_rate": 7.28043002229563e-06,
      "loss": 0.45819454193115233,
      "mean_token_accuracy": 0.8734262466430665,
      "num_tokens": 38736289.0,
      "step": 2950
    },
    {
      "entropy": 4.6875,
      "epoch": 0.8314606741573034,
      "grad_norm": 1.3027217388153076,
      "learning_rate": 7.2585876644535705e-06,
      "loss": 0.44776058197021484,
      "mean_token_accuracy": 0.8765788912773133,
      "num_tokens": 38867232.0,
      "step": 2960
    },
    {
      "entropy": 4.509375,
      "epoch": 0.8342696629213483,
      "grad_norm": 1.195500135421753,
      "learning_rate": 7.236691020239157e-06,
      "loss": 0.4474950313568115,
      "mean_token_accuracy": 0.874479067325592,
      "num_tokens": 38998719.0,
      "step": 2970
    },
    {
      "entropy": 4.58125,
      "epoch": 0.8370786516853933,
      "grad_norm": 1.3842933177947998,
      "learning_rate": 7.214740615950041e-06,
      "loss": 0.44739570617675783,
      "mean_token_accuracy": 0.8759231686592102,
      "num_tokens": 39129324.0,
      "step": 2980
    },
    {
      "entropy": 4.6125,
      "epoch": 0.8398876404494382,
      "grad_norm": 1.178361415863037,
      "learning_rate": 7.192736979176025e-06,
      "loss": 0.46214685440063474,
      "mean_token_accuracy": 0.8718967854976654,
      "num_tokens": 39261280.0,
      "step": 2990
    },
    {
      "entropy": 4.590625,
      "epoch": 0.8426966292134831,
      "grad_norm": 1.1464803218841553,
      "learning_rate": 7.170680638786383e-06,
      "loss": 0.446087121963501,
      "mean_token_accuracy": 0.8766535460948944,
      "num_tokens": 39393139.0,
      "step": 3000
    },
    {
      "epoch": 0.8426966292134831,
      "eval_entropy": 4.737099644128114,
      "eval_loss": 0.45965006947517395,
      "eval_mean_token_accuracy": 0.8729924797163315,
      "eval_num_tokens": 39393139.0,
      "eval_runtime": 118.1885,
      "eval_samples_per_second": 133.109,
      "eval_steps_per_second": 2.378,
      "step": 3000
    },
    {
      "entropy": 4.43125,
      "epoch": 0.8455056179775281,
      "grad_norm": 1.4188731908798218,
      "learning_rate": 7.148572124917148e-06,
      "loss": 0.4457691192626953,
      "mean_token_accuracy": 0.8770259976387024,
      "num_tokens": 39524758.0,
      "step": 3010
    },
    {
      "entropy": 4.39375,
      "epoch": 0.848314606741573,
      "grad_norm": 1.284428358078003,
      "learning_rate": 7.126411968958374e-06,
      "loss": 0.45645918846130373,
      "mean_token_accuracy": 0.8727712094783783,
      "num_tokens": 39655324.0,
      "step": 3020
    },
    {
      "entropy": 4.290625,
      "epoch": 0.851123595505618,
      "grad_norm": 1.1207385063171387,
      "learning_rate": 7.104200703541358e-06,
      "loss": 0.4533103942871094,
      "mean_token_accuracy": 0.872863644361496,
      "num_tokens": 39786808.0,
      "step": 3030
    },
    {
      "entropy": 4.096875,
      "epoch": 0.8539325842696629,
      "grad_norm": 1.2317564487457275,
      "learning_rate": 7.0819388625258385e-06,
      "loss": 0.46272802352905273,
      "mean_token_accuracy": 0.8731277465820313,
      "num_tokens": 39918818.0,
      "step": 3040
    },
    {
      "entropy": 4.375,
      "epoch": 0.8567415730337079,
      "grad_norm": 1.6724387407302856,
      "learning_rate": 7.059626980987172e-06,
      "loss": 0.45761966705322266,
      "mean_token_accuracy": 0.8715218722820282,
      "num_tokens": 40049308.0,
      "step": 3050
    },
    {
      "entropy": 4.3125,
      "epoch": 0.8595505617977528,
      "grad_norm": 1.360580325126648,
      "learning_rate": 7.0372655952034575e-06,
      "loss": 0.46041140556335447,
      "mean_token_accuracy": 0.8729157269001007,
      "num_tokens": 40180466.0,
      "step": 3060
    },
    {
      "entropy": 4.440625,
      "epoch": 0.8623595505617978,
      "grad_norm": 1.3175475597381592,
      "learning_rate": 7.014855242642662e-06,
      "loss": 0.45488624572753905,
      "mean_token_accuracy": 0.8733796298503875,
      "num_tokens": 40310690.0,
      "step": 3070
    },
    {
      "entropy": 4.328125,
      "epoch": 0.8651685393258427,
      "grad_norm": 1.2995195388793945,
      "learning_rate": 6.992396461949693e-06,
      "loss": 0.43857574462890625,
      "mean_token_accuracy": 0.8771652042865753,
      "num_tokens": 40442483.0,
      "step": 3080
    },
    {
      "entropy": 4.471875,
      "epoch": 0.8679775280898876,
      "grad_norm": 1.041253685951233,
      "learning_rate": 6.969889792933454e-06,
      "loss": 0.4632700443267822,
      "mean_token_accuracy": 0.8721619069576263,
      "num_tokens": 40574053.0,
      "step": 3090
    },
    {
      "entropy": 4.725,
      "epoch": 0.8707865168539326,
      "grad_norm": 1.3057990074157715,
      "learning_rate": 6.94733577655387e-06,
      "loss": 0.4591245651245117,
      "mean_token_accuracy": 0.8736916005611419,
      "num_tokens": 40705079.0,
      "step": 3100
    },
    {
      "epoch": 0.8707865168539326,
      "eval_entropy": 4.786032028469751,
      "eval_loss": 0.45755231380462646,
      "eval_mean_token_accuracy": 0.8733045995447559,
      "eval_num_tokens": 40705079.0,
      "eval_runtime": 117.5102,
      "eval_samples_per_second": 133.878,
      "eval_steps_per_second": 2.391,
      "step": 3100
    },
    {
      "entropy": 4.821875,
      "epoch": 0.8735955056179775,
      "grad_norm": 1.248618721961975,
      "learning_rate": 6.924734954908887e-06,
      "loss": 0.4721823692321777,
      "mean_token_accuracy": 0.8694739580154419,
      "num_tokens": 40836996.0,
      "step": 3110
    },
    {
      "entropy": 4.55,
      "epoch": 0.8764044943820225,
      "grad_norm": 1.0796260833740234,
      "learning_rate": 6.902087871221439e-06,
      "loss": 0.4529205322265625,
      "mean_token_accuracy": 0.873122262954712,
      "num_tokens": 40969051.0,
      "step": 3120
    },
    {
      "entropy": 4.253125,
      "epoch": 0.8792134831460674,
      "grad_norm": 1.3218443393707275,
      "learning_rate": 6.879395069826394e-06,
      "loss": 0.4453711986541748,
      "mean_token_accuracy": 0.8770368218421936,
      "num_tokens": 41100590.0,
      "step": 3130
    },
    {
      "entropy": 4.225,
      "epoch": 0.8820224719101124,
      "grad_norm": 1.18660306930542,
      "learning_rate": 6.856657096157469e-06,
      "loss": 0.4539553165435791,
      "mean_token_accuracy": 0.8736288547515869,
      "num_tokens": 41231574.0,
      "step": 3140
    },
    {
      "entropy": 4.4,
      "epoch": 0.8848314606741573,
      "grad_norm": 1.1301614046096802,
      "learning_rate": 6.833874496734122e-06,
      "loss": 0.44768557548522947,
      "mean_token_accuracy": 0.8753218710422516,
      "num_tokens": 41361314.0,
      "step": 3150
    },
    {
      "entropy": 4.40625,
      "epoch": 0.8876404494382022,
      "grad_norm": 1.1133365631103516,
      "learning_rate": 6.811047819148413e-06,
      "loss": 0.43836302757263185,
      "mean_token_accuracy": 0.8791981160640716,
      "num_tokens": 41492036.0,
      "step": 3160
    },
    {
      "entropy": 4.459375,
      "epoch": 0.8904494382022472,
      "grad_norm": 1.1792795658111572,
      "learning_rate": 6.78817761205185e-06,
      "loss": 0.44870815277099607,
      "mean_token_accuracy": 0.8743037939071655,
      "num_tokens": 41622600.0,
      "step": 3170
    },
    {
      "entropy": 4.634375,
      "epoch": 0.8932584269662921,
      "grad_norm": 1.2515537738800049,
      "learning_rate": 6.7652644251421875e-06,
      "loss": 0.44904842376708987,
      "mean_token_accuracy": 0.8767969906330109,
      "num_tokens": 41753824.0,
      "step": 3180
    },
    {
      "entropy": 4.671875,
      "epoch": 0.8960674157303371,
      "grad_norm": 1.1964305639266968,
      "learning_rate": 6.742308809150232e-06,
      "loss": 0.46018013954162595,
      "mean_token_accuracy": 0.8749961256980896,
      "num_tokens": 41885058.0,
      "step": 3190
    },
    {
      "entropy": 4.7375,
      "epoch": 0.898876404494382,
      "grad_norm": 1.1080882549285889,
      "learning_rate": 6.719311315826589e-06,
      "loss": 0.47750205993652345,
      "mean_token_accuracy": 0.870861005783081,
      "num_tokens": 42017164.0,
      "step": 3200
    },
    {
      "epoch": 0.898876404494382,
      "eval_entropy": 4.715858540925267,
      "eval_loss": 0.456572949886322,
      "eval_mean_token_accuracy": 0.8734287447776659,
      "eval_num_tokens": 42017164.0,
      "eval_runtime": 121.1419,
      "eval_samples_per_second": 129.864,
      "eval_steps_per_second": 2.32,
      "step": 3200
    },
    {
      "entropy": 4.6125,
      "epoch": 0.901685393258427,
      "grad_norm": 1.1153473854064941,
      "learning_rate": 6.696272497928411e-06,
      "loss": 0.4418421745300293,
      "mean_token_accuracy": 0.8759481072425842,
      "num_tokens": 42147206.0,
      "step": 3210
    },
    {
      "entropy": 4.578125,
      "epoch": 0.9044943820224719,
      "grad_norm": 1.4467153549194336,
      "learning_rate": 6.673192909206109e-06,
      "loss": 0.47050909996032714,
      "mean_token_accuracy": 0.8722999155521393,
      "num_tokens": 42279333.0,
      "step": 3220
    },
    {
      "entropy": 4.665625,
      "epoch": 0.9073033707865169,
      "grad_norm": 1.0181089639663696,
      "learning_rate": 6.6500731043900425e-06,
      "loss": 0.44898009300231934,
      "mean_token_accuracy": 0.8750045657157898,
      "num_tokens": 42410288.0,
      "step": 3230
    },
    {
      "entropy": 4.709375,
      "epoch": 0.9101123595505618,
      "grad_norm": 1.161218523979187,
      "learning_rate": 6.626913639177189e-06,
      "loss": 0.45329575538635253,
      "mean_token_accuracy": 0.8738721311092377,
      "num_tokens": 42541166.0,
      "step": 3240
    },
    {
      "entropy": 4.834375,
      "epoch": 0.9129213483146067,
      "grad_norm": 1.189934492111206,
      "learning_rate": 6.603715070217779e-06,
      "loss": 0.44208469390869143,
      "mean_token_accuracy": 0.876932692527771,
      "num_tokens": 42673652.0,
      "step": 3250
    },
    {
      "entropy": 4.84375,
      "epoch": 0.9157303370786517,
      "grad_norm": 1.1587965488433838,
      "learning_rate": 6.580477955101927e-06,
      "loss": 0.46686840057373047,
      "mean_token_accuracy": 0.8699420154094696,
      "num_tokens": 42805128.0,
      "step": 3260
    },
    {
      "entropy": 4.971875,
      "epoch": 0.9185393258426966,
      "grad_norm": 1.189774513244629,
      "learning_rate": 6.5572028523462275e-06,
      "loss": 0.45156116485595704,
      "mean_token_accuracy": 0.8750099956989288,
      "num_tokens": 42936825.0,
      "step": 3270
    },
    {
      "entropy": 4.975,
      "epoch": 0.9213483146067416,
      "grad_norm": 1.1389515399932861,
      "learning_rate": 6.53389032138032e-06,
      "loss": 0.4708730697631836,
      "mean_token_accuracy": 0.869329184293747,
      "num_tokens": 43068414.0,
      "step": 3280
    },
    {
      "entropy": 4.98125,
      "epoch": 0.9241573033707865,
      "grad_norm": 1.242422342300415,
      "learning_rate": 6.51054092253346e-06,
      "loss": 0.474805212020874,
      "mean_token_accuracy": 0.8698181211948395,
      "num_tokens": 43199646.0,
      "step": 3290
    },
    {
      "entropy": 4.896875,
      "epoch": 0.9269662921348315,
      "grad_norm": 1.3307466506958008,
      "learning_rate": 6.487155217021039e-06,
      "loss": 0.4436355113983154,
      "mean_token_accuracy": 0.8753653168678284,
      "num_tokens": 43330276.0,
      "step": 3300
    },
    {
      "epoch": 0.9269662921348315,
      "eval_entropy": 4.768572064056939,
      "eval_loss": 0.4548853039741516,
      "eval_mean_token_accuracy": 0.8738842441093879,
      "eval_num_tokens": 43330276.0,
      "eval_runtime": 113.9555,
      "eval_samples_per_second": 138.054,
      "eval_steps_per_second": 2.466,
      "step": 3300
    },
    {
      "entropy": 4.68125,
      "epoch": 0.9297752808988764,
      "grad_norm": 1.2893747091293335,
      "learning_rate": 6.463733766931096e-06,
      "loss": 0.4596144676208496,
      "mean_token_accuracy": 0.8708854496479035,
      "num_tokens": 43462242.0,
      "step": 3310
    },
    {
      "entropy": 4.784375,
      "epoch": 0.9325842696629213,
      "grad_norm": 1.1250450611114502,
      "learning_rate": 6.440277135210815e-06,
      "loss": 0.46097331047058104,
      "mean_token_accuracy": 0.8723002552986145,
      "num_tokens": 43592832.0,
      "step": 3320
    },
    {
      "entropy": 5.003125,
      "epoch": 0.9353932584269663,
      "grad_norm": 1.158905029296875,
      "learning_rate": 6.4167858856529875e-06,
      "loss": 0.4451141834259033,
      "mean_token_accuracy": 0.8782232224941253,
      "num_tokens": 43724027.0,
      "step": 3330
    },
    {
      "entropy": 5.11875,
      "epoch": 0.9382022471910112,
      "grad_norm": 1.3706824779510498,
      "learning_rate": 6.393260582882462e-06,
      "loss": 0.4438319206237793,
      "mean_token_accuracy": 0.8772729396820068,
      "num_tokens": 43855356.0,
      "step": 3340
    },
    {
      "entropy": 5.215625,
      "epoch": 0.9410112359550562,
      "grad_norm": 1.3636852502822876,
      "learning_rate": 6.369701792342576e-06,
      "loss": 0.4574113368988037,
      "mean_token_accuracy": 0.8746276140213013,
      "num_tokens": 43986930.0,
      "step": 3350
    },
    {
      "entropy": 5.175,
      "epoch": 0.9438202247191011,
      "grad_norm": 2.2288269996643066,
      "learning_rate": 6.3461100802815625e-06,
      "loss": 0.45460071563720705,
      "mean_token_accuracy": 0.8740879535675049,
      "num_tokens": 44119176.0,
      "step": 3360
    },
    {
      "entropy": 5.225,
      "epoch": 0.9466292134831461,
      "grad_norm": 1.121077299118042,
      "learning_rate": 6.322486013738942e-06,
      "loss": 0.4636176109313965,
      "mean_token_accuracy": 0.8713511109352112,
      "num_tokens": 44251078.0,
      "step": 3370
    },
    {
      "entropy": 5.209375,
      "epoch": 0.949438202247191,
      "grad_norm": 1.083916187286377,
      "learning_rate": 6.298830160531895e-06,
      "loss": 0.4237912654876709,
      "mean_token_accuracy": 0.8818411886692047,
      "num_tokens": 44382716.0,
      "step": 3380
    },
    {
      "entropy": 5.1,
      "epoch": 0.952247191011236,
      "grad_norm": 1.0926802158355713,
      "learning_rate": 6.275143089241603e-06,
      "loss": 0.4595186710357666,
      "mean_token_accuracy": 0.8746098279953003,
      "num_tokens": 44514902.0,
      "step": 3390
    },
    {
      "entropy": 5.253125,
      "epoch": 0.9550561797752809,
      "grad_norm": 1.2092517614364624,
      "learning_rate": 6.2514253691996e-06,
      "loss": 0.4422555923461914,
      "mean_token_accuracy": 0.8772244393825531,
      "num_tokens": 44646338.0,
      "step": 3400
    },
    {
      "epoch": 0.9550561797752809,
      "eval_entropy": 5.254670818505338,
      "eval_loss": 0.453759104013443,
      "eval_mean_token_accuracy": 0.8738977898482326,
      "eval_num_tokens": 44646338.0,
      "eval_runtime": 117.0497,
      "eval_samples_per_second": 134.404,
      "eval_steps_per_second": 2.401,
      "step": 3400
    },
    {
      "entropy": 5.178125,
      "epoch": 0.9578651685393258,
      "grad_norm": 1.131625771522522,
      "learning_rate": 6.227677570474077e-06,
      "loss": 0.4520420551300049,
      "mean_token_accuracy": 0.8743519842624664,
      "num_tokens": 44778110.0,
      "step": 3410
    },
    {
      "entropy": 4.796875,
      "epoch": 0.9606741573033708,
      "grad_norm": 1.2695562839508057,
      "learning_rate": 6.203900263856177e-06,
      "loss": 0.4519766330718994,
      "mean_token_accuracy": 0.8747868776321411,
      "num_tokens": 44909874.0,
      "step": 3420
    },
    {
      "entropy": 4.85625,
      "epoch": 0.9634831460674157,
      "grad_norm": 1.0476394891738892,
      "learning_rate": 6.180094020846291e-06,
      "loss": 0.44031801223754885,
      "mean_token_accuracy": 0.8773936033248901,
      "num_tokens": 45040932.0,
      "step": 3430
    },
    {
      "entropy": 4.784375,
      "epoch": 0.9662921348314607,
      "grad_norm": 1.2428944110870361,
      "learning_rate": 6.156259413640302e-06,
      "loss": 0.4616535186767578,
      "mean_token_accuracy": 0.8734331488609314,
      "num_tokens": 45173441.0,
      "step": 3440
    },
    {
      "entropy": 4.790625,
      "epoch": 0.9691011235955056,
      "grad_norm": 1.350412368774414,
      "learning_rate": 6.132397015115846e-06,
      "loss": 0.45598635673522947,
      "mean_token_accuracy": 0.8745548248291015,
      "num_tokens": 45303640.0,
      "step": 3450
    },
    {
      "entropy": 4.75625,
      "epoch": 0.9719101123595506,
      "grad_norm": 1.0514638423919678,
      "learning_rate": 6.10850739881854e-06,
      "loss": 0.4448367118835449,
      "mean_token_accuracy": 0.8775962054729461,
      "num_tokens": 45434236.0,
      "step": 3460
    },
    {
      "entropy": 4.83125,
      "epoch": 0.9747191011235955,
      "grad_norm": 1.1635887622833252,
      "learning_rate": 6.084591138948192e-06,
      "loss": 0.4457675933837891,
      "mean_token_accuracy": 0.8755600154399872,
      "num_tokens": 45565825.0,
      "step": 3470
    },
    {
      "entropy": 4.653125,
      "epoch": 0.9775280898876404,
      "grad_norm": 1.2397221326828003,
      "learning_rate": 6.060648810345006e-06,
      "loss": 0.4476353645324707,
      "mean_token_accuracy": 0.8757467627525329,
      "num_tokens": 45697153.0,
      "step": 3480
    },
    {
      "entropy": 4.915625,
      "epoch": 0.9803370786516854,
      "grad_norm": 0.9696674346923828,
      "learning_rate": 6.036680988475756e-06,
      "loss": 0.44799246788024905,
      "mean_token_accuracy": 0.8759482622146606,
      "num_tokens": 45828460.0,
      "step": 3490
    },
    {
      "entropy": 4.93125,
      "epoch": 0.9831460674157303,
      "grad_norm": 1.1220450401306152,
      "learning_rate": 6.012688249419966e-06,
      "loss": 0.44777488708496094,
      "mean_token_accuracy": 0.8757860362529755,
      "num_tokens": 45959799.0,
      "step": 3500
    },
    {
      "epoch": 0.9831460674157303,
      "eval_entropy": 4.858540925266904,
      "eval_loss": 0.4512580335140228,
      "eval_mean_token_accuracy": 0.8745333940533132,
      "eval_num_tokens": 45959799.0,
      "eval_runtime": 116.5426,
      "eval_samples_per_second": 134.989,
      "eval_steps_per_second": 2.411,
      "step": 3500
    },
    {
      "entropy": 4.8125,
      "epoch": 0.9859550561797753,
      "grad_norm": 1.1713377237319946,
      "learning_rate": 5.988671169856056e-06,
      "loss": 0.4368232250213623,
      "mean_token_accuracy": 0.8777824997901916,
      "num_tokens": 46090761.0,
      "step": 3510
    },
    {
      "entropy": 4.696875,
      "epoch": 0.9887640449438202,
      "grad_norm": 1.0816149711608887,
      "learning_rate": 5.964630327047485e-06,
      "loss": 0.44577460289001464,
      "mean_token_accuracy": 0.8757070600986481,
      "num_tokens": 46221018.0,
      "step": 3520
    },
    {
      "entropy": 4.85,
      "epoch": 0.9915730337078652,
      "grad_norm": 1.224841833114624,
      "learning_rate": 5.940566298828871e-06,
      "loss": 0.43196535110473633,
      "mean_token_accuracy": 0.878901869058609,
      "num_tokens": 46352314.0,
      "step": 3530
    },
    {
      "entropy": 4.746875,
      "epoch": 0.9943820224719101,
      "grad_norm": 1.7760214805603027,
      "learning_rate": 5.916479663592107e-06,
      "loss": 0.45603384971618655,
      "mean_token_accuracy": 0.8730259537696838,
      "num_tokens": 46484752.0,
      "step": 3540
    },
    {
      "entropy": 4.759375,
      "epoch": 0.9971910112359551,
      "grad_norm": 1.2012728452682495,
      "learning_rate": 5.8923710002724595e-06,
      "loss": 0.44951772689819336,
      "mean_token_accuracy": 0.8754217982292175,
      "num_tokens": 46616171.0,
      "step": 3550
    },
    {
      "entropy": 4.8,
      "epoch": 1.0,
      "grad_norm": 1.0834134817123413,
      "learning_rate": 5.8682408883346535e-06,
      "loss": 0.4463536739349365,
      "mean_token_accuracy": 0.8752633213996888,
      "num_tokens": 46748171.0,
      "step": 3560
    },
    {
      "entropy": 4.628125,
      "epoch": 1.002808988764045,
      "grad_norm": 1.2700437307357788,
      "learning_rate": 5.844089907758935e-06,
      "loss": 0.3979910373687744,
      "mean_token_accuracy": 0.8852331697940826,
      "num_tokens": 46879319.0,
      "step": 3570
    },
    {
      "entropy": 4.24375,
      "epoch": 1.0056179775280898,
      "grad_norm": 1.1215442419052124,
      "learning_rate": 5.819918639027149e-06,
      "loss": 0.4029043674468994,
      "mean_token_accuracy": 0.882667738199234,
      "num_tokens": 47010970.0,
      "step": 3580
    },
    {
      "entropy": 4.334375,
      "epoch": 1.0084269662921348,
      "grad_norm": 1.0945370197296143,
      "learning_rate": 5.79572766310877e-06,
      "loss": 0.4163822174072266,
      "mean_token_accuracy": 0.8814712107181549,
      "num_tokens": 47141729.0,
      "step": 3590
    },
    {
      "entropy": 4.328125,
      "epoch": 1.0112359550561798,
      "grad_norm": 1.2106223106384277,
      "learning_rate": 5.771517561446949e-06,
      "loss": 0.4136615753173828,
      "mean_token_accuracy": 0.8824999868869782,
      "num_tokens": 47272715.0,
      "step": 3600
    },
    {
      "epoch": 1.0112359550561798,
      "eval_entropy": 4.183607651245552,
      "eval_loss": 0.4540982246398926,
      "eval_mean_token_accuracy": 0.874537681769646,
      "eval_num_tokens": 47272715.0,
      "eval_runtime": 118.0732,
      "eval_samples_per_second": 133.239,
      "eval_steps_per_second": 2.38,
      "step": 3600
    },
    {
      "entropy": 4.190625,
      "epoch": 1.0140449438202248,
      "grad_norm": 1.4473798274993896,
      "learning_rate": 5.747288915944533e-06,
      "loss": 0.4010066032409668,
      "mean_token_accuracy": 0.886745035648346,
      "num_tokens": 47404398.0,
      "step": 3610
    },
    {
      "entropy": 4.296875,
      "epoch": 1.0168539325842696,
      "grad_norm": 1.1612318754196167,
      "learning_rate": 5.7230423089500845e-06,
      "loss": 0.41094355583190917,
      "mean_token_accuracy": 0.8829556941986084,
      "num_tokens": 47535039.0,
      "step": 3620
    },
    {
      "entropy": 4.23125,
      "epoch": 1.0196629213483146,
      "grad_norm": 1.2845969200134277,
      "learning_rate": 5.698778323243871e-06,
      "loss": 0.4121527671813965,
      "mean_token_accuracy": 0.8817197620868683,
      "num_tokens": 47665195.0,
      "step": 3630
    },
    {
      "entropy": 4.06875,
      "epoch": 1.0224719101123596,
      "grad_norm": 1.3181904554367065,
      "learning_rate": 5.674497542023875e-06,
      "loss": 0.41451435089111327,
      "mean_token_accuracy": 0.8813693523406982,
      "num_tokens": 47796231.0,
      "step": 3640
    },
    {
      "entropy": 4.0328125,
      "epoch": 1.0252808988764044,
      "grad_norm": 1.0861999988555908,
      "learning_rate": 5.650200548891764e-06,
      "loss": 0.42156410217285156,
      "mean_token_accuracy": 0.8793133020401,
      "num_tokens": 47927262.0,
      "step": 3650
    },
    {
      "entropy": 3.921875,
      "epoch": 1.0280898876404494,
      "grad_norm": 1.1171104907989502,
      "learning_rate": 5.625887927838872e-06,
      "loss": 0.41171441078186033,
      "mean_token_accuracy": 0.8833256721496582,
      "num_tokens": 48057817.0,
      "step": 3660
    },
    {
      "entropy": 3.8546875,
      "epoch": 1.0308988764044944,
      "grad_norm": 1.070045828819275,
      "learning_rate": 5.601560263232153e-06,
      "loss": 0.40767884254455566,
      "mean_token_accuracy": 0.8828452110290528,
      "num_tokens": 48189759.0,
      "step": 3670
    },
    {
      "entropy": 3.9734375,
      "epoch": 1.0337078651685394,
      "grad_norm": 1.1839975118637085,
      "learning_rate": 5.577218139800143e-06,
      "loss": 0.4215818405151367,
      "mean_token_accuracy": 0.877341341972351,
      "num_tokens": 48321984.0,
      "step": 3680
    },
    {
      "entropy": 3.834375,
      "epoch": 1.0365168539325842,
      "grad_norm": 1.098108172416687,
      "learning_rate": 5.552862142618906e-06,
      "loss": 0.41292123794555663,
      "mean_token_accuracy": 0.8819846153259278,
      "num_tokens": 48453805.0,
      "step": 3690
    },
    {
      "entropy": 3.7828125,
      "epoch": 1.0393258426966292,
      "grad_norm": 1.1380692720413208,
      "learning_rate": 5.528492857097966e-06,
      "loss": 0.43189454078674316,
      "mean_token_accuracy": 0.8770938158035279,
      "num_tokens": 48584941.0,
      "step": 3700
    },
    {
      "epoch": 1.0393258426966292,
      "eval_entropy": 3.804437277580071,
      "eval_loss": 0.4506149888038635,
      "eval_mean_token_accuracy": 0.874865082447215,
      "eval_num_tokens": 48584941.0,
      "eval_runtime": 120.4213,
      "eval_samples_per_second": 130.641,
      "eval_steps_per_second": 2.333,
      "step": 3700
    },
    {
      "entropy": 3.8203125,
      "epoch": 1.0421348314606742,
      "grad_norm": 1.0784122943878174,
      "learning_rate": 5.504110868966239e-06,
      "loss": 0.4063127040863037,
      "mean_token_accuracy": 0.8819452285766601,
      "num_tokens": 48716163.0,
      "step": 3710
    },
    {
      "entropy": 3.7515625,
      "epoch": 1.0449438202247192,
      "grad_norm": 1.080193281173706,
      "learning_rate": 5.479716764257961e-06,
      "loss": 0.4122201919555664,
      "mean_token_accuracy": 0.8820477306842804,
      "num_tokens": 48846835.0,
      "step": 3720
    },
    {
      "entropy": 3.803125,
      "epoch": 1.047752808988764,
      "grad_norm": 1.2241575717926025,
      "learning_rate": 5.455311129298586e-06,
      "loss": 0.4118812084197998,
      "mean_token_accuracy": 0.8808844923973084,
      "num_tokens": 48978194.0,
      "step": 3730
    },
    {
      "entropy": 3.6203125,
      "epoch": 1.050561797752809,
      "grad_norm": 1.1736242771148682,
      "learning_rate": 5.430894550690714e-06,
      "loss": 0.39596149921417234,
      "mean_token_accuracy": 0.8840920090675354,
      "num_tokens": 49109274.0,
      "step": 3740
    },
    {
      "entropy": 3.625,
      "epoch": 1.053370786516854,
      "grad_norm": 1.0754168033599854,
      "learning_rate": 5.4064676152999765e-06,
      "loss": 0.41350374221801756,
      "mean_token_accuracy": 0.8796187341213226,
      "num_tokens": 49239687.0,
      "step": 3750
    },
    {
      "entropy": 3.659375,
      "epoch": 1.0561797752808988,
      "grad_norm": 1.329152226448059,
      "learning_rate": 5.382030910240936e-06,
      "loss": 0.416829776763916,
      "mean_token_accuracy": 0.8807898879051208,
      "num_tokens": 49370766.0,
      "step": 3760
    },
    {
      "entropy": 3.6921875,
      "epoch": 1.0589887640449438,
      "grad_norm": 1.0852371454238892,
      "learning_rate": 5.357585022862977e-06,
      "loss": 0.40590496063232423,
      "mean_token_accuracy": 0.8848334968090057,
      "num_tokens": 49502684.0,
      "step": 3770
    },
    {
      "entropy": 3.5484375,
      "epoch": 1.0617977528089888,
      "grad_norm": 1.2195013761520386,
      "learning_rate": 5.33313054073618e-06,
      "loss": 0.4049705505371094,
      "mean_token_accuracy": 0.8828855991363526,
      "num_tokens": 49633697.0,
      "step": 3780
    },
    {
      "entropy": 3.471875,
      "epoch": 1.0646067415730338,
      "grad_norm": 0.9979469180107117,
      "learning_rate": 5.308668051637213e-06,
      "loss": 0.38985157012939453,
      "mean_token_accuracy": 0.8875100553035736,
      "num_tokens": 49764718.0,
      "step": 3790
    },
    {
      "entropy": 3.60625,
      "epoch": 1.0674157303370786,
      "grad_norm": 1.2056188583374023,
      "learning_rate": 5.284198143535188e-06,
      "loss": 0.4026196956634521,
      "mean_token_accuracy": 0.8845828652381897,
      "num_tokens": 49895680.0,
      "step": 3800
    },
    {
      "epoch": 1.0674157303370786,
      "eval_entropy": 3.5981983985765122,
      "eval_loss": 0.44998064637184143,
      "eval_mean_token_accuracy": 0.8752762802555043,
      "eval_num_tokens": 49895680.0,
      "eval_runtime": 106.5442,
      "eval_samples_per_second": 147.657,
      "eval_steps_per_second": 2.637,
      "step": 3800
    },
    {
      "entropy": 3.578125,
      "epoch": 1.0702247191011236,
      "grad_norm": 1.0685278177261353,
      "learning_rate": 5.259721404577546e-06,
      "loss": 0.3967797040939331,
      "mean_token_accuracy": 0.885551393032074,
      "num_tokens": 50026997.0,
      "step": 3810
    },
    {
      "entropy": 3.53125,
      "epoch": 1.0730337078651686,
      "grad_norm": 1.1308002471923828,
      "learning_rate": 5.235238423075899e-06,
      "loss": 0.3994171380996704,
      "mean_token_accuracy": 0.8845544040203095,
      "num_tokens": 50157748.0,
      "step": 3820
    },
    {
      "entropy": 3.53125,
      "epoch": 1.0758426966292134,
      "grad_norm": 1.2615025043487549,
      "learning_rate": 5.210749787491913e-06,
      "loss": 0.39702210426330564,
      "mean_token_accuracy": 0.8862241923809051,
      "num_tokens": 50289435.0,
      "step": 3830
    },
    {
      "entropy": 3.503125,
      "epoch": 1.0786516853932584,
      "grad_norm": 1.451629400253296,
      "learning_rate": 5.186256086423148e-06,
      "loss": 0.4162610530853271,
      "mean_token_accuracy": 0.882129555940628,
      "num_tokens": 50422060.0,
      "step": 3840
    },
    {
      "entropy": 3.5,
      "epoch": 1.0814606741573034,
      "grad_norm": 1.2850306034088135,
      "learning_rate": 5.161757908588917e-06,
      "loss": 0.40077662467956543,
      "mean_token_accuracy": 0.8841105997562408,
      "num_tokens": 50552062.0,
      "step": 3850
    },
    {
      "entropy": 3.584375,
      "epoch": 1.0842696629213484,
      "grad_norm": 1.1060551404953003,
      "learning_rate": 5.137255842816132e-06,
      "loss": 0.41530394554138184,
      "mean_token_accuracy": 0.8813151955604553,
      "num_tokens": 50684567.0,
      "step": 3860
    },
    {
      "entropy": 3.7484375,
      "epoch": 1.0870786516853932,
      "grad_norm": 1.179275631904602,
      "learning_rate": 5.112750478025156e-06,
      "loss": 0.40700445175170896,
      "mean_token_accuracy": 0.8837471306324005,
      "num_tokens": 50815732.0,
      "step": 3870
    },
    {
      "entropy": 3.6796875,
      "epoch": 1.0898876404494382,
      "grad_norm": 1.329466700553894,
      "learning_rate": 5.088242403215644e-06,
      "loss": 0.37986049652099607,
      "mean_token_accuracy": 0.8896700024604798,
      "num_tokens": 50946975.0,
      "step": 3880
    },
    {
      "entropy": 3.778125,
      "epoch": 1.0926966292134832,
      "grad_norm": 1.1162314414978027,
      "learning_rate": 5.063732207452391e-06,
      "loss": 0.3953003168106079,
      "mean_token_accuracy": 0.8867117524147033,
      "num_tokens": 51078339.0,
      "step": 3890
    },
    {
      "entropy": 4.0234375,
      "epoch": 1.095505617977528,
      "grad_norm": 1.169447422027588,
      "learning_rate": 5.039220479851167e-06,
      "loss": 0.4284778118133545,
      "mean_token_accuracy": 0.8773736596107483,
      "num_tokens": 51208784.0,
      "step": 3900
    },
    {
      "epoch": 1.095505617977528,
      "eval_entropy": 4.0824065836298935,
      "eval_loss": 0.44906312227249146,
      "eval_mean_token_accuracy": 0.8754903105654326,
      "eval_num_tokens": 51208784.0,
      "eval_runtime": 119.7529,
      "eval_samples_per_second": 131.371,
      "eval_steps_per_second": 2.346,
      "step": 3900
    },
    {
      "entropy": 3.9609375,
      "epoch": 1.098314606741573,
      "grad_norm": 1.190915584564209,
      "learning_rate": 5.014707809564562e-06,
      "loss": 0.41518387794494627,
      "mean_token_accuracy": 0.8802760481834412,
      "num_tokens": 51340519.0,
      "step": 3910
    },
    {
      "entropy": 3.821875,
      "epoch": 1.101123595505618,
      "grad_norm": 1.0643631219863892,
      "learning_rate": 4.990194785767824e-06,
      "loss": 0.39606800079345705,
      "mean_token_accuracy": 0.8863142788410187,
      "num_tokens": 51471855.0,
      "step": 3920
    },
    {
      "entropy": 3.709375,
      "epoch": 1.103932584269663,
      "grad_norm": 1.0440655946731567,
      "learning_rate": 4.9656819976446975e-06,
      "loss": 0.41166105270385744,
      "mean_token_accuracy": 0.8818281888961792,
      "num_tokens": 51603760.0,
      "step": 3930
    },
    {
      "entropy": 3.5765625,
      "epoch": 1.1067415730337078,
      "grad_norm": 0.990972101688385,
      "learning_rate": 4.94117003437326e-06,
      "loss": 0.40563225746154785,
      "mean_token_accuracy": 0.8835303544998169,
      "num_tokens": 51735139.0,
      "step": 3940
    },
    {
      "entropy": 3.603125,
      "epoch": 1.1095505617977528,
      "grad_norm": 1.021336555480957,
      "learning_rate": 4.9166594851117696e-06,
      "loss": 0.40679044723510743,
      "mean_token_accuracy": 0.8813790500164032,
      "num_tokens": 51866808.0,
      "step": 3950
    },
    {
      "entropy": 3.75625,
      "epoch": 1.1123595505617978,
      "grad_norm": 1.230919599533081,
      "learning_rate": 4.892150938984491e-06,
      "loss": 0.3935567378997803,
      "mean_token_accuracy": 0.8878830373287201,
      "num_tokens": 51998276.0,
      "step": 3960
    },
    {
      "entropy": 3.79375,
      "epoch": 1.1151685393258428,
      "grad_norm": 1.1783368587493896,
      "learning_rate": 4.867644985067548e-06,
      "loss": 0.4146561622619629,
      "mean_token_accuracy": 0.8810646951198577,
      "num_tokens": 52130049.0,
      "step": 3970
    },
    {
      "entropy": 3.746875,
      "epoch": 1.1179775280898876,
      "grad_norm": 1.3365007638931274,
      "learning_rate": 4.8431422123747524e-06,
      "loss": 0.3943379640579224,
      "mean_token_accuracy": 0.886041784286499,
      "num_tokens": 52262002.0,
      "step": 3980
    },
    {
      "entropy": 3.5375,
      "epoch": 1.1207865168539326,
      "grad_norm": 1.4117707014083862,
      "learning_rate": 4.818643209843463e-06,
      "loss": 0.40580263137817385,
      "mean_token_accuracy": 0.8847820401191712,
      "num_tokens": 52393179.0,
      "step": 3990
    },
    {
      "entropy": 3.6625,
      "epoch": 1.1235955056179776,
      "grad_norm": 1.127669095993042,
      "learning_rate": 4.794148566320412e-06,
      "loss": 0.4044083595275879,
      "mean_token_accuracy": 0.8833491742610932,
      "num_tokens": 52524520.0,
      "step": 4000
    },
    {
      "epoch": 1.1235955056179776,
      "eval_entropy": 3.668038256227758,
      "eval_loss": 0.44782185554504395,
      "eval_mean_token_accuracy": 0.8757484729179709,
      "eval_num_tokens": 52524520.0,
      "eval_runtime": 105.0381,
      "eval_samples_per_second": 149.774,
      "eval_steps_per_second": 2.675,
      "step": 4000
    },
    {
      "entropy": 3.5421875,
      "epoch": 1.1264044943820224,
      "grad_norm": 1.2276097536087036,
      "learning_rate": 4.769658870547567e-06,
      "loss": 0.4037940979003906,
      "mean_token_accuracy": 0.8857294976711273,
      "num_tokens": 52655462.0,
      "step": 4010
    },
    {
      "entropy": 3.6140625,
      "epoch": 1.1292134831460674,
      "grad_norm": 1.430707335472107,
      "learning_rate": 4.745174711147967e-06,
      "loss": 0.4096232891082764,
      "mean_token_accuracy": 0.8828846216201782,
      "num_tokens": 52787537.0,
      "step": 4020
    },
    {
      "entropy": 3.7015625,
      "epoch": 1.1320224719101124,
      "grad_norm": 1.1962041854858398,
      "learning_rate": 4.720696676611589e-06,
      "loss": 0.4125821113586426,
      "mean_token_accuracy": 0.8836497902870178,
      "num_tokens": 52919035.0,
      "step": 4030
    },
    {
      "entropy": 3.6359375,
      "epoch": 1.1348314606741572,
      "grad_norm": 1.1133244037628174,
      "learning_rate": 4.6962253552811885e-06,
      "loss": 0.41582231521606444,
      "mean_token_accuracy": 0.8806864976882934,
      "num_tokens": 53050872.0,
      "step": 4040
    },
    {
      "entropy": 3.6796875,
      "epoch": 1.1376404494382022,
      "grad_norm": 1.2544440031051636,
      "learning_rate": 4.671761335338171e-06,
      "loss": 0.4070328712463379,
      "mean_token_accuracy": 0.8829859495162964,
      "num_tokens": 53182521.0,
      "step": 4050
    },
    {
      "entropy": 3.703125,
      "epoch": 1.1404494382022472,
      "grad_norm": 1.0880098342895508,
      "learning_rate": 4.647305204788445e-06,
      "loss": 0.4128103256225586,
      "mean_token_accuracy": 0.8824875831604004,
      "num_tokens": 53313918.0,
      "step": 4060
    },
    {
      "entropy": 3.546875,
      "epoch": 1.1432584269662922,
      "grad_norm": 1.1713446378707886,
      "learning_rate": 4.622857551448297e-06,
      "loss": 0.40809359550476076,
      "mean_token_accuracy": 0.882917457818985,
      "num_tokens": 53445560.0,
      "step": 4070
    },
    {
      "entropy": 3.4265625,
      "epoch": 1.146067415730337,
      "grad_norm": 1.231393575668335,
      "learning_rate": 4.598418962930258e-06,
      "loss": 0.40450549125671387,
      "mean_token_accuracy": 0.8826418519020081,
      "num_tokens": 53577253.0,
      "step": 4080
    },
    {
      "entropy": 3.45,
      "epoch": 1.148876404494382,
      "grad_norm": 1.1458688974380493,
      "learning_rate": 4.573990026628976e-06,
      "loss": 0.4218902111053467,
      "mean_token_accuracy": 0.8781938314437866,
      "num_tokens": 53709315.0,
      "step": 4090
    },
    {
      "entropy": 3.4328125,
      "epoch": 1.151685393258427,
      "grad_norm": 1.2311592102050781,
      "learning_rate": 4.549571329707113e-06,
      "loss": 0.4104034423828125,
      "mean_token_accuracy": 0.8818806231021881,
      "num_tokens": 53841229.0,
      "step": 4100
    },
    {
      "epoch": 1.151685393258427,
      "eval_entropy": 3.4549599644128115,
      "eval_loss": 0.4456959366798401,
      "eval_mean_token_accuracy": 0.8763092342644824,
      "eval_num_tokens": 53841229.0,
      "eval_runtime": 101.9303,
      "eval_samples_per_second": 154.341,
      "eval_steps_per_second": 2.757,
      "step": 4100
    },
    {
      "entropy": 3.4140625,
      "epoch": 1.154494382022472,
      "grad_norm": 1.20809805393219,
      "learning_rate": 4.52516345908121e-06,
      "loss": 0.42303857803344724,
      "mean_token_accuracy": 0.877154815196991,
      "num_tokens": 53973685.0,
      "step": 4110
    },
    {
      "entropy": 3.3578125,
      "epoch": 1.1573033707865168,
      "grad_norm": 1.1762957572937012,
      "learning_rate": 4.500767001407604e-06,
      "loss": 0.4012744426727295,
      "mean_token_accuracy": 0.8858739852905273,
      "num_tokens": 54104087.0,
      "step": 4120
    },
    {
      "entropy": 3.3515625,
      "epoch": 1.1601123595505618,
      "grad_norm": 1.1495301723480225,
      "learning_rate": 4.4763825430683055e-06,
      "loss": 0.4072721004486084,
      "mean_token_accuracy": 0.8822124481201172,
      "num_tokens": 54235873.0,
      "step": 4130
    },
    {
      "entropy": 3.3578125,
      "epoch": 1.1629213483146068,
      "grad_norm": 1.228175163269043,
      "learning_rate": 4.452010670156922e-06,
      "loss": 0.4043898582458496,
      "mean_token_accuracy": 0.8838089823722839,
      "num_tokens": 54367428.0,
      "step": 4140
    },
    {
      "entropy": 3.403125,
      "epoch": 1.1657303370786516,
      "grad_norm": 1.2128483057022095,
      "learning_rate": 4.427651968464559e-06,
      "loss": 0.42446460723876955,
      "mean_token_accuracy": 0.8771877825260163,
      "num_tokens": 54500396.0,
      "step": 4150
    },
    {
      "entropy": 3.525,
      "epoch": 1.1685393258426966,
      "grad_norm": 1.2157642841339111,
      "learning_rate": 4.403307023465746e-06,
      "loss": 0.4092427730560303,
      "mean_token_accuracy": 0.8819334626197814,
      "num_tokens": 54632102.0,
      "step": 4160
    },
    {
      "entropy": 3.5203125,
      "epoch": 1.1713483146067416,
      "grad_norm": 0.991139829158783,
      "learning_rate": 4.378976420304361e-06,
      "loss": 0.40639586448669435,
      "mean_token_accuracy": 0.8829631984233857,
      "num_tokens": 54763256.0,
      "step": 4170
    },
    {
      "entropy": 3.31875,
      "epoch": 1.1741573033707866,
      "grad_norm": 1.0436687469482422,
      "learning_rate": 4.354660743779575e-06,
      "loss": 0.392691445350647,
      "mean_token_accuracy": 0.8873530805110932,
      "num_tokens": 54894484.0,
      "step": 4180
    },
    {
      "entropy": 3.21875,
      "epoch": 1.1769662921348314,
      "grad_norm": 1.288089632987976,
      "learning_rate": 4.3303605783317794e-06,
      "loss": 0.4123360633850098,
      "mean_token_accuracy": 0.8815446257591247,
      "num_tokens": 55025646.0,
      "step": 4190
    },
    {
      "entropy": 3.221875,
      "epoch": 1.1797752808988764,
      "grad_norm": 1.0607856512069702,
      "learning_rate": 4.306076508028557e-06,
      "loss": 0.3960428714752197,
      "mean_token_accuracy": 0.8846178352832794,
      "num_tokens": 55156200.0,
      "step": 4200
    },
    {
      "epoch": 1.1797752808988764,
      "eval_entropy": 3.2018460854092528,
      "eval_loss": 0.4448745548725128,
      "eval_mean_token_accuracy": 0.8765745665679199,
      "eval_num_tokens": 55156200.0,
      "eval_runtime": 113.1269,
      "eval_samples_per_second": 139.065,
      "eval_steps_per_second": 2.484,
      "step": 4200
    },
    {
      "entropy": 3.2140625,
      "epoch": 1.1825842696629214,
      "grad_norm": 1.3779313564300537,
      "learning_rate": 4.281809116550629e-06,
      "loss": 0.40043025016784667,
      "mean_token_accuracy": 0.882966685295105,
      "num_tokens": 55286679.0,
      "step": 4210
    },
    {
      "entropy": 3.240625,
      "epoch": 1.1853932584269664,
      "grad_norm": 1.2117971181869507,
      "learning_rate": 4.257558987177835e-06,
      "loss": 0.39475986957550047,
      "mean_token_accuracy": 0.8857363998889923,
      "num_tokens": 55417913.0,
      "step": 4220
    },
    {
      "entropy": 3.3734375,
      "epoch": 1.1882022471910112,
      "grad_norm": 1.2620396614074707,
      "learning_rate": 4.2333267027751125e-06,
      "loss": 0.39558069705963134,
      "mean_token_accuracy": 0.8851284503936767,
      "num_tokens": 55548727.0,
      "step": 4230
    },
    {
      "entropy": 3.5890625,
      "epoch": 1.1910112359550562,
      "grad_norm": 1.293059229850769,
      "learning_rate": 4.209112845778481e-06,
      "loss": 0.4009528160095215,
      "mean_token_accuracy": 0.8840288460254669,
      "num_tokens": 55679262.0,
      "step": 4240
    },
    {
      "entropy": 3.546875,
      "epoch": 1.1938202247191012,
      "grad_norm": 1.305606484413147,
      "learning_rate": 4.1849179981810506e-06,
      "loss": 0.41488146781921387,
      "mean_token_accuracy": 0.8813881635665893,
      "num_tokens": 55812027.0,
      "step": 4250
    },
    {
      "entropy": 3.6453125,
      "epoch": 1.196629213483146,
      "grad_norm": 1.2907280921936035,
      "learning_rate": 4.160742741519028e-06,
      "loss": 0.42234272956848146,
      "mean_token_accuracy": 0.8804936349391937,
      "num_tokens": 55943521.0,
      "step": 4260
    },
    {
      "entropy": 3.625,
      "epoch": 1.199438202247191,
      "grad_norm": 1.2587887048721313,
      "learning_rate": 4.136587656857744e-06,
      "loss": 0.39614095687866213,
      "mean_token_accuracy": 0.8857029020786286,
      "num_tokens": 56074834.0,
      "step": 4270
    },
    {
      "entropy": 3.625,
      "epoch": 1.202247191011236,
      "grad_norm": 1.1753331422805786,
      "learning_rate": 4.112453324777683e-06,
      "loss": 0.3984139680862427,
      "mean_token_accuracy": 0.8854274928569794,
      "num_tokens": 56206095.0,
      "step": 4280
    },
    {
      "entropy": 3.575,
      "epoch": 1.2050561797752808,
      "grad_norm": 1.168663501739502,
      "learning_rate": 4.088340325360529e-06,
      "loss": 0.38228046894073486,
      "mean_token_accuracy": 0.8916480660438537,
      "num_tokens": 56336773.0,
      "step": 4290
    },
    {
      "entropy": 3.55,
      "epoch": 1.2078651685393258,
      "grad_norm": 1.613463044166565,
      "learning_rate": 4.064249238175223e-06,
      "loss": 0.40636096000671384,
      "mean_token_accuracy": 0.8843181967735291,
      "num_tokens": 56468198.0,
      "step": 4300
    },
    {
      "epoch": 1.2078651685393258,
      "eval_entropy": 3.6843861209964412,
      "eval_loss": 0.4434632360935211,
      "eval_mean_token_accuracy": 0.8767975882703299,
      "eval_num_tokens": 56468198.0,
      "eval_runtime": 119.1857,
      "eval_samples_per_second": 131.996,
      "eval_steps_per_second": 2.358,
      "step": 4300
    },
    {
      "entropy": 3.765625,
      "epoch": 1.2106741573033708,
      "grad_norm": 1.33748459815979,
      "learning_rate": 4.04018064226404e-06,
      "loss": 0.40183248519897463,
      "mean_token_accuracy": 0.8845306396484375,
      "num_tokens": 56599602.0,
      "step": 4310
    },
    {
      "entropy": 3.671875,
      "epoch": 1.2134831460674158,
      "grad_norm": 1.2077350616455078,
      "learning_rate": 4.016135116128656e-06,
      "loss": 0.4326166152954102,
      "mean_token_accuracy": 0.8777515232563019,
      "num_tokens": 56731393.0,
      "step": 4320
    },
    {
      "entropy": 3.6828125,
      "epoch": 1.2162921348314606,
      "grad_norm": 1.0218945741653442,
      "learning_rate": 3.992113237716261e-06,
      "loss": 0.400331974029541,
      "mean_token_accuracy": 0.8863215804100036,
      "num_tokens": 56863522.0,
      "step": 4330
    },
    {
      "entropy": 3.7078125,
      "epoch": 1.2191011235955056,
      "grad_norm": 1.190545678138733,
      "learning_rate": 3.9681155844056525e-06,
      "loss": 0.39993202686309814,
      "mean_token_accuracy": 0.8841948628425598,
      "num_tokens": 56994774.0,
      "step": 4340
    },
    {
      "entropy": 3.5703125,
      "epoch": 1.2219101123595506,
      "grad_norm": 1.5019413232803345,
      "learning_rate": 3.94414273299337e-06,
      "loss": 0.40560131072998046,
      "mean_token_accuracy": 0.8851337134838104,
      "num_tokens": 57126769.0,
      "step": 4350
    },
    {
      "entropy": 3.6140625,
      "epoch": 1.2247191011235956,
      "grad_norm": 1.3031398057937622,
      "learning_rate": 3.920195259679822e-06,
      "loss": 0.41339969635009766,
      "mean_token_accuracy": 0.879789674282074,
      "num_tokens": 57257878.0,
      "step": 4360
    },
    {
      "entropy": 3.6015625,
      "epoch": 1.2275280898876404,
      "grad_norm": 1.1451210975646973,
      "learning_rate": 3.8962737400554395e-06,
      "loss": 0.4006919860839844,
      "mean_token_accuracy": 0.8831992089748383,
      "num_tokens": 57388777.0,
      "step": 4370
    },
    {
      "entropy": 3.5640625,
      "epoch": 1.2303370786516854,
      "grad_norm": 1.1797364950180054,
      "learning_rate": 3.87237874908685e-06,
      "loss": 0.39904673099517823,
      "mean_token_accuracy": 0.8837074041366577,
      "num_tokens": 57519514.0,
      "step": 4380
    },
    {
      "entropy": 3.66875,
      "epoch": 1.2331460674157304,
      "grad_norm": 1.1160693168640137,
      "learning_rate": 3.8485108611030415e-06,
      "loss": 0.41873650550842284,
      "mean_token_accuracy": 0.8794358730316162,
      "num_tokens": 57651018.0,
      "step": 4390
    },
    {
      "entropy": 3.5546875,
      "epoch": 1.2359550561797752,
      "grad_norm": 5.751740455627441,
      "learning_rate": 3.824670649781576e-06,
      "loss": 0.4053187370300293,
      "mean_token_accuracy": 0.8834964215755463,
      "num_tokens": 57783089.0,
      "step": 4400
    },
    {
      "epoch": 1.2359550561797752,
      "eval_entropy": 3.597586743772242,
      "eval_loss": 0.44261476397514343,
      "eval_mean_token_accuracy": 0.8771198847964141,
      "eval_num_tokens": 57783089.0,
      "eval_runtime": 104.5539,
      "eval_samples_per_second": 150.468,
      "eval_steps_per_second": 2.688,
      "step": 4400
    },
    {
      "entropy": 3.5546875,
      "epoch": 1.2387640449438202,
      "grad_norm": 1.1170510053634644,
      "learning_rate": 3.8008586881347815e-06,
      "loss": 0.4001267433166504,
      "mean_token_accuracy": 0.8854337751865387,
      "num_tokens": 57914121.0,
      "step": 4410
    },
    {
      "entropy": 3.4421875,
      "epoch": 1.2415730337078652,
      "grad_norm": 1.1510826349258423,
      "learning_rate": 3.777075548496001e-06,
      "loss": 0.3900010585784912,
      "mean_token_accuracy": 0.8872870624065399,
      "num_tokens": 58044699.0,
      "step": 4420
    },
    {
      "entropy": 3.3421875,
      "epoch": 1.24438202247191,
      "grad_norm": 1.34503173828125,
      "learning_rate": 3.753321802505817e-06,
      "loss": 0.40116004943847655,
      "mean_token_accuracy": 0.8843511760234832,
      "num_tokens": 58176783.0,
      "step": 4430
    },
    {
      "entropy": 3.3390625,
      "epoch": 1.247191011235955,
      "grad_norm": 1.345521092414856,
      "learning_rate": 3.7295980210983233e-06,
      "loss": 0.40514535903930665,
      "mean_token_accuracy": 0.8822214365005493,
      "num_tokens": 58307848.0,
      "step": 4440
    },
    {
      "entropy": 3.4015625,
      "epoch": 1.25,
      "grad_norm": 0.9972835183143616,
      "learning_rate": 3.705904774487396e-06,
      "loss": 0.40918550491333006,
      "mean_token_accuracy": 0.88335702419281,
      "num_tokens": 58440276.0,
      "step": 4450
    },
    {
      "entropy": 3.5125,
      "epoch": 1.252808988764045,
      "grad_norm": 1.155853033065796,
      "learning_rate": 3.6822426321529967e-06,
      "loss": 0.4174796581268311,
      "mean_token_accuracy": 0.880107456445694,
      "num_tokens": 58572345.0,
      "step": 4460
    },
    {
      "entropy": 3.5,
      "epoch": 1.25561797752809,
      "grad_norm": 1.0794697999954224,
      "learning_rate": 3.658612162827472e-06,
      "loss": 0.41515040397644043,
      "mean_token_accuracy": 0.8810400545597077,
      "num_tokens": 58703115.0,
      "step": 4470
    },
    {
      "entropy": 3.3828125,
      "epoch": 1.2584269662921348,
      "grad_norm": 1.1338727474212646,
      "learning_rate": 3.635013934481895e-06,
      "loss": 0.3853863954544067,
      "mean_token_accuracy": 0.8895859003067017,
      "num_tokens": 58833868.0,
      "step": 4480
    },
    {
      "entropy": 3.4796875,
      "epoch": 1.2612359550561798,
      "grad_norm": 1.3220850229263306,
      "learning_rate": 3.6114485143124068e-06,
      "loss": 0.39600634574890137,
      "mean_token_accuracy": 0.8836991369724274,
      "num_tokens": 58964446.0,
      "step": 4490
    },
    {
      "entropy": 3.4921875,
      "epoch": 1.2640449438202248,
      "grad_norm": 1.304870843887329,
      "learning_rate": 3.58791646872659e-06,
      "loss": 0.4082575798034668,
      "mean_token_accuracy": 0.8822609126567841,
      "num_tokens": 59095948.0,
      "step": 4500
    },
    {
      "epoch": 1.2640449438202248,
      "eval_entropy": 3.543427491103203,
      "eval_loss": 0.44224244356155396,
      "eval_mean_token_accuracy": 0.877214883570145,
      "eval_num_tokens": 59095948.0,
      "eval_runtime": 107.6108,
      "eval_samples_per_second": 146.194,
      "eval_steps_per_second": 2.611,
      "step": 4500
    },
    {
      "entropy": 3.5828125,
      "epoch": 1.2668539325842696,
      "grad_norm": 1.2184215784072876,
      "learning_rate": 3.564418363329848e-06,
      "loss": 0.4023094177246094,
      "mean_token_accuracy": 0.8844042122364044,
      "num_tokens": 59227257.0,
      "step": 4510
    },
    {
      "entropy": 3.4046875,
      "epoch": 1.2696629213483146,
      "grad_norm": 1.187645435333252,
      "learning_rate": 3.5409547629118124e-06,
      "loss": 0.391055703163147,
      "mean_token_accuracy": 0.8877169847488403,
      "num_tokens": 59359018.0,
      "step": 4520
    },
    {
      "entropy": 3.4171875,
      "epoch": 1.2724719101123596,
      "grad_norm": 1.1363829374313354,
      "learning_rate": 3.517526231432775e-06,
      "loss": 0.42122039794921873,
      "mean_token_accuracy": 0.8789879262447358,
      "num_tokens": 59490178.0,
      "step": 4530
    },
    {
      "entropy": 3.5078125,
      "epoch": 1.2752808988764044,
      "grad_norm": 1.2686041593551636,
      "learning_rate": 3.4941333320101173e-06,
      "loss": 0.40302042961120604,
      "mean_token_accuracy": 0.8843392610549927,
      "num_tokens": 59622431.0,
      "step": 4540
    },
    {
      "entropy": 3.4390625,
      "epoch": 1.2780898876404494,
      "grad_norm": 1.1338917016983032,
      "learning_rate": 3.470776626904795e-06,
      "loss": 0.40384559631347655,
      "mean_token_accuracy": 0.8849855244159699,
      "num_tokens": 59753794.0,
      "step": 4550
    },
    {
      "entropy": 3.3515625,
      "epoch": 1.2808988764044944,
      "grad_norm": 1.0716454982757568,
      "learning_rate": 3.4474566775078055e-06,
      "loss": 0.3970206260681152,
      "mean_token_accuracy": 0.8866419434547425,
      "num_tokens": 59885985.0,
      "step": 4560
    },
    {
      "entropy": 3.3234375,
      "epoch": 1.2837078651685394,
      "grad_norm": 1.1167986392974854,
      "learning_rate": 3.4241740443267112e-06,
      "loss": 0.40575480461120605,
      "mean_token_accuracy": 0.8839031279087066,
      "num_tokens": 60018325.0,
      "step": 4570
    },
    {
      "entropy": 3.3609375,
      "epoch": 1.2865168539325842,
      "grad_norm": 1.1201732158660889,
      "learning_rate": 3.4009292869721516e-06,
      "loss": 0.40019211769104,
      "mean_token_accuracy": 0.8843213200569153,
      "num_tokens": 60149512.0,
      "step": 4580
    },
    {
      "entropy": 3.3984375,
      "epoch": 1.2893258426966292,
      "grad_norm": 1.1381447315216064,
      "learning_rate": 3.377722964144405e-06,
      "loss": 0.3991153955459595,
      "mean_token_accuracy": 0.8849389791488648,
      "num_tokens": 60281055.0,
      "step": 4590
    },
    {
      "entropy": 3.4625,
      "epoch": 1.2921348314606742,
      "grad_norm": 1.0801565647125244,
      "learning_rate": 3.35455563361995e-06,
      "loss": 0.3923803806304932,
      "mean_token_accuracy": 0.8869569897651672,
      "num_tokens": 60411540.0,
      "step": 4600
    },
    {
      "epoch": 1.2921348314606742,
      "eval_entropy": 3.420373665480427,
      "eval_loss": 0.4403699040412903,
      "eval_mean_token_accuracy": 0.8776448076305865,
      "eval_num_tokens": 60411540.0,
      "eval_runtime": 123.1634,
      "eval_samples_per_second": 127.733,
      "eval_steps_per_second": 2.282,
      "step": 4600
    },
    {
      "entropy": 3.328125,
      "epoch": 1.2949438202247192,
      "grad_norm": 1.2614485025405884,
      "learning_rate": 3.331427852238073e-06,
      "loss": 0.39527292251586915,
      "mean_token_accuracy": 0.8863827228546143,
      "num_tokens": 60543250.0,
      "step": 4610
    },
    {
      "entropy": 3.3703125,
      "epoch": 1.297752808988764,
      "grad_norm": 1.1612732410430908,
      "learning_rate": 3.3083401758874655e-06,
      "loss": 0.3968764305114746,
      "mean_token_accuracy": 0.8844027042388916,
      "num_tokens": 60674931.0,
      "step": 4620
    },
    {
      "entropy": 3.3609375,
      "epoch": 1.300561797752809,
      "grad_norm": 1.390006422996521,
      "learning_rate": 3.2852931594928804e-06,
      "loss": 0.4070241451263428,
      "mean_token_accuracy": 0.8845610201358796,
      "num_tokens": 60807211.0,
      "step": 4630
    },
    {
      "entropy": 3.3828125,
      "epoch": 1.303370786516854,
      "grad_norm": 1.0587964057922363,
      "learning_rate": 3.262287357001781e-06,
      "loss": 0.3908301115036011,
      "mean_token_accuracy": 0.8868767976760864,
      "num_tokens": 60937623.0,
      "step": 4640
    },
    {
      "entropy": 3.4421875,
      "epoch": 1.3061797752808988,
      "grad_norm": 1.0927470922470093,
      "learning_rate": 3.239323321371039e-06,
      "loss": 0.3899258852005005,
      "mean_token_accuracy": 0.8863810122013092,
      "num_tokens": 61067812.0,
      "step": 4650
    },
    {
      "entropy": 3.3671875,
      "epoch": 1.3089887640449438,
      "grad_norm": 1.1747150421142578,
      "learning_rate": 3.2164016045536306e-06,
      "loss": 0.3925123929977417,
      "mean_token_accuracy": 0.886543619632721,
      "num_tokens": 61198887.0,
      "step": 4660
    },
    {
      "entropy": 3.3875,
      "epoch": 1.3117977528089888,
      "grad_norm": 1.2835659980773926,
      "learning_rate": 3.193522757485378e-06,
      "loss": 0.40700678825378417,
      "mean_token_accuracy": 0.8822829127311707,
      "num_tokens": 61330662.0,
      "step": 4670
    },
    {
      "entropy": 3.346875,
      "epoch": 1.3146067415730336,
      "grad_norm": 1.2220410108566284,
      "learning_rate": 3.1706873300717094e-06,
      "loss": 0.4237339019775391,
      "mean_token_accuracy": 0.8789071083068848,
      "num_tokens": 61463373.0,
      "step": 4680
    },
    {
      "entropy": 3.40625,
      "epoch": 1.3174157303370786,
      "grad_norm": 1.1730097532272339,
      "learning_rate": 3.1478958711744324e-06,
      "loss": 0.3881240367889404,
      "mean_token_accuracy": 0.8876212298870086,
      "num_tokens": 61594147.0,
      "step": 4690
    },
    {
      "entropy": 3.2453125,
      "epoch": 1.3202247191011236,
      "grad_norm": 1.2072031497955322,
      "learning_rate": 3.125148928598554e-06,
      "loss": 0.3998769044876099,
      "mean_token_accuracy": 0.8841192722320557,
      "num_tokens": 61724764.0,
      "step": 4700
    },
    {
      "epoch": 1.3202247191011236,
      "eval_entropy": 3.099199288256228,
      "eval_loss": 0.4401569366455078,
      "eval_mean_token_accuracy": 0.8778069564456193,
      "eval_num_tokens": 61724764.0,
      "eval_runtime": 122.3699,
      "eval_samples_per_second": 128.561,
      "eval_steps_per_second": 2.296,
      "step": 4700
    },
    {
      "entropy": 3.1703125,
      "epoch": 1.3230337078651686,
      "grad_norm": 1.113301396369934,
      "learning_rate": 3.1024470490791027e-06,
      "loss": 0.3848711967468262,
      "mean_token_accuracy": 0.8879956543445587,
      "num_tokens": 61855318.0,
      "step": 4710
    },
    {
      "entropy": 3.2046875,
      "epoch": 1.3258426966292136,
      "grad_norm": 1.2754250764846802,
      "learning_rate": 3.0797907782679944e-06,
      "loss": 0.39142980575561526,
      "mean_token_accuracy": 0.8875797986984253,
      "num_tokens": 61987056.0,
      "step": 4720
    },
    {
      "entropy": 3.2421875,
      "epoch": 1.3286516853932584,
      "grad_norm": 1.2785242795944214,
      "learning_rate": 3.057180660720912e-06,
      "loss": 0.40009379386901855,
      "mean_token_accuracy": 0.885002589225769,
      "num_tokens": 62118577.0,
      "step": 4730
    },
    {
      "entropy": 3.440625,
      "epoch": 1.3314606741573034,
      "grad_norm": 1.2100592851638794,
      "learning_rate": 3.0346172398842254e-06,
      "loss": 0.39901151657104494,
      "mean_token_accuracy": 0.8827756702899933,
      "num_tokens": 62247698.0,
      "step": 4740
    },
    {
      "entropy": 3.4125,
      "epoch": 1.3342696629213484,
      "grad_norm": 1.2566052675247192,
      "learning_rate": 3.012101058081919e-06,
      "loss": 0.4067508697509766,
      "mean_token_accuracy": 0.8827375590801239,
      "num_tokens": 62378814.0,
      "step": 4750
    },
    {
      "entropy": 3.2640625,
      "epoch": 1.3370786516853932,
      "grad_norm": 1.2522542476654053,
      "learning_rate": 2.989632656502564e-06,
      "loss": 0.41240944862365725,
      "mean_token_accuracy": 0.8810955107212066,
      "num_tokens": 62510169.0,
      "step": 4760
    },
    {
      "entropy": 3.2984375,
      "epoch": 1.3398876404494382,
      "grad_norm": 1.3600422143936157,
      "learning_rate": 2.9672125751863067e-06,
      "loss": 0.3957082986831665,
      "mean_token_accuracy": 0.8861571192741394,
      "num_tokens": 62640569.0,
      "step": 4770
    },
    {
      "entropy": 3.19375,
      "epoch": 1.3426966292134832,
      "grad_norm": 1.4984418153762817,
      "learning_rate": 2.9448413530118912e-06,
      "loss": 0.3987653017044067,
      "mean_token_accuracy": 0.8854438006877899,
      "num_tokens": 62771818.0,
      "step": 4780
    },
    {
      "entropy": 3.2859375,
      "epoch": 1.345505617977528,
      "grad_norm": 1.3019371032714844,
      "learning_rate": 2.922519527683706e-06,
      "loss": 0.3944662809371948,
      "mean_token_accuracy": 0.8866125583648682,
      "num_tokens": 62902415.0,
      "step": 4790
    },
    {
      "entropy": 3.33125,
      "epoch": 1.348314606741573,
      "grad_norm": 1.380331039428711,
      "learning_rate": 2.900247635718856e-06,
      "loss": 0.3923499584197998,
      "mean_token_accuracy": 0.8867680847644805,
      "num_tokens": 63033947.0,
      "step": 4800
    },
    {
      "epoch": 1.348314606741573,
      "eval_entropy": 3.4255449288256226,
      "eval_loss": 0.43870997428894043,
      "eval_mean_token_accuracy": 0.8778788390108699,
      "eval_num_tokens": 63033947.0,
      "eval_runtime": 116.9323,
      "eval_samples_per_second": 134.539,
      "eval_steps_per_second": 2.403,
      "step": 4800
    },
    {
      "entropy": 3.4703125,
      "epoch": 1.351123595505618,
      "grad_norm": 1.221097707748413,
      "learning_rate": 2.8780262124342755e-06,
      "loss": 0.3960247993469238,
      "mean_token_accuracy": 0.8840667068958282,
      "num_tokens": 63164730.0,
      "step": 4810
    },
    {
      "entropy": 3.378125,
      "epoch": 1.3539325842696628,
      "grad_norm": 1.4236786365509033,
      "learning_rate": 2.8558557919338537e-06,
      "loss": 0.40567626953125,
      "mean_token_accuracy": 0.8834254801273346,
      "num_tokens": 63295380.0,
      "step": 4820
    },
    {
      "entropy": 3.3609375,
      "epoch": 1.3567415730337078,
      "grad_norm": 1.199161171913147,
      "learning_rate": 2.833736907095604e-06,
      "loss": 0.38303470611572266,
      "mean_token_accuracy": 0.8880647659301758,
      "num_tokens": 63428143.0,
      "step": 4830
    },
    {
      "entropy": 3.45625,
      "epoch": 1.3595505617977528,
      "grad_norm": 1.1476725339889526,
      "learning_rate": 2.8116700895588473e-06,
      "loss": 0.40090270042419435,
      "mean_token_accuracy": 0.8855371713638306,
      "num_tokens": 63558620.0,
      "step": 4840
    },
    {
      "entropy": 3.534375,
      "epoch": 1.3623595505617978,
      "grad_norm": 1.2623145580291748,
      "learning_rate": 2.789655869711445e-06,
      "loss": 0.3990975856781006,
      "mean_token_accuracy": 0.8853765845298767,
      "num_tokens": 63689105.0,
      "step": 4850
    },
    {
      "entropy": 3.3671875,
      "epoch": 1.3651685393258428,
      "grad_norm": 1.2698699235916138,
      "learning_rate": 2.7676947766770367e-06,
      "loss": 0.3995832920074463,
      "mean_token_accuracy": 0.8868639349937439,
      "num_tokens": 63819577.0,
      "step": 4860
    },
    {
      "entropy": 3.3390625,
      "epoch": 1.3679775280898876,
      "grad_norm": 1.2152751684188843,
      "learning_rate": 2.745787338302341e-06,
      "loss": 0.4011357307434082,
      "mean_token_accuracy": 0.8845615327358246,
      "num_tokens": 63951065.0,
      "step": 4870
    },
    {
      "entropy": 3.40625,
      "epoch": 1.3707865168539326,
      "grad_norm": 1.1834627389907837,
      "learning_rate": 2.7239340811444476e-06,
      "loss": 0.38979063034057615,
      "mean_token_accuracy": 0.8886278688907623,
      "num_tokens": 64082320.0,
      "step": 4880
    },
    {
      "entropy": 3.4125,
      "epoch": 1.3735955056179776,
      "grad_norm": 1.3519912958145142,
      "learning_rate": 2.7021355304581765e-06,
      "loss": 0.4010939598083496,
      "mean_token_accuracy": 0.88430255651474,
      "num_tokens": 64213060.0,
      "step": 4890
    },
    {
      "entropy": 3.3609375,
      "epoch": 1.3764044943820224,
      "grad_norm": 1.3043447732925415,
      "learning_rate": 2.680392210183446e-06,
      "loss": 0.4047522068023682,
      "mean_token_accuracy": 0.8813096582889557,
      "num_tokens": 64344151.0,
      "step": 4900
    },
    {
      "epoch": 1.3764044943820224,
      "eval_entropy": 3.3610431494661923,
      "eval_loss": 0.4377325177192688,
      "eval_mean_token_accuracy": 0.8780667523048102,
      "eval_num_tokens": 64344151.0,
      "eval_runtime": 119.074,
      "eval_samples_per_second": 132.119,
      "eval_steps_per_second": 2.36,
      "step": 4900
    },
    {
      "entropy": 3.3140625,
      "epoch": 1.3792134831460674,
      "grad_norm": 1.1788116693496704,
      "learning_rate": 2.6587046429326855e-06,
      "loss": 0.3862060785293579,
      "mean_token_accuracy": 0.8879417955875397,
      "num_tokens": 64475639.0,
      "step": 4910
    },
    {
      "entropy": 3.2953125,
      "epoch": 1.3820224719101124,
      "grad_norm": 1.275259017944336,
      "learning_rate": 2.6370733499782654e-06,
      "loss": 0.40172605514526366,
      "mean_token_accuracy": 0.8852776110172271,
      "num_tokens": 64607333.0,
      "step": 4920
    },
    {
      "entropy": 3.3328125,
      "epoch": 1.3848314606741572,
      "grad_norm": 1.2757631540298462,
      "learning_rate": 2.6154988512399784e-06,
      "loss": 0.39677953720092773,
      "mean_token_accuracy": 0.8852551281452179,
      "num_tokens": 64738529.0,
      "step": 4930
    },
    {
      "entropy": 3.3953125,
      "epoch": 1.3876404494382022,
      "grad_norm": 1.2770277261734009,
      "learning_rate": 2.5939816652725324e-06,
      "loss": 0.3958810091018677,
      "mean_token_accuracy": 0.885463434457779,
      "num_tokens": 64869282.0,
      "step": 4940
    },
    {
      "entropy": 3.38125,
      "epoch": 1.3904494382022472,
      "grad_norm": 1.1037077903747559,
      "learning_rate": 2.5725223092530937e-06,
      "loss": 0.3810693025588989,
      "mean_token_accuracy": 0.8901585161685943,
      "num_tokens": 65000317.0,
      "step": 4950
    },
    {
      "entropy": 3.3171875,
      "epoch": 1.3932584269662922,
      "grad_norm": 1.2318624258041382,
      "learning_rate": 2.5511212989688587e-06,
      "loss": 0.4024513244628906,
      "mean_token_accuracy": 0.8834908306598663,
      "num_tokens": 65131044.0,
      "step": 4960
    },
    {
      "entropy": 3.4140625,
      "epoch": 1.3960674157303372,
      "grad_norm": 1.4876635074615479,
      "learning_rate": 2.5297791488046445e-06,
      "loss": 0.4085329532623291,
      "mean_token_accuracy": 0.880237364768982,
      "num_tokens": 65261612.0,
      "step": 4970
    },
    {
      "entropy": 3.3859375,
      "epoch": 1.398876404494382,
      "grad_norm": 1.246620774269104,
      "learning_rate": 2.508496371730543e-06,
      "loss": 0.39396157264709475,
      "mean_token_accuracy": 0.8874974071979522,
      "num_tokens": 65393546.0,
      "step": 4980
    },
    {
      "entropy": 3.3953125,
      "epoch": 1.401685393258427,
      "grad_norm": 1.2896212339401245,
      "learning_rate": 2.487273479289574e-06,
      "loss": 0.3987667798995972,
      "mean_token_accuracy": 0.8858571469783783,
      "num_tokens": 65525043.0,
      "step": 4990
    },
    {
      "entropy": 3.3625,
      "epoch": 1.404494382022472,
      "grad_norm": 1.1399091482162476,
      "learning_rate": 2.4661109815854005e-06,
      "loss": 0.3843994140625,
      "mean_token_accuracy": 0.8884540855884552,
      "num_tokens": 65655822.0,
      "step": 5000
    },
    {
      "epoch": 1.404494382022472,
      "eval_entropy": 3.32456628113879,
      "eval_loss": 0.436702698469162,
      "eval_mean_token_accuracy": 0.8784497942364513,
      "eval_num_tokens": 65655822.0,
      "eval_runtime": 116.4564,
      "eval_samples_per_second": 135.089,
      "eval_steps_per_second": 2.413,
      "step": 5000
    },
    {
      "entropy": 3.40625,
      "epoch": 1.4073033707865168,
      "grad_norm": 1.430216908454895,
      "learning_rate": 2.4450093872700648e-06,
      "loss": 0.3895522117614746,
      "mean_token_accuracy": 0.8870774149894715,
      "num_tokens": 65786496.0,
      "step": 5010
    },
    {
      "entropy": 3.38125,
      "epoch": 1.4101123595505618,
      "grad_norm": 1.2821283340454102,
      "learning_rate": 2.423969203531768e-06,
      "loss": 0.3995974540710449,
      "mean_token_accuracy": 0.8845342874526978,
      "num_tokens": 65917401.0,
      "step": 5020
    },
    {
      "entropy": 3.4734375,
      "epoch": 1.4129213483146068,
      "grad_norm": 1.3741205930709839,
      "learning_rate": 2.402990936082667e-06,
      "loss": 0.3982215404510498,
      "mean_token_accuracy": 0.8843914210796356,
      "num_tokens": 66048545.0,
      "step": 5030
    },
    {
      "entropy": 3.4296875,
      "epoch": 1.4157303370786516,
      "grad_norm": 2.152801990509033,
      "learning_rate": 2.3820750891467355e-06,
      "loss": 0.4025409698486328,
      "mean_token_accuracy": 0.8853477001190185,
      "num_tokens": 66180647.0,
      "step": 5040
    },
    {
      "entropy": 3.4171875,
      "epoch": 1.4185393258426966,
      "grad_norm": 1.175767183303833,
      "learning_rate": 2.361222165447628e-06,
      "loss": 0.3806973695755005,
      "mean_token_accuracy": 0.8894385099411011,
      "num_tokens": 66311812.0,
      "step": 5050
    },
    {
      "entropy": 3.425,
      "epoch": 1.4213483146067416,
      "grad_norm": 1.2237714529037476,
      "learning_rate": 2.3404326661966148e-06,
      "loss": 0.40421085357666015,
      "mean_token_accuracy": 0.8821461796760559,
      "num_tokens": 66443063.0,
      "step": 5060
    },
    {
      "entropy": 3.390625,
      "epoch": 1.4241573033707864,
      "grad_norm": 1.492456316947937,
      "learning_rate": 2.319707091080517e-06,
      "loss": 0.3955252647399902,
      "mean_token_accuracy": 0.8861067235469818,
      "num_tokens": 66574507.0,
      "step": 5070
    },
    {
      "entropy": 3.2453125,
      "epoch": 1.4269662921348314,
      "grad_norm": 1.2392797470092773,
      "learning_rate": 2.2990459382497086e-06,
      "loss": 0.39186649322509765,
      "mean_token_accuracy": 0.8870264649391174,
      "num_tokens": 66705548.0,
      "step": 5080
    },
    {
      "entropy": 3.3484375,
      "epoch": 1.4297752808988764,
      "grad_norm": 1.2176330089569092,
      "learning_rate": 2.2784497043061384e-06,
      "loss": 0.39931521415710447,
      "mean_token_accuracy": 0.8841040790081024,
      "num_tokens": 66836571.0,
      "step": 5090
    },
    {
      "entropy": 3.3234375,
      "epoch": 1.4325842696629214,
      "grad_norm": 1.1505810022354126,
      "learning_rate": 2.257918884291392e-06,
      "loss": 0.4085914134979248,
      "mean_token_accuracy": 0.8824636995792389,
      "num_tokens": 66968327.0,
      "step": 5100
    },
    {
      "epoch": 1.4325842696629214,
      "eval_entropy": 3.32595640569395,
      "eval_loss": 0.43558910489082336,
      "eval_mean_token_accuracy": 0.8785107420008378,
      "eval_num_tokens": 66968327.0,
      "eval_runtime": 110.0114,
      "eval_samples_per_second": 143.003,
      "eval_steps_per_second": 2.554,
      "step": 5100
    },
    {
      "entropy": 3.26875,
      "epoch": 1.4353932584269664,
      "grad_norm": 1.155544400215149,
      "learning_rate": 2.2374539716748034e-06,
      "loss": 0.3753185749053955,
      "mean_token_accuracy": 0.8922010898590088,
      "num_tokens": 67099321.0,
      "step": 5110
    },
    {
      "entropy": 3.1328125,
      "epoch": 1.4382022471910112,
      "grad_norm": 1.291313886642456,
      "learning_rate": 2.2170554583415782e-06,
      "loss": 0.38300347328186035,
      "mean_token_accuracy": 0.8881960988044739,
      "num_tokens": 67230644.0,
      "step": 5120
    },
    {
      "entropy": 3.1703125,
      "epoch": 1.4410112359550562,
      "grad_norm": 1.3153183460235596,
      "learning_rate": 2.196723834580987e-06,
      "loss": 0.38433027267456055,
      "mean_token_accuracy": 0.8875162780284882,
      "num_tokens": 67362483.0,
      "step": 5130
    },
    {
      "entropy": 3.1296875,
      "epoch": 1.4438202247191012,
      "grad_norm": 1.397641897201538,
      "learning_rate": 2.176459589074566e-06,
      "loss": 0.3869623184204102,
      "mean_token_accuracy": 0.8889076232910156,
      "num_tokens": 67493781.0,
      "step": 5140
    },
    {
      "entropy": 3.1421875,
      "epoch": 1.446629213483146,
      "grad_norm": 1.2871443033218384,
      "learning_rate": 2.156263208884386e-06,
      "loss": 0.4007300853729248,
      "mean_token_accuracy": 0.8820683300495148,
      "num_tokens": 67624676.0,
      "step": 5150
    },
    {
      "entropy": 3.1921875,
      "epoch": 1.449438202247191,
      "grad_norm": 1.259085774421692,
      "learning_rate": 2.1361351794413334e-06,
      "loss": 0.38936259746551516,
      "mean_token_accuracy": 0.8869066596031189,
      "num_tokens": 67756482.0,
      "step": 5160
    },
    {
      "entropy": 3.2140625,
      "epoch": 1.452247191011236,
      "grad_norm": 1.2081668376922607,
      "learning_rate": 2.1160759845334483e-06,
      "loss": 0.3983312129974365,
      "mean_token_accuracy": 0.886744225025177,
      "num_tokens": 67887850.0,
      "step": 5170
    },
    {
      "entropy": 3.1796875,
      "epoch": 1.4550561797752808,
      "grad_norm": 1.3973309993743896,
      "learning_rate": 2.0960861062942956e-06,
      "loss": 0.4065245628356934,
      "mean_token_accuracy": 0.8840452671051026,
      "num_tokens": 68018665.0,
      "step": 5180
    },
    {
      "entropy": 3.2359375,
      "epoch": 1.4578651685393258,
      "grad_norm": 1.1832175254821777,
      "learning_rate": 2.0761660251913795e-06,
      "loss": 0.40360226631164553,
      "mean_token_accuracy": 0.8832993924617767,
      "num_tokens": 68150054.0,
      "step": 5190
    },
    {
      "entropy": 3.2953125,
      "epoch": 1.4606741573033708,
      "grad_norm": 1.2756125926971436,
      "learning_rate": 2.056316220014588e-06,
      "loss": 0.39550175666809084,
      "mean_token_accuracy": 0.8854811549186706,
      "num_tokens": 68281123.0,
      "step": 5200
    },
    {
      "epoch": 1.4606741573033708,
      "eval_entropy": 3.2493327402135233,
      "eval_loss": 0.43479055166244507,
      "eval_mean_token_accuracy": 0.8788050217984834,
      "eval_num_tokens": 68281123.0,
      "eval_runtime": 113.7877,
      "eval_samples_per_second": 138.257,
      "eval_steps_per_second": 2.47,
      "step": 5200
    },
    {
      "entropy": 3.2125,
      "epoch": 1.4634831460674158,
      "grad_norm": 1.1542668342590332,
      "learning_rate": 2.036537167864695e-06,
      "loss": 0.3979199409484863,
      "mean_token_accuracy": 0.8839752078056335,
      "num_tokens": 68413359.0,
      "step": 5210
    },
    {
      "entropy": 3.2953125,
      "epoch": 1.4662921348314606,
      "grad_norm": 1.2714418172836304,
      "learning_rate": 2.0168293441418798e-06,
      "loss": 0.4048189163208008,
      "mean_token_accuracy": 0.8816901683807373,
      "num_tokens": 68544806.0,
      "step": 5220
    },
    {
      "entropy": 3.3046875,
      "epoch": 1.4691011235955056,
      "grad_norm": 1.20458984375,
      "learning_rate": 1.997193222534316e-06,
      "loss": 0.3869774580001831,
      "mean_token_accuracy": 0.8862630307674408,
      "num_tokens": 68676198.0,
      "step": 5230
    },
    {
      "entropy": 3.2859375,
      "epoch": 1.4719101123595506,
      "grad_norm": 1.2865755558013916,
      "learning_rate": 1.977629275006772e-06,
      "loss": 0.37839035987854003,
      "mean_token_accuracy": 0.8904088377952576,
      "num_tokens": 68807652.0,
      "step": 5240
    },
    {
      "entropy": 3.2203125,
      "epoch": 1.4747191011235956,
      "grad_norm": 1.2034536600112915,
      "learning_rate": 1.9581379717892748e-06,
      "loss": 0.40187907218933105,
      "mean_token_accuracy": 0.8828882694244384,
      "num_tokens": 68939020.0,
      "step": 5250
    },
    {
      "entropy": 3.203125,
      "epoch": 1.4775280898876404,
      "grad_norm": 1.2613860368728638,
      "learning_rate": 1.9387197813658092e-06,
      "loss": 0.3905703783035278,
      "mean_token_accuracy": 0.8865590333938599,
      "num_tokens": 69071161.0,
      "step": 5260
    },
    {
      "entropy": 3.2,
      "epoch": 1.4803370786516854,
      "grad_norm": 1.2193385362625122,
      "learning_rate": 1.919375170463052e-06,
      "loss": 0.3922713279724121,
      "mean_token_accuracy": 0.8870265364646912,
      "num_tokens": 69202197.0,
      "step": 5270
    },
    {
      "entropy": 3.115625,
      "epoch": 1.4831460674157304,
      "grad_norm": 1.0939764976501465,
      "learning_rate": 1.9001046040391558e-06,
      "loss": 0.3809997081756592,
      "mean_token_accuracy": 0.8887838184833526,
      "num_tokens": 69334347.0,
      "step": 5280
    },
    {
      "entropy": 3.06875,
      "epoch": 1.4859550561797752,
      "grad_norm": 1.1265596151351929,
      "learning_rate": 1.8809085452725744e-06,
      "loss": 0.38692612648010255,
      "mean_token_accuracy": 0.8882668733596801,
      "num_tokens": 69465420.0,
      "step": 5290
    },
    {
      "entropy": 3.0578125,
      "epoch": 1.4887640449438202,
      "grad_norm": 1.2461986541748047,
      "learning_rate": 1.8617874555509342e-06,
      "loss": 0.39112081527709963,
      "mean_token_accuracy": 0.8852463662624359,
      "num_tokens": 69596303.0,
      "step": 5300
    },
    {
      "epoch": 1.4887640449438202,
      "eval_entropy": 3.13272909252669,
      "eval_loss": 0.43504172563552856,
      "eval_mean_token_accuracy": 0.8788616903311841,
      "eval_num_tokens": 69596303.0,
      "eval_runtime": 114.3425,
      "eval_samples_per_second": 137.587,
      "eval_steps_per_second": 2.458,
      "step": 5300
    },
    {
      "entropy": 3.203125,
      "epoch": 1.4915730337078652,
      "grad_norm": 1.3385220766067505,
      "learning_rate": 1.8427417944599325e-06,
      "loss": 0.3960011720657349,
      "mean_token_accuracy": 0.8853031098842621,
      "num_tokens": 69727973.0,
      "step": 5310
    },
    {
      "entropy": 3.2734375,
      "epoch": 1.49438202247191,
      "grad_norm": 2.045070171356201,
      "learning_rate": 1.8237720197723075e-06,
      "loss": 0.3876666069030762,
      "mean_token_accuracy": 0.8872966825962066,
      "num_tokens": 69858431.0,
      "step": 5320
    },
    {
      "entropy": 3.1734375,
      "epoch": 1.497191011235955,
      "grad_norm": 1.4479190111160278,
      "learning_rate": 1.8048785874368191e-06,
      "loss": 0.3656523942947388,
      "mean_token_accuracy": 0.8931134462356567,
      "num_tokens": 69989615.0,
      "step": 5330
    },
    {
      "entropy": 3.1546875,
      "epoch": 1.5,
      "grad_norm": 1.4309070110321045,
      "learning_rate": 1.7860619515673034e-06,
      "loss": 0.39683573246002196,
      "mean_token_accuracy": 0.8861665070056916,
      "num_tokens": 70121276.0,
      "step": 5340
    },
    {
      "entropy": 3.221875,
      "epoch": 1.5028089887640448,
      "grad_norm": 1.2095012664794922,
      "learning_rate": 1.7673225644317487e-06,
      "loss": 0.3868628263473511,
      "mean_token_accuracy": 0.8882637560367584,
      "num_tokens": 70252723.0,
      "step": 5350
    },
    {
      "entropy": 3.25625,
      "epoch": 1.50561797752809,
      "grad_norm": 1.304080605506897,
      "learning_rate": 1.748660876441428e-06,
      "loss": 0.3885697364807129,
      "mean_token_accuracy": 0.8886791825294494,
      "num_tokens": 70383726.0,
      "step": 5360
    },
    {
      "entropy": 3.2390625,
      "epoch": 1.5084269662921348,
      "grad_norm": 1.2994922399520874,
      "learning_rate": 1.7300773361400746e-06,
      "loss": 0.39111104011535647,
      "mean_token_accuracy": 0.8852004528045654,
      "num_tokens": 70514232.0,
      "step": 5370
    },
    {
      "entropy": 3.246875,
      "epoch": 1.5112359550561798,
      "grad_norm": 1.1440037488937378,
      "learning_rate": 1.711572390193102e-06,
      "loss": 0.39958460330963136,
      "mean_token_accuracy": 0.8859919130802154,
      "num_tokens": 70646084.0,
      "step": 5380
    },
    {
      "entropy": 3.2484375,
      "epoch": 1.5140449438202248,
      "grad_norm": 1.1734848022460938,
      "learning_rate": 1.6931464833768624e-06,
      "loss": 0.410199499130249,
      "mean_token_accuracy": 0.8824083685874939,
      "num_tokens": 70777174.0,
      "step": 5390
    },
    {
      "entropy": 3.28125,
      "epoch": 1.5168539325842696,
      "grad_norm": 1.3031336069107056,
      "learning_rate": 1.6748000585679602e-06,
      "loss": 0.3950439453125,
      "mean_token_accuracy": 0.8837095201015472,
      "num_tokens": 70908218.0,
      "step": 5400
    },
    {
      "epoch": 1.5168539325842696,
      "eval_entropy": 3.2950400355871885,
      "eval_loss": 0.43345770239830017,
      "eval_mean_token_accuracy": 0.8790701277739636,
      "eval_num_tokens": 70908218.0,
      "eval_runtime": 106.7249,
      "eval_samples_per_second": 147.407,
      "eval_steps_per_second": 2.633,
      "step": 5400
    },
    {
      "entropy": 3.30625,
      "epoch": 1.5196629213483146,
      "grad_norm": 1.259217381477356,
      "learning_rate": 1.6565335567326112e-06,
      "loss": 0.4049022674560547,
      "mean_token_accuracy": 0.8820204317569733,
      "num_tokens": 71039523.0,
      "step": 5410
    },
    {
      "entropy": 3.25,
      "epoch": 1.5224719101123596,
      "grad_norm": 1.2236769199371338,
      "learning_rate": 1.6383474169160334e-06,
      "loss": 0.3802346229553223,
      "mean_token_accuracy": 0.8923309803009033,
      "num_tokens": 71172166.0,
      "step": 5420
    },
    {
      "entropy": 3.2828125,
      "epoch": 1.5252808988764044,
      "grad_norm": 1.2794389724731445,
      "learning_rate": 1.6202420762319065e-06,
      "loss": 0.4148292064666748,
      "mean_token_accuracy": 0.8811677753925323,
      "num_tokens": 71303551.0,
      "step": 5430
    },
    {
      "entropy": 3.2640625,
      "epoch": 1.5280898876404494,
      "grad_norm": 1.227354884147644,
      "learning_rate": 1.6022179698518525e-06,
      "loss": 0.4008466720581055,
      "mean_token_accuracy": 0.88446324467659,
      "num_tokens": 71435365.0,
      "step": 5440
    },
    {
      "entropy": 3.296875,
      "epoch": 1.5308988764044944,
      "grad_norm": 1.2043499946594238,
      "learning_rate": 1.584275530994991e-06,
      "loss": 0.3980281352996826,
      "mean_token_accuracy": 0.8865043997764588,
      "num_tokens": 71566448.0,
      "step": 5450
    },
    {
      "entropy": 3.3125,
      "epoch": 1.5337078651685392,
      "grad_norm": 1.234902024269104,
      "learning_rate": 1.5664151909175124e-06,
      "loss": 0.38166399002075196,
      "mean_token_accuracy": 0.8889816045761109,
      "num_tokens": 71698087.0,
      "step": 5460
    },
    {
      "entropy": 3.3109375,
      "epoch": 1.5365168539325844,
      "grad_norm": 1.3395627737045288,
      "learning_rate": 1.5486373789023206e-06,
      "loss": 0.39968094825744627,
      "mean_token_accuracy": 0.8830544352531433,
      "num_tokens": 71829350.0,
      "step": 5470
    },
    {
      "entropy": 3.2625,
      "epoch": 1.5393258426966292,
      "grad_norm": 1.3813813924789429,
      "learning_rate": 1.5309425222487119e-06,
      "loss": 0.39154715538024903,
      "mean_token_accuracy": 0.8846165716648102,
      "num_tokens": 71959575.0,
      "step": 5480
    },
    {
      "entropy": 3.2421875,
      "epoch": 1.5421348314606742,
      "grad_norm": 1.082271933555603,
      "learning_rate": 1.5133310462621103e-06,
      "loss": 0.40457496643066404,
      "mean_token_accuracy": 0.8840779602527619,
      "num_tokens": 72090067.0,
      "step": 5490
    },
    {
      "entropy": 3.3109375,
      "epoch": 1.5449438202247192,
      "grad_norm": 1.3384177684783936,
      "learning_rate": 1.4958033742438348e-06,
      "loss": 0.3898488998413086,
      "mean_token_accuracy": 0.8876750588417053,
      "num_tokens": 72221491.0,
      "step": 5500
    },
    {
      "epoch": 1.5449438202247192,
      "eval_entropy": 3.3295707295373664,
      "eval_loss": 0.43265220522880554,
      "eval_mean_token_accuracy": 0.8791950294131486,
      "eval_num_tokens": 72221491.0,
      "eval_runtime": 119.5129,
      "eval_samples_per_second": 131.634,
      "eval_steps_per_second": 2.351,
      "step": 5500
    },
    {
      "entropy": 3.3,
      "epoch": 1.547752808988764,
      "grad_norm": 1.2310550212860107,
      "learning_rate": 1.478359927480935e-06,
      "loss": 0.3895530700683594,
      "mean_token_accuracy": 0.8870162963867188,
      "num_tokens": 72353688.0,
      "step": 5510
    },
    {
      "entropy": 3.1796875,
      "epoch": 1.550561797752809,
      "grad_norm": 1.2941917181015015,
      "learning_rate": 1.4610011252360594e-06,
      "loss": 0.3931793928146362,
      "mean_token_accuracy": 0.8852518558502197,
      "num_tokens": 72485325.0,
      "step": 5520
    },
    {
      "entropy": 3.1796875,
      "epoch": 1.553370786516854,
      "grad_norm": 1.483259677886963,
      "learning_rate": 1.4437273847373778e-06,
      "loss": 0.4024465084075928,
      "mean_token_accuracy": 0.8847292900085449,
      "num_tokens": 72618058.0,
      "step": 5530
    },
    {
      "entropy": 3.271875,
      "epoch": 1.5561797752808988,
      "grad_norm": 1.3513277769088745,
      "learning_rate": 1.4265391211685597e-06,
      "loss": 0.39669370651245117,
      "mean_token_accuracy": 0.8865042626857758,
      "num_tokens": 72749806.0,
      "step": 5540
    },
    {
      "entropy": 3.290625,
      "epoch": 1.5589887640449438,
      "grad_norm": 1.2254847288131714,
      "learning_rate": 1.4094367476587867e-06,
      "loss": 0.4046636581420898,
      "mean_token_accuracy": 0.8813999354839325,
      "num_tokens": 72880818.0,
      "step": 5550
    },
    {
      "entropy": 3.2171875,
      "epoch": 1.5617977528089888,
      "grad_norm": 0.9283570051193237,
      "learning_rate": 1.3924206752728282e-06,
      "loss": 0.38408446311950684,
      "mean_token_accuracy": 0.8887440204620362,
      "num_tokens": 73012945.0,
      "step": 5560
    },
    {
      "entropy": 3.14375,
      "epoch": 1.5646067415730336,
      "grad_norm": 1.1806893348693848,
      "learning_rate": 1.3754913130011566e-06,
      "loss": 0.3869128704071045,
      "mean_token_accuracy": 0.8883541762828827,
      "num_tokens": 73144155.0,
      "step": 5570
    },
    {
      "entropy": 3.16875,
      "epoch": 1.5674157303370788,
      "grad_norm": 1.1321043968200684,
      "learning_rate": 1.3586490677501269e-06,
      "loss": 0.38425674438476565,
      "mean_token_accuracy": 0.8875288009643555,
      "num_tokens": 73275712.0,
      "step": 5580
    },
    {
      "entropy": 3.2671875,
      "epoch": 1.5702247191011236,
      "grad_norm": 1.1926648616790771,
      "learning_rate": 1.3418943443321807e-06,
      "loss": 0.38610775470733644,
      "mean_token_accuracy": 0.8883664429187774,
      "num_tokens": 73406482.0,
      "step": 5590
    },
    {
      "entropy": 3.2640625,
      "epoch": 1.5730337078651684,
      "grad_norm": 1.1891111135482788,
      "learning_rate": 1.3252275454561337e-06,
      "loss": 0.38506450653076174,
      "mean_token_accuracy": 0.8876772880554199,
      "num_tokens": 73537591.0,
      "step": 5600
    },
    {
      "epoch": 1.5730337078651684,
      "eval_entropy": 3.1947842526690393,
      "eval_loss": 0.43231961131095886,
      "eval_mean_token_accuracy": 0.8794745792273525,
      "eval_num_tokens": 73537591.0,
      "eval_runtime": 118.5968,
      "eval_samples_per_second": 132.651,
      "eval_steps_per_second": 2.369,
      "step": 5600
    },
    {
      "entropy": 3.19375,
      "epoch": 1.5758426966292136,
      "grad_norm": 1.2878915071487427,
      "learning_rate": 1.30864907171748e-06,
      "loss": 0.403409481048584,
      "mean_token_accuracy": 0.8831920087337494,
      "num_tokens": 73668634.0,
      "step": 5610
    },
    {
      "entropy": 3.234375,
      "epoch": 1.5786516853932584,
      "grad_norm": 1.1634392738342285,
      "learning_rate": 1.292159321588778e-06,
      "loss": 0.3802034854888916,
      "mean_token_accuracy": 0.8896283030509948,
      "num_tokens": 73799328.0,
      "step": 5620
    },
    {
      "entropy": 3.178125,
      "epoch": 1.5814606741573034,
      "grad_norm": 1.0587197542190552,
      "learning_rate": 1.2757586914100612e-06,
      "loss": 0.3905323028564453,
      "mean_token_accuracy": 0.8859264731407166,
      "num_tokens": 73930397.0,
      "step": 5630
    },
    {
      "entropy": 3.2578125,
      "epoch": 1.5842696629213484,
      "grad_norm": 1.3114968538284302,
      "learning_rate": 1.2594475753793211e-06,
      "loss": 0.40627012252807615,
      "mean_token_accuracy": 0.8835435330867767,
      "num_tokens": 74062890.0,
      "step": 5640
    },
    {
      "entropy": 3.309375,
      "epoch": 1.5870786516853932,
      "grad_norm": 1.2347261905670166,
      "learning_rate": 1.243226365543026e-06,
      "loss": 0.3782741785049438,
      "mean_token_accuracy": 0.8899897336959839,
      "num_tokens": 74193804.0,
      "step": 5650
    },
    {
      "entropy": 3.20625,
      "epoch": 1.5898876404494382,
      "grad_norm": 1.2206196784973145,
      "learning_rate": 1.2270954517867e-06,
      "loss": 0.3931741714477539,
      "mean_token_accuracy": 0.88639235496521,
      "num_tokens": 74326014.0,
      "step": 5660
    },
    {
      "entropy": 3.14375,
      "epoch": 1.5926966292134832,
      "grad_norm": 1.1728156805038452,
      "learning_rate": 1.211055221825554e-06,
      "loss": 0.3860816478729248,
      "mean_token_accuracy": 0.8863325417041779,
      "num_tokens": 74458084.0,
      "step": 5670
    },
    {
      "entropy": 3.203125,
      "epoch": 1.595505617977528,
      "grad_norm": 1.2966103553771973,
      "learning_rate": 1.1951060611951615e-06,
      "loss": 0.3932689905166626,
      "mean_token_accuracy": 0.8871618807315826,
      "num_tokens": 74589948.0,
      "step": 5680
    },
    {
      "entropy": 3.1953125,
      "epoch": 1.598314606741573,
      "grad_norm": 1.208764672279358,
      "learning_rate": 1.1792483532422021e-06,
      "loss": 0.3951655626296997,
      "mean_token_accuracy": 0.8858476161956788,
      "num_tokens": 74721066.0,
      "step": 5690
    },
    {
      "entropy": 3.1921875,
      "epoch": 1.601123595505618,
      "grad_norm": 1.178815484046936,
      "learning_rate": 1.1634824791152334e-06,
      "loss": 0.39367165565490725,
      "mean_token_accuracy": 0.8862790644168854,
      "num_tokens": 74852887.0,
      "step": 5700
    },
    {
      "epoch": 1.601123595505618,
      "eval_entropy": 3.165202402135231,
      "eval_loss": 0.43144819140434265,
      "eval_mean_token_accuracy": 0.8795525001037163,
      "eval_num_tokens": 74852887.0,
      "eval_runtime": 118.2729,
      "eval_samples_per_second": 133.014,
      "eval_steps_per_second": 2.376,
      "step": 5700
    },
    {
      "entropy": 3.1546875,
      "epoch": 1.6039325842696628,
      "grad_norm": 1.4627737998962402,
      "learning_rate": 1.147808817755544e-06,
      "loss": 0.3850782632827759,
      "mean_token_accuracy": 0.8901103496551513,
      "num_tokens": 74983592.0,
      "step": 5710
    },
    {
      "entropy": 3.1265625,
      "epoch": 1.606741573033708,
      "grad_norm": 1.3125560283660889,
      "learning_rate": 1.1322277458880337e-06,
      "loss": 0.37709972858428953,
      "mean_token_accuracy": 0.8900519132614135,
      "num_tokens": 75114179.0,
      "step": 5720
    },
    {
      "entropy": 3.15,
      "epoch": 1.6095505617977528,
      "grad_norm": 1.2019294500350952,
      "learning_rate": 1.116739638012168e-06,
      "loss": 0.3945604801177979,
      "mean_token_accuracy": 0.8863421976566315,
      "num_tokens": 75244927.0,
      "step": 5730
    },
    {
      "entropy": 3.128125,
      "epoch": 1.6123595505617978,
      "grad_norm": 1.066537618637085,
      "learning_rate": 1.1013448663929704e-06,
      "loss": 0.3904740333557129,
      "mean_token_accuracy": 0.8850284099578858,
      "num_tokens": 75377280.0,
      "step": 5740
    },
    {
      "entropy": 3.2015625,
      "epoch": 1.6151685393258428,
      "grad_norm": 1.2805153131484985,
      "learning_rate": 1.0860438010520773e-06,
      "loss": 0.3831620454788208,
      "mean_token_accuracy": 0.8882053375244141,
      "num_tokens": 75508360.0,
      "step": 5750
    },
    {
      "entropy": 3.1765625,
      "epoch": 1.6179775280898876,
      "grad_norm": 1.3139517307281494,
      "learning_rate": 1.0708368097588435e-06,
      "loss": 0.38374814987182615,
      "mean_token_accuracy": 0.8877224564552307,
      "num_tokens": 75640071.0,
      "step": 5760
    },
    {
      "entropy": 3.1984375,
      "epoch": 1.6207865168539326,
      "grad_norm": 1.832893967628479,
      "learning_rate": 1.0557242580215066e-06,
      "loss": 0.3936143398284912,
      "mean_token_accuracy": 0.8843203365802765,
      "num_tokens": 75771149.0,
      "step": 5770
    },
    {
      "entropy": 3.134375,
      "epoch": 1.6235955056179776,
      "grad_norm": 1.2032394409179688,
      "learning_rate": 1.040706509078394e-06,
      "loss": 0.41098995208740235,
      "mean_token_accuracy": 0.8798267185688019,
      "num_tokens": 75903035.0,
      "step": 5780
    },
    {
      "entropy": 3.1375,
      "epoch": 1.6264044943820224,
      "grad_norm": 1.2080698013305664,
      "learning_rate": 1.025783923889202e-06,
      "loss": 0.4032297611236572,
      "mean_token_accuracy": 0.8838773429393768,
      "num_tokens": 76034783.0,
      "step": 5790
    },
    {
      "entropy": 3.1875,
      "epoch": 1.6292134831460674,
      "grad_norm": 1.1789093017578125,
      "learning_rate": 1.0109568611263094e-06,
      "loss": 0.39084458351135254,
      "mean_token_accuracy": 0.8861207127571106,
      "num_tokens": 76165582.0,
      "step": 5800
    },
    {
      "epoch": 1.6292134831460674,
      "eval_entropy": 3.188723309608541,
      "eval_loss": 0.4310080111026764,
      "eval_mean_token_accuracy": 0.8796484990052057,
      "eval_num_tokens": 76165582.0,
      "eval_runtime": 108.2285,
      "eval_samples_per_second": 145.359,
      "eval_steps_per_second": 2.596,
      "step": 5800
    },
    {
      "entropy": 3.14375,
      "epoch": 1.6320224719101124,
      "grad_norm": 1.3102200031280518,
      "learning_rate": 9.96225677166166e-07,
      "loss": 0.39276602268218996,
      "mean_token_accuracy": 0.8869596183300018,
      "num_tokens": 76297040.0,
      "step": 5810
    },
    {
      "entropy": 3.11875,
      "epoch": 1.6348314606741572,
      "grad_norm": 1.3742560148239136,
      "learning_rate": 9.81590726080721e-07,
      "loss": 0.3828797578811646,
      "mean_token_accuracy": 0.8907691597938537,
      "num_tokens": 76428907.0,
      "step": 5820
    },
    {
      "entropy": 3.1859375,
      "epoch": 1.6376404494382022,
      "grad_norm": 1.3238019943237305,
      "learning_rate": 9.670523596289138e-07,
      "loss": 0.3899382591247559,
      "mean_token_accuracy": 0.8861070334911346,
      "num_tokens": 76559666.0,
      "step": 5830
    },
    {
      "entropy": 3.2125,
      "epoch": 1.6404494382022472,
      "grad_norm": 1.8754230737686157,
      "learning_rate": 9.526109272482237e-07,
      "loss": 0.4060515880584717,
      "mean_token_accuracy": 0.8814957022666932,
      "num_tokens": 76690253.0,
      "step": 5840
    },
    {
      "entropy": 3.3046875,
      "epoch": 1.643258426966292,
      "grad_norm": 1.5277117490768433,
      "learning_rate": 9.382667760462633e-07,
      "loss": 0.39259862899780273,
      "mean_token_accuracy": 0.887138557434082,
      "num_tokens": 76820337.0,
      "step": 5850
    },
    {
      "entropy": 3.2390625,
      "epoch": 1.6460674157303372,
      "grad_norm": 1.2744336128234863,
      "learning_rate": 9.240202507924412e-07,
      "loss": 0.40212836265563967,
      "mean_token_accuracy": 0.8842680633068085,
      "num_tokens": 76951995.0,
      "step": 5860
    },
    {
      "entropy": 3.215625,
      "epoch": 1.648876404494382,
      "grad_norm": 1.2882815599441528,
      "learning_rate": 9.098716939096719e-07,
      "loss": 0.37579348087310793,
      "mean_token_accuracy": 0.89112468957901,
      "num_tokens": 77082825.0,
      "step": 5870
    },
    {
      "entropy": 3.221875,
      "epoch": 1.651685393258427,
      "grad_norm": 1.2383270263671875,
      "learning_rate": 8.958214454661529e-07,
      "loss": 0.3934124231338501,
      "mean_token_accuracy": 0.8869007468223572,
      "num_tokens": 77214596.0,
      "step": 5880
    },
    {
      "entropy": 3.2359375,
      "epoch": 1.654494382022472,
      "grad_norm": 1.3670285940170288,
      "learning_rate": 8.818698431671774e-07,
      "loss": 0.3989357233047485,
      "mean_token_accuracy": 0.8859530508518219,
      "num_tokens": 77346939.0,
      "step": 5890
    },
    {
      "entropy": 3.2546875,
      "epoch": 1.6573033707865168,
      "grad_norm": 1.2166470289230347,
      "learning_rate": 8.680172223470329e-07,
      "loss": 0.383312726020813,
      "mean_token_accuracy": 0.8883917152881622,
      "num_tokens": 77478641.0,
      "step": 5900
    },
    {
      "epoch": 1.6573033707865168,
      "eval_entropy": 3.2660142348754446,
      "eval_loss": 0.43034791946411133,
      "eval_mean_token_accuracy": 0.8797890312731054,
      "eval_num_tokens": 77478641.0,
      "eval_runtime": 116.4481,
      "eval_samples_per_second": 135.099,
      "eval_steps_per_second": 2.413,
      "step": 5900
    },
    {
      "entropy": 3.2515625,
      "epoch": 1.6601123595505618,
      "grad_norm": 1.4173338413238525,
      "learning_rate": 8.542639159609278e-07,
      "loss": 0.38932361602783205,
      "mean_token_accuracy": 0.8867664277553559,
      "num_tokens": 77610398.0,
      "step": 5910
    },
    {
      "entropy": 3.178125,
      "epoch": 1.6629213483146068,
      "grad_norm": 1.0318958759307861,
      "learning_rate": 8.406102545769989e-07,
      "loss": 0.38496315479278564,
      "mean_token_accuracy": 0.8893871486186982,
      "num_tokens": 77741294.0,
      "step": 5920
    },
    {
      "entropy": 3.16875,
      "epoch": 1.6657303370786516,
      "grad_norm": 1.2144213914871216,
      "learning_rate": 8.270565663683583e-07,
      "loss": 0.3768044948577881,
      "mean_token_accuracy": 0.8892023384571075,
      "num_tokens": 77872036.0,
      "step": 5930
    },
    {
      "entropy": 3.178125,
      "epoch": 1.6685393258426966,
      "grad_norm": 1.2177233695983887,
      "learning_rate": 8.13603177105211e-07,
      "loss": 0.3857748985290527,
      "mean_token_accuracy": 0.8877400815486908,
      "num_tokens": 78003496.0,
      "step": 5940
    },
    {
      "entropy": 3.234375,
      "epoch": 1.6713483146067416,
      "grad_norm": 1.2058058977127075,
      "learning_rate": 8.002504101470204e-07,
      "loss": 0.39029593467712403,
      "mean_token_accuracy": 0.8879157900810242,
      "num_tokens": 78134718.0,
      "step": 5950
    },
    {
      "entropy": 3.2375,
      "epoch": 1.6741573033707864,
      "grad_norm": 1.3687621355056763,
      "learning_rate": 7.869985864347424e-07,
      "loss": 0.3992608547210693,
      "mean_token_accuracy": 0.8834168076515198,
      "num_tokens": 78265946.0,
      "step": 5960
    },
    {
      "entropy": 3.2390625,
      "epoch": 1.6769662921348316,
      "grad_norm": 1.0915995836257935,
      "learning_rate": 7.738480244831042e-07,
      "loss": 0.3955454111099243,
      "mean_token_accuracy": 0.883608877658844,
      "num_tokens": 78398230.0,
      "step": 5970
    },
    {
      "entropy": 3.259375,
      "epoch": 1.6797752808988764,
      "grad_norm": 1.162134051322937,
      "learning_rate": 7.607990403729526e-07,
      "loss": 0.37786307334899905,
      "mean_token_accuracy": 0.8896137297153472,
      "num_tokens": 78528518.0,
      "step": 5980
    },
    {
      "entropy": 3.234375,
      "epoch": 1.6825842696629212,
      "grad_norm": 1.211387038230896,
      "learning_rate": 7.47851947743658e-07,
      "loss": 0.3835395097732544,
      "mean_token_accuracy": 0.8892298936843872,
      "num_tokens": 78660097.0,
      "step": 5990
    },
    {
      "entropy": 3.2671875,
      "epoch": 1.6853932584269664,
      "grad_norm": 1.3493833541870117,
      "learning_rate": 7.350070577855716e-07,
      "loss": 0.3885589838027954,
      "mean_token_accuracy": 0.8878313183784485,
      "num_tokens": 78791235.0,
      "step": 6000
    },
    {
      "epoch": 1.6853932584269664,
      "eval_entropy": 3.194228202846975,
      "eval_loss": 0.4304349422454834,
      "eval_mean_token_accuracy": 0.8798338707231542,
      "eval_num_tokens": 78791235.0,
      "eval_runtime": 115.7748,
      "eval_samples_per_second": 135.884,
      "eval_steps_per_second": 2.427,
      "step": 6000
    },
    {
      "entropy": 3.1828125,
      "epoch": 1.6882022471910112,
      "grad_norm": 1.1839690208435059,
      "learning_rate": 7.222646792325516e-07,
      "loss": 0.39657154083251955,
      "mean_token_accuracy": 0.8845458447933197,
      "num_tokens": 78922094.0,
      "step": 6010
    },
    {
      "entropy": 3.1640625,
      "epoch": 1.6910112359550562,
      "grad_norm": 1.3457016944885254,
      "learning_rate": 7.096251183545355e-07,
      "loss": 0.39935686588287356,
      "mean_token_accuracy": 0.8847565472126007,
      "num_tokens": 79052951.0,
      "step": 6020
    },
    {
      "entropy": 3.153125,
      "epoch": 1.6938202247191012,
      "grad_norm": 1.2550835609436035,
      "learning_rate": 6.970886789501851e-07,
      "loss": 0.4017538070678711,
      "mean_token_accuracy": 0.8838404595851899,
      "num_tokens": 79184343.0,
      "step": 6030
    },
    {
      "entropy": 3.19375,
      "epoch": 1.696629213483146,
      "grad_norm": 1.3737390041351318,
      "learning_rate": 6.846556623395795e-07,
      "loss": 0.37827134132385254,
      "mean_token_accuracy": 0.8887855470180511,
      "num_tokens": 79315055.0,
      "step": 6040
    },
    {
      "entropy": 3.1953125,
      "epoch": 1.699438202247191,
      "grad_norm": 1.248664140701294,
      "learning_rate": 6.723263673569796e-07,
      "loss": 0.371117901802063,
      "mean_token_accuracy": 0.8926220059394836,
      "num_tokens": 79445831.0,
      "step": 6050
    },
    {
      "entropy": 3.196875,
      "epoch": 1.702247191011236,
      "grad_norm": 1.2837352752685547,
      "learning_rate": 6.601010903436355e-07,
      "loss": 0.39250712394714354,
      "mean_token_accuracy": 0.8879317522048951,
      "num_tokens": 79576737.0,
      "step": 6060
    },
    {
      "entropy": 3.1984375,
      "epoch": 1.7050561797752808,
      "grad_norm": 2.19681978225708,
      "learning_rate": 6.479801251406748e-07,
      "loss": 0.37683782577514646,
      "mean_token_accuracy": 0.8896573185920715,
      "num_tokens": 79707176.0,
      "step": 6070
    },
    {
      "entropy": 3.19375,
      "epoch": 1.7078651685393258,
      "grad_norm": 1.2631367444992065,
      "learning_rate": 6.359637630820292e-07,
      "loss": 0.3923129320144653,
      "mean_token_accuracy": 0.8865497350692749,
      "num_tokens": 79838385.0,
      "step": 6080
    },
    {
      "entropy": 3.15,
      "epoch": 1.7106741573033708,
      "grad_norm": 1.2925978899002075,
      "learning_rate": 6.24052292987442e-07,
      "loss": 0.40004520416259765,
      "mean_token_accuracy": 0.8843684732913971,
      "num_tokens": 79969761.0,
      "step": 6090
    },
    {
      "entropy": 3.1421875,
      "epoch": 1.7134831460674156,
      "grad_norm": 1.0745518207550049,
      "learning_rate": 6.122460011555187e-07,
      "loss": 0.3853758811950684,
      "mean_token_accuracy": 0.8886780917644501,
      "num_tokens": 80100788.0,
      "step": 6100
    },
    {
      "epoch": 1.7134831460674156,
      "eval_entropy": 3.1719306049822062,
      "eval_loss": 0.4298255443572998,
      "eval_mean_token_accuracy": 0.8799454758599984,
      "eval_num_tokens": 80100788.0,
      "eval_runtime": 112.623,
      "eval_samples_per_second": 139.687,
      "eval_steps_per_second": 2.495,
      "step": 6100
    },
    {
      "entropy": 3.196875,
      "epoch": 1.7162921348314608,
      "grad_norm": 1.3174694776535034,
      "learning_rate": 6.005451713568505e-07,
      "loss": 0.3894709587097168,
      "mean_token_accuracy": 0.8877410292625427,
      "num_tokens": 80232348.0,
      "step": 6110
    },
    {
      "entropy": 3.25625,
      "epoch": 1.7191011235955056,
      "grad_norm": 1.3446450233459473,
      "learning_rate": 5.889500848271901e-07,
      "loss": 0.3868286371231079,
      "mean_token_accuracy": 0.8882597923278809,
      "num_tokens": 80362964.0,
      "step": 6120
    },
    {
      "entropy": 3.2265625,
      "epoch": 1.7219101123595506,
      "grad_norm": 1.275291085243225,
      "learning_rate": 5.774610202606939e-07,
      "loss": 0.3816858768463135,
      "mean_token_accuracy": 0.8896496772766114,
      "num_tokens": 80494174.0,
      "step": 6130
    },
    {
      "entropy": 3.190625,
      "epoch": 1.7247191011235956,
      "grad_norm": 1.1371703147888184,
      "learning_rate": 5.660782538032245e-07,
      "loss": 0.38695707321166994,
      "mean_token_accuracy": 0.8883748471736908,
      "num_tokens": 80626472.0,
      "step": 6140
    },
    {
      "entropy": 3.1984375,
      "epoch": 1.7275280898876404,
      "grad_norm": 1.1520910263061523,
      "learning_rate": 5.548020590457098e-07,
      "loss": 0.3863677024841309,
      "mean_token_accuracy": 0.8871554136276245,
      "num_tokens": 80757861.0,
      "step": 6150
    },
    {
      "entropy": 3.1921875,
      "epoch": 1.7303370786516854,
      "grad_norm": 1.2423131465911865,
      "learning_rate": 5.436327070175729e-07,
      "loss": 0.38500347137451174,
      "mean_token_accuracy": 0.8875517010688782,
      "num_tokens": 80889177.0,
      "step": 6160
    },
    {
      "entropy": 3.184375,
      "epoch": 1.7331460674157304,
      "grad_norm": 1.3326811790466309,
      "learning_rate": 5.325704661802106e-07,
      "loss": 0.38845248222351075,
      "mean_token_accuracy": 0.888093239068985,
      "num_tokens": 81019729.0,
      "step": 6170
    },
    {
      "entropy": 3.134375,
      "epoch": 1.7359550561797752,
      "grad_norm": 1.1854122877120972,
      "learning_rate": 5.216156024205482e-07,
      "loss": 0.3897011041641235,
      "mean_token_accuracy": 0.8856644570827484,
      "num_tokens": 81152423.0,
      "step": 6180
    },
    {
      "entropy": 3.1875,
      "epoch": 1.7387640449438202,
      "grad_norm": 1.1079387664794922,
      "learning_rate": 5.107683790446411e-07,
      "loss": 0.382286524772644,
      "mean_token_accuracy": 0.8891783475875854,
      "num_tokens": 81284409.0,
      "step": 6190
    },
    {
      "entropy": 3.1765625,
      "epoch": 1.7415730337078652,
      "grad_norm": 1.293662190437317,
      "learning_rate": 5.000290567713533e-07,
      "loss": 0.391655421257019,
      "mean_token_accuracy": 0.8870730876922608,
      "num_tokens": 81414870.0,
      "step": 6200
    },
    {
      "epoch": 1.7415730337078652,
      "eval_entropy": 3.1559163701067616,
      "eval_loss": 0.42958641052246094,
      "eval_mean_token_accuracy": 0.8800417337553357,
      "eval_num_tokens": 81414870.0,
      "eval_runtime": 110.2582,
      "eval_samples_per_second": 142.683,
      "eval_steps_per_second": 2.549,
      "step": 6200
    },
    {
      "entropy": 3.1703125,
      "epoch": 1.74438202247191,
      "grad_norm": 1.2860597372055054,
      "learning_rate": 4.893978937260868e-07,
      "loss": 0.3851505994796753,
      "mean_token_accuracy": 0.889102578163147,
      "num_tokens": 81546001.0,
      "step": 6210
    },
    {
      "entropy": 3.1828125,
      "epoch": 1.7471910112359552,
      "grad_norm": 1.2604219913482666,
      "learning_rate": 4.788751454345763e-07,
      "loss": 0.3931628465652466,
      "mean_token_accuracy": 0.8845866084098816,
      "num_tokens": 81676705.0,
      "step": 6220
    },
    {
      "entropy": 3.1921875,
      "epoch": 1.75,
      "grad_norm": 1.337127685546875,
      "learning_rate": 4.6846106481675035e-07,
      "loss": 0.39096114635467527,
      "mean_token_accuracy": 0.8870581388473511,
      "num_tokens": 81808219.0,
      "step": 6230
    },
    {
      "entropy": 3.1625,
      "epoch": 1.7528089887640448,
      "grad_norm": 1.1952024698257446,
      "learning_rate": 4.581559021806542e-07,
      "loss": 0.3884358167648315,
      "mean_token_accuracy": 0.887700867652893,
      "num_tokens": 81940657.0,
      "step": 6240
    },
    {
      "entropy": 3.1828125,
      "epoch": 1.75561797752809,
      "grad_norm": 1.1326738595962524,
      "learning_rate": 4.4795990521642684e-07,
      "loss": 0.38137362003326414,
      "mean_token_accuracy": 0.8890506803989411,
      "num_tokens": 82072564.0,
      "step": 6250
    },
    {
      "entropy": 3.221875,
      "epoch": 1.7584269662921348,
      "grad_norm": 1.1880830526351929,
      "learning_rate": 4.378733189903528e-07,
      "loss": 0.3884089231491089,
      "mean_token_accuracy": 0.8872714757919311,
      "num_tokens": 82203068.0,
      "step": 6260
    },
    {
      "entropy": 3.2125,
      "epoch": 1.7612359550561798,
      "grad_norm": 1.2926839590072632,
      "learning_rate": 4.278963859389723e-07,
      "loss": 0.38527402877807615,
      "mean_token_accuracy": 0.8890871286392212,
      "num_tokens": 82334130.0,
      "step": 6270
    },
    {
      "entropy": 3.2015625,
      "epoch": 1.7640449438202248,
      "grad_norm": 1.3774532079696655,
      "learning_rate": 4.1802934586324897e-07,
      "loss": 0.3873408317565918,
      "mean_token_accuracy": 0.8877808511257171,
      "num_tokens": 82465333.0,
      "step": 6280
    },
    {
      "entropy": 3.1390625,
      "epoch": 1.7668539325842696,
      "grad_norm": 1.210107684135437,
      "learning_rate": 4.0827243592281294e-07,
      "loss": 0.3876020908355713,
      "mean_token_accuracy": 0.8869581997394562,
      "num_tokens": 82597295.0,
      "step": 6290
    },
    {
      "entropy": 3.184375,
      "epoch": 1.7696629213483146,
      "grad_norm": 1.3541264533996582,
      "learning_rate": 3.986258906302543e-07,
      "loss": 0.38852949142456056,
      "mean_token_accuracy": 0.8879195570945739,
      "num_tokens": 82729331.0,
      "step": 6300
    },
    {
      "epoch": 1.7696629213483146,
      "eval_entropy": 3.212967081850534,
      "eval_loss": 0.4293440282344818,
      "eval_mean_token_accuracy": 0.8800719690068337,
      "eval_num_tokens": 82729331.0,
      "eval_runtime": 111.3666,
      "eval_samples_per_second": 141.263,
      "eval_steps_per_second": 2.523,
      "step": 6300
    },
    {
      "entropy": 3.1984375,
      "epoch": 1.7724719101123596,
      "grad_norm": 1.2566618919372559,
      "learning_rate": 3.890899418454913e-07,
      "loss": 0.38900437355041506,
      "mean_token_accuracy": 0.887330812215805,
      "num_tokens": 82860698.0,
      "step": 6310
    },
    {
      "entropy": 3.1921875,
      "epoch": 1.7752808988764044,
      "grad_norm": 1.2893481254577637,
      "learning_rate": 3.796648187701957e-07,
      "loss": 0.3835317850112915,
      "mean_token_accuracy": 0.8876651406288147,
      "num_tokens": 82991286.0,
      "step": 6320
    },
    {
      "entropy": 3.140625,
      "epoch": 1.7780898876404494,
      "grad_norm": 1.3896664381027222,
      "learning_rate": 3.703507479422813e-07,
      "loss": 0.4013826370239258,
      "mean_token_accuracy": 0.8865646898746491,
      "num_tokens": 83122552.0,
      "step": 6330
    },
    {
      "entropy": 3.1453125,
      "epoch": 1.7808988764044944,
      "grad_norm": 1.2666797637939453,
      "learning_rate": 3.611479532304618e-07,
      "loss": 0.39299464225769043,
      "mean_token_accuracy": 0.8859761536121369,
      "num_tokens": 83253798.0,
      "step": 6340
    },
    {
      "entropy": 3.175,
      "epoch": 1.7837078651685392,
      "grad_norm": 1.1284576654434204,
      "learning_rate": 3.5205665582887296e-07,
      "loss": 0.38024125099182127,
      "mean_token_accuracy": 0.8896082520484925,
      "num_tokens": 83384296.0,
      "step": 6350
    },
    {
      "entropy": 3.1640625,
      "epoch": 1.7865168539325844,
      "grad_norm": 1.313480019569397,
      "learning_rate": 3.43077074251747e-07,
      "loss": 0.4098865509033203,
      "mean_token_accuracy": 0.8808252215385437,
      "num_tokens": 83516265.0,
      "step": 6360
    },
    {
      "entropy": 3.1859375,
      "epoch": 1.7893258426966292,
      "grad_norm": 1.2759302854537964,
      "learning_rate": 3.3420942432817127e-07,
      "loss": 0.37343926429748536,
      "mean_token_accuracy": 0.8910487771034241,
      "num_tokens": 83647549.0,
      "step": 6370
    },
    {
      "entropy": 3.1671875,
      "epoch": 1.7921348314606742,
      "grad_norm": 1.1181399822235107,
      "learning_rate": 3.2545391919689193e-07,
      "loss": 0.37969355583190917,
      "mean_token_accuracy": 0.8897684276103973,
      "num_tokens": 83780400.0,
      "step": 6380
    },
    {
      "entropy": 3.2125,
      "epoch": 1.7949438202247192,
      "grad_norm": 1.2292587757110596,
      "learning_rate": 3.1681076930119626e-07,
      "loss": 0.3844544172286987,
      "mean_token_accuracy": 0.889862322807312,
      "num_tokens": 83911422.0,
      "step": 6390
    },
    {
      "entropy": 3.1390625,
      "epoch": 1.797752808988764,
      "grad_norm": 1.3465687036514282,
      "learning_rate": 3.082801823838527e-07,
      "loss": 0.37302532196044924,
      "mean_token_accuracy": 0.8923552215099335,
      "num_tokens": 84042867.0,
      "step": 6400
    },
    {
      "epoch": 1.797752808988764,
      "eval_entropy": 3.1783807829181496,
      "eval_loss": 0.4290313124656677,
      "eval_mean_token_accuracy": 0.8802014063261582,
      "eval_num_tokens": 84042867.0,
      "eval_runtime": 104.8405,
      "eval_samples_per_second": 150.057,
      "eval_steps_per_second": 2.68,
      "step": 6400
    },
    {
      "entropy": 3.1671875,
      "epoch": 1.800561797752809,
      "grad_norm": 1.1289780139923096,
      "learning_rate": 2.9986236348211684e-07,
      "loss": 0.3786012649536133,
      "mean_token_accuracy": 0.8901362180709839,
      "num_tokens": 84174261.0,
      "step": 6410
    },
    {
      "entropy": 3.1859375,
      "epoch": 1.803370786516854,
      "grad_norm": 1.1964200735092163,
      "learning_rate": 2.915575149228056e-07,
      "loss": 0.3843598127365112,
      "mean_token_accuracy": 0.887418657541275,
      "num_tokens": 84305314.0,
      "step": 6420
    },
    {
      "entropy": 3.165625,
      "epoch": 1.8061797752808988,
      "grad_norm": 1.562326192855835,
      "learning_rate": 2.833658363174302e-07,
      "loss": 0.37695701122283937,
      "mean_token_accuracy": 0.8901226997375489,
      "num_tokens": 84436916.0,
      "step": 6430
    },
    {
      "entropy": 3.159375,
      "epoch": 1.8089887640449438,
      "grad_norm": 1.52724289894104,
      "learning_rate": 2.7528752455740606e-07,
      "loss": 0.3888294219970703,
      "mean_token_accuracy": 0.8864633440971375,
      "num_tokens": 84568131.0,
      "step": 6440
    },
    {
      "entropy": 3.1734375,
      "epoch": 1.8117977528089888,
      "grad_norm": 1.3424180746078491,
      "learning_rate": 2.6732277380930873e-07,
      "loss": 0.38145601749420166,
      "mean_token_accuracy": 0.8889956533908844,
      "num_tokens": 84698517.0,
      "step": 6450
    },
    {
      "entropy": 3.1703125,
      "epoch": 1.8146067415730336,
      "grad_norm": 1.2423303127288818,
      "learning_rate": 2.594717755102205e-07,
      "loss": 0.3772451877593994,
      "mean_token_accuracy": 0.8914096295833588,
      "num_tokens": 84830012.0,
      "step": 6460
    },
    {
      "entropy": 3.175,
      "epoch": 1.8174157303370788,
      "grad_norm": 1.3002800941467285,
      "learning_rate": 2.517347183631158e-07,
      "loss": 0.3898607730865479,
      "mean_token_accuracy": 0.8873790919780731,
      "num_tokens": 84960244.0,
      "step": 6470
    },
    {
      "entropy": 3.159375,
      "epoch": 1.8202247191011236,
      "grad_norm": 1.255793571472168,
      "learning_rate": 2.441117883323374e-07,
      "loss": 0.39146358966827394,
      "mean_token_accuracy": 0.8844018936157226,
      "num_tokens": 85091527.0,
      "step": 6480
    },
    {
      "entropy": 3.1890625,
      "epoch": 1.8230337078651684,
      "grad_norm": 1.36159348487854,
      "learning_rate": 2.3660316863911682e-07,
      "loss": 0.382525372505188,
      "mean_token_accuracy": 0.8890960931777954,
      "num_tokens": 85222254.0,
      "step": 6490
    },
    {
      "entropy": 3.1640625,
      "epoch": 1.8258426966292136,
      "grad_norm": 1.1689728498458862,
      "learning_rate": 2.292090397571789e-07,
      "loss": 0.3805016756057739,
      "mean_token_accuracy": 0.8900243997573852,
      "num_tokens": 85354291.0,
      "step": 6500
    },
    {
      "epoch": 1.8258426966292136,
      "eval_entropy": 3.177379893238434,
      "eval_loss": 0.4286893904209137,
      "eval_mean_token_accuracy": 0.8802066364746501,
      "eval_num_tokens": 85354291.0,
      "eval_runtime": 106.6385,
      "eval_samples_per_second": 147.526,
      "eval_steps_per_second": 2.635,
      "step": 6500
    },
    {
      "entropy": 3.159375,
      "epoch": 1.8286516853932584,
      "grad_norm": 1.1547116041183472,
      "learning_rate": 2.21929579408397e-07,
      "loss": 0.38411972522735593,
      "mean_token_accuracy": 0.8896714806556701,
      "num_tokens": 85485964.0,
      "step": 6510
    },
    {
      "entropy": 3.196875,
      "epoch": 1.8314606741573034,
      "grad_norm": 2.100903272628784,
      "learning_rate": 2.1476496255852685e-07,
      "loss": 0.37227981090545653,
      "mean_token_accuracy": 0.8914181232452393,
      "num_tokens": 85615876.0,
      "step": 6520
    },
    {
      "entropy": 3.128125,
      "epoch": 1.8342696629213484,
      "grad_norm": 1.1161167621612549,
      "learning_rate": 2.0771536141299565e-07,
      "loss": 0.392805290222168,
      "mean_token_accuracy": 0.8855752527713776,
      "num_tokens": 85748062.0,
      "step": 6530
    },
    {
      "entropy": 3.125,
      "epoch": 1.8370786516853932,
      "grad_norm": 1.3620573282241821,
      "learning_rate": 2.0078094541277016e-07,
      "loss": 0.3900078535079956,
      "mean_token_accuracy": 0.8874358057975769,
      "num_tokens": 85879819.0,
      "step": 6540
    },
    {
      "entropy": 3.1390625,
      "epoch": 1.8398876404494382,
      "grad_norm": 1.1815170049667358,
      "learning_rate": 1.9396188123027736e-07,
      "loss": 0.3912449598312378,
      "mean_token_accuracy": 0.8856905817985534,
      "num_tokens": 86011509.0,
      "step": 6550
    },
    {
      "entropy": 3.1609375,
      "epoch": 1.8426966292134832,
      "grad_norm": 1.2505420446395874,
      "learning_rate": 1.8725833276540095e-07,
      "loss": 0.3803662061691284,
      "mean_token_accuracy": 0.8909626841545105,
      "num_tokens": 86143565.0,
      "step": 6560
    },
    {
      "entropy": 3.1515625,
      "epoch": 1.845505617977528,
      "grad_norm": 1.1324859857559204,
      "learning_rate": 1.8067046114154386e-07,
      "loss": 0.38357589244842527,
      "mean_token_accuracy": 0.8884536385536194,
      "num_tokens": 86275157.0,
      "step": 6570
    },
    {
      "entropy": 3.1578125,
      "epoch": 1.848314606741573,
      "grad_norm": 1.332536220550537,
      "learning_rate": 1.7419842470175196e-07,
      "loss": 0.38278238773345946,
      "mean_token_accuracy": 0.8884197056293488,
      "num_tokens": 86406091.0,
      "step": 6580
    },
    {
      "entropy": 3.11875,
      "epoch": 1.851123595505618,
      "grad_norm": 1.3262176513671875,
      "learning_rate": 1.6784237900491163e-07,
      "loss": 0.3846198558807373,
      "mean_token_accuracy": 0.887979394197464,
      "num_tokens": 86537560.0,
      "step": 6590
    },
    {
      "entropy": 3.153125,
      "epoch": 1.8539325842696628,
      "grad_norm": 1.3944357633590698,
      "learning_rate": 1.6160247682200813e-07,
      "loss": 0.3969149112701416,
      "mean_token_accuracy": 0.8856332421302795,
      "num_tokens": 86668820.0,
      "step": 6600
    },
    {
      "epoch": 1.8539325842696628,
      "eval_entropy": 3.1421819395017794,
      "eval_loss": 0.4286288619041443,
      "eval_mean_token_accuracy": 0.8802983438417157,
      "eval_num_tokens": 86668820.0,
      "eval_runtime": 119.2373,
      "eval_samples_per_second": 131.939,
      "eval_steps_per_second": 2.357,
      "step": 6600
    },
    {
      "entropy": 3.165625,
      "epoch": 1.856741573033708,
      "grad_norm": 1.243005394935608,
      "learning_rate": 1.554788681324554e-07,
      "loss": 0.3752056360244751,
      "mean_token_accuracy": 0.8933846771717071,
      "num_tokens": 86800003.0,
      "step": 6610
    },
    {
      "entropy": 3.1375,
      "epoch": 1.8595505617977528,
      "grad_norm": 1.4393961429595947,
      "learning_rate": 1.4947170012048872e-07,
      "loss": 0.3869161605834961,
      "mean_token_accuracy": 0.8869595944881439,
      "num_tokens": 86930831.0,
      "step": 6620
    },
    {
      "entropy": 3.1203125,
      "epoch": 1.8623595505617978,
      "grad_norm": 1.4512866735458374,
      "learning_rate": 1.435811171716317e-07,
      "loss": 0.3894651889801025,
      "mean_token_accuracy": 0.888670802116394,
      "num_tokens": 87062325.0,
      "step": 6630
    },
    {
      "entropy": 3.153125,
      "epoch": 1.8651685393258428,
      "grad_norm": 1.2857064008712769,
      "learning_rate": 1.3780726086922103e-07,
      "loss": 0.3884331226348877,
      "mean_token_accuracy": 0.8902813017368316,
      "num_tokens": 87193012.0,
      "step": 6640
    },
    {
      "entropy": 3.1046875,
      "epoch": 1.8679775280898876,
      "grad_norm": 1.4063314199447632,
      "learning_rate": 1.3215026999100655e-07,
      "loss": 0.3881614923477173,
      "mean_token_accuracy": 0.8869928181171417,
      "num_tokens": 87324759.0,
      "step": 6650
    },
    {
      "entropy": 3.1234375,
      "epoch": 1.8707865168539326,
      "grad_norm": 1.3051425218582153,
      "learning_rate": 1.2661028050581446e-07,
      "loss": 0.3982240676879883,
      "mean_token_accuracy": 0.8868953824043274,
      "num_tokens": 87457236.0,
      "step": 6660
    },
    {
      "entropy": 3.1625,
      "epoch": 1.8735955056179776,
      "grad_norm": 1.2355676889419556,
      "learning_rate": 1.2118742557027885e-07,
      "loss": 0.3851290702819824,
      "mean_token_accuracy": 0.8887217998504638,
      "num_tokens": 87588769.0,
      "step": 6670
    },
    {
      "entropy": 3.1359375,
      "epoch": 1.8764044943820224,
      "grad_norm": 1.2653839588165283,
      "learning_rate": 1.1588183552564247e-07,
      "loss": 0.39154644012451173,
      "mean_token_accuracy": 0.8860277712345124,
      "num_tokens": 87719527.0,
      "step": 6680
    },
    {
      "entropy": 3.1625,
      "epoch": 1.8792134831460674,
      "grad_norm": 1.203320026397705,
      "learning_rate": 1.1069363789462273e-07,
      "loss": 0.39923944473266604,
      "mean_token_accuracy": 0.8839006364345551,
      "num_tokens": 87851747.0,
      "step": 6690
    },
    {
      "entropy": 3.175,
      "epoch": 1.8820224719101124,
      "grad_norm": 1.1897104978561401,
      "learning_rate": 1.0562295737834738e-07,
      "loss": 0.3971160173416138,
      "mean_token_accuracy": 0.8863435447216034,
      "num_tokens": 87983162.0,
      "step": 6700
    },
    {
      "epoch": 1.8820224719101124,
      "eval_entropy": 3.1671485765124556,
      "eval_loss": 0.4284597933292389,
      "eval_mean_token_accuracy": 0.8803274932705211,
      "eval_num_tokens": 87983162.0,
      "eval_runtime": 116.5216,
      "eval_samples_per_second": 135.014,
      "eval_steps_per_second": 2.412,
      "step": 6700
    },
    {
      "entropy": 3.1953125,
      "epoch": 1.8848314606741572,
      "grad_norm": 1.4615147113800049,
      "learning_rate": 1.0066991585335583e-07,
      "loss": 0.38571991920471194,
      "mean_token_accuracy": 0.8882844388484955,
      "num_tokens": 88113370.0,
      "step": 6710
    },
    {
      "entropy": 3.15,
      "epoch": 1.8876404494382022,
      "grad_norm": 1.1358058452606201,
      "learning_rate": 9.583463236867318e-08,
      "loss": 0.388935923576355,
      "mean_token_accuracy": 0.8872272610664368,
      "num_tokens": 88244550.0,
      "step": 6720
    },
    {
      "entropy": 3.19375,
      "epoch": 1.8904494382022472,
      "grad_norm": 1.1610984802246094,
      "learning_rate": 9.111722314294358e-08,
      "loss": 0.3826758623123169,
      "mean_token_accuracy": 0.8887862980365753,
      "num_tokens": 88375034.0,
      "step": 6730
    },
    {
      "entropy": 3.1421875,
      "epoch": 1.893258426966292,
      "grad_norm": 1.400122880935669,
      "learning_rate": 8.651780156164302e-08,
      "loss": 0.3786154270172119,
      "mean_token_accuracy": 0.891159200668335,
      "num_tokens": 88506347.0,
      "step": 6740
    },
    {
      "entropy": 3.1609375,
      "epoch": 1.8960674157303372,
      "grad_norm": 1.2274420261383057,
      "learning_rate": 8.203647817434823e-08,
      "loss": 0.3743563413619995,
      "mean_token_accuracy": 0.8918124437332153,
      "num_tokens": 88636871.0,
      "step": 6750
    },
    {
      "entropy": 3.1609375,
      "epoch": 1.898876404494382,
      "grad_norm": 1.273725152015686,
      "learning_rate": 7.76733606920832e-08,
      "loss": 0.38526599407196044,
      "mean_token_accuracy": 0.8878537595272065,
      "num_tokens": 88767927.0,
      "step": 6760
    },
    {
      "entropy": 3.18125,
      "epoch": 1.901685393258427,
      "grad_norm": 1.193949580192566,
      "learning_rate": 7.342855398472958e-08,
      "loss": 0.37825779914855956,
      "mean_token_accuracy": 0.8899289071559906,
      "num_tokens": 88898892.0,
      "step": 6770
    },
    {
      "entropy": 3.1359375,
      "epoch": 1.904494382022472,
      "grad_norm": 1.4476648569107056,
      "learning_rate": 6.930216007850598e-08,
      "loss": 0.38942737579345704,
      "mean_token_accuracy": 0.887349683046341,
      "num_tokens": 89030577.0,
      "step": 6780
    },
    {
      "entropy": 3.1953125,
      "epoch": 1.9073033707865168,
      "grad_norm": 1.2812118530273438,
      "learning_rate": 6.529427815351374e-08,
      "loss": 0.38655614852905273,
      "mean_token_accuracy": 0.8879540503025055,
      "num_tokens": 89162271.0,
      "step": 6790
    },
    {
      "entropy": 3.178125,
      "epoch": 1.9101123595505618,
      "grad_norm": 1.2434537410736084,
      "learning_rate": 6.140500454135668e-08,
      "loss": 0.36540679931640624,
      "mean_token_accuracy": 0.8927718698978424,
      "num_tokens": 89292461.0,
      "step": 6800
    },
    {
      "epoch": 1.9101123595505618,
      "eval_entropy": 3.174321619217082,
      "eval_loss": 0.4284653067588806,
      "eval_mean_token_accuracy": 0.8802866708765674,
      "eval_num_tokens": 89292461.0,
      "eval_runtime": 106.8898,
      "eval_samples_per_second": 147.18,
      "eval_steps_per_second": 2.629,
      "step": 6800
    },
    {
      "entropy": 3.1734375,
      "epoch": 1.9129213483146068,
      "grad_norm": 1.1921577453613281,
      "learning_rate": 5.7634432722822875e-08,
      "loss": 0.3780758142471313,
      "mean_token_accuracy": 0.8893661618232727,
      "num_tokens": 89423424.0,
      "step": 6810
    },
    {
      "entropy": 3.165625,
      "epoch": 1.9157303370786516,
      "grad_norm": 1.2014129161834717,
      "learning_rate": 5.398265332563935e-08,
      "loss": 0.38281033039093015,
      "mean_token_accuracy": 0.8879689037799835,
      "num_tokens": 89554803.0,
      "step": 6820
    },
    {
      "entropy": 3.1625,
      "epoch": 1.9185393258426966,
      "grad_norm": 1.160826325416565,
      "learning_rate": 5.0449754122292585e-08,
      "loss": 0.37180819511413576,
      "mean_token_accuracy": 0.8934452354907989,
      "num_tokens": 89686469.0,
      "step": 6830
    },
    {
      "entropy": 3.1984375,
      "epoch": 1.9213483146067416,
      "grad_norm": 1.4151344299316406,
      "learning_rate": 4.7035820027920284e-08,
      "loss": 0.3923435926437378,
      "mean_token_accuracy": 0.8867287814617157,
      "num_tokens": 89817422.0,
      "step": 6840
    },
    {
      "entropy": 3.140625,
      "epoch": 1.9241573033707864,
      "grad_norm": 1.1328307390213013,
      "learning_rate": 4.37409330982691e-08,
      "loss": 0.3903668880462646,
      "mean_token_accuracy": 0.8868801772594452,
      "num_tokens": 89948524.0,
      "step": 6850
    },
    {
      "entropy": 3.18125,
      "epoch": 1.9269662921348316,
      "grad_norm": 1.3234708309173584,
      "learning_rate": 4.056517252772229e-08,
      "loss": 0.3790267467498779,
      "mean_token_accuracy": 0.8878870725631713,
      "num_tokens": 90078608.0,
      "step": 6860
    },
    {
      "entropy": 3.2078125,
      "epoch": 1.9297752808988764,
      "grad_norm": 1.684568166732788,
      "learning_rate": 3.750861464739908e-08,
      "loss": 0.39711709022521974,
      "mean_token_accuracy": 0.8840954959392547,
      "num_tokens": 90210099.0,
      "step": 6870
    },
    {
      "entropy": 3.153125,
      "epoch": 1.9325842696629212,
      "grad_norm": 1.2920962572097778,
      "learning_rate": 3.457133292331494e-08,
      "loss": 0.3833528757095337,
      "mean_token_accuracy": 0.8881546795368195,
      "num_tokens": 90341139.0,
      "step": 6880
    },
    {
      "entropy": 3.128125,
      "epoch": 1.9353932584269664,
      "grad_norm": 1.2458369731903076,
      "learning_rate": 3.175339795462029e-08,
      "loss": 0.38412041664123536,
      "mean_token_accuracy": 0.8901838004589081,
      "num_tokens": 90472979.0,
      "step": 6890
    },
    {
      "entropy": 3.16875,
      "epoch": 1.9382022471910112,
      "grad_norm": 1.5036685466766357,
      "learning_rate": 2.9054877471901277e-08,
      "loss": 0.38884906768798827,
      "mean_token_accuracy": 0.8876751482486724,
      "num_tokens": 90604738.0,
      "step": 6900
    },
    {
      "epoch": 1.9382022471910112,
      "eval_entropy": 3.176434608540925,
      "eval_loss": 0.42840880155563354,
      "eval_mean_token_accuracy": 0.8803754457375333,
      "eval_num_tokens": 90604738.0,
      "eval_runtime": 117.5563,
      "eval_samples_per_second": 133.825,
      "eval_steps_per_second": 2.39,
      "step": 6900
    },
    {
      "entropy": 3.1015625,
      "epoch": 1.9410112359550562,
      "grad_norm": 1.0810261964797974,
      "learning_rate": 2.6475836335553838e-08,
      "loss": 0.3766488075256348,
      "mean_token_accuracy": 0.8902922570705414,
      "num_tokens": 90737836.0,
      "step": 6910
    },
    {
      "entropy": 3.178125,
      "epoch": 1.9438202247191012,
      "grad_norm": 1.3510791063308716,
      "learning_rate": 2.401633653422053e-08,
      "loss": 0.38427708148956297,
      "mean_token_accuracy": 0.8899927377700806,
      "num_tokens": 90870065.0,
      "step": 6920
    },
    {
      "entropy": 3.1734375,
      "epoch": 1.946629213483146,
      "grad_norm": 1.2950721979141235,
      "learning_rate": 2.1676437183306697e-08,
      "loss": 0.3944436311721802,
      "mean_token_accuracy": 0.8849980354309082,
      "num_tokens": 91000974.0,
      "step": 6930
    },
    {
      "entropy": 3.153125,
      "epoch": 1.949438202247191,
      "grad_norm": 1.3066872358322144,
      "learning_rate": 1.9456194523554404e-08,
      "loss": 0.37736735343933103,
      "mean_token_accuracy": 0.891019755601883,
      "num_tokens": 91131992.0,
      "step": 6940
    },
    {
      "entropy": 3.165625,
      "epoch": 1.952247191011236,
      "grad_norm": 1.1921286582946777,
      "learning_rate": 1.7355661919693513e-08,
      "loss": 0.3845776796340942,
      "mean_token_accuracy": 0.8892274260520935,
      "num_tokens": 91262992.0,
      "step": 6950
    },
    {
      "entropy": 3.1171875,
      "epoch": 1.9550561797752808,
      "grad_norm": 1.1625781059265137,
      "learning_rate": 1.5374889859157137e-08,
      "loss": 0.37804951667785647,
      "mean_token_accuracy": 0.8898682415485382,
      "num_tokens": 91394471.0,
      "step": 6960
    },
    {
      "entropy": 3.1703125,
      "epoch": 1.9578651685393258,
      "grad_norm": 1.4300299882888794,
      "learning_rate": 1.351392595087042e-08,
      "loss": 0.3884760856628418,
      "mean_token_accuracy": 0.8889488577842712,
      "num_tokens": 91525247.0,
      "step": 6970
    },
    {
      "entropy": 3.1828125,
      "epoch": 1.9606741573033708,
      "grad_norm": 1.408408284187317,
      "learning_rate": 1.1772814924103649e-08,
      "loss": 0.3714028596878052,
      "mean_token_accuracy": 0.8909241378307342,
      "num_tokens": 91656534.0,
      "step": 6980
    },
    {
      "entropy": 3.1453125,
      "epoch": 1.9634831460674156,
      "grad_norm": 1.277018666267395,
      "learning_rate": 1.0151598627399784e-08,
      "loss": 0.3755080223083496,
      "mean_token_accuracy": 0.8909967958927154,
      "num_tokens": 91788485.0,
      "step": 6990
    },
    {
      "entropy": 3.1890625,
      "epoch": 1.9662921348314608,
      "grad_norm": 1.4810922145843506,
      "learning_rate": 8.650316027566386e-09,
      "loss": 0.3897495985031128,
      "mean_token_accuracy": 0.8868665277957917,
      "num_tokens": 91919952.0,
      "step": 7000
    },
    {
      "epoch": 1.9662921348314608,
      "eval_entropy": 3.167204181494662,
      "eval_loss": 0.4284047782421112,
      "eval_mean_token_accuracy": 0.8803352319048817,
      "eval_num_tokens": 91919952.0,
      "eval_runtime": 106.7832,
      "eval_samples_per_second": 147.326,
      "eval_steps_per_second": 2.631,
      "step": 7000
    },
    {
      "entropy": 3.1765625,
      "epoch": 1.9691011235955056,
      "grad_norm": 1.258554458618164,
      "learning_rate": 7.269003208740244e-09,
      "loss": 0.37886719703674315,
      "mean_token_accuracy": 0.8881078600883484,
      "num_tokens": 92050564.0,
      "step": 7010
    },
    {
      "entropy": 3.190625,
      "epoch": 1.9719101123595506,
      "grad_norm": 1.180822491645813,
      "learning_rate": 6.00769337151974e-09,
      "loss": 0.3979960918426514,
      "mean_token_accuracy": 0.8843827843666077,
      "num_tokens": 92182293.0,
      "step": 7020
    },
    {
      "entropy": 3.1875,
      "epoch": 1.9747191011235956,
      "grad_norm": 1.1398983001708984,
      "learning_rate": 4.866416832167153e-09,
      "loss": 0.38153297901153566,
      "mean_token_accuracy": 0.888750946521759,
      "num_tokens": 92313936.0,
      "step": 7030
    },
    {
      "entropy": 3.1640625,
      "epoch": 1.9775280898876404,
      "grad_norm": 1.2697659730911255,
      "learning_rate": 3.845201021879241e-09,
      "loss": 0.3808913230895996,
      "mean_token_accuracy": 0.8899268209934235,
      "num_tokens": 92444872.0,
      "step": 7040
    },
    {
      "entropy": 3.1375,
      "epoch": 1.9803370786516854,
      "grad_norm": 1.2345160245895386,
      "learning_rate": 2.9440704861288804e-09,
      "loss": 0.3939173698425293,
      "mean_token_accuracy": 0.8858482539653778,
      "num_tokens": 92577112.0,
      "step": 7050
    },
    {
      "entropy": 3.1625,
      "epoch": 1.9831460674157304,
      "grad_norm": 1.3393018245697021,
      "learning_rate": 2.1630468840738716e-09,
      "loss": 0.39134142398834226,
      "mean_token_accuracy": 0.8865519225597381,
      "num_tokens": 92708107.0,
      "step": 7060
    },
    {
      "entropy": 3.1375,
      "epoch": 1.9859550561797752,
      "grad_norm": 1.4240082502365112,
      "learning_rate": 1.5021489880384653e-09,
      "loss": 0.38851854801177976,
      "mean_token_accuracy": 0.8881351888179779,
      "num_tokens": 92839743.0,
      "step": 7070
    },
    {
      "entropy": 3.1640625,
      "epoch": 1.9887640449438202,
      "grad_norm": 1.2211360931396484,
      "learning_rate": 9.613926830587262e-10,
      "loss": 0.3745984315872192,
      "mean_token_accuracy": 0.8907041132450104,
      "num_tokens": 92970973.0,
      "step": 7080
    },
    {
      "entropy": 3.1640625,
      "epoch": 1.9915730337078652,
      "grad_norm": 1.17845618724823,
      "learning_rate": 5.40790966505611e-10,
      "loss": 0.3786585330963135,
      "mean_token_accuracy": 0.8909171342849731,
      "num_tokens": 93102361.0,
      "step": 7090
    },
    {
      "entropy": 3.1640625,
      "epoch": 1.99438202247191,
      "grad_norm": 1.285343885421753,
      "learning_rate": 2.403539477668915e-10,
      "loss": 0.3837996244430542,
      "mean_token_accuracy": 0.8882774293422699,
      "num_tokens": 93233940.0,
      "step": 7100
    },
    {
      "epoch": 1.99438202247191,
      "eval_entropy": 3.1725422597864767,
      "eval_loss": 0.4284049868583679,
      "eval_mean_token_accuracy": 0.8802941091544263,
      "eval_num_tokens": 93233940.0,
      "eval_runtime": 115.0783,
      "eval_samples_per_second": 136.707,
      "eval_steps_per_second": 2.442,
      "step": 7100
    },
    {
      "entropy": 3.159375,
      "epoch": 1.9971910112359552,
      "grad_norm": 1.2171465158462524,
      "learning_rate": 6.008884800845494e-11,
      "loss": 0.39413464069366455,
      "mean_token_accuracy": 0.883305823802948,
      "num_tokens": 93365082.0,
      "step": 7110
    },
    {
      "entropy": 3.1984375,
      "epoch": 2.0,
      "grad_norm": 1.3144692182540894,
      "learning_rate": 0.0,
      "loss": 0.381104850769043,
      "mean_token_accuracy": 0.890763658285141,
      "num_tokens": 93496453.0,
      "step": 7120
    },
    {
      "epoch": 2.0,
      "eval_entropy": 3.17048487544484,
      "eval_loss": 0.42838677763938904,
      "eval_mean_token_accuracy": 0.8803565705798274,
      "eval_num_tokens": 93496453.0,
      "eval_runtime": 118.1625,
      "eval_samples_per_second": 133.139,
      "eval_steps_per_second": 2.378,
      "step": 7120
    },
    {
      "epoch": 2.0,
      "step": 7120,
      "total_flos": 5.321262511054389e+18,
      "train_loss": 0.4982514748077714,
      "train_runtime": 27435.3786,
      "train_samples_per_second": 14.532,
      "train_steps_per_second": 0.26
    }
  ],
  "logging_steps": 10,
  "max_steps": 7120,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.321262511054389e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}