diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,18178 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 2016,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001488095238095238,
+      "grad_norm": 5.513469831239093,
+      "learning_rate": 2.9761904761904764e-08,
+      "loss": 1.1667,
+      "mean_token_accuracy": 0.7050577402114868,
+      "num_tokens": 378406.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.002976190476190476,
+      "grad_norm": 5.801401919652211,
+      "learning_rate": 5.952380952380953e-08,
+      "loss": 1.2512,
+      "mean_token_accuracy": 0.6819975972175598,
+      "num_tokens": 744927.0,
+      "step": 2
+    },
+    {
+      "epoch": 0.004464285714285714,
+      "grad_norm": 5.785362058197214,
+      "learning_rate": 8.928571428571429e-08,
+      "loss": 1.2703,
+      "mean_token_accuracy": 0.6718945503234863,
+      "num_tokens": 1121938.0,
+      "step": 3
+    },
+    {
+      "epoch": 0.005952380952380952,
+      "grad_norm": 5.95135016721193,
+      "learning_rate": 1.1904761904761906e-07,
+      "loss": 1.3384,
+      "mean_token_accuracy": 0.6634253859519958,
+      "num_tokens": 1499811.0,
+      "step": 4
+    },
+    {
+      "epoch": 0.00744047619047619,
+      "grad_norm": 5.642363752310127,
+      "learning_rate": 1.4880952380952382e-07,
+      "loss": 1.2128,
+      "mean_token_accuracy": 0.6863797903060913,
+      "num_tokens": 1878602.0,
+      "step": 5
+    },
+    {
+      "epoch": 0.008928571428571428,
+      "grad_norm": 5.588971161882852,
+      "learning_rate": 1.7857142857142858e-07,
+      "loss": 1.3124,
+      "mean_token_accuracy": 0.6678875684738159,
+      "num_tokens": 2256084.0,
+      "step": 6
+    },
+    {
+      "epoch": 0.010416666666666666,
+      "grad_norm": 5.687559003211702,
+      "learning_rate": 2.0833333333333333e-07,
+      "loss": 1.1888,
+      "mean_token_accuracy": 0.6916842460632324,
+      "num_tokens": 2624010.0,
+      "step": 7
+    },
+    {
+      "epoch": 0.011904761904761904,
+      "grad_norm": 6.052867660375496,
+      "learning_rate": 2.3809523809523811e-07,
+      "loss": 1.2088,
+      "mean_token_accuracy": 0.6965529322624207,
+      "num_tokens": 2989700.0,
+      "step": 8
+    },
+    {
+      "epoch": 0.013392857142857142,
+      "grad_norm": 6.153533133958832,
+      "learning_rate": 2.6785714285714284e-07,
+      "loss": 1.277,
+      "mean_token_accuracy": 0.6747573018074036,
+      "num_tokens": 3367925.0,
+      "step": 9
+    },
+    {
+      "epoch": 0.01488095238095238,
+      "grad_norm": 5.768939031635095,
+      "learning_rate": 2.9761904761904765e-07,
+      "loss": 1.3304,
+      "mean_token_accuracy": 0.6665447354316711,
+      "num_tokens": 3744955.0,
+      "step": 10
+    },
+    {
+      "epoch": 0.01636904761904762,
+      "grad_norm": 6.001148436447376,
+      "learning_rate": 3.273809523809524e-07,
+      "loss": 1.2655,
+      "mean_token_accuracy": 0.6772600412368774,
+      "num_tokens": 4122491.0,
+      "step": 11
+    },
+    {
+      "epoch": 0.017857142857142856,
+      "grad_norm": 5.708865520758087,
+      "learning_rate": 3.5714285714285716e-07,
+      "loss": 1.1737,
+      "mean_token_accuracy": 0.6950653195381165,
+      "num_tokens": 4500541.0,
+      "step": 12
+    },
+    {
+      "epoch": 0.019345238095238096,
+      "grad_norm": 5.399130893633233,
+      "learning_rate": 3.8690476190476196e-07,
+      "loss": 1.1911,
+      "mean_token_accuracy": 0.699893593788147,
+      "num_tokens": 4879714.0,
+      "step": 13
+    },
+    {
+      "epoch": 0.020833333333333332,
+      "grad_norm": 5.573127013228007,
+      "learning_rate": 4.1666666666666667e-07,
+      "loss": 1.267,
+      "mean_token_accuracy": 0.6768276691436768,
+      "num_tokens": 5257852.0,
+      "step": 14
+    },
+    {
+      "epoch": 0.022321428571428572,
+      "grad_norm": 5.28694197799472,
+      "learning_rate": 4.4642857142857147e-07,
+      "loss": 1.2211,
+      "mean_token_accuracy": 0.6780167818069458,
+      "num_tokens": 5637131.0,
+      "step": 15
+    },
+    {
+      "epoch": 0.023809523809523808,
+      "grad_norm": 5.323315791223196,
+      "learning_rate": 4.7619047619047623e-07,
+      "loss": 1.1859,
+      "mean_token_accuracy": 0.6893203854560852,
+      "num_tokens": 6015677.0,
+      "step": 16
+    },
+    {
+      "epoch": 0.025297619047619048,
+      "grad_norm": 5.611624785364136,
+      "learning_rate": 5.05952380952381e-07,
+      "loss": 1.1831,
+      "mean_token_accuracy": 0.6948133707046509,
+      "num_tokens": 6393267.0,
+      "step": 17
+    },
+    {
+      "epoch": 0.026785714285714284,
+      "grad_norm": 5.308009444063126,
+      "learning_rate": 5.357142857142857e-07,
+      "loss": 1.2437,
+      "mean_token_accuracy": 0.6834532618522644,
+      "num_tokens": 6769591.0,
+      "step": 18
+    },
+    {
+      "epoch": 0.028273809523809524,
+      "grad_norm": 4.295266612442179,
+      "learning_rate": 5.654761904761905e-07,
+      "loss": 1.2202,
+      "mean_token_accuracy": 0.6841848492622375,
+      "num_tokens": 7134007.0,
+      "step": 19
+    },
+    {
+      "epoch": 0.02976190476190476,
+      "grad_norm": 4.7674694276905205,
+      "learning_rate": 5.952380952380953e-07,
+      "loss": 1.248,
+      "mean_token_accuracy": 0.678716778755188,
+      "num_tokens": 7510115.0,
+      "step": 20
+    },
+    {
+      "epoch": 0.03125,
+      "grad_norm": 4.7098658585536555,
+      "learning_rate": 6.25e-07,
+      "loss": 1.1941,
+      "mean_token_accuracy": 0.6996863484382629,
+      "num_tokens": 7865643.0,
+      "step": 21
+    },
+    {
+      "epoch": 0.03273809523809524,
+      "grad_norm": 4.530638374336081,
+      "learning_rate": 6.547619047619048e-07,
+      "loss": 1.2429,
+      "mean_token_accuracy": 0.6736594438552856,
+      "num_tokens": 8220496.0,
+      "step": 22
+    },
+    {
+      "epoch": 0.03422619047619048,
+      "grad_norm": 4.297673733130923,
+      "learning_rate": 6.845238095238097e-07,
+      "loss": 1.1482,
+      "mean_token_accuracy": 0.6925426125526428,
+      "num_tokens": 8574186.0,
+      "step": 23
+    },
+    {
+      "epoch": 0.03571428571428571,
+      "grad_norm": 4.311460485887874,
+      "learning_rate": 7.142857142857143e-07,
+      "loss": 1.1797,
+      "mean_token_accuracy": 0.6984924674034119,
+      "num_tokens": 8951999.0,
+      "step": 24
+    },
+    {
+      "epoch": 0.03720238095238095,
+      "grad_norm": 3.9158842290892726,
+      "learning_rate": 7.440476190476191e-07,
+      "loss": 1.0967,
+      "mean_token_accuracy": 0.7100571393966675,
+      "num_tokens": 9330751.0,
+      "step": 25
+    },
+    {
+      "epoch": 0.03869047619047619,
+      "grad_norm": 4.097965005125217,
+      "learning_rate": 7.738095238095239e-07,
+      "loss": 1.1203,
+      "mean_token_accuracy": 0.701015830039978,
+      "num_tokens": 9695871.0,
+      "step": 26
+    },
+    {
+      "epoch": 0.04017857142857143,
+      "grad_norm": 3.6693382751223917,
+      "learning_rate": 8.035714285714287e-07,
+      "loss": 1.121,
+      "mean_token_accuracy": 0.7068036794662476,
+      "num_tokens": 10074442.0,
+      "step": 27
+    },
+    {
+      "epoch": 0.041666666666666664,
+      "grad_norm": 3.727601697135322,
+      "learning_rate": 8.333333333333333e-07,
+      "loss": 1.0462,
+      "mean_token_accuracy": 0.7200284004211426,
+      "num_tokens": 10454001.0,
+      "step": 28
+    },
+    {
+      "epoch": 0.043154761904761904,
+      "grad_norm": 3.57463965015713,
+      "learning_rate": 8.630952380952382e-07,
+      "loss": 1.0695,
+      "mean_token_accuracy": 0.71123868227005,
+      "num_tokens": 10834168.0,
+      "step": 29
+    },
+    {
+      "epoch": 0.044642857142857144,
+      "grad_norm": 4.263998801418569,
+      "learning_rate": 8.928571428571429e-07,
+      "loss": 1.1605,
+      "mean_token_accuracy": 0.6914666891098022,
+      "num_tokens": 11209621.0,
+      "step": 30
+    },
+    {
+      "epoch": 0.046130952380952384,
+      "grad_norm": 3.8659148486258346,
+      "learning_rate": 9.226190476190477e-07,
+      "loss": 1.142,
+      "mean_token_accuracy": 0.6893674731254578,
+      "num_tokens": 11588436.0,
+      "step": 31
+    },
+    {
+      "epoch": 0.047619047619047616,
+      "grad_norm": 3.710615271516903,
+      "learning_rate": 9.523809523809525e-07,
+      "loss": 1.0502,
+      "mean_token_accuracy": 0.7189852595329285,
+      "num_tokens": 11967647.0,
+      "step": 32
+    },
+    {
+      "epoch": 0.049107142857142856,
+      "grad_norm": 3.8167530420309648,
+      "learning_rate": 9.821428571428572e-07,
+      "loss": 1.1356,
+      "mean_token_accuracy": 0.6983118653297424,
+      "num_tokens": 12345099.0,
+      "step": 33
+    },
+    {
+      "epoch": 0.050595238095238096,
+      "grad_norm": 3.2767633677733667,
+      "learning_rate": 1.011904761904762e-06,
+      "loss": 1.0735,
+      "mean_token_accuracy": 0.709028959274292,
+      "num_tokens": 12723712.0,
+      "step": 34
+    },
+    {
+      "epoch": 0.052083333333333336,
+      "grad_norm": 3.2288238847767508,
+      "learning_rate": 1.0416666666666667e-06,
+      "loss": 1.1303,
+      "mean_token_accuracy": 0.6931818127632141,
+      "num_tokens": 13092650.0,
+      "step": 35
+    },
+    {
+      "epoch": 0.05357142857142857,
+      "grad_norm": 3.2661815344928358,
+      "learning_rate": 1.0714285714285714e-06,
+      "loss": 1.0467,
+      "mean_token_accuracy": 0.7168600559234619,
+      "num_tokens": 13444292.0,
+      "step": 36
+    },
+    {
+      "epoch": 0.05505952380952381,
+      "grad_norm": 3.208322935976092,
+      "learning_rate": 1.1011904761904762e-06,
+      "loss": 1.0277,
+      "mean_token_accuracy": 0.7185325622558594,
+      "num_tokens": 13821920.0,
+      "step": 37
+    },
+    {
+      "epoch": 0.05654761904761905,
+      "grad_norm": 3.215252834203039,
+      "learning_rate": 1.130952380952381e-06,
+      "loss": 1.0126,
+      "mean_token_accuracy": 0.7265944480895996,
+      "num_tokens": 14200347.0,
+      "step": 38
+    },
+    {
+      "epoch": 0.05803571428571429,
+      "grad_norm": 3.3911428142589446,
+      "learning_rate": 1.160714285714286e-06,
+      "loss": 1.062,
+      "mean_token_accuracy": 0.7092524170875549,
+      "num_tokens": 14576967.0,
+      "step": 39
+    },
+    {
+      "epoch": 0.05952380952380952,
+      "grad_norm": 3.023518913745028,
+      "learning_rate": 1.1904761904761906e-06,
+      "loss": 1.0194,
+      "mean_token_accuracy": 0.7236782312393188,
+      "num_tokens": 14955342.0,
+      "step": 40
+    },
+    {
+      "epoch": 0.06101190476190476,
+      "grad_norm": 3.092360723512762,
+      "learning_rate": 1.2202380952380952e-06,
+      "loss": 1.0176,
+      "mean_token_accuracy": 0.719568133354187,
+      "num_tokens": 15325465.0,
+      "step": 41
+    },
+    {
+      "epoch": 0.0625,
+      "grad_norm": 2.8818416712816255,
+      "learning_rate": 1.25e-06,
+      "loss": 1.0146,
+      "mean_token_accuracy": 0.717740535736084,
+      "num_tokens": 15703946.0,
+      "step": 42
+    },
+    {
+      "epoch": 0.06398809523809523,
+      "grad_norm": 3.161781699608857,
+      "learning_rate": 1.2797619047619048e-06,
+      "loss": 0.9927,
+      "mean_token_accuracy": 0.7219191193580627,
+      "num_tokens": 16071292.0,
+      "step": 43
+    },
+    {
+      "epoch": 0.06547619047619048,
+      "grad_norm": 2.541414525224109,
+      "learning_rate": 1.3095238095238096e-06,
+      "loss": 0.9458,
+      "mean_token_accuracy": 0.7387768030166626,
+      "num_tokens": 16451202.0,
+      "step": 44
+    },
+    {
+      "epoch": 0.06696428571428571,
+      "grad_norm": 2.789848289378049,
+      "learning_rate": 1.3392857142857143e-06,
+      "loss": 1.0195,
+      "mean_token_accuracy": 0.7224578261375427,
+      "num_tokens": 16830552.0,
+      "step": 45
+    },
+    {
+      "epoch": 0.06845238095238096,
+      "grad_norm": 2.8181413684377095,
+      "learning_rate": 1.3690476190476193e-06,
+      "loss": 0.9588,
+      "mean_token_accuracy": 0.7289387583732605,
+      "num_tokens": 17208562.0,
+      "step": 46
+    },
+    {
+      "epoch": 0.06994047619047619,
+      "grad_norm": 2.7236138722562697,
+      "learning_rate": 1.398809523809524e-06,
+      "loss": 0.9637,
+      "mean_token_accuracy": 0.7335543632507324,
+      "num_tokens": 17586714.0,
+      "step": 47
+    },
+    {
+      "epoch": 0.07142857142857142,
+      "grad_norm": 2.6479908605041333,
+      "learning_rate": 1.4285714285714286e-06,
+      "loss": 0.9176,
+      "mean_token_accuracy": 0.7388561367988586,
+      "num_tokens": 17966440.0,
+      "step": 48
+    },
+    {
+      "epoch": 0.07291666666666667,
+      "grad_norm": 3.097273745834598,
+      "learning_rate": 1.4583333333333335e-06,
+      "loss": 1.0457,
+      "mean_token_accuracy": 0.7116292715072632,
+      "num_tokens": 18342379.0,
+      "step": 49
+    },
+    {
+      "epoch": 0.0744047619047619,
+      "grad_norm": 2.822618455249215,
+      "learning_rate": 1.4880952380952381e-06,
+      "loss": 0.9624,
+      "mean_token_accuracy": 0.7310827374458313,
+      "num_tokens": 18721884.0,
+      "step": 50
+    },
+    {
+      "epoch": 0.07589285714285714,
+      "grad_norm": 2.660395677208289,
+      "learning_rate": 1.5178571428571428e-06,
+      "loss": 0.9505,
+      "mean_token_accuracy": 0.7380973696708679,
+      "num_tokens": 19099277.0,
+      "step": 51
+    },
+    {
+      "epoch": 0.07738095238095238,
+      "grad_norm": 2.654640495734123,
+      "learning_rate": 1.5476190476190479e-06,
+      "loss": 0.9745,
+      "mean_token_accuracy": 0.7295454740524292,
+      "num_tokens": 19474612.0,
+      "step": 52
+    },
+    {
+      "epoch": 0.07886904761904762,
+      "grad_norm": 2.803754724069493,
+      "learning_rate": 1.5773809523809525e-06,
+      "loss": 0.9498,
+      "mean_token_accuracy": 0.7317637801170349,
+      "num_tokens": 19848889.0,
+      "step": 53
+    },
+    {
+      "epoch": 0.08035714285714286,
+      "grad_norm": 2.638110422571911,
+      "learning_rate": 1.6071428571428574e-06,
+      "loss": 1.0061,
+      "mean_token_accuracy": 0.7163302302360535,
+      "num_tokens": 20228113.0,
+      "step": 54
+    },
+    {
+      "epoch": 0.0818452380952381,
+      "grad_norm": 2.5786439198280324,
+      "learning_rate": 1.636904761904762e-06,
+      "loss": 0.9419,
+      "mean_token_accuracy": 0.7281362414360046,
+      "num_tokens": 20606671.0,
+      "step": 55
+    },
+    {
+      "epoch": 0.08333333333333333,
+      "grad_norm": 2.886808567664751,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 1.0306,
+      "mean_token_accuracy": 0.7134860754013062,
+      "num_tokens": 20984460.0,
+      "step": 56
+    },
+    {
+      "epoch": 0.08482142857142858,
+      "grad_norm": 2.52926097556405,
+      "learning_rate": 1.6964285714285717e-06,
+      "loss": 0.9478,
+      "mean_token_accuracy": 0.7309808731079102,
+      "num_tokens": 21362840.0,
+      "step": 57
+    },
+    {
+      "epoch": 0.08630952380952381,
+      "grad_norm": 2.5948763861743953,
+      "learning_rate": 1.7261904761904764e-06,
+      "loss": 0.9249,
+      "mean_token_accuracy": 0.7368120551109314,
+      "num_tokens": 21742442.0,
+      "step": 58
+    },
+    {
+      "epoch": 0.08779761904761904,
+      "grad_norm": 2.769115350437219,
+      "learning_rate": 1.755952380952381e-06,
+      "loss": 0.9562,
+      "mean_token_accuracy": 0.7301288843154907,
+      "num_tokens": 22120375.0,
+      "step": 59
+    },
+    {
+      "epoch": 0.08928571428571429,
+      "grad_norm": 2.5781896487000817,
+      "learning_rate": 1.7857142857142859e-06,
+      "loss": 0.8992,
+      "mean_token_accuracy": 0.7435463070869446,
+      "num_tokens": 22485025.0,
+      "step": 60
+    },
+    {
+      "epoch": 0.09077380952380952,
+      "grad_norm": 2.6901603657800015,
+      "learning_rate": 1.8154761904761905e-06,
+      "loss": 0.907,
+      "mean_token_accuracy": 0.7422633767127991,
+      "num_tokens": 22863387.0,
+      "step": 61
+    },
+    {
+      "epoch": 0.09226190476190477,
+      "grad_norm": 2.7092101776567428,
+      "learning_rate": 1.8452380952380954e-06,
+      "loss": 0.9352,
+      "mean_token_accuracy": 0.7307118773460388,
+      "num_tokens": 23241007.0,
+      "step": 62
+    },
+    {
+      "epoch": 0.09375,
+      "grad_norm": 2.4947204887618595,
+      "learning_rate": 1.8750000000000003e-06,
+      "loss": 0.9111,
+      "mean_token_accuracy": 0.7398694753646851,
+      "num_tokens": 23620398.0,
+      "step": 63
+    },
+    {
+      "epoch": 0.09523809523809523,
+      "grad_norm": 2.841944490998094,
+      "learning_rate": 1.904761904761905e-06,
+      "loss": 0.9535,
+      "mean_token_accuracy": 0.7319353818893433,
+      "num_tokens": 23997416.0,
+      "step": 64
+    },
+    {
+      "epoch": 0.09672619047619048,
+      "grad_norm": 2.5289595968929954,
+      "learning_rate": 1.9345238095238096e-06,
+      "loss": 0.8884,
+      "mean_token_accuracy": 0.739799976348877,
+      "num_tokens": 24374368.0,
+      "step": 65
+    },
+    {
+      "epoch": 0.09821428571428571,
+      "grad_norm": 2.524052553244891,
+      "learning_rate": 1.9642857142857144e-06,
+      "loss": 0.8835,
+      "mean_token_accuracy": 0.7446730136871338,
+      "num_tokens": 24751883.0,
+      "step": 66
+    },
+    {
+      "epoch": 0.09970238095238096,
+      "grad_norm": 2.64736821850323,
+      "learning_rate": 1.9940476190476193e-06,
+      "loss": 0.9539,
+      "mean_token_accuracy": 0.733038067817688,
+      "num_tokens": 25123942.0,
+      "step": 67
+    },
+    {
+      "epoch": 0.10119047619047619,
+      "grad_norm": 2.8501976688620227,
+      "learning_rate": 2.023809523809524e-06,
+      "loss": 0.9146,
+      "mean_token_accuracy": 0.7414661645889282,
+      "num_tokens": 25501175.0,
+      "step": 68
+    },
+    {
+      "epoch": 0.10267857142857142,
+      "grad_norm": 2.627340397259595,
+      "learning_rate": 2.0535714285714286e-06,
+      "loss": 0.9162,
+      "mean_token_accuracy": 0.7366005182266235,
+      "num_tokens": 25881312.0,
+      "step": 69
+    },
+    {
+      "epoch": 0.10416666666666667,
+      "grad_norm": 2.5678796826383103,
+      "learning_rate": 2.0833333333333334e-06,
+      "loss": 0.9267,
+      "mean_token_accuracy": 0.7437546849250793,
+      "num_tokens": 26259986.0,
+      "step": 70
+    },
+    {
+      "epoch": 0.1056547619047619,
+      "grad_norm": 2.5129927584374556,
+      "learning_rate": 2.1130952380952383e-06,
+      "loss": 0.8883,
+      "mean_token_accuracy": 0.7330623269081116,
+      "num_tokens": 26639522.0,
+      "step": 71
+    },
+    {
+      "epoch": 0.10714285714285714,
+      "grad_norm": 2.662687623268289,
+      "learning_rate": 2.1428571428571427e-06,
+      "loss": 0.9469,
+      "mean_token_accuracy": 0.7353377938270569,
+      "num_tokens": 27017634.0,
+      "step": 72
+    },
+    {
+      "epoch": 0.10863095238095238,
+      "grad_norm": 2.602451085581668,
+      "learning_rate": 2.172619047619048e-06,
+      "loss": 0.8966,
+      "mean_token_accuracy": 0.7531445026397705,
+      "num_tokens": 27394936.0,
+      "step": 73
+    },
+    {
+      "epoch": 0.11011904761904762,
+      "grad_norm": 2.639404839526633,
+      "learning_rate": 2.2023809523809525e-06,
+      "loss": 0.9789,
+      "mean_token_accuracy": 0.7341576218605042,
+      "num_tokens": 27774386.0,
+      "step": 74
+    },
+    {
+      "epoch": 0.11160714285714286,
+      "grad_norm": 2.451323857539176,
+      "learning_rate": 2.2321428571428573e-06,
+      "loss": 0.8931,
+      "mean_token_accuracy": 0.7387523055076599,
+      "num_tokens": 28153546.0,
+      "step": 75
+    },
+    {
+      "epoch": 0.1130952380952381,
+      "grad_norm": 2.4545665733925732,
+      "learning_rate": 2.261904761904762e-06,
+      "loss": 0.828,
+      "mean_token_accuracy": 0.7598411440849304,
+      "num_tokens": 28531809.0,
+      "step": 76
+    },
+    {
+      "epoch": 0.11458333333333333,
+      "grad_norm": 2.6871831465620137,
+      "learning_rate": 2.2916666666666666e-06,
+      "loss": 0.9492,
+      "mean_token_accuracy": 0.7273542881011963,
+      "num_tokens": 28909353.0,
+      "step": 77
+    },
+    {
+      "epoch": 0.11607142857142858,
+      "grad_norm": 2.7218664697046044,
+      "learning_rate": 2.321428571428572e-06,
+      "loss": 0.9084,
+      "mean_token_accuracy": 0.7414243221282959,
+      "num_tokens": 29286528.0,
+      "step": 78
+    },
+    {
+      "epoch": 0.11755952380952381,
+      "grad_norm": 2.5505795995091165,
+      "learning_rate": 2.3511904761904763e-06,
+      "loss": 0.8983,
+      "mean_token_accuracy": 0.7430711984634399,
+      "num_tokens": 29666050.0,
+      "step": 79
+    },
+    {
+      "epoch": 0.11904761904761904,
+      "grad_norm": 2.913122930230481,
+      "learning_rate": 2.380952380952381e-06,
+      "loss": 0.9547,
+      "mean_token_accuracy": 0.723631739616394,
+      "num_tokens": 30043844.0,
+      "step": 80
+    },
+    {
+      "epoch": 0.12053571428571429,
+      "grad_norm": 2.839200456375454,
+      "learning_rate": 2.410714285714286e-06,
+      "loss": 0.9606,
+      "mean_token_accuracy": 0.7303155660629272,
+      "num_tokens": 30413305.0,
+      "step": 81
+    },
+    {
+      "epoch": 0.12202380952380952,
+      "grad_norm": 2.4791152958555323,
+      "learning_rate": 2.4404761904761905e-06,
+      "loss": 0.82,
+      "mean_token_accuracy": 0.7602225542068481,
+      "num_tokens": 30791643.0,
+      "step": 82
+    },
+    {
+      "epoch": 0.12351190476190477,
+      "grad_norm": 2.471151847464406,
+      "learning_rate": 2.4702380952380953e-06,
+      "loss": 0.835,
+      "mean_token_accuracy": 0.7576201558113098,
+      "num_tokens": 31172813.0,
+      "step": 83
+    },
+    {
+      "epoch": 0.125,
+      "grad_norm": 2.624964117707357,
+      "learning_rate": 2.5e-06,
+      "loss": 0.8581,
+      "mean_token_accuracy": 0.7497380375862122,
+      "num_tokens": 31551237.0,
+      "step": 84
+    },
+    {
+      "epoch": 0.12648809523809523,
+      "grad_norm": 2.5845412376055403,
+      "learning_rate": 2.529761904761905e-06,
+      "loss": 0.8962,
+      "mean_token_accuracy": 0.7427788972854614,
+      "num_tokens": 31929607.0,
+      "step": 85
+    },
+    {
+      "epoch": 0.12797619047619047,
+      "grad_norm": 2.570749781142318,
+      "learning_rate": 2.5595238095238095e-06,
+      "loss": 0.8848,
+      "mean_token_accuracy": 0.7500653862953186,
+      "num_tokens": 32308127.0,
+      "step": 86
+    },
+    {
+      "epoch": 0.12946428571428573,
+      "grad_norm": 2.7905855155232384,
+      "learning_rate": 2.5892857142857148e-06,
+      "loss": 0.9217,
+      "mean_token_accuracy": 0.7346823215484619,
+      "num_tokens": 32685384.0,
+      "step": 87
+    },
+    {
+      "epoch": 0.13095238095238096,
+      "grad_norm": 2.4583671760303307,
+      "learning_rate": 2.6190476190476192e-06,
+      "loss": 0.874,
+      "mean_token_accuracy": 0.7487618327140808,
+      "num_tokens": 33064567.0,
+      "step": 88
+    },
+    {
+      "epoch": 0.1324404761904762,
+      "grad_norm": 2.5571438536112896,
+      "learning_rate": 2.648809523809524e-06,
+      "loss": 0.8689,
+      "mean_token_accuracy": 0.753685474395752,
+      "num_tokens": 33442807.0,
+      "step": 89
+    },
+    {
+      "epoch": 0.13392857142857142,
+      "grad_norm": 2.2379761043521063,
+      "learning_rate": 2.6785714285714285e-06,
+      "loss": 0.7951,
+      "mean_token_accuracy": 0.7653989791870117,
+      "num_tokens": 33822723.0,
+      "step": 90
+    },
+    {
+      "epoch": 0.13541666666666666,
+      "grad_norm": 2.891183499090045,
+      "learning_rate": 2.7083333333333334e-06,
+      "loss": 0.9772,
+      "mean_token_accuracy": 0.7336057424545288,
+      "num_tokens": 34199709.0,
+      "step": 91
+    },
+    {
+      "epoch": 0.13690476190476192,
+      "grad_norm": 2.5092804528184205,
+      "learning_rate": 2.7380952380952387e-06,
+      "loss": 0.8292,
+      "mean_token_accuracy": 0.7546103000640869,
+      "num_tokens": 34576933.0,
+      "step": 92
+    },
+    {
+      "epoch": 0.13839285714285715,
+      "grad_norm": 2.439502958141385,
+      "learning_rate": 2.767857142857143e-06,
+      "loss": 0.8445,
+      "mean_token_accuracy": 0.7546067237854004,
+      "num_tokens": 34954698.0,
+      "step": 93
+    },
+    {
+      "epoch": 0.13988095238095238,
+      "grad_norm": 2.3582400342387415,
+      "learning_rate": 2.797619047619048e-06,
+      "loss": 0.8797,
+      "mean_token_accuracy": 0.7417332530021667,
+      "num_tokens": 35333060.0,
+      "step": 94
+    },
+    {
+      "epoch": 0.14136904761904762,
+      "grad_norm": 2.5552651642665545,
+      "learning_rate": 2.8273809523809524e-06,
+      "loss": 0.9269,
+      "mean_token_accuracy": 0.7363049387931824,
+      "num_tokens": 35700235.0,
+      "step": 95
+    },
+    {
+      "epoch": 0.14285714285714285,
+      "grad_norm": 2.39302992670196,
+      "learning_rate": 2.8571428571428573e-06,
+      "loss": 0.8736,
+      "mean_token_accuracy": 0.750278890132904,
+      "num_tokens": 36066328.0,
+      "step": 96
+    },
+    {
+      "epoch": 0.14434523809523808,
+      "grad_norm": 2.3341510874172657,
+      "learning_rate": 2.8869047619047617e-06,
+      "loss": 0.8237,
+      "mean_token_accuracy": 0.7601529359817505,
+      "num_tokens": 36444522.0,
+      "step": 97
+    },
+    {
+      "epoch": 0.14583333333333334,
+      "grad_norm": 2.5142340750518315,
+      "learning_rate": 2.916666666666667e-06,
+      "loss": 0.8373,
+      "mean_token_accuracy": 0.7585219144821167,
+      "num_tokens": 36822292.0,
+      "step": 98
+    },
+    {
+      "epoch": 0.14732142857142858,
+      "grad_norm": 2.4075148113912617,
+      "learning_rate": 2.946428571428572e-06,
+      "loss": 0.8375,
+      "mean_token_accuracy": 0.7553053498268127,
+      "num_tokens": 37201008.0,
+      "step": 99
+    },
+    {
+      "epoch": 0.1488095238095238,
+      "grad_norm": 2.5962199119138094,
+      "learning_rate": 2.9761904761904763e-06,
+      "loss": 0.9198,
+      "mean_token_accuracy": 0.7385632991790771,
+      "num_tokens": 37578834.0,
+      "step": 100
+    },
+    {
+      "epoch": 0.15029761904761904,
+      "grad_norm": 2.6282414727722454,
+      "learning_rate": 3.005952380952381e-06,
+      "loss": 0.8461,
+      "mean_token_accuracy": 0.757478654384613,
+      "num_tokens": 37941159.0,
+      "step": 101
+    },
+    {
+      "epoch": 0.15178571428571427,
+      "grad_norm": 2.523666093656296,
+      "learning_rate": 3.0357142857142856e-06,
+      "loss": 0.8718,
+      "mean_token_accuracy": 0.7391644716262817,
+      "num_tokens": 38306297.0,
+      "step": 102
+    },
+    {
+      "epoch": 0.15327380952380953,
+      "grad_norm": 2.617584607331214,
+      "learning_rate": 3.065476190476191e-06,
+      "loss": 0.8175,
+      "mean_token_accuracy": 0.7598081231117249,
+      "num_tokens": 38683633.0,
+      "step": 103
+    },
+    {
+      "epoch": 0.15476190476190477,
+      "grad_norm": 2.5439578772339746,
+      "learning_rate": 3.0952380952380957e-06,
+      "loss": 0.9195,
+      "mean_token_accuracy": 0.7376594543457031,
+      "num_tokens": 39050672.0,
+      "step": 104
+    },
+    {
+      "epoch": 0.15625,
+      "grad_norm": 2.6546325489523452,
+      "learning_rate": 3.125e-06,
+      "loss": 0.8636,
+      "mean_token_accuracy": 0.7441918253898621,
+      "num_tokens": 39417290.0,
+      "step": 105
+    },
+    {
+      "epoch": 0.15773809523809523,
+      "grad_norm": 2.7433025166099445,
+      "learning_rate": 3.154761904761905e-06,
+      "loss": 0.8784,
+      "mean_token_accuracy": 0.7494934797286987,
+      "num_tokens": 39795225.0,
+      "step": 106
+    },
+    {
+      "epoch": 0.15922619047619047,
+      "grad_norm": 2.6352314112716426,
+      "learning_rate": 3.1845238095238094e-06,
+      "loss": 0.812,
+      "mean_token_accuracy": 0.7585965991020203,
+      "num_tokens": 40173964.0,
+      "step": 107
+    },
+    {
+      "epoch": 0.16071428571428573,
+      "grad_norm": 2.8771741417162606,
+      "learning_rate": 3.2142857142857147e-06,
+      "loss": 0.9361,
+      "mean_token_accuracy": 0.7409167289733887,
+      "num_tokens": 40550582.0,
+      "step": 108
+    },
+    {
+      "epoch": 0.16220238095238096,
+      "grad_norm": 2.5004366767398003,
+      "learning_rate": 3.2440476190476196e-06,
+      "loss": 0.8669,
+      "mean_token_accuracy": 0.7517943382263184,
+      "num_tokens": 40928254.0,
+      "step": 109
+    },
+    {
+      "epoch": 0.1636904761904762,
+      "grad_norm": 2.424402097239488,
+      "learning_rate": 3.273809523809524e-06,
+      "loss": 0.8485,
+      "mean_token_accuracy": 0.7469754219055176,
+      "num_tokens": 41308385.0,
+      "step": 110
+    },
+    {
+      "epoch": 0.16517857142857142,
+      "grad_norm": 2.4352826024387912,
+      "learning_rate": 3.303571428571429e-06,
+      "loss": 0.7611,
+      "mean_token_accuracy": 0.7783715128898621,
+      "num_tokens": 41665330.0,
+      "step": 111
+    },
+    {
+      "epoch": 0.16666666666666666,
+      "grad_norm": 2.42230598433068,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 0.7851,
+      "mean_token_accuracy": 0.768338143825531,
+      "num_tokens": 42044655.0,
+      "step": 112
+    },
+    {
+      "epoch": 0.16815476190476192,
+      "grad_norm": 2.9323127152926327,
+      "learning_rate": 3.3630952380952386e-06,
+      "loss": 0.8588,
+      "mean_token_accuracy": 0.7500655055046082,
+      "num_tokens": 42411273.0,
+      "step": 113
+    },
+    {
+      "epoch": 0.16964285714285715,
+      "grad_norm": 2.8210709588770575,
+      "learning_rate": 3.3928571428571435e-06,
+      "loss": 0.8387,
+      "mean_token_accuracy": 0.7521870732307434,
+      "num_tokens": 42789269.0,
+      "step": 114
+    },
+    {
+      "epoch": 0.17113095238095238,
+      "grad_norm": 2.5922089370796084,
+      "learning_rate": 3.422619047619048e-06,
+      "loss": 0.854,
+      "mean_token_accuracy": 0.7513460516929626,
+      "num_tokens": 43168599.0,
+      "step": 115
+    },
+    {
+      "epoch": 0.17261904761904762,
+      "grad_norm": 2.379546532150223,
+      "learning_rate": 3.4523809523809528e-06,
+      "loss": 0.8364,
+      "mean_token_accuracy": 0.7613523602485657,
+      "num_tokens": 43547958.0,
+      "step": 116
+    },
+    {
+      "epoch": 0.17410714285714285,
+      "grad_norm": 2.452103413273037,
+      "learning_rate": 3.482142857142857e-06,
+      "loss": 0.8279,
+      "mean_token_accuracy": 0.7496170997619629,
+      "num_tokens": 43923671.0,
+      "step": 117
+    },
+    {
+      "epoch": 0.17559523809523808,
+      "grad_norm": 2.8856427297700065,
+      "learning_rate": 3.511904761904762e-06,
+      "loss": 0.9326,
+      "mean_token_accuracy": 0.7441740036010742,
+      "num_tokens": 44301996.0,
+      "step": 118
+    },
+    {
+      "epoch": 0.17708333333333334,
+      "grad_norm": 2.5135058872407714,
+      "learning_rate": 3.5416666666666673e-06,
+      "loss": 0.8155,
+      "mean_token_accuracy": 0.7623541355133057,
+      "num_tokens": 44681277.0,
+      "step": 119
+    },
+    {
+      "epoch": 0.17857142857142858,
+      "grad_norm": 2.930451768695696,
+      "learning_rate": 3.5714285714285718e-06,
+      "loss": 0.8687,
+      "mean_token_accuracy": 0.7515465021133423,
+      "num_tokens": 45040302.0,
+      "step": 120
+    },
+    {
+      "epoch": 0.1800595238095238,
+      "grad_norm": 2.552460310857417,
+      "learning_rate": 3.6011904761904766e-06,
+      "loss": 0.8092,
+      "mean_token_accuracy": 0.7635560035705566,
+      "num_tokens": 45415975.0,
+      "step": 121
+    },
+    {
+      "epoch": 0.18154761904761904,
+      "grad_norm": 2.69790456631497,
+      "learning_rate": 3.630952380952381e-06,
+      "loss": 0.8263,
+      "mean_token_accuracy": 0.7550548911094666,
+      "num_tokens": 45792196.0,
+      "step": 122
+    },
+    {
+      "epoch": 0.18303571428571427,
+      "grad_norm": 2.8444248623524637,
+      "learning_rate": 3.660714285714286e-06,
+      "loss": 0.8691,
+      "mean_token_accuracy": 0.753216564655304,
+      "num_tokens": 46171503.0,
+      "step": 123
+    },
+    {
+      "epoch": 0.18452380952380953,
+      "grad_norm": 2.6914788197935375,
+      "learning_rate": 3.690476190476191e-06,
+      "loss": 0.8217,
+      "mean_token_accuracy": 0.7654385566711426,
+      "num_tokens": 46550056.0,
+      "step": 124
+    },
+    {
+      "epoch": 0.18601190476190477,
+      "grad_norm": 2.4656448679879293,
+      "learning_rate": 3.7202380952380957e-06,
+      "loss": 0.8007,
+      "mean_token_accuracy": 0.7636629939079285,
+      "num_tokens": 46929141.0,
+      "step": 125
+    },
+    {
+      "epoch": 0.1875,
+      "grad_norm": 2.543999601831726,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.8533,
+      "mean_token_accuracy": 0.7497294545173645,
+      "num_tokens": 47307516.0,
+      "step": 126
+    },
+    {
+      "epoch": 0.18898809523809523,
+      "grad_norm": 2.5068883154355093,
+      "learning_rate": 3.779761904761905e-06,
+      "loss": 0.8614,
+      "mean_token_accuracy": 0.7482050061225891,
+      "num_tokens": 47686667.0,
+      "step": 127
+    },
+    {
+      "epoch": 0.19047619047619047,
+      "grad_norm": 2.5868226612196965,
+      "learning_rate": 3.80952380952381e-06,
+      "loss": 0.8505,
+      "mean_token_accuracy": 0.7543191909790039,
+      "num_tokens": 48063581.0,
+      "step": 128
+    },
+    {
+      "epoch": 0.19196428571428573,
+      "grad_norm": 2.5484837127722537,
+      "learning_rate": 3.839285714285715e-06,
+      "loss": 0.7802,
+      "mean_token_accuracy": 0.7647309899330139,
+      "num_tokens": 48442051.0,
+      "step": 129
+    },
+    {
+      "epoch": 0.19345238095238096,
+      "grad_norm": 2.540341688067511,
+      "learning_rate": 3.869047619047619e-06,
+      "loss": 0.8576,
+      "mean_token_accuracy": 0.7504828572273254,
+      "num_tokens": 48820538.0,
+      "step": 130
+    },
+    {
+      "epoch": 0.1949404761904762,
+      "grad_norm": 2.579461871111694,
+      "learning_rate": 3.898809523809524e-06,
+      "loss": 0.8471,
+      "mean_token_accuracy": 0.7492630481719971,
+      "num_tokens": 49199348.0,
+      "step": 131
+    },
+    {
+      "epoch": 0.19642857142857142,
+      "grad_norm": 2.3893618577948605,
+      "learning_rate": 3.928571428571429e-06,
+      "loss": 0.8568,
+      "mean_token_accuracy": 0.7500627040863037,
+      "num_tokens": 49563092.0,
+      "step": 132
+    },
+    {
+      "epoch": 0.19791666666666666,
+      "grad_norm": 2.5757227769032918,
+      "learning_rate": 3.958333333333333e-06,
+      "loss": 0.9128,
+      "mean_token_accuracy": 0.7354434728622437,
+      "num_tokens": 49923330.0,
+      "step": 133
+    },
+    {
+      "epoch": 0.19940476190476192,
+      "grad_norm": 2.8257342527609404,
+      "learning_rate": 3.9880952380952386e-06,
+      "loss": 0.8936,
+      "mean_token_accuracy": 0.7421853542327881,
+      "num_tokens": 50288427.0,
+      "step": 134
+    },
+    {
+      "epoch": 0.20089285714285715,
+      "grad_norm": 2.496092042638026,
+      "learning_rate": 4.017857142857143e-06,
+      "loss": 0.723,
+      "mean_token_accuracy": 0.7821406126022339,
+      "num_tokens": 50666873.0,
+      "step": 135
+    },
+    {
+      "epoch": 0.20238095238095238,
+      "grad_norm": 2.6296777424499695,
+      "learning_rate": 4.047619047619048e-06,
+      "loss": 0.9154,
+      "mean_token_accuracy": 0.7520933747291565,
+      "num_tokens": 51033757.0,
+      "step": 136
+    },
+    {
+      "epoch": 0.20386904761904762,
+      "grad_norm": 2.309172823936338,
+      "learning_rate": 4.077380952380953e-06,
+      "loss": 0.7969,
+      "mean_token_accuracy": 0.7643204927444458,
+      "num_tokens": 51412918.0,
+      "step": 137
+    },
+    {
+      "epoch": 0.20535714285714285,
+      "grad_norm": 2.647871169645622,
+      "learning_rate": 4.107142857142857e-06,
+      "loss": 0.8422,
+      "mean_token_accuracy": 0.7529441118240356,
+      "num_tokens": 51792495.0,
+      "step": 138
+    },
+    {
+      "epoch": 0.20684523809523808,
+      "grad_norm": 2.5916529740748206,
+      "learning_rate": 4.136904761904762e-06,
+      "loss": 0.8236,
+      "mean_token_accuracy": 0.7626055479049683,
+      "num_tokens": 52171709.0,
+      "step": 139
+    },
+    {
+      "epoch": 0.20833333333333334,
+      "grad_norm": 2.4815357228570516,
+      "learning_rate": 4.166666666666667e-06,
+      "loss": 0.8499,
+      "mean_token_accuracy": 0.7603694796562195,
+      "num_tokens": 52548032.0,
+      "step": 140
+    },
+    {
+      "epoch": 0.20982142857142858,
+      "grad_norm": 2.894944587689418,
+      "learning_rate": 4.196428571428572e-06,
+      "loss": 0.8688,
+      "mean_token_accuracy": 0.7600411176681519,
+      "num_tokens": 52924787.0,
+      "step": 141
+    },
+    {
+      "epoch": 0.2113095238095238,
+      "grad_norm": 2.7605639124686956,
+      "learning_rate": 4.226190476190477e-06,
+      "loss": 0.8898,
+      "mean_token_accuracy": 0.7485986351966858,
+      "num_tokens": 53302113.0,
+      "step": 142
+    },
+    {
+      "epoch": 0.21279761904761904,
+      "grad_norm": 2.5080726791165566,
+      "learning_rate": 4.255952380952381e-06,
+      "loss": 0.8847,
+      "mean_token_accuracy": 0.7459037899971008,
+      "num_tokens": 53680048.0,
+      "step": 143
+    },
+    {
+      "epoch": 0.21428571428571427,
+      "grad_norm": 2.5096744041458408,
+      "learning_rate": 4.2857142857142855e-06,
+      "loss": 0.8304,
+      "mean_token_accuracy": 0.7560586929321289,
+      "num_tokens": 54059035.0,
+      "step": 144
+    },
+    {
+      "epoch": 0.21577380952380953,
+      "grad_norm": 2.5669486460139948,
+      "learning_rate": 4.315476190476191e-06,
+      "loss": 0.9268,
+      "mean_token_accuracy": 0.7420302033424377,
+      "num_tokens": 54423125.0,
+      "step": 145
+    },
+    {
+      "epoch": 0.21726190476190477,
+      "grad_norm": 2.462708570197547,
+      "learning_rate": 4.345238095238096e-06,
+      "loss": 0.7821,
+      "mean_token_accuracy": 0.7641040086746216,
+      "num_tokens": 54801027.0,
+      "step": 146
+    },
+    {
+      "epoch": 0.21875,
+      "grad_norm": 2.4749698408326672,
+      "learning_rate": 4.3750000000000005e-06,
+      "loss": 0.837,
+      "mean_token_accuracy": 0.7593703269958496,
+      "num_tokens": 55179709.0,
+      "step": 147
+    },
+    {
+      "epoch": 0.22023809523809523,
+      "grad_norm": 2.3764197361785544,
+      "learning_rate": 4.404761904761905e-06,
+      "loss": 0.8105,
+      "mean_token_accuracy": 0.7632536888122559,
+      "num_tokens": 55558424.0,
+      "step": 148
+    },
+    {
+      "epoch": 0.22172619047619047,
+      "grad_norm": 2.707910873701133,
+      "learning_rate": 4.434523809523809e-06,
+      "loss": 0.8681,
+      "mean_token_accuracy": 0.7455480694770813,
+      "num_tokens": 55935581.0,
+      "step": 149
+    },
+    {
+      "epoch": 0.22321428571428573,
+      "grad_norm": 2.6084895186256234,
+      "learning_rate": 4.464285714285715e-06,
+      "loss": 0.8484,
+      "mean_token_accuracy": 0.7552898526191711,
+      "num_tokens": 56314308.0,
+      "step": 150
+    },
+    {
+      "epoch": 0.22470238095238096,
+      "grad_norm": 2.7178679336574625,
+      "learning_rate": 4.49404761904762e-06,
+      "loss": 0.8475,
+      "mean_token_accuracy": 0.7507035136222839,
+      "num_tokens": 56690716.0,
+      "step": 151
+    },
+    {
+      "epoch": 0.2261904761904762,
+      "grad_norm": 2.587952068413652,
+      "learning_rate": 4.523809523809524e-06,
+      "loss": 0.8856,
+      "mean_token_accuracy": 0.7468295693397522,
+      "num_tokens": 57065370.0,
+      "step": 152
+    },
+    {
+      "epoch": 0.22767857142857142,
+      "grad_norm": 2.601620953610116,
+      "learning_rate": 4.553571428571429e-06,
+      "loss": 0.8259,
+      "mean_token_accuracy": 0.751610517501831,
+      "num_tokens": 57441749.0,
+      "step": 153
+    },
+    {
+      "epoch": 0.22916666666666666,
+      "grad_norm": 2.3654995379859716,
+      "learning_rate": 4.583333333333333e-06,
+      "loss": 0.8168,
+      "mean_token_accuracy": 0.7572703957557678,
+      "num_tokens": 57821186.0,
+      "step": 154
+    },
+    {
+      "epoch": 0.23065476190476192,
+      "grad_norm": 2.458105059491224,
+      "learning_rate": 4.6130952380952385e-06,
+      "loss": 0.7675,
+      "mean_token_accuracy": 0.768574059009552,
+      "num_tokens": 58200218.0,
+      "step": 155
+    },
+    {
+      "epoch": 0.23214285714285715,
+      "grad_norm": 2.589306507727725,
+      "learning_rate": 4.642857142857144e-06,
+      "loss": 0.8454,
+      "mean_token_accuracy": 0.756946861743927,
+      "num_tokens": 58577601.0,
+      "step": 156
+    },
+    {
+      "epoch": 0.23363095238095238,
+      "grad_norm": 2.517074788438444,
+      "learning_rate": 4.672619047619048e-06,
+      "loss": 0.8469,
+      "mean_token_accuracy": 0.7490100860595703,
+      "num_tokens": 58956203.0,
+      "step": 157
+    },
+    {
+      "epoch": 0.23511904761904762,
+      "grad_norm": 2.5993580827273948,
+      "learning_rate": 4.702380952380953e-06,
+      "loss": 0.8862,
+      "mean_token_accuracy": 0.7457659840583801,
+      "num_tokens": 59333008.0,
+      "step": 158
+    },
+    {
+      "epoch": 0.23660714285714285,
+      "grad_norm": 2.3013151065024804,
+      "learning_rate": 4.732142857142857e-06,
+      "loss": 0.7934,
+      "mean_token_accuracy": 0.7663149833679199,
+      "num_tokens": 59712739.0,
+      "step": 159
+    },
+    {
+      "epoch": 0.23809523809523808,
+      "grad_norm": 2.5244919206877205,
+      "learning_rate": 4.761904761904762e-06,
+      "loss": 0.7799,
+      "mean_token_accuracy": 0.7717186212539673,
+      "num_tokens": 60090427.0,
+      "step": 160
+    },
+    {
+      "epoch": 0.23958333333333334,
+      "grad_norm": 2.63307390362606,
+      "learning_rate": 4.791666666666668e-06,
+      "loss": 0.8622,
+      "mean_token_accuracy": 0.7497954368591309,
+      "num_tokens": 60468559.0,
+      "step": 161
+    },
+    {
+      "epoch": 0.24107142857142858,
+      "grad_norm": 2.7619577864265668,
+      "learning_rate": 4.821428571428572e-06,
+      "loss": 0.8468,
+      "mean_token_accuracy": 0.7511264085769653,
+      "num_tokens": 60846853.0,
+      "step": 162
+    },
+    {
+      "epoch": 0.2425595238095238,
+      "grad_norm": 2.386996456881852,
+      "learning_rate": 4.8511904761904765e-06,
+      "loss": 0.8107,
+      "mean_token_accuracy": 0.7610884308815002,
+      "num_tokens": 61225947.0,
+      "step": 163
+    },
+    {
+      "epoch": 0.24404761904761904,
+      "grad_norm": 2.4501387911918306,
+      "learning_rate": 4.880952380952381e-06,
+      "loss": 0.8064,
+      "mean_token_accuracy": 0.7615894079208374,
+      "num_tokens": 61588543.0,
+      "step": 164
+    },
+    {
+      "epoch": 0.24553571428571427,
+      "grad_norm": 2.507027536228417,
+      "learning_rate": 4.910714285714286e-06,
+      "loss": 0.837,
+      "mean_token_accuracy": 0.761777937412262,
+      "num_tokens": 61965744.0,
+      "step": 165
+    },
+    {
+      "epoch": 0.24702380952380953,
+      "grad_norm": 2.5220263260443443,
+      "learning_rate": 4.940476190476191e-06,
+      "loss": 0.8848,
+      "mean_token_accuracy": 0.7406700253486633,
+      "num_tokens": 62344388.0,
+      "step": 166
+    },
+    {
+      "epoch": 0.24851190476190477,
+      "grad_norm": 2.45243601470016,
+      "learning_rate": 4.970238095238096e-06,
+      "loss": 0.8332,
+      "mean_token_accuracy": 0.7574658393859863,
+      "num_tokens": 62723174.0,
+      "step": 167
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 2.3889490876464285,
+      "learning_rate": 5e-06,
+      "loss": 0.8252,
+      "mean_token_accuracy": 0.7514283061027527,
+      "num_tokens": 63101238.0,
+      "step": 168
+    },
+    {
+      "epoch": 0.25148809523809523,
+      "grad_norm": 2.4091207524554843,
+      "learning_rate": 5.029761904761905e-06,
+      "loss": 0.7734,
+      "mean_token_accuracy": 0.769940197467804,
+      "num_tokens": 63480704.0,
+      "step": 169
+    },
+    {
+      "epoch": 0.25297619047619047,
+      "grad_norm": 2.5356896523500403,
+      "learning_rate": 5.05952380952381e-06,
+      "loss": 0.8823,
+      "mean_token_accuracy": 0.7395924925804138,
+      "num_tokens": 63856858.0,
+      "step": 170
+    },
+    {
+      "epoch": 0.2544642857142857,
+      "grad_norm": 2.401274170632147,
+      "learning_rate": 5.0892857142857146e-06,
+      "loss": 0.8515,
+      "mean_token_accuracy": 0.7510275840759277,
+      "num_tokens": 64236049.0,
+      "step": 171
+    },
+    {
+      "epoch": 0.25595238095238093,
+      "grad_norm": 2.247719392230819,
+      "learning_rate": 5.119047619047619e-06,
+      "loss": 0.7771,
+      "mean_token_accuracy": 0.7671481966972351,
+      "num_tokens": 64610528.0,
+      "step": 172
+    },
+    {
+      "epoch": 0.25744047619047616,
+      "grad_norm": 2.3728101821585916,
+      "learning_rate": 5.1488095238095234e-06,
+      "loss": 0.7977,
+      "mean_token_accuracy": 0.7650033831596375,
+      "num_tokens": 64990039.0,
+      "step": 173
+    },
+    {
+      "epoch": 0.25892857142857145,
+      "grad_norm": 2.4930631897765787,
+      "learning_rate": 5.1785714285714296e-06,
+      "loss": 0.8631,
+      "mean_token_accuracy": 0.7459099292755127,
+      "num_tokens": 65368866.0,
+      "step": 174
+    },
+    {
+      "epoch": 0.2604166666666667,
+      "grad_norm": 2.2840228803863587,
+      "learning_rate": 5.208333333333334e-06,
+      "loss": 0.8279,
+      "mean_token_accuracy": 0.7553481459617615,
+      "num_tokens": 65747941.0,
+      "step": 175
+    },
+    {
+      "epoch": 0.2619047619047619,
+      "grad_norm": 2.2718341435925398,
+      "learning_rate": 5.2380952380952384e-06,
+      "loss": 0.7694,
+      "mean_token_accuracy": 0.7744379639625549,
+      "num_tokens": 66127086.0,
+      "step": 176
+    },
+    {
+      "epoch": 0.26339285714285715,
+      "grad_norm": 2.3403627196071843,
+      "learning_rate": 5.267857142857144e-06,
+      "loss": 0.8458,
+      "mean_token_accuracy": 0.7499790191650391,
+      "num_tokens": 66506076.0,
+      "step": 177
+    },
+    {
+      "epoch": 0.2648809523809524,
+      "grad_norm": 2.4436913630274217,
+      "learning_rate": 5.297619047619048e-06,
+      "loss": 0.7973,
+      "mean_token_accuracy": 0.7622070908546448,
+      "num_tokens": 66875984.0,
+      "step": 178
+    },
+    {
+      "epoch": 0.2663690476190476,
+      "grad_norm": 2.715656301104326,
+      "learning_rate": 5.327380952380953e-06,
+      "loss": 0.8281,
+      "mean_token_accuracy": 0.7544952630996704,
+      "num_tokens": 67256017.0,
+      "step": 179
+    },
+    {
+      "epoch": 0.26785714285714285,
+      "grad_norm": 2.480811958560633,
+      "learning_rate": 5.357142857142857e-06,
+      "loss": 0.8721,
+      "mean_token_accuracy": 0.7439382672309875,
+      "num_tokens": 67633350.0,
+      "step": 180
+    },
+    {
+      "epoch": 0.2693452380952381,
+      "grad_norm": 2.6019758308611216,
+      "learning_rate": 5.386904761904762e-06,
+      "loss": 0.8506,
+      "mean_token_accuracy": 0.7464002966880798,
+      "num_tokens": 68010694.0,
+      "step": 181
+    },
+    {
+      "epoch": 0.2708333333333333,
+      "grad_norm": 2.250147559533335,
+      "learning_rate": 5.416666666666667e-06,
+      "loss": 0.8098,
+      "mean_token_accuracy": 0.7606633305549622,
+      "num_tokens": 68389284.0,
+      "step": 182
+    },
+    {
+      "epoch": 0.27232142857142855,
+      "grad_norm": 2.249952330316731,
+      "learning_rate": 5.446428571428571e-06,
+      "loss": 0.8279,
+      "mean_token_accuracy": 0.7603970170021057,
+      "num_tokens": 68769044.0,
+      "step": 183
+    },
+    {
+      "epoch": 0.27380952380952384,
+      "grad_norm": 2.7166773109028117,
+      "learning_rate": 5.476190476190477e-06,
+      "loss": 0.8722,
+      "mean_token_accuracy": 0.7538506388664246,
+      "num_tokens": 69145642.0,
+      "step": 184
+    },
+    {
+      "epoch": 0.27529761904761907,
+      "grad_norm": 2.644003818016007,
+      "learning_rate": 5.505952380952382e-06,
+      "loss": 0.8416,
+      "mean_token_accuracy": 0.766242265701294,
+      "num_tokens": 69513037.0,
+      "step": 185
+    },
+    {
+      "epoch": 0.2767857142857143,
+      "grad_norm": 2.2578534849297967,
+      "learning_rate": 5.535714285714286e-06,
+      "loss": 0.7894,
+      "mean_token_accuracy": 0.7643928527832031,
+      "num_tokens": 69893108.0,
+      "step": 186
+    },
+    {
+      "epoch": 0.27827380952380953,
+      "grad_norm": 2.3780537750557316,
+      "learning_rate": 5.5654761904761915e-06,
+      "loss": 0.768,
+      "mean_token_accuracy": 0.7724405527114868,
+      "num_tokens": 70271884.0,
+      "step": 187
+    },
+    {
+      "epoch": 0.27976190476190477,
+      "grad_norm": 2.5442741169436944,
+      "learning_rate": 5.595238095238096e-06,
+      "loss": 0.8148,
+      "mean_token_accuracy": 0.7549163699150085,
+      "num_tokens": 70648661.0,
+      "step": 188
+    },
+    {
+      "epoch": 0.28125,
+      "grad_norm": 2.2463982117413255,
+      "learning_rate": 5.625e-06,
+      "loss": 0.79,
+      "mean_token_accuracy": 0.7712557315826416,
+      "num_tokens": 71028002.0,
+      "step": 189
+    },
+    {
+      "epoch": 0.28273809523809523,
+      "grad_norm": 2.63453263050719,
+      "learning_rate": 5.654761904761905e-06,
+      "loss": 0.8428,
+      "mean_token_accuracy": 0.7525055408477783,
+      "num_tokens": 71405933.0,
+      "step": 190
+    },
+    {
+      "epoch": 0.28422619047619047,
+      "grad_norm": 2.4934209444353077,
+      "learning_rate": 5.68452380952381e-06,
+      "loss": 0.8492,
+      "mean_token_accuracy": 0.749284029006958,
+      "num_tokens": 71785312.0,
+      "step": 191
+    },
+    {
+      "epoch": 0.2857142857142857,
+      "grad_norm": 2.360965792478758,
+      "learning_rate": 5.7142857142857145e-06,
+      "loss": 0.8319,
+      "mean_token_accuracy": 0.7514162659645081,
+      "num_tokens": 72163439.0,
+      "step": 192
+    },
+    {
+      "epoch": 0.28720238095238093,
+      "grad_norm": 2.6737361728690336,
+      "learning_rate": 5.744047619047619e-06,
+      "loss": 0.8749,
+      "mean_token_accuracy": 0.7489073276519775,
+      "num_tokens": 72541270.0,
+      "step": 193
+    },
+    {
+      "epoch": 0.28869047619047616,
+      "grad_norm": 2.6261867293562275,
+      "learning_rate": 5.773809523809523e-06,
+      "loss": 0.7692,
+      "mean_token_accuracy": 0.7691885232925415,
+      "num_tokens": 72898199.0,
+      "step": 194
+    },
+    {
+      "epoch": 0.29017857142857145,
+      "grad_norm": 2.4486712261241292,
+      "learning_rate": 5.8035714285714295e-06,
+      "loss": 0.8655,
+      "mean_token_accuracy": 0.7440395355224609,
+      "num_tokens": 73276125.0,
+      "step": 195
+    },
+    {
+      "epoch": 0.2916666666666667,
+      "grad_norm": 2.7944215591214188,
+      "learning_rate": 5.833333333333334e-06,
+      "loss": 0.8433,
+      "mean_token_accuracy": 0.7611925601959229,
+      "num_tokens": 73652902.0,
+      "step": 196
+    },
+    {
+      "epoch": 0.2931547619047619,
+      "grad_norm": 2.788500347784768,
+      "learning_rate": 5.863095238095239e-06,
+      "loss": 0.8517,
+      "mean_token_accuracy": 0.7453721761703491,
+      "num_tokens": 74025227.0,
+      "step": 197
+    },
+    {
+      "epoch": 0.29464285714285715,
+      "grad_norm": 2.504213446936013,
+      "learning_rate": 5.892857142857144e-06,
+      "loss": 0.8511,
+      "mean_token_accuracy": 0.7494570016860962,
+      "num_tokens": 74404106.0,
+      "step": 198
+    },
+    {
+      "epoch": 0.2961309523809524,
+      "grad_norm": 2.630002334992354,
+      "learning_rate": 5.922619047619048e-06,
+      "loss": 0.8112,
+      "mean_token_accuracy": 0.7553975582122803,
+      "num_tokens": 74782607.0,
+      "step": 199
+    },
+    {
+      "epoch": 0.2976190476190476,
+      "grad_norm": 2.469048751454046,
+      "learning_rate": 5.9523809523809525e-06,
+      "loss": 0.8566,
+      "mean_token_accuracy": 0.7469901442527771,
+      "num_tokens": 75160305.0,
+      "step": 200
+    },
+    {
+      "epoch": 0.29910714285714285,
+      "grad_norm": 2.694635898243874,
+      "learning_rate": 5.982142857142858e-06,
+      "loss": 0.7565,
+      "mean_token_accuracy": 0.778057336807251,
+      "num_tokens": 75526395.0,
+      "step": 201
+    },
+    {
+      "epoch": 0.3005952380952381,
+      "grad_norm": 3.057804449140092,
+      "learning_rate": 6.011904761904762e-06,
+      "loss": 0.8239,
+      "mean_token_accuracy": 0.7573689222335815,
+      "num_tokens": 75896097.0,
+      "step": 202
+    },
+    {
+      "epoch": 0.3020833333333333,
+      "grad_norm": 2.495247296713956,
+      "learning_rate": 6.041666666666667e-06,
+      "loss": 0.8389,
+      "mean_token_accuracy": 0.753683865070343,
+      "num_tokens": 76274239.0,
+      "step": 203
+    },
+    {
+      "epoch": 0.30357142857142855,
+      "grad_norm": 2.638549747929536,
+      "learning_rate": 6.071428571428571e-06,
+      "loss": 0.809,
+      "mean_token_accuracy": 0.7603899240493774,
+      "num_tokens": 76640833.0,
+      "step": 204
+    },
+    {
+      "epoch": 0.30505952380952384,
+      "grad_norm": 2.421811733909648,
+      "learning_rate": 6.101190476190477e-06,
+      "loss": 0.8411,
+      "mean_token_accuracy": 0.7532062530517578,
+      "num_tokens": 77019138.0,
+      "step": 205
+    },
+    {
+      "epoch": 0.30654761904761907,
+      "grad_norm": 2.5577407333252524,
+      "learning_rate": 6.130952380952382e-06,
+      "loss": 0.7994,
+      "mean_token_accuracy": 0.752764880657196,
+      "num_tokens": 77397435.0,
+      "step": 206
+    },
+    {
+      "epoch": 0.3080357142857143,
+      "grad_norm": 2.4439282359356254,
+      "learning_rate": 6.160714285714286e-06,
+      "loss": 0.8489,
+      "mean_token_accuracy": 0.7530348300933838,
+      "num_tokens": 77777131.0,
+      "step": 207
+    },
+    {
+      "epoch": 0.30952380952380953,
+      "grad_norm": 2.5862347748096632,
+      "learning_rate": 6.1904761904761914e-06,
+      "loss": 0.8515,
+      "mean_token_accuracy": 0.7514484524726868,
+      "num_tokens": 78156371.0,
+      "step": 208
+    },
+    {
+      "epoch": 0.31101190476190477,
+      "grad_norm": 2.5222000636008373,
+      "learning_rate": 6.220238095238096e-06,
+      "loss": 0.8396,
+      "mean_token_accuracy": 0.74796062707901,
+      "num_tokens": 78534021.0,
+      "step": 209
+    },
+    {
+      "epoch": 0.3125,
+      "grad_norm": 2.3232024952508596,
+      "learning_rate": 6.25e-06,
+      "loss": 0.763,
+      "mean_token_accuracy": 0.7641589641571045,
+      "num_tokens": 78912998.0,
+      "step": 210
+    },
+    {
+      "epoch": 0.31398809523809523,
+      "grad_norm": 2.2764974138266614,
+      "learning_rate": 6.279761904761906e-06,
+      "loss": 0.7843,
+      "mean_token_accuracy": 0.7640148401260376,
+      "num_tokens": 79291706.0,
+      "step": 211
+    },
+    {
+      "epoch": 0.31547619047619047,
+      "grad_norm": 2.598628050142767,
+      "learning_rate": 6.30952380952381e-06,
+      "loss": 0.8786,
+      "mean_token_accuracy": 0.7431068420410156,
+      "num_tokens": 79665519.0,
+      "step": 212
+    },
+    {
+      "epoch": 0.3169642857142857,
+      "grad_norm": 2.6194548876358574,
+      "learning_rate": 6.3392857142857145e-06,
+      "loss": 0.8222,
+      "mean_token_accuracy": 0.7587214708328247,
+      "num_tokens": 80042777.0,
+      "step": 213
+    },
+    {
+      "epoch": 0.31845238095238093,
+      "grad_norm": 2.4845118868814047,
+      "learning_rate": 6.369047619047619e-06,
+      "loss": 0.8421,
+      "mean_token_accuracy": 0.748102068901062,
+      "num_tokens": 80419984.0,
+      "step": 214
+    },
+    {
+      "epoch": 0.31994047619047616,
+      "grad_norm": 2.502645831082269,
+      "learning_rate": 6.398809523809524e-06,
+      "loss": 0.8468,
+      "mean_token_accuracy": 0.7525883316993713,
+      "num_tokens": 80798329.0,
+      "step": 215
+    },
+    {
+      "epoch": 0.32142857142857145,
+      "grad_norm": 2.6480216109569557,
+      "learning_rate": 6.4285714285714295e-06,
+      "loss": 0.7665,
+      "mean_token_accuracy": 0.7753902673721313,
+      "num_tokens": 81175639.0,
+      "step": 216
+    },
+    {
+      "epoch": 0.3229166666666667,
+      "grad_norm": 2.2697655250254396,
+      "learning_rate": 6.458333333333334e-06,
+      "loss": 0.7948,
+      "mean_token_accuracy": 0.7668435573577881,
+      "num_tokens": 81556801.0,
+      "step": 217
+    },
+    {
+      "epoch": 0.3244047619047619,
+      "grad_norm": 2.4267453137456028,
+      "learning_rate": 6.488095238095239e-06,
+      "loss": 0.77,
+      "mean_token_accuracy": 0.7686731219291687,
+      "num_tokens": 81936900.0,
+      "step": 218
+    },
+    {
+      "epoch": 0.32589285714285715,
+      "grad_norm": 2.219647888508299,
+      "learning_rate": 6.517857142857144e-06,
+      "loss": 0.8201,
+      "mean_token_accuracy": 0.7530809640884399,
+      "num_tokens": 82317522.0,
+      "step": 219
+    },
+    {
+      "epoch": 0.3273809523809524,
+      "grad_norm": 2.508896818112407,
+      "learning_rate": 6.547619047619048e-06,
+      "loss": 0.844,
+      "mean_token_accuracy": 0.7458573579788208,
+      "num_tokens": 82682479.0,
+      "step": 220
+    },
+    {
+      "epoch": 0.3288690476190476,
+      "grad_norm": 2.4278615882117487,
+      "learning_rate": 6.5773809523809525e-06,
+      "loss": 0.8389,
+      "mean_token_accuracy": 0.7632617950439453,
+      "num_tokens": 83060938.0,
+      "step": 221
+    },
+    {
+      "epoch": 0.33035714285714285,
+      "grad_norm": 2.7613970299323514,
+      "learning_rate": 6.607142857142858e-06,
+      "loss": 0.882,
+      "mean_token_accuracy": 0.7445062398910522,
+      "num_tokens": 83425002.0,
+      "step": 222
+    },
+    {
+      "epoch": 0.3318452380952381,
+      "grad_norm": 2.3814414305206473,
+      "learning_rate": 6.636904761904762e-06,
+      "loss": 0.802,
+      "mean_token_accuracy": 0.7601584196090698,
+      "num_tokens": 83803342.0,
+      "step": 223
+    },
+    {
+      "epoch": 0.3333333333333333,
+      "grad_norm": 2.6406199281006533,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 0.8751,
+      "mean_token_accuracy": 0.7502114176750183,
+      "num_tokens": 84180557.0,
+      "step": 224
+    },
+    {
+      "epoch": 0.33482142857142855,
+      "grad_norm": 2.6703842553168124,
+      "learning_rate": 6.696428571428571e-06,
+      "loss": 0.8435,
+      "mean_token_accuracy": 0.7586020827293396,
+      "num_tokens": 84557957.0,
+      "step": 225
+    },
+    {
+      "epoch": 0.33630952380952384,
+      "grad_norm": 2.4922424498050675,
+      "learning_rate": 6.726190476190477e-06,
+      "loss": 0.8243,
+      "mean_token_accuracy": 0.7565697431564331,
+      "num_tokens": 84922875.0,
+      "step": 226
+    },
+    {
+      "epoch": 0.33779761904761907,
+      "grad_norm": 2.504394000228559,
+      "learning_rate": 6.755952380952382e-06,
+      "loss": 0.7932,
+      "mean_token_accuracy": 0.7642784714698792,
+      "num_tokens": 85300410.0,
+      "step": 227
+    },
+    {
+      "epoch": 0.3392857142857143,
+      "grad_norm": 2.439042859362159,
+      "learning_rate": 6.785714285714287e-06,
+      "loss": 0.8194,
+      "mean_token_accuracy": 0.7606461048126221,
+      "num_tokens": 85676789.0,
+      "step": 228
+    },
+    {
+      "epoch": 0.34077380952380953,
+      "grad_norm": 2.8979021452121905,
+      "learning_rate": 6.815476190476191e-06,
+      "loss": 0.8307,
+      "mean_token_accuracy": 0.7548035383224487,
+      "num_tokens": 86053888.0,
+      "step": 229
+    },
+    {
+      "epoch": 0.34226190476190477,
+      "grad_norm": 2.8596325037018095,
+      "learning_rate": 6.845238095238096e-06,
+      "loss": 0.8585,
+      "mean_token_accuracy": 0.7573807835578918,
+      "num_tokens": 86425967.0,
+      "step": 230
+    },
+    {
+      "epoch": 0.34375,
+      "grad_norm": 2.60732080505383,
+      "learning_rate": 6.875e-06,
+      "loss": 0.8398,
+      "mean_token_accuracy": 0.7562230229377747,
+      "num_tokens": 86803442.0,
+      "step": 231
+    },
+    {
+      "epoch": 0.34523809523809523,
+      "grad_norm": 2.4792448323953282,
+      "learning_rate": 6.9047619047619055e-06,
+      "loss": 0.791,
+      "mean_token_accuracy": 0.7641351819038391,
+      "num_tokens": 87182623.0,
+      "step": 232
+    },
+    {
+      "epoch": 0.34672619047619047,
+      "grad_norm": 2.384019387404919,
+      "learning_rate": 6.93452380952381e-06,
+      "loss": 0.8236,
+      "mean_token_accuracy": 0.7501693964004517,
+      "num_tokens": 87556580.0,
+      "step": 233
+    },
+    {
+      "epoch": 0.3482142857142857,
+      "grad_norm": 2.380271455199982,
+      "learning_rate": 6.964285714285714e-06,
+      "loss": 0.784,
+      "mean_token_accuracy": 0.7669869065284729,
+      "num_tokens": 87935128.0,
+      "step": 234
+    },
+    {
+      "epoch": 0.34970238095238093,
+      "grad_norm": 2.370970336590661,
+      "learning_rate": 6.994047619047619e-06,
+      "loss": 0.7802,
+      "mean_token_accuracy": 0.7660841345787048,
+      "num_tokens": 88311692.0,
+      "step": 235
+    },
+    {
+      "epoch": 0.35119047619047616,
+      "grad_norm": 2.4032578556540813,
+      "learning_rate": 7.023809523809524e-06,
+      "loss": 0.8291,
+      "mean_token_accuracy": 0.7511368989944458,
+      "num_tokens": 88689801.0,
+      "step": 236
+    },
+    {
+      "epoch": 0.35267857142857145,
+      "grad_norm": 2.795888500140925,
+      "learning_rate": 7.053571428571429e-06,
+      "loss": 0.8377,
+      "mean_token_accuracy": 0.7553403377532959,
+      "num_tokens": 89067740.0,
+      "step": 237
+    },
+    {
+      "epoch": 0.3541666666666667,
+      "grad_norm": 2.8335654806232573,
+      "learning_rate": 7.083333333333335e-06,
+      "loss": 0.8684,
+      "mean_token_accuracy": 0.751986026763916,
+      "num_tokens": 89444588.0,
+      "step": 238
+    },
+    {
+      "epoch": 0.3556547619047619,
+      "grad_norm": 2.285655167055083,
+      "learning_rate": 7.113095238095239e-06,
+      "loss": 0.834,
+      "mean_token_accuracy": 0.7545310258865356,
+      "num_tokens": 89823573.0,
+      "step": 239
+    },
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 2.5059437222090266,
+      "learning_rate": 7.1428571428571436e-06,
+      "loss": 0.8504,
+      "mean_token_accuracy": 0.7593204975128174,
+      "num_tokens": 90200836.0,
+      "step": 240
+    },
+    {
+      "epoch": 0.3586309523809524,
+      "grad_norm": 2.6838015536470263,
+      "learning_rate": 7.172619047619048e-06,
+      "loss": 0.8527,
+      "mean_token_accuracy": 0.7529047131538391,
+      "num_tokens": 90577790.0,
+      "step": 241
+    },
+    {
+      "epoch": 0.3601190476190476,
+      "grad_norm": 3.008999474644847,
+      "learning_rate": 7.202380952380953e-06,
+      "loss": 0.8543,
+      "mean_token_accuracy": 0.758245587348938,
+      "num_tokens": 90936379.0,
+      "step": 242
+    },
+    {
+      "epoch": 0.36160714285714285,
+      "grad_norm": 2.4389457477336696,
+      "learning_rate": 7.232142857142858e-06,
+      "loss": 0.8057,
+      "mean_token_accuracy": 0.7557859420776367,
+      "num_tokens": 91315994.0,
+      "step": 243
+    },
+    {
+      "epoch": 0.3630952380952381,
+      "grad_norm": 2.4227527183471467,
+      "learning_rate": 7.261904761904762e-06,
+      "loss": 0.8205,
+      "mean_token_accuracy": 0.7598336338996887,
+      "num_tokens": 91689890.0,
+      "step": 244
+    },
+    {
+      "epoch": 0.3645833333333333,
+      "grad_norm": 2.406389056151078,
+      "learning_rate": 7.291666666666667e-06,
+      "loss": 0.7982,
+      "mean_token_accuracy": 0.7642269730567932,
+      "num_tokens": 92055688.0,
+      "step": 245
+    },
+    {
+      "epoch": 0.36607142857142855,
+      "grad_norm": 2.740632564185024,
+      "learning_rate": 7.321428571428572e-06,
+      "loss": 0.8638,
+      "mean_token_accuracy": 0.7570995092391968,
+      "num_tokens": 92420964.0,
+      "step": 246
+    },
+    {
+      "epoch": 0.36755952380952384,
+      "grad_norm": 2.8185156723168996,
+      "learning_rate": 7.351190476190477e-06,
+      "loss": 0.8437,
+      "mean_token_accuracy": 0.7587153911590576,
+      "num_tokens": 92800114.0,
+      "step": 247
+    },
+    {
+      "epoch": 0.36904761904761907,
+      "grad_norm": 2.421924266565702,
+      "learning_rate": 7.380952380952382e-06,
+      "loss": 0.7819,
+      "mean_token_accuracy": 0.7699534893035889,
+      "num_tokens": 93164487.0,
+      "step": 248
+    },
+    {
+      "epoch": 0.3705357142857143,
+      "grad_norm": 2.511346856168042,
+      "learning_rate": 7.410714285714287e-06,
+      "loss": 0.7991,
+      "mean_token_accuracy": 0.7762194871902466,
+      "num_tokens": 93543392.0,
+      "step": 249
+    },
+    {
+      "epoch": 0.37202380952380953,
+      "grad_norm": 2.3855781030618024,
+      "learning_rate": 7.440476190476191e-06,
+      "loss": 0.7582,
+      "mean_token_accuracy": 0.7722495794296265,
+      "num_tokens": 93923405.0,
+      "step": 250
+    },
+    {
+      "epoch": 0.37351190476190477,
+      "grad_norm": 2.6559476552222687,
+      "learning_rate": 7.470238095238096e-06,
+      "loss": 0.8482,
+      "mean_token_accuracy": 0.7598265409469604,
+      "num_tokens": 94301661.0,
+      "step": 251
+    },
+    {
+      "epoch": 0.375,
+      "grad_norm": 2.417086406396309,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.7929,
+      "mean_token_accuracy": 0.7671769857406616,
+      "num_tokens": 94677601.0,
+      "step": 252
+    },
+    {
+      "epoch": 0.37648809523809523,
+      "grad_norm": 2.431904930131837,
+      "learning_rate": 7.5297619047619055e-06,
+      "loss": 0.7714,
+      "mean_token_accuracy": 0.7757912874221802,
+      "num_tokens": 95055768.0,
+      "step": 253
+    },
+    {
+      "epoch": 0.37797619047619047,
+      "grad_norm": 2.7856985954691362,
+      "learning_rate": 7.55952380952381e-06,
+      "loss": 0.8606,
+      "mean_token_accuracy": 0.7502493262290955,
+      "num_tokens": 95434338.0,
+      "step": 254
+    },
+    {
+      "epoch": 0.3794642857142857,
+      "grad_norm": 2.448108438645311,
+      "learning_rate": 7.589285714285714e-06,
+      "loss": 0.7667,
+      "mean_token_accuracy": 0.7730329036712646,
+      "num_tokens": 95800953.0,
+      "step": 255
+    },
+    {
+      "epoch": 0.38095238095238093,
+      "grad_norm": 2.3160152972826373,
+      "learning_rate": 7.61904761904762e-06,
+      "loss": 0.7967,
+      "mean_token_accuracy": 0.7635219097137451,
+      "num_tokens": 96179674.0,
+      "step": 256
+    },
+    {
+      "epoch": 0.38244047619047616,
+      "grad_norm": 2.5593748944165426,
+      "learning_rate": 7.648809523809523e-06,
+      "loss": 0.8556,
+      "mean_token_accuracy": 0.7541589736938477,
+      "num_tokens": 96538372.0,
+      "step": 257
+    },
+    {
+      "epoch": 0.38392857142857145,
+      "grad_norm": 2.3472280144362347,
+      "learning_rate": 7.67857142857143e-06,
+      "loss": 0.7513,
+      "mean_token_accuracy": 0.7745699286460876,
+      "num_tokens": 96917464.0,
+      "step": 258
+    },
+    {
+      "epoch": 0.3854166666666667,
+      "grad_norm": 2.3378455824992557,
+      "learning_rate": 7.708333333333334e-06,
+      "loss": 0.8076,
+      "mean_token_accuracy": 0.7535193562507629,
+      "num_tokens": 97296399.0,
+      "step": 259
+    },
+    {
+      "epoch": 0.3869047619047619,
+      "grad_norm": 3.0520887911989067,
+      "learning_rate": 7.738095238095238e-06,
+      "loss": 0.7992,
+      "mean_token_accuracy": 0.7679130434989929,
+      "num_tokens": 97674265.0,
+      "step": 260
+    },
+    {
+      "epoch": 0.38839285714285715,
+      "grad_norm": 2.44499258637831,
+      "learning_rate": 7.767857142857144e-06,
+      "loss": 0.8012,
+      "mean_token_accuracy": 0.7589735388755798,
+      "num_tokens": 98053531.0,
+      "step": 261
+    },
+    {
+      "epoch": 0.3898809523809524,
+      "grad_norm": 2.2310264979940357,
+      "learning_rate": 7.797619047619049e-06,
+      "loss": 0.7857,
+      "mean_token_accuracy": 0.7711564302444458,
+      "num_tokens": 98432641.0,
+      "step": 262
+    },
+    {
+      "epoch": 0.3913690476190476,
+      "grad_norm": 2.9539261154703076,
+      "learning_rate": 7.827380952380953e-06,
+      "loss": 0.8835,
+      "mean_token_accuracy": 0.7436355352401733,
+      "num_tokens": 98809999.0,
+      "step": 263
+    },
+    {
+      "epoch": 0.39285714285714285,
+      "grad_norm": 2.750502103191404,
+      "learning_rate": 7.857142857142858e-06,
+      "loss": 0.8517,
+      "mean_token_accuracy": 0.7549232244491577,
+      "num_tokens": 99183077.0,
+      "step": 264
+    },
+    {
+      "epoch": 0.3943452380952381,
+      "grad_norm": 2.5866873276630913,
+      "learning_rate": 7.886904761904762e-06,
+      "loss": 0.8131,
+      "mean_token_accuracy": 0.76121586561203,
+      "num_tokens": 99547726.0,
+      "step": 265
+    },
+    {
+      "epoch": 0.3958333333333333,
+      "grad_norm": 2.7409354868563134,
+      "learning_rate": 7.916666666666667e-06,
+      "loss": 0.8781,
+      "mean_token_accuracy": 0.7446550726890564,
+      "num_tokens": 99925782.0,
+      "step": 266
+    },
+    {
+      "epoch": 0.39732142857142855,
+      "grad_norm": 2.415633517692654,
+      "learning_rate": 7.946428571428571e-06,
+      "loss": 0.8391,
+      "mean_token_accuracy": 0.7539689540863037,
+      "num_tokens": 100304434.0,
+      "step": 267
+    },
+    {
+      "epoch": 0.39880952380952384,
+      "grad_norm": 2.251010110725441,
+      "learning_rate": 7.976190476190477e-06,
+      "loss": 0.7672,
+      "mean_token_accuracy": 0.7711908221244812,
+      "num_tokens": 100683320.0,
+      "step": 268
+    },
+    {
+      "epoch": 0.40029761904761907,
+      "grad_norm": 2.7135231067901087,
+      "learning_rate": 8.005952380952382e-06,
+      "loss": 0.8417,
+      "mean_token_accuracy": 0.7505620121955872,
+      "num_tokens": 101061361.0,
+      "step": 269
+    },
+    {
+      "epoch": 0.4017857142857143,
+      "grad_norm": 2.407427825032665,
+      "learning_rate": 8.035714285714286e-06,
+      "loss": 0.8165,
+      "mean_token_accuracy": 0.7541680932044983,
+      "num_tokens": 101440129.0,
+      "step": 270
+    },
+    {
+      "epoch": 0.40327380952380953,
+      "grad_norm": 2.218663960142416,
+      "learning_rate": 8.065476190476192e-06,
+      "loss": 0.8057,
+      "mean_token_accuracy": 0.7608489394187927,
+      "num_tokens": 101818508.0,
+      "step": 271
+    },
+    {
+      "epoch": 0.40476190476190477,
+      "grad_norm": 2.205674941906529,
+      "learning_rate": 8.095238095238097e-06,
+      "loss": 0.8296,
+      "mean_token_accuracy": 0.7496461868286133,
+      "num_tokens": 102197629.0,
+      "step": 272
+    },
+    {
+      "epoch": 0.40625,
+      "grad_norm": 2.5516149768406717,
+      "learning_rate": 8.125000000000001e-06,
+      "loss": 0.7419,
+      "mean_token_accuracy": 0.7768390774726868,
+      "num_tokens": 102576113.0,
+      "step": 273
+    },
+    {
+      "epoch": 0.40773809523809523,
+      "grad_norm": 2.7030255636927185,
+      "learning_rate": 8.154761904761905e-06,
+      "loss": 0.7889,
+      "mean_token_accuracy": 0.771920382976532,
+      "num_tokens": 102955164.0,
+      "step": 274
+    },
+    {
+      "epoch": 0.40922619047619047,
+      "grad_norm": 2.622117645033341,
+      "learning_rate": 8.18452380952381e-06,
+      "loss": 0.8183,
+      "mean_token_accuracy": 0.7670851945877075,
+      "num_tokens": 103332301.0,
+      "step": 275
+    },
+    {
+      "epoch": 0.4107142857142857,
+      "grad_norm": 2.5582208816205516,
+      "learning_rate": 8.214285714285714e-06,
+      "loss": 0.8399,
+      "mean_token_accuracy": 0.759028971195221,
+      "num_tokens": 103710931.0,
+      "step": 276
+    },
+    {
+      "epoch": 0.41220238095238093,
+      "grad_norm": 2.585602439845819,
+      "learning_rate": 8.244047619047619e-06,
+      "loss": 0.8142,
+      "mean_token_accuracy": 0.7512217164039612,
+      "num_tokens": 104088449.0,
+      "step": 277
+    },
+    {
+      "epoch": 0.41369047619047616,
+      "grad_norm": 2.5017797715485757,
+      "learning_rate": 8.273809523809523e-06,
+      "loss": 0.8578,
+      "mean_token_accuracy": 0.748291552066803,
+      "num_tokens": 104466985.0,
+      "step": 278
+    },
+    {
+      "epoch": 0.41517857142857145,
+      "grad_norm": 2.6345576660151875,
+      "learning_rate": 8.30357142857143e-06,
+      "loss": 0.8029,
+      "mean_token_accuracy": 0.7705637812614441,
+      "num_tokens": 104846006.0,
+      "step": 279
+    },
+    {
+      "epoch": 0.4166666666666667,
+      "grad_norm": 2.3749189235251755,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.7777,
+      "mean_token_accuracy": 0.7746248245239258,
+      "num_tokens": 105220174.0,
+      "step": 280
+    },
+    {
+      "epoch": 0.4181547619047619,
+      "grad_norm": 2.9515035344620157,
+      "learning_rate": 8.36309523809524e-06,
+      "loss": 0.8477,
+      "mean_token_accuracy": 0.7499563694000244,
+      "num_tokens": 105597476.0,
+      "step": 281
+    },
+    {
+      "epoch": 0.41964285714285715,
+      "grad_norm": 2.505322307865475,
+      "learning_rate": 8.392857142857144e-06,
+      "loss": 0.8665,
+      "mean_token_accuracy": 0.7396718859672546,
+      "num_tokens": 105975943.0,
+      "step": 282
+    },
+    {
+      "epoch": 0.4211309523809524,
+      "grad_norm": 2.4200561964280016,
+      "learning_rate": 8.422619047619049e-06,
+      "loss": 0.7668,
+      "mean_token_accuracy": 0.7729743123054504,
+      "num_tokens": 106342485.0,
+      "step": 283
+    },
+    {
+      "epoch": 0.4226190476190476,
+      "grad_norm": 2.580232614140847,
+      "learning_rate": 8.452380952380953e-06,
+      "loss": 0.8106,
+      "mean_token_accuracy": 0.7594707608222961,
+      "num_tokens": 106721523.0,
+      "step": 284
+    },
+    {
+      "epoch": 0.42410714285714285,
+      "grad_norm": 2.7188427162898905,
+      "learning_rate": 8.482142857142858e-06,
+      "loss": 0.8027,
+      "mean_token_accuracy": 0.7612557411193848,
+      "num_tokens": 107097989.0,
+      "step": 285
+    },
+    {
+      "epoch": 0.4255952380952381,
+      "grad_norm": 2.47645964651674,
+      "learning_rate": 8.511904761904762e-06,
+      "loss": 0.806,
+      "mean_token_accuracy": 0.7560685873031616,
+      "num_tokens": 107465007.0,
+      "step": 286
+    },
+    {
+      "epoch": 0.4270833333333333,
+      "grad_norm": 2.5348960738315602,
+      "learning_rate": 8.541666666666666e-06,
+      "loss": 0.7704,
+      "mean_token_accuracy": 0.7714031338691711,
+      "num_tokens": 107843563.0,
+      "step": 287
+    },
+    {
+      "epoch": 0.42857142857142855,
+      "grad_norm": 2.4550713880123936,
+      "learning_rate": 8.571428571428571e-06,
+      "loss": 0.7753,
+      "mean_token_accuracy": 0.7643919587135315,
+      "num_tokens": 108210371.0,
+      "step": 288
+    },
+    {
+      "epoch": 0.43005952380952384,
+      "grad_norm": 2.7338153425609617,
+      "learning_rate": 8.601190476190477e-06,
+      "loss": 0.7585,
+      "mean_token_accuracy": 0.7696078419685364,
+      "num_tokens": 108587848.0,
+      "step": 289
+    },
+    {
+      "epoch": 0.43154761904761907,
+      "grad_norm": 2.3147295455821673,
+      "learning_rate": 8.630952380952381e-06,
+      "loss": 0.7469,
+      "mean_token_accuracy": 0.7760506868362427,
+      "num_tokens": 108967368.0,
+      "step": 290
+    },
+    {
+      "epoch": 0.4330357142857143,
+      "grad_norm": 3.0516645962398496,
+      "learning_rate": 8.660714285714286e-06,
+      "loss": 0.8192,
+      "mean_token_accuracy": 0.7515782117843628,
+      "num_tokens": 109345648.0,
+      "step": 291
+    },
+    {
+      "epoch": 0.43452380952380953,
+      "grad_norm": 3.3447415752519216,
+      "learning_rate": 8.690476190476192e-06,
+      "loss": 0.7874,
+      "mean_token_accuracy": 0.7668451070785522,
+      "num_tokens": 109722867.0,
+      "step": 292
+    },
+    {
+      "epoch": 0.43601190476190477,
+      "grad_norm": 2.7914543160377137,
+      "learning_rate": 8.720238095238096e-06,
+      "loss": 0.8062,
+      "mean_token_accuracy": 0.7687567472457886,
+      "num_tokens": 110099843.0,
+      "step": 293
+    },
+    {
+      "epoch": 0.4375,
+      "grad_norm": 2.4795931369083117,
+      "learning_rate": 8.750000000000001e-06,
+      "loss": 0.806,
+      "mean_token_accuracy": 0.7543814778327942,
+      "num_tokens": 110478553.0,
+      "step": 294
+    },
+    {
+      "epoch": 0.43898809523809523,
+      "grad_norm": 2.728828905551046,
+      "learning_rate": 8.779761904761905e-06,
+      "loss": 0.7893,
+      "mean_token_accuracy": 0.7634230852127075,
+      "num_tokens": 110857203.0,
+      "step": 295
+    },
+    {
+      "epoch": 0.44047619047619047,
+      "grad_norm": 2.564594905793595,
+      "learning_rate": 8.80952380952381e-06,
+      "loss": 0.8124,
+      "mean_token_accuracy": 0.7622863054275513,
+      "num_tokens": 111236517.0,
+      "step": 296
+    },
+    {
+      "epoch": 0.4419642857142857,
+      "grad_norm": 2.4539300306331597,
+      "learning_rate": 8.839285714285714e-06,
+      "loss": 0.7894,
+      "mean_token_accuracy": 0.7641110420227051,
+      "num_tokens": 111615969.0,
+      "step": 297
+    },
+    {
+      "epoch": 0.44345238095238093,
+      "grad_norm": 2.5675362963984574,
+      "learning_rate": 8.869047619047619e-06,
+      "loss": 0.8468,
+      "mean_token_accuracy": 0.7471016645431519,
+      "num_tokens": 111987027.0,
+      "step": 298
+    },
+    {
+      "epoch": 0.44494047619047616,
+      "grad_norm": 2.5204192052838676,
+      "learning_rate": 8.898809523809525e-06,
+      "loss": 0.8568,
+      "mean_token_accuracy": 0.7498636841773987,
+      "num_tokens": 112363967.0,
+      "step": 299
+    },
+    {
+      "epoch": 0.44642857142857145,
+      "grad_norm": 2.6765137046290275,
+      "learning_rate": 8.92857142857143e-06,
+      "loss": 0.8246,
+      "mean_token_accuracy": 0.7621666193008423,
+      "num_tokens": 112740629.0,
+      "step": 300
+    },
+    {
+      "epoch": 0.4479166666666667,
+      "grad_norm": 2.378625251770564,
+      "learning_rate": 8.958333333333334e-06,
+      "loss": 0.7923,
+      "mean_token_accuracy": 0.7606196403503418,
+      "num_tokens": 113120361.0,
+      "step": 301
+    },
+    {
+      "epoch": 0.4494047619047619,
+      "grad_norm": 2.3950051376679173,
+      "learning_rate": 8.98809523809524e-06,
+      "loss": 0.7969,
+      "mean_token_accuracy": 0.7660772800445557,
+      "num_tokens": 113498845.0,
+      "step": 302
+    },
+    {
+      "epoch": 0.45089285714285715,
+      "grad_norm": 2.4374622882867865,
+      "learning_rate": 9.017857142857144e-06,
+      "loss": 0.7744,
+      "mean_token_accuracy": 0.7617419362068176,
+      "num_tokens": 113876623.0,
+      "step": 303
+    },
+    {
+      "epoch": 0.4523809523809524,
+      "grad_norm": 2.5937496779113416,
+      "learning_rate": 9.047619047619049e-06,
+      "loss": 0.8236,
+      "mean_token_accuracy": 0.7511516809463501,
+      "num_tokens": 114254030.0,
+      "step": 304
+    },
+    {
+      "epoch": 0.4538690476190476,
+      "grad_norm": 2.5822460244245873,
+      "learning_rate": 9.077380952380953e-06,
+      "loss": 0.8162,
+      "mean_token_accuracy": 0.7601204514503479,
+      "num_tokens": 114632038.0,
+      "step": 305
+    },
+    {
+      "epoch": 0.45535714285714285,
+      "grad_norm": 2.4281280599561224,
+      "learning_rate": 9.107142857142858e-06,
+      "loss": 0.7671,
+      "mean_token_accuracy": 0.7707247138023376,
+      "num_tokens": 115011306.0,
+      "step": 306
+    },
+    {
+      "epoch": 0.4568452380952381,
+      "grad_norm": 2.614903222814987,
+      "learning_rate": 9.136904761904762e-06,
+      "loss": 0.7972,
+      "mean_token_accuracy": 0.7623823881149292,
+      "num_tokens": 115387152.0,
+      "step": 307
+    },
+    {
+      "epoch": 0.4583333333333333,
+      "grad_norm": 2.420904650982505,
+      "learning_rate": 9.166666666666666e-06,
+      "loss": 0.8078,
+      "mean_token_accuracy": 0.758929431438446,
+      "num_tokens": 115760410.0,
+      "step": 308
+    },
+    {
+      "epoch": 0.45982142857142855,
+      "grad_norm": 2.4997673251313706,
+      "learning_rate": 9.196428571428571e-06,
+      "loss": 0.843,
+      "mean_token_accuracy": 0.7522587180137634,
+      "num_tokens": 116139506.0,
+      "step": 309
+    },
+    {
+      "epoch": 0.46130952380952384,
+      "grad_norm": 2.5580481025161474,
+      "learning_rate": 9.226190476190477e-06,
+      "loss": 0.7614,
+      "mean_token_accuracy": 0.772212564945221,
+      "num_tokens": 116518666.0,
+      "step": 310
+    },
+    {
+      "epoch": 0.46279761904761907,
+      "grad_norm": 2.704956247614411,
+      "learning_rate": 9.255952380952381e-06,
+      "loss": 0.8014,
+      "mean_token_accuracy": 0.7641128897666931,
+      "num_tokens": 116883185.0,
+      "step": 311
+    },
+    {
+      "epoch": 0.4642857142857143,
+      "grad_norm": 2.4011730340408475,
+      "learning_rate": 9.285714285714288e-06,
+      "loss": 0.8304,
+      "mean_token_accuracy": 0.7591840624809265,
+      "num_tokens": 117256332.0,
+      "step": 312
+    },
+    {
+      "epoch": 0.46577380952380953,
+      "grad_norm": 2.4316360080402526,
+      "learning_rate": 9.315476190476192e-06,
+      "loss": 0.8126,
+      "mean_token_accuracy": 0.7647058963775635,
+      "num_tokens": 117634780.0,
+      "step": 313
+    },
+    {
+      "epoch": 0.46726190476190477,
+      "grad_norm": 2.2994724464370155,
+      "learning_rate": 9.345238095238096e-06,
+      "loss": 0.785,
+      "mean_token_accuracy": 0.7661627531051636,
+      "num_tokens": 118013036.0,
+      "step": 314
+    },
+    {
+      "epoch": 0.46875,
+      "grad_norm": 2.160307128288804,
+      "learning_rate": 9.375000000000001e-06,
+      "loss": 0.8331,
+      "mean_token_accuracy": 0.7550109624862671,
+      "num_tokens": 118380810.0,
+      "step": 315
+    },
+    {
+      "epoch": 0.47023809523809523,
+      "grad_norm": 2.275191199019259,
+      "learning_rate": 9.404761904761905e-06,
+      "loss": 0.7852,
+      "mean_token_accuracy": 0.7672451734542847,
+      "num_tokens": 118759282.0,
+      "step": 316
+    },
+    {
+      "epoch": 0.47172619047619047,
+      "grad_norm": 2.4197082795342917,
+      "learning_rate": 9.43452380952381e-06,
+      "loss": 0.7606,
+      "mean_token_accuracy": 0.7666339874267578,
+      "num_tokens": 119137162.0,
+      "step": 317
+    },
+    {
+      "epoch": 0.4732142857142857,
+      "grad_norm": 2.8814998056855896,
+      "learning_rate": 9.464285714285714e-06,
+      "loss": 0.8268,
+      "mean_token_accuracy": 0.7589877843856812,
+      "num_tokens": 119516260.0,
+      "step": 318
+    },
+    {
+      "epoch": 0.47470238095238093,
+      "grad_norm": 2.4699732562885717,
+      "learning_rate": 9.494047619047619e-06,
+      "loss": 0.8593,
+      "mean_token_accuracy": 0.7460074424743652,
+      "num_tokens": 119892606.0,
+      "step": 319
+    },
+    {
+      "epoch": 0.47619047619047616,
+      "grad_norm": 2.263989511278806,
+      "learning_rate": 9.523809523809525e-06,
+      "loss": 0.8645,
+      "mean_token_accuracy": 0.7498471140861511,
+      "num_tokens": 120270897.0,
+      "step": 320
+    },
+    {
+      "epoch": 0.47767857142857145,
+      "grad_norm": 2.4837358234767395,
+      "learning_rate": 9.55357142857143e-06,
+      "loss": 0.7912,
+      "mean_token_accuracy": 0.7661221623420715,
+      "num_tokens": 120649887.0,
+      "step": 321
+    },
+    {
+      "epoch": 0.4791666666666667,
+      "grad_norm": 2.299736051077147,
+      "learning_rate": 9.583333333333335e-06,
+      "loss": 0.7864,
+      "mean_token_accuracy": 0.771015465259552,
+      "num_tokens": 121026948.0,
+      "step": 322
+    },
+    {
+      "epoch": 0.4806547619047619,
+      "grad_norm": 2.410959453237422,
+      "learning_rate": 9.61309523809524e-06,
+      "loss": 0.7905,
+      "mean_token_accuracy": 0.7657426595687866,
+      "num_tokens": 121405177.0,
+      "step": 323
+    },
+    {
+      "epoch": 0.48214285714285715,
+      "grad_norm": 2.405718207679544,
+      "learning_rate": 9.642857142857144e-06,
+      "loss": 0.8036,
+      "mean_token_accuracy": 0.7620789408683777,
+      "num_tokens": 121783271.0,
+      "step": 324
+    },
+    {
+      "epoch": 0.4836309523809524,
+      "grad_norm": 2.549719960390193,
+      "learning_rate": 9.672619047619049e-06,
+      "loss": 0.8268,
+      "mean_token_accuracy": 0.7544642686843872,
+      "num_tokens": 122160488.0,
+      "step": 325
+    },
+    {
+      "epoch": 0.4851190476190476,
+      "grad_norm": 2.213245915712409,
+      "learning_rate": 9.702380952380953e-06,
+      "loss": 0.7584,
+      "mean_token_accuracy": 0.7742464542388916,
+      "num_tokens": 122541266.0,
+      "step": 326
+    },
+    {
+      "epoch": 0.48660714285714285,
+      "grad_norm": 2.514196163897498,
+      "learning_rate": 9.732142857142858e-06,
+      "loss": 0.7997,
+      "mean_token_accuracy": 0.7584723234176636,
+      "num_tokens": 122919718.0,
+      "step": 327
+    },
+    {
+      "epoch": 0.4880952380952381,
+      "grad_norm": 2.8206928333700754,
+      "learning_rate": 9.761904761904762e-06,
+      "loss": 0.8188,
+      "mean_token_accuracy": 0.755731463432312,
+      "num_tokens": 123297661.0,
+      "step": 328
+    },
+    {
+      "epoch": 0.4895833333333333,
+      "grad_norm": 2.6007753558238473,
+      "learning_rate": 9.791666666666666e-06,
+      "loss": 0.7815,
+      "mean_token_accuracy": 0.7654410600662231,
+      "num_tokens": 123675432.0,
+      "step": 329
+    },
+    {
+      "epoch": 0.49107142857142855,
+      "grad_norm": 2.447194748924661,
+      "learning_rate": 9.821428571428573e-06,
+      "loss": 0.7964,
+      "mean_token_accuracy": 0.7639085650444031,
+      "num_tokens": 124053357.0,
+      "step": 330
+    },
+    {
+      "epoch": 0.49255952380952384,
+      "grad_norm": 2.2856672882903144,
+      "learning_rate": 9.851190476190477e-06,
+      "loss": 0.7495,
+      "mean_token_accuracy": 0.7749939560890198,
+      "num_tokens": 124431071.0,
+      "step": 331
+    },
+    {
+      "epoch": 0.49404761904761907,
+      "grad_norm": 2.5311349852406995,
+      "learning_rate": 9.880952380952381e-06,
+      "loss": 0.7932,
+      "mean_token_accuracy": 0.7659996747970581,
+      "num_tokens": 124810261.0,
+      "step": 332
+    },
+    {
+      "epoch": 0.4955357142857143,
+      "grad_norm": 2.7962144344527067,
+      "learning_rate": 9.910714285714288e-06,
+      "loss": 0.8126,
+      "mean_token_accuracy": 0.7666003108024597,
+      "num_tokens": 125187945.0,
+      "step": 333
+    },
+    {
+      "epoch": 0.49702380952380953,
+      "grad_norm": 2.4575337340621854,
+      "learning_rate": 9.940476190476192e-06,
+      "loss": 0.8068,
+      "mean_token_accuracy": 0.7561527490615845,
+      "num_tokens": 125565972.0,
+      "step": 334
+    },
+    {
+      "epoch": 0.49851190476190477,
+      "grad_norm": 2.3379501641682463,
+      "learning_rate": 9.970238095238096e-06,
+      "loss": 0.8392,
+      "mean_token_accuracy": 0.7554892301559448,
+      "num_tokens": 125946626.0,
+      "step": 335
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 2.569994488439497,
+      "learning_rate": 1e-05,
+      "loss": 0.729,
+      "mean_token_accuracy": 0.7780126929283142,
+      "num_tokens": 126325317.0,
+      "step": 336
+    },
+    {
+      "epoch": 0.5014880952380952,
+      "grad_norm": 2.4118639403688924,
+      "learning_rate": 9.999997301787554e-06,
+      "loss": 0.8162,
+      "mean_token_accuracy": 0.7540923357009888,
+      "num_tokens": 126705897.0,
+      "step": 337
+    },
+    {
+      "epoch": 0.5029761904761905,
+      "grad_norm": 2.4955679801131954,
+      "learning_rate": 9.999989207153126e-06,
+      "loss": 0.7903,
+      "mean_token_accuracy": 0.7719624638557434,
+      "num_tokens": 127082545.0,
+      "step": 338
+    },
+    {
+      "epoch": 0.5044642857142857,
+      "grad_norm": 2.39829862645412,
+      "learning_rate": 9.999975716105453e-06,
+      "loss": 0.8332,
+      "mean_token_accuracy": 0.7481508851051331,
+      "num_tokens": 127456011.0,
+      "step": 339
+    },
+    {
+      "epoch": 0.5059523809523809,
+      "grad_norm": 2.586551856051117,
+      "learning_rate": 9.999956828659095e-06,
+      "loss": 0.7826,
+      "mean_token_accuracy": 0.7699803113937378,
+      "num_tokens": 127835062.0,
+      "step": 340
+    },
+    {
+      "epoch": 0.5074404761904762,
+      "grad_norm": 2.600295438122129,
+      "learning_rate": 9.99993254483444e-06,
+      "loss": 0.821,
+      "mean_token_accuracy": 0.7547000050544739,
+      "num_tokens": 128213115.0,
+      "step": 341
+    },
+    {
+      "epoch": 0.5089285714285714,
+      "grad_norm": 2.271967689304244,
+      "learning_rate": 9.999902864657691e-06,
+      "loss": 0.785,
+      "mean_token_accuracy": 0.7636333703994751,
+      "num_tokens": 128591379.0,
+      "step": 342
+    },
+    {
+      "epoch": 0.5104166666666666,
+      "grad_norm": 2.538432457008736,
+      "learning_rate": 9.999867788160888e-06,
+      "loss": 0.7872,
+      "mean_token_accuracy": 0.7678649425506592,
+      "num_tokens": 128969998.0,
+      "step": 343
+    },
+    {
+      "epoch": 0.5119047619047619,
+      "grad_norm": 2.419817858950018,
+      "learning_rate": 9.999827315381885e-06,
+      "loss": 0.8445,
+      "mean_token_accuracy": 0.7490323781967163,
+      "num_tokens": 129348172.0,
+      "step": 344
+    },
+    {
+      "epoch": 0.5133928571428571,
+      "grad_norm": 2.2953213079111014,
+      "learning_rate": 9.999781446364366e-06,
+      "loss": 0.7532,
+      "mean_token_accuracy": 0.772921085357666,
+      "num_tokens": 129726610.0,
+      "step": 345
+    },
+    {
+      "epoch": 0.5148809523809523,
+      "grad_norm": 2.2085808117745374,
+      "learning_rate": 9.999730181157832e-06,
+      "loss": 0.7375,
+      "mean_token_accuracy": 0.7793495059013367,
+      "num_tokens": 130103715.0,
+      "step": 346
+    },
+    {
+      "epoch": 0.5163690476190477,
+      "grad_norm": 2.424210465766237,
+      "learning_rate": 9.999673519817618e-06,
+      "loss": 0.8149,
+      "mean_token_accuracy": 0.7558252215385437,
+      "num_tokens": 130482160.0,
+      "step": 347
+    },
+    {
+      "epoch": 0.5178571428571429,
+      "grad_norm": 2.149910178843274,
+      "learning_rate": 9.999611462404874e-06,
+      "loss": 0.7468,
+      "mean_token_accuracy": 0.7753230333328247,
+      "num_tokens": 130860162.0,
+      "step": 348
+    },
+    {
+      "epoch": 0.5193452380952381,
+      "grad_norm": 2.266412484591112,
+      "learning_rate": 9.999544008986581e-06,
+      "loss": 0.7513,
+      "mean_token_accuracy": 0.771677553653717,
+      "num_tokens": 131237911.0,
+      "step": 349
+    },
+    {
+      "epoch": 0.5208333333333334,
+      "grad_norm": 2.4086947381657215,
+      "learning_rate": 9.999471159635538e-06,
+      "loss": 0.7585,
+      "mean_token_accuracy": 0.7731324434280396,
+      "num_tokens": 131616573.0,
+      "step": 350
+    },
+    {
+      "epoch": 0.5223214285714286,
+      "grad_norm": 2.4396056414557794,
+      "learning_rate": 9.999392914430372e-06,
+      "loss": 0.7989,
+      "mean_token_accuracy": 0.7677114009857178,
+      "num_tokens": 131991795.0,
+      "step": 351
+    },
+    {
+      "epoch": 0.5238095238095238,
+      "grad_norm": 2.6646147521871506,
+      "learning_rate": 9.99930927345553e-06,
+      "loss": 0.8389,
+      "mean_token_accuracy": 0.7588778734207153,
+      "num_tokens": 132370551.0,
+      "step": 352
+    },
+    {
+      "epoch": 0.5252976190476191,
+      "grad_norm": 2.047331068856892,
+      "learning_rate": 9.999220236801285e-06,
+      "loss": 0.7331,
+      "mean_token_accuracy": 0.7789658308029175,
+      "num_tokens": 132750528.0,
+      "step": 353
+    },
+    {
+      "epoch": 0.5267857142857143,
+      "grad_norm": 2.4360846319097225,
+      "learning_rate": 9.999125804563732e-06,
+      "loss": 0.7916,
+      "mean_token_accuracy": 0.7664345502853394,
+      "num_tokens": 133116828.0,
+      "step": 354
+    },
+    {
+      "epoch": 0.5282738095238095,
+      "grad_norm": 2.355660249088419,
+      "learning_rate": 9.999025976844794e-06,
+      "loss": 0.7568,
+      "mean_token_accuracy": 0.7717447876930237,
+      "num_tokens": 133495725.0,
+      "step": 355
+    },
+    {
+      "epoch": 0.5297619047619048,
+      "grad_norm": 2.776167050620993,
+      "learning_rate": 9.99892075375221e-06,
+      "loss": 0.7884,
+      "mean_token_accuracy": 0.7689514756202698,
+      "num_tokens": 133866199.0,
+      "step": 356
+    },
+    {
+      "epoch": 0.53125,
+      "grad_norm": 2.3095270455454044,
+      "learning_rate": 9.998810135399545e-06,
+      "loss": 0.8046,
+      "mean_token_accuracy": 0.7665999531745911,
+      "num_tokens": 134244588.0,
+      "step": 357
+    },
+    {
+      "epoch": 0.5327380952380952,
+      "grad_norm": 2.2264397255772663,
+      "learning_rate": 9.998694121906191e-06,
+      "loss": 0.8096,
+      "mean_token_accuracy": 0.7567612528800964,
+      "num_tokens": 134622978.0,
+      "step": 358
+    },
+    {
+      "epoch": 0.5342261904761905,
+      "grad_norm": 2.8287520932707633,
+      "learning_rate": 9.99857271339736e-06,
+      "loss": 0.8117,
+      "mean_token_accuracy": 0.7631921172142029,
+      "num_tokens": 135001209.0,
+      "step": 359
+    },
+    {
+      "epoch": 0.5357142857142857,
+      "grad_norm": 2.5064653242995854,
+      "learning_rate": 9.998445910004082e-06,
+      "loss": 0.8297,
+      "mean_token_accuracy": 0.7538565993309021,
+      "num_tokens": 135379721.0,
+      "step": 360
+    },
+    {
+      "epoch": 0.5372023809523809,
+      "grad_norm": 2.5740727767853104,
+      "learning_rate": 9.998313711863217e-06,
+      "loss": 0.8181,
+      "mean_token_accuracy": 0.7580350637435913,
+      "num_tokens": 135758131.0,
+      "step": 361
+    },
+    {
+      "epoch": 0.5386904761904762,
+      "grad_norm": 2.267927407952649,
+      "learning_rate": 9.998176119117445e-06,
+      "loss": 0.7609,
+      "mean_token_accuracy": 0.7719283699989319,
+      "num_tokens": 136137052.0,
+      "step": 362
+    },
+    {
+      "epoch": 0.5401785714285714,
+      "grad_norm": 2.2332994046261434,
+      "learning_rate": 9.998033131915266e-06,
+      "loss": 0.7818,
+      "mean_token_accuracy": 0.7636044025421143,
+      "num_tokens": 136515341.0,
+      "step": 363
+    },
+    {
+      "epoch": 0.5416666666666666,
+      "grad_norm": 2.926404492264879,
+      "learning_rate": 9.997884750411004e-06,
+      "loss": 0.8463,
+      "mean_token_accuracy": 0.7637007832527161,
+      "num_tokens": 136893088.0,
+      "step": 364
+    },
+    {
+      "epoch": 0.5431547619047619,
+      "grad_norm": 2.466238103593642,
+      "learning_rate": 9.997730974764805e-06,
+      "loss": 0.8298,
+      "mean_token_accuracy": 0.7566924691200256,
+      "num_tokens": 137271707.0,
+      "step": 365
+    },
+    {
+      "epoch": 0.5446428571428571,
+      "grad_norm": 2.2976526646347466,
+      "learning_rate": 9.997571805142638e-06,
+      "loss": 0.7461,
+      "mean_token_accuracy": 0.7752527594566345,
+      "num_tokens": 137649724.0,
+      "step": 366
+    },
+    {
+      "epoch": 0.5461309523809523,
+      "grad_norm": 2.563552892319996,
+      "learning_rate": 9.997407241716292e-06,
+      "loss": 0.8081,
+      "mean_token_accuracy": 0.764133870601654,
+      "num_tokens": 138028071.0,
+      "step": 367
+    },
+    {
+      "epoch": 0.5476190476190477,
+      "grad_norm": 2.80757274390489,
+      "learning_rate": 9.99723728466338e-06,
+      "loss": 0.8842,
+      "mean_token_accuracy": 0.740105152130127,
+      "num_tokens": 138405751.0,
+      "step": 368
+    },
+    {
+      "epoch": 0.5491071428571429,
+      "grad_norm": 2.4705809301504895,
+      "learning_rate": 9.997061934167328e-06,
+      "loss": 0.8467,
+      "mean_token_accuracy": 0.7475184202194214,
+      "num_tokens": 138784552.0,
+      "step": 369
+    },
+    {
+      "epoch": 0.5505952380952381,
+      "grad_norm": 2.5148552472649714,
+      "learning_rate": 9.996881190417393e-06,
+      "loss": 0.8727,
+      "mean_token_accuracy": 0.7485411763191223,
+      "num_tokens": 139162215.0,
+      "step": 370
+    },
+    {
+      "epoch": 0.5520833333333334,
+      "grad_norm": 2.47420457472063,
+      "learning_rate": 9.996695053608651e-06,
+      "loss": 0.7781,
+      "mean_token_accuracy": 0.7645584940910339,
+      "num_tokens": 139541199.0,
+      "step": 371
+    },
+    {
+      "epoch": 0.5535714285714286,
+      "grad_norm": 3.047469055558771,
+      "learning_rate": 9.996503523941994e-06,
+      "loss": 0.8302,
+      "mean_token_accuracy": 0.7639842629432678,
+      "num_tokens": 139917031.0,
+      "step": 372
+    },
+    {
+      "epoch": 0.5550595238095238,
+      "grad_norm": 2.6774942937706174,
+      "learning_rate": 9.996306601624137e-06,
+      "loss": 0.8262,
+      "mean_token_accuracy": 0.7508988380432129,
+      "num_tokens": 140295803.0,
+      "step": 373
+    },
+    {
+      "epoch": 0.5565476190476191,
+      "grad_norm": 2.4070801707301217,
+      "learning_rate": 9.996104286867616e-06,
+      "loss": 0.8165,
+      "mean_token_accuracy": 0.7569167017936707,
+      "num_tokens": 140675781.0,
+      "step": 374
+    },
+    {
+      "epoch": 0.5580357142857143,
+      "grad_norm": 2.6564663015672516,
+      "learning_rate": 9.995896579890785e-06,
+      "loss": 0.8075,
+      "mean_token_accuracy": 0.7594267725944519,
+      "num_tokens": 141054432.0,
+      "step": 375
+    },
+    {
+      "epoch": 0.5595238095238095,
+      "grad_norm": 2.5270042764580984,
+      "learning_rate": 9.995683480917821e-06,
+      "loss": 0.7803,
+      "mean_token_accuracy": 0.7670716643333435,
+      "num_tokens": 141434071.0,
+      "step": 376
+    },
+    {
+      "epoch": 0.5610119047619048,
+      "grad_norm": 2.7957183036698847,
+      "learning_rate": 9.995464990178718e-06,
+      "loss": 0.8665,
+      "mean_token_accuracy": 0.7420497536659241,
+      "num_tokens": 141811681.0,
+      "step": 377
+    },
+    {
+      "epoch": 0.5625,
+      "grad_norm": 2.4638207741805975,
+      "learning_rate": 9.99524110790929e-06,
+      "loss": 0.8959,
+      "mean_token_accuracy": 0.7386406064033508,
+      "num_tokens": 142185697.0,
+      "step": 378
+    },
+    {
+      "epoch": 0.5639880952380952,
+      "grad_norm": 2.242612681540062,
+      "learning_rate": 9.995011834351168e-06,
+      "loss": 0.7993,
+      "mean_token_accuracy": 0.7624521255493164,
+      "num_tokens": 142555353.0,
+      "step": 379
+    },
+    {
+      "epoch": 0.5654761904761905,
+      "grad_norm": 2.5972370090093495,
+      "learning_rate": 9.994777169751806e-06,
+      "loss": 0.7839,
+      "mean_token_accuracy": 0.7624208331108093,
+      "num_tokens": 142934985.0,
+      "step": 380
+    },
+    {
+      "epoch": 0.5669642857142857,
+      "grad_norm": 2.6397163714897234,
+      "learning_rate": 9.994537114364471e-06,
+      "loss": 0.8504,
+      "mean_token_accuracy": 0.7537856101989746,
+      "num_tokens": 143313392.0,
+      "step": 381
+    },
+    {
+      "epoch": 0.5684523809523809,
+      "grad_norm": 2.1910873371166697,
+      "learning_rate": 9.994291668448255e-06,
+      "loss": 0.7587,
+      "mean_token_accuracy": 0.773768424987793,
+      "num_tokens": 143691834.0,
+      "step": 382
+    },
+    {
+      "epoch": 0.5699404761904762,
+      "grad_norm": 2.24375580465422,
+      "learning_rate": 9.994040832268059e-06,
+      "loss": 0.7622,
+      "mean_token_accuracy": 0.7697617411613464,
+      "num_tokens": 144070826.0,
+      "step": 383
+    },
+    {
+      "epoch": 0.5714285714285714,
+      "grad_norm": 2.7116771719877892,
+      "learning_rate": 9.993784606094612e-06,
+      "loss": 0.8263,
+      "mean_token_accuracy": 0.7554301619529724,
+      "num_tokens": 144446634.0,
+      "step": 384
+    },
+    {
+      "epoch": 0.5729166666666666,
+      "grad_norm": 2.61995935788304,
+      "learning_rate": 9.993522990204453e-06,
+      "loss": 0.7639,
+      "mean_token_accuracy": 0.7737082242965698,
+      "num_tokens": 144824982.0,
+      "step": 385
+    },
+    {
+      "epoch": 0.5744047619047619,
+      "grad_norm": 2.4082078600209167,
+      "learning_rate": 9.99325598487994e-06,
+      "loss": 0.8085,
+      "mean_token_accuracy": 0.757756233215332,
+      "num_tokens": 145202847.0,
+      "step": 386
+    },
+    {
+      "epoch": 0.5758928571428571,
+      "grad_norm": 2.4804420031957726,
+      "learning_rate": 9.992983590409246e-06,
+      "loss": 0.7676,
+      "mean_token_accuracy": 0.7673775553703308,
+      "num_tokens": 145577740.0,
+      "step": 387
+    },
+    {
+      "epoch": 0.5773809523809523,
+      "grad_norm": 2.2491057434194506,
+      "learning_rate": 9.992705807086364e-06,
+      "loss": 0.8426,
+      "mean_token_accuracy": 0.7512651085853577,
+      "num_tokens": 145942329.0,
+      "step": 388
+    },
+    {
+      "epoch": 0.5788690476190477,
+      "grad_norm": 2.257006945551683,
+      "learning_rate": 9.992422635211103e-06,
+      "loss": 0.8019,
+      "mean_token_accuracy": 0.7553571462631226,
+      "num_tokens": 146320786.0,
+      "step": 389
+    },
+    {
+      "epoch": 0.5803571428571429,
+      "grad_norm": 2.4499140212268147,
+      "learning_rate": 9.992134075089085e-06,
+      "loss": 0.8049,
+      "mean_token_accuracy": 0.755973756313324,
+      "num_tokens": 146697330.0,
+      "step": 390
+    },
+    {
+      "epoch": 0.5818452380952381,
+      "grad_norm": 2.4216831781664268,
+      "learning_rate": 9.991840127031746e-06,
+      "loss": 0.7903,
+      "mean_token_accuracy": 0.7613458633422852,
+      "num_tokens": 147074988.0,
+      "step": 391
+    },
+    {
+      "epoch": 0.5833333333333334,
+      "grad_norm": 2.530572751908938,
+      "learning_rate": 9.991540791356342e-06,
+      "loss": 0.7876,
+      "mean_token_accuracy": 0.756466269493103,
+      "num_tokens": 147453642.0,
+      "step": 392
+    },
+    {
+      "epoch": 0.5848214285714286,
+      "grad_norm": 2.3720062124471917,
+      "learning_rate": 9.99123606838594e-06,
+      "loss": 0.7971,
+      "mean_token_accuracy": 0.7601131796836853,
+      "num_tokens": 147820662.0,
+      "step": 393
+    },
+    {
+      "epoch": 0.5863095238095238,
+      "grad_norm": 2.4028093068759424,
+      "learning_rate": 9.990925958449428e-06,
+      "loss": 0.8593,
+      "mean_token_accuracy": 0.7485013604164124,
+      "num_tokens": 148184252.0,
+      "step": 394
+    },
+    {
+      "epoch": 0.5877976190476191,
+      "grad_norm": 2.428625977170398,
+      "learning_rate": 9.990610461881497e-06,
+      "loss": 0.7919,
+      "mean_token_accuracy": 0.7672779560089111,
+      "num_tokens": 148563673.0,
+      "step": 395
+    },
+    {
+      "epoch": 0.5892857142857143,
+      "grad_norm": 2.3385612817916748,
+      "learning_rate": 9.990289579022661e-06,
+      "loss": 0.8443,
+      "mean_token_accuracy": 0.7472571730613708,
+      "num_tokens": 148943010.0,
+      "step": 396
+    },
+    {
+      "epoch": 0.5907738095238095,
+      "grad_norm": 2.5276050259047453,
+      "learning_rate": 9.989963310219242e-06,
+      "loss": 0.8265,
+      "mean_token_accuracy": 0.7647454738616943,
+      "num_tokens": 149319608.0,
+      "step": 397
+    },
+    {
+      "epoch": 0.5922619047619048,
+      "grad_norm": 2.360338112738273,
+      "learning_rate": 9.989631655823377e-06,
+      "loss": 0.7399,
+      "mean_token_accuracy": 0.7736846804618835,
+      "num_tokens": 149691064.0,
+      "step": 398
+    },
+    {
+      "epoch": 0.59375,
+      "grad_norm": 2.420700691067952,
+      "learning_rate": 9.989294616193018e-06,
+      "loss": 0.8118,
+      "mean_token_accuracy": 0.7629058361053467,
+      "num_tokens": 150069418.0,
+      "step": 399
+    },
+    {
+      "epoch": 0.5952380952380952,
+      "grad_norm": 2.505040412484911,
+      "learning_rate": 9.988952191691925e-06,
+      "loss": 0.8588,
+      "mean_token_accuracy": 0.7512460947036743,
+      "num_tokens": 150446391.0,
+      "step": 400
+    },
+    {
+      "epoch": 0.5967261904761905,
+      "grad_norm": 2.5923956220474884,
+      "learning_rate": 9.988604382689672e-06,
+      "loss": 0.8527,
+      "mean_token_accuracy": 0.7494648694992065,
+      "num_tokens": 150824582.0,
+      "step": 401
+    },
+    {
+      "epoch": 0.5982142857142857,
+      "grad_norm": 2.313028337915683,
+      "learning_rate": 9.988251189561645e-06,
+      "loss": 0.8403,
+      "mean_token_accuracy": 0.7523447275161743,
+      "num_tokens": 151203293.0,
+      "step": 402
+    },
+    {
+      "epoch": 0.5997023809523809,
+      "grad_norm": 2.352133875777309,
+      "learning_rate": 9.987892612689039e-06,
+      "loss": 0.8232,
+      "mean_token_accuracy": 0.7607864737510681,
+      "num_tokens": 151582887.0,
+      "step": 403
+    },
+    {
+      "epoch": 0.6011904761904762,
+      "grad_norm": 2.440404547512422,
+      "learning_rate": 9.98752865245886e-06,
+      "loss": 0.851,
+      "mean_token_accuracy": 0.7486730217933655,
+      "num_tokens": 151960674.0,
+      "step": 404
+    },
+    {
+      "epoch": 0.6026785714285714,
+      "grad_norm": 2.7237825875308257,
+      "learning_rate": 9.987159309263926e-06,
+      "loss": 0.8692,
+      "mean_token_accuracy": 0.7472294569015503,
+      "num_tokens": 152337931.0,
+      "step": 405
+    },
+    {
+      "epoch": 0.6041666666666666,
+      "grad_norm": 2.4224912360371884,
+      "learning_rate": 9.986784583502863e-06,
+      "loss": 0.7234,
+      "mean_token_accuracy": 0.780518651008606,
+      "num_tokens": 152704418.0,
+      "step": 406
+    },
+    {
+      "epoch": 0.6056547619047619,
+      "grad_norm": 2.4365701476983697,
+      "learning_rate": 9.986404475580107e-06,
+      "loss": 0.7836,
+      "mean_token_accuracy": 0.7690278887748718,
+      "num_tokens": 153082369.0,
+      "step": 407
+    },
+    {
+      "epoch": 0.6071428571428571,
+      "grad_norm": 2.4428618043548984,
+      "learning_rate": 9.986018985905901e-06,
+      "loss": 0.8093,
+      "mean_token_accuracy": 0.7690455317497253,
+      "num_tokens": 153459268.0,
+      "step": 408
+    },
+    {
+      "epoch": 0.6086309523809523,
+      "grad_norm": 2.3550778945747526,
+      "learning_rate": 9.985628114896302e-06,
+      "loss": 0.7924,
+      "mean_token_accuracy": 0.7648941874504089,
+      "num_tokens": 153826921.0,
+      "step": 409
+    },
+    {
+      "epoch": 0.6101190476190477,
+      "grad_norm": 2.322300012241885,
+      "learning_rate": 9.985231862973167e-06,
+      "loss": 0.7912,
+      "mean_token_accuracy": 0.7654976844787598,
+      "num_tokens": 154187442.0,
+      "step": 410
+    },
+    {
+      "epoch": 0.6116071428571429,
+      "grad_norm": 2.5460194051792007,
+      "learning_rate": 9.98483023056417e-06,
+      "loss": 0.8363,
+      "mean_token_accuracy": 0.7568179965019226,
+      "num_tokens": 154561336.0,
+      "step": 411
+    },
+    {
+      "epoch": 0.6130952380952381,
+      "grad_norm": 2.672408719101405,
+      "learning_rate": 9.984423218102783e-06,
+      "loss": 0.83,
+      "mean_token_accuracy": 0.7565404772758484,
+      "num_tokens": 154938159.0,
+      "step": 412
+    },
+    {
+      "epoch": 0.6145833333333334,
+      "grad_norm": 2.219708087761756,
+      "learning_rate": 9.984010826028289e-06,
+      "loss": 0.7167,
+      "mean_token_accuracy": 0.7817323207855225,
+      "num_tokens": 155303667.0,
+      "step": 413
+    },
+    {
+      "epoch": 0.6160714285714286,
+      "grad_norm": 2.4125197094727118,
+      "learning_rate": 9.983593054785776e-06,
+      "loss": 0.7435,
+      "mean_token_accuracy": 0.7794288992881775,
+      "num_tokens": 155663105.0,
+      "step": 414
+    },
+    {
+      "epoch": 0.6175595238095238,
+      "grad_norm": 3.1354047315649134,
+      "learning_rate": 9.98316990482614e-06,
+      "loss": 0.7909,
+      "mean_token_accuracy": 0.7686386704444885,
+      "num_tokens": 156041823.0,
+      "step": 415
+    },
+    {
+      "epoch": 0.6190476190476191,
+      "grad_norm": 2.2176506767031907,
+      "learning_rate": 9.982741376606077e-06,
+      "loss": 0.8106,
+      "mean_token_accuracy": 0.7625644207000732,
+      "num_tokens": 156420090.0,
+      "step": 416
+    },
+    {
+      "epoch": 0.6205357142857143,
+      "grad_norm": 2.317799091265602,
+      "learning_rate": 9.982307470588097e-06,
+      "loss": 0.7552,
+      "mean_token_accuracy": 0.7706335783004761,
+      "num_tokens": 156798195.0,
+      "step": 417
+    },
+    {
+      "epoch": 0.6220238095238095,
+      "grad_norm": 2.221899616179134,
+      "learning_rate": 9.981868187240501e-06,
+      "loss": 0.7871,
+      "mean_token_accuracy": 0.7630293965339661,
+      "num_tokens": 157176946.0,
+      "step": 418
+    },
+    {
+      "epoch": 0.6235119047619048,
+      "grad_norm": 2.2177109312185492,
+      "learning_rate": 9.981423527037409e-06,
+      "loss": 0.7759,
+      "mean_token_accuracy": 0.7675144672393799,
+      "num_tokens": 157554180.0,
+      "step": 419
+    },
+    {
+      "epoch": 0.625,
+      "grad_norm": 2.2671260836103877,
+      "learning_rate": 9.980973490458728e-06,
+      "loss": 0.7783,
+      "mean_token_accuracy": 0.7719668745994568,
+      "num_tokens": 157932023.0,
+      "step": 420
+    },
+    {
+      "epoch": 0.6264880952380952,
+      "grad_norm": 2.335445219839476,
+      "learning_rate": 9.98051807799018e-06,
+      "loss": 0.8071,
+      "mean_token_accuracy": 0.7643492221832275,
+      "num_tokens": 158306061.0,
+      "step": 421
+    },
+    {
+      "epoch": 0.6279761904761905,
+      "grad_norm": 2.3011564420918855,
+      "learning_rate": 9.980057290123286e-06,
+      "loss": 0.7943,
+      "mean_token_accuracy": 0.7696520090103149,
+      "num_tokens": 158683986.0,
+      "step": 422
+    },
+    {
+      "epoch": 0.6294642857142857,
+      "grad_norm": 2.241023775129557,
+      "learning_rate": 9.979591127355365e-06,
+      "loss": 0.7593,
+      "mean_token_accuracy": 0.7748115658760071,
+      "num_tokens": 159060870.0,
+      "step": 423
+    },
+    {
+      "epoch": 0.6309523809523809,
+      "grad_norm": 2.4944378203246584,
+      "learning_rate": 9.97911959018954e-06,
+      "loss": 0.7931,
+      "mean_token_accuracy": 0.7732247710227966,
+      "num_tokens": 159437802.0,
+      "step": 424
+    },
+    {
+      "epoch": 0.6324404761904762,
+      "grad_norm": 2.2953739177882215,
+      "learning_rate": 9.978642679134733e-06,
+      "loss": 0.8613,
+      "mean_token_accuracy": 0.7465783953666687,
+      "num_tokens": 159813340.0,
+      "step": 425
+    },
+    {
+      "epoch": 0.6339285714285714,
+      "grad_norm": 2.4466237589807744,
+      "learning_rate": 9.978160394705669e-06,
+      "loss": 0.7946,
+      "mean_token_accuracy": 0.7669371962547302,
+      "num_tokens": 160191059.0,
+      "step": 426
+    },
+    {
+      "epoch": 0.6354166666666666,
+      "grad_norm": 2.1643762175841053,
+      "learning_rate": 9.97767273742287e-06,
+      "loss": 0.7838,
+      "mean_token_accuracy": 0.7665217518806458,
+      "num_tokens": 160569245.0,
+      "step": 427
+    },
+    {
+      "epoch": 0.6369047619047619,
+      "grad_norm": 2.2534906145974154,
+      "learning_rate": 9.977179707812656e-06,
+      "loss": 0.8244,
+      "mean_token_accuracy": 0.7555648684501648,
+      "num_tokens": 160947726.0,
+      "step": 428
+    },
+    {
+      "epoch": 0.6383928571428571,
+      "grad_norm": 2.2610777620877385,
+      "learning_rate": 9.976681306407147e-06,
+      "loss": 0.8545,
+      "mean_token_accuracy": 0.7479950189590454,
+      "num_tokens": 161325530.0,
+      "step": 429
+    },
+    {
+      "epoch": 0.6398809523809523,
+      "grad_norm": 2.3880389833018554,
+      "learning_rate": 9.97617753374426e-06,
+      "loss": 0.8514,
+      "mean_token_accuracy": 0.7529988884925842,
+      "num_tokens": 161703143.0,
+      "step": 430
+    },
+    {
+      "epoch": 0.6413690476190477,
+      "grad_norm": 2.219594190648269,
+      "learning_rate": 9.975668390367711e-06,
+      "loss": 0.8207,
+      "mean_token_accuracy": 0.7598025798797607,
+      "num_tokens": 162081048.0,
+      "step": 431
+    },
+    {
+      "epoch": 0.6428571428571429,
+      "grad_norm": 2.5109899090423893,
+      "learning_rate": 9.975153876827008e-06,
+      "loss": 0.8515,
+      "mean_token_accuracy": 0.7555494904518127,
+      "num_tokens": 162458535.0,
+      "step": 432
+    },
+    {
+      "epoch": 0.6443452380952381,
+      "grad_norm": 2.107961637223581,
+      "learning_rate": 9.974633993677459e-06,
+      "loss": 0.7649,
+      "mean_token_accuracy": 0.7717664241790771,
+      "num_tokens": 162838362.0,
+      "step": 433
+    },
+    {
+      "epoch": 0.6458333333333334,
+      "grad_norm": 2.0967033434004962,
+      "learning_rate": 9.974108741480167e-06,
+      "loss": 0.8334,
+      "mean_token_accuracy": 0.754153311252594,
+      "num_tokens": 163219136.0,
+      "step": 434
+    },
+    {
+      "epoch": 0.6473214285714286,
+      "grad_norm": 2.4386419684188527,
+      "learning_rate": 9.973578120802025e-06,
+      "loss": 0.837,
+      "mean_token_accuracy": 0.7590125799179077,
+      "num_tokens": 163596593.0,
+      "step": 435
+    },
+    {
+      "epoch": 0.6488095238095238,
+      "grad_norm": 2.5853061110071187,
+      "learning_rate": 9.97304213221573e-06,
+      "loss": 0.8683,
+      "mean_token_accuracy": 0.7562277317047119,
+      "num_tokens": 163972946.0,
+      "step": 436
+    },
+    {
+      "epoch": 0.6502976190476191,
+      "grad_norm": 2.3282865884918538,
+      "learning_rate": 9.972500776299763e-06,
+      "loss": 0.8107,
+      "mean_token_accuracy": 0.76071697473526,
+      "num_tokens": 164350452.0,
+      "step": 437
+    },
+    {
+      "epoch": 0.6517857142857143,
+      "grad_norm": 2.1075782097688154,
+      "learning_rate": 9.9719540536384e-06,
+      "loss": 0.7753,
+      "mean_token_accuracy": 0.7655922770500183,
+      "num_tokens": 164730005.0,
+      "step": 438
+    },
+    {
+      "epoch": 0.6532738095238095,
+      "grad_norm": 2.136849081521859,
+      "learning_rate": 9.971401964821712e-06,
+      "loss": 0.7908,
+      "mean_token_accuracy": 0.7645189166069031,
+      "num_tokens": 165108003.0,
+      "step": 439
+    },
+    {
+      "epoch": 0.6547619047619048,
+      "grad_norm": 2.4440489177958304,
+      "learning_rate": 9.97084451044556e-06,
+      "loss": 0.8113,
+      "mean_token_accuracy": 0.7557384967803955,
+      "num_tokens": 165481250.0,
+      "step": 440
+    },
+    {
+      "epoch": 0.65625,
+      "grad_norm": 2.3709515878104117,
+      "learning_rate": 9.970281691111598e-06,
+      "loss": 0.7846,
+      "mean_token_accuracy": 0.766571581363678,
+      "num_tokens": 165860615.0,
+      "step": 441
+    },
+    {
+      "epoch": 0.6577380952380952,
+      "grad_norm": 2.2571639853919,
+      "learning_rate": 9.969713507427268e-06,
+      "loss": 0.8055,
+      "mean_token_accuracy": 0.7627787590026855,
+      "num_tokens": 166226000.0,
+      "step": 442
+    },
+    {
+      "epoch": 0.6592261904761905,
+      "grad_norm": 2.6090009548344866,
+      "learning_rate": 9.969139960005798e-06,
+      "loss": 0.8068,
+      "mean_token_accuracy": 0.7622710466384888,
+      "num_tokens": 166602546.0,
+      "step": 443
+    },
+    {
+      "epoch": 0.6607142857142857,
+      "grad_norm": 2.269894946643627,
+      "learning_rate": 9.968561049466214e-06,
+      "loss": 0.8177,
+      "mean_token_accuracy": 0.7578491568565369,
+      "num_tokens": 166980370.0,
+      "step": 444
+    },
+    {
+      "epoch": 0.6622023809523809,
+      "grad_norm": 2.405940469284995,
+      "learning_rate": 9.967976776433323e-06,
+      "loss": 0.864,
+      "mean_token_accuracy": 0.7478210926055908,
+      "num_tokens": 167355958.0,
+      "step": 445
+    },
+    {
+      "epoch": 0.6636904761904762,
+      "grad_norm": 2.0254279270966054,
+      "learning_rate": 9.967387141537721e-06,
+      "loss": 0.7539,
+      "mean_token_accuracy": 0.7708075046539307,
+      "num_tokens": 167708856.0,
+      "step": 446
+    },
+    {
+      "epoch": 0.6651785714285714,
+      "grad_norm": 2.4490992955749764,
+      "learning_rate": 9.966792145415796e-06,
+      "loss": 0.7604,
+      "mean_token_accuracy": 0.7726967930793762,
+      "num_tokens": 168085656.0,
+      "step": 447
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 2.455028274555107,
+      "learning_rate": 9.966191788709716e-06,
+      "loss": 0.8198,
+      "mean_token_accuracy": 0.755533754825592,
+      "num_tokens": 168463740.0,
+      "step": 448
+    },
+    {
+      "epoch": 0.6681547619047619,
+      "grad_norm": 2.244320314663738,
+      "learning_rate": 9.965586072067437e-06,
+      "loss": 0.7851,
+      "mean_token_accuracy": 0.7652745246887207,
+      "num_tokens": 168840914.0,
+      "step": 449
+    },
+    {
+      "epoch": 0.6696428571428571,
+      "grad_norm": 2.5139650868154657,
+      "learning_rate": 9.964974996142699e-06,
+      "loss": 0.8502,
+      "mean_token_accuracy": 0.7446865439414978,
+      "num_tokens": 169218697.0,
+      "step": 450
+    },
+    {
+      "epoch": 0.6711309523809523,
+      "grad_norm": 2.0271194981796206,
+      "learning_rate": 9.964358561595029e-06,
+      "loss": 0.7521,
+      "mean_token_accuracy": 0.772848904132843,
+      "num_tokens": 169597468.0,
+      "step": 451
+    },
+    {
+      "epoch": 0.6726190476190477,
+      "grad_norm": 2.0957372633203772,
+      "learning_rate": 9.963736769089734e-06,
+      "loss": 0.7781,
+      "mean_token_accuracy": 0.7713422179222107,
+      "num_tokens": 169976217.0,
+      "step": 452
+    },
+    {
+      "epoch": 0.6741071428571429,
+      "grad_norm": 2.2763037794877934,
+      "learning_rate": 9.963109619297905e-06,
+      "loss": 0.7625,
+      "mean_token_accuracy": 0.7682791352272034,
+      "num_tokens": 170355103.0,
+      "step": 453
+    },
+    {
+      "epoch": 0.6755952380952381,
+      "grad_norm": 2.5840555079329617,
+      "learning_rate": 9.962477112896417e-06,
+      "loss": 0.7971,
+      "mean_token_accuracy": 0.7557924389839172,
+      "num_tokens": 170731228.0,
+      "step": 454
+    },
+    {
+      "epoch": 0.6770833333333334,
+      "grad_norm": 2.3098065933931715,
+      "learning_rate": 9.961839250567925e-06,
+      "loss": 0.7733,
+      "mean_token_accuracy": 0.7639572620391846,
+      "num_tokens": 171110167.0,
+      "step": 455
+    },
+    {
+      "epoch": 0.6785714285714286,
+      "grad_norm": 2.311197921276789,
+      "learning_rate": 9.961196033000862e-06,
+      "loss": 0.7775,
+      "mean_token_accuracy": 0.7739092111587524,
+      "num_tokens": 171488333.0,
+      "step": 456
+    },
+    {
+      "epoch": 0.6800595238095238,
+      "grad_norm": 2.4760294845803394,
+      "learning_rate": 9.960547460889443e-06,
+      "loss": 0.8349,
+      "mean_token_accuracy": 0.7532545924186707,
+      "num_tokens": 171857473.0,
+      "step": 457
+    },
+    {
+      "epoch": 0.6815476190476191,
+      "grad_norm": 2.425913284527865,
+      "learning_rate": 9.959893534933665e-06,
+      "loss": 0.8103,
+      "mean_token_accuracy": 0.755517303943634,
+      "num_tokens": 172237400.0,
+      "step": 458
+    },
+    {
+      "epoch": 0.6830357142857143,
+      "grad_norm": 2.0537550046343287,
+      "learning_rate": 9.9592342558393e-06,
+      "loss": 0.7563,
+      "mean_token_accuracy": 0.7692558169364929,
+      "num_tokens": 172617234.0,
+      "step": 459
+    },
+    {
+      "epoch": 0.6845238095238095,
+      "grad_norm": 2.4130539607872117,
+      "learning_rate": 9.958569624317894e-06,
+      "loss": 0.7787,
+      "mean_token_accuracy": 0.7700129151344299,
+      "num_tokens": 172980692.0,
+      "step": 460
+    },
+    {
+      "epoch": 0.6860119047619048,
+      "grad_norm": 2.158721025170757,
+      "learning_rate": 9.957899641086777e-06,
+      "loss": 0.7747,
+      "mean_token_accuracy": 0.7712138295173645,
+      "num_tokens": 173360695.0,
+      "step": 461
+    },
+    {
+      "epoch": 0.6875,
+      "grad_norm": 2.4892540354477104,
+      "learning_rate": 9.957224306869053e-06,
+      "loss": 0.8246,
+      "mean_token_accuracy": 0.7516012191772461,
+      "num_tokens": 173723466.0,
+      "step": 462
+    },
+    {
+      "epoch": 0.6889880952380952,
+      "grad_norm": 2.5724991707703335,
+      "learning_rate": 9.956543622393597e-06,
+      "loss": 0.8068,
+      "mean_token_accuracy": 0.7630789279937744,
+      "num_tokens": 174100606.0,
+      "step": 463
+    },
+    {
+      "epoch": 0.6904761904761905,
+      "grad_norm": 2.3466415019716607,
+      "learning_rate": 9.955857588395065e-06,
+      "loss": 0.7841,
+      "mean_token_accuracy": 0.7606614232063293,
+      "num_tokens": 174479191.0,
+      "step": 464
+    },
+    {
+      "epoch": 0.6919642857142857,
+      "grad_norm": 2.1837378709266844,
+      "learning_rate": 9.95516620561388e-06,
+      "loss": 0.825,
+      "mean_token_accuracy": 0.748581051826477,
+      "num_tokens": 174857822.0,
+      "step": 465
+    },
+    {
+      "epoch": 0.6934523809523809,
+      "grad_norm": 2.3243963495571074,
+      "learning_rate": 9.954469474796242e-06,
+      "loss": 0.759,
+      "mean_token_accuracy": 0.7778409123420715,
+      "num_tokens": 175237620.0,
+      "step": 466
+    },
+    {
+      "epoch": 0.6949404761904762,
+      "grad_norm": 2.2063928930371897,
+      "learning_rate": 9.953767396694123e-06,
+      "loss": 0.7488,
+      "mean_token_accuracy": 0.7763053178787231,
+      "num_tokens": 175615945.0,
+      "step": 467
+    },
+    {
+      "epoch": 0.6964285714285714,
+      "grad_norm": 2.3462591387341765,
+      "learning_rate": 9.953059972065264e-06,
+      "loss": 0.8138,
+      "mean_token_accuracy": 0.7595534324645996,
+      "num_tokens": 175994233.0,
+      "step": 468
+    },
+    {
+      "epoch": 0.6979166666666666,
+      "grad_norm": 2.310659771915979,
+      "learning_rate": 9.952347201673181e-06,
+      "loss": 0.8065,
+      "mean_token_accuracy": 0.7571525573730469,
+      "num_tokens": 176373806.0,
+      "step": 469
+    },
+    {
+      "epoch": 0.6994047619047619,
+      "grad_norm": 2.5171934777343905,
+      "learning_rate": 9.95162908628715e-06,
+      "loss": 0.78,
+      "mean_token_accuracy": 0.7694202661514282,
+      "num_tokens": 176752551.0,
+      "step": 470
+    },
+    {
+      "epoch": 0.7008928571428571,
+      "grad_norm": 2.4004813477165086,
+      "learning_rate": 9.950905626682229e-06,
+      "loss": 0.858,
+      "mean_token_accuracy": 0.7454370856285095,
+      "num_tokens": 177129583.0,
+      "step": 471
+    },
+    {
+      "epoch": 0.7023809523809523,
+      "grad_norm": 2.1800621718006625,
+      "learning_rate": 9.950176823639233e-06,
+      "loss": 0.7489,
+      "mean_token_accuracy": 0.7770506143569946,
+      "num_tokens": 177508144.0,
+      "step": 472
+    },
+    {
+      "epoch": 0.7038690476190477,
+      "grad_norm": 2.8377683881735334,
+      "learning_rate": 9.949442677944749e-06,
+      "loss": 0.8057,
+      "mean_token_accuracy": 0.760857880115509,
+      "num_tokens": 177886115.0,
+      "step": 473
+    },
+    {
+      "epoch": 0.7053571428571429,
+      "grad_norm": 2.724980097579837,
+      "learning_rate": 9.948703190391131e-06,
+      "loss": 0.8619,
+      "mean_token_accuracy": 0.7536898851394653,
+      "num_tokens": 178262915.0,
+      "step": 474
+    },
+    {
+      "epoch": 0.7068452380952381,
+      "grad_norm": 2.363388776898293,
+      "learning_rate": 9.947958361776495e-06,
+      "loss": 0.8243,
+      "mean_token_accuracy": 0.7559298276901245,
+      "num_tokens": 178641484.0,
+      "step": 475
+    },
+    {
+      "epoch": 0.7083333333333334,
+      "grad_norm": 2.217777837832058,
+      "learning_rate": 9.947208192904722e-06,
+      "loss": 0.8047,
+      "mean_token_accuracy": 0.7636073231697083,
+      "num_tokens": 179019119.0,
+      "step": 476
+    },
+    {
+      "epoch": 0.7098214285714286,
+      "grad_norm": 2.3930860310900375,
+      "learning_rate": 9.946452684585464e-06,
+      "loss": 0.8244,
+      "mean_token_accuracy": 0.7533463835716248,
+      "num_tokens": 179397692.0,
+      "step": 477
+    },
+    {
+      "epoch": 0.7113095238095238,
+      "grad_norm": 2.3337987056548135,
+      "learning_rate": 9.945691837634121e-06,
+      "loss": 0.7295,
+      "mean_token_accuracy": 0.7769878506660461,
+      "num_tokens": 179769011.0,
+      "step": 478
+    },
+    {
+      "epoch": 0.7127976190476191,
+      "grad_norm": 2.077186740035528,
+      "learning_rate": 9.944925652871871e-06,
+      "loss": 0.7867,
+      "mean_token_accuracy": 0.7658684253692627,
+      "num_tokens": 180150995.0,
+      "step": 479
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 2.2685137295760005,
+      "learning_rate": 9.944154131125643e-06,
+      "loss": 0.8032,
+      "mean_token_accuracy": 0.7661324143409729,
+      "num_tokens": 180529971.0,
+      "step": 480
+    },
+    {
+      "epoch": 0.7157738095238095,
+      "grad_norm": 2.4847952700238376,
+      "learning_rate": 9.94337727322813e-06,
+      "loss": 0.7706,
+      "mean_token_accuracy": 0.7721290588378906,
+      "num_tokens": 180908143.0,
+      "step": 481
+    },
+    {
+      "epoch": 0.7172619047619048,
+      "grad_norm": 2.3656246107497236,
+      "learning_rate": 9.94259508001778e-06,
+      "loss": 0.7703,
+      "mean_token_accuracy": 0.7784695625305176,
+      "num_tokens": 181285814.0,
+      "step": 482
+    },
+    {
+      "epoch": 0.71875,
+      "grad_norm": 2.0988666876326887,
+      "learning_rate": 9.941807552338805e-06,
+      "loss": 0.7341,
+      "mean_token_accuracy": 0.7784519791603088,
+      "num_tokens": 181665737.0,
+      "step": 483
+    },
+    {
+      "epoch": 0.7202380952380952,
+      "grad_norm": 2.4710171148170836,
+      "learning_rate": 9.941014691041171e-06,
+      "loss": 0.795,
+      "mean_token_accuracy": 0.7650205492973328,
+      "num_tokens": 182042518.0,
+      "step": 484
+    },
+    {
+      "epoch": 0.7217261904761905,
+      "grad_norm": 2.1622478137516064,
+      "learning_rate": 9.940216496980602e-06,
+      "loss": 0.7759,
+      "mean_token_accuracy": 0.766381025314331,
+      "num_tokens": 182421295.0,
+      "step": 485
+    },
+    {
+      "epoch": 0.7232142857142857,
+      "grad_norm": 2.4069108932289716,
+      "learning_rate": 9.939412971018574e-06,
+      "loss": 0.8138,
+      "mean_token_accuracy": 0.758122444152832,
+      "num_tokens": 182798984.0,
+      "step": 486
+    },
+    {
+      "epoch": 0.7247023809523809,
+      "grad_norm": 2.0737899159169455,
+      "learning_rate": 9.938604114022325e-06,
+      "loss": 0.7812,
+      "mean_token_accuracy": 0.7665718197822571,
+      "num_tokens": 183177473.0,
+      "step": 487
+    },
+    {
+      "epoch": 0.7261904761904762,
+      "grad_norm": 2.203812731959564,
+      "learning_rate": 9.937789926864838e-06,
+      "loss": 0.7967,
+      "mean_token_accuracy": 0.7614662647247314,
+      "num_tokens": 183555239.0,
+      "step": 488
+    },
+    {
+      "epoch": 0.7276785714285714,
+      "grad_norm": 2.3181361255249446,
+      "learning_rate": 9.936970410424857e-06,
+      "loss": 0.7948,
+      "mean_token_accuracy": 0.7587428689002991,
+      "num_tokens": 183932907.0,
+      "step": 489
+    },
+    {
+      "epoch": 0.7291666666666666,
+      "grad_norm": 2.385899248092011,
+      "learning_rate": 9.936145565586871e-06,
+      "loss": 0.8412,
+      "mean_token_accuracy": 0.7606973052024841,
+      "num_tokens": 184311038.0,
+      "step": 490
+    },
+    {
+      "epoch": 0.7306547619047619,
+      "grad_norm": 2.1656300372768182,
+      "learning_rate": 9.935315393241123e-06,
+      "loss": 0.7547,
+      "mean_token_accuracy": 0.7719976902008057,
+      "num_tokens": 184678349.0,
+      "step": 491
+    },
+    {
+      "epoch": 0.7321428571428571,
+      "grad_norm": 2.325819183969089,
+      "learning_rate": 9.934479894283607e-06,
+      "loss": 0.7742,
+      "mean_token_accuracy": 0.7688069939613342,
+      "num_tokens": 185049252.0,
+      "step": 492
+    },
+    {
+      "epoch": 0.7336309523809523,
+      "grad_norm": 2.3246069793626822,
+      "learning_rate": 9.933639069616062e-06,
+      "loss": 0.777,
+      "mean_token_accuracy": 0.7637931108474731,
+      "num_tokens": 185427356.0,
+      "step": 493
+    },
+    {
+      "epoch": 0.7351190476190477,
+      "grad_norm": 2.4858872787611284,
+      "learning_rate": 9.932792920145979e-06,
+      "loss": 0.7609,
+      "mean_token_accuracy": 0.7681053280830383,
+      "num_tokens": 185797668.0,
+      "step": 494
+    },
+    {
+      "epoch": 0.7366071428571429,
+      "grad_norm": 2.167284410846178,
+      "learning_rate": 9.931941446786594e-06,
+      "loss": 0.7931,
+      "mean_token_accuracy": 0.7620806097984314,
+      "num_tokens": 186176980.0,
+      "step": 495
+    },
+    {
+      "epoch": 0.7380952380952381,
+      "grad_norm": 2.0024339921625813,
+      "learning_rate": 9.931084650456892e-06,
+      "loss": 0.7607,
+      "mean_token_accuracy": 0.7740536332130432,
+      "num_tokens": 186557099.0,
+      "step": 496
+    },
+    {
+      "epoch": 0.7395833333333334,
+      "grad_norm": 2.1366906941927297,
+      "learning_rate": 9.930222532081597e-06,
+      "loss": 0.7871,
+      "mean_token_accuracy": 0.7682004570960999,
+      "num_tokens": 186936983.0,
+      "step": 497
+    },
+    {
+      "epoch": 0.7410714285714286,
+      "grad_norm": 2.1229423092303215,
+      "learning_rate": 9.92935509259118e-06,
+      "loss": 0.7597,
+      "mean_token_accuracy": 0.7717771530151367,
+      "num_tokens": 187297671.0,
+      "step": 498
+    },
+    {
+      "epoch": 0.7425595238095238,
+      "grad_norm": 2.2444374062703742,
+      "learning_rate": 9.928482332921859e-06,
+      "loss": 0.8105,
+      "mean_token_accuracy": 0.7554982900619507,
+      "num_tokens": 187677421.0,
+      "step": 499
+    },
+    {
+      "epoch": 0.7440476190476191,
+      "grad_norm": 2.2827974019913557,
+      "learning_rate": 9.927604254015586e-06,
+      "loss": 0.8278,
+      "mean_token_accuracy": 0.758293867111206,
+      "num_tokens": 188055912.0,
+      "step": 500
+    },
+    {
+      "epoch": 0.7455357142857143,
+      "grad_norm": 2.140819097296577,
+      "learning_rate": 9.926720856820061e-06,
+      "loss": 0.761,
+      "mean_token_accuracy": 0.775010347366333,
+      "num_tokens": 188435050.0,
+      "step": 501
+    },
+    {
+      "epoch": 0.7470238095238095,
+      "grad_norm": 2.159764975168133,
+      "learning_rate": 9.925832142288723e-06,
+      "loss": 0.778,
+      "mean_token_accuracy": 0.7647901773452759,
+      "num_tokens": 188813627.0,
+      "step": 502
+    },
+    {
+      "epoch": 0.7485119047619048,
+      "grad_norm": 2.3382782560822712,
+      "learning_rate": 9.924938111380744e-06,
+      "loss": 0.7564,
+      "mean_token_accuracy": 0.7692853212356567,
+      "num_tokens": 189191219.0,
+      "step": 503
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 2.0719475451381837,
+      "learning_rate": 9.924038765061042e-06,
+      "loss": 0.7464,
+      "mean_token_accuracy": 0.7747620344161987,
+      "num_tokens": 189557708.0,
+      "step": 504
+    },
+    {
+      "epoch": 0.7514880952380952,
+      "grad_norm": 2.162319303612145,
+      "learning_rate": 9.923134104300266e-06,
+      "loss": 0.7585,
+      "mean_token_accuracy": 0.7674521207809448,
+      "num_tokens": 189936187.0,
+      "step": 505
+    },
+    {
+      "epoch": 0.7529761904761905,
+      "grad_norm": 2.1134695006032502,
+      "learning_rate": 9.922224130074801e-06,
+      "loss": 0.7092,
+      "mean_token_accuracy": 0.7849007844924927,
+      "num_tokens": 190284133.0,
+      "step": 506
+    },
+    {
+      "epoch": 0.7544642857142857,
+      "grad_norm": 2.1970336885346957,
+      "learning_rate": 9.921308843366773e-06,
+      "loss": 0.8009,
+      "mean_token_accuracy": 0.766215980052948,
+      "num_tokens": 190661373.0,
+      "step": 507
+    },
+    {
+      "epoch": 0.7559523809523809,
+      "grad_norm": 2.1636651479102973,
+      "learning_rate": 9.920388245164033e-06,
+      "loss": 0.7689,
+      "mean_token_accuracy": 0.7714235186576843,
+      "num_tokens": 191039786.0,
+      "step": 508
+    },
+    {
+      "epoch": 0.7574404761904762,
+      "grad_norm": 2.2454182282175297,
+      "learning_rate": 9.919462336460172e-06,
+      "loss": 0.7393,
+      "mean_token_accuracy": 0.7744503617286682,
+      "num_tokens": 191417242.0,
+      "step": 509
+    },
+    {
+      "epoch": 0.7589285714285714,
+      "grad_norm": 2.2785807263660134,
+      "learning_rate": 9.918531118254507e-06,
+      "loss": 0.791,
+      "mean_token_accuracy": 0.7650524973869324,
+      "num_tokens": 191796090.0,
+      "step": 510
+    },
+    {
+      "epoch": 0.7604166666666666,
+      "grad_norm": 2.3143841865838706,
+      "learning_rate": 9.91759459155209e-06,
+      "loss": 0.8094,
+      "mean_token_accuracy": 0.7543814778327942,
+      "num_tokens": 192163249.0,
+      "step": 511
+    },
+    {
+      "epoch": 0.7619047619047619,
+      "grad_norm": 2.235890902591328,
+      "learning_rate": 9.916652757363698e-06,
+      "loss": 0.7848,
+      "mean_token_accuracy": 0.7704679369926453,
+      "num_tokens": 192529954.0,
+      "step": 512
+    },
+    {
+      "epoch": 0.7633928571428571,
+      "grad_norm": 2.113844101382049,
+      "learning_rate": 9.91570561670584e-06,
+      "loss": 0.7825,
+      "mean_token_accuracy": 0.7660157680511475,
+      "num_tokens": 192908911.0,
+      "step": 513
+    },
+    {
+      "epoch": 0.7648809523809523,
+      "grad_norm": 2.0960205561030483,
+      "learning_rate": 9.91475317060075e-06,
+      "loss": 0.7285,
+      "mean_token_accuracy": 0.7831937074661255,
+      "num_tokens": 193287828.0,
+      "step": 514
+    },
+    {
+      "epoch": 0.7663690476190477,
+      "grad_norm": 2.2587628664797035,
+      "learning_rate": 9.91379542007639e-06,
+      "loss": 0.8219,
+      "mean_token_accuracy": 0.7632508873939514,
+      "num_tokens": 193646530.0,
+      "step": 515
+    },
+    {
+      "epoch": 0.7678571428571429,
+      "grad_norm": 2.32401769884902,
+      "learning_rate": 9.912832366166443e-06,
+      "loss": 0.8174,
+      "mean_token_accuracy": 0.7580294013023376,
+      "num_tokens": 194024288.0,
+      "step": 516
+    },
+    {
+      "epoch": 0.7693452380952381,
+      "grad_norm": 2.3948349272697924,
+      "learning_rate": 9.91186400991032e-06,
+      "loss": 0.8439,
+      "mean_token_accuracy": 0.7617377042770386,
+      "num_tokens": 194402154.0,
+      "step": 517
+    },
+    {
+      "epoch": 0.7708333333333334,
+      "grad_norm": 2.3727467895854604,
+      "learning_rate": 9.910890352353154e-06,
+      "loss": 0.8028,
+      "mean_token_accuracy": 0.7669249176979065,
+      "num_tokens": 194779167.0,
+      "step": 518
+    },
+    {
+      "epoch": 0.7723214285714286,
+      "grad_norm": 2.097228968186795,
+      "learning_rate": 9.909911394545799e-06,
+      "loss": 0.7666,
+      "mean_token_accuracy": 0.7719298005104065,
+      "num_tokens": 195145787.0,
+      "step": 519
+    },
+    {
+      "epoch": 0.7738095238095238,
+      "grad_norm": 2.002104506937028,
+      "learning_rate": 9.90892713754483e-06,
+      "loss": 0.6949,
+      "mean_token_accuracy": 0.7835907936096191,
+      "num_tokens": 195525117.0,
+      "step": 520
+    },
+    {
+      "epoch": 0.7752976190476191,
+      "grad_norm": 2.260955940597372,
+      "learning_rate": 9.90793758241254e-06,
+      "loss": 0.8538,
+      "mean_token_accuracy": 0.7533246278762817,
+      "num_tokens": 195904074.0,
+      "step": 521
+    },
+    {
+      "epoch": 0.7767857142857143,
+      "grad_norm": 2.1175335703057026,
+      "learning_rate": 9.90694273021694e-06,
+      "loss": 0.7669,
+      "mean_token_accuracy": 0.7693830132484436,
+      "num_tokens": 196255588.0,
+      "step": 522
+    },
+    {
+      "epoch": 0.7782738095238095,
+      "grad_norm": 2.0901928799781935,
+      "learning_rate": 9.905942582031759e-06,
+      "loss": 0.7559,
+      "mean_token_accuracy": 0.7698140144348145,
+      "num_tokens": 196635435.0,
+      "step": 523
+    },
+    {
+      "epoch": 0.7797619047619048,
+      "grad_norm": 2.1704616335570077,
+      "learning_rate": 9.904937138936443e-06,
+      "loss": 0.7728,
+      "mean_token_accuracy": 0.7663898468017578,
+      "num_tokens": 197013988.0,
+      "step": 524
+    },
+    {
+      "epoch": 0.78125,
+      "grad_norm": 2.115353300712371,
+      "learning_rate": 9.903926402016153e-06,
+      "loss": 0.7828,
+      "mean_token_accuracy": 0.7678216099739075,
+      "num_tokens": 197393208.0,
+      "step": 525
+    },
+    {
+      "epoch": 0.7827380952380952,
+      "grad_norm": 2.24986349613249,
+      "learning_rate": 9.90291037236176e-06,
+      "loss": 0.8443,
+      "mean_token_accuracy": 0.7554673552513123,
+      "num_tokens": 197771611.0,
+      "step": 526
+    },
+    {
+      "epoch": 0.7842261904761905,
+      "grad_norm": 2.451377163190707,
+      "learning_rate": 9.90188905106985e-06,
+      "loss": 0.9001,
+      "mean_token_accuracy": 0.7373397946357727,
+      "num_tokens": 198134497.0,
+      "step": 527
+    },
+    {
+      "epoch": 0.7857142857142857,
+      "grad_norm": 2.4657912577996313,
+      "learning_rate": 9.900862439242719e-06,
+      "loss": 0.8132,
+      "mean_token_accuracy": 0.7685887813568115,
+      "num_tokens": 198511571.0,
+      "step": 528
+    },
+    {
+      "epoch": 0.7872023809523809,
+      "grad_norm": 2.1710832613371878,
+      "learning_rate": 9.899830537988377e-06,
+      "loss": 0.7909,
+      "mean_token_accuracy": 0.759983241558075,
+      "num_tokens": 198878165.0,
+      "step": 529
+    },
+    {
+      "epoch": 0.7886904761904762,
+      "grad_norm": 2.448170586329768,
+      "learning_rate": 9.898793348420535e-06,
+      "loss": 0.8513,
+      "mean_token_accuracy": 0.7608509659767151,
+      "num_tokens": 199254916.0,
+      "step": 530
+    },
+    {
+      "epoch": 0.7901785714285714,
+      "grad_norm": 2.1191718721570094,
+      "learning_rate": 9.89775087165862e-06,
+      "loss": 0.7489,
+      "mean_token_accuracy": 0.775558352470398,
+      "num_tokens": 199635192.0,
+      "step": 531
+    },
+    {
+      "epoch": 0.7916666666666666,
+      "grad_norm": 2.16943502485521,
+      "learning_rate": 9.896703108827758e-06,
+      "loss": 0.7995,
+      "mean_token_accuracy": 0.7631037831306458,
+      "num_tokens": 200013905.0,
+      "step": 532
+    },
+    {
+      "epoch": 0.7931547619047619,
+      "grad_norm": 2.432244714250809,
+      "learning_rate": 9.895650061058787e-06,
+      "loss": 0.8227,
+      "mean_token_accuracy": 0.7554945349693298,
+      "num_tokens": 200391175.0,
+      "step": 533
+    },
+    {
+      "epoch": 0.7946428571428571,
+      "grad_norm": 2.2212903426649255,
+      "learning_rate": 9.894591729488243e-06,
+      "loss": 0.7726,
+      "mean_token_accuracy": 0.7669621706008911,
+      "num_tokens": 200764422.0,
+      "step": 534
+    },
+    {
+      "epoch": 0.7961309523809523,
+      "grad_norm": 2.1665339191834474,
+      "learning_rate": 9.893528115258368e-06,
+      "loss": 0.7579,
+      "mean_token_accuracy": 0.7772925496101379,
+      "num_tokens": 201135874.0,
+      "step": 535
+    },
+    {
+      "epoch": 0.7976190476190477,
+      "grad_norm": 2.3670113277330223,
+      "learning_rate": 9.892459219517108e-06,
+      "loss": 0.8113,
+      "mean_token_accuracy": 0.7661784887313843,
+      "num_tokens": 201512835.0,
+      "step": 536
+    },
+    {
+      "epoch": 0.7991071428571429,
+      "grad_norm": 2.465231998726611,
+      "learning_rate": 9.8913850434181e-06,
+      "loss": 0.7686,
+      "mean_token_accuracy": 0.7664206027984619,
+      "num_tokens": 201890372.0,
+      "step": 537
+    },
+    {
+      "epoch": 0.8005952380952381,
+      "grad_norm": 2.263329767906382,
+      "learning_rate": 9.890305588120692e-06,
+      "loss": 0.7594,
+      "mean_token_accuracy": 0.7709291577339172,
+      "num_tokens": 202268088.0,
+      "step": 538
+    },
+    {
+      "epoch": 0.8020833333333334,
+      "grad_norm": 2.1574662345044042,
+      "learning_rate": 9.88922085478992e-06,
+      "loss": 0.8105,
+      "mean_token_accuracy": 0.7567829489707947,
+      "num_tokens": 202636145.0,
+      "step": 539
+    },
+    {
+      "epoch": 0.8035714285714286,
+      "grad_norm": 2.157389903638308,
+      "learning_rate": 9.888130844596525e-06,
+      "loss": 0.8117,
+      "mean_token_accuracy": 0.7525520920753479,
+      "num_tokens": 203014526.0,
+      "step": 540
+    },
+    {
+      "epoch": 0.8050595238095238,
+      "grad_norm": 2.3225737368226285,
+      "learning_rate": 9.887035558716931e-06,
+      "loss": 0.7586,
+      "mean_token_accuracy": 0.7728502154350281,
+      "num_tokens": 203392505.0,
+      "step": 541
+    },
+    {
+      "epoch": 0.8065476190476191,
+      "grad_norm": 2.2592085183137596,
+      "learning_rate": 9.885934998333272e-06,
+      "loss": 0.7649,
+      "mean_token_accuracy": 0.7741418480873108,
+      "num_tokens": 203771647.0,
+      "step": 542
+    },
+    {
+      "epoch": 0.8080357142857143,
+      "grad_norm": 2.2198022119230956,
+      "learning_rate": 9.884829164633359e-06,
+      "loss": 0.7619,
+      "mean_token_accuracy": 0.7788913249969482,
+      "num_tokens": 204149758.0,
+      "step": 543
+    },
+    {
+      "epoch": 0.8095238095238095,
+      "grad_norm": 2.2409434811684292,
+      "learning_rate": 9.883718058810708e-06,
+      "loss": 0.8535,
+      "mean_token_accuracy": 0.7503941059112549,
+      "num_tokens": 204527853.0,
+      "step": 544
+    },
+    {
+      "epoch": 0.8110119047619048,
+      "grad_norm": 2.53659094218269,
+      "learning_rate": 9.882601682064514e-06,
+      "loss": 0.8119,
+      "mean_token_accuracy": 0.7576950192451477,
+      "num_tokens": 204906146.0,
+      "step": 545
+    },
+    {
+      "epoch": 0.8125,
+      "grad_norm": 2.5924597977749464,
+      "learning_rate": 9.881480035599667e-06,
+      "loss": 0.8358,
+      "mean_token_accuracy": 0.7506717443466187,
+      "num_tokens": 205271553.0,
+      "step": 546
+    },
+    {
+      "epoch": 0.8139880952380952,
+      "grad_norm": 2.4539734784763425,
+      "learning_rate": 9.880353120626746e-06,
+      "loss": 0.8545,
+      "mean_token_accuracy": 0.7589715123176575,
+      "num_tokens": 205648009.0,
+      "step": 547
+    },
+    {
+      "epoch": 0.8154761904761905,
+      "grad_norm": 2.200158733084308,
+      "learning_rate": 9.87922093836201e-06,
+      "loss": 0.8181,
+      "mean_token_accuracy": 0.7702013254165649,
+      "num_tokens": 206025867.0,
+      "step": 548
+    },
+    {
+      "epoch": 0.8169642857142857,
+      "grad_norm": 2.8215665662992855,
+      "learning_rate": 9.878083490027406e-06,
+      "loss": 0.7597,
+      "mean_token_accuracy": 0.7702596783638,
+      "num_tokens": 206404601.0,
+      "step": 549
+    },
+    {
+      "epoch": 0.8184523809523809,
+      "grad_norm": 2.4796769104250296,
+      "learning_rate": 9.876940776850568e-06,
+      "loss": 0.7992,
+      "mean_token_accuracy": 0.7586883306503296,
+      "num_tokens": 206784323.0,
+      "step": 550
+    },
+    {
+      "epoch": 0.8199404761904762,
+      "grad_norm": 2.2865263925889723,
+      "learning_rate": 9.875792800064807e-06,
+      "loss": 0.7916,
+      "mean_token_accuracy": 0.7746691107749939,
+      "num_tokens": 207163668.0,
+      "step": 551
+    },
+    {
+      "epoch": 0.8214285714285714,
+      "grad_norm": 2.5280384907603244,
+      "learning_rate": 9.874639560909118e-06,
+      "loss": 0.8365,
+      "mean_token_accuracy": 0.7559518218040466,
+      "num_tokens": 207540500.0,
+      "step": 552
+    },
+    {
+      "epoch": 0.8229166666666666,
+      "grad_norm": 2.573200246471243,
+      "learning_rate": 9.873481060628175e-06,
+      "loss": 0.8131,
+      "mean_token_accuracy": 0.7575067281723022,
+      "num_tokens": 207919725.0,
+      "step": 553
+    },
+    {
+      "epoch": 0.8244047619047619,
+      "grad_norm": 2.3111765489750944,
+      "learning_rate": 9.87231730047233e-06,
+      "loss": 0.7832,
+      "mean_token_accuracy": 0.763755738735199,
+      "num_tokens": 208298268.0,
+      "step": 554
+    },
+    {
+      "epoch": 0.8258928571428571,
+      "grad_norm": 2.184879213202074,
+      "learning_rate": 9.871148281697609e-06,
+      "loss": 0.7575,
+      "mean_token_accuracy": 0.7666813135147095,
+      "num_tokens": 208666592.0,
+      "step": 555
+    },
+    {
+      "epoch": 0.8273809523809523,
+      "grad_norm": 2.3289515453521332,
+      "learning_rate": 9.869974005565719e-06,
+      "loss": 0.797,
+      "mean_token_accuracy": 0.7621490359306335,
+      "num_tokens": 209031455.0,
+      "step": 556
+    },
+    {
+      "epoch": 0.8288690476190477,
+      "grad_norm": 2.446015284082495,
+      "learning_rate": 9.868794473344039e-06,
+      "loss": 0.8661,
+      "mean_token_accuracy": 0.7558775544166565,
+      "num_tokens": 209408872.0,
+      "step": 557
+    },
+    {
+      "epoch": 0.8303571428571429,
+      "grad_norm": 2.1646738606888163,
+      "learning_rate": 9.867609686305616e-06,
+      "loss": 0.8176,
+      "mean_token_accuracy": 0.7581487894058228,
+      "num_tokens": 209783813.0,
+      "step": 558
+    },
+    {
+      "epoch": 0.8318452380952381,
+      "grad_norm": 2.15928663581817,
+      "learning_rate": 9.86641964572918e-06,
+      "loss": 0.8002,
+      "mean_token_accuracy": 0.7610766291618347,
+      "num_tokens": 210150217.0,
+      "step": 559
+    },
+    {
+      "epoch": 0.8333333333333334,
+      "grad_norm": 2.1231284892135265,
+      "learning_rate": 9.86522435289912e-06,
+      "loss": 0.7411,
+      "mean_token_accuracy": 0.7793722748756409,
+      "num_tokens": 210530111.0,
+      "step": 560
+    },
+    {
+      "epoch": 0.8348214285714286,
+      "grad_norm": 2.2212618325624383,
+      "learning_rate": 9.864023809105497e-06,
+      "loss": 0.7891,
+      "mean_token_accuracy": 0.7690379023551941,
+      "num_tokens": 210907412.0,
+      "step": 561
+    },
+    {
+      "epoch": 0.8363095238095238,
+      "grad_norm": 2.2210340176436727,
+      "learning_rate": 9.862818015644042e-06,
+      "loss": 0.8013,
+      "mean_token_accuracy": 0.75701504945755,
+      "num_tokens": 211286071.0,
+      "step": 562
+    },
+    {
+      "epoch": 0.8377976190476191,
+      "grad_norm": 2.548391240682889,
+      "learning_rate": 9.861606973816146e-06,
+      "loss": 0.7656,
+      "mean_token_accuracy": 0.7738215923309326,
+      "num_tokens": 211663873.0,
+      "step": 563
+    },
+    {
+      "epoch": 0.8392857142857143,
+      "grad_norm": 2.9909901512448545,
+      "learning_rate": 9.860390684928873e-06,
+      "loss": 0.8378,
+      "mean_token_accuracy": 0.749320924282074,
+      "num_tokens": 212040482.0,
+      "step": 564
+    },
+    {
+      "epoch": 0.8407738095238095,
+      "grad_norm": 2.1964228229605376,
+      "learning_rate": 9.859169150294943e-06,
+      "loss": 0.7755,
+      "mean_token_accuracy": 0.7660459280014038,
+      "num_tokens": 212417571.0,
+      "step": 565
+    },
+    {
+      "epoch": 0.8422619047619048,
+      "grad_norm": 2.357263043881747,
+      "learning_rate": 9.85794237123274e-06,
+      "loss": 0.769,
+      "mean_token_accuracy": 0.7643123269081116,
+      "num_tokens": 212796464.0,
+      "step": 566
+    },
+    {
+      "epoch": 0.84375,
+      "grad_norm": 2.441764283191768,
+      "learning_rate": 9.856710349066307e-06,
+      "loss": 0.7856,
+      "mean_token_accuracy": 0.7616475224494934,
+      "num_tokens": 213175976.0,
+      "step": 567
+    },
+    {
+      "epoch": 0.8452380952380952,
+      "grad_norm": 2.3173169410079795,
+      "learning_rate": 9.855473085125351e-06,
+      "loss": 0.8459,
+      "mean_token_accuracy": 0.7502025961875916,
+      "num_tokens": 213554089.0,
+      "step": 568
+    },
+    {
+      "epoch": 0.8467261904761905,
+      "grad_norm": 2.196352383783416,
+      "learning_rate": 9.854230580745228e-06,
+      "loss": 0.7624,
+      "mean_token_accuracy": 0.7761070728302002,
+      "num_tokens": 213932106.0,
+      "step": 569
+    },
+    {
+      "epoch": 0.8482142857142857,
+      "grad_norm": 2.0996234511001655,
+      "learning_rate": 9.852982837266955e-06,
+      "loss": 0.7553,
+      "mean_token_accuracy": 0.7674927115440369,
+      "num_tokens": 214293340.0,
+      "step": 570
+    },
+    {
+      "epoch": 0.8497023809523809,
+      "grad_norm": 2.3110663220605927,
+      "learning_rate": 9.851729856037206e-06,
+      "loss": 0.7842,
+      "mean_token_accuracy": 0.7697689533233643,
+      "num_tokens": 214671305.0,
+      "step": 571
+    },
+    {
+      "epoch": 0.8511904761904762,
+      "grad_norm": 2.2605440319189527,
+      "learning_rate": 9.850471638408301e-06,
+      "loss": 0.8071,
+      "mean_token_accuracy": 0.7625566720962524,
+      "num_tokens": 215049727.0,
+      "step": 572
+    },
+    {
+      "epoch": 0.8526785714285714,
+      "grad_norm": 2.1435850594674872,
+      "learning_rate": 9.849208185738217e-06,
+      "loss": 0.6864,
+      "mean_token_accuracy": 0.788354754447937,
+      "num_tokens": 215427774.0,
+      "step": 573
+    },
+    {
+      "epoch": 0.8541666666666666,
+      "grad_norm": 2.034655320353168,
+      "learning_rate": 9.847939499390581e-06,
+      "loss": 0.773,
+      "mean_token_accuracy": 0.7661179900169373,
+      "num_tokens": 215806305.0,
+      "step": 574
+    },
+    {
+      "epoch": 0.8556547619047619,
+      "grad_norm": 2.289956128756729,
+      "learning_rate": 9.846665580734665e-06,
+      "loss": 0.8261,
+      "mean_token_accuracy": 0.7635886073112488,
+      "num_tokens": 216184509.0,
+      "step": 575
+    },
+    {
+      "epoch": 0.8571428571428571,
+      "grad_norm": 2.297139248002308,
+      "learning_rate": 9.84538643114539e-06,
+      "loss": 0.7249,
+      "mean_token_accuracy": 0.7777121067047119,
+      "num_tokens": 216563691.0,
+      "step": 576
+    },
+    {
+      "epoch": 0.8586309523809523,
+      "grad_norm": 2.140493638736509,
+      "learning_rate": 9.844102052003325e-06,
+      "loss": 0.754,
+      "mean_token_accuracy": 0.7739826440811157,
+      "num_tokens": 216942347.0,
+      "step": 577
+    },
+    {
+      "epoch": 0.8601190476190477,
+      "grad_norm": 2.113093610085358,
+      "learning_rate": 9.84281244469468e-06,
+      "loss": 0.7968,
+      "mean_token_accuracy": 0.7597280740737915,
+      "num_tokens": 217320521.0,
+      "step": 578
+    },
+    {
+      "epoch": 0.8616071428571429,
+      "grad_norm": 2.1494725554191514,
+      "learning_rate": 9.841517610611309e-06,
+      "loss": 0.7589,
+      "mean_token_accuracy": 0.7735381722450256,
+      "num_tokens": 217686740.0,
+      "step": 579
+    },
+    {
+      "epoch": 0.8630952380952381,
+      "grad_norm": 2.4710769984422694,
+      "learning_rate": 9.840217551150706e-06,
+      "loss": 0.7924,
+      "mean_token_accuracy": 0.7744138836860657,
+      "num_tokens": 218065627.0,
+      "step": 580
+    },
+    {
+      "epoch": 0.8645833333333334,
+      "grad_norm": 2.4666982051264044,
+      "learning_rate": 9.838912267716006e-06,
+      "loss": 0.7727,
+      "mean_token_accuracy": 0.7639898061752319,
+      "num_tokens": 218443215.0,
+      "step": 581
+    },
+    {
+      "epoch": 0.8660714285714286,
+      "grad_norm": 2.175636450068978,
+      "learning_rate": 9.837601761715982e-06,
+      "loss": 0.7337,
+      "mean_token_accuracy": 0.7814536690711975,
+      "num_tokens": 218821913.0,
+      "step": 582
+    },
+    {
+      "epoch": 0.8675595238095238,
+      "grad_norm": 2.3654429373057035,
+      "learning_rate": 9.836286034565044e-06,
+      "loss": 0.7833,
+      "mean_token_accuracy": 0.7668706178665161,
+      "num_tokens": 219199278.0,
+      "step": 583
+    },
+    {
+      "epoch": 0.8690476190476191,
+      "grad_norm": 2.5701320061052457,
+      "learning_rate": 9.834965087683237e-06,
+      "loss": 0.86,
+      "mean_token_accuracy": 0.7558269500732422,
+      "num_tokens": 219562262.0,
+      "step": 584
+    },
+    {
+      "epoch": 0.8705357142857143,
+      "grad_norm": 2.248097375891916,
+      "learning_rate": 9.833638922496238e-06,
+      "loss": 0.7801,
+      "mean_token_accuracy": 0.7654374837875366,
+      "num_tokens": 219940900.0,
+      "step": 585
+    },
+    {
+      "epoch": 0.8720238095238095,
+      "grad_norm": 2.3603181659134105,
+      "learning_rate": 9.832307540435357e-06,
+      "loss": 0.801,
+      "mean_token_accuracy": 0.7588149905204773,
+      "num_tokens": 220307580.0,
+      "step": 586
+    },
+    {
+      "epoch": 0.8735119047619048,
+      "grad_norm": 2.5116607805283704,
+      "learning_rate": 9.830970942937534e-06,
+      "loss": 0.8016,
+      "mean_token_accuracy": 0.7587926387786865,
+      "num_tokens": 220669256.0,
+      "step": 587
+    },
+    {
+      "epoch": 0.875,
+      "grad_norm": 2.3018492042823873,
+      "learning_rate": 9.829629131445342e-06,
+      "loss": 0.7831,
+      "mean_token_accuracy": 0.7742421627044678,
+      "num_tokens": 221049003.0,
+      "step": 588
+    },
+    {
+      "epoch": 0.8764880952380952,
+      "grad_norm": 2.4816649806341853,
+      "learning_rate": 9.828282107406974e-06,
+      "loss": 0.826,
+      "mean_token_accuracy": 0.7521160840988159,
+      "num_tokens": 221426310.0,
+      "step": 589
+    },
+    {
+      "epoch": 0.8779761904761905,
+      "grad_norm": 2.529775352062679,
+      "learning_rate": 9.826929872276257e-06,
+      "loss": 0.8119,
+      "mean_token_accuracy": 0.7562154531478882,
+      "num_tokens": 221791369.0,
+      "step": 590
+    },
+    {
+      "epoch": 0.8794642857142857,
+      "grad_norm": 2.1994292885283593,
+      "learning_rate": 9.825572427512632e-06,
+      "loss": 0.8093,
+      "mean_token_accuracy": 0.7577499151229858,
+      "num_tokens": 222170034.0,
+      "step": 591
+    },
+    {
+      "epoch": 0.8809523809523809,
+      "grad_norm": 2.1233586657989068,
+      "learning_rate": 9.824209774581176e-06,
+      "loss": 0.7375,
+      "mean_token_accuracy": 0.778951108455658,
+      "num_tokens": 222542189.0,
+      "step": 592
+    },
+    {
+      "epoch": 0.8824404761904762,
+      "grad_norm": 2.494411887436483,
+      "learning_rate": 9.822841914952575e-06,
+      "loss": 0.7892,
+      "mean_token_accuracy": 0.7623745203018188,
+      "num_tokens": 222920680.0,
+      "step": 593
+    },
+    {
+      "epoch": 0.8839285714285714,
+      "grad_norm": 2.6445538962635906,
+      "learning_rate": 9.82146885010314e-06,
+      "loss": 0.8349,
+      "mean_token_accuracy": 0.7498632073402405,
+      "num_tokens": 223288855.0,
+      "step": 594
+    },
+    {
+      "epoch": 0.8854166666666666,
+      "grad_norm": 2.2582479125061985,
+      "learning_rate": 9.820090581514799e-06,
+      "loss": 0.8482,
+      "mean_token_accuracy": 0.749934732913971,
+      "num_tokens": 223667010.0,
+      "step": 595
+    },
+    {
+      "epoch": 0.8869047619047619,
+      "grad_norm": 2.191358249325834,
+      "learning_rate": 9.818707110675099e-06,
+      "loss": 0.8139,
+      "mean_token_accuracy": 0.7565962076187134,
+      "num_tokens": 224045375.0,
+      "step": 596
+    },
+    {
+      "epoch": 0.8883928571428571,
+      "grad_norm": 2.4937145101367713,
+      "learning_rate": 9.817318439077197e-06,
+      "loss": 0.8027,
+      "mean_token_accuracy": 0.7604925632476807,
+      "num_tokens": 224423926.0,
+      "step": 597
+    },
+    {
+      "epoch": 0.8898809523809523,
+      "grad_norm": 2.2162905885528152,
+      "learning_rate": 9.815924568219865e-06,
+      "loss": 0.7728,
+      "mean_token_accuracy": 0.7695112824440002,
+      "num_tokens": 224801223.0,
+      "step": 598
+    },
+    {
+      "epoch": 0.8913690476190477,
+      "grad_norm": 2.2535466179063652,
+      "learning_rate": 9.814525499607489e-06,
+      "loss": 0.8072,
+      "mean_token_accuracy": 0.7652223110198975,
+      "num_tokens": 225178406.0,
+      "step": 599
+    },
+    {
+      "epoch": 0.8928571428571429,
+      "grad_norm": 2.389752359546262,
+      "learning_rate": 9.81312123475006e-06,
+      "loss": 0.8265,
+      "mean_token_accuracy": 0.7548670172691345,
+      "num_tokens": 225556286.0,
+      "step": 600
+    },
+    {
+      "epoch": 0.8943452380952381,
+      "grad_norm": 2.179587252697458,
+      "learning_rate": 9.811711775163184e-06,
+      "loss": 0.8123,
+      "mean_token_accuracy": 0.7531967163085938,
+      "num_tokens": 225934217.0,
+      "step": 601
+    },
+    {
+      "epoch": 0.8958333333333334,
+      "grad_norm": 2.403049222029006,
+      "learning_rate": 9.810297122368067e-06,
+      "loss": 0.87,
+      "mean_token_accuracy": 0.7445114850997925,
+      "num_tokens": 226312219.0,
+      "step": 602
+    },
+    {
+      "epoch": 0.8973214285714286,
+      "grad_norm": 2.1444716624887565,
+      "learning_rate": 9.808877277891522e-06,
+      "loss": 0.7919,
+      "mean_token_accuracy": 0.7663387060165405,
+      "num_tokens": 226691943.0,
+      "step": 603
+    },
+    {
+      "epoch": 0.8988095238095238,
+      "grad_norm": 2.3897589533579238,
+      "learning_rate": 9.807452243265968e-06,
+      "loss": 0.8029,
+      "mean_token_accuracy": 0.7620751261711121,
+      "num_tokens": 227069140.0,
+      "step": 604
+    },
+    {
+      "epoch": 0.9002976190476191,
+      "grad_norm": 2.1794502732353327,
+      "learning_rate": 9.80602202002942e-06,
+      "loss": 0.7511,
+      "mean_token_accuracy": 0.7754650115966797,
+      "num_tokens": 227447221.0,
+      "step": 605
+    },
+    {
+      "epoch": 0.9017857142857143,
+      "grad_norm": 2.2940210633364635,
+      "learning_rate": 9.804586609725499e-06,
+      "loss": 0.7995,
+      "mean_token_accuracy": 0.7548350691795349,
+      "num_tokens": 227824553.0,
+      "step": 606
+    },
+    {
+      "epoch": 0.9032738095238095,
+      "grad_norm": 2.0929498141009226,
+      "learning_rate": 9.803146013903421e-06,
+      "loss": 0.7524,
+      "mean_token_accuracy": 0.7762200236320496,
+      "num_tokens": 228202254.0,
+      "step": 607
+    },
+    {
+      "epoch": 0.9047619047619048,
+      "grad_norm": 2.0903155288427597,
+      "learning_rate": 9.801700234118e-06,
+      "loss": 0.7619,
+      "mean_token_accuracy": 0.7743849754333496,
+      "num_tokens": 228580860.0,
+      "step": 608
+    },
+    {
+      "epoch": 0.90625,
+      "grad_norm": 2.220664820860115,
+      "learning_rate": 9.800249271929645e-06,
+      "loss": 0.7967,
+      "mean_token_accuracy": 0.7644937038421631,
+      "num_tokens": 228959177.0,
+      "step": 609
+    },
+    {
+      "epoch": 0.9077380952380952,
+      "grad_norm": 1.9762299692602587,
+      "learning_rate": 9.798793128904356e-06,
+      "loss": 0.796,
+      "mean_token_accuracy": 0.7640405893325806,
+      "num_tokens": 229338944.0,
+      "step": 610
+    },
+    {
+      "epoch": 0.9092261904761905,
+      "grad_norm": 1.932698938334983,
+      "learning_rate": 9.797331806613726e-06,
+      "loss": 0.7641,
+      "mean_token_accuracy": 0.7716626524925232,
+      "num_tokens": 229719600.0,
+      "step": 611
+    },
+    {
+      "epoch": 0.9107142857142857,
+      "grad_norm": 2.2535349332778383,
+      "learning_rate": 9.795865306634939e-06,
+      "loss": 0.827,
+      "mean_token_accuracy": 0.7575547695159912,
+      "num_tokens": 230097873.0,
+      "step": 612
+    },
+    {
+      "epoch": 0.9122023809523809,
+      "grad_norm": 2.5446237338669366,
+      "learning_rate": 9.794393630550769e-06,
+      "loss": 0.7913,
+      "mean_token_accuracy": 0.7666345238685608,
+      "num_tokens": 230468344.0,
+      "step": 613
+    },
+    {
+      "epoch": 0.9136904761904762,
+      "grad_norm": 2.062047757279412,
+      "learning_rate": 9.79291677994957e-06,
+      "loss": 0.7442,
+      "mean_token_accuracy": 0.7749908566474915,
+      "num_tokens": 230846446.0,
+      "step": 614
+    },
+    {
+      "epoch": 0.9151785714285714,
+      "grad_norm": 2.032862938253999,
+      "learning_rate": 9.791434756425288e-06,
+      "loss": 0.7487,
+      "mean_token_accuracy": 0.7740064263343811,
+      "num_tokens": 231226066.0,
+      "step": 615
+    },
+    {
+      "epoch": 0.9166666666666666,
+      "grad_norm": 2.1138260091365333,
+      "learning_rate": 9.789947561577445e-06,
+      "loss": 0.7794,
+      "mean_token_accuracy": 0.765358030796051,
+      "num_tokens": 231604897.0,
+      "step": 616
+    },
+    {
+      "epoch": 0.9181547619047619,
+      "grad_norm": 2.181070392234823,
+      "learning_rate": 9.78845519701115e-06,
+      "loss": 0.7337,
+      "mean_token_accuracy": 0.7772727012634277,
+      "num_tokens": 231982449.0,
+      "step": 617
+    },
+    {
+      "epoch": 0.9196428571428571,
+      "grad_norm": 2.295953515853044,
+      "learning_rate": 9.78695766433709e-06,
+      "loss": 0.8415,
+      "mean_token_accuracy": 0.7485480904579163,
+      "num_tokens": 232360183.0,
+      "step": 618
+    },
+    {
+      "epoch": 0.9211309523809523,
+      "grad_norm": 2.3967286711501026,
+      "learning_rate": 9.78545496517153e-06,
+      "loss": 0.822,
+      "mean_token_accuracy": 0.7575286626815796,
+      "num_tokens": 232738191.0,
+      "step": 619
+    },
+    {
+      "epoch": 0.9226190476190477,
+      "grad_norm": 2.203486799273358,
+      "learning_rate": 9.783947101136311e-06,
+      "loss": 0.7706,
+      "mean_token_accuracy": 0.7731975317001343,
+      "num_tokens": 233117762.0,
+      "step": 620
+    },
+    {
+      "epoch": 0.9241071428571429,
+      "grad_norm": 2.201565963557121,
+      "learning_rate": 9.782434073858845e-06,
+      "loss": 0.8332,
+      "mean_token_accuracy": 0.7554592490196228,
+      "num_tokens": 233496139.0,
+      "step": 621
+    },
+    {
+      "epoch": 0.9255952380952381,
+      "grad_norm": 2.2404983184485423,
+      "learning_rate": 9.780915884972121e-06,
+      "loss": 0.8262,
+      "mean_token_accuracy": 0.7495020031929016,
+      "num_tokens": 233875630.0,
+      "step": 622
+    },
+    {
+      "epoch": 0.9270833333333334,
+      "grad_norm": 2.171156143647477,
+      "learning_rate": 9.779392536114698e-06,
+      "loss": 0.847,
+      "mean_token_accuracy": 0.7478122711181641,
+      "num_tokens": 234243477.0,
+      "step": 623
+    },
+    {
+      "epoch": 0.9285714285714286,
+      "grad_norm": 2.3973201527773322,
+      "learning_rate": 9.777864028930705e-06,
+      "loss": 0.7861,
+      "mean_token_accuracy": 0.7641556859016418,
+      "num_tokens": 234619628.0,
+      "step": 624
+    },
+    {
+      "epoch": 0.9300595238095238,
+      "grad_norm": 2.4379191341330326,
+      "learning_rate": 9.776330365069833e-06,
+      "loss": 0.8167,
+      "mean_token_accuracy": 0.7569529414176941,
+      "num_tokens": 234997707.0,
+      "step": 625
+    },
+    {
+      "epoch": 0.9315476190476191,
+      "grad_norm": 2.0149402527109084,
+      "learning_rate": 9.774791546187346e-06,
+      "loss": 0.7564,
+      "mean_token_accuracy": 0.771775484085083,
+      "num_tokens": 235377086.0,
+      "step": 626
+    },
+    {
+      "epoch": 0.9330357142857143,
+      "grad_norm": 2.7902241707594446,
+      "learning_rate": 9.773247573944067e-06,
+      "loss": 0.8587,
+      "mean_token_accuracy": 0.7430628538131714,
+      "num_tokens": 235733790.0,
+      "step": 627
+    },
+    {
+      "epoch": 0.9345238095238095,
+      "grad_norm": 2.5408868449781026,
+      "learning_rate": 9.771698450006382e-06,
+      "loss": 0.8609,
+      "mean_token_accuracy": 0.7483543157577515,
+      "num_tokens": 236111242.0,
+      "step": 628
+    },
+    {
+      "epoch": 0.9360119047619048,
+      "grad_norm": 2.178221370952544,
+      "learning_rate": 9.770144176046237e-06,
+      "loss": 0.7326,
+      "mean_token_accuracy": 0.7789002060890198,
+      "num_tokens": 236477722.0,
+      "step": 629
+    },
+    {
+      "epoch": 0.9375,
+      "grad_norm": 2.322348264474308,
+      "learning_rate": 9.768584753741134e-06,
+      "loss": 0.7792,
+      "mean_token_accuracy": 0.7700604200363159,
+      "num_tokens": 236855552.0,
+      "step": 630
+    },
+    {
+      "epoch": 0.9389880952380952,
+      "grad_norm": 2.2522323809790805,
+      "learning_rate": 9.76702018477414e-06,
+      "loss": 0.7778,
+      "mean_token_accuracy": 0.7665531635284424,
+      "num_tokens": 237230756.0,
+      "step": 631
+    },
+    {
+      "epoch": 0.9404761904761905,
+      "grad_norm": 2.466690934595837,
+      "learning_rate": 9.765450470833867e-06,
+      "loss": 0.7671,
+      "mean_token_accuracy": 0.7687544822692871,
+      "num_tokens": 237608953.0,
+      "step": 632
+    },
+    {
+      "epoch": 0.9419642857142857,
+      "grad_norm": 2.2816535746013282,
+      "learning_rate": 9.763875613614482e-06,
+      "loss": 0.7968,
+      "mean_token_accuracy": 0.7673419117927551,
+      "num_tokens": 237986605.0,
+      "step": 633
+    },
+    {
+      "epoch": 0.9434523809523809,
+      "grad_norm": 2.592636373212804,
+      "learning_rate": 9.762295614815708e-06,
+      "loss": 0.8255,
+      "mean_token_accuracy": 0.7615466117858887,
+      "num_tokens": 238363735.0,
+      "step": 634
+    },
+    {
+      "epoch": 0.9449404761904762,
+      "grad_norm": 2.308252081486983,
+      "learning_rate": 9.760710476142812e-06,
+      "loss": 0.7777,
+      "mean_token_accuracy": 0.7670193910598755,
+      "num_tokens": 238732588.0,
+      "step": 635
+    },
+    {
+      "epoch": 0.9464285714285714,
+      "grad_norm": 2.2831048229253765,
+      "learning_rate": 9.759120199306613e-06,
+      "loss": 0.77,
+      "mean_token_accuracy": 0.7722880244255066,
+      "num_tokens": 239111482.0,
+      "step": 636
+    },
+    {
+      "epoch": 0.9479166666666666,
+      "grad_norm": 2.2553789920896463,
+      "learning_rate": 9.75752478602347e-06,
+      "loss": 0.748,
+      "mean_token_accuracy": 0.7804667949676514,
+      "num_tokens": 239490253.0,
+      "step": 637
+    },
+    {
+      "epoch": 0.9494047619047619,
+      "grad_norm": 2.1343795651181474,
+      "learning_rate": 9.755924238015289e-06,
+      "loss": 0.7994,
+      "mean_token_accuracy": 0.7633252739906311,
+      "num_tokens": 239867327.0,
+      "step": 638
+    },
+    {
+      "epoch": 0.9508928571428571,
+      "grad_norm": 2.0774360100400826,
+      "learning_rate": 9.754318557009518e-06,
+      "loss": 0.7184,
+      "mean_token_accuracy": 0.7795575261116028,
+      "num_tokens": 240245954.0,
+      "step": 639
+    },
+    {
+      "epoch": 0.9523809523809523,
+      "grad_norm": 2.2252512591908933,
+      "learning_rate": 9.752707744739146e-06,
+      "loss": 0.8409,
+      "mean_token_accuracy": 0.7465044856071472,
+      "num_tokens": 240612097.0,
+      "step": 640
+    },
+    {
+      "epoch": 0.9538690476190477,
+      "grad_norm": 2.4787922418138586,
+      "learning_rate": 9.751091802942696e-06,
+      "loss": 0.8624,
+      "mean_token_accuracy": 0.7462185025215149,
+      "num_tokens": 240990130.0,
+      "step": 641
+    },
+    {
+      "epoch": 0.9553571428571429,
+      "grad_norm": 2.0505609986684865,
+      "learning_rate": 9.749470733364231e-06,
+      "loss": 0.7474,
+      "mean_token_accuracy": 0.7711857557296753,
+      "num_tokens": 241368931.0,
+      "step": 642
+    },
+    {
+      "epoch": 0.9568452380952381,
+      "grad_norm": 2.392735373877206,
+      "learning_rate": 9.747844537753346e-06,
+      "loss": 0.7737,
+      "mean_token_accuracy": 0.767175555229187,
+      "num_tokens": 241746558.0,
+      "step": 643
+    },
+    {
+      "epoch": 0.9583333333333334,
+      "grad_norm": 2.2855891099542953,
+      "learning_rate": 9.74621321786517e-06,
+      "loss": 0.8447,
+      "mean_token_accuracy": 0.7434188723564148,
+      "num_tokens": 242125577.0,
+      "step": 644
+    },
+    {
+      "epoch": 0.9598214285714286,
+      "grad_norm": 2.293740884186301,
+      "learning_rate": 9.744576775460363e-06,
+      "loss": 0.8027,
+      "mean_token_accuracy": 0.7599866390228271,
+      "num_tokens": 242492669.0,
+      "step": 645
+    },
+    {
+      "epoch": 0.9613095238095238,
+      "grad_norm": 2.5172562548448516,
+      "learning_rate": 9.742935212305112e-06,
+      "loss": 0.8399,
+      "mean_token_accuracy": 0.7582898736000061,
+      "num_tokens": 242856273.0,
+      "step": 646
+    },
+    {
+      "epoch": 0.9627976190476191,
+      "grad_norm": 2.303068999081195,
+      "learning_rate": 9.741288530171132e-06,
+      "loss": 0.8593,
+      "mean_token_accuracy": 0.7454009056091309,
+      "num_tokens": 243219453.0,
+      "step": 647
+    },
+    {
+      "epoch": 0.9642857142857143,
+      "grad_norm": 2.1483895048949284,
+      "learning_rate": 9.73963673083566e-06,
+      "loss": 0.7895,
+      "mean_token_accuracy": 0.7664793729782104,
+      "num_tokens": 243599183.0,
+      "step": 648
+    },
+    {
+      "epoch": 0.9657738095238095,
+      "grad_norm": 2.364473178990349,
+      "learning_rate": 9.73797981608146e-06,
+      "loss": 0.8658,
+      "mean_token_accuracy": 0.7494881749153137,
+      "num_tokens": 243976906.0,
+      "step": 649
+    },
+    {
+      "epoch": 0.9672619047619048,
+      "grad_norm": 2.1943720165969545,
+      "learning_rate": 9.736317787696816e-06,
+      "loss": 0.7569,
+      "mean_token_accuracy": 0.7704520225524902,
+      "num_tokens": 244353908.0,
+      "step": 650
+    },
+    {
+      "epoch": 0.96875,
+      "grad_norm": 2.226830296656961,
+      "learning_rate": 9.73465064747553e-06,
+      "loss": 0.7534,
+      "mean_token_accuracy": 0.7790178656578064,
+      "num_tokens": 244731432.0,
+      "step": 651
+    },
+    {
+      "epoch": 0.9702380952380952,
+      "grad_norm": 2.110416615468736,
+      "learning_rate": 9.732978397216918e-06,
+      "loss": 0.7681,
+      "mean_token_accuracy": 0.7740501761436462,
+      "num_tokens": 245109945.0,
+      "step": 652
+    },
+    {
+      "epoch": 0.9717261904761905,
+      "grad_norm": 2.0527398623522264,
+      "learning_rate": 9.73130103872582e-06,
+      "loss": 0.8001,
+      "mean_token_accuracy": 0.7634296417236328,
+      "num_tokens": 245487863.0,
+      "step": 653
+    },
+    {
+      "epoch": 0.9732142857142857,
+      "grad_norm": 2.1417338422402685,
+      "learning_rate": 9.72961857381258e-06,
+      "loss": 0.7915,
+      "mean_token_accuracy": 0.7581797242164612,
+      "num_tokens": 245866876.0,
+      "step": 654
+    },
+    {
+      "epoch": 0.9747023809523809,
+      "grad_norm": 2.2245265476455343,
+      "learning_rate": 9.727931004293058e-06,
+      "loss": 0.7211,
+      "mean_token_accuracy": 0.7869523763656616,
+      "num_tokens": 246245628.0,
+      "step": 655
+    },
+    {
+      "epoch": 0.9761904761904762,
+      "grad_norm": 2.027740349713491,
+      "learning_rate": 9.726238331988625e-06,
+      "loss": 0.7546,
+      "mean_token_accuracy": 0.7768011093139648,
+      "num_tokens": 246623995.0,
+      "step": 656
+    },
+    {
+      "epoch": 0.9776785714285714,
+      "grad_norm": 1.9957445333993171,
+      "learning_rate": 9.724540558726152e-06,
+      "loss": 0.7565,
+      "mean_token_accuracy": 0.7732576727867126,
+      "num_tokens": 247004042.0,
+      "step": 657
+    },
+    {
+      "epoch": 0.9791666666666666,
+      "grad_norm": 2.0331534061566123,
+      "learning_rate": 9.722837686338025e-06,
+      "loss": 0.7363,
+      "mean_token_accuracy": 0.7787255644798279,
+      "num_tokens": 247371134.0,
+      "step": 658
+    },
+    {
+      "epoch": 0.9806547619047619,
+      "grad_norm": 2.19598876743861,
+      "learning_rate": 9.721129716662124e-06,
+      "loss": 0.7919,
+      "mean_token_accuracy": 0.7679522633552551,
+      "num_tokens": 247749400.0,
+      "step": 659
+    },
+    {
+      "epoch": 0.9821428571428571,
+      "grad_norm": 2.403576154561992,
+      "learning_rate": 9.719416651541839e-06,
+      "loss": 0.7889,
+      "mean_token_accuracy": 0.7591047286987305,
+      "num_tokens": 248126174.0,
+      "step": 660
+    },
+    {
+      "epoch": 0.9836309523809523,
+      "grad_norm": 2.049900528079302,
+      "learning_rate": 9.717698492826054e-06,
+      "loss": 0.7749,
+      "mean_token_accuracy": 0.7674819231033325,
+      "num_tokens": 248504946.0,
+      "step": 661
+    },
+    {
+      "epoch": 0.9851190476190477,
+      "grad_norm": 2.386565119376524,
+      "learning_rate": 9.715975242369148e-06,
+      "loss": 0.8497,
+      "mean_token_accuracy": 0.7467985153198242,
+      "num_tokens": 248881810.0,
+      "step": 662
+    },
+    {
+      "epoch": 0.9866071428571429,
+      "grad_norm": 2.1763010992624774,
+      "learning_rate": 9.714246902031007e-06,
+      "loss": 0.7861,
+      "mean_token_accuracy": 0.7671101689338684,
+      "num_tokens": 249260220.0,
+      "step": 663
+    },
+    {
+      "epoch": 0.9880952380952381,
+      "grad_norm": 2.241573271808781,
+      "learning_rate": 9.712513473676997e-06,
+      "loss": 0.8219,
+      "mean_token_accuracy": 0.7587485313415527,
+      "num_tokens": 249637789.0,
+      "step": 664
+    },
+    {
+      "epoch": 0.9895833333333334,
+      "grad_norm": 2.0296195215683612,
+      "learning_rate": 9.710774959177983e-06,
+      "loss": 0.7696,
+      "mean_token_accuracy": 0.7752209901809692,
+      "num_tokens": 250017394.0,
+      "step": 665
+    },
+    {
+      "epoch": 0.9910714285714286,
+      "grad_norm": 2.295136278260645,
+      "learning_rate": 9.709031360410318e-06,
+      "loss": 0.7521,
+      "mean_token_accuracy": 0.7728481888771057,
+      "num_tokens": 250396794.0,
+      "step": 666
+    },
+    {
+      "epoch": 0.9925595238095238,
+      "grad_norm": 2.078296441856573,
+      "learning_rate": 9.707282679255841e-06,
+      "loss": 0.7292,
+      "mean_token_accuracy": 0.7777870893478394,
+      "num_tokens": 250775342.0,
+      "step": 667
+    },
+    {
+      "epoch": 0.9940476190476191,
+      "grad_norm": 2.143515010869908,
+      "learning_rate": 9.705528917601878e-06,
+      "loss": 0.7842,
+      "mean_token_accuracy": 0.7730729579925537,
+      "num_tokens": 251154089.0,
+      "step": 668
+    },
+    {
+      "epoch": 0.9955357142857143,
+      "grad_norm": 2.2115181893681273,
+      "learning_rate": 9.703770077341236e-06,
+      "loss": 0.7835,
+      "mean_token_accuracy": 0.7684409618377686,
+      "num_tokens": 251531432.0,
+      "step": 669
+    },
+    {
+      "epoch": 0.9970238095238095,
+      "grad_norm": 2.2880602667930017,
+      "learning_rate": 9.70200616037221e-06,
+      "loss": 0.7558,
+      "mean_token_accuracy": 0.7742716073989868,
+      "num_tokens": 251909485.0,
+      "step": 670
+    },
+    {
+      "epoch": 0.9985119047619048,
+      "grad_norm": 2.20220348400776,
+      "learning_rate": 9.700237168598563e-06,
+      "loss": 0.8148,
+      "mean_token_accuracy": 0.7532491087913513,
+      "num_tokens": 252287753.0,
+      "step": 671
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.172579766441655,
+      "learning_rate": 9.698463103929542e-06,
+      "loss": 0.8359,
+      "mean_token_accuracy": 0.7570339441299438,
+      "num_tokens": 252650780.0,
+      "step": 672
+    },
+    {
+      "epoch": 1.0014880952380953,
+      "grad_norm": 2.362105837880461,
+      "learning_rate": 9.696683968279873e-06,
+      "loss": 0.6005,
+      "mean_token_accuracy": 0.8042909502983093,
+      "num_tokens": 253019642.0,
+      "step": 673
+    },
+    {
+      "epoch": 1.0029761904761905,
+      "grad_norm": 1.8036153722928137,
+      "learning_rate": 9.694899763569744e-06,
+      "loss": 0.5543,
+      "mean_token_accuracy": 0.8223846554756165,
+      "num_tokens": 253399164.0,
+      "step": 674
+    },
+    {
+      "epoch": 1.0044642857142858,
+      "grad_norm": 2.0759425628930077,
+      "learning_rate": 9.693110491724826e-06,
+      "loss": 0.573,
+      "mean_token_accuracy": 0.8190493583679199,
+      "num_tokens": 253775754.0,
+      "step": 675
+    },
+    {
+      "epoch": 1.005952380952381,
+      "grad_norm": 2.3294588647903613,
+      "learning_rate": 9.691316154676251e-06,
+      "loss": 0.6714,
+      "mean_token_accuracy": 0.8029769062995911,
+      "num_tokens": 254153706.0,
+      "step": 676
+    },
+    {
+      "epoch": 1.0074404761904763,
+      "grad_norm": 1.99629871192852,
+      "learning_rate": 9.689516754360619e-06,
+      "loss": 0.5348,
+      "mean_token_accuracy": 0.8280665874481201,
+      "num_tokens": 254531929.0,
+      "step": 677
+    },
+    {
+      "epoch": 1.0089285714285714,
+      "grad_norm": 2.009270870966838,
+      "learning_rate": 9.687712292719997e-06,
+      "loss": 0.549,
+      "mean_token_accuracy": 0.8248860836029053,
+      "num_tokens": 254909189.0,
+      "step": 678
+    },
+    {
+      "epoch": 1.0104166666666667,
+      "grad_norm": 2.2964113232301058,
+      "learning_rate": 9.685902771701913e-06,
+      "loss": 0.672,
+      "mean_token_accuracy": 0.7939180731773376,
+      "num_tokens": 255280680.0,
+      "step": 679
+    },
+    {
+      "epoch": 1.0119047619047619,
+      "grad_norm": 1.960998086756381,
+      "learning_rate": 9.684088193259356e-06,
+      "loss": 0.5138,
+      "mean_token_accuracy": 0.8355591297149658,
+      "num_tokens": 255658962.0,
+      "step": 680
+    },
+    {
+      "epoch": 1.0133928571428572,
+      "grad_norm": 2.1928956485965565,
+      "learning_rate": 9.682268559350771e-06,
+      "loss": 0.5549,
+      "mean_token_accuracy": 0.8256645798683167,
+      "num_tokens": 256024805.0,
+      "step": 681
+    },
+    {
+      "epoch": 1.0148809523809523,
+      "grad_norm": 2.083638552768012,
+      "learning_rate": 9.680443871940067e-06,
+      "loss": 0.5629,
+      "mean_token_accuracy": 0.81871497631073,
+      "num_tokens": 256391842.0,
+      "step": 682
+    },
+    {
+      "epoch": 1.0163690476190477,
+      "grad_norm": 2.2812926499626265,
+      "learning_rate": 9.678614132996596e-06,
+      "loss": 0.6564,
+      "mean_token_accuracy": 0.7929259538650513,
+      "num_tokens": 256771013.0,
+      "step": 683
+    },
+    {
+      "epoch": 1.0178571428571428,
+      "grad_norm": 2.234353343269859,
+      "learning_rate": 9.67677934449517e-06,
+      "loss": 0.6935,
+      "mean_token_accuracy": 0.7870613932609558,
+      "num_tokens": 257150492.0,
+      "step": 684
+    },
+    {
+      "epoch": 1.0193452380952381,
+      "grad_norm": 2.118374844636065,
+      "learning_rate": 9.67493950841605e-06,
+      "loss": 0.5266,
+      "mean_token_accuracy": 0.8308270573616028,
+      "num_tokens": 257526390.0,
+      "step": 685
+    },
+    {
+      "epoch": 1.0208333333333333,
+      "grad_norm": 2.1484381123235026,
+      "learning_rate": 9.673094626744944e-06,
+      "loss": 0.5836,
+      "mean_token_accuracy": 0.8156471252441406,
+      "num_tokens": 257892623.0,
+      "step": 686
+    },
+    {
+      "epoch": 1.0223214285714286,
+      "grad_norm": 2.0979480141371094,
+      "learning_rate": 9.671244701472999e-06,
+      "loss": 0.606,
+      "mean_token_accuracy": 0.815895140171051,
+      "num_tokens": 258271472.0,
+      "step": 687
+    },
+    {
+      "epoch": 1.0238095238095237,
+      "grad_norm": 2.3050001686081254,
+      "learning_rate": 9.669389734596819e-06,
+      "loss": 0.5318,
+      "mean_token_accuracy": 0.8284277319908142,
+      "num_tokens": 258649841.0,
+      "step": 688
+    },
+    {
+      "epoch": 1.025297619047619,
+      "grad_norm": 2.5088216949969615,
+      "learning_rate": 9.667529728118439e-06,
+      "loss": 0.5636,
+      "mean_token_accuracy": 0.8241778016090393,
+      "num_tokens": 259027938.0,
+      "step": 689
+    },
+    {
+      "epoch": 1.0267857142857142,
+      "grad_norm": 2.166114391335827,
+      "learning_rate": 9.665664684045332e-06,
+      "loss": 0.5536,
+      "mean_token_accuracy": 0.8183578848838806,
+      "num_tokens": 259405112.0,
+      "step": 690
+    },
+    {
+      "epoch": 1.0282738095238095,
+      "grad_norm": 1.9880412071612021,
+      "learning_rate": 9.66379460439042e-06,
+      "loss": 0.5413,
+      "mean_token_accuracy": 0.8249602317810059,
+      "num_tokens": 259773859.0,
+      "step": 691
+    },
+    {
+      "epoch": 1.0297619047619047,
+      "grad_norm": 2.0617094467296,
+      "learning_rate": 9.661919491172046e-06,
+      "loss": 0.5304,
+      "mean_token_accuracy": 0.8231899738311768,
+      "num_tokens": 260151591.0,
+      "step": 692
+    },
+    {
+      "epoch": 1.03125,
+      "grad_norm": 2.2049262837212744,
+      "learning_rate": 9.660039346413994e-06,
+      "loss": 0.5822,
+      "mean_token_accuracy": 0.808900773525238,
+      "num_tokens": 260529843.0,
+      "step": 693
+    },
+    {
+      "epoch": 1.0327380952380953,
+      "grad_norm": 2.2813260079490933,
+      "learning_rate": 9.658154172145475e-06,
+      "loss": 0.6075,
+      "mean_token_accuracy": 0.8094258904457092,
+      "num_tokens": 260896675.0,
+      "step": 694
+    },
+    {
+      "epoch": 1.0342261904761905,
+      "grad_norm": 2.17796695567597,
+      "learning_rate": 9.656263970401127e-06,
+      "loss": 0.5715,
+      "mean_token_accuracy": 0.8178147673606873,
+      "num_tokens": 261276042.0,
+      "step": 695
+    },
+    {
+      "epoch": 1.0357142857142858,
+      "grad_norm": 2.047802306971016,
+      "learning_rate": 9.654368743221022e-06,
+      "loss": 0.5489,
+      "mean_token_accuracy": 0.8245279788970947,
+      "num_tokens": 261655098.0,
+      "step": 696
+    },
+    {
+      "epoch": 1.037202380952381,
+      "grad_norm": 1.9336977062709475,
+      "learning_rate": 9.652468492650646e-06,
+      "loss": 0.5409,
+      "mean_token_accuracy": 0.8314597010612488,
+      "num_tokens": 262034145.0,
+      "step": 697
+    },
+    {
+      "epoch": 1.0386904761904763,
+      "grad_norm": 2.051520542839783,
+      "learning_rate": 9.650563220740912e-06,
+      "loss": 0.5684,
+      "mean_token_accuracy": 0.823256254196167,
+      "num_tokens": 262400365.0,
+      "step": 698
+    },
+    {
+      "epoch": 1.0401785714285714,
+      "grad_norm": 1.9488931958499036,
+      "learning_rate": 9.648652929548153e-06,
+      "loss": 0.5544,
+      "mean_token_accuracy": 0.8204989433288574,
+      "num_tokens": 262777513.0,
+      "step": 699
+    },
+    {
+      "epoch": 1.0416666666666667,
+      "grad_norm": 2.0552795743044507,
+      "learning_rate": 9.646737621134112e-06,
+      "loss": 0.5748,
+      "mean_token_accuracy": 0.8185328245162964,
+      "num_tokens": 263155929.0,
+      "step": 700
+    },
+    {
+      "epoch": 1.0431547619047619,
+      "grad_norm": 2.206871191768218,
+      "learning_rate": 9.64481729756596e-06,
+      "loss": 0.6385,
+      "mean_token_accuracy": 0.7984189987182617,
+      "num_tokens": 263534874.0,
+      "step": 701
+    },
+    {
+      "epoch": 1.0446428571428572,
+      "grad_norm": 2.237237030977722,
+      "learning_rate": 9.642891960916269e-06,
+      "loss": 0.5388,
+      "mean_token_accuracy": 0.8270334005355835,
+      "num_tokens": 263910924.0,
+      "step": 702
+    },
+    {
+      "epoch": 1.0461309523809523,
+      "grad_norm": 2.1255625657761166,
+      "learning_rate": 9.640961613263028e-06,
+      "loss": 0.5883,
+      "mean_token_accuracy": 0.8139474987983704,
+      "num_tokens": 264289347.0,
+      "step": 703
+    },
+    {
+      "epoch": 1.0476190476190477,
+      "grad_norm": 2.0287953867951227,
+      "learning_rate": 9.639026256689628e-06,
+      "loss": 0.6088,
+      "mean_token_accuracy": 0.8056619763374329,
+      "num_tokens": 264661002.0,
+      "step": 704
+    },
+    {
+      "epoch": 1.0491071428571428,
+      "grad_norm": 2.18000852333234,
+      "learning_rate": 9.637085893284875e-06,
+      "loss": 0.5242,
+      "mean_token_accuracy": 0.8287292718887329,
+      "num_tokens": 265037337.0,
+      "step": 705
+    },
+    {
+      "epoch": 1.0505952380952381,
+      "grad_norm": 2.454057509790487,
+      "learning_rate": 9.635140525142971e-06,
+      "loss": 0.6083,
+      "mean_token_accuracy": 0.8083675503730774,
+      "num_tokens": 265414207.0,
+      "step": 706
+    },
+    {
+      "epoch": 1.0520833333333333,
+      "grad_norm": 2.143921229520474,
+      "learning_rate": 9.633190154363527e-06,
+      "loss": 0.603,
+      "mean_token_accuracy": 0.8081603646278381,
+      "num_tokens": 265780196.0,
+      "step": 707
+    },
+    {
+      "epoch": 1.0535714285714286,
+      "grad_norm": 2.197078573437343,
+      "learning_rate": 9.631234783051544e-06,
+      "loss": 0.6039,
+      "mean_token_accuracy": 0.8100722432136536,
+      "num_tokens": 266156871.0,
+      "step": 708
+    },
+    {
+      "epoch": 1.0550595238095237,
+      "grad_norm": 1.905731519479051,
+      "learning_rate": 9.62927441331743e-06,
+      "loss": 0.5473,
+      "mean_token_accuracy": 0.8251760005950928,
+      "num_tokens": 266535926.0,
+      "step": 709
+    },
+    {
+      "epoch": 1.056547619047619,
+      "grad_norm": 1.8992805903491818,
+      "learning_rate": 9.627309047276974e-06,
+      "loss": 0.5006,
+      "mean_token_accuracy": 0.8385787606239319,
+      "num_tokens": 266906437.0,
+      "step": 710
+    },
+    {
+      "epoch": 1.0580357142857142,
+      "grad_norm": 2.0489055640082188,
+      "learning_rate": 9.625338687051376e-06,
+      "loss": 0.5637,
+      "mean_token_accuracy": 0.8187175989151001,
+      "num_tokens": 267267341.0,
+      "step": 711
+    },
+    {
+      "epoch": 1.0595238095238095,
+      "grad_norm": 2.458806104973723,
+      "learning_rate": 9.623363334767208e-06,
+      "loss": 0.6716,
+      "mean_token_accuracy": 0.7943827509880066,
+      "num_tokens": 267644274.0,
+      "step": 712
+    },
+    {
+      "epoch": 1.0610119047619047,
+      "grad_norm": 2.0525912599185245,
+      "learning_rate": 9.621382992556443e-06,
+      "loss": 0.5594,
+      "mean_token_accuracy": 0.8225187659263611,
+      "num_tokens": 268023229.0,
+      "step": 713
+    },
+    {
+      "epoch": 1.0625,
+      "grad_norm": 2.0767531156502903,
+      "learning_rate": 9.619397662556434e-06,
+      "loss": 0.5694,
+      "mean_token_accuracy": 0.8160852789878845,
+      "num_tokens": 268402499.0,
+      "step": 714
+    },
+    {
+      "epoch": 1.0639880952380953,
+      "grad_norm": 2.4651364726010803,
+      "learning_rate": 9.617407346909917e-06,
+      "loss": 0.5808,
+      "mean_token_accuracy": 0.8186082243919373,
+      "num_tokens": 268779965.0,
+      "step": 715
+    },
+    {
+      "epoch": 1.0654761904761905,
+      "grad_norm": 1.9733667245733477,
+      "learning_rate": 9.61541204776501e-06,
+      "loss": 0.5635,
+      "mean_token_accuracy": 0.8203259110450745,
+      "num_tokens": 269158413.0,
+      "step": 716
+    },
+    {
+      "epoch": 1.0669642857142858,
+      "grad_norm": 2.011071529652574,
+      "learning_rate": 9.613411767275209e-06,
+      "loss": 0.5346,
+      "mean_token_accuracy": 0.8278758525848389,
+      "num_tokens": 269537013.0,
+      "step": 717
+    },
+    {
+      "epoch": 1.068452380952381,
+      "grad_norm": 2.202401288807394,
+      "learning_rate": 9.611406507599388e-06,
+      "loss": 0.6357,
+      "mean_token_accuracy": 0.801872193813324,
+      "num_tokens": 269916068.0,
+      "step": 718
+    },
+    {
+      "epoch": 1.0699404761904763,
+      "grad_norm": 2.0442923776356405,
+      "learning_rate": 9.609396270901791e-06,
+      "loss": 0.5422,
+      "mean_token_accuracy": 0.8293498754501343,
+      "num_tokens": 270280771.0,
+      "step": 719
+    },
+    {
+      "epoch": 1.0714285714285714,
+      "grad_norm": 1.973742115746249,
+      "learning_rate": 9.60738105935204e-06,
+      "loss": 0.5593,
+      "mean_token_accuracy": 0.8201768398284912,
+      "num_tokens": 270660360.0,
+      "step": 720
+    },
+    {
+      "epoch": 1.0729166666666667,
+      "grad_norm": 2.147062166066114,
+      "learning_rate": 9.605360875125119e-06,
+      "loss": 0.5643,
+      "mean_token_accuracy": 0.8155251145362854,
+      "num_tokens": 271024110.0,
+      "step": 721
+    },
+    {
+      "epoch": 1.0744047619047619,
+      "grad_norm": 2.23475409843562,
+      "learning_rate": 9.603335720401384e-06,
+      "loss": 0.5945,
+      "mean_token_accuracy": 0.8107507824897766,
+      "num_tokens": 271401432.0,
+      "step": 722
+    },
+    {
+      "epoch": 1.0758928571428572,
+      "grad_norm": 2.0889543302086944,
+      "learning_rate": 9.601305597366553e-06,
+      "loss": 0.5954,
+      "mean_token_accuracy": 0.8185188174247742,
+      "num_tokens": 271769031.0,
+      "step": 723
+    },
+    {
+      "epoch": 1.0773809523809523,
+      "grad_norm": 2.2432158774631077,
+      "learning_rate": 9.59927050821171e-06,
+      "loss": 0.5917,
+      "mean_token_accuracy": 0.8160815238952637,
+      "num_tokens": 272147261.0,
+      "step": 724
+    },
+    {
+      "epoch": 1.0788690476190477,
+      "grad_norm": 2.0418430636012994,
+      "learning_rate": 9.597230455133294e-06,
+      "loss": 0.5212,
+      "mean_token_accuracy": 0.8374292850494385,
+      "num_tokens": 272524592.0,
+      "step": 725
+    },
+    {
+      "epoch": 1.0803571428571428,
+      "grad_norm": 1.9162790264094096,
+      "learning_rate": 9.595185440333103e-06,
+      "loss": 0.5076,
+      "mean_token_accuracy": 0.8351022601127625,
+      "num_tokens": 272902928.0,
+      "step": 726
+    },
+    {
+      "epoch": 1.0818452380952381,
+      "grad_norm": 2.2661600545805682,
+      "learning_rate": 9.593135466018295e-06,
+      "loss": 0.5428,
+      "mean_token_accuracy": 0.8249918222427368,
+      "num_tokens": 273281876.0,
+      "step": 727
+    },
+    {
+      "epoch": 1.0833333333333333,
+      "grad_norm": 2.2487822391367636,
+      "learning_rate": 9.591080534401371e-06,
+      "loss": 0.5465,
+      "mean_token_accuracy": 0.8200560212135315,
+      "num_tokens": 273660625.0,
+      "step": 728
+    },
+    {
+      "epoch": 1.0848214285714286,
+      "grad_norm": 2.242244908237911,
+      "learning_rate": 9.589020647700191e-06,
+      "loss": 0.5393,
+      "mean_token_accuracy": 0.8250343799591064,
+      "num_tokens": 274038582.0,
+      "step": 729
+    },
+    {
+      "epoch": 1.0863095238095237,
+      "grad_norm": 2.1329024194447266,
+      "learning_rate": 9.586955808137958e-06,
+      "loss": 0.6022,
+      "mean_token_accuracy": 0.8050121665000916,
+      "num_tokens": 274415610.0,
+      "step": 730
+    },
+    {
+      "epoch": 1.087797619047619,
+      "grad_norm": 2.2738855276725816,
+      "learning_rate": 9.584886017943224e-06,
+      "loss": 0.5867,
+      "mean_token_accuracy": 0.8121277689933777,
+      "num_tokens": 274780495.0,
+      "step": 731
+    },
+    {
+      "epoch": 1.0892857142857142,
+      "grad_norm": 2.183875722212954,
+      "learning_rate": 9.582811279349881e-06,
+      "loss": 0.5603,
+      "mean_token_accuracy": 0.8222165107727051,
+      "num_tokens": 275158086.0,
+      "step": 732
+    },
+    {
+      "epoch": 1.0907738095238095,
+      "grad_norm": 2.287744359191613,
+      "learning_rate": 9.580731594597165e-06,
+      "loss": 0.5859,
+      "mean_token_accuracy": 0.8129780292510986,
+      "num_tokens": 275537834.0,
+      "step": 733
+    },
+    {
+      "epoch": 1.0922619047619047,
+      "grad_norm": 1.9288066579621348,
+      "learning_rate": 9.578646965929648e-06,
+      "loss": 0.527,
+      "mean_token_accuracy": 0.828852653503418,
+      "num_tokens": 275915846.0,
+      "step": 734
+    },
+    {
+      "epoch": 1.09375,
+      "grad_norm": 2.1151334461522655,
+      "learning_rate": 9.576557395597237e-06,
+      "loss": 0.5955,
+      "mean_token_accuracy": 0.8059327602386475,
+      "num_tokens": 276292838.0,
+      "step": 735
+    },
+    {
+      "epoch": 1.0952380952380953,
+      "grad_norm": 2.107802968115475,
+      "learning_rate": 9.574462885855173e-06,
+      "loss": 0.5422,
+      "mean_token_accuracy": 0.8217831254005432,
+      "num_tokens": 276670592.0,
+      "step": 736
+    },
+    {
+      "epoch": 1.0967261904761905,
+      "grad_norm": 2.3081463525885066,
+      "learning_rate": 9.572363438964034e-06,
+      "loss": 0.5814,
+      "mean_token_accuracy": 0.8077414631843567,
+      "num_tokens": 277047897.0,
+      "step": 737
+    },
+    {
+      "epoch": 1.0982142857142858,
+      "grad_norm": 2.041711636714166,
+      "learning_rate": 9.570259057189716e-06,
+      "loss": 0.5355,
+      "mean_token_accuracy": 0.8184574246406555,
+      "num_tokens": 277424483.0,
+      "step": 738
+    },
+    {
+      "epoch": 1.099702380952381,
+      "grad_norm": 2.125249298606362,
+      "learning_rate": 9.568149742803449e-06,
+      "loss": 0.5616,
+      "mean_token_accuracy": 0.8307952880859375,
+      "num_tokens": 277802663.0,
+      "step": 739
+    },
+    {
+      "epoch": 1.1011904761904763,
+      "grad_norm": 1.964138826805037,
+      "learning_rate": 9.566035498081785e-06,
+      "loss": 0.5723,
+      "mean_token_accuracy": 0.8183988332748413,
+      "num_tokens": 278182675.0,
+      "step": 740
+    },
+    {
+      "epoch": 1.1026785714285714,
+      "grad_norm": 2.1609217471747644,
+      "learning_rate": 9.563916325306595e-06,
+      "loss": 0.5007,
+      "mean_token_accuracy": 0.8376145958900452,
+      "num_tokens": 278559242.0,
+      "step": 741
+    },
+    {
+      "epoch": 1.1041666666666667,
+      "grad_norm": 2.007114204133707,
+      "learning_rate": 9.561792226765072e-06,
+      "loss": 0.5472,
+      "mean_token_accuracy": 0.8247008323669434,
+      "num_tokens": 278938691.0,
+      "step": 742
+    },
+    {
+      "epoch": 1.1056547619047619,
+      "grad_norm": 2.0210224190528967,
+      "learning_rate": 9.55966320474972e-06,
+      "loss": 0.5719,
+      "mean_token_accuracy": 0.8194056153297424,
+      "num_tokens": 279296967.0,
+      "step": 743
+    },
+    {
+      "epoch": 1.1071428571428572,
+      "grad_norm": 2.3466806464544816,
+      "learning_rate": 9.557529261558367e-06,
+      "loss": 0.5697,
+      "mean_token_accuracy": 0.8179101943969727,
+      "num_tokens": 279675707.0,
+      "step": 744
+    },
+    {
+      "epoch": 1.1086309523809523,
+      "grad_norm": 2.0956744873319346,
+      "learning_rate": 9.555390399494139e-06,
+      "loss": 0.6026,
+      "mean_token_accuracy": 0.8090660572052002,
+      "num_tokens": 280053721.0,
+      "step": 745
+    },
+    {
+      "epoch": 1.1101190476190477,
+      "grad_norm": 1.996157388090437,
+      "learning_rate": 9.553246620865482e-06,
+      "loss": 0.5706,
+      "mean_token_accuracy": 0.8201414346694946,
+      "num_tokens": 280433569.0,
+      "step": 746
+    },
+    {
+      "epoch": 1.1116071428571428,
+      "grad_norm": 2.085722643463357,
+      "learning_rate": 9.551097927986143e-06,
+      "loss": 0.5833,
+      "mean_token_accuracy": 0.8170064687728882,
+      "num_tokens": 280813031.0,
+      "step": 747
+    },
+    {
+      "epoch": 1.1130952380952381,
+      "grad_norm": 2.048463574858198,
+      "learning_rate": 9.548944323175173e-06,
+      "loss": 0.5946,
+      "mean_token_accuracy": 0.8162456750869751,
+      "num_tokens": 281191955.0,
+      "step": 748
+    },
+    {
+      "epoch": 1.1145833333333333,
+      "grad_norm": 1.9175697992998826,
+      "learning_rate": 9.546785808756926e-06,
+      "loss": 0.5637,
+      "mean_token_accuracy": 0.8146288990974426,
+      "num_tokens": 281569071.0,
+      "step": 749
+    },
+    {
+      "epoch": 1.1160714285714286,
+      "grad_norm": 2.188159375176344,
+      "learning_rate": 9.544622387061055e-06,
+      "loss": 0.5511,
+      "mean_token_accuracy": 0.8204147219657898,
+      "num_tokens": 281942444.0,
+      "step": 750
+    },
+    {
+      "epoch": 1.1175595238095237,
+      "grad_norm": 1.983817677911432,
+      "learning_rate": 9.542454060422509e-06,
+      "loss": 0.6104,
+      "mean_token_accuracy": 0.8010402917861938,
+      "num_tokens": 282308766.0,
+      "step": 751
+    },
+    {
+      "epoch": 1.119047619047619,
+      "grad_norm": 2.069415668508082,
+      "learning_rate": 9.540280831181525e-06,
+      "loss": 0.6151,
+      "mean_token_accuracy": 0.8046026825904846,
+      "num_tokens": 282687020.0,
+      "step": 752
+    },
+    {
+      "epoch": 1.1205357142857142,
+      "grad_norm": 2.064081054041995,
+      "learning_rate": 9.538102701683643e-06,
+      "loss": 0.636,
+      "mean_token_accuracy": 0.799895167350769,
+      "num_tokens": 283065249.0,
+      "step": 753
+    },
+    {
+      "epoch": 1.1220238095238095,
+      "grad_norm": 2.2718555775203746,
+      "learning_rate": 9.535919674279685e-06,
+      "loss": 0.6213,
+      "mean_token_accuracy": 0.808293342590332,
+      "num_tokens": 283441116.0,
+      "step": 754
+    },
+    {
+      "epoch": 1.1235119047619047,
+      "grad_norm": 2.0094340452727164,
+      "learning_rate": 9.533731751325757e-06,
+      "loss": 0.6589,
+      "mean_token_accuracy": 0.7937465906143188,
+      "num_tokens": 283821519.0,
+      "step": 755
+    },
+    {
+      "epoch": 1.125,
+      "grad_norm": 1.9083118966878658,
+      "learning_rate": 9.531538935183252e-06,
+      "loss": 0.5462,
+      "mean_token_accuracy": 0.8283535242080688,
+      "num_tokens": 284199908.0,
+      "step": 756
+    },
+    {
+      "epoch": 1.1264880952380953,
+      "grad_norm": 1.9192774075855983,
+      "learning_rate": 9.529341228218842e-06,
+      "loss": 0.5731,
+      "mean_token_accuracy": 0.817098081111908,
+      "num_tokens": 284573448.0,
+      "step": 757
+    },
+    {
+      "epoch": 1.1279761904761905,
+      "grad_norm": 2.2907194187596365,
+      "learning_rate": 9.527138632804482e-06,
+      "loss": 0.5824,
+      "mean_token_accuracy": 0.8192100524902344,
+      "num_tokens": 284951477.0,
+      "step": 758
+    },
+    {
+      "epoch": 1.1294642857142858,
+      "grad_norm": 1.9119016672439488,
+      "learning_rate": 9.5249311513174e-06,
+      "loss": 0.5842,
+      "mean_token_accuracy": 0.8150478005409241,
+      "num_tokens": 285330683.0,
+      "step": 759
+    },
+    {
+      "epoch": 1.130952380952381,
+      "grad_norm": 2.1058874286665015,
+      "learning_rate": 9.522718786140096e-06,
+      "loss": 0.6204,
+      "mean_token_accuracy": 0.8024632930755615,
+      "num_tokens": 285707395.0,
+      "step": 760
+    },
+    {
+      "epoch": 1.1324404761904763,
+      "grad_norm": 2.024311570284342,
+      "learning_rate": 9.520501539660345e-06,
+      "loss": 0.6067,
+      "mean_token_accuracy": 0.8096943497657776,
+      "num_tokens": 286085343.0,
+      "step": 761
+    },
+    {
+      "epoch": 1.1339285714285714,
+      "grad_norm": 2.4009594876478926,
+      "learning_rate": 9.518279414271184e-06,
+      "loss": 0.5315,
+      "mean_token_accuracy": 0.8288061618804932,
+      "num_tokens": 286462663.0,
+      "step": 762
+    },
+    {
+      "epoch": 1.1354166666666667,
+      "grad_norm": 2.1484142798377523,
+      "learning_rate": 9.516052412370922e-06,
+      "loss": 0.5561,
+      "mean_token_accuracy": 0.823197603225708,
+      "num_tokens": 286841058.0,
+      "step": 763
+    },
+    {
+      "epoch": 1.1369047619047619,
+      "grad_norm": 2.0398038272267422,
+      "learning_rate": 9.51382053636313e-06,
+      "loss": 0.5755,
+      "mean_token_accuracy": 0.8172533512115479,
+      "num_tokens": 287218658.0,
+      "step": 764
+    },
+    {
+      "epoch": 1.1383928571428572,
+      "grad_norm": 2.14520021901154,
+      "learning_rate": 9.511583788656633e-06,
+      "loss": 0.5936,
+      "mean_token_accuracy": 0.815679669380188,
+      "num_tokens": 287596088.0,
+      "step": 765
+    },
+    {
+      "epoch": 1.1398809523809523,
+      "grad_norm": 2.0538695218162686,
+      "learning_rate": 9.509342171665525e-06,
+      "loss": 0.5693,
+      "mean_token_accuracy": 0.8165031671524048,
+      "num_tokens": 287974843.0,
+      "step": 766
+    },
+    {
+      "epoch": 1.1413690476190477,
+      "grad_norm": 1.867397699878955,
+      "learning_rate": 9.507095687809148e-06,
+      "loss": 0.5228,
+      "mean_token_accuracy": 0.8289710879325867,
+      "num_tokens": 288353834.0,
+      "step": 767
+    },
+    {
+      "epoch": 1.1428571428571428,
+      "grad_norm": 2.023052593757089,
+      "learning_rate": 9.504844339512096e-06,
+      "loss": 0.5876,
+      "mean_token_accuracy": 0.8078992962837219,
+      "num_tokens": 288732545.0,
+      "step": 768
+    },
+    {
+      "epoch": 1.1443452380952381,
+      "grad_norm": 1.8281320967645833,
+      "learning_rate": 9.502588129204217e-06,
+      "loss": 0.547,
+      "mean_token_accuracy": 0.8222616910934448,
+      "num_tokens": 289112235.0,
+      "step": 769
+    },
+    {
+      "epoch": 1.1458333333333333,
+      "grad_norm": 1.899405296276924,
+      "learning_rate": 9.500327059320606e-06,
+      "loss": 0.5873,
+      "mean_token_accuracy": 0.8154640793800354,
+      "num_tokens": 289476843.0,
+      "step": 770
+    },
+    {
+      "epoch": 1.1473214285714286,
+      "grad_norm": 1.8593991065962259,
+      "learning_rate": 9.498061132301601e-06,
+      "loss": 0.5626,
+      "mean_token_accuracy": 0.8219048976898193,
+      "num_tokens": 289856019.0,
+      "step": 771
+    },
+    {
+      "epoch": 1.1488095238095237,
+      "grad_norm": 2.1941941807217185,
+      "learning_rate": 9.495790350592782e-06,
+      "loss": 0.5476,
+      "mean_token_accuracy": 0.8237773776054382,
+      "num_tokens": 290233254.0,
+      "step": 772
+    },
+    {
+      "epoch": 1.150297619047619,
+      "grad_norm": 2.179649585049504,
+      "learning_rate": 9.49351471664497e-06,
+      "loss": 0.5873,
+      "mean_token_accuracy": 0.8100578188896179,
+      "num_tokens": 290597779.0,
+      "step": 773
+    },
+    {
+      "epoch": 1.1517857142857142,
+      "grad_norm": 2.1672953148879537,
+      "learning_rate": 9.49123423291422e-06,
+      "loss": 0.6117,
+      "mean_token_accuracy": 0.8002078533172607,
+      "num_tokens": 290976673.0,
+      "step": 774
+    },
+    {
+      "epoch": 1.1532738095238095,
+      "grad_norm": 1.9052362487103944,
+      "learning_rate": 9.488948901861832e-06,
+      "loss": 0.5458,
+      "mean_token_accuracy": 0.8195371627807617,
+      "num_tokens": 291355193.0,
+      "step": 775
+    },
+    {
+      "epoch": 1.1547619047619047,
+      "grad_norm": 2.0230511321540923,
+      "learning_rate": 9.486658725954321e-06,
+      "loss": 0.5419,
+      "mean_token_accuracy": 0.8269920945167542,
+      "num_tokens": 291733291.0,
+      "step": 776
+    },
+    {
+      "epoch": 1.15625,
+      "grad_norm": 1.9825560767110557,
+      "learning_rate": 9.484363707663443e-06,
+      "loss": 0.5892,
+      "mean_token_accuracy": 0.8171806335449219,
+      "num_tokens": 292111707.0,
+      "step": 777
+    },
+    {
+      "epoch": 1.1577380952380953,
+      "grad_norm": 2.17773477665293,
+      "learning_rate": 9.482063849466175e-06,
+      "loss": 0.5963,
+      "mean_token_accuracy": 0.8199836611747742,
+      "num_tokens": 292489952.0,
+      "step": 778
+    },
+    {
+      "epoch": 1.1592261904761905,
+      "grad_norm": 2.259471819835252,
+      "learning_rate": 9.479759153844722e-06,
+      "loss": 0.6464,
+      "mean_token_accuracy": 0.8054420351982117,
+      "num_tokens": 292855505.0,
+      "step": 779
+    },
+    {
+      "epoch": 1.1607142857142858,
+      "grad_norm": 1.9675552575293522,
+      "learning_rate": 9.477449623286505e-06,
+      "loss": 0.5352,
+      "mean_token_accuracy": 0.8228188157081604,
+      "num_tokens": 293234491.0,
+      "step": 780
+    },
+    {
+      "epoch": 1.162202380952381,
+      "grad_norm": 2.1421954084052026,
+      "learning_rate": 9.475135260284169e-06,
+      "loss": 0.5917,
+      "mean_token_accuracy": 0.8146676421165466,
+      "num_tokens": 293611351.0,
+      "step": 781
+    },
+    {
+      "epoch": 1.1636904761904763,
+      "grad_norm": 1.8865177056364566,
+      "learning_rate": 9.472816067335566e-06,
+      "loss": 0.5315,
+      "mean_token_accuracy": 0.8293294310569763,
+      "num_tokens": 293990089.0,
+      "step": 782
+    },
+    {
+      "epoch": 1.1651785714285714,
+      "grad_norm": 2.1570809414097125,
+      "learning_rate": 9.47049204694377e-06,
+      "loss": 0.5742,
+      "mean_token_accuracy": 0.813149094581604,
+      "num_tokens": 294366838.0,
+      "step": 783
+    },
+    {
+      "epoch": 1.1666666666666667,
+      "grad_norm": 2.073006854363427,
+      "learning_rate": 9.468163201617063e-06,
+      "loss": 0.564,
+      "mean_token_accuracy": 0.822139322757721,
+      "num_tokens": 294732837.0,
+      "step": 784
+    },
+    {
+      "epoch": 1.1681547619047619,
+      "grad_norm": 2.0098650199735686,
+      "learning_rate": 9.465829533868927e-06,
+      "loss": 0.6236,
+      "mean_token_accuracy": 0.8058885931968689,
+      "num_tokens": 295098076.0,
+      "step": 785
+    },
+    {
+      "epoch": 1.1696428571428572,
+      "grad_norm": 2.135348997786266,
+      "learning_rate": 9.463491046218058e-06,
+      "loss": 0.6109,
+      "mean_token_accuracy": 0.8046916127204895,
+      "num_tokens": 295476812.0,
+      "step": 786
+    },
+    {
+      "epoch": 1.1711309523809523,
+      "grad_norm": 2.094102872738001,
+      "learning_rate": 9.46114774118835e-06,
+      "loss": 0.5262,
+      "mean_token_accuracy": 0.8299300670623779,
+      "num_tokens": 295838370.0,
+      "step": 787
+    },
+    {
+      "epoch": 1.1726190476190477,
+      "grad_norm": 1.9946618124829638,
+      "learning_rate": 9.458799621308898e-06,
+      "loss": 0.5757,
+      "mean_token_accuracy": 0.8213324546813965,
+      "num_tokens": 296209989.0,
+      "step": 788
+    },
+    {
+      "epoch": 1.1741071428571428,
+      "grad_norm": 2.075761436606432,
+      "learning_rate": 9.456446689113992e-06,
+      "loss": 0.5573,
+      "mean_token_accuracy": 0.8200976848602295,
+      "num_tokens": 296588959.0,
+      "step": 789
+    },
+    {
+      "epoch": 1.1755952380952381,
+      "grad_norm": 2.1890201204033124,
+      "learning_rate": 9.454088947143116e-06,
+      "loss": 0.5543,
+      "mean_token_accuracy": 0.818461537361145,
+      "num_tokens": 296968029.0,
+      "step": 790
+    },
+    {
+      "epoch": 1.1770833333333333,
+      "grad_norm": 2.3654814753395694,
+      "learning_rate": 9.451726397940946e-06,
+      "loss": 0.624,
+      "mean_token_accuracy": 0.800469696521759,
+      "num_tokens": 297347125.0,
+      "step": 791
+    },
+    {
+      "epoch": 1.1785714285714286,
+      "grad_norm": 2.1084648348109467,
+      "learning_rate": 9.449359044057344e-06,
+      "loss": 0.603,
+      "mean_token_accuracy": 0.8135844469070435,
+      "num_tokens": 297725607.0,
+      "step": 792
+    },
+    {
+      "epoch": 1.1800595238095237,
+      "grad_norm": 2.250197269608781,
+      "learning_rate": 9.446986888047361e-06,
+      "loss": 0.5626,
+      "mean_token_accuracy": 0.812006950378418,
+      "num_tokens": 298094669.0,
+      "step": 793
+    },
+    {
+      "epoch": 1.181547619047619,
+      "grad_norm": 2.2188815579431744,
+      "learning_rate": 9.44460993247123e-06,
+      "loss": 0.5917,
+      "mean_token_accuracy": 0.8109529614448547,
+      "num_tokens": 298472761.0,
+      "step": 794
+    },
+    {
+      "epoch": 1.1830357142857142,
+      "grad_norm": 2.036919267948368,
+      "learning_rate": 9.442228179894362e-06,
+      "loss": 0.5953,
+      "mean_token_accuracy": 0.8138692378997803,
+      "num_tokens": 298851774.0,
+      "step": 795
+    },
+    {
+      "epoch": 1.1845238095238095,
+      "grad_norm": 2.0100770977536735,
+      "learning_rate": 9.439841632887348e-06,
+      "loss": 0.5958,
+      "mean_token_accuracy": 0.8095526099205017,
+      "num_tokens": 299227228.0,
+      "step": 796
+    },
+    {
+      "epoch": 1.1860119047619047,
+      "grad_norm": 2.0389588710347173,
+      "learning_rate": 9.437450294025952e-06,
+      "loss": 0.5557,
+      "mean_token_accuracy": 0.8186625838279724,
+      "num_tokens": 299605969.0,
+      "step": 797
+    },
+    {
+      "epoch": 1.1875,
+      "grad_norm": 1.9233775343762343,
+      "learning_rate": 9.43505416589111e-06,
+      "loss": 0.5419,
+      "mean_token_accuracy": 0.823127031326294,
+      "num_tokens": 299985141.0,
+      "step": 798
+    },
+    {
+      "epoch": 1.1889880952380953,
+      "grad_norm": 2.0365940957053126,
+      "learning_rate": 9.432653251068926e-06,
+      "loss": 0.5659,
+      "mean_token_accuracy": 0.8190015554428101,
+      "num_tokens": 300362579.0,
+      "step": 799
+    },
+    {
+      "epoch": 1.1904761904761905,
+      "grad_norm": 2.307196514014825,
+      "learning_rate": 9.430247552150673e-06,
+      "loss": 0.5849,
+      "mean_token_accuracy": 0.8170263171195984,
+      "num_tokens": 300739216.0,
+      "step": 800
+    },
+    {
+      "epoch": 1.1919642857142858,
+      "grad_norm": 2.101875360478469,
+      "learning_rate": 9.427837071732784e-06,
+      "loss": 0.5666,
+      "mean_token_accuracy": 0.823324978351593,
+      "num_tokens": 301116254.0,
+      "step": 801
+    },
+    {
+      "epoch": 1.193452380952381,
+      "grad_norm": 2.1538928023145956,
+      "learning_rate": 9.425421812416857e-06,
+      "loss": 0.5232,
+      "mean_token_accuracy": 0.8298402428627014,
+      "num_tokens": 301493320.0,
+      "step": 802
+    },
+    {
+      "epoch": 1.1949404761904763,
+      "grad_norm": 2.0272911010822052,
+      "learning_rate": 9.423001776809642e-06,
+      "loss": 0.5382,
+      "mean_token_accuracy": 0.8237878084182739,
+      "num_tokens": 301871302.0,
+      "step": 803
+    },
+    {
+      "epoch": 1.1964285714285714,
+      "grad_norm": 2.1612165849032565,
+      "learning_rate": 9.420576967523049e-06,
+      "loss": 0.547,
+      "mean_token_accuracy": 0.8195620179176331,
+      "num_tokens": 302249292.0,
+      "step": 804
+    },
+    {
+      "epoch": 1.1979166666666667,
+      "grad_norm": 2.254327969613661,
+      "learning_rate": 9.41814738717414e-06,
+      "loss": 0.6343,
+      "mean_token_accuracy": 0.8072141408920288,
+      "num_tokens": 302627667.0,
+      "step": 805
+    },
+    {
+      "epoch": 1.1994047619047619,
+      "grad_norm": 2.1647427206890675,
+      "learning_rate": 9.41571303838512e-06,
+      "loss": 0.6196,
+      "mean_token_accuracy": 0.8067978620529175,
+      "num_tokens": 303006137.0,
+      "step": 806
+    },
+    {
+      "epoch": 1.2008928571428572,
+      "grad_norm": 2.1052230488243238,
+      "learning_rate": 9.413273923783346e-06,
+      "loss": 0.6021,
+      "mean_token_accuracy": 0.8040948510169983,
+      "num_tokens": 303381077.0,
+      "step": 807
+    },
+    {
+      "epoch": 1.2023809523809523,
+      "grad_norm": 2.063565795405654,
+      "learning_rate": 9.410830046001321e-06,
+      "loss": 0.564,
+      "mean_token_accuracy": 0.8219788670539856,
+      "num_tokens": 303758524.0,
+      "step": 808
+    },
+    {
+      "epoch": 1.2038690476190477,
+      "grad_norm": 2.165662728985711,
+      "learning_rate": 9.408381407676684e-06,
+      "loss": 0.6462,
+      "mean_token_accuracy": 0.8085319399833679,
+      "num_tokens": 304135700.0,
+      "step": 809
+    },
+    {
+      "epoch": 1.2053571428571428,
+      "grad_norm": 1.9000241502217419,
+      "learning_rate": 9.405928011452211e-06,
+      "loss": 0.5512,
+      "mean_token_accuracy": 0.8172851800918579,
+      "num_tokens": 304514560.0,
+      "step": 810
+    },
+    {
+      "epoch": 1.2068452380952381,
+      "grad_norm": 1.988706683416688,
+      "learning_rate": 9.403469859975822e-06,
+      "loss": 0.535,
+      "mean_token_accuracy": 0.8231908679008484,
+      "num_tokens": 304879780.0,
+      "step": 811
+    },
+    {
+      "epoch": 1.2083333333333333,
+      "grad_norm": 2.101601469875732,
+      "learning_rate": 9.401006955900555e-06,
+      "loss": 0.5624,
+      "mean_token_accuracy": 0.8202975988388062,
+      "num_tokens": 305256666.0,
+      "step": 812
+    },
+    {
+      "epoch": 1.2098214285714286,
+      "grad_norm": 2.0857602611695976,
+      "learning_rate": 9.398539301884592e-06,
+      "loss": 0.5797,
+      "mean_token_accuracy": 0.8178226351737976,
+      "num_tokens": 305634638.0,
+      "step": 813
+    },
+    {
+      "epoch": 1.2113095238095237,
+      "grad_norm": 2.0879306338923485,
+      "learning_rate": 9.396066900591232e-06,
+      "loss": 0.5661,
+      "mean_token_accuracy": 0.8226495981216431,
+      "num_tokens": 306014246.0,
+      "step": 814
+    },
+    {
+      "epoch": 1.212797619047619,
+      "grad_norm": 2.13872524280848,
+      "learning_rate": 9.3935897546889e-06,
+      "loss": 0.6066,
+      "mean_token_accuracy": 0.805692732334137,
+      "num_tokens": 306380122.0,
+      "step": 815
+    },
+    {
+      "epoch": 1.2142857142857142,
+      "grad_norm": 1.9451719855670677,
+      "learning_rate": 9.391107866851143e-06,
+      "loss": 0.5897,
+      "mean_token_accuracy": 0.8119219541549683,
+      "num_tokens": 306758567.0,
+      "step": 816
+    },
+    {
+      "epoch": 1.2157738095238095,
+      "grad_norm": 1.8601793856777138,
+      "learning_rate": 9.388621239756626e-06,
+      "loss": 0.5587,
+      "mean_token_accuracy": 0.8239633440971375,
+      "num_tokens": 307137266.0,
+      "step": 817
+    },
+    {
+      "epoch": 1.2172619047619047,
+      "grad_norm": 1.9461290677755778,
+      "learning_rate": 9.386129876089128e-06,
+      "loss": 0.5512,
+      "mean_token_accuracy": 0.8207309246063232,
+      "num_tokens": 307515099.0,
+      "step": 818
+    },
+    {
+      "epoch": 1.21875,
+      "grad_norm": 2.0394003296472967,
+      "learning_rate": 9.38363377853754e-06,
+      "loss": 0.5633,
+      "mean_token_accuracy": 0.8170075416564941,
+      "num_tokens": 307893377.0,
+      "step": 819
+    },
+    {
+      "epoch": 1.2202380952380953,
+      "grad_norm": 2.0858348119949155,
+      "learning_rate": 9.381132949795862e-06,
+      "loss": 0.5185,
+      "mean_token_accuracy": 0.8359746932983398,
+      "num_tokens": 308270652.0,
+      "step": 820
+    },
+    {
+      "epoch": 1.2217261904761905,
+      "grad_norm": 1.9942413711742708,
+      "learning_rate": 9.378627392563202e-06,
+      "loss": 0.5338,
+      "mean_token_accuracy": 0.8256261944770813,
+      "num_tokens": 308649634.0,
+      "step": 821
+    },
+    {
+      "epoch": 1.2232142857142858,
+      "grad_norm": 2.067775867915946,
+      "learning_rate": 9.376117109543769e-06,
+      "loss": 0.5744,
+      "mean_token_accuracy": 0.8128575086593628,
+      "num_tokens": 309028553.0,
+      "step": 822
+    },
+    {
+      "epoch": 1.224702380952381,
+      "grad_norm": 2.1724838969984512,
+      "learning_rate": 9.373602103446875e-06,
+      "loss": 0.5986,
+      "mean_token_accuracy": 0.8113699555397034,
+      "num_tokens": 309406952.0,
+      "step": 823
+    },
+    {
+      "epoch": 1.2261904761904763,
+      "grad_norm": 2.1450963662241547,
+      "learning_rate": 9.37108237698693e-06,
+      "loss": 0.6227,
+      "mean_token_accuracy": 0.8063614964485168,
+      "num_tokens": 309781784.0,
+      "step": 824
+    },
+    {
+      "epoch": 1.2276785714285714,
+      "grad_norm": 2.2058803313968998,
+      "learning_rate": 9.368557932883432e-06,
+      "loss": 0.5734,
+      "mean_token_accuracy": 0.8186208009719849,
+      "num_tokens": 310144426.0,
+      "step": 825
+    },
+    {
+      "epoch": 1.2291666666666667,
+      "grad_norm": 2.084850168908297,
+      "learning_rate": 9.366028773860981e-06,
+      "loss": 0.6262,
+      "mean_token_accuracy": 0.8008291721343994,
+      "num_tokens": 310510711.0,
+      "step": 826
+    },
+    {
+      "epoch": 1.2306547619047619,
+      "grad_norm": 2.1978022995372224,
+      "learning_rate": 9.363494902649256e-06,
+      "loss": 0.5934,
+      "mean_token_accuracy": 0.8111493587493896,
+      "num_tokens": 310888245.0,
+      "step": 827
+    },
+    {
+      "epoch": 1.2321428571428572,
+      "grad_norm": 1.9574881466615788,
+      "learning_rate": 9.360956321983028e-06,
+      "loss": 0.5666,
+      "mean_token_accuracy": 0.8162776231765747,
+      "num_tokens": 311266573.0,
+      "step": 828
+    },
+    {
+      "epoch": 1.2336309523809523,
+      "grad_norm": 1.9236785253690962,
+      "learning_rate": 9.358413034602149e-06,
+      "loss": 0.5372,
+      "mean_token_accuracy": 0.8300283551216125,
+      "num_tokens": 311644171.0,
+      "step": 829
+    },
+    {
+      "epoch": 1.2351190476190477,
+      "grad_norm": 2.0993445791479437,
+      "learning_rate": 9.355865043251551e-06,
+      "loss": 0.594,
+      "mean_token_accuracy": 0.8120816946029663,
+      "num_tokens": 312007972.0,
+      "step": 830
+    },
+    {
+      "epoch": 1.2366071428571428,
+      "grad_norm": 2.0478242368236224,
+      "learning_rate": 9.353312350681242e-06,
+      "loss": 0.5155,
+      "mean_token_accuracy": 0.8334861397743225,
+      "num_tokens": 312385027.0,
+      "step": 831
+    },
+    {
+      "epoch": 1.2380952380952381,
+      "grad_norm": 1.9570121049331055,
+      "learning_rate": 9.350754959646306e-06,
+      "loss": 0.5028,
+      "mean_token_accuracy": 0.8429684638977051,
+      "num_tokens": 312763039.0,
+      "step": 832
+    },
+    {
+      "epoch": 1.2395833333333333,
+      "grad_norm": 1.9778527985040357,
+      "learning_rate": 9.348192872906896e-06,
+      "loss": 0.5396,
+      "mean_token_accuracy": 0.8312588930130005,
+      "num_tokens": 313143030.0,
+      "step": 833
+    },
+    {
+      "epoch": 1.2410714285714286,
+      "grad_norm": 2.027112558173173,
+      "learning_rate": 9.345626093228233e-06,
+      "loss": 0.5997,
+      "mean_token_accuracy": 0.8092144131660461,
+      "num_tokens": 313510048.0,
+      "step": 834
+    },
+    {
+      "epoch": 1.2425595238095237,
+      "grad_norm": 2.098152554542254,
+      "learning_rate": 9.343054623380606e-06,
+      "loss": 0.6165,
+      "mean_token_accuracy": 0.8073710203170776,
+      "num_tokens": 313888393.0,
+      "step": 835
+    },
+    {
+      "epoch": 1.244047619047619,
+      "grad_norm": 1.9393222774026213,
+      "learning_rate": 9.34047846613936e-06,
+      "loss": 0.5067,
+      "mean_token_accuracy": 0.8374354243278503,
+      "num_tokens": 314256861.0,
+      "step": 836
+    },
+    {
+      "epoch": 1.2455357142857142,
+      "grad_norm": 2.2054786865727114,
+      "learning_rate": 9.337897624284908e-06,
+      "loss": 0.5856,
+      "mean_token_accuracy": 0.8111719489097595,
+      "num_tokens": 314633734.0,
+      "step": 837
+    },
+    {
+      "epoch": 1.2470238095238095,
+      "grad_norm": 2.040840418838966,
+      "learning_rate": 9.33531210060271e-06,
+      "loss": 0.5701,
+      "mean_token_accuracy": 0.8210543990135193,
+      "num_tokens": 315012968.0,
+      "step": 838
+    },
+    {
+      "epoch": 1.2485119047619047,
+      "grad_norm": 1.989434408323571,
+      "learning_rate": 9.332721897883283e-06,
+      "loss": 0.5666,
+      "mean_token_accuracy": 0.8197983503341675,
+      "num_tokens": 315393470.0,
+      "step": 839
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 2.0756941978847387,
+      "learning_rate": 9.330127018922195e-06,
+      "loss": 0.5554,
+      "mean_token_accuracy": 0.8235189914703369,
+      "num_tokens": 315771547.0,
+      "step": 840
+    },
+    {
+      "epoch": 1.2514880952380953,
+      "grad_norm": 1.9047939139117398,
+      "learning_rate": 9.327527466520059e-06,
+      "loss": 0.5031,
+      "mean_token_accuracy": 0.8344039916992188,
+      "num_tokens": 316149875.0,
+      "step": 841
+    },
+    {
+      "epoch": 1.2529761904761905,
+      "grad_norm": 2.0798741614236333,
+      "learning_rate": 9.324923243482533e-06,
+      "loss": 0.5551,
+      "mean_token_accuracy": 0.8273730874061584,
+      "num_tokens": 316528326.0,
+      "step": 842
+    },
+    {
+      "epoch": 1.2544642857142856,
+      "grad_norm": 1.9916624836523125,
+      "learning_rate": 9.322314352620318e-06,
+      "loss": 0.5376,
+      "mean_token_accuracy": 0.8222545981407166,
+      "num_tokens": 316906649.0,
+      "step": 843
+    },
+    {
+      "epoch": 1.255952380952381,
+      "grad_norm": 1.9840168098242779,
+      "learning_rate": 9.31970079674915e-06,
+      "loss": 0.5464,
+      "mean_token_accuracy": 0.8237038254737854,
+      "num_tokens": 317285685.0,
+      "step": 844
+    },
+    {
+      "epoch": 1.2574404761904763,
+      "grad_norm": 2.0591435628305255,
+      "learning_rate": 9.317082578689796e-06,
+      "loss": 0.5619,
+      "mean_token_accuracy": 0.8207365274429321,
+      "num_tokens": 317651401.0,
+      "step": 845
+    },
+    {
+      "epoch": 1.2589285714285714,
+      "grad_norm": 1.9850095207926712,
+      "learning_rate": 9.314459701268065e-06,
+      "loss": 0.5773,
+      "mean_token_accuracy": 0.8142967224121094,
+      "num_tokens": 318030127.0,
+      "step": 846
+    },
+    {
+      "epoch": 1.2604166666666667,
+      "grad_norm": 2.0626891765581363,
+      "learning_rate": 9.311832167314788e-06,
+      "loss": 0.603,
+      "mean_token_accuracy": 0.8123400211334229,
+      "num_tokens": 318408052.0,
+      "step": 847
+    },
+    {
+      "epoch": 1.2619047619047619,
+      "grad_norm": 2.0519748450911037,
+      "learning_rate": 9.309199979665821e-06,
+      "loss": 0.5331,
+      "mean_token_accuracy": 0.8214254379272461,
+      "num_tokens": 318786133.0,
+      "step": 848
+    },
+    {
+      "epoch": 1.2633928571428572,
+      "grad_norm": 2.1515940548382524,
+      "learning_rate": 9.306563141162046e-06,
+      "loss": 0.6198,
+      "mean_token_accuracy": 0.8129520416259766,
+      "num_tokens": 319164853.0,
+      "step": 849
+    },
+    {
+      "epoch": 1.2648809523809523,
+      "grad_norm": 1.906058184990987,
+      "learning_rate": 9.30392165464936e-06,
+      "loss": 0.5163,
+      "mean_token_accuracy": 0.8309871554374695,
+      "num_tokens": 319543557.0,
+      "step": 850
+    },
+    {
+      "epoch": 1.2663690476190477,
+      "grad_norm": 2.1268940771617117,
+      "learning_rate": 9.301275522978687e-06,
+      "loss": 0.6251,
+      "mean_token_accuracy": 0.8040232062339783,
+      "num_tokens": 319922153.0,
+      "step": 851
+    },
+    {
+      "epoch": 1.2678571428571428,
+      "grad_norm": 2.2234566671756784,
+      "learning_rate": 9.298624749005953e-06,
+      "loss": 0.5796,
+      "mean_token_accuracy": 0.8122282028198242,
+      "num_tokens": 320300264.0,
+      "step": 852
+    },
+    {
+      "epoch": 1.2693452380952381,
+      "grad_norm": 1.957549651990241,
+      "learning_rate": 9.295969335592096e-06,
+      "loss": 0.5589,
+      "mean_token_accuracy": 0.8205019235610962,
+      "num_tokens": 320678641.0,
+      "step": 853
+    },
+    {
+      "epoch": 1.2708333333333333,
+      "grad_norm": 1.823140941937733,
+      "learning_rate": 9.293309285603066e-06,
+      "loss": 0.5118,
+      "mean_token_accuracy": 0.8299343585968018,
+      "num_tokens": 321050783.0,
+      "step": 854
+    },
+    {
+      "epoch": 1.2723214285714286,
+      "grad_norm": 1.9512439156171293,
+      "learning_rate": 9.290644601909816e-06,
+      "loss": 0.5614,
+      "mean_token_accuracy": 0.8226248025894165,
+      "num_tokens": 321430104.0,
+      "step": 855
+    },
+    {
+      "epoch": 1.2738095238095237,
+      "grad_norm": 2.0736624788991045,
+      "learning_rate": 9.287975287388297e-06,
+      "loss": 0.6661,
+      "mean_token_accuracy": 0.7898349165916443,
+      "num_tokens": 321808422.0,
+      "step": 856
+    },
+    {
+      "epoch": 1.275297619047619,
+      "grad_norm": 2.0610591909921774,
+      "learning_rate": 9.285301344919463e-06,
+      "loss": 0.6072,
+      "mean_token_accuracy": 0.807233989238739,
+      "num_tokens": 322175143.0,
+      "step": 857
+    },
+    {
+      "epoch": 1.2767857142857144,
+      "grad_norm": 2.165236972400485,
+      "learning_rate": 9.282622777389258e-06,
+      "loss": 0.5736,
+      "mean_token_accuracy": 0.8122250437736511,
+      "num_tokens": 322539874.0,
+      "step": 858
+    },
+    {
+      "epoch": 1.2782738095238095,
+      "grad_norm": 1.9765181161397263,
+      "learning_rate": 9.27993958768862e-06,
+      "loss": 0.5923,
+      "mean_token_accuracy": 0.8148206472396851,
+      "num_tokens": 322919878.0,
+      "step": 859
+    },
+    {
+      "epoch": 1.2797619047619047,
+      "grad_norm": 2.043577849529747,
+      "learning_rate": 9.277251778713475e-06,
+      "loss": 0.5438,
+      "mean_token_accuracy": 0.8180894255638123,
+      "num_tokens": 323291276.0,
+      "step": 860
+    },
+    {
+      "epoch": 1.28125,
+      "grad_norm": 2.0569098123992418,
+      "learning_rate": 9.274559353364734e-06,
+      "loss": 0.5664,
+      "mean_token_accuracy": 0.8195359706878662,
+      "num_tokens": 323669646.0,
+      "step": 861
+    },
+    {
+      "epoch": 1.2827380952380953,
+      "grad_norm": 1.8981100362461902,
+      "learning_rate": 9.271862314548292e-06,
+      "loss": 0.5664,
+      "mean_token_accuracy": 0.8162866830825806,
+      "num_tokens": 324048748.0,
+      "step": 862
+    },
+    {
+      "epoch": 1.2842261904761905,
+      "grad_norm": 2.1049790137416773,
+      "learning_rate": 9.269160665175026e-06,
+      "loss": 0.612,
+      "mean_token_accuracy": 0.8034878969192505,
+      "num_tokens": 324427427.0,
+      "step": 863
+    },
+    {
+      "epoch": 1.2857142857142856,
+      "grad_norm": 1.9768055438068797,
+      "learning_rate": 9.266454408160779e-06,
+      "loss": 0.5819,
+      "mean_token_accuracy": 0.815247118473053,
+      "num_tokens": 324807016.0,
+      "step": 864
+    },
+    {
+      "epoch": 1.287202380952381,
+      "grad_norm": 2.1198349523471514,
+      "learning_rate": 9.263743546426379e-06,
+      "loss": 0.62,
+      "mean_token_accuracy": 0.8045099377632141,
+      "num_tokens": 325172651.0,
+      "step": 865
+    },
+    {
+      "epoch": 1.2886904761904763,
+      "grad_norm": 1.96392173482627,
+      "learning_rate": 9.261028082897615e-06,
+      "loss": 0.6325,
+      "mean_token_accuracy": 0.8025031089782715,
+      "num_tokens": 325551706.0,
+      "step": 866
+    },
+    {
+      "epoch": 1.2901785714285714,
+      "grad_norm": 2.0590004315556434,
+      "learning_rate": 9.258308020505247e-06,
+      "loss": 0.6446,
+      "mean_token_accuracy": 0.8022189140319824,
+      "num_tokens": 325930170.0,
+      "step": 867
+    },
+    {
+      "epoch": 1.2916666666666667,
+      "grad_norm": 2.043254409157718,
+      "learning_rate": 9.255583362184998e-06,
+      "loss": 0.587,
+      "mean_token_accuracy": 0.8109766840934753,
+      "num_tokens": 326308570.0,
+      "step": 868
+    },
+    {
+      "epoch": 1.2931547619047619,
+      "grad_norm": 2.0791821957030128,
+      "learning_rate": 9.252854110877553e-06,
+      "loss": 0.6298,
+      "mean_token_accuracy": 0.8034330606460571,
+      "num_tokens": 326684935.0,
+      "step": 869
+    },
+    {
+      "epoch": 1.2946428571428572,
+      "grad_norm": 1.9505503425574522,
+      "learning_rate": 9.250120269528546e-06,
+      "loss": 0.5947,
+      "mean_token_accuracy": 0.8091248273849487,
+      "num_tokens": 327063725.0,
+      "step": 870
+    },
+    {
+      "epoch": 1.2961309523809523,
+      "grad_norm": 1.9156802201612508,
+      "learning_rate": 9.247381841088577e-06,
+      "loss": 0.5722,
+      "mean_token_accuracy": 0.8150380849838257,
+      "num_tokens": 327442832.0,
+      "step": 871
+    },
+    {
+      "epoch": 1.2976190476190477,
+      "grad_norm": 1.9075297641296547,
+      "learning_rate": 9.244638828513189e-06,
+      "loss": 0.5888,
+      "mean_token_accuracy": 0.8136942386627197,
+      "num_tokens": 327823083.0,
+      "step": 872
+    },
+    {
+      "epoch": 1.2991071428571428,
+      "grad_norm": 2.0149657551757327,
+      "learning_rate": 9.24189123476287e-06,
+      "loss": 0.5734,
+      "mean_token_accuracy": 0.8161783218383789,
+      "num_tokens": 328201742.0,
+      "step": 873
+    },
+    {
+      "epoch": 1.3005952380952381,
+      "grad_norm": 2.211872039562608,
+      "learning_rate": 9.239139062803062e-06,
+      "loss": 0.6324,
+      "mean_token_accuracy": 0.798663318157196,
+      "num_tokens": 328580625.0,
+      "step": 874
+    },
+    {
+      "epoch": 1.3020833333333333,
+      "grad_norm": 2.2012719474384026,
+      "learning_rate": 9.23638231560414e-06,
+      "loss": 0.5666,
+      "mean_token_accuracy": 0.8089052438735962,
+      "num_tokens": 328951701.0,
+      "step": 875
+    },
+    {
+      "epoch": 1.3035714285714286,
+      "grad_norm": 2.221818140323112,
+      "learning_rate": 9.233620996141421e-06,
+      "loss": 0.6146,
+      "mean_token_accuracy": 0.8100502490997314,
+      "num_tokens": 329331086.0,
+      "step": 876
+    },
+    {
+      "epoch": 1.3050595238095237,
+      "grad_norm": 2.0733862942143153,
+      "learning_rate": 9.230855107395155e-06,
+      "loss": 0.5916,
+      "mean_token_accuracy": 0.8142661452293396,
+      "num_tokens": 329701484.0,
+      "step": 877
+    },
+    {
+      "epoch": 1.306547619047619,
+      "grad_norm": 1.9418025385158602,
+      "learning_rate": 9.228084652350526e-06,
+      "loss": 0.5517,
+      "mean_token_accuracy": 0.8178122043609619,
+      "num_tokens": 330063739.0,
+      "step": 878
+    },
+    {
+      "epoch": 1.3080357142857144,
+      "grad_norm": 1.968847289176696,
+      "learning_rate": 9.225309633997641e-06,
+      "loss": 0.5404,
+      "mean_token_accuracy": 0.8256440758705139,
+      "num_tokens": 330442774.0,
+      "step": 879
+    },
+    {
+      "epoch": 1.3095238095238095,
+      "grad_norm": 2.162622043857718,
+      "learning_rate": 9.22253005533154e-06,
+      "loss": 0.5631,
+      "mean_token_accuracy": 0.8205040693283081,
+      "num_tokens": 330821667.0,
+      "step": 880
+    },
+    {
+      "epoch": 1.3110119047619047,
+      "grad_norm": 1.9862998656715212,
+      "learning_rate": 9.219745919352177e-06,
+      "loss": 0.5323,
+      "mean_token_accuracy": 0.8266732692718506,
+      "num_tokens": 331201010.0,
+      "step": 881
+    },
+    {
+      "epoch": 1.3125,
+      "grad_norm": 2.0850240597389647,
+      "learning_rate": 9.21695722906443e-06,
+      "loss": 0.5711,
+      "mean_token_accuracy": 0.8133552074432373,
+      "num_tokens": 331580101.0,
+      "step": 882
+    },
+    {
+      "epoch": 1.3139880952380953,
+      "grad_norm": 2.049545697936541,
+      "learning_rate": 9.214163987478089e-06,
+      "loss": 0.553,
+      "mean_token_accuracy": 0.8176780939102173,
+      "num_tokens": 331958350.0,
+      "step": 883
+    },
+    {
+      "epoch": 1.3154761904761905,
+      "grad_norm": 2.438500027286287,
+      "learning_rate": 9.211366197607858e-06,
+      "loss": 0.6297,
+      "mean_token_accuracy": 0.8087493181228638,
+      "num_tokens": 332323623.0,
+      "step": 884
+    },
+    {
+      "epoch": 1.3169642857142856,
+      "grad_norm": 2.062264338194174,
+      "learning_rate": 9.20856386247335e-06,
+      "loss": 0.5408,
+      "mean_token_accuracy": 0.8261436820030212,
+      "num_tokens": 332704218.0,
+      "step": 885
+    },
+    {
+      "epoch": 1.318452380952381,
+      "grad_norm": 2.321462998466097,
+      "learning_rate": 9.205756985099084e-06,
+      "loss": 0.6691,
+      "mean_token_accuracy": 0.7909261584281921,
+      "num_tokens": 333082490.0,
+      "step": 886
+    },
+    {
+      "epoch": 1.3199404761904763,
+      "grad_norm": 1.988690012243049,
+      "learning_rate": 9.202945568514482e-06,
+      "loss": 0.5866,
+      "mean_token_accuracy": 0.8083061575889587,
+      "num_tokens": 333460269.0,
+      "step": 887
+    },
+    {
+      "epoch": 1.3214285714285714,
+      "grad_norm": 2.0181479711573767,
+      "learning_rate": 9.200129615753858e-06,
+      "loss": 0.5281,
+      "mean_token_accuracy": 0.8294509053230286,
+      "num_tokens": 333838883.0,
+      "step": 888
+    },
+    {
+      "epoch": 1.3229166666666667,
+      "grad_norm": 2.171170239417196,
+      "learning_rate": 9.197309129856433e-06,
+      "loss": 0.5688,
+      "mean_token_accuracy": 0.8152944445610046,
+      "num_tokens": 334216172.0,
+      "step": 889
+    },
+    {
+      "epoch": 1.3244047619047619,
+      "grad_norm": 2.041603378438486,
+      "learning_rate": 9.194484113866313e-06,
+      "loss": 0.5532,
+      "mean_token_accuracy": 0.8197807669639587,
+      "num_tokens": 334595036.0,
+      "step": 890
+    },
+    {
+      "epoch": 1.3258928571428572,
+      "grad_norm": 2.0915594097852725,
+      "learning_rate": 9.191654570832496e-06,
+      "loss": 0.5313,
+      "mean_token_accuracy": 0.8288635611534119,
+      "num_tokens": 334970921.0,
+      "step": 891
+    },
+    {
+      "epoch": 1.3273809523809523,
+      "grad_norm": 2.119690011472503,
+      "learning_rate": 9.188820503808862e-06,
+      "loss": 0.5862,
+      "mean_token_accuracy": 0.8167128562927246,
+      "num_tokens": 335335756.0,
+      "step": 892
+    },
+    {
+      "epoch": 1.3288690476190477,
+      "grad_norm": 2.0631534410670027,
+      "learning_rate": 9.185981915854182e-06,
+      "loss": 0.5366,
+      "mean_token_accuracy": 0.8294718861579895,
+      "num_tokens": 335714451.0,
+      "step": 893
+    },
+    {
+      "epoch": 1.3303571428571428,
+      "grad_norm": 2.238785801026008,
+      "learning_rate": 9.1831388100321e-06,
+      "loss": 0.6134,
+      "mean_token_accuracy": 0.8080028295516968,
+      "num_tokens": 336091788.0,
+      "step": 894
+    },
+    {
+      "epoch": 1.3318452380952381,
+      "grad_norm": 2.264734569137033,
+      "learning_rate": 9.180291189411134e-06,
+      "loss": 0.5815,
+      "mean_token_accuracy": 0.8221818804740906,
+      "num_tokens": 336469463.0,
+      "step": 895
+    },
+    {
+      "epoch": 1.3333333333333333,
+      "grad_norm": 1.9808825002887371,
+      "learning_rate": 9.177439057064684e-06,
+      "loss": 0.5356,
+      "mean_token_accuracy": 0.8297649025917053,
+      "num_tokens": 336848331.0,
+      "step": 896
+    },
+    {
+      "epoch": 1.3348214285714286,
+      "grad_norm": 1.9464840234256502,
+      "learning_rate": 9.174582416071008e-06,
+      "loss": 0.5442,
+      "mean_token_accuracy": 0.8137832283973694,
+      "num_tokens": 337227699.0,
+      "step": 897
+    },
+    {
+      "epoch": 1.3363095238095237,
+      "grad_norm": 2.1750473888411785,
+      "learning_rate": 9.171721269513237e-06,
+      "loss": 0.6058,
+      "mean_token_accuracy": 0.8058117032051086,
+      "num_tokens": 337606119.0,
+      "step": 898
+    },
+    {
+      "epoch": 1.337797619047619,
+      "grad_norm": 1.9748382327326937,
+      "learning_rate": 9.168855620479367e-06,
+      "loss": 0.5779,
+      "mean_token_accuracy": 0.8126193284988403,
+      "num_tokens": 337983232.0,
+      "step": 899
+    },
+    {
+      "epoch": 1.3392857142857144,
+      "grad_norm": 2.2924780152522004,
+      "learning_rate": 9.165985472062245e-06,
+      "loss": 0.6407,
+      "mean_token_accuracy": 0.8017559051513672,
+      "num_tokens": 338361483.0,
+      "step": 900
+    },
+    {
+      "epoch": 1.3407738095238095,
+      "grad_norm": 1.9955672611641067,
+      "learning_rate": 9.163110827359583e-06,
+      "loss": 0.5168,
+      "mean_token_accuracy": 0.8341668248176575,
+      "num_tokens": 338727627.0,
+      "step": 901
+    },
+    {
+      "epoch": 1.3422619047619047,
+      "grad_norm": 1.985842203281484,
+      "learning_rate": 9.16023168947394e-06,
+      "loss": 0.6145,
+      "mean_token_accuracy": 0.8112062811851501,
+      "num_tokens": 339104963.0,
+      "step": 902
+    },
+    {
+      "epoch": 1.34375,
+      "grad_norm": 2.0858729381841745,
+      "learning_rate": 9.157348061512728e-06,
+      "loss": 0.5805,
+      "mean_token_accuracy": 0.8119556307792664,
+      "num_tokens": 339483200.0,
+      "step": 903
+    },
+    {
+      "epoch": 1.3452380952380953,
+      "grad_norm": 2.1805775802710596,
+      "learning_rate": 9.154459946588199e-06,
+      "loss": 0.5317,
+      "mean_token_accuracy": 0.8278106451034546,
+      "num_tokens": 339859973.0,
+      "step": 904
+    },
+    {
+      "epoch": 1.3467261904761905,
+      "grad_norm": 1.9649569601010906,
+      "learning_rate": 9.151567347817458e-06,
+      "loss": 0.5605,
+      "mean_token_accuracy": 0.8156877160072327,
+      "num_tokens": 340236952.0,
+      "step": 905
+    },
+    {
+      "epoch": 1.3482142857142856,
+      "grad_norm": 2.180347230876114,
+      "learning_rate": 9.148670268322439e-06,
+      "loss": 0.5797,
+      "mean_token_accuracy": 0.812561571598053,
+      "num_tokens": 340614129.0,
+      "step": 906
+    },
+    {
+      "epoch": 1.349702380952381,
+      "grad_norm": 2.0357007751658642,
+      "learning_rate": 9.145768711229917e-06,
+      "loss": 0.5749,
+      "mean_token_accuracy": 0.817155122756958,
+      "num_tokens": 340991946.0,
+      "step": 907
+    },
+    {
+      "epoch": 1.3511904761904763,
+      "grad_norm": 2.1763331406165443,
+      "learning_rate": 9.1428626796715e-06,
+      "loss": 0.5822,
+      "mean_token_accuracy": 0.8160229325294495,
+      "num_tokens": 341360486.0,
+      "step": 908
+    },
+    {
+      "epoch": 1.3526785714285714,
+      "grad_norm": 1.9296790373565793,
+      "learning_rate": 9.139952176783626e-06,
+      "loss": 0.5263,
+      "mean_token_accuracy": 0.8255400061607361,
+      "num_tokens": 341739442.0,
+      "step": 909
+    },
+    {
+      "epoch": 1.3541666666666667,
+      "grad_norm": 1.7909628231960504,
+      "learning_rate": 9.137037205707552e-06,
+      "loss": 0.551,
+      "mean_token_accuracy": 0.8210278749465942,
+      "num_tokens": 342119022.0,
+      "step": 910
+    },
+    {
+      "epoch": 1.3556547619047619,
+      "grad_norm": 2.2920039863256285,
+      "learning_rate": 9.134117769589368e-06,
+      "loss": 0.6161,
+      "mean_token_accuracy": 0.8082638382911682,
+      "num_tokens": 342495677.0,
+      "step": 911
+    },
+    {
+      "epoch": 1.3571428571428572,
+      "grad_norm": 2.433180196637211,
+      "learning_rate": 9.131193871579975e-06,
+      "loss": 0.6579,
+      "mean_token_accuracy": 0.7987515330314636,
+      "num_tokens": 342873294.0,
+      "step": 912
+    },
+    {
+      "epoch": 1.3586309523809523,
+      "grad_norm": 2.111827400818793,
+      "learning_rate": 9.128265514835092e-06,
+      "loss": 0.5051,
+      "mean_token_accuracy": 0.8354260325431824,
+      "num_tokens": 343243031.0,
+      "step": 913
+    },
+    {
+      "epoch": 1.3601190476190477,
+      "grad_norm": 2.116122012272247,
+      "learning_rate": 9.12533270251525e-06,
+      "loss": 0.5908,
+      "mean_token_accuracy": 0.8111584186553955,
+      "num_tokens": 343622118.0,
+      "step": 914
+    },
+    {
+      "epoch": 1.3616071428571428,
+      "grad_norm": 2.3949365994686147,
+      "learning_rate": 9.122395437785791e-06,
+      "loss": 0.6692,
+      "mean_token_accuracy": 0.7898256182670593,
+      "num_tokens": 343999656.0,
+      "step": 915
+    },
+    {
+      "epoch": 1.3630952380952381,
+      "grad_norm": 1.9885457831245044,
+      "learning_rate": 9.119453723816858e-06,
+      "loss": 0.5896,
+      "mean_token_accuracy": 0.8155817985534668,
+      "num_tokens": 344378411.0,
+      "step": 916
+    },
+    {
+      "epoch": 1.3645833333333333,
+      "grad_norm": 2.1220601503827714,
+      "learning_rate": 9.116507563783402e-06,
+      "loss": 0.6467,
+      "mean_token_accuracy": 0.7946654558181763,
+      "num_tokens": 344755773.0,
+      "step": 917
+    },
+    {
+      "epoch": 1.3660714285714286,
+      "grad_norm": 1.8827403262778473,
+      "learning_rate": 9.113556960865168e-06,
+      "loss": 0.5357,
+      "mean_token_accuracy": 0.82438725233078,
+      "num_tokens": 345134174.0,
+      "step": 918
+    },
+    {
+      "epoch": 1.3675595238095237,
+      "grad_norm": 1.9556294989341199,
+      "learning_rate": 9.110601918246695e-06,
+      "loss": 0.5807,
+      "mean_token_accuracy": 0.8110441565513611,
+      "num_tokens": 345513843.0,
+      "step": 919
+    },
+    {
+      "epoch": 1.369047619047619,
+      "grad_norm": 2.0902603636277823,
+      "learning_rate": 9.107642439117322e-06,
+      "loss": 0.6248,
+      "mean_token_accuracy": 0.8022447824478149,
+      "num_tokens": 345892302.0,
+      "step": 920
+    },
+    {
+      "epoch": 1.3705357142857144,
+      "grad_norm": 2.0544010921158558,
+      "learning_rate": 9.104678526671162e-06,
+      "loss": 0.6163,
+      "mean_token_accuracy": 0.8112020492553711,
+      "num_tokens": 346271006.0,
+      "step": 921
+    },
+    {
+      "epoch": 1.3720238095238095,
+      "grad_norm": 2.06528573275091,
+      "learning_rate": 9.10171018410713e-06,
+      "loss": 0.5542,
+      "mean_token_accuracy": 0.81886225938797,
+      "num_tokens": 346649415.0,
+      "step": 922
+    },
+    {
+      "epoch": 1.3735119047619047,
+      "grad_norm": 2.128307113034042,
+      "learning_rate": 9.098737414628905e-06,
+      "loss": 0.5967,
+      "mean_token_accuracy": 0.8042250871658325,
+      "num_tokens": 347027018.0,
+      "step": 923
+    },
+    {
+      "epoch": 1.375,
+      "grad_norm": 2.0604607825161643,
+      "learning_rate": 9.09576022144496e-06,
+      "loss": 0.5653,
+      "mean_token_accuracy": 0.8214524984359741,
+      "num_tokens": 347390407.0,
+      "step": 924
+    },
+    {
+      "epoch": 1.3764880952380953,
+      "grad_norm": 2.0587544639916575,
+      "learning_rate": 9.09277860776853e-06,
+      "loss": 0.6029,
+      "mean_token_accuracy": 0.802998423576355,
+      "num_tokens": 347769105.0,
+      "step": 925
+    },
+    {
+      "epoch": 1.3779761904761905,
+      "grad_norm": 2.0950430290715083,
+      "learning_rate": 9.089792576817628e-06,
+      "loss": 0.6036,
+      "mean_token_accuracy": 0.807928204536438,
+      "num_tokens": 348146286.0,
+      "step": 926
+    },
+    {
+      "epoch": 1.3794642857142856,
+      "grad_norm": 2.053536303387354,
+      "learning_rate": 9.086802131815033e-06,
+      "loss": 0.6039,
+      "mean_token_accuracy": 0.8103991746902466,
+      "num_tokens": 348511934.0,
+      "step": 927
+    },
+    {
+      "epoch": 1.380952380952381,
+      "grad_norm": 2.054391299906357,
+      "learning_rate": 9.083807275988285e-06,
+      "loss": 0.5731,
+      "mean_token_accuracy": 0.8136516213417053,
+      "num_tokens": 348889150.0,
+      "step": 928
+    },
+    {
+      "epoch": 1.3824404761904763,
+      "grad_norm": 2.2255969642407374,
+      "learning_rate": 9.080808012569687e-06,
+      "loss": 0.592,
+      "mean_token_accuracy": 0.8094304203987122,
+      "num_tokens": 349265997.0,
+      "step": 929
+    },
+    {
+      "epoch": 1.3839285714285714,
+      "grad_norm": 2.081022426783968,
+      "learning_rate": 9.077804344796302e-06,
+      "loss": 0.6042,
+      "mean_token_accuracy": 0.8107444643974304,
+      "num_tokens": 349631797.0,
+      "step": 930
+    },
+    {
+      "epoch": 1.3854166666666667,
+      "grad_norm": 2.150713376548924,
+      "learning_rate": 9.074796275909941e-06,
+      "loss": 0.5687,
+      "mean_token_accuracy": 0.8137410879135132,
+      "num_tokens": 350008053.0,
+      "step": 931
+    },
+    {
+      "epoch": 1.3869047619047619,
+      "grad_norm": 1.8231478756841673,
+      "learning_rate": 9.071783809157168e-06,
+      "loss": 0.561,
+      "mean_token_accuracy": 0.8141212463378906,
+      "num_tokens": 350387592.0,
+      "step": 932
+    },
+    {
+      "epoch": 1.3883928571428572,
+      "grad_norm": 1.9132598363769375,
+      "learning_rate": 9.068766947789292e-06,
+      "loss": 0.554,
+      "mean_token_accuracy": 0.8212330341339111,
+      "num_tokens": 350765867.0,
+      "step": 933
+    },
+    {
+      "epoch": 1.3898809523809523,
+      "grad_norm": 2.1338948251498415,
+      "learning_rate": 9.065745695062368e-06,
+      "loss": 0.6126,
+      "mean_token_accuracy": 0.8047276735305786,
+      "num_tokens": 351144122.0,
+      "step": 934
+    },
+    {
+      "epoch": 1.3913690476190477,
+      "grad_norm": 1.9005330304517485,
+      "learning_rate": 9.06272005423719e-06,
+      "loss": 0.5705,
+      "mean_token_accuracy": 0.817179262638092,
+      "num_tokens": 351524501.0,
+      "step": 935
+    },
+    {
+      "epoch": 1.3928571428571428,
+      "grad_norm": 1.9662267332841061,
+      "learning_rate": 9.059690028579285e-06,
+      "loss": 0.5907,
+      "mean_token_accuracy": 0.8109153509140015,
+      "num_tokens": 351903893.0,
+      "step": 936
+    },
+    {
+      "epoch": 1.3943452380952381,
+      "grad_norm": 2.067329867261314,
+      "learning_rate": 9.05665562135891e-06,
+      "loss": 0.585,
+      "mean_token_accuracy": 0.8114866614341736,
+      "num_tokens": 352281632.0,
+      "step": 937
+    },
+    {
+      "epoch": 1.3958333333333333,
+      "grad_norm": 2.188903154123561,
+      "learning_rate": 9.053616835851062e-06,
+      "loss": 0.6019,
+      "mean_token_accuracy": 0.8119270205497742,
+      "num_tokens": 352658687.0,
+      "step": 938
+    },
+    {
+      "epoch": 1.3973214285714286,
+      "grad_norm": 1.936791232939388,
+      "learning_rate": 9.050573675335453e-06,
+      "loss": 0.5607,
+      "mean_token_accuracy": 0.8207476735115051,
+      "num_tokens": 353037535.0,
+      "step": 939
+    },
+    {
+      "epoch": 1.3988095238095237,
+      "grad_norm": 2.0027682643007227,
+      "learning_rate": 9.047526143096522e-06,
+      "loss": 0.6118,
+      "mean_token_accuracy": 0.8092567920684814,
+      "num_tokens": 353401743.0,
+      "step": 940
+    },
+    {
+      "epoch": 1.400297619047619,
+      "grad_norm": 1.9487143491364562,
+      "learning_rate": 9.044474242423422e-06,
+      "loss": 0.507,
+      "mean_token_accuracy": 0.8319924473762512,
+      "num_tokens": 353780568.0,
+      "step": 941
+    },
+    {
+      "epoch": 1.4017857142857144,
+      "grad_norm": 2.0064298859551988,
+      "learning_rate": 9.041417976610028e-06,
+      "loss": 0.5649,
+      "mean_token_accuracy": 0.8169262409210205,
+      "num_tokens": 354141414.0,
+      "step": 942
+    },
+    {
+      "epoch": 1.4032738095238095,
+      "grad_norm": 1.896723036957113,
+      "learning_rate": 9.038357348954917e-06,
+      "loss": 0.5737,
+      "mean_token_accuracy": 0.8161155581474304,
+      "num_tokens": 354521030.0,
+      "step": 943
+    },
+    {
+      "epoch": 1.4047619047619047,
+      "grad_norm": 2.0615840663823994,
+      "learning_rate": 9.035292362761382e-06,
+      "loss": 0.5984,
+      "mean_token_accuracy": 0.8066644072532654,
+      "num_tokens": 354899089.0,
+      "step": 944
+    },
+    {
+      "epoch": 1.40625,
+      "grad_norm": 2.089826519106595,
+      "learning_rate": 9.032223021337415e-06,
+      "loss": 0.5652,
+      "mean_token_accuracy": 0.8180536031723022,
+      "num_tokens": 355277093.0,
+      "step": 945
+    },
+    {
+      "epoch": 1.4077380952380953,
+      "grad_norm": 2.106822935917691,
+      "learning_rate": 9.029149327995708e-06,
+      "loss": 0.6043,
+      "mean_token_accuracy": 0.807967483997345,
+      "num_tokens": 355655320.0,
+      "step": 946
+    },
+    {
+      "epoch": 1.4092261904761905,
+      "grad_norm": 2.1959020957310353,
+      "learning_rate": 9.026071286053657e-06,
+      "loss": 0.6442,
+      "mean_token_accuracy": 0.7919508218765259,
+      "num_tokens": 356034113.0,
+      "step": 947
+    },
+    {
+      "epoch": 1.4107142857142856,
+      "grad_norm": 2.075903389466246,
+      "learning_rate": 9.022988898833342e-06,
+      "loss": 0.5799,
+      "mean_token_accuracy": 0.814646303653717,
+      "num_tokens": 356412364.0,
+      "step": 948
+    },
+    {
+      "epoch": 1.412202380952381,
+      "grad_norm": 1.8758680394636567,
+      "learning_rate": 9.019902169661542e-06,
+      "loss": 0.524,
+      "mean_token_accuracy": 0.8251973986625671,
+      "num_tokens": 356791401.0,
+      "step": 949
+    },
+    {
+      "epoch": 1.4136904761904763,
+      "grad_norm": 2.168539335554128,
+      "learning_rate": 9.01681110186971e-06,
+      "loss": 0.5973,
+      "mean_token_accuracy": 0.803349494934082,
+      "num_tokens": 357169488.0,
+      "step": 950
+    },
+    {
+      "epoch": 1.4151785714285714,
+      "grad_norm": 2.098352324929902,
+      "learning_rate": 9.013715698793997e-06,
+      "loss": 0.6244,
+      "mean_token_accuracy": 0.8010642528533936,
+      "num_tokens": 357547194.0,
+      "step": 951
+    },
+    {
+      "epoch": 1.4166666666666667,
+      "grad_norm": 1.947228377051347,
+      "learning_rate": 9.01061596377522e-06,
+      "loss": 0.5825,
+      "mean_token_accuracy": 0.8147368431091309,
+      "num_tokens": 357912026.0,
+      "step": 952
+    },
+    {
+      "epoch": 1.4181547619047619,
+      "grad_norm": 1.9411671257750542,
+      "learning_rate": 9.007511900158879e-06,
+      "loss": 0.6288,
+      "mean_token_accuracy": 0.7998065948486328,
+      "num_tokens": 358278180.0,
+      "step": 953
+    },
+    {
+      "epoch": 1.4196428571428572,
+      "grad_norm": 2.2213691546375207,
+      "learning_rate": 9.004403511295141e-06,
+      "loss": 0.6909,
+      "mean_token_accuracy": 0.7880409955978394,
+      "num_tokens": 358655361.0,
+      "step": 954
+    },
+    {
+      "epoch": 1.4211309523809523,
+      "grad_norm": 2.0277350779179732,
+      "learning_rate": 9.001290800538844e-06,
+      "loss": 0.6063,
+      "mean_token_accuracy": 0.8076726794242859,
+      "num_tokens": 359034798.0,
+      "step": 955
+    },
+    {
+      "epoch": 1.4226190476190477,
+      "grad_norm": 1.968574059286171,
+      "learning_rate": 8.998173771249491e-06,
+      "loss": 0.5646,
+      "mean_token_accuracy": 0.8116809725761414,
+      "num_tokens": 359413716.0,
+      "step": 956
+    },
+    {
+      "epoch": 1.4241071428571428,
+      "grad_norm": 2.057750463130908,
+      "learning_rate": 8.995052426791247e-06,
+      "loss": 0.5975,
+      "mean_token_accuracy": 0.8081323504447937,
+      "num_tokens": 359791910.0,
+      "step": 957
+    },
+    {
+      "epoch": 1.4255952380952381,
+      "grad_norm": 2.064018746827887,
+      "learning_rate": 8.991926770532927e-06,
+      "loss": 0.5471,
+      "mean_token_accuracy": 0.8193418979644775,
+      "num_tokens": 360169393.0,
+      "step": 958
+    },
+    {
+      "epoch": 1.4270833333333333,
+      "grad_norm": 2.0205252966328424,
+      "learning_rate": 8.988796805848008e-06,
+      "loss": 0.5837,
+      "mean_token_accuracy": 0.8186607360839844,
+      "num_tokens": 360547729.0,
+      "step": 959
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 2.2211987724491697,
+      "learning_rate": 8.985662536114614e-06,
+      "loss": 0.5846,
+      "mean_token_accuracy": 0.8127372860908508,
+      "num_tokens": 360924929.0,
+      "step": 960
+    },
+    {
+      "epoch": 1.4300595238095237,
+      "grad_norm": 2.0715520321361076,
+      "learning_rate": 8.982523964715513e-06,
+      "loss": 0.5848,
+      "mean_token_accuracy": 0.8124944567680359,
+      "num_tokens": 361303064.0,
+      "step": 961
+    },
+    {
+      "epoch": 1.431547619047619,
+      "grad_norm": 1.9238766504894993,
+      "learning_rate": 8.97938109503812e-06,
+      "loss": 0.5445,
+      "mean_token_accuracy": 0.830237627029419,
+      "num_tokens": 361668020.0,
+      "step": 962
+    },
+    {
+      "epoch": 1.4330357142857144,
+      "grad_norm": 1.966321178399939,
+      "learning_rate": 8.976233930474487e-06,
+      "loss": 0.5397,
+      "mean_token_accuracy": 0.8209401965141296,
+      "num_tokens": 362032559.0,
+      "step": 963
+    },
+    {
+      "epoch": 1.4345238095238095,
+      "grad_norm": 2.018389860472175,
+      "learning_rate": 8.973082474421302e-06,
+      "loss": 0.5663,
+      "mean_token_accuracy": 0.8176785707473755,
+      "num_tokens": 362410225.0,
+      "step": 964
+    },
+    {
+      "epoch": 1.4360119047619047,
+      "grad_norm": 1.9638773483318201,
+      "learning_rate": 8.969926730279884e-06,
+      "loss": 0.6004,
+      "mean_token_accuracy": 0.8081903457641602,
+      "num_tokens": 362789568.0,
+      "step": 965
+    },
+    {
+      "epoch": 1.4375,
+      "grad_norm": 2.175253072554456,
+      "learning_rate": 8.966766701456177e-06,
+      "loss": 0.6256,
+      "mean_token_accuracy": 0.8042507171630859,
+      "num_tokens": 363168617.0,
+      "step": 966
+    },
+    {
+      "epoch": 1.4389880952380953,
+      "grad_norm": 2.1229357126720148,
+      "learning_rate": 8.963602391360755e-06,
+      "loss": 0.5499,
+      "mean_token_accuracy": 0.8214806318283081,
+      "num_tokens": 363546841.0,
+      "step": 967
+    },
+    {
+      "epoch": 1.4404761904761905,
+      "grad_norm": 2.1188715928433544,
+      "learning_rate": 8.960433803408813e-06,
+      "loss": 0.5745,
+      "mean_token_accuracy": 0.8163644671440125,
+      "num_tokens": 363924769.0,
+      "step": 968
+    },
+    {
+      "epoch": 1.4419642857142856,
+      "grad_norm": 2.0213499193914486,
+      "learning_rate": 8.957260941020155e-06,
+      "loss": 0.5943,
+      "mean_token_accuracy": 0.8062469363212585,
+      "num_tokens": 364303241.0,
+      "step": 969
+    },
+    {
+      "epoch": 1.443452380952381,
+      "grad_norm": 2.3742834844772585,
+      "learning_rate": 8.954083807619208e-06,
+      "loss": 0.5889,
+      "mean_token_accuracy": 0.8125795722007751,
+      "num_tokens": 364669723.0,
+      "step": 970
+    },
+    {
+      "epoch": 1.4449404761904763,
+      "grad_norm": 2.3102370536135433,
+      "learning_rate": 8.950902406635002e-06,
+      "loss": 0.6172,
+      "mean_token_accuracy": 0.8085501790046692,
+      "num_tokens": 365046386.0,
+      "step": 971
+    },
+    {
+      "epoch": 1.4464285714285714,
+      "grad_norm": 1.920719509807201,
+      "learning_rate": 8.947716741501178e-06,
+      "loss": 0.5141,
+      "mean_token_accuracy": 0.8310546875,
+      "num_tokens": 365424761.0,
+      "step": 972
+    },
+    {
+      "epoch": 1.4479166666666667,
+      "grad_norm": 2.067036805944019,
+      "learning_rate": 8.944526815655974e-06,
+      "loss": 0.5136,
+      "mean_token_accuracy": 0.8350096940994263,
+      "num_tokens": 365802136.0,
+      "step": 973
+    },
+    {
+      "epoch": 1.4494047619047619,
+      "grad_norm": 2.1072234548388997,
+      "learning_rate": 8.941332632542227e-06,
+      "loss": 0.5616,
+      "mean_token_accuracy": 0.8155573606491089,
+      "num_tokens": 366180363.0,
+      "step": 974
+    },
+    {
+      "epoch": 1.4508928571428572,
+      "grad_norm": 2.057016386166144,
+      "learning_rate": 8.938134195607378e-06,
+      "loss": 0.5424,
+      "mean_token_accuracy": 0.8245453834533691,
+      "num_tokens": 366548998.0,
+      "step": 975
+    },
+    {
+      "epoch": 1.4523809523809523,
+      "grad_norm": 1.973709335455194,
+      "learning_rate": 8.934931508303446e-06,
+      "loss": 0.5703,
+      "mean_token_accuracy": 0.8144937753677368,
+      "num_tokens": 366926915.0,
+      "step": 976
+    },
+    {
+      "epoch": 1.4538690476190477,
+      "grad_norm": 1.958838890535617,
+      "learning_rate": 8.931724574087044e-06,
+      "loss": 0.5702,
+      "mean_token_accuracy": 0.8136070966720581,
+      "num_tokens": 367305649.0,
+      "step": 977
+    },
+    {
+      "epoch": 1.4553571428571428,
+      "grad_norm": 2.101956360437897,
+      "learning_rate": 8.92851339641937e-06,
+      "loss": 0.5664,
+      "mean_token_accuracy": 0.8204391598701477,
+      "num_tokens": 367675385.0,
+      "step": 978
+    },
+    {
+      "epoch": 1.4568452380952381,
+      "grad_norm": 1.927792724588282,
+      "learning_rate": 8.925297978766197e-06,
+      "loss": 0.5612,
+      "mean_token_accuracy": 0.8144436478614807,
+      "num_tokens": 368053931.0,
+      "step": 979
+    },
+    {
+      "epoch": 1.4583333333333333,
+      "grad_norm": 2.087558764016022,
+      "learning_rate": 8.92207832459788e-06,
+      "loss": 0.5792,
+      "mean_token_accuracy": 0.8152221441268921,
+      "num_tokens": 368431766.0,
+      "step": 980
+    },
+    {
+      "epoch": 1.4598214285714286,
+      "grad_norm": 2.0903964615132455,
+      "learning_rate": 8.918854437389341e-06,
+      "loss": 0.5848,
+      "mean_token_accuracy": 0.8201613426208496,
+      "num_tokens": 368810042.0,
+      "step": 981
+    },
+    {
+      "epoch": 1.4613095238095237,
+      "grad_norm": 2.0970846364668105,
+      "learning_rate": 8.915626320620074e-06,
+      "loss": 0.5911,
+      "mean_token_accuracy": 0.8131499886512756,
+      "num_tokens": 369188322.0,
+      "step": 982
+    },
+    {
+      "epoch": 1.462797619047619,
+      "grad_norm": 2.121353360301938,
+      "learning_rate": 8.912393977774138e-06,
+      "loss": 0.5621,
+      "mean_token_accuracy": 0.8197847008705139,
+      "num_tokens": 369565758.0,
+      "step": 983
+    },
+    {
+      "epoch": 1.4642857142857144,
+      "grad_norm": 2.291006144364,
+      "learning_rate": 8.90915741234015e-06,
+      "loss": 0.6317,
+      "mean_token_accuracy": 0.8010531067848206,
+      "num_tokens": 369943195.0,
+      "step": 984
+    },
+    {
+      "epoch": 1.4657738095238095,
+      "grad_norm": 2.1062165897734473,
+      "learning_rate": 8.905916627811287e-06,
+      "loss": 0.6237,
+      "mean_token_accuracy": 0.8022703528404236,
+      "num_tokens": 370320532.0,
+      "step": 985
+    },
+    {
+      "epoch": 1.4672619047619047,
+      "grad_norm": 2.0490207599308143,
+      "learning_rate": 8.902671627685281e-06,
+      "loss": 0.5763,
+      "mean_token_accuracy": 0.8089155554771423,
+      "num_tokens": 370698916.0,
+      "step": 986
+    },
+    {
+      "epoch": 1.46875,
+      "grad_norm": 2.185498361161548,
+      "learning_rate": 8.899422415464409e-06,
+      "loss": 0.6554,
+      "mean_token_accuracy": 0.790718674659729,
+      "num_tokens": 371076991.0,
+      "step": 987
+    },
+    {
+      "epoch": 1.4702380952380953,
+      "grad_norm": 1.930342219120099,
+      "learning_rate": 8.8961689946555e-06,
+      "loss": 0.5564,
+      "mean_token_accuracy": 0.8186808824539185,
+      "num_tokens": 371444046.0,
+      "step": 988
+    },
+    {
+      "epoch": 1.4717261904761905,
+      "grad_norm": 2.0037601842553214,
+      "learning_rate": 8.89291136876992e-06,
+      "loss": 0.5916,
+      "mean_token_accuracy": 0.8028879761695862,
+      "num_tokens": 371822758.0,
+      "step": 989
+    },
+    {
+      "epoch": 1.4732142857142856,
+      "grad_norm": 1.8425962394514293,
+      "learning_rate": 8.889649541323575e-06,
+      "loss": 0.5218,
+      "mean_token_accuracy": 0.8315027952194214,
+      "num_tokens": 372189622.0,
+      "step": 990
+    },
+    {
+      "epoch": 1.474702380952381,
+      "grad_norm": 2.1754577246092164,
+      "learning_rate": 8.886383515836909e-06,
+      "loss": 0.5391,
+      "mean_token_accuracy": 0.8165239095687866,
+      "num_tokens": 372567579.0,
+      "step": 991
+    },
+    {
+      "epoch": 1.4761904761904763,
+      "grad_norm": 1.9709575699652038,
+      "learning_rate": 8.883113295834893e-06,
+      "loss": 0.5537,
+      "mean_token_accuracy": 0.8221470713615417,
+      "num_tokens": 372940729.0,
+      "step": 992
+    },
+    {
+      "epoch": 1.4776785714285714,
+      "grad_norm": 1.9461069628692482,
+      "learning_rate": 8.879838884847025e-06,
+      "loss": 0.5796,
+      "mean_token_accuracy": 0.8093616366386414,
+      "num_tokens": 373318608.0,
+      "step": 993
+    },
+    {
+      "epoch": 1.4791666666666667,
+      "grad_norm": 2.0289337862434174,
+      "learning_rate": 8.87656028640733e-06,
+      "loss": 0.5553,
+      "mean_token_accuracy": 0.8185749053955078,
+      "num_tokens": 373697367.0,
+      "step": 994
+    },
+    {
+      "epoch": 1.4806547619047619,
+      "grad_norm": 2.1793913246734546,
+      "learning_rate": 8.873277504054349e-06,
+      "loss": 0.6351,
+      "mean_token_accuracy": 0.7958551049232483,
+      "num_tokens": 374075247.0,
+      "step": 995
+    },
+    {
+      "epoch": 1.4821428571428572,
+      "grad_norm": 1.8738815115396452,
+      "learning_rate": 8.869990541331137e-06,
+      "loss": 0.5764,
+      "mean_token_accuracy": 0.8196361064910889,
+      "num_tokens": 374451639.0,
+      "step": 996
+    },
+    {
+      "epoch": 1.4836309523809523,
+      "grad_norm": 2.138083201707517,
+      "learning_rate": 8.866699401785269e-06,
+      "loss": 0.5402,
+      "mean_token_accuracy": 0.8210342526435852,
+      "num_tokens": 374829501.0,
+      "step": 997
+    },
+    {
+      "epoch": 1.4851190476190477,
+      "grad_norm": 2.4371773960319723,
+      "learning_rate": 8.863404088968819e-06,
+      "loss": 0.5761,
+      "mean_token_accuracy": 0.8192372918128967,
+      "num_tokens": 375208916.0,
+      "step": 998
+    },
+    {
+      "epoch": 1.4866071428571428,
+      "grad_norm": 2.0440520521325185,
+      "learning_rate": 8.86010460643837e-06,
+      "loss": 0.6223,
+      "mean_token_accuracy": 0.8013082146644592,
+      "num_tokens": 375587822.0,
+      "step": 999
+    },
+    {
+      "epoch": 1.4880952380952381,
+      "grad_norm": 1.9470179748587526,
+      "learning_rate": 8.856800957755e-06,
+      "loss": 0.5512,
+      "mean_token_accuracy": 0.8196926712989807,
+      "num_tokens": 375966934.0,
+      "step": 1000
+    },
+    {
+      "epoch": 1.4895833333333333,
+      "grad_norm": 1.9649720745264034,
+      "learning_rate": 8.85349314648429e-06,
+      "loss": 0.5745,
+      "mean_token_accuracy": 0.8065350651741028,
+      "num_tokens": 376343564.0,
+      "step": 1001
+    },
+    {
+      "epoch": 1.4910714285714286,
+      "grad_norm": 1.8060711774635145,
+      "learning_rate": 8.850181176196316e-06,
+      "loss": 0.5186,
+      "mean_token_accuracy": 0.8298832774162292,
+      "num_tokens": 376710167.0,
+      "step": 1002
+    },
+    {
+      "epoch": 1.4925595238095237,
+      "grad_norm": 1.9120016468501362,
+      "learning_rate": 8.84686505046563e-06,
+      "loss": 0.56,
+      "mean_token_accuracy": 0.8206192851066589,
+      "num_tokens": 377069942.0,
+      "step": 1003
+    },
+    {
+      "epoch": 1.494047619047619,
+      "grad_norm": 2.0967548203593003,
+      "learning_rate": 8.84354477287128e-06,
+      "loss": 0.5649,
+      "mean_token_accuracy": 0.8202626705169678,
+      "num_tokens": 377446858.0,
+      "step": 1004
+    },
+    {
+      "epoch": 1.4955357142857144,
+      "grad_norm": 2.418990304155378,
+      "learning_rate": 8.840220346996793e-06,
+      "loss": 0.6168,
+      "mean_token_accuracy": 0.8060134649276733,
+      "num_tokens": 377823540.0,
+      "step": 1005
+    },
+    {
+      "epoch": 1.4970238095238095,
+      "grad_norm": 1.970796832434918,
+      "learning_rate": 8.83689177643017e-06,
+      "loss": 0.5606,
+      "mean_token_accuracy": 0.8172839283943176,
+      "num_tokens": 378175692.0,
+      "step": 1006
+    },
+    {
+      "epoch": 1.4985119047619047,
+      "grad_norm": 2.064494218528196,
+      "learning_rate": 8.833559064763888e-06,
+      "loss": 0.5683,
+      "mean_token_accuracy": 0.8174393773078918,
+      "num_tokens": 378554537.0,
+      "step": 1007
+    },
+    {
+      "epoch": 1.5,
+      "grad_norm": 2.256322392326293,
+      "learning_rate": 8.83022221559489e-06,
+      "loss": 0.644,
+      "mean_token_accuracy": 0.7972933650016785,
+      "num_tokens": 378931622.0,
+      "step": 1008
+    },
+    {
+      "epoch": 1.5014880952380953,
+      "grad_norm": 1.92957763175785,
+      "learning_rate": 8.826881232524593e-06,
+      "loss": 0.5839,
+      "mean_token_accuracy": 0.8142237663269043,
+      "num_tokens": 379304874.0,
+      "step": 1009
+    },
+    {
+      "epoch": 1.5029761904761905,
+      "grad_norm": 2.1973304297466885,
+      "learning_rate": 8.823536119158865e-06,
+      "loss": 0.636,
+      "mean_token_accuracy": 0.802836000919342,
+      "num_tokens": 379682470.0,
+      "step": 1010
+    },
+    {
+      "epoch": 1.5044642857142856,
+      "grad_norm": 2.174249690911212,
+      "learning_rate": 8.820186879108038e-06,
+      "loss": 0.5556,
+      "mean_token_accuracy": 0.8173628449440002,
+      "num_tokens": 380056777.0,
+      "step": 1011
+    },
+    {
+      "epoch": 1.505952380952381,
+      "grad_norm": 2.0859867792669213,
+      "learning_rate": 8.816833515986896e-06,
+      "loss": 0.5788,
+      "mean_token_accuracy": 0.818721354007721,
+      "num_tokens": 380435051.0,
+      "step": 1012
+    },
+    {
+      "epoch": 1.5074404761904763,
+      "grad_norm": 2.0503925862344965,
+      "learning_rate": 8.813476033414674e-06,
+      "loss": 0.6532,
+      "mean_token_accuracy": 0.795214831829071,
+      "num_tokens": 380813575.0,
+      "step": 1013
+    },
+    {
+      "epoch": 1.5089285714285714,
+      "grad_norm": 2.205159017267369,
+      "learning_rate": 8.810114435015055e-06,
+      "loss": 0.6037,
+      "mean_token_accuracy": 0.809969425201416,
+      "num_tokens": 381191234.0,
+      "step": 1014
+    },
+    {
+      "epoch": 1.5104166666666665,
+      "grad_norm": 2.184674922713484,
+      "learning_rate": 8.806748724416156e-06,
+      "loss": 0.6283,
+      "mean_token_accuracy": 0.8019483685493469,
+      "num_tokens": 381569026.0,
+      "step": 1015
+    },
+    {
+      "epoch": 1.5119047619047619,
+      "grad_norm": 2.0370217505518124,
+      "learning_rate": 8.803378905250544e-06,
+      "loss": 0.5648,
+      "mean_token_accuracy": 0.8220676183700562,
+      "num_tokens": 381947152.0,
+      "step": 1016
+    },
+    {
+      "epoch": 1.5133928571428572,
+      "grad_norm": 1.9415784913043082,
+      "learning_rate": 8.800004981155208e-06,
+      "loss": 0.5755,
+      "mean_token_accuracy": 0.8149347901344299,
+      "num_tokens": 382326183.0,
+      "step": 1017
+    },
+    {
+      "epoch": 1.5148809523809523,
+      "grad_norm": 2.035295496153888,
+      "learning_rate": 8.796626955771579e-06,
+      "loss": 0.5551,
+      "mean_token_accuracy": 0.8251479268074036,
+      "num_tokens": 382703635.0,
+      "step": 1018
+    },
+    {
+      "epoch": 1.5163690476190477,
+      "grad_norm": 1.9152809287479726,
+      "learning_rate": 8.793244832745507e-06,
+      "loss": 0.532,
+      "mean_token_accuracy": 0.8271799087524414,
+      "num_tokens": 383070607.0,
+      "step": 1019
+    },
+    {
+      "epoch": 1.5178571428571428,
+      "grad_norm": 1.9948212259071254,
+      "learning_rate": 8.789858615727266e-06,
+      "loss": 0.5872,
+      "mean_token_accuracy": 0.812663733959198,
+      "num_tokens": 383449127.0,
+      "step": 1020
+    },
+    {
+      "epoch": 1.5193452380952381,
+      "grad_norm": 1.8945194089663027,
+      "learning_rate": 8.78646830837155e-06,
+      "loss": 0.5635,
+      "mean_token_accuracy": 0.815504252910614,
+      "num_tokens": 383828855.0,
+      "step": 1021
+    },
+    {
+      "epoch": 1.5208333333333335,
+      "grad_norm": 2.082550110716788,
+      "learning_rate": 8.783073914337465e-06,
+      "loss": 0.6501,
+      "mean_token_accuracy": 0.791801393032074,
+      "num_tokens": 384207415.0,
+      "step": 1022
+    },
+    {
+      "epoch": 1.5223214285714286,
+      "grad_norm": 2.0035754777210877,
+      "learning_rate": 8.779675437288533e-06,
+      "loss": 0.557,
+      "mean_token_accuracy": 0.8228863477706909,
+      "num_tokens": 384585993.0,
+      "step": 1023
+    },
+    {
+      "epoch": 1.5238095238095237,
+      "grad_norm": 2.409552582337914,
+      "learning_rate": 8.776272880892675e-06,
+      "loss": 0.5764,
+      "mean_token_accuracy": 0.8185468912124634,
+      "num_tokens": 384964325.0,
+      "step": 1024
+    },
+    {
+      "epoch": 1.525297619047619,
+      "grad_norm": 1.9319467378832982,
+      "learning_rate": 8.772866248822222e-06,
+      "loss": 0.5905,
+      "mean_token_accuracy": 0.810677707195282,
+      "num_tokens": 385344084.0,
+      "step": 1025
+    },
+    {
+      "epoch": 1.5267857142857144,
+      "grad_norm": 1.970979713406338,
+      "learning_rate": 8.7694555447539e-06,
+      "loss": 0.5775,
+      "mean_token_accuracy": 0.811985194683075,
+      "num_tokens": 385712644.0,
+      "step": 1026
+    },
+    {
+      "epoch": 1.5282738095238095,
+      "grad_norm": 2.1744783205797775,
+      "learning_rate": 8.766040772368831e-06,
+      "loss": 0.5931,
+      "mean_token_accuracy": 0.811234712600708,
+      "num_tokens": 386086615.0,
+      "step": 1027
+    },
+    {
+      "epoch": 1.5297619047619047,
+      "grad_norm": 1.8758628596598586,
+      "learning_rate": 8.762621935352526e-06,
+      "loss": 0.5559,
+      "mean_token_accuracy": 0.8196225166320801,
+      "num_tokens": 386466443.0,
+      "step": 1028
+    },
+    {
+      "epoch": 1.53125,
+      "grad_norm": 1.9403035541528135,
+      "learning_rate": 8.759199037394888e-06,
+      "loss": 0.5714,
+      "mean_token_accuracy": 0.8172135949134827,
+      "num_tokens": 386815508.0,
+      "step": 1029
+    },
+    {
+      "epoch": 1.5327380952380953,
+      "grad_norm": 1.8195144775122047,
+      "learning_rate": 8.755772082190196e-06,
+      "loss": 0.5439,
+      "mean_token_accuracy": 0.8226571083068848,
+      "num_tokens": 387194141.0,
+      "step": 1030
+    },
+    {
+      "epoch": 1.5342261904761905,
+      "grad_norm": 2.0697384582529224,
+      "learning_rate": 8.752341073437112e-06,
+      "loss": 0.5939,
+      "mean_token_accuracy": 0.8105159401893616,
+      "num_tokens": 387565292.0,
+      "step": 1031
+    },
+    {
+      "epoch": 1.5357142857142856,
+      "grad_norm": 1.8422952589066723,
+      "learning_rate": 8.748906014838672e-06,
+      "loss": 0.5864,
+      "mean_token_accuracy": 0.815313458442688,
+      "num_tokens": 387944508.0,
+      "step": 1032
+    },
+    {
+      "epoch": 1.537202380952381,
+      "grad_norm": 2.062443064200462,
+      "learning_rate": 8.745466910102283e-06,
+      "loss": 0.5977,
+      "mean_token_accuracy": 0.8057560324668884,
+      "num_tokens": 388323364.0,
+      "step": 1033
+    },
+    {
+      "epoch": 1.5386904761904763,
+      "grad_norm": 2.1025241431079738,
+      "learning_rate": 8.742023762939723e-06,
+      "loss": 0.5696,
+      "mean_token_accuracy": 0.816728949546814,
+      "num_tokens": 388700334.0,
+      "step": 1034
+    },
+    {
+      "epoch": 1.5401785714285714,
+      "grad_norm": 1.8684830919468352,
+      "learning_rate": 8.738576577067122e-06,
+      "loss": 0.5451,
+      "mean_token_accuracy": 0.8207224607467651,
+      "num_tokens": 389079702.0,
+      "step": 1035
+    },
+    {
+      "epoch": 1.5416666666666665,
+      "grad_norm": 1.9236191487162366,
+      "learning_rate": 8.735125356204982e-06,
+      "loss": 0.528,
+      "mean_token_accuracy": 0.825704038143158,
+      "num_tokens": 389457502.0,
+      "step": 1036
+    },
+    {
+      "epoch": 1.5431547619047619,
+      "grad_norm": 2.162129727939349,
+      "learning_rate": 8.73167010407815e-06,
+      "loss": 0.5808,
+      "mean_token_accuracy": 0.8157687187194824,
+      "num_tokens": 389835732.0,
+      "step": 1037
+    },
+    {
+      "epoch": 1.5446428571428572,
+      "grad_norm": 2.2340301686834785,
+      "learning_rate": 8.728210824415829e-06,
+      "loss": 0.6362,
+      "mean_token_accuracy": 0.7987167835235596,
+      "num_tokens": 390189768.0,
+      "step": 1038
+    },
+    {
+      "epoch": 1.5461309523809523,
+      "grad_norm": 1.948883884868676,
+      "learning_rate": 8.724747520951568e-06,
+      "loss": 0.5369,
+      "mean_token_accuracy": 0.8221677541732788,
+      "num_tokens": 390552144.0,
+      "step": 1039
+    },
+    {
+      "epoch": 1.5476190476190477,
+      "grad_norm": 1.7864106277219751,
+      "learning_rate": 8.721280197423259e-06,
+      "loss": 0.5545,
+      "mean_token_accuracy": 0.8283538818359375,
+      "num_tokens": 390934379.0,
+      "step": 1040
+    },
+    {
+      "epoch": 1.5491071428571428,
+      "grad_norm": 2.0183420366515707,
+      "learning_rate": 8.71780885757313e-06,
+      "loss": 0.5592,
+      "mean_token_accuracy": 0.820923388004303,
+      "num_tokens": 391298647.0,
+      "step": 1041
+    },
+    {
+      "epoch": 1.5505952380952381,
+      "grad_norm": 2.23238565983906,
+      "learning_rate": 8.71433350514775e-06,
+      "loss": 0.5811,
+      "mean_token_accuracy": 0.816837728023529,
+      "num_tokens": 391675380.0,
+      "step": 1042
+    },
+    {
+      "epoch": 1.5520833333333335,
+      "grad_norm": 1.9238913431603986,
+      "learning_rate": 8.710854143898008e-06,
+      "loss": 0.5407,
+      "mean_token_accuracy": 0.822956383228302,
+      "num_tokens": 392054149.0,
+      "step": 1043
+    },
+    {
+      "epoch": 1.5535714285714286,
+      "grad_norm": 2.0552658323328874,
+      "learning_rate": 8.707370777579134e-06,
+      "loss": 0.6174,
+      "mean_token_accuracy": 0.8113636374473572,
+      "num_tokens": 392431592.0,
+      "step": 1044
+    },
+    {
+      "epoch": 1.5550595238095237,
+      "grad_norm": 1.9731035527943999,
+      "learning_rate": 8.703883409950668e-06,
+      "loss": 0.5603,
+      "mean_token_accuracy": 0.814519464969635,
+      "num_tokens": 392810719.0,
+      "step": 1045
+    },
+    {
+      "epoch": 1.556547619047619,
+      "grad_norm": 2.0080279828470817,
+      "learning_rate": 8.700392044776474e-06,
+      "loss": 0.5693,
+      "mean_token_accuracy": 0.8114112615585327,
+      "num_tokens": 393176920.0,
+      "step": 1046
+    },
+    {
+      "epoch": 1.5580357142857144,
+      "grad_norm": 2.124891754557483,
+      "learning_rate": 8.69689668582473e-06,
+      "loss": 0.6022,
+      "mean_token_accuracy": 0.8103915452957153,
+      "num_tokens": 393554055.0,
+      "step": 1047
+    },
+    {
+      "epoch": 1.5595238095238095,
+      "grad_norm": 2.161937442020458,
+      "learning_rate": 8.69339733686793e-06,
+      "loss": 0.6114,
+      "mean_token_accuracy": 0.8068233132362366,
+      "num_tokens": 393931741.0,
+      "step": 1048
+    },
+    {
+      "epoch": 1.5610119047619047,
+      "grad_norm": 2.0543058931045994,
+      "learning_rate": 8.68989400168286e-06,
+      "loss": 0.6065,
+      "mean_token_accuracy": 0.8073909282684326,
+      "num_tokens": 394309894.0,
+      "step": 1049
+    },
+    {
+      "epoch": 1.5625,
+      "grad_norm": 1.9771394148490922,
+      "learning_rate": 8.68638668405062e-06,
+      "loss": 0.555,
+      "mean_token_accuracy": 0.8205527067184448,
+      "num_tokens": 394688748.0,
+      "step": 1050
+    },
+    {
+      "epoch": 1.5639880952380953,
+      "grad_norm": 2.061899032080458,
+      "learning_rate": 8.682875387756609e-06,
+      "loss": 0.6367,
+      "mean_token_accuracy": 0.7961591482162476,
+      "num_tokens": 395068098.0,
+      "step": 1051
+    },
+    {
+      "epoch": 1.5654761904761905,
+      "grad_norm": 1.9835879377156134,
+      "learning_rate": 8.679360116590512e-06,
+      "loss": 0.6029,
+      "mean_token_accuracy": 0.8109660744667053,
+      "num_tokens": 395446420.0,
+      "step": 1052
+    },
+    {
+      "epoch": 1.5669642857142856,
+      "grad_norm": 2.121248270235506,
+      "learning_rate": 8.675840874346311e-06,
+      "loss": 0.6002,
+      "mean_token_accuracy": 0.8069432973861694,
+      "num_tokens": 395823712.0,
+      "step": 1053
+    },
+    {
+      "epoch": 1.568452380952381,
+      "grad_norm": 1.8371093070520832,
+      "learning_rate": 8.672317664822269e-06,
+      "loss": 0.5656,
+      "mean_token_accuracy": 0.8180558085441589,
+      "num_tokens": 396192950.0,
+      "step": 1054
+    },
+    {
+      "epoch": 1.5699404761904763,
+      "grad_norm": 1.98005095577594,
+      "learning_rate": 8.668790491820936e-06,
+      "loss": 0.5481,
+      "mean_token_accuracy": 0.8194254636764526,
+      "num_tokens": 396570792.0,
+      "step": 1055
+    },
+    {
+      "epoch": 1.5714285714285714,
+      "grad_norm": 1.9532101327131173,
+      "learning_rate": 8.665259359149132e-06,
+      "loss": 0.5664,
+      "mean_token_accuracy": 0.8163641095161438,
+      "num_tokens": 396943705.0,
+      "step": 1056
+    },
+    {
+      "epoch": 1.5729166666666665,
+      "grad_norm": 1.9046159082381071,
+      "learning_rate": 8.661724270617961e-06,
+      "loss": 0.5602,
+      "mean_token_accuracy": 0.8247724771499634,
+      "num_tokens": 397322791.0,
+      "step": 1057
+    },
+    {
+      "epoch": 1.5744047619047619,
+      "grad_norm": 2.0216340073104426,
+      "learning_rate": 8.658185230042788e-06,
+      "loss": 0.6191,
+      "mean_token_accuracy": 0.8010520339012146,
+      "num_tokens": 397695227.0,
+      "step": 1058
+    },
+    {
+      "epoch": 1.5758928571428572,
+      "grad_norm": 2.011158456249419,
+      "learning_rate": 8.654642241243247e-06,
+      "loss": 0.5907,
+      "mean_token_accuracy": 0.8097497820854187,
+      "num_tokens": 398073876.0,
+      "step": 1059
+    },
+    {
+      "epoch": 1.5773809523809523,
+      "grad_norm": 2.2259011238458664,
+      "learning_rate": 8.651095308043232e-06,
+      "loss": 0.6025,
+      "mean_token_accuracy": 0.8068798184394836,
+      "num_tokens": 398432707.0,
+      "step": 1060
+    },
+    {
+      "epoch": 1.5788690476190477,
+      "grad_norm": 1.961628099424687,
+      "learning_rate": 8.647544434270895e-06,
+      "loss": 0.6132,
+      "mean_token_accuracy": 0.8111147284507751,
+      "num_tokens": 398811727.0,
+      "step": 1061
+    },
+    {
+      "epoch": 1.5803571428571428,
+      "grad_norm": 1.9001609064144733,
+      "learning_rate": 8.643989623758642e-06,
+      "loss": 0.5349,
+      "mean_token_accuracy": 0.8265753388404846,
+      "num_tokens": 399191904.0,
+      "step": 1062
+    },
+    {
+      "epoch": 1.5818452380952381,
+      "grad_norm": 2.1636427319017137,
+      "learning_rate": 8.640430880343128e-06,
+      "loss": 0.5648,
+      "mean_token_accuracy": 0.8191553950309753,
+      "num_tokens": 399570423.0,
+      "step": 1063
+    },
+    {
+      "epoch": 1.5833333333333335,
+      "grad_norm": 2.0547759440956312,
+      "learning_rate": 8.636868207865244e-06,
+      "loss": 0.5453,
+      "mean_token_accuracy": 0.8252778053283691,
+      "num_tokens": 399947953.0,
+      "step": 1064
+    },
+    {
+      "epoch": 1.5848214285714286,
+      "grad_norm": 2.0119462570310676,
+      "learning_rate": 8.633301610170136e-06,
+      "loss": 0.5551,
+      "mean_token_accuracy": 0.8207641243934631,
+      "num_tokens": 400325965.0,
+      "step": 1065
+    },
+    {
+      "epoch": 1.5863095238095237,
+      "grad_norm": 1.9925456138984057,
+      "learning_rate": 8.629731091107177e-06,
+      "loss": 0.5746,
+      "mean_token_accuracy": 0.8164874911308289,
+      "num_tokens": 400704652.0,
+      "step": 1066
+    },
+    {
+      "epoch": 1.587797619047619,
+      "grad_norm": 1.8967817449072688,
+      "learning_rate": 8.626156654529973e-06,
+      "loss": 0.5816,
+      "mean_token_accuracy": 0.8117607235908508,
+      "num_tokens": 401063786.0,
+      "step": 1067
+    },
+    {
+      "epoch": 1.5892857142857144,
+      "grad_norm": 2.0109976091974087,
+      "learning_rate": 8.622578304296364e-06,
+      "loss": 0.6025,
+      "mean_token_accuracy": 0.8085476160049438,
+      "num_tokens": 401442816.0,
+      "step": 1068
+    },
+    {
+      "epoch": 1.5907738095238095,
+      "grad_norm": 1.9501315701302877,
+      "learning_rate": 8.618996044268406e-06,
+      "loss": 0.5877,
+      "mean_token_accuracy": 0.8076061010360718,
+      "num_tokens": 401821423.0,
+      "step": 1069
+    },
+    {
+      "epoch": 1.5922619047619047,
+      "grad_norm": 1.9727906304739844,
+      "learning_rate": 8.61540987831238e-06,
+      "loss": 0.5888,
+      "mean_token_accuracy": 0.8110653162002563,
+      "num_tokens": 402200257.0,
+      "step": 1070
+    },
+    {
+      "epoch": 1.59375,
+      "grad_norm": 2.091321913804192,
+      "learning_rate": 8.611819810298778e-06,
+      "loss": 0.5997,
+      "mean_token_accuracy": 0.7998606562614441,
+      "num_tokens": 402578849.0,
+      "step": 1071
+    },
+    {
+      "epoch": 1.5952380952380953,
+      "grad_norm": 1.9402964319035043,
+      "learning_rate": 8.608225844102312e-06,
+      "loss": 0.5163,
+      "mean_token_accuracy": 0.8278570771217346,
+      "num_tokens": 402956869.0,
+      "step": 1072
+    },
+    {
+      "epoch": 1.5967261904761905,
+      "grad_norm": 1.9120190304576605,
+      "learning_rate": 8.60462798360189e-06,
+      "loss": 0.5772,
+      "mean_token_accuracy": 0.8163877129554749,
+      "num_tokens": 403336083.0,
+      "step": 1073
+    },
+    {
+      "epoch": 1.5982142857142856,
+      "grad_norm": 2.2434949471455505,
+      "learning_rate": 8.601026232680634e-06,
+      "loss": 0.6872,
+      "mean_token_accuracy": 0.7897831797599792,
+      "num_tokens": 403714192.0,
+      "step": 1074
+    },
+    {
+      "epoch": 1.599702380952381,
+      "grad_norm": 2.168265234708881,
+      "learning_rate": 8.597420595225856e-06,
+      "loss": 0.6037,
+      "mean_token_accuracy": 0.8115103840827942,
+      "num_tokens": 404091814.0,
+      "step": 1075
+    },
+    {
+      "epoch": 1.6011904761904763,
+      "grad_norm": 1.8348424941790538,
+      "learning_rate": 8.593811075129065e-06,
+      "loss": 0.5879,
+      "mean_token_accuracy": 0.8105973601341248,
+      "num_tokens": 404471200.0,
+      "step": 1076
+    },
+    {
+      "epoch": 1.6026785714285714,
+      "grad_norm": 2.2904372105703183,
+      "learning_rate": 8.59019767628597e-06,
+      "loss": 0.6004,
+      "mean_token_accuracy": 0.8174962401390076,
+      "num_tokens": 404849893.0,
+      "step": 1077
+    },
+    {
+      "epoch": 1.6041666666666665,
+      "grad_norm": 2.0938577652978494,
+      "learning_rate": 8.586580402596448e-06,
+      "loss": 0.603,
+      "mean_token_accuracy": 0.8081400394439697,
+      "num_tokens": 405226693.0,
+      "step": 1078
+    },
+    {
+      "epoch": 1.6056547619047619,
+      "grad_norm": 2.2180486370759622,
+      "learning_rate": 8.582959257964574e-06,
+      "loss": 0.6183,
+      "mean_token_accuracy": 0.8137274384498596,
+      "num_tokens": 405603302.0,
+      "step": 1079
+    },
+    {
+      "epoch": 1.6071428571428572,
+      "grad_norm": 1.9836056082873068,
+      "learning_rate": 8.579334246298593e-06,
+      "loss": 0.6172,
+      "mean_token_accuracy": 0.8119497299194336,
+      "num_tokens": 405982200.0,
+      "step": 1080
+    },
+    {
+      "epoch": 1.6086309523809523,
+      "grad_norm": 2.084036835455078,
+      "learning_rate": 8.575705371510926e-06,
+      "loss": 0.5985,
+      "mean_token_accuracy": 0.8064572215080261,
+      "num_tokens": 406356204.0,
+      "step": 1081
+    },
+    {
+      "epoch": 1.6101190476190477,
+      "grad_norm": 1.8978849049707773,
+      "learning_rate": 8.572072637518166e-06,
+      "loss": 0.5463,
+      "mean_token_accuracy": 0.823622465133667,
+      "num_tokens": 406736176.0,
+      "step": 1082
+    },
+    {
+      "epoch": 1.6116071428571428,
+      "grad_norm": 1.945806165001805,
+      "learning_rate": 8.568436048241062e-06,
+      "loss": 0.5738,
+      "mean_token_accuracy": 0.8158851265907288,
+      "num_tokens": 407096675.0,
+      "step": 1083
+    },
+    {
+      "epoch": 1.6130952380952381,
+      "grad_norm": 2.184380136622664,
+      "learning_rate": 8.564795607604535e-06,
+      "loss": 0.5907,
+      "mean_token_accuracy": 0.8134475946426392,
+      "num_tokens": 407474182.0,
+      "step": 1084
+    },
+    {
+      "epoch": 1.6145833333333335,
+      "grad_norm": 2.0285237062203616,
+      "learning_rate": 8.561151319537656e-06,
+      "loss": 0.542,
+      "mean_token_accuracy": 0.8204033970832825,
+      "num_tokens": 407851500.0,
+      "step": 1085
+    },
+    {
+      "epoch": 1.6160714285714286,
+      "grad_norm": 2.067329370796049,
+      "learning_rate": 8.557503187973652e-06,
+      "loss": 0.5763,
+      "mean_token_accuracy": 0.81388258934021,
+      "num_tokens": 408230232.0,
+      "step": 1086
+    },
+    {
+      "epoch": 1.6175595238095237,
+      "grad_norm": 1.9220027376109157,
+      "learning_rate": 8.553851216849895e-06,
+      "loss": 0.5538,
+      "mean_token_accuracy": 0.8212290406227112,
+      "num_tokens": 408609064.0,
+      "step": 1087
+    },
+    {
+      "epoch": 1.619047619047619,
+      "grad_norm": 2.030967380622667,
+      "learning_rate": 8.550195410107903e-06,
+      "loss": 0.6096,
+      "mean_token_accuracy": 0.8071711659431458,
+      "num_tokens": 408988306.0,
+      "step": 1088
+    },
+    {
+      "epoch": 1.6205357142857144,
+      "grad_norm": 1.9263867709427172,
+      "learning_rate": 8.546535771693333e-06,
+      "loss": 0.582,
+      "mean_token_accuracy": 0.8132321834564209,
+      "num_tokens": 409354651.0,
+      "step": 1089
+    },
+    {
+      "epoch": 1.6220238095238095,
+      "grad_norm": 2.06091714380366,
+      "learning_rate": 8.542872305555978e-06,
+      "loss": 0.6141,
+      "mean_token_accuracy": 0.8050181269645691,
+      "num_tokens": 409720989.0,
+      "step": 1090
+    },
+    {
+      "epoch": 1.6235119047619047,
+      "grad_norm": 2.127873316305744,
+      "learning_rate": 8.539205015649764e-06,
+      "loss": 0.6103,
+      "mean_token_accuracy": 0.8049986958503723,
+      "num_tokens": 410091107.0,
+      "step": 1091
+    },
+    {
+      "epoch": 1.625,
+      "grad_norm": 1.9856064510644837,
+      "learning_rate": 8.535533905932739e-06,
+      "loss": 0.5692,
+      "mean_token_accuracy": 0.8156865835189819,
+      "num_tokens": 410470129.0,
+      "step": 1092
+    },
+    {
+      "epoch": 1.6264880952380953,
+      "grad_norm": 1.9964219325748904,
+      "learning_rate": 8.531858980367077e-06,
+      "loss": 0.585,
+      "mean_token_accuracy": 0.8107528686523438,
+      "num_tokens": 410849251.0,
+      "step": 1093
+    },
+    {
+      "epoch": 1.6279761904761905,
+      "grad_norm": 1.8576329062237686,
+      "learning_rate": 8.52818024291907e-06,
+      "loss": 0.5363,
+      "mean_token_accuracy": 0.8235026597976685,
+      "num_tokens": 411227268.0,
+      "step": 1094
+    },
+    {
+      "epoch": 1.6294642857142856,
+      "grad_norm": 2.0463810808210026,
+      "learning_rate": 8.524497697559126e-06,
+      "loss": 0.6501,
+      "mean_token_accuracy": 0.7956905364990234,
+      "num_tokens": 411605344.0,
+      "step": 1095
+    },
+    {
+      "epoch": 1.630952380952381,
+      "grad_norm": 1.873400771038805,
+      "learning_rate": 8.52081134826176e-06,
+      "loss": 0.5835,
+      "mean_token_accuracy": 0.8111101984977722,
+      "num_tokens": 411984723.0,
+      "step": 1096
+    },
+    {
+      "epoch": 1.6324404761904763,
+      "grad_norm": 2.0391908162654944,
+      "learning_rate": 8.517121199005593e-06,
+      "loss": 0.5708,
+      "mean_token_accuracy": 0.8212676048278809,
+      "num_tokens": 412361921.0,
+      "step": 1097
+    },
+    {
+      "epoch": 1.6339285714285714,
+      "grad_norm": 1.9575247960634798,
+      "learning_rate": 8.513427253773347e-06,
+      "loss": 0.5762,
+      "mean_token_accuracy": 0.8180903792381287,
+      "num_tokens": 412739841.0,
+      "step": 1098
+    },
+    {
+      "epoch": 1.6354166666666665,
+      "grad_norm": 1.8800190929039007,
+      "learning_rate": 8.509729516551842e-06,
+      "loss": 0.5613,
+      "mean_token_accuracy": 0.8204314112663269,
+      "num_tokens": 413120127.0,
+      "step": 1099
+    },
+    {
+      "epoch": 1.6369047619047619,
+      "grad_norm": 1.8484894557792184,
+      "learning_rate": 8.50602799133199e-06,
+      "loss": 0.5475,
+      "mean_token_accuracy": 0.8255562782287598,
+      "num_tokens": 413499519.0,
+      "step": 1100
+    },
+    {
+      "epoch": 1.6383928571428572,
+      "grad_norm": 2.008211394219434,
+      "learning_rate": 8.502322682108792e-06,
+      "loss": 0.5612,
+      "mean_token_accuracy": 0.814922571182251,
+      "num_tokens": 413877805.0,
+      "step": 1101
+    },
+    {
+      "epoch": 1.6398809523809523,
+      "grad_norm": 2.063353425706062,
+      "learning_rate": 8.498613592881334e-06,
+      "loss": 0.6197,
+      "mean_token_accuracy": 0.8026809692382812,
+      "num_tokens": 414255236.0,
+      "step": 1102
+    },
+    {
+      "epoch": 1.6413690476190477,
+      "grad_norm": 2.100864544947686,
+      "learning_rate": 8.494900727652776e-06,
+      "loss": 0.605,
+      "mean_token_accuracy": 0.8083034157752991,
+      "num_tokens": 414635603.0,
+      "step": 1103
+    },
+    {
+      "epoch": 1.6428571428571428,
+      "grad_norm": 1.97097391797393,
+      "learning_rate": 8.491184090430365e-06,
+      "loss": 0.5856,
+      "mean_token_accuracy": 0.8039779663085938,
+      "num_tokens": 415014879.0,
+      "step": 1104
+    },
+    {
+      "epoch": 1.6443452380952381,
+      "grad_norm": 2.057407256767728,
+      "learning_rate": 8.487463685225406e-06,
+      "loss": 0.6119,
+      "mean_token_accuracy": 0.8051044344902039,
+      "num_tokens": 415392993.0,
+      "step": 1105
+    },
+    {
+      "epoch": 1.6458333333333335,
+      "grad_norm": 1.992354926162808,
+      "learning_rate": 8.483739516053276e-06,
+      "loss": 0.6299,
+      "mean_token_accuracy": 0.8036115765571594,
+      "num_tokens": 415772549.0,
+      "step": 1106
+    },
+    {
+      "epoch": 1.6473214285714286,
+      "grad_norm": 1.8734057285905397,
+      "learning_rate": 8.480011586933417e-06,
+      "loss": 0.5411,
+      "mean_token_accuracy": 0.8269230723381042,
+      "num_tokens": 416152404.0,
+      "step": 1107
+    },
+    {
+      "epoch": 1.6488095238095237,
+      "grad_norm": 1.9626670968100377,
+      "learning_rate": 8.476279901889328e-06,
+      "loss": 0.4972,
+      "mean_token_accuracy": 0.836223304271698,
+      "num_tokens": 416529565.0,
+      "step": 1108
+    },
+    {
+      "epoch": 1.650297619047619,
+      "grad_norm": 2.198389615225553,
+      "learning_rate": 8.472544464948558e-06,
+      "loss": 0.6018,
+      "mean_token_accuracy": 0.8084786534309387,
+      "num_tokens": 416908097.0,
+      "step": 1109
+    },
+    {
+      "epoch": 1.6517857142857144,
+      "grad_norm": 2.0874226728472838,
+      "learning_rate": 8.46880528014271e-06,
+      "loss": 0.5878,
+      "mean_token_accuracy": 0.8083170652389526,
+      "num_tokens": 417285248.0,
+      "step": 1110
+    },
+    {
+      "epoch": 1.6532738095238095,
+      "grad_norm": 1.9526185207946185,
+      "learning_rate": 8.465062351507427e-06,
+      "loss": 0.5689,
+      "mean_token_accuracy": 0.8244274854660034,
+      "num_tokens": 417664084.0,
+      "step": 1111
+    },
+    {
+      "epoch": 1.6547619047619047,
+      "grad_norm": 2.0661220441623804,
+      "learning_rate": 8.461315683082398e-06,
+      "loss": 0.616,
+      "mean_token_accuracy": 0.7952308058738708,
+      "num_tokens": 418042468.0,
+      "step": 1112
+    },
+    {
+      "epoch": 1.65625,
+      "grad_norm": 1.967929214587319,
+      "learning_rate": 8.457565278911349e-06,
+      "loss": 0.5818,
+      "mean_token_accuracy": 0.8086165189743042,
+      "num_tokens": 418420012.0,
+      "step": 1113
+    },
+    {
+      "epoch": 1.6577380952380953,
+      "grad_norm": 1.9571868904139773,
+      "learning_rate": 8.453811143042027e-06,
+      "loss": 0.561,
+      "mean_token_accuracy": 0.8169890642166138,
+      "num_tokens": 418799148.0,
+      "step": 1114
+    },
+    {
+      "epoch": 1.6592261904761905,
+      "grad_norm": 1.9751157015463348,
+      "learning_rate": 8.450053279526223e-06,
+      "loss": 0.5911,
+      "mean_token_accuracy": 0.8106662034988403,
+      "num_tokens": 419177974.0,
+      "step": 1115
+    },
+    {
+      "epoch": 1.6607142857142856,
+      "grad_norm": 1.9626085997321654,
+      "learning_rate": 8.446291692419735e-06,
+      "loss": 0.5797,
+      "mean_token_accuracy": 0.8110519647598267,
+      "num_tokens": 419551629.0,
+      "step": 1116
+    },
+    {
+      "epoch": 1.662202380952381,
+      "grad_norm": 2.152262473960765,
+      "learning_rate": 8.442526385782395e-06,
+      "loss": 0.6002,
+      "mean_token_accuracy": 0.801326334476471,
+      "num_tokens": 419916808.0,
+      "step": 1117
+    },
+    {
+      "epoch": 1.6636904761904763,
+      "grad_norm": 2.005104323506233,
+      "learning_rate": 8.438757363678036e-06,
+      "loss": 0.6063,
+      "mean_token_accuracy": 0.8028563857078552,
+      "num_tokens": 420295487.0,
+      "step": 1118
+    },
+    {
+      "epoch": 1.6651785714285714,
+      "grad_norm": 1.9948147678260426,
+      "learning_rate": 8.43498463017451e-06,
+      "loss": 0.555,
+      "mean_token_accuracy": 0.8205313682556152,
+      "num_tokens": 420668202.0,
+      "step": 1119
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 2.133831533726892,
+      "learning_rate": 8.43120818934367e-06,
+      "loss": 0.5304,
+      "mean_token_accuracy": 0.8260053396224976,
+      "num_tokens": 421044851.0,
+      "step": 1120
+    },
+    {
+      "epoch": 1.6681547619047619,
+      "grad_norm": 1.9672167094089952,
+      "learning_rate": 8.427428045261371e-06,
+      "loss": 0.5224,
+      "mean_token_accuracy": 0.8303547501564026,
+      "num_tokens": 421422838.0,
+      "step": 1121
+    },
+    {
+      "epoch": 1.6696428571428572,
+      "grad_norm": 1.8928244385880482,
+      "learning_rate": 8.423644202007468e-06,
+      "loss": 0.5889,
+      "mean_token_accuracy": 0.8106966018676758,
+      "num_tokens": 421792398.0,
+      "step": 1122
+    },
+    {
+      "epoch": 1.6711309523809523,
+      "grad_norm": 2.100341453682681,
+      "learning_rate": 8.419856663665807e-06,
+      "loss": 0.5664,
+      "mean_token_accuracy": 0.8168962597846985,
+      "num_tokens": 422169778.0,
+      "step": 1123
+    },
+    {
+      "epoch": 1.6726190476190477,
+      "grad_norm": 2.0402801595433133,
+      "learning_rate": 8.416065434324216e-06,
+      "loss": 0.5902,
+      "mean_token_accuracy": 0.8078931570053101,
+      "num_tokens": 422545048.0,
+      "step": 1124
+    },
+    {
+      "epoch": 1.6741071428571428,
+      "grad_norm": 2.0486510926027615,
+      "learning_rate": 8.41227051807452e-06,
+      "loss": 0.6095,
+      "mean_token_accuracy": 0.8070071339607239,
+      "num_tokens": 422923484.0,
+      "step": 1125
+    },
+    {
+      "epoch": 1.6755952380952381,
+      "grad_norm": 2.2455706542847413,
+      "learning_rate": 8.408471919012508e-06,
+      "loss": 0.6321,
+      "mean_token_accuracy": 0.8013505935668945,
+      "num_tokens": 423301054.0,
+      "step": 1126
+    },
+    {
+      "epoch": 1.6770833333333335,
+      "grad_norm": 2.1407785502374175,
+      "learning_rate": 8.404669641237952e-06,
+      "loss": 0.6531,
+      "mean_token_accuracy": 0.7902893424034119,
+      "num_tokens": 423678091.0,
+      "step": 1127
+    },
+    {
+      "epoch": 1.6785714285714286,
+      "grad_norm": 2.098354377568745,
+      "learning_rate": 8.400863688854598e-06,
+      "loss": 0.6066,
+      "mean_token_accuracy": 0.8070899844169617,
+      "num_tokens": 424056482.0,
+      "step": 1128
+    },
+    {
+      "epoch": 1.6800595238095237,
+      "grad_norm": 2.1037184643561515,
+      "learning_rate": 8.39705406597015e-06,
+      "loss": 0.6046,
+      "mean_token_accuracy": 0.8058099746704102,
+      "num_tokens": 424421308.0,
+      "step": 1129
+    },
+    {
+      "epoch": 1.681547619047619,
+      "grad_norm": 1.9508572200772354,
+      "learning_rate": 8.393240776696274e-06,
+      "loss": 0.5681,
+      "mean_token_accuracy": 0.8201195597648621,
+      "num_tokens": 424800461.0,
+      "step": 1130
+    },
+    {
+      "epoch": 1.6830357142857144,
+      "grad_norm": 1.920286792561504,
+      "learning_rate": 8.3894238251486e-06,
+      "loss": 0.5827,
+      "mean_token_accuracy": 0.813940167427063,
+      "num_tokens": 425175227.0,
+      "step": 1131
+    },
+    {
+      "epoch": 1.6845238095238095,
+      "grad_norm": 2.1530787118058274,
+      "learning_rate": 8.385603215446704e-06,
+      "loss": 0.5866,
+      "mean_token_accuracy": 0.8142374157905579,
+      "num_tokens": 425539838.0,
+      "step": 1132
+    },
+    {
+      "epoch": 1.6860119047619047,
+      "grad_norm": 1.909217345258858,
+      "learning_rate": 8.381778951714115e-06,
+      "loss": 0.5735,
+      "mean_token_accuracy": 0.8136690258979797,
+      "num_tokens": 425919533.0,
+      "step": 1133
+    },
+    {
+      "epoch": 1.6875,
+      "grad_norm": 2.0133015177925335,
+      "learning_rate": 8.377951038078303e-06,
+      "loss": 0.6083,
+      "mean_token_accuracy": 0.8002554178237915,
+      "num_tokens": 426284998.0,
+      "step": 1134
+    },
+    {
+      "epoch": 1.6889880952380953,
+      "grad_norm": 1.8254412772636095,
+      "learning_rate": 8.374119478670675e-06,
+      "loss": 0.5605,
+      "mean_token_accuracy": 0.8172242045402527,
+      "num_tokens": 426665004.0,
+      "step": 1135
+    },
+    {
+      "epoch": 1.6904761904761905,
+      "grad_norm": 2.1706787262389025,
+      "learning_rate": 8.370284277626576e-06,
+      "loss": 0.6134,
+      "mean_token_accuracy": 0.8086296916007996,
+      "num_tokens": 427042067.0,
+      "step": 1136
+    },
+    {
+      "epoch": 1.6919642857142856,
+      "grad_norm": 1.9511071368766764,
+      "learning_rate": 8.366445439085286e-06,
+      "loss": 0.5352,
+      "mean_token_accuracy": 0.8221896290779114,
+      "num_tokens": 427412418.0,
+      "step": 1137
+    },
+    {
+      "epoch": 1.693452380952381,
+      "grad_norm": 2.2535987826038064,
+      "learning_rate": 8.36260296719e-06,
+      "loss": 0.6302,
+      "mean_token_accuracy": 0.7918727993965149,
+      "num_tokens": 427775279.0,
+      "step": 1138
+    },
+    {
+      "epoch": 1.6949404761904763,
+      "grad_norm": 1.769600527947353,
+      "learning_rate": 8.358756866087843e-06,
+      "loss": 0.5149,
+      "mean_token_accuracy": 0.8296626210212708,
+      "num_tokens": 428149871.0,
+      "step": 1139
+    },
+    {
+      "epoch": 1.6964285714285714,
+      "grad_norm": 1.9578196486501838,
+      "learning_rate": 8.35490713992985e-06,
+      "loss": 0.5265,
+      "mean_token_accuracy": 0.8289915323257446,
+      "num_tokens": 428523441.0,
+      "step": 1140
+    },
+    {
+      "epoch": 1.6979166666666665,
+      "grad_norm": 2.1286046869403106,
+      "learning_rate": 8.35105379287098e-06,
+      "loss": 0.5986,
+      "mean_token_accuracy": 0.7984551191329956,
+      "num_tokens": 428901384.0,
+      "step": 1141
+    },
+    {
+      "epoch": 1.6994047619047619,
+      "grad_norm": 2.0335562525023874,
+      "learning_rate": 8.347196829070088e-06,
+      "loss": 0.5819,
+      "mean_token_accuracy": 0.8131868243217468,
+      "num_tokens": 429278877.0,
+      "step": 1142
+    },
+    {
+      "epoch": 1.7008928571428572,
+      "grad_norm": 2.0135610346738364,
+      "learning_rate": 8.343336252689935e-06,
+      "loss": 0.5708,
+      "mean_token_accuracy": 0.8131372332572937,
+      "num_tokens": 429655915.0,
+      "step": 1143
+    },
+    {
+      "epoch": 1.7023809523809523,
+      "grad_norm": 2.0700640731307245,
+      "learning_rate": 8.339472067897187e-06,
+      "loss": 0.6153,
+      "mean_token_accuracy": 0.8020216226577759,
+      "num_tokens": 430034609.0,
+      "step": 1144
+    },
+    {
+      "epoch": 1.7038690476190477,
+      "grad_norm": 1.8850203689825868,
+      "learning_rate": 8.3356042788624e-06,
+      "loss": 0.5682,
+      "mean_token_accuracy": 0.8214452862739563,
+      "num_tokens": 430414905.0,
+      "step": 1145
+    },
+    {
+      "epoch": 1.7053571428571428,
+      "grad_norm": 2.116209395708487,
+      "learning_rate": 8.331732889760021e-06,
+      "loss": 0.5656,
+      "mean_token_accuracy": 0.8169969916343689,
+      "num_tokens": 430774200.0,
+      "step": 1146
+    },
+    {
+      "epoch": 1.7068452380952381,
+      "grad_norm": 1.9608017667778472,
+      "learning_rate": 8.32785790476838e-06,
+      "loss": 0.586,
+      "mean_token_accuracy": 0.8142356872558594,
+      "num_tokens": 431152837.0,
+      "step": 1147
+    },
+    {
+      "epoch": 1.7083333333333335,
+      "grad_norm": 1.8636568069117343,
+      "learning_rate": 8.323979328069689e-06,
+      "loss": 0.5572,
+      "mean_token_accuracy": 0.8210409283638,
+      "num_tokens": 431531692.0,
+      "step": 1148
+    },
+    {
+      "epoch": 1.7098214285714286,
+      "grad_norm": 2.0508216243028468,
+      "learning_rate": 8.320097163850043e-06,
+      "loss": 0.5565,
+      "mean_token_accuracy": 0.8235242366790771,
+      "num_tokens": 431910790.0,
+      "step": 1149
+    },
+    {
+      "epoch": 1.7113095238095237,
+      "grad_norm": 2.0005935050279153,
+      "learning_rate": 8.316211416299397e-06,
+      "loss": 0.6033,
+      "mean_token_accuracy": 0.80824214220047,
+      "num_tokens": 432276893.0,
+      "step": 1150
+    },
+    {
+      "epoch": 1.712797619047619,
+      "grad_norm": 2.0170749465388425,
+      "learning_rate": 8.312322089611587e-06,
+      "loss": 0.5425,
+      "mean_token_accuracy": 0.8283124566078186,
+      "num_tokens": 432655963.0,
+      "step": 1151
+    },
+    {
+      "epoch": 1.7142857142857144,
+      "grad_norm": 2.066502104620739,
+      "learning_rate": 8.308429187984298e-06,
+      "loss": 0.5801,
+      "mean_token_accuracy": 0.8196271657943726,
+      "num_tokens": 433021852.0,
+      "step": 1152
+    },
+    {
+      "epoch": 1.7157738095238095,
+      "grad_norm": 2.196699800338595,
+      "learning_rate": 8.304532715619085e-06,
+      "loss": 0.6622,
+      "mean_token_accuracy": 0.7861239314079285,
+      "num_tokens": 433399185.0,
+      "step": 1153
+    },
+    {
+      "epoch": 1.7172619047619047,
+      "grad_norm": 2.1083268652990683,
+      "learning_rate": 8.300632676721348e-06,
+      "loss": 0.5917,
+      "mean_token_accuracy": 0.8150710463523865,
+      "num_tokens": 433776448.0,
+      "step": 1154
+    },
+    {
+      "epoch": 1.71875,
+      "grad_norm": 2.293493801756693,
+      "learning_rate": 8.296729075500345e-06,
+      "loss": 0.6513,
+      "mean_token_accuracy": 0.7942511439323425,
+      "num_tokens": 434153455.0,
+      "step": 1155
+    },
+    {
+      "epoch": 1.7202380952380953,
+      "grad_norm": 2.063225375898075,
+      "learning_rate": 8.292821916169171e-06,
+      "loss": 0.5975,
+      "mean_token_accuracy": 0.8072620630264282,
+      "num_tokens": 434531463.0,
+      "step": 1156
+    },
+    {
+      "epoch": 1.7217261904761905,
+      "grad_norm": 2.0051783450177507,
+      "learning_rate": 8.288911202944765e-06,
+      "loss": 0.518,
+      "mean_token_accuracy": 0.8374160528182983,
+      "num_tokens": 434909297.0,
+      "step": 1157
+    },
+    {
+      "epoch": 1.7232142857142856,
+      "grad_norm": 2.0802661550192867,
+      "learning_rate": 8.284996940047904e-06,
+      "loss": 0.5994,
+      "mean_token_accuracy": 0.8098604083061218,
+      "num_tokens": 435287488.0,
+      "step": 1158
+    },
+    {
+      "epoch": 1.724702380952381,
+      "grad_norm": 1.7575115520858242,
+      "learning_rate": 8.281079131703188e-06,
+      "loss": 0.521,
+      "mean_token_accuracy": 0.8287022113800049,
+      "num_tokens": 435662484.0,
+      "step": 1159
+    },
+    {
+      "epoch": 1.7261904761904763,
+      "grad_norm": 2.0949192320230114,
+      "learning_rate": 8.277157782139051e-06,
+      "loss": 0.6442,
+      "mean_token_accuracy": 0.7989017367362976,
+      "num_tokens": 436039838.0,
+      "step": 1160
+    },
+    {
+      "epoch": 1.7276785714285714,
+      "grad_norm": 1.921522222296609,
+      "learning_rate": 8.273232895587749e-06,
+      "loss": 0.5552,
+      "mean_token_accuracy": 0.8222280144691467,
+      "num_tokens": 436418680.0,
+      "step": 1161
+    },
+    {
+      "epoch": 1.7291666666666665,
+      "grad_norm": 2.0941844879238016,
+      "learning_rate": 8.269304476285349e-06,
+      "loss": 0.5583,
+      "mean_token_accuracy": 0.818236231803894,
+      "num_tokens": 436794889.0,
+      "step": 1162
+    },
+    {
+      "epoch": 1.7306547619047619,
+      "grad_norm": 1.9456738044987192,
+      "learning_rate": 8.265372528471737e-06,
+      "loss": 0.5617,
+      "mean_token_accuracy": 0.8218032717704773,
+      "num_tokens": 437174105.0,
+      "step": 1163
+    },
+    {
+      "epoch": 1.7321428571428572,
+      "grad_norm": 2.1092873078465866,
+      "learning_rate": 8.261437056390607e-06,
+      "loss": 0.573,
+      "mean_token_accuracy": 0.8152728080749512,
+      "num_tokens": 437550942.0,
+      "step": 1164
+    },
+    {
+      "epoch": 1.7336309523809523,
+      "grad_norm": 1.89071135189947,
+      "learning_rate": 8.257498064289451e-06,
+      "loss": 0.547,
+      "mean_token_accuracy": 0.8204737901687622,
+      "num_tokens": 437929536.0,
+      "step": 1165
+    },
+    {
+      "epoch": 1.7351190476190477,
+      "grad_norm": 2.0072705475434316,
+      "learning_rate": 8.253555556419569e-06,
+      "loss": 0.5943,
+      "mean_token_accuracy": 0.8033447861671448,
+      "num_tokens": 438307976.0,
+      "step": 1166
+    },
+    {
+      "epoch": 1.7366071428571428,
+      "grad_norm": 1.9043428200239698,
+      "learning_rate": 8.249609537036044e-06,
+      "loss": 0.5485,
+      "mean_token_accuracy": 0.8209397196769714,
+      "num_tokens": 438687579.0,
+      "step": 1167
+    },
+    {
+      "epoch": 1.7380952380952381,
+      "grad_norm": 2.1292342180231225,
+      "learning_rate": 8.24566001039776e-06,
+      "loss": 0.6272,
+      "mean_token_accuracy": 0.7992300987243652,
+      "num_tokens": 439066331.0,
+      "step": 1168
+    },
+    {
+      "epoch": 1.7395833333333335,
+      "grad_norm": 2.167174579108422,
+      "learning_rate": 8.241706980767382e-06,
+      "loss": 0.6274,
+      "mean_token_accuracy": 0.8039179444313049,
+      "num_tokens": 439443740.0,
+      "step": 1169
+    },
+    {
+      "epoch": 1.7410714285714286,
+      "grad_norm": 1.8287596271091187,
+      "learning_rate": 8.237750452411353e-06,
+      "loss": 0.515,
+      "mean_token_accuracy": 0.8334439396858215,
+      "num_tokens": 439822678.0,
+      "step": 1170
+    },
+    {
+      "epoch": 1.7425595238095237,
+      "grad_norm": 1.808569046294732,
+      "learning_rate": 8.233790429599896e-06,
+      "loss": 0.5137,
+      "mean_token_accuracy": 0.8300333023071289,
+      "num_tokens": 440202185.0,
+      "step": 1171
+    },
+    {
+      "epoch": 1.744047619047619,
+      "grad_norm": 1.9064945908414483,
+      "learning_rate": 8.229826916607005e-06,
+      "loss": 0.5413,
+      "mean_token_accuracy": 0.8232839107513428,
+      "num_tokens": 440580659.0,
+      "step": 1172
+    },
+    {
+      "epoch": 1.7455357142857144,
+      "grad_norm": 1.9991907528330553,
+      "learning_rate": 8.22585991771044e-06,
+      "loss": 0.5426,
+      "mean_token_accuracy": 0.8223409056663513,
+      "num_tokens": 440958894.0,
+      "step": 1173
+    },
+    {
+      "epoch": 1.7470238095238095,
+      "grad_norm": 2.009393899063029,
+      "learning_rate": 8.22188943719172e-06,
+      "loss": 0.5625,
+      "mean_token_accuracy": 0.824764609336853,
+      "num_tokens": 441338304.0,
+      "step": 1174
+    },
+    {
+      "epoch": 1.7485119047619047,
+      "grad_norm": 2.2128253318623776,
+      "learning_rate": 8.217915479336129e-06,
+      "loss": 0.5895,
+      "mean_token_accuracy": 0.8097500801086426,
+      "num_tokens": 441716323.0,
+      "step": 1175
+    },
+    {
+      "epoch": 1.75,
+      "grad_norm": 2.195316058939976,
+      "learning_rate": 8.213938048432697e-06,
+      "loss": 0.5746,
+      "mean_token_accuracy": 0.8188071250915527,
+      "num_tokens": 442092623.0,
+      "step": 1176
+    },
+    {
+      "epoch": 1.7514880952380953,
+      "grad_norm": 2.266891605954161,
+      "learning_rate": 8.209957148774209e-06,
+      "loss": 0.6351,
+      "mean_token_accuracy": 0.7963201403617859,
+      "num_tokens": 442458307.0,
+      "step": 1177
+    },
+    {
+      "epoch": 1.7529761904761905,
+      "grad_norm": 1.864014281241484,
+      "learning_rate": 8.205972784657186e-06,
+      "loss": 0.5404,
+      "mean_token_accuracy": 0.8240038156509399,
+      "num_tokens": 442838128.0,
+      "step": 1178
+    },
+    {
+      "epoch": 1.7544642857142856,
+      "grad_norm": 1.9289699659117079,
+      "learning_rate": 8.201984960381895e-06,
+      "loss": 0.5926,
+      "mean_token_accuracy": 0.8152459263801575,
+      "num_tokens": 443215266.0,
+      "step": 1179
+    },
+    {
+      "epoch": 1.755952380952381,
+      "grad_norm": 1.927280726798787,
+      "learning_rate": 8.197993680252334e-06,
+      "loss": 0.5709,
+      "mean_token_accuracy": 0.8211793303489685,
+      "num_tokens": 443593550.0,
+      "step": 1180
+    },
+    {
+      "epoch": 1.7574404761904763,
+      "grad_norm": 1.9737445848986004,
+      "learning_rate": 8.193998948576234e-06,
+      "loss": 0.5883,
+      "mean_token_accuracy": 0.8107857704162598,
+      "num_tokens": 443971211.0,
+      "step": 1181
+    },
+    {
+      "epoch": 1.7589285714285714,
+      "grad_norm": 1.9391773862942423,
+      "learning_rate": 8.190000769665044e-06,
+      "loss": 0.5599,
+      "mean_token_accuracy": 0.8236546516418457,
+      "num_tokens": 444350199.0,
+      "step": 1182
+    },
+    {
+      "epoch": 1.7604166666666665,
+      "grad_norm": 1.9024025907690503,
+      "learning_rate": 8.185999147833943e-06,
+      "loss": 0.5516,
+      "mean_token_accuracy": 0.8222843408584595,
+      "num_tokens": 444725056.0,
+      "step": 1183
+    },
+    {
+      "epoch": 1.7619047619047619,
+      "grad_norm": 1.8207351377203222,
+      "learning_rate": 8.181994087401819e-06,
+      "loss": 0.5596,
+      "mean_token_accuracy": 0.8150975108146667,
+      "num_tokens": 445103508.0,
+      "step": 1184
+    },
+    {
+      "epoch": 1.7633928571428572,
+      "grad_norm": 1.931507716453512,
+      "learning_rate": 8.177985592691273e-06,
+      "loss": 0.539,
+      "mean_token_accuracy": 0.825424075126648,
+      "num_tokens": 445465199.0,
+      "step": 1185
+    },
+    {
+      "epoch": 1.7648809523809523,
+      "grad_norm": 1.9853676533200466,
+      "learning_rate": 8.173973668028615e-06,
+      "loss": 0.5945,
+      "mean_token_accuracy": 0.8076888918876648,
+      "num_tokens": 445841269.0,
+      "step": 1186
+    },
+    {
+      "epoch": 1.7663690476190477,
+      "grad_norm": 2.148529345733289,
+      "learning_rate": 8.169958317743856e-06,
+      "loss": 0.6505,
+      "mean_token_accuracy": 0.7912144064903259,
+      "num_tokens": 446219674.0,
+      "step": 1187
+    },
+    {
+      "epoch": 1.7678571428571428,
+      "grad_norm": 2.0246840656742546,
+      "learning_rate": 8.165939546170701e-06,
+      "loss": 0.5827,
+      "mean_token_accuracy": 0.8129094839096069,
+      "num_tokens": 446584886.0,
+      "step": 1188
+    },
+    {
+      "epoch": 1.7693452380952381,
+      "grad_norm": 1.923084524041107,
+      "learning_rate": 8.161917357646547e-06,
+      "loss": 0.5551,
+      "mean_token_accuracy": 0.8168905377388,
+      "num_tokens": 446949682.0,
+      "step": 1189
+    },
+    {
+      "epoch": 1.7708333333333335,
+      "grad_norm": 1.998888088646245,
+      "learning_rate": 8.157891756512488e-06,
+      "loss": 0.5326,
+      "mean_token_accuracy": 0.8266361355781555,
+      "num_tokens": 447327834.0,
+      "step": 1190
+    },
+    {
+      "epoch": 1.7723214285714286,
+      "grad_norm": 2.0408687292301777,
+      "learning_rate": 8.153862747113293e-06,
+      "loss": 0.613,
+      "mean_token_accuracy": 0.8154041767120361,
+      "num_tokens": 447705941.0,
+      "step": 1191
+    },
+    {
+      "epoch": 1.7738095238095237,
+      "grad_norm": 1.9761536739300531,
+      "learning_rate": 8.149830333797407e-06,
+      "loss": 0.5302,
+      "mean_token_accuracy": 0.8278260827064514,
+      "num_tokens": 448065300.0,
+      "step": 1192
+    },
+    {
+      "epoch": 1.775297619047619,
+      "grad_norm": 2.067123588549256,
+      "learning_rate": 8.14579452091696e-06,
+      "loss": 0.5793,
+      "mean_token_accuracy": 0.8097115159034729,
+      "num_tokens": 448442487.0,
+      "step": 1193
+    },
+    {
+      "epoch": 1.7767857142857144,
+      "grad_norm": 2.0081165740924756,
+      "learning_rate": 8.141755312827737e-06,
+      "loss": 0.555,
+      "mean_token_accuracy": 0.8154452443122864,
+      "num_tokens": 448821274.0,
+      "step": 1194
+    },
+    {
+      "epoch": 1.7782738095238095,
+      "grad_norm": 1.9607509493840534,
+      "learning_rate": 8.137712713889198e-06,
+      "loss": 0.5997,
+      "mean_token_accuracy": 0.8115331530570984,
+      "num_tokens": 449200129.0,
+      "step": 1195
+    },
+    {
+      "epoch": 1.7797619047619047,
+      "grad_norm": 2.0639864573263993,
+      "learning_rate": 8.13366672846446e-06,
+      "loss": 0.5774,
+      "mean_token_accuracy": 0.8153858780860901,
+      "num_tokens": 449579249.0,
+      "step": 1196
+    },
+    {
+      "epoch": 1.78125,
+      "grad_norm": 2.1084947614327594,
+      "learning_rate": 8.129617360920297e-06,
+      "loss": 0.5546,
+      "mean_token_accuracy": 0.8178058862686157,
+      "num_tokens": 449958730.0,
+      "step": 1197
+    },
+    {
+      "epoch": 1.7827380952380953,
+      "grad_norm": 1.8179131375206494,
+      "learning_rate": 8.125564615627124e-06,
+      "loss": 0.56,
+      "mean_token_accuracy": 0.8153490424156189,
+      "num_tokens": 450337991.0,
+      "step": 1198
+    },
+    {
+      "epoch": 1.7842261904761905,
+      "grad_norm": 1.8081628696846244,
+      "learning_rate": 8.121508496959011e-06,
+      "loss": 0.5509,
+      "mean_token_accuracy": 0.8263647556304932,
+      "num_tokens": 450717860.0,
+      "step": 1199
+    },
+    {
+      "epoch": 1.7857142857142856,
+      "grad_norm": 1.995856475934769,
+      "learning_rate": 8.117449009293668e-06,
+      "loss": 0.5734,
+      "mean_token_accuracy": 0.8161625266075134,
+      "num_tokens": 451094589.0,
+      "step": 1200
+    },
+    {
+      "epoch": 1.787202380952381,
+      "grad_norm": 1.981674741369985,
+      "learning_rate": 8.113386157012437e-06,
+      "loss": 0.5565,
+      "mean_token_accuracy": 0.8226265907287598,
+      "num_tokens": 451460253.0,
+      "step": 1201
+    },
+    {
+      "epoch": 1.7886904761904763,
+      "grad_norm": 1.9823574348804707,
+      "learning_rate": 8.109319944500294e-06,
+      "loss": 0.5403,
+      "mean_token_accuracy": 0.8243327736854553,
+      "num_tokens": 451838283.0,
+      "step": 1202
+    },
+    {
+      "epoch": 1.7901785714285714,
+      "grad_norm": 2.0866742908283853,
+      "learning_rate": 8.105250376145841e-06,
+      "loss": 0.5409,
+      "mean_token_accuracy": 0.8225579261779785,
+      "num_tokens": 452213967.0,
+      "step": 1203
+    },
+    {
+      "epoch": 1.7916666666666665,
+      "grad_norm": 1.8955903021021725,
+      "learning_rate": 8.101177456341301e-06,
+      "loss": 0.5241,
+      "mean_token_accuracy": 0.8319122195243835,
+      "num_tokens": 452591775.0,
+      "step": 1204
+    },
+    {
+      "epoch": 1.7931547619047619,
+      "grad_norm": 2.131475560909074,
+      "learning_rate": 8.097101189482517e-06,
+      "loss": 0.5904,
+      "mean_token_accuracy": 0.8078776001930237,
+      "num_tokens": 452969036.0,
+      "step": 1205
+    },
+    {
+      "epoch": 1.7946428571428572,
+      "grad_norm": 2.1171833264782847,
+      "learning_rate": 8.093021579968942e-06,
+      "loss": 0.5418,
+      "mean_token_accuracy": 0.8268375396728516,
+      "num_tokens": 453347313.0,
+      "step": 1206
+    },
+    {
+      "epoch": 1.7961309523809523,
+      "grad_norm": 1.9775829522705728,
+      "learning_rate": 8.088938632203635e-06,
+      "loss": 0.5343,
+      "mean_token_accuracy": 0.8233675956726074,
+      "num_tokens": 453725081.0,
+      "step": 1207
+    },
+    {
+      "epoch": 1.7976190476190477,
+      "grad_norm": 1.8389787049873427,
+      "learning_rate": 8.084852350593264e-06,
+      "loss": 0.5239,
+      "mean_token_accuracy": 0.827527642250061,
+      "num_tokens": 454104104.0,
+      "step": 1208
+    },
+    {
+      "epoch": 1.7991071428571428,
+      "grad_norm": 1.9932874914880157,
+      "learning_rate": 8.08076273954809e-06,
+      "loss": 0.5088,
+      "mean_token_accuracy": 0.832739531993866,
+      "num_tokens": 454472227.0,
+      "step": 1209
+    },
+    {
+      "epoch": 1.8005952380952381,
+      "grad_norm": 2.0479012616465906,
+      "learning_rate": 8.076669803481966e-06,
+      "loss": 0.5452,
+      "mean_token_accuracy": 0.8226569890975952,
+      "num_tokens": 454849987.0,
+      "step": 1210
+    },
+    {
+      "epoch": 1.8020833333333335,
+      "grad_norm": 1.899836874542363,
+      "learning_rate": 8.072573546812338e-06,
+      "loss": 0.5033,
+      "mean_token_accuracy": 0.8356059193611145,
+      "num_tokens": 455229145.0,
+      "step": 1211
+    },
+    {
+      "epoch": 1.8035714285714286,
+      "grad_norm": 1.9879542000798442,
+      "learning_rate": 8.068473973960238e-06,
+      "loss": 0.5785,
+      "mean_token_accuracy": 0.8150304555892944,
+      "num_tokens": 455607459.0,
+      "step": 1212
+    },
+    {
+      "epoch": 1.8050595238095237,
+      "grad_norm": 1.893706010206685,
+      "learning_rate": 8.06437108935027e-06,
+      "loss": 0.5559,
+      "mean_token_accuracy": 0.8228822946548462,
+      "num_tokens": 455974896.0,
+      "step": 1213
+    },
+    {
+      "epoch": 1.806547619047619,
+      "grad_norm": 1.9332808873945955,
+      "learning_rate": 8.060264897410614e-06,
+      "loss": 0.5846,
+      "mean_token_accuracy": 0.8182030916213989,
+      "num_tokens": 456355479.0,
+      "step": 1214
+    },
+    {
+      "epoch": 1.8080357142857144,
+      "grad_norm": 2.2090798237838882,
+      "learning_rate": 8.056155402573024e-06,
+      "loss": 0.614,
+      "mean_token_accuracy": 0.8017867803573608,
+      "num_tokens": 456733579.0,
+      "step": 1215
+    },
+    {
+      "epoch": 1.8095238095238095,
+      "grad_norm": 2.148057139717735,
+      "learning_rate": 8.052042609272817e-06,
+      "loss": 0.5726,
+      "mean_token_accuracy": 0.8136582374572754,
+      "num_tokens": 457110620.0,
+      "step": 1216
+    },
+    {
+      "epoch": 1.8110119047619047,
+      "grad_norm": 2.182050881241156,
+      "learning_rate": 8.047926521948865e-06,
+      "loss": 0.5622,
+      "mean_token_accuracy": 0.8239149451255798,
+      "num_tokens": 457489242.0,
+      "step": 1217
+    },
+    {
+      "epoch": 1.8125,
+      "grad_norm": 1.919621187162155,
+      "learning_rate": 8.043807145043604e-06,
+      "loss": 0.5814,
+      "mean_token_accuracy": 0.8155609965324402,
+      "num_tokens": 457868328.0,
+      "step": 1218
+    },
+    {
+      "epoch": 1.8139880952380953,
+      "grad_norm": 2.0607638131256643,
+      "learning_rate": 8.039684483003011e-06,
+      "loss": 0.5915,
+      "mean_token_accuracy": 0.8046875,
+      "num_tokens": 458247015.0,
+      "step": 1219
+    },
+    {
+      "epoch": 1.8154761904761905,
+      "grad_norm": 2.109766815459668,
+      "learning_rate": 8.035558540276618e-06,
+      "loss": 0.5594,
+      "mean_token_accuracy": 0.8216959834098816,
+      "num_tokens": 458624131.0,
+      "step": 1220
+    },
+    {
+      "epoch": 1.8169642857142856,
+      "grad_norm": 1.939688913084686,
+      "learning_rate": 8.03142932131749e-06,
+      "loss": 0.543,
+      "mean_token_accuracy": 0.8278715014457703,
+      "num_tokens": 459000257.0,
+      "step": 1221
+    },
+    {
+      "epoch": 1.818452380952381,
+      "grad_norm": 2.1065899612756858,
+      "learning_rate": 8.02729683058223e-06,
+      "loss": 0.6226,
+      "mean_token_accuracy": 0.8031985759735107,
+      "num_tokens": 459378078.0,
+      "step": 1222
+    },
+    {
+      "epoch": 1.8199404761904763,
+      "grad_norm": 2.037906370409132,
+      "learning_rate": 8.023161072530977e-06,
+      "loss": 0.5845,
+      "mean_token_accuracy": 0.8081044554710388,
+      "num_tokens": 459742091.0,
+      "step": 1223
+    },
+    {
+      "epoch": 1.8214285714285714,
+      "grad_norm": 1.8943166881090758,
+      "learning_rate": 8.019022051627387e-06,
+      "loss": 0.5606,
+      "mean_token_accuracy": 0.818062961101532,
+      "num_tokens": 460121137.0,
+      "step": 1224
+    },
+    {
+      "epoch": 1.8229166666666665,
+      "grad_norm": 1.9731180256655325,
+      "learning_rate": 8.014879772338649e-06,
+      "loss": 0.5544,
+      "mean_token_accuracy": 0.820517361164093,
+      "num_tokens": 460499558.0,
+      "step": 1225
+    },
+    {
+      "epoch": 1.8244047619047619,
+      "grad_norm": 2.068246382749694,
+      "learning_rate": 8.010734239135459e-06,
+      "loss": 0.5522,
+      "mean_token_accuracy": 0.8211267590522766,
+      "num_tokens": 460878286.0,
+      "step": 1226
+    },
+    {
+      "epoch": 1.8258928571428572,
+      "grad_norm": 2.092670659403887,
+      "learning_rate": 8.00658545649203e-06,
+      "loss": 0.6252,
+      "mean_token_accuracy": 0.8002672791481018,
+      "num_tokens": 461255914.0,
+      "step": 1227
+    },
+    {
+      "epoch": 1.8273809523809523,
+      "grad_norm": 1.8935627272216926,
+      "learning_rate": 8.00243342888608e-06,
+      "loss": 0.564,
+      "mean_token_accuracy": 0.8164966702461243,
+      "num_tokens": 461634305.0,
+      "step": 1228
+    },
+    {
+      "epoch": 1.8288690476190477,
+      "grad_norm": 1.8623053659456894,
+      "learning_rate": 7.998278160798831e-06,
+      "loss": 0.5581,
+      "mean_token_accuracy": 0.8232148885726929,
+      "num_tokens": 462001407.0,
+      "step": 1229
+    },
+    {
+      "epoch": 1.8303571428571428,
+      "grad_norm": 2.1839842486645016,
+      "learning_rate": 7.994119656715002e-06,
+      "loss": 0.6277,
+      "mean_token_accuracy": 0.7997850179672241,
+      "num_tokens": 462379847.0,
+      "step": 1230
+    },
+    {
+      "epoch": 1.8318452380952381,
+      "grad_norm": 2.036460023104678,
+      "learning_rate": 7.989957921122801e-06,
+      "loss": 0.5787,
+      "mean_token_accuracy": 0.8132817149162292,
+      "num_tokens": 462745432.0,
+      "step": 1231
+    },
+    {
+      "epoch": 1.8333333333333335,
+      "grad_norm": 1.9767996777269012,
+      "learning_rate": 7.985792958513932e-06,
+      "loss": 0.5732,
+      "mean_token_accuracy": 0.8199608325958252,
+      "num_tokens": 463124009.0,
+      "step": 1232
+    },
+    {
+      "epoch": 1.8348214285714286,
+      "grad_norm": 1.9728962847793974,
+      "learning_rate": 7.981624773383572e-06,
+      "loss": 0.5379,
+      "mean_token_accuracy": 0.8327535390853882,
+      "num_tokens": 463501879.0,
+      "step": 1233
+    },
+    {
+      "epoch": 1.8363095238095237,
+      "grad_norm": 1.8253989013018634,
+      "learning_rate": 7.97745337023038e-06,
+      "loss": 0.4995,
+      "mean_token_accuracy": 0.8345713019371033,
+      "num_tokens": 463880432.0,
+      "step": 1234
+    },
+    {
+      "epoch": 1.837797619047619,
+      "grad_norm": 1.9290433840573138,
+      "learning_rate": 7.973278753556492e-06,
+      "loss": 0.6045,
+      "mean_token_accuracy": 0.8080054521560669,
+      "num_tokens": 464258384.0,
+      "step": 1235
+    },
+    {
+      "epoch": 1.8392857142857144,
+      "grad_norm": 1.9203497428802945,
+      "learning_rate": 7.969100927867508e-06,
+      "loss": 0.5541,
+      "mean_token_accuracy": 0.8211458325386047,
+      "num_tokens": 464638574.0,
+      "step": 1236
+    },
+    {
+      "epoch": 1.8407738095238095,
+      "grad_norm": 1.8664206695745675,
+      "learning_rate": 7.964919897672493e-06,
+      "loss": 0.568,
+      "mean_token_accuracy": 0.8157297968864441,
+      "num_tokens": 465017773.0,
+      "step": 1237
+    },
+    {
+      "epoch": 1.8422619047619047,
+      "grad_norm": 1.9644180272501237,
+      "learning_rate": 7.960735667483969e-06,
+      "loss": 0.5544,
+      "mean_token_accuracy": 0.8176294565200806,
+      "num_tokens": 465382983.0,
+      "step": 1238
+    },
+    {
+      "epoch": 1.84375,
+      "grad_norm": 1.8482010450851838,
+      "learning_rate": 7.956548241817914e-06,
+      "loss": 0.5507,
+      "mean_token_accuracy": 0.8206971287727356,
+      "num_tokens": 465762189.0,
+      "step": 1239
+    },
+    {
+      "epoch": 1.8452380952380953,
+      "grad_norm": 2.034294228059538,
+      "learning_rate": 7.952357625193749e-06,
+      "loss": 0.5411,
+      "mean_token_accuracy": 0.8213125467300415,
+      "num_tokens": 466139181.0,
+      "step": 1240
+    },
+    {
+      "epoch": 1.8467261904761905,
+      "grad_norm": 2.0568638581619827,
+      "learning_rate": 7.948163822134351e-06,
+      "loss": 0.5864,
+      "mean_token_accuracy": 0.8133439421653748,
+      "num_tokens": 466515896.0,
+      "step": 1241
+    },
+    {
+      "epoch": 1.8482142857142856,
+      "grad_norm": 1.9956349223298435,
+      "learning_rate": 7.943966837166024e-06,
+      "loss": 0.5451,
+      "mean_token_accuracy": 0.8232644200325012,
+      "num_tokens": 466894539.0,
+      "step": 1242
+    },
+    {
+      "epoch": 1.849702380952381,
+      "grad_norm": 1.9219553598567136,
+      "learning_rate": 7.939766674818511e-06,
+      "loss": 0.5004,
+      "mean_token_accuracy": 0.8279094099998474,
+      "num_tokens": 467272596.0,
+      "step": 1243
+    },
+    {
+      "epoch": 1.8511904761904763,
+      "grad_norm": 1.930906317772096,
+      "learning_rate": 7.935563339624988e-06,
+      "loss": 0.5124,
+      "mean_token_accuracy": 0.8318811655044556,
+      "num_tokens": 467651693.0,
+      "step": 1244
+    },
+    {
+      "epoch": 1.8526785714285714,
+      "grad_norm": 1.8927308278782229,
+      "learning_rate": 7.931356836122046e-06,
+      "loss": 0.5599,
+      "mean_token_accuracy": 0.8206768035888672,
+      "num_tokens": 468031504.0,
+      "step": 1245
+    },
+    {
+      "epoch": 1.8541666666666665,
+      "grad_norm": 2.315358484776198,
+      "learning_rate": 7.927147168849704e-06,
+      "loss": 0.6328,
+      "mean_token_accuracy": 0.8075326085090637,
+      "num_tokens": 468408250.0,
+      "step": 1246
+    },
+    {
+      "epoch": 1.8556547619047619,
+      "grad_norm": 2.1305414161100744,
+      "learning_rate": 7.922934342351391e-06,
+      "loss": 0.5994,
+      "mean_token_accuracy": 0.8060150146484375,
+      "num_tokens": 468785604.0,
+      "step": 1247
+    },
+    {
+      "epoch": 1.8571428571428572,
+      "grad_norm": 2.3006551647107236,
+      "learning_rate": 7.918718361173951e-06,
+      "loss": 0.6257,
+      "mean_token_accuracy": 0.8057081699371338,
+      "num_tokens": 469163572.0,
+      "step": 1248
+    },
+    {
+      "epoch": 1.8586309523809523,
+      "grad_norm": 1.9127602631630096,
+      "learning_rate": 7.914499229867624e-06,
+      "loss": 0.5311,
+      "mean_token_accuracy": 0.8286060094833374,
+      "num_tokens": 469542284.0,
+      "step": 1249
+    },
+    {
+      "epoch": 1.8601190476190477,
+      "grad_norm": 1.9691725642502511,
+      "learning_rate": 7.91027695298606e-06,
+      "loss": 0.5927,
+      "mean_token_accuracy": 0.8144518136978149,
+      "num_tokens": 469920230.0,
+      "step": 1250
+    },
+    {
+      "epoch": 1.8616071428571428,
+      "grad_norm": 2.089934605696488,
+      "learning_rate": 7.906051535086296e-06,
+      "loss": 0.5471,
+      "mean_token_accuracy": 0.8191666603088379,
+      "num_tokens": 470297975.0,
+      "step": 1251
+    },
+    {
+      "epoch": 1.8630952380952381,
+      "grad_norm": 2.112632965554361,
+      "learning_rate": 7.901822980728761e-06,
+      "loss": 0.5441,
+      "mean_token_accuracy": 0.8249337077140808,
+      "num_tokens": 470676138.0,
+      "step": 1252
+    },
+    {
+      "epoch": 1.8645833333333335,
+      "grad_norm": 1.8905693525108547,
+      "learning_rate": 7.897591294477276e-06,
+      "loss": 0.5541,
+      "mean_token_accuracy": 0.8208906054496765,
+      "num_tokens": 471055433.0,
+      "step": 1253
+    },
+    {
+      "epoch": 1.8660714285714286,
+      "grad_norm": 2.1796258975340064,
+      "learning_rate": 7.89335648089903e-06,
+      "loss": 0.6222,
+      "mean_token_accuracy": 0.8016613721847534,
+      "num_tokens": 471416219.0,
+      "step": 1254
+    },
+    {
+      "epoch": 1.8675595238095237,
+      "grad_norm": 1.9536643329716783,
+      "learning_rate": 7.889118544564597e-06,
+      "loss": 0.5725,
+      "mean_token_accuracy": 0.8135423064231873,
+      "num_tokens": 471794722.0,
+      "step": 1255
+    },
+    {
+      "epoch": 1.869047619047619,
+      "grad_norm": 1.9459158255252806,
+      "learning_rate": 7.884877490047915e-06,
+      "loss": 0.6184,
+      "mean_token_accuracy": 0.8058637976646423,
+      "num_tokens": 472174125.0,
+      "step": 1256
+    },
+    {
+      "epoch": 1.8705357142857144,
+      "grad_norm": 1.9306508155126203,
+      "learning_rate": 7.880633321926294e-06,
+      "loss": 0.5735,
+      "mean_token_accuracy": 0.8208132386207581,
+      "num_tokens": 472551758.0,
+      "step": 1257
+    },
+    {
+      "epoch": 1.8720238095238095,
+      "grad_norm": 1.8693894679868357,
+      "learning_rate": 7.8763860447804e-06,
+      "loss": 0.4899,
+      "mean_token_accuracy": 0.8358315229415894,
+      "num_tokens": 472930170.0,
+      "step": 1258
+    },
+    {
+      "epoch": 1.8735119047619047,
+      "grad_norm": 1.9505313161661222,
+      "learning_rate": 7.872135663194254e-06,
+      "loss": 0.5843,
+      "mean_token_accuracy": 0.8130030632019043,
+      "num_tokens": 473309596.0,
+      "step": 1259
+    },
+    {
+      "epoch": 1.875,
+      "grad_norm": 1.8630497994449664,
+      "learning_rate": 7.86788218175523e-06,
+      "loss": 0.5326,
+      "mean_token_accuracy": 0.8224184513092041,
+      "num_tokens": 473685410.0,
+      "step": 1260
+    },
+    {
+      "epoch": 1.8764880952380953,
+      "grad_norm": 2.0013917829671244,
+      "learning_rate": 7.863625605054047e-06,
+      "loss": 0.5905,
+      "mean_token_accuracy": 0.8102604150772095,
+      "num_tokens": 474063756.0,
+      "step": 1261
+    },
+    {
+      "epoch": 1.8779761904761905,
+      "grad_norm": 1.8727632306593651,
+      "learning_rate": 7.859365937684764e-06,
+      "loss": 0.5777,
+      "mean_token_accuracy": 0.8202592730522156,
+      "num_tokens": 474443360.0,
+      "step": 1262
+    },
+    {
+      "epoch": 1.8794642857142856,
+      "grad_norm": 2.0300188152222427,
+      "learning_rate": 7.855103184244777e-06,
+      "loss": 0.6202,
+      "mean_token_accuracy": 0.8001197576522827,
+      "num_tokens": 474820311.0,
+      "step": 1263
+    },
+    {
+      "epoch": 1.880952380952381,
+      "grad_norm": 2.255620395062468,
+      "learning_rate": 7.85083734933481e-06,
+      "loss": 0.6493,
+      "mean_token_accuracy": 0.7961801290512085,
+      "num_tokens": 475195353.0,
+      "step": 1264
+    },
+    {
+      "epoch": 1.8824404761904763,
+      "grad_norm": 2.012773483360538,
+      "learning_rate": 7.846568437558914e-06,
+      "loss": 0.5452,
+      "mean_token_accuracy": 0.8200330138206482,
+      "num_tokens": 475573080.0,
+      "step": 1265
+    },
+    {
+      "epoch": 1.8839285714285714,
+      "grad_norm": 2.2045712740698793,
+      "learning_rate": 7.842296453524462e-06,
+      "loss": 0.5746,
+      "mean_token_accuracy": 0.816071093082428,
+      "num_tokens": 475950257.0,
+      "step": 1266
+    },
+    {
+      "epoch": 1.8854166666666665,
+      "grad_norm": 2.049135148995267,
+      "learning_rate": 7.838021401842145e-06,
+      "loss": 0.5599,
+      "mean_token_accuracy": 0.8231121301651001,
+      "num_tokens": 476328239.0,
+      "step": 1267
+    },
+    {
+      "epoch": 1.8869047619047619,
+      "grad_norm": 1.9554538846305902,
+      "learning_rate": 7.833743287125958e-06,
+      "loss": 0.5666,
+      "mean_token_accuracy": 0.8248260617256165,
+      "num_tokens": 476706296.0,
+      "step": 1268
+    },
+    {
+      "epoch": 1.8883928571428572,
+      "grad_norm": 1.8557112502334183,
+      "learning_rate": 7.829462113993207e-06,
+      "loss": 0.5652,
+      "mean_token_accuracy": 0.814319372177124,
+      "num_tokens": 477085765.0,
+      "step": 1269
+    },
+    {
+      "epoch": 1.8898809523809523,
+      "grad_norm": 1.8695852943996685,
+      "learning_rate": 7.8251778870645e-06,
+      "loss": 0.5499,
+      "mean_token_accuracy": 0.8203825950622559,
+      "num_tokens": 477465040.0,
+      "step": 1270
+    },
+    {
+      "epoch": 1.8913690476190477,
+      "grad_norm": 2.1952813687574952,
+      "learning_rate": 7.820890610963738e-06,
+      "loss": 0.638,
+      "mean_token_accuracy": 0.7998321056365967,
+      "num_tokens": 477842215.0,
+      "step": 1271
+    },
+    {
+      "epoch": 1.8928571428571428,
+      "grad_norm": 1.9848223231242832,
+      "learning_rate": 7.81660029031811e-06,
+      "loss": 0.5758,
+      "mean_token_accuracy": 0.8111428618431091,
+      "num_tokens": 478220451.0,
+      "step": 1272
+    },
+    {
+      "epoch": 1.8943452380952381,
+      "grad_norm": 1.9618857918403882,
+      "learning_rate": 7.8123069297581e-06,
+      "loss": 0.5734,
+      "mean_token_accuracy": 0.8135434985160828,
+      "num_tokens": 478599839.0,
+      "step": 1273
+    },
+    {
+      "epoch": 1.8958333333333335,
+      "grad_norm": 1.9651995571145127,
+      "learning_rate": 7.808010533917465e-06,
+      "loss": 0.5342,
+      "mean_token_accuracy": 0.8244004249572754,
+      "num_tokens": 478978236.0,
+      "step": 1274
+    },
+    {
+      "epoch": 1.8973214285714286,
+      "grad_norm": 2.042832463957176,
+      "learning_rate": 7.80371110743324e-06,
+      "loss": 0.5228,
+      "mean_token_accuracy": 0.8291952610015869,
+      "num_tokens": 479356098.0,
+      "step": 1275
+    },
+    {
+      "epoch": 1.8988095238095237,
+      "grad_norm": 2.1927417605807693,
+      "learning_rate": 7.799408654945733e-06,
+      "loss": 0.6025,
+      "mean_token_accuracy": 0.8079530596733093,
+      "num_tokens": 479729946.0,
+      "step": 1276
+    },
+    {
+      "epoch": 1.900297619047619,
+      "grad_norm": 1.9362338889571336,
+      "learning_rate": 7.795103181098516e-06,
+      "loss": 0.5383,
+      "mean_token_accuracy": 0.8240844011306763,
+      "num_tokens": 480108821.0,
+      "step": 1277
+    },
+    {
+      "epoch": 1.9017857142857144,
+      "grad_norm": 1.9577769521910833,
+      "learning_rate": 7.790794690538422e-06,
+      "loss": 0.5944,
+      "mean_token_accuracy": 0.8120417594909668,
+      "num_tokens": 480487556.0,
+      "step": 1278
+    },
+    {
+      "epoch": 1.9032738095238095,
+      "grad_norm": 2.0494842875428056,
+      "learning_rate": 7.786483187915536e-06,
+      "loss": 0.5573,
+      "mean_token_accuracy": 0.8180989623069763,
+      "num_tokens": 480865326.0,
+      "step": 1279
+    },
+    {
+      "epoch": 1.9047619047619047,
+      "grad_norm": 1.9581928745483261,
+      "learning_rate": 7.782168677883206e-06,
+      "loss": 0.5419,
+      "mean_token_accuracy": 0.821442186832428,
+      "num_tokens": 481242084.0,
+      "step": 1280
+    },
+    {
+      "epoch": 1.90625,
+      "grad_norm": 1.9860303772808061,
+      "learning_rate": 7.777851165098012e-06,
+      "loss": 0.5946,
+      "mean_token_accuracy": 0.8108065128326416,
+      "num_tokens": 481621285.0,
+      "step": 1281
+    },
+    {
+      "epoch": 1.9077380952380953,
+      "grad_norm": 2.0984878535925837,
+      "learning_rate": 7.773530654219783e-06,
+      "loss": 0.5404,
+      "mean_token_accuracy": 0.8242946267127991,
+      "num_tokens": 481999711.0,
+      "step": 1282
+    },
+    {
+      "epoch": 1.9092261904761905,
+      "grad_norm": 2.0755333654504438,
+      "learning_rate": 7.76920714991158e-06,
+      "loss": 0.6269,
+      "mean_token_accuracy": 0.8022474050521851,
+      "num_tokens": 482378612.0,
+      "step": 1283
+    },
+    {
+      "epoch": 1.9107142857142856,
+      "grad_norm": 2.2001894554331587,
+      "learning_rate": 7.764880656839698e-06,
+      "loss": 0.5719,
+      "mean_token_accuracy": 0.8183867335319519,
+      "num_tokens": 482755267.0,
+      "step": 1284
+    },
+    {
+      "epoch": 1.912202380952381,
+      "grad_norm": 2.035821235885804,
+      "learning_rate": 7.760551179673655e-06,
+      "loss": 0.6503,
+      "mean_token_accuracy": 0.7942676544189453,
+      "num_tokens": 483134729.0,
+      "step": 1285
+    },
+    {
+      "epoch": 1.9136904761904763,
+      "grad_norm": 1.8596214876844475,
+      "learning_rate": 7.75621872308619e-06,
+      "loss": 0.5561,
+      "mean_token_accuracy": 0.8208030462265015,
+      "num_tokens": 483514429.0,
+      "step": 1286
+    },
+    {
+      "epoch": 1.9151785714285714,
+      "grad_norm": 1.8750848470027064,
+      "learning_rate": 7.751883291753261e-06,
+      "loss": 0.5547,
+      "mean_token_accuracy": 0.8191795349121094,
+      "num_tokens": 483892684.0,
+      "step": 1287
+    },
+    {
+      "epoch": 1.9166666666666665,
+      "grad_norm": 1.8865599516844263,
+      "learning_rate": 7.747544890354031e-06,
+      "loss": 0.5324,
+      "mean_token_accuracy": 0.8288127183914185,
+      "num_tokens": 484272112.0,
+      "step": 1288
+    },
+    {
+      "epoch": 1.9181547619047619,
+      "grad_norm": 1.96264157846834,
+      "learning_rate": 7.743203523570872e-06,
+      "loss": 0.5607,
+      "mean_token_accuracy": 0.8118677139282227,
+      "num_tokens": 484650976.0,
+      "step": 1289
+    },
+    {
+      "epoch": 1.9196428571428572,
+      "grad_norm": 2.0214175231811216,
+      "learning_rate": 7.738859196089358e-06,
+      "loss": 0.5943,
+      "mean_token_accuracy": 0.8135051131248474,
+      "num_tokens": 485029410.0,
+      "step": 1290
+    },
+    {
+      "epoch": 1.9211309523809523,
+      "grad_norm": 1.9005820290042545,
+      "learning_rate": 7.734511912598253e-06,
+      "loss": 0.5388,
+      "mean_token_accuracy": 0.8290879130363464,
+      "num_tokens": 485380771.0,
+      "step": 1291
+    },
+    {
+      "epoch": 1.9226190476190477,
+      "grad_norm": 1.8526862741459365,
+      "learning_rate": 7.730161677789518e-06,
+      "loss": 0.5801,
+      "mean_token_accuracy": 0.8187178373336792,
+      "num_tokens": 485746739.0,
+      "step": 1292
+    },
+    {
+      "epoch": 1.9241071428571428,
+      "grad_norm": 2.0131244230299425,
+      "learning_rate": 7.725808496358295e-06,
+      "loss": 0.5801,
+      "mean_token_accuracy": 0.8140555024147034,
+      "num_tokens": 486125200.0,
+      "step": 1293
+    },
+    {
+      "epoch": 1.9255952380952381,
+      "grad_norm": 1.9462576879153597,
+      "learning_rate": 7.721452373002906e-06,
+      "loss": 0.5815,
+      "mean_token_accuracy": 0.8110445141792297,
+      "num_tokens": 486504042.0,
+      "step": 1294
+    },
+    {
+      "epoch": 1.9270833333333335,
+      "grad_norm": 2.091674559149258,
+      "learning_rate": 7.71709331242485e-06,
+      "loss": 0.604,
+      "mean_token_accuracy": 0.8109750151634216,
+      "num_tokens": 486880828.0,
+      "step": 1295
+    },
+    {
+      "epoch": 1.9285714285714286,
+      "grad_norm": 1.8608715189728768,
+      "learning_rate": 7.712731319328798e-06,
+      "loss": 0.5366,
+      "mean_token_accuracy": 0.8279799222946167,
+      "num_tokens": 487259524.0,
+      "step": 1296
+    },
+    {
+      "epoch": 1.9300595238095237,
+      "grad_norm": 1.9165160368224836,
+      "learning_rate": 7.708366398422581e-06,
+      "loss": 0.5164,
+      "mean_token_accuracy": 0.828984260559082,
+      "num_tokens": 487637931.0,
+      "step": 1297
+    },
+    {
+      "epoch": 1.931547619047619,
+      "grad_norm": 2.254559471177039,
+      "learning_rate": 7.703998554417193e-06,
+      "loss": 0.6357,
+      "mean_token_accuracy": 0.8039718866348267,
+      "num_tokens": 487994980.0,
+      "step": 1298
+    },
+    {
+      "epoch": 1.9330357142857144,
+      "grad_norm": 2.0984407820975695,
+      "learning_rate": 7.699627792026784e-06,
+      "loss": 0.6071,
+      "mean_token_accuracy": 0.8026819825172424,
+      "num_tokens": 488372767.0,
+      "step": 1299
+    },
+    {
+      "epoch": 1.9345238095238095,
+      "grad_norm": 1.9784246177450229,
+      "learning_rate": 7.69525411596865e-06,
+      "loss": 0.5452,
+      "mean_token_accuracy": 0.8181507587432861,
+      "num_tokens": 488751181.0,
+      "step": 1300
+    },
+    {
+      "epoch": 1.9360119047619047,
+      "grad_norm": 1.9624532603806863,
+      "learning_rate": 7.690877530963233e-06,
+      "loss": 0.5308,
+      "mean_token_accuracy": 0.8237151503562927,
+      "num_tokens": 489129008.0,
+      "step": 1301
+    },
+    {
+      "epoch": 1.9375,
+      "grad_norm": 2.0339710544696072,
+      "learning_rate": 7.686498041734121e-06,
+      "loss": 0.5993,
+      "mean_token_accuracy": 0.808647871017456,
+      "num_tokens": 489509028.0,
+      "step": 1302
+    },
+    {
+      "epoch": 1.9389880952380953,
+      "grad_norm": 2.000268254989905,
+      "learning_rate": 7.682115653008024e-06,
+      "loss": 0.5868,
+      "mean_token_accuracy": 0.812822699546814,
+      "num_tokens": 489887596.0,
+      "step": 1303
+    },
+    {
+      "epoch": 1.9404761904761905,
+      "grad_norm": 1.9404127374059075,
+      "learning_rate": 7.677730369514792e-06,
+      "loss": 0.5423,
+      "mean_token_accuracy": 0.8269782066345215,
+      "num_tokens": 490266732.0,
+      "step": 1304
+    },
+    {
+      "epoch": 1.9419642857142856,
+      "grad_norm": 2.0403131031542414,
+      "learning_rate": 7.673342195987396e-06,
+      "loss": 0.5882,
+      "mean_token_accuracy": 0.8133824467658997,
+      "num_tokens": 490644975.0,
+      "step": 1305
+    },
+    {
+      "epoch": 1.943452380952381,
+      "grad_norm": 1.9955018686101151,
+      "learning_rate": 7.668951137161925e-06,
+      "loss": 0.564,
+      "mean_token_accuracy": 0.8163415789604187,
+      "num_tokens": 491022779.0,
+      "step": 1306
+    },
+    {
+      "epoch": 1.9449404761904763,
+      "grad_norm": 2.206144422893642,
+      "learning_rate": 7.664557197777584e-06,
+      "loss": 0.6333,
+      "mean_token_accuracy": 0.7998166680335999,
+      "num_tokens": 491400628.0,
+      "step": 1307
+    },
+    {
+      "epoch": 1.9464285714285714,
+      "grad_norm": 2.0226542323624646,
+      "learning_rate": 7.660160382576683e-06,
+      "loss": 0.5378,
+      "mean_token_accuracy": 0.8236154317855835,
+      "num_tokens": 491777422.0,
+      "step": 1308
+    },
+    {
+      "epoch": 1.9479166666666665,
+      "grad_norm": 1.881773616664525,
+      "learning_rate": 7.655760696304642e-06,
+      "loss": 0.5366,
+      "mean_token_accuracy": 0.8252695202827454,
+      "num_tokens": 492155244.0,
+      "step": 1309
+    },
+    {
+      "epoch": 1.9494047619047619,
+      "grad_norm": 2.2569842291937126,
+      "learning_rate": 7.651358143709972e-06,
+      "loss": 0.5854,
+      "mean_token_accuracy": 0.8120014667510986,
+      "num_tokens": 492533828.0,
+      "step": 1310
+    },
+    {
+      "epoch": 1.9508928571428572,
+      "grad_norm": 1.9454763594737177,
+      "learning_rate": 7.646952729544285e-06,
+      "loss": 0.5595,
+      "mean_token_accuracy": 0.8181136250495911,
+      "num_tokens": 492913251.0,
+      "step": 1311
+    },
+    {
+      "epoch": 1.9523809523809523,
+      "grad_norm": 2.0659854214168316,
+      "learning_rate": 7.642544458562278e-06,
+      "loss": 0.5354,
+      "mean_token_accuracy": 0.8241797089576721,
+      "num_tokens": 493290900.0,
+      "step": 1312
+    },
+    {
+      "epoch": 1.9538690476190477,
+      "grad_norm": 1.919832739490874,
+      "learning_rate": 7.638133335521732e-06,
+      "loss": 0.5944,
+      "mean_token_accuracy": 0.8063623905181885,
+      "num_tokens": 493658033.0,
+      "step": 1313
+    },
+    {
+      "epoch": 1.9553571428571428,
+      "grad_norm": 1.9949534081165072,
+      "learning_rate": 7.633719365183505e-06,
+      "loss": 0.5402,
+      "mean_token_accuracy": 0.8243052363395691,
+      "num_tokens": 494035037.0,
+      "step": 1314
+    },
+    {
+      "epoch": 1.9568452380952381,
+      "grad_norm": 2.10691385333705,
+      "learning_rate": 7.6293025523115284e-06,
+      "loss": 0.5745,
+      "mean_token_accuracy": 0.8136231303215027,
+      "num_tokens": 494412774.0,
+      "step": 1315
+    },
+    {
+      "epoch": 1.9583333333333335,
+      "grad_norm": 1.8844894279526119,
+      "learning_rate": 7.624882901672801e-06,
+      "loss": 0.564,
+      "mean_token_accuracy": 0.815826952457428,
+      "num_tokens": 494791551.0,
+      "step": 1316
+    },
+    {
+      "epoch": 1.9598214285714286,
+      "grad_norm": 1.8696424466199115,
+      "learning_rate": 7.620460418037388e-06,
+      "loss": 0.5054,
+      "mean_token_accuracy": 0.832766056060791,
+      "num_tokens": 495168686.0,
+      "step": 1317
+    },
+    {
+      "epoch": 1.9613095238095237,
+      "grad_norm": 2.0222736066377807,
+      "learning_rate": 7.616035106178411e-06,
+      "loss": 0.6004,
+      "mean_token_accuracy": 0.8044286966323853,
+      "num_tokens": 495546832.0,
+      "step": 1318
+    },
+    {
+      "epoch": 1.962797619047619,
+      "grad_norm": 2.0857539998632206,
+      "learning_rate": 7.611606970872039e-06,
+      "loss": 0.5656,
+      "mean_token_accuracy": 0.8152457475662231,
+      "num_tokens": 495924012.0,
+      "step": 1319
+    },
+    {
+      "epoch": 1.9642857142857144,
+      "grad_norm": 2.133822559330401,
+      "learning_rate": 7.607176016897491e-06,
+      "loss": 0.6452,
+      "mean_token_accuracy": 0.7994279861450195,
+      "num_tokens": 496301492.0,
+      "step": 1320
+    },
+    {
+      "epoch": 1.9657738095238095,
+      "grad_norm": 1.9365821142571067,
+      "learning_rate": 7.602742249037032e-06,
+      "loss": 0.5876,
+      "mean_token_accuracy": 0.8134016394615173,
+      "num_tokens": 496668635.0,
+      "step": 1321
+    },
+    {
+      "epoch": 1.9672619047619047,
+      "grad_norm": 1.9257652907725007,
+      "learning_rate": 7.598305672075961e-06,
+      "loss": 0.5928,
+      "mean_token_accuracy": 0.8086913228034973,
+      "num_tokens": 497032837.0,
+      "step": 1322
+    },
+    {
+      "epoch": 1.96875,
+      "grad_norm": 1.986666410955196,
+      "learning_rate": 7.593866290802608e-06,
+      "loss": 0.5413,
+      "mean_token_accuracy": 0.825728714466095,
+      "num_tokens": 497410150.0,
+      "step": 1323
+    },
+    {
+      "epoch": 1.9702380952380953,
+      "grad_norm": 1.9115727963116118,
+      "learning_rate": 7.58942411000833e-06,
+      "loss": 0.5995,
+      "mean_token_accuracy": 0.803578794002533,
+      "num_tokens": 497789201.0,
+      "step": 1324
+    },
+    {
+      "epoch": 1.9717261904761905,
+      "grad_norm": 2.160243116831774,
+      "learning_rate": 7.5849791344875066e-06,
+      "loss": 0.6124,
+      "mean_token_accuracy": 0.8111006617546082,
+      "num_tokens": 498166776.0,
+      "step": 1325
+    },
+    {
+      "epoch": 1.9732142857142856,
+      "grad_norm": 2.0097760689203827,
+      "learning_rate": 7.580531369037534e-06,
+      "loss": 0.5773,
+      "mean_token_accuracy": 0.8183972835540771,
+      "num_tokens": 498544266.0,
+      "step": 1326
+    },
+    {
+      "epoch": 1.974702380952381,
+      "grad_norm": 2.0811061456710087,
+      "learning_rate": 7.576080818458817e-06,
+      "loss": 0.5608,
+      "mean_token_accuracy": 0.8236406445503235,
+      "num_tokens": 498921076.0,
+      "step": 1327
+    },
+    {
+      "epoch": 1.9761904761904763,
+      "grad_norm": 2.0589938153218985,
+      "learning_rate": 7.571627487554769e-06,
+      "loss": 0.6044,
+      "mean_token_accuracy": 0.8061929941177368,
+      "num_tokens": 499298666.0,
+      "step": 1328
+    },
+    {
+      "epoch": 1.9776785714285714,
+      "grad_norm": 2.080328523833742,
+      "learning_rate": 7.567171381131803e-06,
+      "loss": 0.6078,
+      "mean_token_accuracy": 0.8013732433319092,
+      "num_tokens": 499677642.0,
+      "step": 1329
+    },
+    {
+      "epoch": 1.9791666666666665,
+      "grad_norm": 1.9470625955454375,
+      "learning_rate": 7.562712503999327e-06,
+      "loss": 0.6207,
+      "mean_token_accuracy": 0.8030290007591248,
+      "num_tokens": 500057231.0,
+      "step": 1330
+    },
+    {
+      "epoch": 1.9806547619047619,
+      "grad_norm": 1.7063183052759534,
+      "learning_rate": 7.55825086096974e-06,
+      "loss": 0.5481,
+      "mean_token_accuracy": 0.8212650418281555,
+      "num_tokens": 500428070.0,
+      "step": 1331
+    },
+    {
+      "epoch": 1.9821428571428572,
+      "grad_norm": 2.0035803209134637,
+      "learning_rate": 7.553786456858429e-06,
+      "loss": 0.5845,
+      "mean_token_accuracy": 0.8134439587593079,
+      "num_tokens": 500805760.0,
+      "step": 1332
+    },
+    {
+      "epoch": 1.9836309523809523,
+      "grad_norm": 1.8765329358920986,
+      "learning_rate": 7.5493192964837545e-06,
+      "loss": 0.55,
+      "mean_token_accuracy": 0.8211174011230469,
+      "num_tokens": 501173773.0,
+      "step": 1333
+    },
+    {
+      "epoch": 1.9851190476190477,
+      "grad_norm": 1.9397691542685125,
+      "learning_rate": 7.544849384667059e-06,
+      "loss": 0.4959,
+      "mean_token_accuracy": 0.8352277874946594,
+      "num_tokens": 501551456.0,
+      "step": 1334
+    },
+    {
+      "epoch": 1.9866071428571428,
+      "grad_norm": 1.9597313547846957,
+      "learning_rate": 7.540376726232648e-06,
+      "loss": 0.5724,
+      "mean_token_accuracy": 0.8132508397102356,
+      "num_tokens": 501929244.0,
+      "step": 1335
+    },
+    {
+      "epoch": 1.9880952380952381,
+      "grad_norm": 2.2540457629665203,
+      "learning_rate": 7.535901326007796e-06,
+      "loss": 0.6116,
+      "mean_token_accuracy": 0.7990365028381348,
+      "num_tokens": 502295119.0,
+      "step": 1336
+    },
+    {
+      "epoch": 1.9895833333333335,
+      "grad_norm": 2.1739396294525046,
+      "learning_rate": 7.531423188822738e-06,
+      "loss": 0.5737,
+      "mean_token_accuracy": 0.8129657506942749,
+      "num_tokens": 502670194.0,
+      "step": 1337
+    },
+    {
+      "epoch": 1.9910714285714286,
+      "grad_norm": 2.0436193722258396,
+      "learning_rate": 7.526942319510655e-06,
+      "loss": 0.5958,
+      "mean_token_accuracy": 0.8075598478317261,
+      "num_tokens": 503047719.0,
+      "step": 1338
+    },
+    {
+      "epoch": 1.9925595238095237,
+      "grad_norm": 1.9521377059082243,
+      "learning_rate": 7.522458722907685e-06,
+      "loss": 0.5752,
+      "mean_token_accuracy": 0.814071536064148,
+      "num_tokens": 503415540.0,
+      "step": 1339
+    },
+    {
+      "epoch": 1.994047619047619,
+      "grad_norm": 1.9794655717471314,
+      "learning_rate": 7.517972403852905e-06,
+      "loss": 0.5312,
+      "mean_token_accuracy": 0.8273248076438904,
+      "num_tokens": 503792599.0,
+      "step": 1340
+    },
+    {
+      "epoch": 1.9955357142857144,
+      "grad_norm": 2.006190088228822,
+      "learning_rate": 7.513483367188335e-06,
+      "loss": 0.601,
+      "mean_token_accuracy": 0.8113552927970886,
+      "num_tokens": 504171691.0,
+      "step": 1341
+    },
+    {
+      "epoch": 1.9970238095238095,
+      "grad_norm": 2.1357849136687657,
+      "learning_rate": 7.508991617758923e-06,
+      "loss": 0.623,
+      "mean_token_accuracy": 0.7989252805709839,
+      "num_tokens": 504545379.0,
+      "step": 1342
+    },
+    {
+      "epoch": 1.9985119047619047,
+      "grad_norm": 2.037829631952838,
+      "learning_rate": 7.504497160412543e-06,
+      "loss": 0.5825,
+      "mean_token_accuracy": 0.8118058443069458,
+      "num_tokens": 504923509.0,
+      "step": 1343
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.9231490298714036,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.5805,
+      "mean_token_accuracy": 0.8199220299720764,
+      "num_tokens": 505301560.0,
+      "step": 1344
+    },
+    {
+      "epoch": 2.0014880952380953,
+      "grad_norm": 1.937015242486606,
+      "learning_rate": 7.495500141375012e-06,
+      "loss": 0.3654,
+      "mean_token_accuracy": 0.8791698813438416,
+      "num_tokens": 505679137.0,
+      "step": 1345
+    },
+    {
+      "epoch": 2.0029761904761907,
+      "grad_norm": 1.8374667339866972,
+      "learning_rate": 7.4909975893942045e-06,
+      "loss": 0.3253,
+      "mean_token_accuracy": 0.8921577334403992,
+      "num_tokens": 506057229.0,
+      "step": 1346
+    },
+    {
+      "epoch": 2.0044642857142856,
+      "grad_norm": 1.6289057337213202,
+      "learning_rate": 7.486492348917119e-06,
+      "loss": 0.3236,
+      "mean_token_accuracy": 0.8931748270988464,
+      "num_tokens": 506429117.0,
+      "step": 1347
+    },
+    {
+      "epoch": 2.005952380952381,
+      "grad_norm": 1.6505084118609863,
+      "learning_rate": 7.48198442480619e-06,
+      "loss": 0.2933,
+      "mean_token_accuracy": 0.9025298953056335,
+      "num_tokens": 506799370.0,
+      "step": 1348
+    },
+    {
+      "epoch": 2.0074404761904763,
+      "grad_norm": 1.6538818454434743,
+      "learning_rate": 7.477473821926754e-06,
+      "loss": 0.2887,
+      "mean_token_accuracy": 0.8970656991004944,
+      "num_tokens": 507176997.0,
+      "step": 1349
+    },
+    {
+      "epoch": 2.0089285714285716,
+      "grad_norm": 1.7581005903580427,
+      "learning_rate": 7.472960545147038e-06,
+      "loss": 0.2873,
+      "mean_token_accuracy": 0.9016923308372498,
+      "num_tokens": 507554556.0,
+      "step": 1350
+    },
+    {
+      "epoch": 2.0104166666666665,
+      "grad_norm": 2.3232948407819665,
+      "learning_rate": 7.468444599338152e-06,
+      "loss": 0.2893,
+      "mean_token_accuracy": 0.9026336669921875,
+      "num_tokens": 507931762.0,
+      "step": 1351
+    },
+    {
+      "epoch": 2.011904761904762,
+      "grad_norm": 2.596411953273291,
+      "learning_rate": 7.463925989374089e-06,
+      "loss": 0.2971,
+      "mean_token_accuracy": 0.9002493619918823,
+      "num_tokens": 508306328.0,
+      "step": 1352
+    },
+    {
+      "epoch": 2.013392857142857,
+      "grad_norm": 2.268752295097305,
+      "learning_rate": 7.459404720131717e-06,
+      "loss": 0.3097,
+      "mean_token_accuracy": 0.8916487097740173,
+      "num_tokens": 508683066.0,
+      "step": 1353
+    },
+    {
+      "epoch": 2.0148809523809526,
+      "grad_norm": 2.1860818395216293,
+      "learning_rate": 7.454880796490776e-06,
+      "loss": 0.2792,
+      "mean_token_accuracy": 0.9065107107162476,
+      "num_tokens": 509059967.0,
+      "step": 1354
+    },
+    {
+      "epoch": 2.0163690476190474,
+      "grad_norm": 1.8089584036588426,
+      "learning_rate": 7.4503542233338635e-06,
+      "loss": 0.2775,
+      "mean_token_accuracy": 0.9019590020179749,
+      "num_tokens": 509438434.0,
+      "step": 1355
+    },
+    {
+      "epoch": 2.017857142857143,
+      "grad_norm": 1.707524881607386,
+      "learning_rate": 7.445825005546448e-06,
+      "loss": 0.2915,
+      "mean_token_accuracy": 0.8978238701820374,
+      "num_tokens": 509816907.0,
+      "step": 1356
+    },
+    {
+      "epoch": 2.019345238095238,
+      "grad_norm": 1.6757407058851055,
+      "learning_rate": 7.441293148016842e-06,
+      "loss": 0.2669,
+      "mean_token_accuracy": 0.9099091291427612,
+      "num_tokens": 510196282.0,
+      "step": 1357
+    },
+    {
+      "epoch": 2.0208333333333335,
+      "grad_norm": 1.73006011724291,
+      "learning_rate": 7.4367586556362125e-06,
+      "loss": 0.32,
+      "mean_token_accuracy": 0.8898230195045471,
+      "num_tokens": 510575893.0,
+      "step": 1358
+    },
+    {
+      "epoch": 2.0223214285714284,
+      "grad_norm": 1.7772092033461206,
+      "learning_rate": 7.43222153329857e-06,
+      "loss": 0.2874,
+      "mean_token_accuracy": 0.9002301692962646,
+      "num_tokens": 510937294.0,
+      "step": 1359
+    },
+    {
+      "epoch": 2.0238095238095237,
+      "grad_norm": 1.6858897137487399,
+      "learning_rate": 7.4276817859007615e-06,
+      "loss": 0.2999,
+      "mean_token_accuracy": 0.8994669914245605,
+      "num_tokens": 511316109.0,
+      "step": 1360
+    },
+    {
+      "epoch": 2.025297619047619,
+      "grad_norm": 1.6439868172266314,
+      "learning_rate": 7.423139418342469e-06,
+      "loss": 0.2879,
+      "mean_token_accuracy": 0.902799129486084,
+      "num_tokens": 511682147.0,
+      "step": 1361
+    },
+    {
+      "epoch": 2.0267857142857144,
+      "grad_norm": 1.650651967174933,
+      "learning_rate": 7.4185944355261996e-06,
+      "loss": 0.2982,
+      "mean_token_accuracy": 0.8965124487876892,
+      "num_tokens": 512060553.0,
+      "step": 1362
+    },
+    {
+      "epoch": 2.0282738095238093,
+      "grad_norm": 1.785927900270129,
+      "learning_rate": 7.414046842357288e-06,
+      "loss": 0.2999,
+      "mean_token_accuracy": 0.8985840678215027,
+      "num_tokens": 512438817.0,
+      "step": 1363
+    },
+    {
+      "epoch": 2.0297619047619047,
+      "grad_norm": 1.6991236364371443,
+      "learning_rate": 7.40949664374388e-06,
+      "loss": 0.2984,
+      "mean_token_accuracy": 0.8984381556510925,
+      "num_tokens": 512809104.0,
+      "step": 1364
+    },
+    {
+      "epoch": 2.03125,
+      "grad_norm": 1.761568072227793,
+      "learning_rate": 7.404943844596939e-06,
+      "loss": 0.3113,
+      "mean_token_accuracy": 0.8947325348854065,
+      "num_tokens": 513183796.0,
+      "step": 1365
+    },
+    {
+      "epoch": 2.0327380952380953,
+      "grad_norm": 1.7706269620154853,
+      "learning_rate": 7.4003884498302315e-06,
+      "loss": 0.3089,
+      "mean_token_accuracy": 0.892314076423645,
+      "num_tokens": 513562355.0,
+      "step": 1366
+    },
+    {
+      "epoch": 2.0342261904761907,
+      "grad_norm": 1.677065483499712,
+      "learning_rate": 7.395830464360329e-06,
+      "loss": 0.2696,
+      "mean_token_accuracy": 0.9071820974349976,
+      "num_tokens": 513940721.0,
+      "step": 1367
+    },
+    {
+      "epoch": 2.0357142857142856,
+      "grad_norm": 1.837196273894349,
+      "learning_rate": 7.391269893106592e-06,
+      "loss": 0.2672,
+      "mean_token_accuracy": 0.9060103297233582,
+      "num_tokens": 514305706.0,
+      "step": 1368
+    },
+    {
+      "epoch": 2.037202380952381,
+      "grad_norm": 1.8622080717788052,
+      "learning_rate": 7.3867067409911796e-06,
+      "loss": 0.2904,
+      "mean_token_accuracy": 0.8895711302757263,
+      "num_tokens": 514683184.0,
+      "step": 1369
+    },
+    {
+      "epoch": 2.0386904761904763,
+      "grad_norm": 1.8601598301838298,
+      "learning_rate": 7.382141012939034e-06,
+      "loss": 0.3016,
+      "mean_token_accuracy": 0.9007481932640076,
+      "num_tokens": 515060666.0,
+      "step": 1370
+    },
+    {
+      "epoch": 2.0401785714285716,
+      "grad_norm": 1.9217186719358688,
+      "learning_rate": 7.3775727138778776e-06,
+      "loss": 0.2908,
+      "mean_token_accuracy": 0.9029501676559448,
+      "num_tokens": 515438248.0,
+      "step": 1371
+    },
+    {
+      "epoch": 2.0416666666666665,
+      "grad_norm": 1.7562111348644942,
+      "learning_rate": 7.373001848738203e-06,
+      "loss": 0.3046,
+      "mean_token_accuracy": 0.8944799900054932,
+      "num_tokens": 515817255.0,
+      "step": 1372
+    },
+    {
+      "epoch": 2.043154761904762,
+      "grad_norm": 1.7810592497002646,
+      "learning_rate": 7.36842842245328e-06,
+      "loss": 0.2706,
+      "mean_token_accuracy": 0.9054414629936218,
+      "num_tokens": 516195238.0,
+      "step": 1373
+    },
+    {
+      "epoch": 2.044642857142857,
+      "grad_norm": 1.7283124561912095,
+      "learning_rate": 7.363852439959136e-06,
+      "loss": 0.2994,
+      "mean_token_accuracy": 0.8973888754844666,
+      "num_tokens": 516573466.0,
+      "step": 1374
+    },
+    {
+      "epoch": 2.0461309523809526,
+      "grad_norm": 1.8384605079340577,
+      "learning_rate": 7.359273906194562e-06,
+      "loss": 0.3154,
+      "mean_token_accuracy": 0.8948285579681396,
+      "num_tokens": 516951352.0,
+      "step": 1375
+    },
+    {
+      "epoch": 2.0476190476190474,
+      "grad_norm": 1.7496598639679,
+      "learning_rate": 7.354692826101102e-06,
+      "loss": 0.2942,
+      "mean_token_accuracy": 0.8966860175132751,
+      "num_tokens": 517331121.0,
+      "step": 1376
+    },
+    {
+      "epoch": 2.049107142857143,
+      "grad_norm": 1.8780904889476715,
+      "learning_rate": 7.350109204623044e-06,
+      "loss": 0.2803,
+      "mean_token_accuracy": 0.9057727456092834,
+      "num_tokens": 517707968.0,
+      "step": 1377
+    },
+    {
+      "epoch": 2.050595238095238,
+      "grad_norm": 1.675012503970546,
+      "learning_rate": 7.345523046707423e-06,
+      "loss": 0.2687,
+      "mean_token_accuracy": 0.9074685573577881,
+      "num_tokens": 518086597.0,
+      "step": 1378
+    },
+    {
+      "epoch": 2.0520833333333335,
+      "grad_norm": 1.685457476574697,
+      "learning_rate": 7.340934357304011e-06,
+      "loss": 0.2881,
+      "mean_token_accuracy": 0.899284303188324,
+      "num_tokens": 518465067.0,
+      "step": 1379
+    },
+    {
+      "epoch": 2.0535714285714284,
+      "grad_norm": 1.8327256626316681,
+      "learning_rate": 7.336343141365311e-06,
+      "loss": 0.301,
+      "mean_token_accuracy": 0.8930439352989197,
+      "num_tokens": 518843389.0,
+      "step": 1380
+    },
+    {
+      "epoch": 2.0550595238095237,
+      "grad_norm": 1.7111079944919707,
+      "learning_rate": 7.331749403846553e-06,
+      "loss": 0.2864,
+      "mean_token_accuracy": 0.8980436325073242,
+      "num_tokens": 519219658.0,
+      "step": 1381
+    },
+    {
+      "epoch": 2.056547619047619,
+      "grad_norm": 1.8089455826872933,
+      "learning_rate": 7.32715314970569e-06,
+      "loss": 0.3042,
+      "mean_token_accuracy": 0.8938557505607605,
+      "num_tokens": 519597396.0,
+      "step": 1382
+    },
+    {
+      "epoch": 2.0580357142857144,
+      "grad_norm": 1.8669970790160828,
+      "learning_rate": 7.322554383903389e-06,
+      "loss": 0.2955,
+      "mean_token_accuracy": 0.9003272652626038,
+      "num_tokens": 519974460.0,
+      "step": 1383
+    },
+    {
+      "epoch": 2.0595238095238093,
+      "grad_norm": 1.8669746020869826,
+      "learning_rate": 7.317953111403029e-06,
+      "loss": 0.3173,
+      "mean_token_accuracy": 0.8930256962776184,
+      "num_tokens": 520352527.0,
+      "step": 1384
+    },
+    {
+      "epoch": 2.0610119047619047,
+      "grad_norm": 1.8238628370855847,
+      "learning_rate": 7.3133493371706946e-06,
+      "loss": 0.2377,
+      "mean_token_accuracy": 0.9152621626853943,
+      "num_tokens": 520729696.0,
+      "step": 1385
+    },
+    {
+      "epoch": 2.0625,
+      "grad_norm": 1.8920906313810442,
+      "learning_rate": 7.308743066175172e-06,
+      "loss": 0.3595,
+      "mean_token_accuracy": 0.8773263692855835,
+      "num_tokens": 521109904.0,
+      "step": 1386
+    },
+    {
+      "epoch": 2.0639880952380953,
+      "grad_norm": 1.732623294611645,
+      "learning_rate": 7.304134303387936e-06,
+      "loss": 0.2663,
+      "mean_token_accuracy": 0.9099399447441101,
+      "num_tokens": 521476478.0,
+      "step": 1387
+    },
+    {
+      "epoch": 2.0654761904761907,
+      "grad_norm": 2.0431770402036866,
+      "learning_rate": 7.299523053783157e-06,
+      "loss": 0.2953,
+      "mean_token_accuracy": 0.9028152823448181,
+      "num_tokens": 521854252.0,
+      "step": 1388
+    },
+    {
+      "epoch": 2.0669642857142856,
+      "grad_norm": 1.8290126135557332,
+      "learning_rate": 7.294909322337689e-06,
+      "loss": 0.2972,
+      "mean_token_accuracy": 0.8957341909408569,
+      "num_tokens": 522232399.0,
+      "step": 1389
+    },
+    {
+      "epoch": 2.068452380952381,
+      "grad_norm": 1.7459437884559834,
+      "learning_rate": 7.290293114031062e-06,
+      "loss": 0.2786,
+      "mean_token_accuracy": 0.8999309539794922,
+      "num_tokens": 522610599.0,
+      "step": 1390
+    },
+    {
+      "epoch": 2.0699404761904763,
+      "grad_norm": 1.706881095264249,
+      "learning_rate": 7.28567443384548e-06,
+      "loss": 0.3345,
+      "mean_token_accuracy": 0.8868281841278076,
+      "num_tokens": 522990510.0,
+      "step": 1391
+    },
+    {
+      "epoch": 2.0714285714285716,
+      "grad_norm": 1.6299118288931922,
+      "learning_rate": 7.281053286765816e-06,
+      "loss": 0.3016,
+      "mean_token_accuracy": 0.8968909978866577,
+      "num_tokens": 523369927.0,
+      "step": 1392
+    },
+    {
+      "epoch": 2.0729166666666665,
+      "grad_norm": 1.6857251571864384,
+      "learning_rate": 7.276429677779603e-06,
+      "loss": 0.268,
+      "mean_token_accuracy": 0.9049781560897827,
+      "num_tokens": 523747973.0,
+      "step": 1393
+    },
+    {
+      "epoch": 2.074404761904762,
+      "grad_norm": 1.881665438028438,
+      "learning_rate": 7.271803611877033e-06,
+      "loss": 0.3289,
+      "mean_token_accuracy": 0.8844859004020691,
+      "num_tokens": 524126231.0,
+      "step": 1394
+    },
+    {
+      "epoch": 2.075892857142857,
+      "grad_norm": 1.739421206913304,
+      "learning_rate": 7.2671750940509526e-06,
+      "loss": 0.2992,
+      "mean_token_accuracy": 0.9004305005073547,
+      "num_tokens": 524505799.0,
+      "step": 1395
+    },
+    {
+      "epoch": 2.0773809523809526,
+      "grad_norm": 1.7949818278093266,
+      "learning_rate": 7.262544129296848e-06,
+      "loss": 0.27,
+      "mean_token_accuracy": 0.8971434235572815,
+      "num_tokens": 524881335.0,
+      "step": 1396
+    },
+    {
+      "epoch": 2.0788690476190474,
+      "grad_norm": 1.7734753851815557,
+      "learning_rate": 7.25791072261285e-06,
+      "loss": 0.2868,
+      "mean_token_accuracy": 0.8995347023010254,
+      "num_tokens": 525259087.0,
+      "step": 1397
+    },
+    {
+      "epoch": 2.080357142857143,
+      "grad_norm": 1.730491011658241,
+      "learning_rate": 7.253274878999728e-06,
+      "loss": 0.3062,
+      "mean_token_accuracy": 0.8912984132766724,
+      "num_tokens": 525636928.0,
+      "step": 1398
+    },
+    {
+      "epoch": 2.081845238095238,
+      "grad_norm": 1.7249374150903425,
+      "learning_rate": 7.2486366034608734e-06,
+      "loss": 0.2964,
+      "mean_token_accuracy": 0.901947557926178,
+      "num_tokens": 526015387.0,
+      "step": 1399
+    },
+    {
+      "epoch": 2.0833333333333335,
+      "grad_norm": 1.6462001232826764,
+      "learning_rate": 7.243995901002312e-06,
+      "loss": 0.2641,
+      "mean_token_accuracy": 0.9054259657859802,
+      "num_tokens": 526393776.0,
+      "step": 1400
+    },
+    {
+      "epoch": 2.0848214285714284,
+      "grad_norm": 1.8083307951406014,
+      "learning_rate": 7.239352776632681e-06,
+      "loss": 0.2859,
+      "mean_token_accuracy": 0.9002729058265686,
+      "num_tokens": 526773968.0,
+      "step": 1401
+    },
+    {
+      "epoch": 2.0863095238095237,
+      "grad_norm": 1.7475961708638852,
+      "learning_rate": 7.234707235363238e-06,
+      "loss": 0.2812,
+      "mean_token_accuracy": 0.9025484323501587,
+      "num_tokens": 527140354.0,
+      "step": 1402
+    },
+    {
+      "epoch": 2.087797619047619,
+      "grad_norm": 1.7350750614356545,
+      "learning_rate": 7.230059282207842e-06,
+      "loss": 0.3107,
+      "mean_token_accuracy": 0.894589900970459,
+      "num_tokens": 527518907.0,
+      "step": 1403
+    },
+    {
+      "epoch": 2.0892857142857144,
+      "grad_norm": 1.6717620281275023,
+      "learning_rate": 7.225408922182962e-06,
+      "loss": 0.2635,
+      "mean_token_accuracy": 0.9091618657112122,
+      "num_tokens": 527889484.0,
+      "step": 1404
+    },
+    {
+      "epoch": 2.0907738095238093,
+      "grad_norm": 1.8540721752752265,
+      "learning_rate": 7.2207561603076625e-06,
+      "loss": 0.2875,
+      "mean_token_accuracy": 0.9002428650856018,
+      "num_tokens": 528268173.0,
+      "step": 1405
+    },
+    {
+      "epoch": 2.0922619047619047,
+      "grad_norm": 1.8201999424606299,
+      "learning_rate": 7.216101001603596e-06,
+      "loss": 0.2708,
+      "mean_token_accuracy": 0.9093462228775024,
+      "num_tokens": 528644060.0,
+      "step": 1406
+    },
+    {
+      "epoch": 2.09375,
+      "grad_norm": 1.7492285730021733,
+      "learning_rate": 7.211443451095007e-06,
+      "loss": 0.2831,
+      "mean_token_accuracy": 0.8986604809761047,
+      "num_tokens": 529015109.0,
+      "step": 1407
+    },
+    {
+      "epoch": 2.0952380952380953,
+      "grad_norm": 1.8749647065547936,
+      "learning_rate": 7.206783513808721e-06,
+      "loss": 0.3161,
+      "mean_token_accuracy": 0.8914902210235596,
+      "num_tokens": 529394659.0,
+      "step": 1408
+    },
+    {
+      "epoch": 2.0967261904761907,
+      "grad_norm": 1.8617673966346209,
+      "learning_rate": 7.202121194774137e-06,
+      "loss": 0.2836,
+      "mean_token_accuracy": 0.9019750356674194,
+      "num_tokens": 529758304.0,
+      "step": 1409
+    },
+    {
+      "epoch": 2.0982142857142856,
+      "grad_norm": 1.7507587766434816,
+      "learning_rate": 7.197456499023226e-06,
+      "loss": 0.3424,
+      "mean_token_accuracy": 0.8829914927482605,
+      "num_tokens": 530138887.0,
+      "step": 1410
+    },
+    {
+      "epoch": 2.099702380952381,
+      "grad_norm": 1.9242985365360887,
+      "learning_rate": 7.1927894315905246e-06,
+      "loss": 0.304,
+      "mean_token_accuracy": 0.8962695002555847,
+      "num_tokens": 530516125.0,
+      "step": 1411
+    },
+    {
+      "epoch": 2.1011904761904763,
+      "grad_norm": 1.7441631630642547,
+      "learning_rate": 7.188119997513127e-06,
+      "loss": 0.3347,
+      "mean_token_accuracy": 0.8808237314224243,
+      "num_tokens": 530895839.0,
+      "step": 1412
+    },
+    {
+      "epoch": 2.1026785714285716,
+      "grad_norm": 1.8243547348799227,
+      "learning_rate": 7.183448201830685e-06,
+      "loss": 0.308,
+      "mean_token_accuracy": 0.8969010710716248,
+      "num_tokens": 531274842.0,
+      "step": 1413
+    },
+    {
+      "epoch": 2.1041666666666665,
+      "grad_norm": 1.9004326613370692,
+      "learning_rate": 7.178774049585397e-06,
+      "loss": 0.3356,
+      "mean_token_accuracy": 0.8839803338050842,
+      "num_tokens": 531652974.0,
+      "step": 1414
+    },
+    {
+      "epoch": 2.105654761904762,
+      "grad_norm": 1.7886281490555331,
+      "learning_rate": 7.174097545822006e-06,
+      "loss": 0.2667,
+      "mean_token_accuracy": 0.908664882183075,
+      "num_tokens": 532030372.0,
+      "step": 1415
+    },
+    {
+      "epoch": 2.107142857142857,
+      "grad_norm": 1.9193138690262899,
+      "learning_rate": 7.169418695587791e-06,
+      "loss": 0.2964,
+      "mean_token_accuracy": 0.8983780741691589,
+      "num_tokens": 532395321.0,
+      "step": 1416
+    },
+    {
+      "epoch": 2.1086309523809526,
+      "grad_norm": 1.7636593612804257,
+      "learning_rate": 7.164737503932565e-06,
+      "loss": 0.2871,
+      "mean_token_accuracy": 0.8984881043434143,
+      "num_tokens": 532774523.0,
+      "step": 1417
+    },
+    {
+      "epoch": 2.1101190476190474,
+      "grad_norm": 1.6834451947261173,
+      "learning_rate": 7.16005397590867e-06,
+      "loss": 0.2958,
+      "mean_token_accuracy": 0.8985792994499207,
+      "num_tokens": 533154342.0,
+      "step": 1418
+    },
+    {
+      "epoch": 2.111607142857143,
+      "grad_norm": 1.7097681221119538,
+      "learning_rate": 7.155368116570963e-06,
+      "loss": 0.2999,
+      "mean_token_accuracy": 0.8940982222557068,
+      "num_tokens": 533534143.0,
+      "step": 1419
+    },
+    {
+      "epoch": 2.113095238095238,
+      "grad_norm": 1.748128420585396,
+      "learning_rate": 7.150679930976826e-06,
+      "loss": 0.2971,
+      "mean_token_accuracy": 0.8986480832099915,
+      "num_tokens": 533912813.0,
+      "step": 1420
+    },
+    {
+      "epoch": 2.1145833333333335,
+      "grad_norm": 1.8064548796089281,
+      "learning_rate": 7.145989424186146e-06,
+      "loss": 0.3028,
+      "mean_token_accuracy": 0.8967759609222412,
+      "num_tokens": 534290195.0,
+      "step": 1421
+    },
+    {
+      "epoch": 2.1160714285714284,
+      "grad_norm": 1.9038278710111085,
+      "learning_rate": 7.1412966012613135e-06,
+      "loss": 0.2915,
+      "mean_token_accuracy": 0.8958351016044617,
+      "num_tokens": 534668600.0,
+      "step": 1422
+    },
+    {
+      "epoch": 2.1175595238095237,
+      "grad_norm": 1.8448394071004248,
+      "learning_rate": 7.136601467267226e-06,
+      "loss": 0.3256,
+      "mean_token_accuracy": 0.8871756792068481,
+      "num_tokens": 535048202.0,
+      "step": 1423
+    },
+    {
+      "epoch": 2.119047619047619,
+      "grad_norm": 1.8398441541084367,
+      "learning_rate": 7.1319040272712705e-06,
+      "loss": 0.3084,
+      "mean_token_accuracy": 0.8969117403030396,
+      "num_tokens": 535426046.0,
+      "step": 1424
+    },
+    {
+      "epoch": 2.1205357142857144,
+      "grad_norm": 1.7954323926777143,
+      "learning_rate": 7.127204286343321e-06,
+      "loss": 0.2979,
+      "mean_token_accuracy": 0.9014877676963806,
+      "num_tokens": 535792606.0,
+      "step": 1425
+    },
+    {
+      "epoch": 2.1220238095238093,
+      "grad_norm": 1.7082818365862866,
+      "learning_rate": 7.122502249555738e-06,
+      "loss": 0.3047,
+      "mean_token_accuracy": 0.8938140869140625,
+      "num_tokens": 536172087.0,
+      "step": 1426
+    },
+    {
+      "epoch": 2.1235119047619047,
+      "grad_norm": 1.7055534178107867,
+      "learning_rate": 7.117797921983361e-06,
+      "loss": 0.3234,
+      "mean_token_accuracy": 0.8906872868537903,
+      "num_tokens": 536552701.0,
+      "step": 1427
+    },
+    {
+      "epoch": 2.125,
+      "grad_norm": 1.7044693816122842,
+      "learning_rate": 7.113091308703498e-06,
+      "loss": 0.3108,
+      "mean_token_accuracy": 0.8935633301734924,
+      "num_tokens": 536931916.0,
+      "step": 1428
+    },
+    {
+      "epoch": 2.1264880952380953,
+      "grad_norm": 1.717873959024712,
+      "learning_rate": 7.108382414795927e-06,
+      "loss": 0.2333,
+      "mean_token_accuracy": 0.9158697724342346,
+      "num_tokens": 537309005.0,
+      "step": 1429
+    },
+    {
+      "epoch": 2.1279761904761907,
+      "grad_norm": 1.7655406956603161,
+      "learning_rate": 7.103671245342887e-06,
+      "loss": 0.2967,
+      "mean_token_accuracy": 0.8956810235977173,
+      "num_tokens": 537687695.0,
+      "step": 1430
+    },
+    {
+      "epoch": 2.1294642857142856,
+      "grad_norm": 1.6709794256287938,
+      "learning_rate": 7.098957805429072e-06,
+      "loss": 0.3016,
+      "mean_token_accuracy": 0.896006166934967,
+      "num_tokens": 538062213.0,
+      "step": 1431
+    },
+    {
+      "epoch": 2.130952380952381,
+      "grad_norm": 1.8737818799921475,
+      "learning_rate": 7.094242100141625e-06,
+      "loss": 0.3278,
+      "mean_token_accuracy": 0.8882167935371399,
+      "num_tokens": 538440741.0,
+      "step": 1432
+    },
+    {
+      "epoch": 2.1324404761904763,
+      "grad_norm": 1.6506092968996666,
+      "learning_rate": 7.089524134570138e-06,
+      "loss": 0.2992,
+      "mean_token_accuracy": 0.8982983231544495,
+      "num_tokens": 538820767.0,
+      "step": 1433
+    },
+    {
+      "epoch": 2.1339285714285716,
+      "grad_norm": 1.6314277054881146,
+      "learning_rate": 7.084803913806642e-06,
+      "loss": 0.3211,
+      "mean_token_accuracy": 0.8904447555541992,
+      "num_tokens": 539200960.0,
+      "step": 1434
+    },
+    {
+      "epoch": 2.1354166666666665,
+      "grad_norm": 1.7299574276940568,
+      "learning_rate": 7.080081442945597e-06,
+      "loss": 0.3162,
+      "mean_token_accuracy": 0.8898965716362,
+      "num_tokens": 539580470.0,
+      "step": 1435
+    },
+    {
+      "epoch": 2.136904761904762,
+      "grad_norm": 1.7664178959467265,
+      "learning_rate": 7.075356727083894e-06,
+      "loss": 0.2941,
+      "mean_token_accuracy": 0.9030594825744629,
+      "num_tokens": 539959353.0,
+      "step": 1436
+    },
+    {
+      "epoch": 2.138392857142857,
+      "grad_norm": 1.8345054316409037,
+      "learning_rate": 7.070629771320853e-06,
+      "loss": 0.3036,
+      "mean_token_accuracy": 0.8956897854804993,
+      "num_tokens": 540337620.0,
+      "step": 1437
+    },
+    {
+      "epoch": 2.1398809523809526,
+      "grad_norm": 1.8534450757861471,
+      "learning_rate": 7.065900580758202e-06,
+      "loss": 0.286,
+      "mean_token_accuracy": 0.9028458595275879,
+      "num_tokens": 540713806.0,
+      "step": 1438
+    },
+    {
+      "epoch": 2.1413690476190474,
+      "grad_norm": 1.8826775536403104,
+      "learning_rate": 7.061169160500088e-06,
+      "loss": 0.3243,
+      "mean_token_accuracy": 0.8840871453285217,
+      "num_tokens": 541091983.0,
+      "step": 1439
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 1.7115875313347844,
+      "learning_rate": 7.056435515653059e-06,
+      "loss": 0.2883,
+      "mean_token_accuracy": 0.9017356038093567,
+      "num_tokens": 541470812.0,
+      "step": 1440
+    },
+    {
+      "epoch": 2.144345238095238,
+      "grad_norm": 1.8708768818801123,
+      "learning_rate": 7.051699651326068e-06,
+      "loss": 0.3233,
+      "mean_token_accuracy": 0.8894736766815186,
+      "num_tokens": 541848566.0,
+      "step": 1441
+    },
+    {
+      "epoch": 2.1458333333333335,
+      "grad_norm": 1.715038877360646,
+      "learning_rate": 7.046961572630463e-06,
+      "loss": 0.288,
+      "mean_token_accuracy": 0.901122510433197,
+      "num_tokens": 542215545.0,
+      "step": 1442
+    },
+    {
+      "epoch": 2.1473214285714284,
+      "grad_norm": 1.725809561388768,
+      "learning_rate": 7.042221284679982e-06,
+      "loss": 0.2413,
+      "mean_token_accuracy": 0.9143549799919128,
+      "num_tokens": 542592377.0,
+      "step": 1443
+    },
+    {
+      "epoch": 2.1488095238095237,
+      "grad_norm": 1.870416208797295,
+      "learning_rate": 7.037478792590744e-06,
+      "loss": 0.2991,
+      "mean_token_accuracy": 0.897419810295105,
+      "num_tokens": 542970373.0,
+      "step": 1444
+    },
+    {
+      "epoch": 2.150297619047619,
+      "grad_norm": 2.020107812475765,
+      "learning_rate": 7.0327341014812515e-06,
+      "loss": 0.2814,
+      "mean_token_accuracy": 0.8982741236686707,
+      "num_tokens": 543347327.0,
+      "step": 1445
+    },
+    {
+      "epoch": 2.1517857142857144,
+      "grad_norm": 1.7532194275631914,
+      "learning_rate": 7.027987216472376e-06,
+      "loss": 0.2928,
+      "mean_token_accuracy": 0.8984127044677734,
+      "num_tokens": 543713450.0,
+      "step": 1446
+    },
+    {
+      "epoch": 2.1532738095238093,
+      "grad_norm": 1.7306464710480598,
+      "learning_rate": 7.023238142687362e-06,
+      "loss": 0.2819,
+      "mean_token_accuracy": 0.9004485011100769,
+      "num_tokens": 544087420.0,
+      "step": 1447
+    },
+    {
+      "epoch": 2.1547619047619047,
+      "grad_norm": 1.8501764831317549,
+      "learning_rate": 7.0184868852518114e-06,
+      "loss": 0.2896,
+      "mean_token_accuracy": 0.8975953459739685,
+      "num_tokens": 544453114.0,
+      "step": 1448
+    },
+    {
+      "epoch": 2.15625,
+      "grad_norm": 1.7779717881925226,
+      "learning_rate": 7.0137334492936875e-06,
+      "loss": 0.305,
+      "mean_token_accuracy": 0.8931586146354675,
+      "num_tokens": 544829250.0,
+      "step": 1449
+    },
+    {
+      "epoch": 2.1577380952380953,
+      "grad_norm": 1.8458221169685805,
+      "learning_rate": 7.0089778399433e-06,
+      "loss": 0.2983,
+      "mean_token_accuracy": 0.893975019454956,
+      "num_tokens": 545192205.0,
+      "step": 1450
+    },
+    {
+      "epoch": 2.1592261904761907,
+      "grad_norm": 1.7405145892621534,
+      "learning_rate": 7.004220062333306e-06,
+      "loss": 0.2859,
+      "mean_token_accuracy": 0.9005154371261597,
+      "num_tokens": 545570227.0,
+      "step": 1451
+    },
+    {
+      "epoch": 2.1607142857142856,
+      "grad_norm": 1.7097662671167706,
+      "learning_rate": 6.999460121598704e-06,
+      "loss": 0.3098,
+      "mean_token_accuracy": 0.8966336250305176,
+      "num_tokens": 545949402.0,
+      "step": 1452
+    },
+    {
+      "epoch": 2.162202380952381,
+      "grad_norm": 1.7211086070668662,
+      "learning_rate": 6.99469802287683e-06,
+      "loss": 0.299,
+      "mean_token_accuracy": 0.8952515125274658,
+      "num_tokens": 546327596.0,
+      "step": 1453
+    },
+    {
+      "epoch": 2.1636904761904763,
+      "grad_norm": 1.6476739232308233,
+      "learning_rate": 6.989933771307342e-06,
+      "loss": 0.2872,
+      "mean_token_accuracy": 0.900929868221283,
+      "num_tokens": 546695498.0,
+      "step": 1454
+    },
+    {
+      "epoch": 2.1651785714285716,
+      "grad_norm": 1.8337011112598078,
+      "learning_rate": 6.985167372032226e-06,
+      "loss": 0.3208,
+      "mean_token_accuracy": 0.8955374956130981,
+      "num_tokens": 547074415.0,
+      "step": 1455
+    },
+    {
+      "epoch": 2.1666666666666665,
+      "grad_norm": 1.6958374787836457,
+      "learning_rate": 6.980398830195785e-06,
+      "loss": 0.2884,
+      "mean_token_accuracy": 0.9010039567947388,
+      "num_tokens": 547452742.0,
+      "step": 1456
+    },
+    {
+      "epoch": 2.168154761904762,
+      "grad_norm": 1.7825360674808397,
+      "learning_rate": 6.9756281509446345e-06,
+      "loss": 0.3022,
+      "mean_token_accuracy": 0.8936963677406311,
+      "num_tokens": 547831207.0,
+      "step": 1457
+    },
+    {
+      "epoch": 2.169642857142857,
+      "grad_norm": 1.770333561768529,
+      "learning_rate": 6.970855339427698e-06,
+      "loss": 0.2964,
+      "mean_token_accuracy": 0.8950756788253784,
+      "num_tokens": 548209523.0,
+      "step": 1458
+    },
+    {
+      "epoch": 2.1711309523809526,
+      "grad_norm": 1.7893140208664968,
+      "learning_rate": 6.9660804007961976e-06,
+      "loss": 0.2814,
+      "mean_token_accuracy": 0.902541995048523,
+      "num_tokens": 548587930.0,
+      "step": 1459
+    },
+    {
+      "epoch": 2.1726190476190474,
+      "grad_norm": 1.8150541040326036,
+      "learning_rate": 6.961303340203653e-06,
+      "loss": 0.2776,
+      "mean_token_accuracy": 0.9056746363639832,
+      "num_tokens": 548965451.0,
+      "step": 1460
+    },
+    {
+      "epoch": 2.174107142857143,
+      "grad_norm": 1.5983398751619409,
+      "learning_rate": 6.956524162805875e-06,
+      "loss": 0.2931,
+      "mean_token_accuracy": 0.8966838717460632,
+      "num_tokens": 549345719.0,
+      "step": 1461
+    },
+    {
+      "epoch": 2.175595238095238,
+      "grad_norm": 1.8363571265520886,
+      "learning_rate": 6.951742873760959e-06,
+      "loss": 0.3107,
+      "mean_token_accuracy": 0.89678555727005,
+      "num_tokens": 549722830.0,
+      "step": 1462
+    },
+    {
+      "epoch": 2.1770833333333335,
+      "grad_norm": 1.7368992598214081,
+      "learning_rate": 6.946959478229277e-06,
+      "loss": 0.3215,
+      "mean_token_accuracy": 0.8914645314216614,
+      "num_tokens": 550068908.0,
+      "step": 1463
+    },
+    {
+      "epoch": 2.1785714285714284,
+      "grad_norm": 1.6926512691139195,
+      "learning_rate": 6.942173981373474e-06,
+      "loss": 0.2885,
+      "mean_token_accuracy": 0.9025985598564148,
+      "num_tokens": 550435618.0,
+      "step": 1464
+    },
+    {
+      "epoch": 2.1800595238095237,
+      "grad_norm": 1.6589047548983804,
+      "learning_rate": 6.937386388358468e-06,
+      "loss": 0.2982,
+      "mean_token_accuracy": 0.8970934748649597,
+      "num_tokens": 550815540.0,
+      "step": 1465
+    },
+    {
+      "epoch": 2.181547619047619,
+      "grad_norm": 1.71852899821708,
+      "learning_rate": 6.932596704351435e-06,
+      "loss": 0.2953,
+      "mean_token_accuracy": 0.8979154229164124,
+      "num_tokens": 551194236.0,
+      "step": 1466
+    },
+    {
+      "epoch": 2.1830357142857144,
+      "grad_norm": 1.7494163134877987,
+      "learning_rate": 6.927804934521809e-06,
+      "loss": 0.3037,
+      "mean_token_accuracy": 0.8924182057380676,
+      "num_tokens": 551572475.0,
+      "step": 1467
+    },
+    {
+      "epoch": 2.1845238095238093,
+      "grad_norm": 1.7991929242389164,
+      "learning_rate": 6.923011084041276e-06,
+      "loss": 0.3423,
+      "mean_token_accuracy": 0.8809891939163208,
+      "num_tokens": 551952025.0,
+      "step": 1468
+    },
+    {
+      "epoch": 2.1860119047619047,
+      "grad_norm": 1.8078677568749337,
+      "learning_rate": 6.918215158083767e-06,
+      "loss": 0.2875,
+      "mean_token_accuracy": 0.9003677368164062,
+      "num_tokens": 552330175.0,
+      "step": 1469
+    },
+    {
+      "epoch": 2.1875,
+      "grad_norm": 1.7144629577414148,
+      "learning_rate": 6.913417161825449e-06,
+      "loss": 0.3229,
+      "mean_token_accuracy": 0.8878143429756165,
+      "num_tokens": 552707502.0,
+      "step": 1470
+    },
+    {
+      "epoch": 2.1889880952380953,
+      "grad_norm": 1.8290776886286215,
+      "learning_rate": 6.908617100444732e-06,
+      "loss": 0.3114,
+      "mean_token_accuracy": 0.8888502717018127,
+      "num_tokens": 553086109.0,
+      "step": 1471
+    },
+    {
+      "epoch": 2.1904761904761907,
+      "grad_norm": 1.9034539753842554,
+      "learning_rate": 6.903814979122249e-06,
+      "loss": 0.3236,
+      "mean_token_accuracy": 0.891053318977356,
+      "num_tokens": 553464075.0,
+      "step": 1472
+    },
+    {
+      "epoch": 2.1919642857142856,
+      "grad_norm": 1.752721568892251,
+      "learning_rate": 6.8990108030408565e-06,
+      "loss": 0.2807,
+      "mean_token_accuracy": 0.9043166637420654,
+      "num_tokens": 553842483.0,
+      "step": 1473
+    },
+    {
+      "epoch": 2.193452380952381,
+      "grad_norm": 1.7788060878536878,
+      "learning_rate": 6.894204577385628e-06,
+      "loss": 0.2683,
+      "mean_token_accuracy": 0.9080449342727661,
+      "num_tokens": 554220764.0,
+      "step": 1474
+    },
+    {
+      "epoch": 2.1949404761904763,
+      "grad_norm": 1.80427615009062,
+      "learning_rate": 6.8893963073438555e-06,
+      "loss": 0.2926,
+      "mean_token_accuracy": 0.9012938737869263,
+      "num_tokens": 554594256.0,
+      "step": 1475
+    },
+    {
+      "epoch": 2.1964285714285716,
+      "grad_norm": 1.83243690234834,
+      "learning_rate": 6.8845859981050265e-06,
+      "loss": 0.2993,
+      "mean_token_accuracy": 0.8965156078338623,
+      "num_tokens": 554971445.0,
+      "step": 1476
+    },
+    {
+      "epoch": 2.1979166666666665,
+      "grad_norm": 1.834138532508054,
+      "learning_rate": 6.8797736548608405e-06,
+      "loss": 0.304,
+      "mean_token_accuracy": 0.8928667902946472,
+      "num_tokens": 555349016.0,
+      "step": 1477
+    },
+    {
+      "epoch": 2.199404761904762,
+      "grad_norm": 1.9407668353285925,
+      "learning_rate": 6.874959282805186e-06,
+      "loss": 0.2928,
+      "mean_token_accuracy": 0.9006533622741699,
+      "num_tokens": 555725425.0,
+      "step": 1478
+    },
+    {
+      "epoch": 2.200892857142857,
+      "grad_norm": 1.6806511873408772,
+      "learning_rate": 6.870142887134141e-06,
+      "loss": 0.2989,
+      "mean_token_accuracy": 0.8989517092704773,
+      "num_tokens": 556105633.0,
+      "step": 1479
+    },
+    {
+      "epoch": 2.2023809523809526,
+      "grad_norm": 1.7249149261158896,
+      "learning_rate": 6.86532447304597e-06,
+      "loss": 0.2995,
+      "mean_token_accuracy": 0.893403172492981,
+      "num_tokens": 556471980.0,
+      "step": 1480
+    },
+    {
+      "epoch": 2.2038690476190474,
+      "grad_norm": 1.7382324705522583,
+      "learning_rate": 6.860504045741114e-06,
+      "loss": 0.2782,
+      "mean_token_accuracy": 0.9046133160591125,
+      "num_tokens": 556851076.0,
+      "step": 1481
+    },
+    {
+      "epoch": 2.205357142857143,
+      "grad_norm": 1.7709110889394462,
+      "learning_rate": 6.85568161042219e-06,
+      "loss": 0.2885,
+      "mean_token_accuracy": 0.9013537764549255,
+      "num_tokens": 557228985.0,
+      "step": 1482
+    },
+    {
+      "epoch": 2.206845238095238,
+      "grad_norm": 1.6560428575496133,
+      "learning_rate": 6.850857172293979e-06,
+      "loss": 0.2935,
+      "mean_token_accuracy": 0.8967317342758179,
+      "num_tokens": 557603756.0,
+      "step": 1483
+    },
+    {
+      "epoch": 2.2083333333333335,
+      "grad_norm": 1.7150907206071713,
+      "learning_rate": 6.8460307365634225e-06,
+      "loss": 0.283,
+      "mean_token_accuracy": 0.9028594493865967,
+      "num_tokens": 557983558.0,
+      "step": 1484
+    },
+    {
+      "epoch": 2.2098214285714284,
+      "grad_norm": 1.6379691942896881,
+      "learning_rate": 6.841202308439623e-06,
+      "loss": 0.2642,
+      "mean_token_accuracy": 0.9035190343856812,
+      "num_tokens": 558363306.0,
+      "step": 1485
+    },
+    {
+      "epoch": 2.2113095238095237,
+      "grad_norm": 1.831517502172383,
+      "learning_rate": 6.8363718931338286e-06,
+      "loss": 0.3109,
+      "mean_token_accuracy": 0.895564615726471,
+      "num_tokens": 558729376.0,
+      "step": 1486
+    },
+    {
+      "epoch": 2.212797619047619,
+      "grad_norm": 1.7634953673699967,
+      "learning_rate": 6.8315394958594364e-06,
+      "loss": 0.2828,
+      "mean_token_accuracy": 0.9002426862716675,
+      "num_tokens": 559092781.0,
+      "step": 1487
+    },
+    {
+      "epoch": 2.2142857142857144,
+      "grad_norm": 1.6942351778887181,
+      "learning_rate": 6.8267051218319766e-06,
+      "loss": 0.2864,
+      "mean_token_accuracy": 0.9014855623245239,
+      "num_tokens": 559470793.0,
+      "step": 1488
+    },
+    {
+      "epoch": 2.2157738095238093,
+      "grad_norm": 1.7971276281059094,
+      "learning_rate": 6.8218687762691186e-06,
+      "loss": 0.3151,
+      "mean_token_accuracy": 0.8908408880233765,
+      "num_tokens": 559848447.0,
+      "step": 1489
+    },
+    {
+      "epoch": 2.2172619047619047,
+      "grad_norm": 1.714219638010471,
+      "learning_rate": 6.817030464390656e-06,
+      "loss": 0.2838,
+      "mean_token_accuracy": 0.9006674289703369,
+      "num_tokens": 560226781.0,
+      "step": 1490
+    },
+    {
+      "epoch": 2.21875,
+      "grad_norm": 1.73805486920442,
+      "learning_rate": 6.812190191418508e-06,
+      "loss": 0.242,
+      "mean_token_accuracy": 0.9131200909614563,
+      "num_tokens": 560571383.0,
+      "step": 1491
+    },
+    {
+      "epoch": 2.2202380952380953,
+      "grad_norm": 1.7854032808311673,
+      "learning_rate": 6.80734796257671e-06,
+      "loss": 0.3239,
+      "mean_token_accuracy": 0.892177939414978,
+      "num_tokens": 560950819.0,
+      "step": 1492
+    },
+    {
+      "epoch": 2.2217261904761907,
+      "grad_norm": 1.803959221918087,
+      "learning_rate": 6.802503783091403e-06,
+      "loss": 0.3089,
+      "mean_token_accuracy": 0.8939521312713623,
+      "num_tokens": 561318049.0,
+      "step": 1493
+    },
+    {
+      "epoch": 2.2232142857142856,
+      "grad_norm": 1.880272472583035,
+      "learning_rate": 6.797657658190838e-06,
+      "loss": 0.2905,
+      "mean_token_accuracy": 0.8978311419487,
+      "num_tokens": 561676560.0,
+      "step": 1494
+    },
+    {
+      "epoch": 2.224702380952381,
+      "grad_norm": 1.6566257954930947,
+      "learning_rate": 6.7928095931053675e-06,
+      "loss": 0.2308,
+      "mean_token_accuracy": 0.9199057817459106,
+      "num_tokens": 562054138.0,
+      "step": 1495
+    },
+    {
+      "epoch": 2.2261904761904763,
+      "grad_norm": 1.7788116681726913,
+      "learning_rate": 6.787959593067431e-06,
+      "loss": 0.3005,
+      "mean_token_accuracy": 0.8963861465454102,
+      "num_tokens": 562432539.0,
+      "step": 1496
+    },
+    {
+      "epoch": 2.2276785714285716,
+      "grad_norm": 1.759618297069151,
+      "learning_rate": 6.783107663311566e-06,
+      "loss": 0.2956,
+      "mean_token_accuracy": 0.8968501687049866,
+      "num_tokens": 562810465.0,
+      "step": 1497
+    },
+    {
+      "epoch": 2.2291666666666665,
+      "grad_norm": 1.7571824189192735,
+      "learning_rate": 6.778253809074384e-06,
+      "loss": 0.3101,
+      "mean_token_accuracy": 0.8924510478973389,
+      "num_tokens": 563189196.0,
+      "step": 1498
+    },
+    {
+      "epoch": 2.230654761904762,
+      "grad_norm": 1.7683953591601806,
+      "learning_rate": 6.7733980355945765e-06,
+      "loss": 0.2747,
+      "mean_token_accuracy": 0.9047020673751831,
+      "num_tokens": 563566720.0,
+      "step": 1499
+    },
+    {
+      "epoch": 2.232142857142857,
+      "grad_norm": 1.7611421800486327,
+      "learning_rate": 6.768540348112908e-06,
+      "loss": 0.3136,
+      "mean_token_accuracy": 0.8910354971885681,
+      "num_tokens": 563946859.0,
+      "step": 1500
+    },
+    {
+      "epoch": 2.2336309523809526,
+      "grad_norm": 1.7233895180941585,
+      "learning_rate": 6.763680751872208e-06,
+      "loss": 0.2927,
+      "mean_token_accuracy": 0.8932012915611267,
+      "num_tokens": 564324675.0,
+      "step": 1501
+    },
+    {
+      "epoch": 2.2351190476190474,
+      "grad_norm": 1.7970188719508788,
+      "learning_rate": 6.758819252117365e-06,
+      "loss": 0.301,
+      "mean_token_accuracy": 0.896121084690094,
+      "num_tokens": 564703828.0,
+      "step": 1502
+    },
+    {
+      "epoch": 2.236607142857143,
+      "grad_norm": 1.7144129611191374,
+      "learning_rate": 6.753955854095323e-06,
+      "loss": 0.3103,
+      "mean_token_accuracy": 0.888518750667572,
+      "num_tokens": 565083335.0,
+      "step": 1503
+    },
+    {
+      "epoch": 2.238095238095238,
+      "grad_norm": 1.869014802102367,
+      "learning_rate": 6.749090563055075e-06,
+      "loss": 0.3031,
+      "mean_token_accuracy": 0.8955521583557129,
+      "num_tokens": 565460681.0,
+      "step": 1504
+    },
+    {
+      "epoch": 2.2395833333333335,
+      "grad_norm": 1.784196342324606,
+      "learning_rate": 6.7442233842476545e-06,
+      "loss": 0.2959,
+      "mean_token_accuracy": 0.8993188142776489,
+      "num_tokens": 565838183.0,
+      "step": 1505
+    },
+    {
+      "epoch": 2.2410714285714284,
+      "grad_norm": 1.7110104825864432,
+      "learning_rate": 6.739354322926136e-06,
+      "loss": 0.2956,
+      "mean_token_accuracy": 0.8990276455879211,
+      "num_tokens": 566217753.0,
+      "step": 1506
+    },
+    {
+      "epoch": 2.2425595238095237,
+      "grad_norm": 1.7152662752875354,
+      "learning_rate": 6.734483384345627e-06,
+      "loss": 0.2994,
+      "mean_token_accuracy": 0.8977074027061462,
+      "num_tokens": 566596072.0,
+      "step": 1507
+    },
+    {
+      "epoch": 2.244047619047619,
+      "grad_norm": 1.7035708959790212,
+      "learning_rate": 6.729610573763252e-06,
+      "loss": 0.2675,
+      "mean_token_accuracy": 0.9046041369438171,
+      "num_tokens": 566973114.0,
+      "step": 1508
+    },
+    {
+      "epoch": 2.2455357142857144,
+      "grad_norm": 1.7740679263615717,
+      "learning_rate": 6.724735896438167e-06,
+      "loss": 0.2686,
+      "mean_token_accuracy": 0.9068275690078735,
+      "num_tokens": 567337903.0,
+      "step": 1509
+    },
+    {
+      "epoch": 2.2470238095238093,
+      "grad_norm": 1.7914140828799108,
+      "learning_rate": 6.719859357631536e-06,
+      "loss": 0.2769,
+      "mean_token_accuracy": 0.9011978507041931,
+      "num_tokens": 567716114.0,
+      "step": 1510
+    },
+    {
+      "epoch": 2.2485119047619047,
+      "grad_norm": 1.997157117527715,
+      "learning_rate": 6.714980962606537e-06,
+      "loss": 0.3163,
+      "mean_token_accuracy": 0.8913900852203369,
+      "num_tokens": 568093273.0,
+      "step": 1511
+    },
+    {
+      "epoch": 2.25,
+      "grad_norm": 1.9139030555847028,
+      "learning_rate": 6.710100716628345e-06,
+      "loss": 0.3268,
+      "mean_token_accuracy": 0.8852980732917786,
+      "num_tokens": 568470042.0,
+      "step": 1512
+    },
+    {
+      "epoch": 2.2514880952380953,
+      "grad_norm": 1.7494389000781214,
+      "learning_rate": 6.705218624964136e-06,
+      "loss": 0.3155,
+      "mean_token_accuracy": 0.8898788690567017,
+      "num_tokens": 568848868.0,
+      "step": 1513
+    },
+    {
+      "epoch": 2.2529761904761907,
+      "grad_norm": 1.8388769843169606,
+      "learning_rate": 6.700334692883081e-06,
+      "loss": 0.2747,
+      "mean_token_accuracy": 0.906151294708252,
+      "num_tokens": 569226654.0,
+      "step": 1514
+    },
+    {
+      "epoch": 2.2544642857142856,
+      "grad_norm": 1.7711961866711443,
+      "learning_rate": 6.6954489256563334e-06,
+      "loss": 0.2872,
+      "mean_token_accuracy": 0.8993120789527893,
+      "num_tokens": 569604670.0,
+      "step": 1515
+    },
+    {
+      "epoch": 2.255952380952381,
+      "grad_norm": 1.797113018011751,
+      "learning_rate": 6.6905613285570295e-06,
+      "loss": 0.2997,
+      "mean_token_accuracy": 0.9010282754898071,
+      "num_tokens": 569983686.0,
+      "step": 1516
+    },
+    {
+      "epoch": 2.2574404761904763,
+      "grad_norm": 1.8139615965300147,
+      "learning_rate": 6.685671906860276e-06,
+      "loss": 0.2818,
+      "mean_token_accuracy": 0.9063178896903992,
+      "num_tokens": 570360985.0,
+      "step": 1517
+    },
+    {
+      "epoch": 2.2589285714285716,
+      "grad_norm": 1.8758022222616186,
+      "learning_rate": 6.680780665843155e-06,
+      "loss": 0.2895,
+      "mean_token_accuracy": 0.8971971273422241,
+      "num_tokens": 570727830.0,
+      "step": 1518
+    },
+    {
+      "epoch": 2.2604166666666665,
+      "grad_norm": 1.7873191523107874,
+      "learning_rate": 6.675887610784708e-06,
+      "loss": 0.3095,
+      "mean_token_accuracy": 0.8958892226219177,
+      "num_tokens": 571106359.0,
+      "step": 1519
+    },
+    {
+      "epoch": 2.261904761904762,
+      "grad_norm": 1.7087171933993983,
+      "learning_rate": 6.6709927469659385e-06,
+      "loss": 0.2956,
+      "mean_token_accuracy": 0.8969104886054993,
+      "num_tokens": 571486126.0,
+      "step": 1520
+    },
+    {
+      "epoch": 2.263392857142857,
+      "grad_norm": 1.7692050830977974,
+      "learning_rate": 6.6660960796697975e-06,
+      "loss": 0.3311,
+      "mean_token_accuracy": 0.8837263584136963,
+      "num_tokens": 571866060.0,
+      "step": 1521
+    },
+    {
+      "epoch": 2.2648809523809526,
+      "grad_norm": 1.7033435670088592,
+      "learning_rate": 6.661197614181183e-06,
+      "loss": 0.3284,
+      "mean_token_accuracy": 0.8859270811080933,
+      "num_tokens": 572245965.0,
+      "step": 1522
+    },
+    {
+      "epoch": 2.2663690476190474,
+      "grad_norm": 1.7377265705679439,
+      "learning_rate": 6.6562973557869384e-06,
+      "loss": 0.3009,
+      "mean_token_accuracy": 0.8941596150398254,
+      "num_tokens": 572612065.0,
+      "step": 1523
+    },
+    {
+      "epoch": 2.267857142857143,
+      "grad_norm": 1.8097377807834827,
+      "learning_rate": 6.651395309775837e-06,
+      "loss": 0.2872,
+      "mean_token_accuracy": 0.9056746363639832,
+      "num_tokens": 572988910.0,
+      "step": 1524
+    },
+    {
+      "epoch": 2.269345238095238,
+      "grad_norm": 1.8182477372414847,
+      "learning_rate": 6.6464914814385815e-06,
+      "loss": 0.2662,
+      "mean_token_accuracy": 0.908289909362793,
+      "num_tokens": 573365430.0,
+      "step": 1525
+    },
+    {
+      "epoch": 2.2708333333333335,
+      "grad_norm": 1.7955777355882923,
+      "learning_rate": 6.641585876067807e-06,
+      "loss": 0.3314,
+      "mean_token_accuracy": 0.8860759735107422,
+      "num_tokens": 573744858.0,
+      "step": 1526
+    },
+    {
+      "epoch": 2.2723214285714284,
+      "grad_norm": 2.0195815336424805,
+      "learning_rate": 6.636678498958053e-06,
+      "loss": 0.3263,
+      "mean_token_accuracy": 0.8852204084396362,
+      "num_tokens": 574123575.0,
+      "step": 1527
+    },
+    {
+      "epoch": 2.2738095238095237,
+      "grad_norm": 1.7839395717689759,
+      "learning_rate": 6.631769355405779e-06,
+      "loss": 0.2755,
+      "mean_token_accuracy": 0.9048349261283875,
+      "num_tokens": 574487773.0,
+      "step": 1528
+    },
+    {
+      "epoch": 2.275297619047619,
+      "grad_norm": 1.8119251779408594,
+      "learning_rate": 6.626858450709351e-06,
+      "loss": 0.2727,
+      "mean_token_accuracy": 0.9076152443885803,
+      "num_tokens": 574852086.0,
+      "step": 1529
+    },
+    {
+      "epoch": 2.2767857142857144,
+      "grad_norm": 1.7669373744880665,
+      "learning_rate": 6.621945790169037e-06,
+      "loss": 0.2964,
+      "mean_token_accuracy": 0.891386091709137,
+      "num_tokens": 575225847.0,
+      "step": 1530
+    },
+    {
+      "epoch": 2.2782738095238093,
+      "grad_norm": 1.8283753908470948,
+      "learning_rate": 6.617031379086994e-06,
+      "loss": 0.2862,
+      "mean_token_accuracy": 0.8992363810539246,
+      "num_tokens": 575603210.0,
+      "step": 1531
+    },
+    {
+      "epoch": 2.2797619047619047,
+      "grad_norm": 1.8037720147740022,
+      "learning_rate": 6.612115222767272e-06,
+      "loss": 0.3091,
+      "mean_token_accuracy": 0.8952134251594543,
+      "num_tokens": 575982891.0,
+      "step": 1532
+    },
+    {
+      "epoch": 2.28125,
+      "grad_norm": 1.9349904179726574,
+      "learning_rate": 6.607197326515808e-06,
+      "loss": 0.2807,
+      "mean_token_accuracy": 0.8958517909049988,
+      "num_tokens": 576359541.0,
+      "step": 1533
+    },
+    {
+      "epoch": 2.2827380952380953,
+      "grad_norm": 1.8453488389556356,
+      "learning_rate": 6.602277695640412e-06,
+      "loss": 0.3052,
+      "mean_token_accuracy": 0.8983398675918579,
+      "num_tokens": 576737876.0,
+      "step": 1534
+    },
+    {
+      "epoch": 2.2842261904761907,
+      "grad_norm": 1.7171416699323319,
+      "learning_rate": 6.597356335450767e-06,
+      "loss": 0.2902,
+      "mean_token_accuracy": 0.8982994556427002,
+      "num_tokens": 577116733.0,
+      "step": 1535
+    },
+    {
+      "epoch": 2.2857142857142856,
+      "grad_norm": 1.8273832849029823,
+      "learning_rate": 6.592433251258423e-06,
+      "loss": 0.308,
+      "mean_token_accuracy": 0.887218713760376,
+      "num_tokens": 577494584.0,
+      "step": 1536
+    },
+    {
+      "epoch": 2.287202380952381,
+      "grad_norm": 1.8936384071405636,
+      "learning_rate": 6.587508448376792e-06,
+      "loss": 0.3069,
+      "mean_token_accuracy": 0.8948836326599121,
+      "num_tokens": 577871389.0,
+      "step": 1537
+    },
+    {
+      "epoch": 2.2886904761904763,
+      "grad_norm": 1.8826211639523127,
+      "learning_rate": 6.582581932121138e-06,
+      "loss": 0.3069,
+      "mean_token_accuracy": 0.8968490958213806,
+      "num_tokens": 578237969.0,
+      "step": 1538
+    },
+    {
+      "epoch": 2.2901785714285716,
+      "grad_norm": 1.8564603095325822,
+      "learning_rate": 6.577653707808578e-06,
+      "loss": 0.3125,
+      "mean_token_accuracy": 0.8917756676673889,
+      "num_tokens": 578616694.0,
+      "step": 1539
+    },
+    {
+      "epoch": 2.2916666666666665,
+      "grad_norm": 1.7966977162510027,
+      "learning_rate": 6.572723780758069e-06,
+      "loss": 0.2978,
+      "mean_token_accuracy": 0.8960292339324951,
+      "num_tokens": 578990679.0,
+      "step": 1540
+    },
+    {
+      "epoch": 2.293154761904762,
+      "grad_norm": 1.7580532832967997,
+      "learning_rate": 6.567792156290409e-06,
+      "loss": 0.2647,
+      "mean_token_accuracy": 0.9062381386756897,
+      "num_tokens": 579367415.0,
+      "step": 1541
+    },
+    {
+      "epoch": 2.294642857142857,
+      "grad_norm": 1.9753202680702056,
+      "learning_rate": 6.562858839728224e-06,
+      "loss": 0.313,
+      "mean_token_accuracy": 0.8911552429199219,
+      "num_tokens": 579745672.0,
+      "step": 1542
+    },
+    {
+      "epoch": 2.2961309523809526,
+      "grad_norm": 1.7296354813697772,
+      "learning_rate": 6.557923836395969e-06,
+      "loss": 0.2612,
+      "mean_token_accuracy": 0.9104681015014648,
+      "num_tokens": 580124076.0,
+      "step": 1543
+    },
+    {
+      "epoch": 2.2976190476190474,
+      "grad_norm": 1.8775093036674408,
+      "learning_rate": 6.552987151619919e-06,
+      "loss": 0.3079,
+      "mean_token_accuracy": 0.894731879234314,
+      "num_tokens": 580483784.0,
+      "step": 1544
+    },
+    {
+      "epoch": 2.299107142857143,
+      "grad_norm": 1.7715035109528399,
+      "learning_rate": 6.548048790728165e-06,
+      "loss": 0.2944,
+      "mean_token_accuracy": 0.8995496034622192,
+      "num_tokens": 580862755.0,
+      "step": 1545
+    },
+    {
+      "epoch": 2.300595238095238,
+      "grad_norm": 1.75861657046792,
+      "learning_rate": 6.543108759050606e-06,
+      "loss": 0.3355,
+      "mean_token_accuracy": 0.8859857320785522,
+      "num_tokens": 581242512.0,
+      "step": 1546
+    },
+    {
+      "epoch": 2.3020833333333335,
+      "grad_norm": 1.7029673325368408,
+      "learning_rate": 6.538167061918942e-06,
+      "loss": 0.2997,
+      "mean_token_accuracy": 0.8961706757545471,
+      "num_tokens": 581605678.0,
+      "step": 1547
+    },
+    {
+      "epoch": 2.3035714285714284,
+      "grad_norm": 1.7456335914518157,
+      "learning_rate": 6.5332237046666725e-06,
+      "loss": 0.2953,
+      "mean_token_accuracy": 0.8963414430618286,
+      "num_tokens": 581982300.0,
+      "step": 1548
+    },
+    {
+      "epoch": 2.3050595238095237,
+      "grad_norm": 1.763488926243676,
+      "learning_rate": 6.5282786926290906e-06,
+      "loss": 0.2861,
+      "mean_token_accuracy": 0.9014447927474976,
+      "num_tokens": 582360714.0,
+      "step": 1549
+    },
+    {
+      "epoch": 2.306547619047619,
+      "grad_norm": 1.8051584203826343,
+      "learning_rate": 6.523332031143273e-06,
+      "loss": 0.3107,
+      "mean_token_accuracy": 0.8944741487503052,
+      "num_tokens": 582738870.0,
+      "step": 1550
+    },
+    {
+      "epoch": 2.3080357142857144,
+      "grad_norm": 1.8325081651594282,
+      "learning_rate": 6.518383725548074e-06,
+      "loss": 0.2958,
+      "mean_token_accuracy": 0.8980638384819031,
+      "num_tokens": 583117339.0,
+      "step": 1551
+    },
+    {
+      "epoch": 2.3095238095238093,
+      "grad_norm": 1.6666109616351492,
+      "learning_rate": 6.513433781184131e-06,
+      "loss": 0.3054,
+      "mean_token_accuracy": 0.893904447555542,
+      "num_tokens": 583496246.0,
+      "step": 1552
+    },
+    {
+      "epoch": 2.3110119047619047,
+      "grad_norm": 2.127050074190029,
+      "learning_rate": 6.50848220339384e-06,
+      "loss": 0.3019,
+      "mean_token_accuracy": 0.8980281949043274,
+      "num_tokens": 583873227.0,
+      "step": 1553
+    },
+    {
+      "epoch": 2.3125,
+      "grad_norm": 1.7362459851214214,
+      "learning_rate": 6.503528997521365e-06,
+      "loss": 0.3244,
+      "mean_token_accuracy": 0.8838769197463989,
+      "num_tokens": 584251435.0,
+      "step": 1554
+    },
+    {
+      "epoch": 2.3139880952380953,
+      "grad_norm": 1.753146009914963,
+      "learning_rate": 6.49857416891263e-06,
+      "loss": 0.2793,
+      "mean_token_accuracy": 0.9028576612472534,
+      "num_tokens": 584629449.0,
+      "step": 1555
+    },
+    {
+      "epoch": 2.3154761904761907,
+      "grad_norm": 1.6640042585553505,
+      "learning_rate": 6.493617722915306e-06,
+      "loss": 0.2829,
+      "mean_token_accuracy": 0.9003638625144958,
+      "num_tokens": 585005819.0,
+      "step": 1556
+    },
+    {
+      "epoch": 2.3169642857142856,
+      "grad_norm": 1.6836613978579864,
+      "learning_rate": 6.488659664878808e-06,
+      "loss": 0.2838,
+      "mean_token_accuracy": 0.9002774953842163,
+      "num_tokens": 585384080.0,
+      "step": 1557
+    },
+    {
+      "epoch": 2.318452380952381,
+      "grad_norm": 1.74573248997913,
+      "learning_rate": 6.483700000154295e-06,
+      "loss": 0.3,
+      "mean_token_accuracy": 0.8977903127670288,
+      "num_tokens": 585762643.0,
+      "step": 1558
+    },
+    {
+      "epoch": 2.3199404761904763,
+      "grad_norm": 1.7885977554415076,
+      "learning_rate": 6.478738734094661e-06,
+      "loss": 0.2664,
+      "mean_token_accuracy": 0.9062384366989136,
+      "num_tokens": 586140274.0,
+      "step": 1559
+    },
+    {
+      "epoch": 2.3214285714285716,
+      "grad_norm": 1.7278860768218893,
+      "learning_rate": 6.473775872054522e-06,
+      "loss": 0.2809,
+      "mean_token_accuracy": 0.9033494591712952,
+      "num_tokens": 586515081.0,
+      "step": 1560
+    },
+    {
+      "epoch": 2.3229166666666665,
+      "grad_norm": 1.7369239269803265,
+      "learning_rate": 6.468811419390222e-06,
+      "loss": 0.2629,
+      "mean_token_accuracy": 0.9083880186080933,
+      "num_tokens": 586884856.0,
+      "step": 1561
+    },
+    {
+      "epoch": 2.324404761904762,
+      "grad_norm": 1.8613110746754669,
+      "learning_rate": 6.463845381459822e-06,
+      "loss": 0.3036,
+      "mean_token_accuracy": 0.8938334584236145,
+      "num_tokens": 587248433.0,
+      "step": 1562
+    },
+    {
+      "epoch": 2.325892857142857,
+      "grad_norm": 1.856961773414769,
+      "learning_rate": 6.458877763623089e-06,
+      "loss": 0.2938,
+      "mean_token_accuracy": 0.8982795476913452,
+      "num_tokens": 587612983.0,
+      "step": 1563
+    },
+    {
+      "epoch": 2.3273809523809526,
+      "grad_norm": 1.6807985731981905,
+      "learning_rate": 6.453908571241501e-06,
+      "loss": 0.3277,
+      "mean_token_accuracy": 0.8860095739364624,
+      "num_tokens": 587994305.0,
+      "step": 1564
+    },
+    {
+      "epoch": 2.3288690476190474,
+      "grad_norm": 1.8599755871497323,
+      "learning_rate": 6.448937809678233e-06,
+      "loss": 0.2988,
+      "mean_token_accuracy": 0.9037248492240906,
+      "num_tokens": 588373558.0,
+      "step": 1565
+    },
+    {
+      "epoch": 2.330357142857143,
+      "grad_norm": 1.766584831245552,
+      "learning_rate": 6.44396548429815e-06,
+      "loss": 0.2992,
+      "mean_token_accuracy": 0.895976722240448,
+      "num_tokens": 588751671.0,
+      "step": 1566
+    },
+    {
+      "epoch": 2.331845238095238,
+      "grad_norm": 1.728773443168273,
+      "learning_rate": 6.43899160046781e-06,
+      "loss": 0.3202,
+      "mean_token_accuracy": 0.8930633664131165,
+      "num_tokens": 589131388.0,
+      "step": 1567
+    },
+    {
+      "epoch": 2.3333333333333335,
+      "grad_norm": 1.7812809763050113,
+      "learning_rate": 6.434016163555452e-06,
+      "loss": 0.2974,
+      "mean_token_accuracy": 0.8947981595993042,
+      "num_tokens": 589510695.0,
+      "step": 1568
+    },
+    {
+      "epoch": 2.3348214285714284,
+      "grad_norm": 1.604925174350134,
+      "learning_rate": 6.429039178930989e-06,
+      "loss": 0.2927,
+      "mean_token_accuracy": 0.8999224901199341,
+      "num_tokens": 589890897.0,
+      "step": 1569
+    },
+    {
+      "epoch": 2.3363095238095237,
+      "grad_norm": 1.8302606228663767,
+      "learning_rate": 6.424060651966006e-06,
+      "loss": 0.28,
+      "mean_token_accuracy": 0.901836633682251,
+      "num_tokens": 590267366.0,
+      "step": 1570
+    },
+    {
+      "epoch": 2.337797619047619,
+      "grad_norm": 1.7470023186511943,
+      "learning_rate": 6.419080588033754e-06,
+      "loss": 0.2725,
+      "mean_token_accuracy": 0.8985618352890015,
+      "num_tokens": 590644178.0,
+      "step": 1571
+    },
+    {
+      "epoch": 2.3392857142857144,
+      "grad_norm": 1.9482411481688109,
+      "learning_rate": 6.414098992509138e-06,
+      "loss": 0.2954,
+      "mean_token_accuracy": 0.8997372984886169,
+      "num_tokens": 591022432.0,
+      "step": 1572
+    },
+    {
+      "epoch": 2.3407738095238093,
+      "grad_norm": 1.7726325217213295,
+      "learning_rate": 6.409115870768721e-06,
+      "loss": 0.2715,
+      "mean_token_accuracy": 0.9071167707443237,
+      "num_tokens": 591400147.0,
+      "step": 1573
+    },
+    {
+      "epoch": 2.3422619047619047,
+      "grad_norm": 1.880567707726188,
+      "learning_rate": 6.404131228190713e-06,
+      "loss": 0.2727,
+      "mean_token_accuracy": 0.9024321436882019,
+      "num_tokens": 591764628.0,
+      "step": 1574
+    },
+    {
+      "epoch": 2.34375,
+      "grad_norm": 1.8714057348657038,
+      "learning_rate": 6.399145070154962e-06,
+      "loss": 0.2968,
+      "mean_token_accuracy": 0.8947145938873291,
+      "num_tokens": 592143195.0,
+      "step": 1575
+    },
+    {
+      "epoch": 2.3452380952380953,
+      "grad_norm": 1.8320138529400027,
+      "learning_rate": 6.394157402042952e-06,
+      "loss": 0.2912,
+      "mean_token_accuracy": 0.902353048324585,
+      "num_tokens": 592521784.0,
+      "step": 1576
+    },
+    {
+      "epoch": 2.3467261904761907,
+      "grad_norm": 1.8761199110766917,
+      "learning_rate": 6.389168229237801e-06,
+      "loss": 0.3101,
+      "mean_token_accuracy": 0.8952310085296631,
+      "num_tokens": 592890421.0,
+      "step": 1577
+    },
+    {
+      "epoch": 2.3482142857142856,
+      "grad_norm": 1.9001858723913239,
+      "learning_rate": 6.384177557124247e-06,
+      "loss": 0.333,
+      "mean_token_accuracy": 0.8868517279624939,
+      "num_tokens": 593268745.0,
+      "step": 1578
+    },
+    {
+      "epoch": 2.349702380952381,
+      "grad_norm": 1.8190218277577428,
+      "learning_rate": 6.379185391088648e-06,
+      "loss": 0.274,
+      "mean_token_accuracy": 0.9038100838661194,
+      "num_tokens": 593646033.0,
+      "step": 1579
+    },
+    {
+      "epoch": 2.3511904761904763,
+      "grad_norm": 1.8176111743046248,
+      "learning_rate": 6.374191736518974e-06,
+      "loss": 0.2635,
+      "mean_token_accuracy": 0.9073020815849304,
+      "num_tokens": 594023741.0,
+      "step": 1580
+    },
+    {
+      "epoch": 2.3526785714285716,
+      "grad_norm": 1.9397160525021886,
+      "learning_rate": 6.369196598804801e-06,
+      "loss": 0.3276,
+      "mean_token_accuracy": 0.8889627456665039,
+      "num_tokens": 594402155.0,
+      "step": 1581
+    },
+    {
+      "epoch": 2.3541666666666665,
+      "grad_norm": 1.9372146032878763,
+      "learning_rate": 6.364199983337306e-06,
+      "loss": 0.3016,
+      "mean_token_accuracy": 0.8955693244934082,
+      "num_tokens": 594780002.0,
+      "step": 1582
+    },
+    {
+      "epoch": 2.355654761904762,
+      "grad_norm": 2.0355954031973242,
+      "learning_rate": 6.35920189550926e-06,
+      "loss": 0.345,
+      "mean_token_accuracy": 0.8827377557754517,
+      "num_tokens": 595158470.0,
+      "step": 1583
+    },
+    {
+      "epoch": 2.357142857142857,
+      "grad_norm": 1.9223013513295306,
+      "learning_rate": 6.354202340715027e-06,
+      "loss": 0.2898,
+      "mean_token_accuracy": 0.8978800177574158,
+      "num_tokens": 595535804.0,
+      "step": 1584
+    },
+    {
+      "epoch": 2.3586309523809526,
+      "grad_norm": 1.8113249309940869,
+      "learning_rate": 6.349201324350548e-06,
+      "loss": 0.3198,
+      "mean_token_accuracy": 0.8916695713996887,
+      "num_tokens": 595914344.0,
+      "step": 1585
+    },
+    {
+      "epoch": 2.3601190476190474,
+      "grad_norm": 1.9034556619579521,
+      "learning_rate": 6.344198851813346e-06,
+      "loss": 0.3037,
+      "mean_token_accuracy": 0.8895255327224731,
+      "num_tokens": 596292169.0,
+      "step": 1586
+    },
+    {
+      "epoch": 2.361607142857143,
+      "grad_norm": 1.9204823727177855,
+      "learning_rate": 6.339194928502516e-06,
+      "loss": 0.2693,
+      "mean_token_accuracy": 0.9051165580749512,
+      "num_tokens": 596668286.0,
+      "step": 1587
+    },
+    {
+      "epoch": 2.363095238095238,
+      "grad_norm": 1.8614862327069035,
+      "learning_rate": 6.334189559818715e-06,
+      "loss": 0.2644,
+      "mean_token_accuracy": 0.910612940788269,
+      "num_tokens": 597019612.0,
+      "step": 1588
+    },
+    {
+      "epoch": 2.3645833333333335,
+      "grad_norm": 1.9393433130161444,
+      "learning_rate": 6.329182751164164e-06,
+      "loss": 0.2804,
+      "mean_token_accuracy": 0.9094610214233398,
+      "num_tokens": 597396466.0,
+      "step": 1589
+    },
+    {
+      "epoch": 2.3660714285714284,
+      "grad_norm": 1.832831288992839,
+      "learning_rate": 6.324174507942636e-06,
+      "loss": 0.279,
+      "mean_token_accuracy": 0.8998178243637085,
+      "num_tokens": 597774061.0,
+      "step": 1590
+    },
+    {
+      "epoch": 2.3675595238095237,
+      "grad_norm": 1.859783455127377,
+      "learning_rate": 6.3191648355594526e-06,
+      "loss": 0.2963,
+      "mean_token_accuracy": 0.9005036950111389,
+      "num_tokens": 598140971.0,
+      "step": 1591
+    },
+    {
+      "epoch": 2.369047619047619,
+      "grad_norm": 1.7811185380144747,
+      "learning_rate": 6.314153739421477e-06,
+      "loss": 0.2793,
+      "mean_token_accuracy": 0.9009984731674194,
+      "num_tokens": 598511800.0,
+      "step": 1592
+    },
+    {
+      "epoch": 2.3705357142857144,
+      "grad_norm": 1.9460664272337063,
+      "learning_rate": 6.3091412249371106e-06,
+      "loss": 0.3639,
+      "mean_token_accuracy": 0.8783703446388245,
+      "num_tokens": 598890781.0,
+      "step": 1593
+    },
+    {
+      "epoch": 2.3720238095238093,
+      "grad_norm": 1.7392702965421476,
+      "learning_rate": 6.304127297516287e-06,
+      "loss": 0.2989,
+      "mean_token_accuracy": 0.897097647190094,
+      "num_tokens": 599266555.0,
+      "step": 1594
+    },
+    {
+      "epoch": 2.3735119047619047,
+      "grad_norm": 1.754443403679712,
+      "learning_rate": 6.299111962570459e-06,
+      "loss": 0.3033,
+      "mean_token_accuracy": 0.8908308148384094,
+      "num_tokens": 599645576.0,
+      "step": 1595
+    },
+    {
+      "epoch": 2.375,
+      "grad_norm": 1.6139828408304366,
+      "learning_rate": 6.294095225512604e-06,
+      "loss": 0.2584,
+      "mean_token_accuracy": 0.9084944128990173,
+      "num_tokens": 600023542.0,
+      "step": 1596
+    },
+    {
+      "epoch": 2.3764880952380953,
+      "grad_norm": 1.8575155455893417,
+      "learning_rate": 6.2890770917572115e-06,
+      "loss": 0.2999,
+      "mean_token_accuracy": 0.8954662084579468,
+      "num_tokens": 600401490.0,
+      "step": 1597
+    },
+    {
+      "epoch": 2.3779761904761907,
+      "grad_norm": 1.722347017166096,
+      "learning_rate": 6.284057566720279e-06,
+      "loss": 0.261,
+      "mean_token_accuracy": 0.9068851470947266,
+      "num_tokens": 600780063.0,
+      "step": 1598
+    },
+    {
+      "epoch": 2.3794642857142856,
+      "grad_norm": 1.8887401860570152,
+      "learning_rate": 6.279036655819302e-06,
+      "loss": 0.3152,
+      "mean_token_accuracy": 0.8917312026023865,
+      "num_tokens": 601157827.0,
+      "step": 1599
+    },
+    {
+      "epoch": 2.380952380952381,
+      "grad_norm": 1.8047584918977808,
+      "learning_rate": 6.274014364473274e-06,
+      "loss": 0.2996,
+      "mean_token_accuracy": 0.9031727313995361,
+      "num_tokens": 601533665.0,
+      "step": 1600
+    },
+    {
+      "epoch": 2.3824404761904763,
+      "grad_norm": 1.8537419349038335,
+      "learning_rate": 6.26899069810268e-06,
+      "loss": 0.2909,
+      "mean_token_accuracy": 0.9040141105651855,
+      "num_tokens": 601898441.0,
+      "step": 1601
+    },
+    {
+      "epoch": 2.3839285714285716,
+      "grad_norm": 1.708105149534495,
+      "learning_rate": 6.2639656621294874e-06,
+      "loss": 0.2983,
+      "mean_token_accuracy": 0.8968329429626465,
+      "num_tokens": 602276644.0,
+      "step": 1602
+    },
+    {
+      "epoch": 2.3854166666666665,
+      "grad_norm": 1.7924456337530659,
+      "learning_rate": 6.2589392619771435e-06,
+      "loss": 0.3204,
+      "mean_token_accuracy": 0.8883713483810425,
+      "num_tokens": 602655185.0,
+      "step": 1603
+    },
+    {
+      "epoch": 2.386904761904762,
+      "grad_norm": 1.8559866295661627,
+      "learning_rate": 6.253911503070564e-06,
+      "loss": 0.3197,
+      "mean_token_accuracy": 0.8875786066055298,
+      "num_tokens": 603032414.0,
+      "step": 1604
+    },
+    {
+      "epoch": 2.388392857142857,
+      "grad_norm": 1.786915394821799,
+      "learning_rate": 6.248882390836135e-06,
+      "loss": 0.289,
+      "mean_token_accuracy": 0.9014261960983276,
+      "num_tokens": 603410187.0,
+      "step": 1605
+    },
+    {
+      "epoch": 2.3898809523809526,
+      "grad_norm": 1.8121930913843767,
+      "learning_rate": 6.243851930701701e-06,
+      "loss": 0.2794,
+      "mean_token_accuracy": 0.9050611257553101,
+      "num_tokens": 603787148.0,
+      "step": 1606
+    },
+    {
+      "epoch": 2.3913690476190474,
+      "grad_norm": 1.892701981994749,
+      "learning_rate": 6.238820128096563e-06,
+      "loss": 0.3398,
+      "mean_token_accuracy": 0.8843955397605896,
+      "num_tokens": 604165258.0,
+      "step": 1607
+    },
+    {
+      "epoch": 2.392857142857143,
+      "grad_norm": 1.6717953422426355,
+      "learning_rate": 6.233786988451468e-06,
+      "loss": 0.2896,
+      "mean_token_accuracy": 0.9012871980667114,
+      "num_tokens": 604545444.0,
+      "step": 1608
+    },
+    {
+      "epoch": 2.394345238095238,
+      "grad_norm": 1.8403193438729084,
+      "learning_rate": 6.2287525171986114e-06,
+      "loss": 0.3062,
+      "mean_token_accuracy": 0.8960589170455933,
+      "num_tokens": 604923867.0,
+      "step": 1609
+    },
+    {
+      "epoch": 2.3958333333333335,
+      "grad_norm": 1.839686647470915,
+      "learning_rate": 6.2237167197716195e-06,
+      "loss": 0.3076,
+      "mean_token_accuracy": 0.8900142312049866,
+      "num_tokens": 605302111.0,
+      "step": 1610
+    },
+    {
+      "epoch": 2.3973214285714284,
+      "grad_norm": 1.8164474511520645,
+      "learning_rate": 6.218679601605554e-06,
+      "loss": 0.3107,
+      "mean_token_accuracy": 0.8900123238563538,
+      "num_tokens": 605679821.0,
+      "step": 1611
+    },
+    {
+      "epoch": 2.3988095238095237,
+      "grad_norm": 1.757010703012436,
+      "learning_rate": 6.2136411681369e-06,
+      "loss": 0.3119,
+      "mean_token_accuracy": 0.8908622860908508,
+      "num_tokens": 606059104.0,
+      "step": 1612
+    },
+    {
+      "epoch": 2.400297619047619,
+      "grad_norm": 1.8390616514073814,
+      "learning_rate": 6.208601424803564e-06,
+      "loss": 0.2927,
+      "mean_token_accuracy": 0.8973137140274048,
+      "num_tokens": 606436634.0,
+      "step": 1613
+    },
+    {
+      "epoch": 2.4017857142857144,
+      "grad_norm": 1.812580330266152,
+      "learning_rate": 6.2035603770448664e-06,
+      "loss": 0.3508,
+      "mean_token_accuracy": 0.8803563117980957,
+      "num_tokens": 606816543.0,
+      "step": 1614
+    },
+    {
+      "epoch": 2.4032738095238093,
+      "grad_norm": 1.8035493856598819,
+      "learning_rate": 6.198518030301531e-06,
+      "loss": 0.3249,
+      "mean_token_accuracy": 0.8889856934547424,
+      "num_tokens": 607195495.0,
+      "step": 1615
+    },
+    {
+      "epoch": 2.4047619047619047,
+      "grad_norm": 1.8461513850506341,
+      "learning_rate": 6.19347439001569e-06,
+      "loss": 0.2828,
+      "mean_token_accuracy": 0.9046324491500854,
+      "num_tokens": 607563949.0,
+      "step": 1616
+    },
+    {
+      "epoch": 2.40625,
+      "grad_norm": 1.7781327482919775,
+      "learning_rate": 6.188429461630866e-06,
+      "loss": 0.3095,
+      "mean_token_accuracy": 0.8962130546569824,
+      "num_tokens": 607941849.0,
+      "step": 1617
+    },
+    {
+      "epoch": 2.4077380952380953,
+      "grad_norm": 1.806760515521878,
+      "learning_rate": 6.183383250591977e-06,
+      "loss": 0.3089,
+      "mean_token_accuracy": 0.8900715708732605,
+      "num_tokens": 608320025.0,
+      "step": 1618
+    },
+    {
+      "epoch": 2.4092261904761907,
+      "grad_norm": 1.7392479933250578,
+      "learning_rate": 6.178335762345322e-06,
+      "loss": 0.292,
+      "mean_token_accuracy": 0.8966689109802246,
+      "num_tokens": 608699163.0,
+      "step": 1619
+    },
+    {
+      "epoch": 2.4107142857142856,
+      "grad_norm": 1.6697115107522582,
+      "learning_rate": 6.173287002338577e-06,
+      "loss": 0.2891,
+      "mean_token_accuracy": 0.8975928425788879,
+      "num_tokens": 609075412.0,
+      "step": 1620
+    },
+    {
+      "epoch": 2.412202380952381,
+      "grad_norm": 1.8025904074766275,
+      "learning_rate": 6.168236976020796e-06,
+      "loss": 0.3444,
+      "mean_token_accuracy": 0.8819101452827454,
+      "num_tokens": 609454983.0,
+      "step": 1621
+    },
+    {
+      "epoch": 2.4136904761904763,
+      "grad_norm": 1.7204188327402221,
+      "learning_rate": 6.163185688842393e-06,
+      "loss": 0.3231,
+      "mean_token_accuracy": 0.8803812265396118,
+      "num_tokens": 609809379.0,
+      "step": 1622
+    },
+    {
+      "epoch": 2.4151785714285716,
+      "grad_norm": 1.9290451490514242,
+      "learning_rate": 6.158133146255153e-06,
+      "loss": 0.3463,
+      "mean_token_accuracy": 0.884025514125824,
+      "num_tokens": 610187850.0,
+      "step": 1623
+    },
+    {
+      "epoch": 2.4166666666666665,
+      "grad_norm": 1.8245562235693273,
+      "learning_rate": 6.153079353712201e-06,
+      "loss": 0.2827,
+      "mean_token_accuracy": 0.9014006853103638,
+      "num_tokens": 610564988.0,
+      "step": 1624
+    },
+    {
+      "epoch": 2.418154761904762,
+      "grad_norm": 1.7005109334988395,
+      "learning_rate": 6.148024316668025e-06,
+      "loss": 0.3528,
+      "mean_token_accuracy": 0.8788145780563354,
+      "num_tokens": 610936524.0,
+      "step": 1625
+    },
+    {
+      "epoch": 2.419642857142857,
+      "grad_norm": 1.8233810761122269,
+      "learning_rate": 6.1429680405784485e-06,
+      "loss": 0.2863,
+      "mean_token_accuracy": 0.896590530872345,
+      "num_tokens": 611315309.0,
+      "step": 1626
+    },
+    {
+      "epoch": 2.4211309523809526,
+      "grad_norm": 1.9553443812685494,
+      "learning_rate": 6.137910530900635e-06,
+      "loss": 0.2767,
+      "mean_token_accuracy": 0.9048783183097839,
+      "num_tokens": 611676884.0,
+      "step": 1627
+    },
+    {
+      "epoch": 2.4226190476190474,
+      "grad_norm": 1.7492519284786385,
+      "learning_rate": 6.132851793093079e-06,
+      "loss": 0.2877,
+      "mean_token_accuracy": 0.8952134251594543,
+      "num_tokens": 612054745.0,
+      "step": 1628
+    },
+    {
+      "epoch": 2.424107142857143,
+      "grad_norm": 1.7981926168270301,
+      "learning_rate": 6.127791832615598e-06,
+      "loss": 0.2765,
+      "mean_token_accuracy": 0.9013665318489075,
+      "num_tokens": 612432390.0,
+      "step": 1629
+    },
+    {
+      "epoch": 2.425595238095238,
+      "grad_norm": 1.7521220342838968,
+      "learning_rate": 6.122730654929334e-06,
+      "loss": 0.2779,
+      "mean_token_accuracy": 0.9046804904937744,
+      "num_tokens": 612809994.0,
+      "step": 1630
+    },
+    {
+      "epoch": 2.4270833333333335,
+      "grad_norm": 1.890118842978789,
+      "learning_rate": 6.117668265496738e-06,
+      "loss": 0.2942,
+      "mean_token_accuracy": 0.9008746147155762,
+      "num_tokens": 613188453.0,
+      "step": 1631
+    },
+    {
+      "epoch": 2.4285714285714284,
+      "grad_norm": 1.7981461158655991,
+      "learning_rate": 6.112604669781572e-06,
+      "loss": 0.2939,
+      "mean_token_accuracy": 0.9004108905792236,
+      "num_tokens": 613566443.0,
+      "step": 1632
+    },
+    {
+      "epoch": 2.4300595238095237,
+      "grad_norm": 1.7938011458129568,
+      "learning_rate": 6.1075398732489e-06,
+      "loss": 0.2858,
+      "mean_token_accuracy": 0.8954578042030334,
+      "num_tokens": 613944455.0,
+      "step": 1633
+    },
+    {
+      "epoch": 2.431547619047619,
+      "grad_norm": 1.8483207874161494,
+      "learning_rate": 6.102473881365079e-06,
+      "loss": 0.2767,
+      "mean_token_accuracy": 0.9056122303009033,
+      "num_tokens": 614321900.0,
+      "step": 1634
+    },
+    {
+      "epoch": 2.4330357142857144,
+      "grad_norm": 1.7371569508489153,
+      "learning_rate": 6.097406699597759e-06,
+      "loss": 0.3023,
+      "mean_token_accuracy": 0.8945847749710083,
+      "num_tokens": 614699789.0,
+      "step": 1635
+    },
+    {
+      "epoch": 2.4345238095238093,
+      "grad_norm": 1.905601036815521,
+      "learning_rate": 6.0923383334158704e-06,
+      "loss": 0.3217,
+      "mean_token_accuracy": 0.8857463002204895,
+      "num_tokens": 615078405.0,
+      "step": 1636
+    },
+    {
+      "epoch": 2.4360119047619047,
+      "grad_norm": 1.8070764475624528,
+      "learning_rate": 6.087268788289628e-06,
+      "loss": 0.3293,
+      "mean_token_accuracy": 0.8864133358001709,
+      "num_tokens": 615457843.0,
+      "step": 1637
+    },
+    {
+      "epoch": 2.4375,
+      "grad_norm": 1.8001579218249353,
+      "learning_rate": 6.0821980696905145e-06,
+      "loss": 0.2665,
+      "mean_token_accuracy": 0.9038892388343811,
+      "num_tokens": 615834699.0,
+      "step": 1638
+    },
+    {
+      "epoch": 2.4389880952380953,
+      "grad_norm": 1.769310139865939,
+      "learning_rate": 6.077126183091282e-06,
+      "loss": 0.3017,
+      "mean_token_accuracy": 0.8954843282699585,
+      "num_tokens": 616213805.0,
+      "step": 1639
+    },
+    {
+      "epoch": 2.4404761904761907,
+      "grad_norm": 1.7952540509520438,
+      "learning_rate": 6.0720531339659386e-06,
+      "loss": 0.2786,
+      "mean_token_accuracy": 0.8983606696128845,
+      "num_tokens": 616578885.0,
+      "step": 1640
+    },
+    {
+      "epoch": 2.4419642857142856,
+      "grad_norm": 1.7236166333733645,
+      "learning_rate": 6.066978927789751e-06,
+      "loss": 0.2429,
+      "mean_token_accuracy": 0.9129654765129089,
+      "num_tokens": 616957120.0,
+      "step": 1641
+    },
+    {
+      "epoch": 2.443452380952381,
+      "grad_norm": 1.866279945291814,
+      "learning_rate": 6.061903570039234e-06,
+      "loss": 0.3072,
+      "mean_token_accuracy": 0.8862151503562927,
+      "num_tokens": 617324134.0,
+      "step": 1642
+    },
+    {
+      "epoch": 2.4449404761904763,
+      "grad_norm": 1.9995581344878435,
+      "learning_rate": 6.056827066192146e-06,
+      "loss": 0.2525,
+      "mean_token_accuracy": 0.9112522602081299,
+      "num_tokens": 617701114.0,
+      "step": 1643
+    },
+    {
+      "epoch": 2.4464285714285716,
+      "grad_norm": 1.7242035515928258,
+      "learning_rate": 6.05174942172748e-06,
+      "loss": 0.3025,
+      "mean_token_accuracy": 0.8951914310455322,
+      "num_tokens": 618079866.0,
+      "step": 1644
+    },
+    {
+      "epoch": 2.4479166666666665,
+      "grad_norm": 1.7669359215459983,
+      "learning_rate": 6.046670642125461e-06,
+      "loss": 0.2946,
+      "mean_token_accuracy": 0.8982452750205994,
+      "num_tokens": 618447950.0,
+      "step": 1645
+    },
+    {
+      "epoch": 2.449404761904762,
+      "grad_norm": 1.6601032140599334,
+      "learning_rate": 6.04159073286754e-06,
+      "loss": 0.2759,
+      "mean_token_accuracy": 0.9049105644226074,
+      "num_tokens": 618826553.0,
+      "step": 1646
+    },
+    {
+      "epoch": 2.450892857142857,
+      "grad_norm": 1.8439145939566377,
+      "learning_rate": 6.036509699436389e-06,
+      "loss": 0.2988,
+      "mean_token_accuracy": 0.8962531089782715,
+      "num_tokens": 619189130.0,
+      "step": 1647
+    },
+    {
+      "epoch": 2.4523809523809526,
+      "grad_norm": 1.8096849127758667,
+      "learning_rate": 6.031427547315889e-06,
+      "loss": 0.3053,
+      "mean_token_accuracy": 0.8921962380409241,
+      "num_tokens": 619567734.0,
+      "step": 1648
+    },
+    {
+      "epoch": 2.4538690476190474,
+      "grad_norm": 1.882522580159858,
+      "learning_rate": 6.026344281991129e-06,
+      "loss": 0.3075,
+      "mean_token_accuracy": 0.8887346982955933,
+      "num_tokens": 619946604.0,
+      "step": 1649
+    },
+    {
+      "epoch": 2.455357142857143,
+      "grad_norm": 1.9579252309381692,
+      "learning_rate": 6.0212599089484026e-06,
+      "loss": 0.2676,
+      "mean_token_accuracy": 0.9087077379226685,
+      "num_tokens": 620323954.0,
+      "step": 1650
+    },
+    {
+      "epoch": 2.456845238095238,
+      "grad_norm": 1.7388725794154207,
+      "learning_rate": 6.016174433675197e-06,
+      "loss": 0.2781,
+      "mean_token_accuracy": 0.9024098515510559,
+      "num_tokens": 620703792.0,
+      "step": 1651
+    },
+    {
+      "epoch": 2.4583333333333335,
+      "grad_norm": 1.855084875087374,
+      "learning_rate": 6.011087861660191e-06,
+      "loss": 0.302,
+      "mean_token_accuracy": 0.8977838754653931,
+      "num_tokens": 621082541.0,
+      "step": 1652
+    },
+    {
+      "epoch": 2.4598214285714284,
+      "grad_norm": 1.7642085101516054,
+      "learning_rate": 6.0060001983932425e-06,
+      "loss": 0.2702,
+      "mean_token_accuracy": 0.9084990620613098,
+      "num_tokens": 621460447.0,
+      "step": 1653
+    },
+    {
+      "epoch": 2.4613095238095237,
+      "grad_norm": 1.752306546746091,
+      "learning_rate": 6.0009114493653905e-06,
+      "loss": 0.3258,
+      "mean_token_accuracy": 0.8896963596343994,
+      "num_tokens": 621840087.0,
+      "step": 1654
+    },
+    {
+      "epoch": 2.462797619047619,
+      "grad_norm": 1.7016191377460201,
+      "learning_rate": 5.995821620068848e-06,
+      "loss": 0.2825,
+      "mean_token_accuracy": 0.9012324810028076,
+      "num_tokens": 622218947.0,
+      "step": 1655
+    },
+    {
+      "epoch": 2.4642857142857144,
+      "grad_norm": 1.7767827914164733,
+      "learning_rate": 5.990730715996989e-06,
+      "loss": 0.2904,
+      "mean_token_accuracy": 0.8963857889175415,
+      "num_tokens": 622596410.0,
+      "step": 1656
+    },
+    {
+      "epoch": 2.4657738095238093,
+      "grad_norm": 1.8859583701200444,
+      "learning_rate": 5.98563874264435e-06,
+      "loss": 0.3348,
+      "mean_token_accuracy": 0.8836649060249329,
+      "num_tokens": 622961373.0,
+      "step": 1657
+    },
+    {
+      "epoch": 2.4672619047619047,
+      "grad_norm": 1.879556522662271,
+      "learning_rate": 5.980545705506622e-06,
+      "loss": 0.2779,
+      "mean_token_accuracy": 0.903836727142334,
+      "num_tokens": 623337496.0,
+      "step": 1658
+    },
+    {
+      "epoch": 2.46875,
+      "grad_norm": 1.7225001899163055,
+      "learning_rate": 5.975451610080643e-06,
+      "loss": 0.293,
+      "mean_token_accuracy": 0.9016960859298706,
+      "num_tokens": 623715693.0,
+      "step": 1659
+    },
+    {
+      "epoch": 2.4702380952380953,
+      "grad_norm": 1.7002594054191618,
+      "learning_rate": 5.970356461864392e-06,
+      "loss": 0.2789,
+      "mean_token_accuracy": 0.903401255607605,
+      "num_tokens": 624093069.0,
+      "step": 1660
+    },
+    {
+      "epoch": 2.4717261904761907,
+      "grad_norm": 1.8063267075567253,
+      "learning_rate": 5.9652602663569905e-06,
+      "loss": 0.3461,
+      "mean_token_accuracy": 0.8788776397705078,
+      "num_tokens": 624471356.0,
+      "step": 1661
+    },
+    {
+      "epoch": 2.4732142857142856,
+      "grad_norm": 1.7154225920996458,
+      "learning_rate": 5.960163029058682e-06,
+      "loss": 0.2702,
+      "mean_token_accuracy": 0.902540385723114,
+      "num_tokens": 624848934.0,
+      "step": 1662
+    },
+    {
+      "epoch": 2.474702380952381,
+      "grad_norm": 1.7816626648747573,
+      "learning_rate": 5.955064755470839e-06,
+      "loss": 0.3054,
+      "mean_token_accuracy": 0.8941747546195984,
+      "num_tokens": 625227187.0,
+      "step": 1663
+    },
+    {
+      "epoch": 2.4761904761904763,
+      "grad_norm": 1.7952199959864077,
+      "learning_rate": 5.949965451095952e-06,
+      "loss": 0.3072,
+      "mean_token_accuracy": 0.8937128186225891,
+      "num_tokens": 625598743.0,
+      "step": 1664
+    },
+    {
+      "epoch": 2.4776785714285716,
+      "grad_norm": 1.8268535848528424,
+      "learning_rate": 5.944865121437622e-06,
+      "loss": 0.2973,
+      "mean_token_accuracy": 0.8968846201896667,
+      "num_tokens": 625977325.0,
+      "step": 1665
+    },
+    {
+      "epoch": 2.4791666666666665,
+      "grad_norm": 1.7710067922166424,
+      "learning_rate": 5.93976377200056e-06,
+      "loss": 0.2978,
+      "mean_token_accuracy": 0.8961119055747986,
+      "num_tokens": 626340935.0,
+      "step": 1666
+    },
+    {
+      "epoch": 2.480654761904762,
+      "grad_norm": 1.7189070192921012,
+      "learning_rate": 5.934661408290576e-06,
+      "loss": 0.2516,
+      "mean_token_accuracy": 0.912659764289856,
+      "num_tokens": 626718726.0,
+      "step": 1667
+    },
+    {
+      "epoch": 2.482142857142857,
+      "grad_norm": 1.8278906953534702,
+      "learning_rate": 5.929558035814574e-06,
+      "loss": 0.2431,
+      "mean_token_accuracy": 0.9128516912460327,
+      "num_tokens": 627095224.0,
+      "step": 1668
+    },
+    {
+      "epoch": 2.4836309523809526,
+      "grad_norm": 1.810582321542149,
+      "learning_rate": 5.924453660080546e-06,
+      "loss": 0.3383,
+      "mean_token_accuracy": 0.8816319704055786,
+      "num_tokens": 627474805.0,
+      "step": 1669
+    },
+    {
+      "epoch": 2.4851190476190474,
+      "grad_norm": 1.916189590429685,
+      "learning_rate": 5.919348286597569e-06,
+      "loss": 0.3108,
+      "mean_token_accuracy": 0.892701268196106,
+      "num_tokens": 627852962.0,
+      "step": 1670
+    },
+    {
+      "epoch": 2.486607142857143,
+      "grad_norm": 1.856538691923473,
+      "learning_rate": 5.914241920875799e-06,
+      "loss": 0.3021,
+      "mean_token_accuracy": 0.8977933526039124,
+      "num_tokens": 628230679.0,
+      "step": 1671
+    },
+    {
+      "epoch": 2.488095238095238,
+      "grad_norm": 1.7933681883489667,
+      "learning_rate": 5.909134568426455e-06,
+      "loss": 0.2864,
+      "mean_token_accuracy": 0.8941745758056641,
+      "num_tokens": 628608198.0,
+      "step": 1672
+    },
+    {
+      "epoch": 2.4895833333333335,
+      "grad_norm": 1.9084275750067443,
+      "learning_rate": 5.904026234761827e-06,
+      "loss": 0.283,
+      "mean_token_accuracy": 0.9005307555198669,
+      "num_tokens": 628986062.0,
+      "step": 1673
+    },
+    {
+      "epoch": 2.4910714285714284,
+      "grad_norm": 1.756805297592627,
+      "learning_rate": 5.8989169253952635e-06,
+      "loss": 0.2582,
+      "mean_token_accuracy": 0.9080148339271545,
+      "num_tokens": 629364229.0,
+      "step": 1674
+    },
+    {
+      "epoch": 2.4925595238095237,
+      "grad_norm": 1.922543949620882,
+      "learning_rate": 5.893806645841167e-06,
+      "loss": 0.3329,
+      "mean_token_accuracy": 0.8835130333900452,
+      "num_tokens": 629717247.0,
+      "step": 1675
+    },
+    {
+      "epoch": 2.494047619047619,
+      "grad_norm": 1.8968257167763432,
+      "learning_rate": 5.888695401614982e-06,
+      "loss": 0.314,
+      "mean_token_accuracy": 0.8888983726501465,
+      "num_tokens": 630096293.0,
+      "step": 1676
+    },
+    {
+      "epoch": 2.4955357142857144,
+      "grad_norm": 1.6565603684973977,
+      "learning_rate": 5.883583198233202e-06,
+      "loss": 0.249,
+      "mean_token_accuracy": 0.9092637896537781,
+      "num_tokens": 630473885.0,
+      "step": 1677
+    },
+    {
+      "epoch": 2.4970238095238093,
+      "grad_norm": 1.8595573010864712,
+      "learning_rate": 5.878470041213345e-06,
+      "loss": 0.3191,
+      "mean_token_accuracy": 0.8868603706359863,
+      "num_tokens": 630852056.0,
+      "step": 1678
+    },
+    {
+      "epoch": 2.4985119047619047,
+      "grad_norm": 1.7372469619858653,
+      "learning_rate": 5.87335593607397e-06,
+      "loss": 0.3086,
+      "mean_token_accuracy": 0.8930187225341797,
+      "num_tokens": 631231485.0,
+      "step": 1679
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 1.7804683196265199,
+      "learning_rate": 5.8682408883346535e-06,
+      "loss": 0.2904,
+      "mean_token_accuracy": 0.8965833187103271,
+      "num_tokens": 631609229.0,
+      "step": 1680
+    },
+    {
+      "epoch": 2.5014880952380953,
+      "grad_norm": 1.7821638911155675,
+      "learning_rate": 5.863124903515986e-06,
+      "loss": 0.3055,
+      "mean_token_accuracy": 0.8940420150756836,
+      "num_tokens": 631988092.0,
+      "step": 1681
+    },
+    {
+      "epoch": 2.5029761904761907,
+      "grad_norm": 1.8011825706900124,
+      "learning_rate": 5.8580079871395766e-06,
+      "loss": 0.3196,
+      "mean_token_accuracy": 0.8871703743934631,
+      "num_tokens": 632366808.0,
+      "step": 1682
+    },
+    {
+      "epoch": 2.5044642857142856,
+      "grad_norm": 1.7837222643953234,
+      "learning_rate": 5.852890144728034e-06,
+      "loss": 0.2935,
+      "mean_token_accuracy": 0.8967921733856201,
+      "num_tokens": 632745694.0,
+      "step": 1683
+    },
+    {
+      "epoch": 2.505952380952381,
+      "grad_norm": 1.8750107199117505,
+      "learning_rate": 5.847771381804972e-06,
+      "loss": 0.335,
+      "mean_token_accuracy": 0.8846786022186279,
+      "num_tokens": 633123966.0,
+      "step": 1684
+    },
+    {
+      "epoch": 2.5074404761904763,
+      "grad_norm": 1.7823681739456114,
+      "learning_rate": 5.842651703894989e-06,
+      "loss": 0.2874,
+      "mean_token_accuracy": 0.8991588950157166,
+      "num_tokens": 633500833.0,
+      "step": 1685
+    },
+    {
+      "epoch": 2.508928571428571,
+      "grad_norm": 1.690472440775116,
+      "learning_rate": 5.837531116523683e-06,
+      "loss": 0.3188,
+      "mean_token_accuracy": 0.8875058889389038,
+      "num_tokens": 633880395.0,
+      "step": 1686
+    },
+    {
+      "epoch": 2.5104166666666665,
+      "grad_norm": 1.843573311988417,
+      "learning_rate": 5.832409625217623e-06,
+      "loss": 0.2763,
+      "mean_token_accuracy": 0.9022058248519897,
+      "num_tokens": 634257859.0,
+      "step": 1687
+    },
+    {
+      "epoch": 2.511904761904762,
+      "grad_norm": 1.7709225974936704,
+      "learning_rate": 5.827287235504356e-06,
+      "loss": 0.2567,
+      "mean_token_accuracy": 0.9101659059524536,
+      "num_tokens": 634635287.0,
+      "step": 1688
+    },
+    {
+      "epoch": 2.513392857142857,
+      "grad_norm": 1.8065300068625862,
+      "learning_rate": 5.822163952912404e-06,
+      "loss": 0.2951,
+      "mean_token_accuracy": 0.8921980857849121,
+      "num_tokens": 635014088.0,
+      "step": 1689
+    },
+    {
+      "epoch": 2.5148809523809526,
+      "grad_norm": 1.8271394430735246,
+      "learning_rate": 5.817039782971249e-06,
+      "loss": 0.3088,
+      "mean_token_accuracy": 0.89557945728302,
+      "num_tokens": 635392045.0,
+      "step": 1690
+    },
+    {
+      "epoch": 2.5163690476190474,
+      "grad_norm": 1.7157310195428894,
+      "learning_rate": 5.811914731211329e-06,
+      "loss": 0.2727,
+      "mean_token_accuracy": 0.8970840573310852,
+      "num_tokens": 635770620.0,
+      "step": 1691
+    },
+    {
+      "epoch": 2.517857142857143,
+      "grad_norm": 1.9467395709844209,
+      "learning_rate": 5.806788803164034e-06,
+      "loss": 0.2706,
+      "mean_token_accuracy": 0.9032716751098633,
+      "num_tokens": 636149045.0,
+      "step": 1692
+    },
+    {
+      "epoch": 2.519345238095238,
+      "grad_norm": 1.7980806919024137,
+      "learning_rate": 5.801662004361703e-06,
+      "loss": 0.3285,
+      "mean_token_accuracy": 0.8900434970855713,
+      "num_tokens": 636527974.0,
+      "step": 1693
+    },
+    {
+      "epoch": 2.5208333333333335,
+      "grad_norm": 1.7674257342569182,
+      "learning_rate": 5.796534340337614e-06,
+      "loss": 0.2792,
+      "mean_token_accuracy": 0.9018733501434326,
+      "num_tokens": 636905478.0,
+      "step": 1694
+    },
+    {
+      "epoch": 2.522321428571429,
+      "grad_norm": 1.776887872404245,
+      "learning_rate": 5.791405816625974e-06,
+      "loss": 0.3159,
+      "mean_token_accuracy": 0.8900159001350403,
+      "num_tokens": 637284666.0,
+      "step": 1695
+    },
+    {
+      "epoch": 2.5238095238095237,
+      "grad_norm": 1.7725027435435181,
+      "learning_rate": 5.786276438761928e-06,
+      "loss": 0.2893,
+      "mean_token_accuracy": 0.8992780447006226,
+      "num_tokens": 637663003.0,
+      "step": 1696
+    },
+    {
+      "epoch": 2.525297619047619,
+      "grad_norm": 1.8420412105683435,
+      "learning_rate": 5.78114621228153e-06,
+      "loss": 0.3017,
+      "mean_token_accuracy": 0.8964527249336243,
+      "num_tokens": 638041514.0,
+      "step": 1697
+    },
+    {
+      "epoch": 2.5267857142857144,
+      "grad_norm": 1.7961673928149309,
+      "learning_rate": 5.776015142721758e-06,
+      "loss": 0.321,
+      "mean_token_accuracy": 0.8936859965324402,
+      "num_tokens": 638420610.0,
+      "step": 1698
+    },
+    {
+      "epoch": 2.5282738095238093,
+      "grad_norm": 1.803183183948107,
+      "learning_rate": 5.7708832356205e-06,
+      "loss": 0.3115,
+      "mean_token_accuracy": 0.8958680629730225,
+      "num_tokens": 638798570.0,
+      "step": 1699
+    },
+    {
+      "epoch": 2.5297619047619047,
+      "grad_norm": 1.8064616294468916,
+      "learning_rate": 5.765750496516547e-06,
+      "loss": 0.3277,
+      "mean_token_accuracy": 0.8827222585678101,
+      "num_tokens": 639177765.0,
+      "step": 1700
+    },
+    {
+      "epoch": 2.53125,
+      "grad_norm": 1.7610075074433607,
+      "learning_rate": 5.760616930949584e-06,
+      "loss": 0.3022,
+      "mean_token_accuracy": 0.8969072103500366,
+      "num_tokens": 639555912.0,
+      "step": 1701
+    },
+    {
+      "epoch": 2.5327380952380953,
+      "grad_norm": 1.8541658977393052,
+      "learning_rate": 5.755482544460193e-06,
+      "loss": 0.3101,
+      "mean_token_accuracy": 0.8953251242637634,
+      "num_tokens": 639932974.0,
+      "step": 1702
+    },
+    {
+      "epoch": 2.5342261904761907,
+      "grad_norm": 1.6916531819207605,
+      "learning_rate": 5.750347342589841e-06,
+      "loss": 0.316,
+      "mean_token_accuracy": 0.8862359523773193,
+      "num_tokens": 640305220.0,
+      "step": 1703
+    },
+    {
+      "epoch": 2.5357142857142856,
+      "grad_norm": 2.552376209080908,
+      "learning_rate": 5.745211330880872e-06,
+      "loss": 0.2891,
+      "mean_token_accuracy": 0.8966350555419922,
+      "num_tokens": 640671830.0,
+      "step": 1704
+    },
+    {
+      "epoch": 2.537202380952381,
+      "grad_norm": 1.7213297479288747,
+      "learning_rate": 5.740074514876509e-06,
+      "loss": 0.2626,
+      "mean_token_accuracy": 0.9083141684532166,
+      "num_tokens": 641049581.0,
+      "step": 1705
+    },
+    {
+      "epoch": 2.5386904761904763,
+      "grad_norm": 1.7651336343237451,
+      "learning_rate": 5.73493690012084e-06,
+      "loss": 0.322,
+      "mean_token_accuracy": 0.8866589069366455,
+      "num_tokens": 641429072.0,
+      "step": 1706
+    },
+    {
+      "epoch": 2.540178571428571,
+      "grad_norm": 1.8085082826665166,
+      "learning_rate": 5.729798492158811e-06,
+      "loss": 0.2789,
+      "mean_token_accuracy": 0.8972324728965759,
+      "num_tokens": 641806579.0,
+      "step": 1707
+    },
+    {
+      "epoch": 2.5416666666666665,
+      "grad_norm": 1.7590733998010601,
+      "learning_rate": 5.724659296536234e-06,
+      "loss": 0.3099,
+      "mean_token_accuracy": 0.8943755030632019,
+      "num_tokens": 642185802.0,
+      "step": 1708
+    },
+    {
+      "epoch": 2.543154761904762,
+      "grad_norm": 1.8132015927029361,
+      "learning_rate": 5.719519318799764e-06,
+      "loss": 0.2898,
+      "mean_token_accuracy": 0.8995474576950073,
+      "num_tokens": 642562978.0,
+      "step": 1709
+    },
+    {
+      "epoch": 2.544642857142857,
+      "grad_norm": 1.829348043285556,
+      "learning_rate": 5.7143785644969005e-06,
+      "loss": 0.3357,
+      "mean_token_accuracy": 0.8852957487106323,
+      "num_tokens": 642942591.0,
+      "step": 1710
+    },
+    {
+      "epoch": 2.5461309523809526,
+      "grad_norm": 1.8099812966891802,
+      "learning_rate": 5.7092370391759835e-06,
+      "loss": 0.3101,
+      "mean_token_accuracy": 0.8939916491508484,
+      "num_tokens": 643320272.0,
+      "step": 1711
+    },
+    {
+      "epoch": 2.5476190476190474,
+      "grad_norm": 1.7591034204572489,
+      "learning_rate": 5.7040947483861845e-06,
+      "loss": 0.2867,
+      "mean_token_accuracy": 0.9027193784713745,
+      "num_tokens": 643687576.0,
+      "step": 1712
+    },
+    {
+      "epoch": 2.549107142857143,
+      "grad_norm": 1.9384872524664853,
+      "learning_rate": 5.698951697677498e-06,
+      "loss": 0.331,
+      "mean_token_accuracy": 0.8866644501686096,
+      "num_tokens": 644066007.0,
+      "step": 1713
+    },
+    {
+      "epoch": 2.550595238095238,
+      "grad_norm": 1.8175161404316325,
+      "learning_rate": 5.693807892600744e-06,
+      "loss": 0.294,
+      "mean_token_accuracy": 0.8988448977470398,
+      "num_tokens": 644444859.0,
+      "step": 1714
+    },
+    {
+      "epoch": 2.5520833333333335,
+      "grad_norm": 1.693641964158492,
+      "learning_rate": 5.688663338707554e-06,
+      "loss": 0.2899,
+      "mean_token_accuracy": 0.8974041938781738,
+      "num_tokens": 644825112.0,
+      "step": 1715
+    },
+    {
+      "epoch": 2.553571428571429,
+      "grad_norm": 1.7456379490073448,
+      "learning_rate": 5.683518041550368e-06,
+      "loss": 0.3018,
+      "mean_token_accuracy": 0.893416166305542,
+      "num_tokens": 645203817.0,
+      "step": 1716
+    },
+    {
+      "epoch": 2.5550595238095237,
+      "grad_norm": 1.7790527182719411,
+      "learning_rate": 5.678372006682426e-06,
+      "loss": 0.3101,
+      "mean_token_accuracy": 0.8920033574104309,
+      "num_tokens": 645582248.0,
+      "step": 1717
+    },
+    {
+      "epoch": 2.556547619047619,
+      "grad_norm": 1.8253547207277452,
+      "learning_rate": 5.673225239657767e-06,
+      "loss": 0.3013,
+      "mean_token_accuracy": 0.8939446806907654,
+      "num_tokens": 645960454.0,
+      "step": 1718
+    },
+    {
+      "epoch": 2.5580357142857144,
+      "grad_norm": 1.7534002511423084,
+      "learning_rate": 5.6680777460312195e-06,
+      "loss": 0.2935,
+      "mean_token_accuracy": 0.8987917900085449,
+      "num_tokens": 646339201.0,
+      "step": 1719
+    },
+    {
+      "epoch": 2.5595238095238093,
+      "grad_norm": 1.847621394020163,
+      "learning_rate": 5.6629295313583975e-06,
+      "loss": 0.3002,
+      "mean_token_accuracy": 0.8979237079620361,
+      "num_tokens": 646717768.0,
+      "step": 1720
+    },
+    {
+      "epoch": 2.5610119047619047,
+      "grad_norm": 1.733177335991377,
+      "learning_rate": 5.65778060119569e-06,
+      "loss": 0.3016,
+      "mean_token_accuracy": 0.8944068551063538,
+      "num_tokens": 647084486.0,
+      "step": 1721
+    },
+    {
+      "epoch": 2.5625,
+      "grad_norm": 1.7047079106707776,
+      "learning_rate": 5.65263096110026e-06,
+      "loss": 0.2994,
+      "mean_token_accuracy": 0.8932868242263794,
+      "num_tokens": 647463012.0,
+      "step": 1722
+    },
+    {
+      "epoch": 2.5639880952380953,
+      "grad_norm": 1.7177668229705543,
+      "learning_rate": 5.6474806166300355e-06,
+      "loss": 0.2845,
+      "mean_token_accuracy": 0.8997309803962708,
+      "num_tokens": 647840540.0,
+      "step": 1723
+    },
+    {
+      "epoch": 2.5654761904761907,
+      "grad_norm": 1.7382029130319054,
+      "learning_rate": 5.642329573343708e-06,
+      "loss": 0.3359,
+      "mean_token_accuracy": 0.8842544555664062,
+      "num_tokens": 648220130.0,
+      "step": 1724
+    },
+    {
+      "epoch": 2.5669642857142856,
+      "grad_norm": 1.7081345796317353,
+      "learning_rate": 5.637177836800723e-06,
+      "loss": 0.2671,
+      "mean_token_accuracy": 0.9063478112220764,
+      "num_tokens": 648598436.0,
+      "step": 1725
+    },
+    {
+      "epoch": 2.568452380952381,
+      "grad_norm": 1.8066036926054578,
+      "learning_rate": 5.6320254125612685e-06,
+      "loss": 0.3127,
+      "mean_token_accuracy": 0.8916666507720947,
+      "num_tokens": 648977528.0,
+      "step": 1726
+    },
+    {
+      "epoch": 2.5699404761904763,
+      "grad_norm": 1.7660859340937063,
+      "learning_rate": 5.626872306186278e-06,
+      "loss": 0.3183,
+      "mean_token_accuracy": 0.8853877186775208,
+      "num_tokens": 649357822.0,
+      "step": 1727
+    },
+    {
+      "epoch": 2.571428571428571,
+      "grad_norm": 1.8747080269883298,
+      "learning_rate": 5.621718523237427e-06,
+      "loss": 0.3086,
+      "mean_token_accuracy": 0.8947267532348633,
+      "num_tokens": 649734862.0,
+      "step": 1728
+    },
+    {
+      "epoch": 2.5729166666666665,
+      "grad_norm": 1.7960879406145815,
+      "learning_rate": 5.616564069277111e-06,
+      "loss": 0.2779,
+      "mean_token_accuracy": 0.9058892130851746,
+      "num_tokens": 650099889.0,
+      "step": 1729
+    },
+    {
+      "epoch": 2.574404761904762,
+      "grad_norm": 1.6720540894641913,
+      "learning_rate": 5.611408949868457e-06,
+      "loss": 0.2812,
+      "mean_token_accuracy": 0.8986939191818237,
+      "num_tokens": 650478949.0,
+      "step": 1730
+    },
+    {
+      "epoch": 2.575892857142857,
+      "grad_norm": 1.7423383147059197,
+      "learning_rate": 5.6062531705753075e-06,
+      "loss": 0.2819,
+      "mean_token_accuracy": 0.9001215696334839,
+      "num_tokens": 650856777.0,
+      "step": 1731
+    },
+    {
+      "epoch": 2.5773809523809526,
+      "grad_norm": 1.843469081182871,
+      "learning_rate": 5.60109673696222e-06,
+      "loss": 0.2908,
+      "mean_token_accuracy": 0.9017924070358276,
+      "num_tokens": 651219224.0,
+      "step": 1732
+    },
+    {
+      "epoch": 2.5788690476190474,
+      "grad_norm": 1.7912755049127764,
+      "learning_rate": 5.59593965459445e-06,
+      "loss": 0.2976,
+      "mean_token_accuracy": 0.895260751247406,
+      "num_tokens": 651596665.0,
+      "step": 1733
+    },
+    {
+      "epoch": 2.580357142857143,
+      "grad_norm": 1.8150622367433014,
+      "learning_rate": 5.590781929037965e-06,
+      "loss": 0.3283,
+      "mean_token_accuracy": 0.8872528672218323,
+      "num_tokens": 651975678.0,
+      "step": 1734
+    },
+    {
+      "epoch": 2.581845238095238,
+      "grad_norm": 1.8622840586632836,
+      "learning_rate": 5.5856235658594184e-06,
+      "loss": 0.3374,
+      "mean_token_accuracy": 0.8870338201522827,
+      "num_tokens": 652354895.0,
+      "step": 1735
+    },
+    {
+      "epoch": 2.5833333333333335,
+      "grad_norm": 1.766557991775029,
+      "learning_rate": 5.5804645706261515e-06,
+      "loss": 0.2803,
+      "mean_token_accuracy": 0.8989169597625732,
+      "num_tokens": 652732908.0,
+      "step": 1736
+    },
+    {
+      "epoch": 2.584821428571429,
+      "grad_norm": 1.744256026489447,
+      "learning_rate": 5.5753049489061935e-06,
+      "loss": 0.2888,
+      "mean_token_accuracy": 0.9034085869789124,
+      "num_tokens": 653110368.0,
+      "step": 1737
+    },
+    {
+      "epoch": 2.5863095238095237,
+      "grad_norm": 1.7873706986750497,
+      "learning_rate": 5.570144706268248e-06,
+      "loss": 0.3085,
+      "mean_token_accuracy": 0.896795928478241,
+      "num_tokens": 653486894.0,
+      "step": 1738
+    },
+    {
+      "epoch": 2.587797619047619,
+      "grad_norm": 1.7766722002936832,
+      "learning_rate": 5.564983848281685e-06,
+      "loss": 0.278,
+      "mean_token_accuracy": 0.9010931253433228,
+      "num_tokens": 653865629.0,
+      "step": 1739
+    },
+    {
+      "epoch": 2.5892857142857144,
+      "grad_norm": 1.752302647709912,
+      "learning_rate": 5.559822380516539e-06,
+      "loss": 0.2863,
+      "mean_token_accuracy": 0.9016407132148743,
+      "num_tokens": 654235853.0,
+      "step": 1740
+    },
+    {
+      "epoch": 2.5907738095238093,
+      "grad_norm": 1.758578744852073,
+      "learning_rate": 5.5546603085435105e-06,
+      "loss": 0.2557,
+      "mean_token_accuracy": 0.9069262146949768,
+      "num_tokens": 654611846.0,
+      "step": 1741
+    },
+    {
+      "epoch": 2.5922619047619047,
+      "grad_norm": 1.7383170330790485,
+      "learning_rate": 5.549497637933939e-06,
+      "loss": 0.2958,
+      "mean_token_accuracy": 0.8999999761581421,
+      "num_tokens": 654978532.0,
+      "step": 1742
+    },
+    {
+      "epoch": 2.59375,
+      "grad_norm": 1.8407728629723186,
+      "learning_rate": 5.544334374259823e-06,
+      "loss": 0.2905,
+      "mean_token_accuracy": 0.8993387222290039,
+      "num_tokens": 655350745.0,
+      "step": 1743
+    },
+    {
+      "epoch": 2.5952380952380953,
+      "grad_norm": 1.696720097697988,
+      "learning_rate": 5.539170523093794e-06,
+      "loss": 0.281,
+      "mean_token_accuracy": 0.8984104990959167,
+      "num_tokens": 655729508.0,
+      "step": 1744
+    },
+    {
+      "epoch": 2.5967261904761907,
+      "grad_norm": 1.9009632933524543,
+      "learning_rate": 5.534006090009119e-06,
+      "loss": 0.3455,
+      "mean_token_accuracy": 0.8830766677856445,
+      "num_tokens": 656105629.0,
+      "step": 1745
+    },
+    {
+      "epoch": 2.5982142857142856,
+      "grad_norm": 1.9732041854943654,
+      "learning_rate": 5.52884108057969e-06,
+      "loss": 0.2875,
+      "mean_token_accuracy": 0.9001321792602539,
+      "num_tokens": 656482997.0,
+      "step": 1746
+    },
+    {
+      "epoch": 2.599702380952381,
+      "grad_norm": 1.9235874740154761,
+      "learning_rate": 5.523675500380027e-06,
+      "loss": 0.311,
+      "mean_token_accuracy": 0.8874262571334839,
+      "num_tokens": 656859594.0,
+      "step": 1747
+    },
+    {
+      "epoch": 2.6011904761904763,
+      "grad_norm": 1.6848668209952489,
+      "learning_rate": 5.518509354985264e-06,
+      "loss": 0.2628,
+      "mean_token_accuracy": 0.9045011401176453,
+      "num_tokens": 657236369.0,
+      "step": 1748
+    },
+    {
+      "epoch": 2.602678571428571,
+      "grad_norm": 1.886104225242948,
+      "learning_rate": 5.513342649971143e-06,
+      "loss": 0.3066,
+      "mean_token_accuracy": 0.8928264379501343,
+      "num_tokens": 657613523.0,
+      "step": 1749
+    },
+    {
+      "epoch": 2.6041666666666665,
+      "grad_norm": 1.8562163356318038,
+      "learning_rate": 5.50817539091401e-06,
+      "loss": 0.2693,
+      "mean_token_accuracy": 0.9048970341682434,
+      "num_tokens": 657991464.0,
+      "step": 1750
+    },
+    {
+      "epoch": 2.605654761904762,
+      "grad_norm": 1.753722080768625,
+      "learning_rate": 5.5030075833908114e-06,
+      "loss": 0.2594,
+      "mean_token_accuracy": 0.9056510925292969,
+      "num_tokens": 658356190.0,
+      "step": 1751
+    },
+    {
+      "epoch": 2.607142857142857,
+      "grad_norm": 1.8502340809059437,
+      "learning_rate": 5.497839232979084e-06,
+      "loss": 0.2951,
+      "mean_token_accuracy": 0.8987883925437927,
+      "num_tokens": 658734045.0,
+      "step": 1752
+    },
+    {
+      "epoch": 2.6086309523809526,
+      "grad_norm": 1.9194048135096915,
+      "learning_rate": 5.492670345256949e-06,
+      "loss": 0.2943,
+      "mean_token_accuracy": 0.8937073349952698,
+      "num_tokens": 659111831.0,
+      "step": 1753
+    },
+    {
+      "epoch": 2.6101190476190474,
+      "grad_norm": 1.7358035126750253,
+      "learning_rate": 5.487500925803115e-06,
+      "loss": 0.2886,
+      "mean_token_accuracy": 0.9031521081924438,
+      "num_tokens": 659489542.0,
+      "step": 1754
+    },
+    {
+      "epoch": 2.611607142857143,
+      "grad_norm": 1.7370459330105956,
+      "learning_rate": 5.4823309801968515e-06,
+      "loss": 0.3161,
+      "mean_token_accuracy": 0.8903383612632751,
+      "num_tokens": 659868277.0,
+      "step": 1755
+    },
+    {
+      "epoch": 2.613095238095238,
+      "grad_norm": 1.7795437344896512,
+      "learning_rate": 5.477160514018007e-06,
+      "loss": 0.294,
+      "mean_token_accuracy": 0.8962568640708923,
+      "num_tokens": 660245691.0,
+      "step": 1756
+    },
+    {
+      "epoch": 2.6145833333333335,
+      "grad_norm": 1.7432555544019537,
+      "learning_rate": 5.471989532846987e-06,
+      "loss": 0.3096,
+      "mean_token_accuracy": 0.8904429078102112,
+      "num_tokens": 660624664.0,
+      "step": 1757
+    },
+    {
+      "epoch": 2.616071428571429,
+      "grad_norm": 1.8317537306609348,
+      "learning_rate": 5.466818042264754e-06,
+      "loss": 0.2983,
+      "mean_token_accuracy": 0.892346203327179,
+      "num_tokens": 661002564.0,
+      "step": 1758
+    },
+    {
+      "epoch": 2.6175595238095237,
+      "grad_norm": 1.8577893599932316,
+      "learning_rate": 5.461646047852821e-06,
+      "loss": 0.318,
+      "mean_token_accuracy": 0.887017548084259,
+      "num_tokens": 661381240.0,
+      "step": 1759
+    },
+    {
+      "epoch": 2.619047619047619,
+      "grad_norm": 1.8452508591657342,
+      "learning_rate": 5.456473555193242e-06,
+      "loss": 0.3086,
+      "mean_token_accuracy": 0.8929417133331299,
+      "num_tokens": 661759191.0,
+      "step": 1760
+    },
+    {
+      "epoch": 2.6205357142857144,
+      "grad_norm": 1.7601506085501508,
+      "learning_rate": 5.4513005698686124e-06,
+      "loss": 0.2844,
+      "mean_token_accuracy": 0.8994066715240479,
+      "num_tokens": 662136562.0,
+      "step": 1761
+    },
+    {
+      "epoch": 2.6220238095238093,
+      "grad_norm": 1.8164923230484717,
+      "learning_rate": 5.446127097462056e-06,
+      "loss": 0.3626,
+      "mean_token_accuracy": 0.8744558095932007,
+      "num_tokens": 662516238.0,
+      "step": 1762
+    },
+    {
+      "epoch": 2.6235119047619047,
+      "grad_norm": 1.7684886231249672,
+      "learning_rate": 5.440953143557226e-06,
+      "loss": 0.2607,
+      "mean_token_accuracy": 0.9094974398612976,
+      "num_tokens": 662892578.0,
+      "step": 1763
+    },
+    {
+      "epoch": 2.625,
+      "grad_norm": 1.8429888061658377,
+      "learning_rate": 5.435778713738292e-06,
+      "loss": 0.2886,
+      "mean_token_accuracy": 0.9018617272377014,
+      "num_tokens": 663269756.0,
+      "step": 1764
+    },
+    {
+      "epoch": 2.6264880952380953,
+      "grad_norm": 1.6283300134033296,
+      "learning_rate": 5.430603813589938e-06,
+      "loss": 0.2871,
+      "mean_token_accuracy": 0.9044865965843201,
+      "num_tokens": 663638945.0,
+      "step": 1765
+    },
+    {
+      "epoch": 2.6279761904761907,
+      "grad_norm": 1.7142504743928833,
+      "learning_rate": 5.425428448697354e-06,
+      "loss": 0.2825,
+      "mean_token_accuracy": 0.9006226658821106,
+      "num_tokens": 664018070.0,
+      "step": 1766
+    },
+    {
+      "epoch": 2.6294642857142856,
+      "grad_norm": 1.7838337281534673,
+      "learning_rate": 5.420252624646238e-06,
+      "loss": 0.2945,
+      "mean_token_accuracy": 0.9023056626319885,
+      "num_tokens": 664396422.0,
+      "step": 1767
+    },
+    {
+      "epoch": 2.630952380952381,
+      "grad_norm": 1.810575972192507,
+      "learning_rate": 5.415076347022777e-06,
+      "loss": 0.3026,
+      "mean_token_accuracy": 0.8947454690933228,
+      "num_tokens": 664776085.0,
+      "step": 1768
+    },
+    {
+      "epoch": 2.6324404761904763,
+      "grad_norm": 1.8126021912538441,
+      "learning_rate": 5.409899621413649e-06,
+      "loss": 0.3053,
+      "mean_token_accuracy": 0.8947128653526306,
+      "num_tokens": 665153248.0,
+      "step": 1769
+    },
+    {
+      "epoch": 2.633928571428571,
+      "grad_norm": 1.755635283808318,
+      "learning_rate": 5.404722453406017e-06,
+      "loss": 0.3468,
+      "mean_token_accuracy": 0.8829869627952576,
+      "num_tokens": 665512992.0,
+      "step": 1770
+    },
+    {
+      "epoch": 2.6354166666666665,
+      "grad_norm": 1.7518047735483864,
+      "learning_rate": 5.3995448485875205e-06,
+      "loss": 0.3087,
+      "mean_token_accuracy": 0.8956295847892761,
+      "num_tokens": 665891252.0,
+      "step": 1771
+    },
+    {
+      "epoch": 2.636904761904762,
+      "grad_norm": 1.6383137452014087,
+      "learning_rate": 5.39436681254627e-06,
+      "loss": 0.298,
+      "mean_token_accuracy": 0.8938294053077698,
+      "num_tokens": 666270261.0,
+      "step": 1772
+    },
+    {
+      "epoch": 2.638392857142857,
+      "grad_norm": 1.80585344448709,
+      "learning_rate": 5.389188350870844e-06,
+      "loss": 0.2749,
+      "mean_token_accuracy": 0.9050989151000977,
+      "num_tokens": 666648211.0,
+      "step": 1773
+    },
+    {
+      "epoch": 2.6398809523809526,
+      "grad_norm": 1.679616437438045,
+      "learning_rate": 5.384009469150276e-06,
+      "loss": 0.2737,
+      "mean_token_accuracy": 0.9061932563781738,
+      "num_tokens": 667014109.0,
+      "step": 1774
+    },
+    {
+      "epoch": 2.6413690476190474,
+      "grad_norm": 1.723126815268501,
+      "learning_rate": 5.378830172974056e-06,
+      "loss": 0.3282,
+      "mean_token_accuracy": 0.8840758800506592,
+      "num_tokens": 667391066.0,
+      "step": 1775
+    },
+    {
+      "epoch": 2.642857142857143,
+      "grad_norm": 1.6146086944463787,
+      "learning_rate": 5.373650467932122e-06,
+      "loss": 0.2808,
+      "mean_token_accuracy": 0.8988300561904907,
+      "num_tokens": 667770791.0,
+      "step": 1776
+    },
+    {
+      "epoch": 2.644345238095238,
+      "grad_norm": 1.7252970391127618,
+      "learning_rate": 5.36847035961485e-06,
+      "loss": 0.2924,
+      "mean_token_accuracy": 0.9006772041320801,
+      "num_tokens": 668149848.0,
+      "step": 1777
+    },
+    {
+      "epoch": 2.6458333333333335,
+      "grad_norm": 1.7690421200594657,
+      "learning_rate": 5.363289853613054e-06,
+      "loss": 0.2744,
+      "mean_token_accuracy": 0.9027790427207947,
+      "num_tokens": 668527765.0,
+      "step": 1778
+    },
+    {
+      "epoch": 2.647321428571429,
+      "grad_norm": 1.7423469407522603,
+      "learning_rate": 5.358108955517976e-06,
+      "loss": 0.2903,
+      "mean_token_accuracy": 0.9022829532623291,
+      "num_tokens": 668905535.0,
+      "step": 1779
+    },
+    {
+      "epoch": 2.6488095238095237,
+      "grad_norm": 1.8227102771982364,
+      "learning_rate": 5.352927670921282e-06,
+      "loss": 0.3089,
+      "mean_token_accuracy": 0.8901191353797913,
+      "num_tokens": 669275796.0,
+      "step": 1780
+    },
+    {
+      "epoch": 2.650297619047619,
+      "grad_norm": 1.7782015211294142,
+      "learning_rate": 5.3477460054150525e-06,
+      "loss": 0.3147,
+      "mean_token_accuracy": 0.8929716348648071,
+      "num_tokens": 669654526.0,
+      "step": 1781
+    },
+    {
+      "epoch": 2.6517857142857144,
+      "grad_norm": 1.703216310598722,
+      "learning_rate": 5.342563964591783e-06,
+      "loss": 0.2866,
+      "mean_token_accuracy": 0.9001873731613159,
+      "num_tokens": 670033281.0,
+      "step": 1782
+    },
+    {
+      "epoch": 2.6532738095238093,
+      "grad_norm": 1.7801940381175274,
+      "learning_rate": 5.337381554044375e-06,
+      "loss": 0.2944,
+      "mean_token_accuracy": 0.8990245461463928,
+      "num_tokens": 670411186.0,
+      "step": 1783
+    },
+    {
+      "epoch": 2.6547619047619047,
+      "grad_norm": 1.8690218171792368,
+      "learning_rate": 5.332198779366123e-06,
+      "loss": 0.2951,
+      "mean_token_accuracy": 0.9014416933059692,
+      "num_tokens": 670773712.0,
+      "step": 1784
+    },
+    {
+      "epoch": 2.65625,
+      "grad_norm": 1.8795568148829414,
+      "learning_rate": 5.327015646150716e-06,
+      "loss": 0.2914,
+      "mean_token_accuracy": 0.8990954756736755,
+      "num_tokens": 671152029.0,
+      "step": 1785
+    },
+    {
+      "epoch": 2.6577380952380953,
+      "grad_norm": 1.7419360731442954,
+      "learning_rate": 5.321832159992236e-06,
+      "loss": 0.24,
+      "mean_token_accuracy": 0.914436399936676,
+      "num_tokens": 671516802.0,
+      "step": 1786
+    },
+    {
+      "epoch": 2.6592261904761907,
+      "grad_norm": 1.7569519244190788,
+      "learning_rate": 5.316648326485138e-06,
+      "loss": 0.3431,
+      "mean_token_accuracy": 0.8758851289749146,
+      "num_tokens": 671896736.0,
+      "step": 1787
+    },
+    {
+      "epoch": 2.6607142857142856,
+      "grad_norm": 1.7277357747584368,
+      "learning_rate": 5.311464151224261e-06,
+      "loss": 0.3093,
+      "mean_token_accuracy": 0.8947585225105286,
+      "num_tokens": 672276124.0,
+      "step": 1788
+    },
+    {
+      "epoch": 2.662202380952381,
+      "grad_norm": 1.8705821041946253,
+      "learning_rate": 5.306279639804803e-06,
+      "loss": 0.2998,
+      "mean_token_accuracy": 0.8942960500717163,
+      "num_tokens": 672654457.0,
+      "step": 1789
+    },
+    {
+      "epoch": 2.6636904761904763,
+      "grad_norm": 1.7760113940941669,
+      "learning_rate": 5.3010947978223305e-06,
+      "loss": 0.2795,
+      "mean_token_accuracy": 0.9026110768318176,
+      "num_tokens": 673033074.0,
+      "step": 1790
+    },
+    {
+      "epoch": 2.665178571428571,
+      "grad_norm": 1.905092722527128,
+      "learning_rate": 5.295909630872764e-06,
+      "loss": 0.2643,
+      "mean_token_accuracy": 0.909879207611084,
+      "num_tokens": 673408566.0,
+      "step": 1791
+    },
+    {
+      "epoch": 2.6666666666666665,
+      "grad_norm": 1.6557462680396153,
+      "learning_rate": 5.290724144552379e-06,
+      "loss": 0.2808,
+      "mean_token_accuracy": 0.9035109877586365,
+      "num_tokens": 673787080.0,
+      "step": 1792
+    },
+    {
+      "epoch": 2.668154761904762,
+      "grad_norm": 1.8884136012590789,
+      "learning_rate": 5.285538344457793e-06,
+      "loss": 0.3105,
+      "mean_token_accuracy": 0.8922552466392517,
+      "num_tokens": 674164381.0,
+      "step": 1793
+    },
+    {
+      "epoch": 2.669642857142857,
+      "grad_norm": 1.7502642953382956,
+      "learning_rate": 5.2803522361859596e-06,
+      "loss": 0.2561,
+      "mean_token_accuracy": 0.9111534953117371,
+      "num_tokens": 674541448.0,
+      "step": 1794
+    },
+    {
+      "epoch": 2.6711309523809526,
+      "grad_norm": 1.854223184611147,
+      "learning_rate": 5.275165825334169e-06,
+      "loss": 0.3321,
+      "mean_token_accuracy": 0.8844642043113708,
+      "num_tokens": 674917844.0,
+      "step": 1795
+    },
+    {
+      "epoch": 2.6726190476190474,
+      "grad_norm": 1.808190934757633,
+      "learning_rate": 5.269979117500037e-06,
+      "loss": 0.3107,
+      "mean_token_accuracy": 0.896062970161438,
+      "num_tokens": 675296233.0,
+      "step": 1796
+    },
+    {
+      "epoch": 2.674107142857143,
+      "grad_norm": 1.773149478423271,
+      "learning_rate": 5.264792118281497e-06,
+      "loss": 0.3089,
+      "mean_token_accuracy": 0.8895601034164429,
+      "num_tokens": 675675483.0,
+      "step": 1797
+    },
+    {
+      "epoch": 2.675595238095238,
+      "grad_norm": 1.71294236476505,
+      "learning_rate": 5.259604833276805e-06,
+      "loss": 0.302,
+      "mean_token_accuracy": 0.8919471502304077,
+      "num_tokens": 676054356.0,
+      "step": 1798
+    },
+    {
+      "epoch": 2.6770833333333335,
+      "grad_norm": 1.8718634824641223,
+      "learning_rate": 5.254417268084514e-06,
+      "loss": 0.282,
+      "mean_token_accuracy": 0.9024921655654907,
+      "num_tokens": 676432806.0,
+      "step": 1799
+    },
+    {
+      "epoch": 2.678571428571429,
+      "grad_norm": 1.7254267356377817,
+      "learning_rate": 5.249229428303486e-06,
+      "loss": 0.3111,
+      "mean_token_accuracy": 0.8956742882728577,
+      "num_tokens": 676800621.0,
+      "step": 1800
+    },
+    {
+      "epoch": 2.6800595238095237,
+      "grad_norm": 1.8373470137175858,
+      "learning_rate": 5.244041319532881e-06,
+      "loss": 0.3015,
+      "mean_token_accuracy": 0.892895519733429,
+      "num_tokens": 677179681.0,
+      "step": 1801
+    },
+    {
+      "epoch": 2.681547619047619,
+      "grad_norm": 1.7441372489994105,
+      "learning_rate": 5.238852947372146e-06,
+      "loss": 0.2747,
+      "mean_token_accuracy": 0.9007620215415955,
+      "num_tokens": 677558176.0,
+      "step": 1802
+    },
+    {
+      "epoch": 2.6830357142857144,
+      "grad_norm": 1.666690525439584,
+      "learning_rate": 5.233664317421012e-06,
+      "loss": 0.2979,
+      "mean_token_accuracy": 0.8974594473838806,
+      "num_tokens": 677933691.0,
+      "step": 1803
+    },
+    {
+      "epoch": 2.6845238095238093,
+      "grad_norm": 1.8250546212992413,
+      "learning_rate": 5.228475435279491e-06,
+      "loss": 0.3106,
+      "mean_token_accuracy": 0.8913161158561707,
+      "num_tokens": 678312529.0,
+      "step": 1804
+    },
+    {
+      "epoch": 2.6860119047619047,
+      "grad_norm": 1.6967845504344123,
+      "learning_rate": 5.223286306547865e-06,
+      "loss": 0.2638,
+      "mean_token_accuracy": 0.9059560894966125,
+      "num_tokens": 678690918.0,
+      "step": 1805
+    },
+    {
+      "epoch": 2.6875,
+      "grad_norm": 1.7933983111210838,
+      "learning_rate": 5.218096936826681e-06,
+      "loss": 0.2705,
+      "mean_token_accuracy": 0.9074429273605347,
+      "num_tokens": 679068167.0,
+      "step": 1806
+    },
+    {
+      "epoch": 2.6889880952380953,
+      "grad_norm": 1.9097449672102271,
+      "learning_rate": 5.21290733171675e-06,
+      "loss": 0.2824,
+      "mean_token_accuracy": 0.900542676448822,
+      "num_tokens": 679425853.0,
+      "step": 1807
+    },
+    {
+      "epoch": 2.6904761904761907,
+      "grad_norm": 1.8778863149397718,
+      "learning_rate": 5.207717496819134e-06,
+      "loss": 0.2681,
+      "mean_token_accuracy": 0.9038113951683044,
+      "num_tokens": 679803062.0,
+      "step": 1808
+    },
+    {
+      "epoch": 2.6919642857142856,
+      "grad_norm": 1.706551472635396,
+      "learning_rate": 5.202527437735143e-06,
+      "loss": 0.3006,
+      "mean_token_accuracy": 0.8958194255828857,
+      "num_tokens": 680181343.0,
+      "step": 1809
+    },
+    {
+      "epoch": 2.693452380952381,
+      "grad_norm": 1.938946289590715,
+      "learning_rate": 5.1973371600663315e-06,
+      "loss": 0.3116,
+      "mean_token_accuracy": 0.8884263038635254,
+      "num_tokens": 680546914.0,
+      "step": 1810
+    },
+    {
+      "epoch": 2.6949404761904763,
+      "grad_norm": 1.7543515638210492,
+      "learning_rate": 5.192146669414486e-06,
+      "loss": 0.2688,
+      "mean_token_accuracy": 0.9037473201751709,
+      "num_tokens": 680924321.0,
+      "step": 1811
+    },
+    {
+      "epoch": 2.696428571428571,
+      "grad_norm": 1.785286842695711,
+      "learning_rate": 5.18695597138163e-06,
+      "loss": 0.3133,
+      "mean_token_accuracy": 0.8892757296562195,
+      "num_tokens": 681303780.0,
+      "step": 1812
+    },
+    {
+      "epoch": 2.6979166666666665,
+      "grad_norm": 1.9502431281900314,
+      "learning_rate": 5.18176507157e-06,
+      "loss": 0.28,
+      "mean_token_accuracy": 0.8951128125190735,
+      "num_tokens": 681681441.0,
+      "step": 1813
+    },
+    {
+      "epoch": 2.699404761904762,
+      "grad_norm": 1.6734695007751406,
+      "learning_rate": 5.176573975582058e-06,
+      "loss": 0.247,
+      "mean_token_accuracy": 0.9141467213630676,
+      "num_tokens": 682058384.0,
+      "step": 1814
+    },
+    {
+      "epoch": 2.700892857142857,
+      "grad_norm": 1.8263273279723142,
+      "learning_rate": 5.1713826890204765e-06,
+      "loss": 0.2777,
+      "mean_token_accuracy": 0.9009387493133545,
+      "num_tokens": 682436712.0,
+      "step": 1815
+    },
+    {
+      "epoch": 2.7023809523809526,
+      "grad_norm": 1.6645605651260449,
+      "learning_rate": 5.166191217488134e-06,
+      "loss": 0.2982,
+      "mean_token_accuracy": 0.8938940763473511,
+      "num_tokens": 682816098.0,
+      "step": 1816
+    },
+    {
+      "epoch": 2.7038690476190474,
+      "grad_norm": 1.7265154631726853,
+      "learning_rate": 5.160999566588107e-06,
+      "loss": 0.282,
+      "mean_token_accuracy": 0.9025143980979919,
+      "num_tokens": 683188826.0,
+      "step": 1817
+    },
+    {
+      "epoch": 2.705357142857143,
+      "grad_norm": 1.6706859557813205,
+      "learning_rate": 5.155807741923666e-06,
+      "loss": 0.2902,
+      "mean_token_accuracy": 0.9028702974319458,
+      "num_tokens": 683552562.0,
+      "step": 1818
+    },
+    {
+      "epoch": 2.706845238095238,
+      "grad_norm": 1.689173170705253,
+      "learning_rate": 5.15061574909827e-06,
+      "loss": 0.2862,
+      "mean_token_accuracy": 0.8999322652816772,
+      "num_tokens": 683931489.0,
+      "step": 1819
+    },
+    {
+      "epoch": 2.7083333333333335,
+      "grad_norm": 1.8194145062009615,
+      "learning_rate": 5.145423593715558e-06,
+      "loss": 0.2832,
+      "mean_token_accuracy": 0.8990654349327087,
+      "num_tokens": 684310373.0,
+      "step": 1820
+    },
+    {
+      "epoch": 2.709821428571429,
+      "grad_norm": 1.8037634385598735,
+      "learning_rate": 5.140231281379345e-06,
+      "loss": 0.281,
+      "mean_token_accuracy": 0.9053159356117249,
+      "num_tokens": 684687790.0,
+      "step": 1821
+    },
+    {
+      "epoch": 2.7113095238095237,
+      "grad_norm": 1.7949602021979707,
+      "learning_rate": 5.135038817693618e-06,
+      "loss": 0.2701,
+      "mean_token_accuracy": 0.9067232608795166,
+      "num_tokens": 685064535.0,
+      "step": 1822
+    },
+    {
+      "epoch": 2.712797619047619,
+      "grad_norm": 1.87074931406753,
+      "learning_rate": 5.129846208262522e-06,
+      "loss": 0.305,
+      "mean_token_accuracy": 0.8944651484489441,
+      "num_tokens": 685443165.0,
+      "step": 1823
+    },
+    {
+      "epoch": 2.7142857142857144,
+      "grad_norm": 1.7823400211469127,
+      "learning_rate": 5.1246534586903655e-06,
+      "loss": 0.2926,
+      "mean_token_accuracy": 0.8953758478164673,
+      "num_tokens": 685822922.0,
+      "step": 1824
+    },
+    {
+      "epoch": 2.7157738095238093,
+      "grad_norm": 1.7665629054181986,
+      "learning_rate": 5.119460574581602e-06,
+      "loss": 0.3335,
+      "mean_token_accuracy": 0.8859786987304688,
+      "num_tokens": 686203148.0,
+      "step": 1825
+    },
+    {
+      "epoch": 2.7172619047619047,
+      "grad_norm": 1.6957181592171784,
+      "learning_rate": 5.1142675615408355e-06,
+      "loss": 0.2894,
+      "mean_token_accuracy": 0.8981020450592041,
+      "num_tokens": 686582392.0,
+      "step": 1826
+    },
+    {
+      "epoch": 2.71875,
+      "grad_norm": 1.8676778746461367,
+      "learning_rate": 5.109074425172806e-06,
+      "loss": 0.3224,
+      "mean_token_accuracy": 0.8919008374214172,
+      "num_tokens": 686961505.0,
+      "step": 1827
+    },
+    {
+      "epoch": 2.7202380952380953,
+      "grad_norm": 1.7742756968347262,
+      "learning_rate": 5.103881171082389e-06,
+      "loss": 0.3236,
+      "mean_token_accuracy": 0.8879454135894775,
+      "num_tokens": 687339679.0,
+      "step": 1828
+    },
+    {
+      "epoch": 2.7217261904761907,
+      "grad_norm": 1.7983848764853576,
+      "learning_rate": 5.098687804874582e-06,
+      "loss": 0.3141,
+      "mean_token_accuracy": 0.8887469172477722,
+      "num_tokens": 687717998.0,
+      "step": 1829
+    },
+    {
+      "epoch": 2.7232142857142856,
+      "grad_norm": 1.803986609067667,
+      "learning_rate": 5.093494332154511e-06,
+      "loss": 0.3128,
+      "mean_token_accuracy": 0.8928754329681396,
+      "num_tokens": 688082384.0,
+      "step": 1830
+    },
+    {
+      "epoch": 2.724702380952381,
+      "grad_norm": 1.7569756514804362,
+      "learning_rate": 5.088300758527412e-06,
+      "loss": 0.2733,
+      "mean_token_accuracy": 0.9016363024711609,
+      "num_tokens": 688450306.0,
+      "step": 1831
+    },
+    {
+      "epoch": 2.7261904761904763,
+      "grad_norm": 1.7439920109240452,
+      "learning_rate": 5.083107089598632e-06,
+      "loss": 0.308,
+      "mean_token_accuracy": 0.891133189201355,
+      "num_tokens": 688828980.0,
+      "step": 1832
+    },
+    {
+      "epoch": 2.727678571428571,
+      "grad_norm": 1.757407975456081,
+      "learning_rate": 5.077913330973616e-06,
+      "loss": 0.3312,
+      "mean_token_accuracy": 0.8866891264915466,
+      "num_tokens": 689209140.0,
+      "step": 1833
+    },
+    {
+      "epoch": 2.7291666666666665,
+      "grad_norm": 1.741633490140426,
+      "learning_rate": 5.072719488257915e-06,
+      "loss": 0.2882,
+      "mean_token_accuracy": 0.8997660875320435,
+      "num_tokens": 689587342.0,
+      "step": 1834
+    },
+    {
+      "epoch": 2.730654761904762,
+      "grad_norm": 1.7713364552337938,
+      "learning_rate": 5.06752556705716e-06,
+      "loss": 0.3049,
+      "mean_token_accuracy": 0.8907409310340881,
+      "num_tokens": 689953214.0,
+      "step": 1835
+    },
+    {
+      "epoch": 2.732142857142857,
+      "grad_norm": 1.666536293469601,
+      "learning_rate": 5.062331572977076e-06,
+      "loss": 0.2871,
+      "mean_token_accuracy": 0.9028900861740112,
+      "num_tokens": 690332173.0,
+      "step": 1836
+    },
+    {
+      "epoch": 2.7336309523809526,
+      "grad_norm": 1.6781557573963586,
+      "learning_rate": 5.0571375116234625e-06,
+      "loss": 0.3265,
+      "mean_token_accuracy": 0.8911386132240295,
+      "num_tokens": 690711093.0,
+      "step": 1837
+    },
+    {
+      "epoch": 2.7351190476190474,
+      "grad_norm": 1.7960770445301049,
+      "learning_rate": 5.051943388602189e-06,
+      "loss": 0.3286,
+      "mean_token_accuracy": 0.8878365159034729,
+      "num_tokens": 691088392.0,
+      "step": 1838
+    },
+    {
+      "epoch": 2.736607142857143,
+      "grad_norm": 1.8866949626510123,
+      "learning_rate": 5.046749209519197e-06,
+      "loss": 0.3024,
+      "mean_token_accuracy": 0.8939533233642578,
+      "num_tokens": 691465238.0,
+      "step": 1839
+    },
+    {
+      "epoch": 2.738095238095238,
+      "grad_norm": 1.9633365709763284,
+      "learning_rate": 5.041554979980487e-06,
+      "loss": 0.2889,
+      "mean_token_accuracy": 0.8982660174369812,
+      "num_tokens": 691842081.0,
+      "step": 1840
+    },
+    {
+      "epoch": 2.7395833333333335,
+      "grad_norm": 1.8189925597902066,
+      "learning_rate": 5.03636070559211e-06,
+      "loss": 0.2722,
+      "mean_token_accuracy": 0.9033613204956055,
+      "num_tokens": 692219412.0,
+      "step": 1841
+    },
+    {
+      "epoch": 2.741071428571429,
+      "grad_norm": 1.7707453600459733,
+      "learning_rate": 5.0311663919601685e-06,
+      "loss": 0.3179,
+      "mean_token_accuracy": 0.8911228179931641,
+      "num_tokens": 692599360.0,
+      "step": 1842
+    },
+    {
+      "epoch": 2.7425595238095237,
+      "grad_norm": 1.7234848178724043,
+      "learning_rate": 5.025972044690809e-06,
+      "loss": 0.2928,
+      "mean_token_accuracy": 0.89906245470047,
+      "num_tokens": 692977932.0,
+      "step": 1843
+    },
+    {
+      "epoch": 2.744047619047619,
+      "grad_norm": 1.7961664922579572,
+      "learning_rate": 5.020777669390213e-06,
+      "loss": 0.3072,
+      "mean_token_accuracy": 0.8930897116661072,
+      "num_tokens": 693357083.0,
+      "step": 1844
+    },
+    {
+      "epoch": 2.7455357142857144,
+      "grad_norm": 1.8685263651475261,
+      "learning_rate": 5.0155832716645895e-06,
+      "loss": 0.2992,
+      "mean_token_accuracy": 0.8988752961158752,
+      "num_tokens": 693734052.0,
+      "step": 1845
+    },
+    {
+      "epoch": 2.7470238095238093,
+      "grad_norm": 1.8302963029269264,
+      "learning_rate": 5.010388857120175e-06,
+      "loss": 0.2766,
+      "mean_token_accuracy": 0.9030085802078247,
+      "num_tokens": 694110650.0,
+      "step": 1846
+    },
+    {
+      "epoch": 2.7485119047619047,
+      "grad_norm": 1.7798694184367458,
+      "learning_rate": 5.005194431363225e-06,
+      "loss": 0.3435,
+      "mean_token_accuracy": 0.8788515329360962,
+      "num_tokens": 694465902.0,
+      "step": 1847
+    },
+    {
+      "epoch": 2.75,
+      "grad_norm": 1.7983511392648406,
+      "learning_rate": 5e-06,
+      "loss": 0.2953,
+      "mean_token_accuracy": 0.895936906337738,
+      "num_tokens": 694844269.0,
+      "step": 1848
+    },
+    {
+      "epoch": 2.7514880952380953,
+      "grad_norm": 1.922417294860574,
+      "learning_rate": 4.994805568636778e-06,
+      "loss": 0.3095,
+      "mean_token_accuracy": 0.8907435536384583,
+      "num_tokens": 695222658.0,
+      "step": 1849
+    },
+    {
+      "epoch": 2.7529761904761907,
+      "grad_norm": 1.8206275842960213,
+      "learning_rate": 4.9896111428798254e-06,
+      "loss": 0.2887,
+      "mean_token_accuracy": 0.901672899723053,
+      "num_tokens": 695601357.0,
+      "step": 1850
+    },
+    {
+      "epoch": 2.7544642857142856,
+      "grad_norm": 1.883254889577856,
+      "learning_rate": 4.984416728335412e-06,
+      "loss": 0.2936,
+      "mean_token_accuracy": 0.9036615490913391,
+      "num_tokens": 695978753.0,
+      "step": 1851
+    },
+    {
+      "epoch": 2.755952380952381,
+      "grad_norm": 1.7799623285774253,
+      "learning_rate": 4.979222330609789e-06,
+      "loss": 0.319,
+      "mean_token_accuracy": 0.8872193098068237,
+      "num_tokens": 696350779.0,
+      "step": 1852
+    },
+    {
+      "epoch": 2.7574404761904763,
+      "grad_norm": 1.747696888883626,
+      "learning_rate": 4.974027955309192e-06,
+      "loss": 0.2753,
+      "mean_token_accuracy": 0.9056736826896667,
+      "num_tokens": 696728851.0,
+      "step": 1853
+    },
+    {
+      "epoch": 2.758928571428571,
+      "grad_norm": 1.7788397073180438,
+      "learning_rate": 4.968833608039832e-06,
+      "loss": 0.2745,
+      "mean_token_accuracy": 0.906212329864502,
+      "num_tokens": 697104098.0,
+      "step": 1854
+    },
+    {
+      "epoch": 2.7604166666666665,
+      "grad_norm": 1.8901158744956292,
+      "learning_rate": 4.963639294407893e-06,
+      "loss": 0.3158,
+      "mean_token_accuracy": 0.8917267322540283,
+      "num_tokens": 697481551.0,
+      "step": 1855
+    },
+    {
+      "epoch": 2.761904761904762,
+      "grad_norm": 1.8087205531219168,
+      "learning_rate": 4.958445020019516e-06,
+      "loss": 0.3154,
+      "mean_token_accuracy": 0.8935686945915222,
+      "num_tokens": 697859490.0,
+      "step": 1856
+    },
+    {
+      "epoch": 2.763392857142857,
+      "grad_norm": 1.8641035324818376,
+      "learning_rate": 4.953250790480805e-06,
+      "loss": 0.2699,
+      "mean_token_accuracy": 0.9062195420265198,
+      "num_tokens": 698236017.0,
+      "step": 1857
+    },
+    {
+      "epoch": 2.7648809523809526,
+      "grad_norm": 1.786865571913578,
+      "learning_rate": 4.948056611397814e-06,
+      "loss": 0.2949,
+      "mean_token_accuracy": 0.8955920338630676,
+      "num_tokens": 698613759.0,
+      "step": 1858
+    },
+    {
+      "epoch": 2.7663690476190474,
+      "grad_norm": 1.8001153048722565,
+      "learning_rate": 4.942862488376539e-06,
+      "loss": 0.2858,
+      "mean_token_accuracy": 0.9012506008148193,
+      "num_tokens": 698990818.0,
+      "step": 1859
+    },
+    {
+      "epoch": 2.767857142857143,
+      "grad_norm": 1.7113209287631834,
+      "learning_rate": 4.937668427022925e-06,
+      "loss": 0.266,
+      "mean_token_accuracy": 0.9028113484382629,
+      "num_tokens": 699367808.0,
+      "step": 1860
+    },
+    {
+      "epoch": 2.769345238095238,
+      "grad_norm": 1.7281172272806602,
+      "learning_rate": 4.932474432942841e-06,
+      "loss": 0.3086,
+      "mean_token_accuracy": 0.8953900933265686,
+      "num_tokens": 699748005.0,
+      "step": 1861
+    },
+    {
+      "epoch": 2.7708333333333335,
+      "grad_norm": 1.8109074566779475,
+      "learning_rate": 4.927280511742087e-06,
+      "loss": 0.2819,
+      "mean_token_accuracy": 0.9045467376708984,
+      "num_tokens": 700112564.0,
+      "step": 1862
+    },
+    {
+      "epoch": 2.772321428571429,
+      "grad_norm": 1.8181184734932219,
+      "learning_rate": 4.922086669026386e-06,
+      "loss": 0.3074,
+      "mean_token_accuracy": 0.8864282369613647,
+      "num_tokens": 700490944.0,
+      "step": 1863
+    },
+    {
+      "epoch": 2.7738095238095237,
+      "grad_norm": 1.8822812398820215,
+      "learning_rate": 4.916892910401369e-06,
+      "loss": 0.3423,
+      "mean_token_accuracy": 0.885728657245636,
+      "num_tokens": 700869551.0,
+      "step": 1864
+    },
+    {
+      "epoch": 2.775297619047619,
+      "grad_norm": 1.6991336435286408,
+      "learning_rate": 4.9116992414725885e-06,
+      "loss": 0.2961,
+      "mean_token_accuracy": 0.892196774482727,
+      "num_tokens": 701249343.0,
+      "step": 1865
+    },
+    {
+      "epoch": 2.7767857142857144,
+      "grad_norm": 1.7068571667604133,
+      "learning_rate": 4.90650566784549e-06,
+      "loss": 0.2874,
+      "mean_token_accuracy": 0.9004318118095398,
+      "num_tokens": 701628439.0,
+      "step": 1866
+    },
+    {
+      "epoch": 2.7782738095238093,
+      "grad_norm": 1.8159512141445961,
+      "learning_rate": 4.901312195125418e-06,
+      "loss": 0.2986,
+      "mean_token_accuracy": 0.8997336626052856,
+      "num_tokens": 702005140.0,
+      "step": 1867
+    },
+    {
+      "epoch": 2.7797619047619047,
+      "grad_norm": 1.8522615301502425,
+      "learning_rate": 4.8961188289176125e-06,
+      "loss": 0.3197,
+      "mean_token_accuracy": 0.8896350860595703,
+      "num_tokens": 702379716.0,
+      "step": 1868
+    },
+    {
+      "epoch": 2.78125,
+      "grad_norm": 1.8532823318597038,
+      "learning_rate": 4.890925574827195e-06,
+      "loss": 0.3184,
+      "mean_token_accuracy": 0.8926246166229248,
+      "num_tokens": 702751892.0,
+      "step": 1869
+    },
+    {
+      "epoch": 2.7827380952380953,
+      "grad_norm": 1.904524573991097,
+      "learning_rate": 4.885732438459165e-06,
+      "loss": 0.3103,
+      "mean_token_accuracy": 0.8917756676673889,
+      "num_tokens": 703129391.0,
+      "step": 1870
+    },
+    {
+      "epoch": 2.7842261904761907,
+      "grad_norm": 1.739889848994138,
+      "learning_rate": 4.880539425418399e-06,
+      "loss": 0.3058,
+      "mean_token_accuracy": 0.887668788433075,
+      "num_tokens": 703509346.0,
+      "step": 1871
+    },
+    {
+      "epoch": 2.7857142857142856,
+      "grad_norm": 1.8551483365878272,
+      "learning_rate": 4.875346541309637e-06,
+      "loss": 0.3124,
+      "mean_token_accuracy": 0.8944531679153442,
+      "num_tokens": 703887625.0,
+      "step": 1872
+    },
+    {
+      "epoch": 2.787202380952381,
+      "grad_norm": 1.8509960107650718,
+      "learning_rate": 4.870153791737478e-06,
+      "loss": 0.2744,
+      "mean_token_accuracy": 0.9021085500717163,
+      "num_tokens": 704252292.0,
+      "step": 1873
+    },
+    {
+      "epoch": 2.7886904761904763,
+      "grad_norm": 1.686042854620599,
+      "learning_rate": 4.864961182306383e-06,
+      "loss": 0.2888,
+      "mean_token_accuracy": 0.8977165222167969,
+      "num_tokens": 704617929.0,
+      "step": 1874
+    },
+    {
+      "epoch": 2.790178571428571,
+      "grad_norm": 1.9264556277912914,
+      "learning_rate": 4.859768718620656e-06,
+      "loss": 0.256,
+      "mean_token_accuracy": 0.9081382751464844,
+      "num_tokens": 704982777.0,
+      "step": 1875
+    },
+    {
+      "epoch": 2.7916666666666665,
+      "grad_norm": 1.8496492968957374,
+      "learning_rate": 4.854576406284443e-06,
+      "loss": 0.2997,
+      "mean_token_accuracy": 0.8983615636825562,
+      "num_tokens": 705360471.0,
+      "step": 1876
+    },
+    {
+      "epoch": 2.793154761904762,
+      "grad_norm": 1.8092016717523074,
+      "learning_rate": 4.849384250901732e-06,
+      "loss": 0.2817,
+      "mean_token_accuracy": 0.9004379510879517,
+      "num_tokens": 705737776.0,
+      "step": 1877
+    },
+    {
+      "epoch": 2.794642857142857,
+      "grad_norm": 1.8090147025872705,
+      "learning_rate": 4.844192258076335e-06,
+      "loss": 0.2715,
+      "mean_token_accuracy": 0.90294349193573,
+      "num_tokens": 706114112.0,
+      "step": 1878
+    },
+    {
+      "epoch": 2.7961309523809526,
+      "grad_norm": 1.7394711745984002,
+      "learning_rate": 4.839000433411894e-06,
+      "loss": 0.2599,
+      "mean_token_accuracy": 0.9115842580795288,
+      "num_tokens": 706492132.0,
+      "step": 1879
+    },
+    {
+      "epoch": 2.7976190476190474,
+      "grad_norm": 1.8804553408567048,
+      "learning_rate": 4.833808782511867e-06,
+      "loss": 0.331,
+      "mean_token_accuracy": 0.8902726173400879,
+      "num_tokens": 706870888.0,
+      "step": 1880
+    },
+    {
+      "epoch": 2.799107142857143,
+      "grad_norm": 1.7507290271078557,
+      "learning_rate": 4.828617310979525e-06,
+      "loss": 0.2993,
+      "mean_token_accuracy": 0.8890262842178345,
+      "num_tokens": 707232654.0,
+      "step": 1881
+    },
+    {
+      "epoch": 2.800595238095238,
+      "grad_norm": 1.7520917090230006,
+      "learning_rate": 4.823426024417943e-06,
+      "loss": 0.272,
+      "mean_token_accuracy": 0.9023929238319397,
+      "num_tokens": 707610595.0,
+      "step": 1882
+    },
+    {
+      "epoch": 2.8020833333333335,
+      "grad_norm": 1.9187642281410215,
+      "learning_rate": 4.818234928430003e-06,
+      "loss": 0.3257,
+      "mean_token_accuracy": 0.8861500024795532,
+      "num_tokens": 707987967.0,
+      "step": 1883
+    },
+    {
+      "epoch": 2.803571428571429,
+      "grad_norm": 1.8398127622822968,
+      "learning_rate": 4.8130440286183725e-06,
+      "loss": 0.2722,
+      "mean_token_accuracy": 0.9042415022850037,
+      "num_tokens": 708362652.0,
+      "step": 1884
+    },
+    {
+      "epoch": 2.8050595238095237,
+      "grad_norm": 2.0511192753727068,
+      "learning_rate": 4.807853330585514e-06,
+      "loss": 0.313,
+      "mean_token_accuracy": 0.8957384824752808,
+      "num_tokens": 708739836.0,
+      "step": 1885
+    },
+    {
+      "epoch": 2.806547619047619,
+      "grad_norm": 1.7208636273166833,
+      "learning_rate": 4.80266283993367e-06,
+      "loss": 0.2918,
+      "mean_token_accuracy": 0.8947650194168091,
+      "num_tokens": 709106103.0,
+      "step": 1886
+    },
+    {
+      "epoch": 2.8080357142857144,
+      "grad_norm": 1.861515317654157,
+      "learning_rate": 4.797472562264859e-06,
+      "loss": 0.283,
+      "mean_token_accuracy": 0.8996806144714355,
+      "num_tokens": 709470510.0,
+      "step": 1887
+    },
+    {
+      "epoch": 2.8095238095238093,
+      "grad_norm": 1.7698440392546955,
+      "learning_rate": 4.792282503180867e-06,
+      "loss": 0.3143,
+      "mean_token_accuracy": 0.8905964493751526,
+      "num_tokens": 709848382.0,
+      "step": 1888
+    },
+    {
+      "epoch": 2.8110119047619047,
+      "grad_norm": 1.7029282778505481,
+      "learning_rate": 4.787092668283252e-06,
+      "loss": 0.2611,
+      "mean_token_accuracy": 0.9113073348999023,
+      "num_tokens": 710225777.0,
+      "step": 1889
+    },
+    {
+      "epoch": 2.8125,
+      "grad_norm": 1.733263807156413,
+      "learning_rate": 4.781903063173321e-06,
+      "loss": 0.2928,
+      "mean_token_accuracy": 0.8948770761489868,
+      "num_tokens": 710604948.0,
+      "step": 1890
+    },
+    {
+      "epoch": 2.8139880952380953,
+      "grad_norm": 1.831623661485758,
+      "learning_rate": 4.776713693452137e-06,
+      "loss": 0.3021,
+      "mean_token_accuracy": 0.8956916332244873,
+      "num_tokens": 710971366.0,
+      "step": 1891
+    },
+    {
+      "epoch": 2.8154761904761907,
+      "grad_norm": 1.8156423332738139,
+      "learning_rate": 4.7715245647205105e-06,
+      "loss": 0.2929,
+      "mean_token_accuracy": 0.8992076516151428,
+      "num_tokens": 711350111.0,
+      "step": 1892
+    },
+    {
+      "epoch": 2.8169642857142856,
+      "grad_norm": 1.7949995139976112,
+      "learning_rate": 4.7663356825789894e-06,
+      "loss": 0.3295,
+      "mean_token_accuracy": 0.8877202272415161,
+      "num_tokens": 711728238.0,
+      "step": 1893
+    },
+    {
+      "epoch": 2.818452380952381,
+      "grad_norm": 1.9165587175156926,
+      "learning_rate": 4.761147052627856e-06,
+      "loss": 0.3329,
+      "mean_token_accuracy": 0.8891113996505737,
+      "num_tokens": 712107014.0,
+      "step": 1894
+    },
+    {
+      "epoch": 2.8199404761904763,
+      "grad_norm": 1.7511972965469764,
+      "learning_rate": 4.755958680467121e-06,
+      "loss": 0.2653,
+      "mean_token_accuracy": 0.9082791209220886,
+      "num_tokens": 712485658.0,
+      "step": 1895
+    },
+    {
+      "epoch": 2.821428571428571,
+      "grad_norm": 1.8782112285932824,
+      "learning_rate": 4.750770571696514e-06,
+      "loss": 0.3136,
+      "mean_token_accuracy": 0.89570552110672,
+      "num_tokens": 712844668.0,
+      "step": 1896
+    },
+    {
+      "epoch": 2.8229166666666665,
+      "grad_norm": 1.7889366270647997,
+      "learning_rate": 4.745582731915488e-06,
+      "loss": 0.2907,
+      "mean_token_accuracy": 0.8944143056869507,
+      "num_tokens": 713222589.0,
+      "step": 1897
+    },
+    {
+      "epoch": 2.824404761904762,
+      "grad_norm": 1.721118794173896,
+      "learning_rate": 4.740395166723198e-06,
+      "loss": 0.2835,
+      "mean_token_accuracy": 0.9036904573440552,
+      "num_tokens": 713601545.0,
+      "step": 1898
+    },
+    {
+      "epoch": 2.825892857142857,
+      "grad_norm": 1.9104822163545925,
+      "learning_rate": 4.735207881718502e-06,
+      "loss": 0.3047,
+      "mean_token_accuracy": 0.8945749402046204,
+      "num_tokens": 713978904.0,
+      "step": 1899
+    },
+    {
+      "epoch": 2.8273809523809526,
+      "grad_norm": 1.866351099556449,
+      "learning_rate": 4.730020882499964e-06,
+      "loss": 0.2753,
+      "mean_token_accuracy": 0.902582585811615,
+      "num_tokens": 714356402.0,
+      "step": 1900
+    },
+    {
+      "epoch": 2.8288690476190474,
+      "grad_norm": 1.7928026742590188,
+      "learning_rate": 4.724834174665833e-06,
+      "loss": 0.2616,
+      "mean_token_accuracy": 0.9082840085029602,
+      "num_tokens": 714734066.0,
+      "step": 1901
+    },
+    {
+      "epoch": 2.830357142857143,
+      "grad_norm": 1.858909397213124,
+      "learning_rate": 4.719647763814041e-06,
+      "loss": 0.3352,
+      "mean_token_accuracy": 0.8862828016281128,
+      "num_tokens": 715099036.0,
+      "step": 1902
+    },
+    {
+      "epoch": 2.831845238095238,
+      "grad_norm": 1.8692515007085855,
+      "learning_rate": 4.714461655542208e-06,
+      "loss": 0.3122,
+      "mean_token_accuracy": 0.8911147713661194,
+      "num_tokens": 715477892.0,
+      "step": 1903
+    },
+    {
+      "epoch": 2.8333333333333335,
+      "grad_norm": 1.9408401297874334,
+      "learning_rate": 4.7092758554476215e-06,
+      "loss": 0.3338,
+      "mean_token_accuracy": 0.8844146728515625,
+      "num_tokens": 715856612.0,
+      "step": 1904
+    },
+    {
+      "epoch": 2.834821428571429,
+      "grad_norm": 1.8156436034551073,
+      "learning_rate": 4.704090369127236e-06,
+      "loss": 0.3134,
+      "mean_token_accuracy": 0.8909181952476501,
+      "num_tokens": 716231696.0,
+      "step": 1905
+    },
+    {
+      "epoch": 2.8363095238095237,
+      "grad_norm": 1.9544748937177479,
+      "learning_rate": 4.698905202177671e-06,
+      "loss": 0.3061,
+      "mean_token_accuracy": 0.8930070996284485,
+      "num_tokens": 716609651.0,
+      "step": 1906
+    },
+    {
+      "epoch": 2.837797619047619,
+      "grad_norm": 1.8195160127021823,
+      "learning_rate": 4.693720360195199e-06,
+      "loss": 0.2973,
+      "mean_token_accuracy": 0.8932477831840515,
+      "num_tokens": 716987190.0,
+      "step": 1907
+    },
+    {
+      "epoch": 2.8392857142857144,
+      "grad_norm": 1.7620156551947106,
+      "learning_rate": 4.6885358487757396e-06,
+      "loss": 0.2765,
+      "mean_token_accuracy": 0.9029448628425598,
+      "num_tokens": 717364727.0,
+      "step": 1908
+    },
+    {
+      "epoch": 2.8407738095238093,
+      "grad_norm": 1.8186106004705695,
+      "learning_rate": 4.6833516735148625e-06,
+      "loss": 0.3157,
+      "mean_token_accuracy": 0.8925612568855286,
+      "num_tokens": 717730876.0,
+      "step": 1909
+    },
+    {
+      "epoch": 2.8422619047619047,
+      "grad_norm": 1.8274132791418454,
+      "learning_rate": 4.678167840007767e-06,
+      "loss": 0.272,
+      "mean_token_accuracy": 0.9053066968917847,
+      "num_tokens": 718108423.0,
+      "step": 1910
+    },
+    {
+      "epoch": 2.84375,
+      "grad_norm": 1.768457934120635,
+      "learning_rate": 4.672984353849285e-06,
+      "loss": 0.2712,
+      "mean_token_accuracy": 0.9059233665466309,
+      "num_tokens": 718484441.0,
+      "step": 1911
+    },
+    {
+      "epoch": 2.8452380952380953,
+      "grad_norm": 1.763797648049534,
+      "learning_rate": 4.66780122063388e-06,
+      "loss": 0.3143,
+      "mean_token_accuracy": 0.8898670673370361,
+      "num_tokens": 718863336.0,
+      "step": 1912
+    },
+    {
+      "epoch": 2.8467261904761907,
+      "grad_norm": 1.7903824710532417,
+      "learning_rate": 4.662618445955628e-06,
+      "loss": 0.286,
+      "mean_token_accuracy": 0.8966982364654541,
+      "num_tokens": 719210192.0,
+      "step": 1913
+    },
+    {
+      "epoch": 2.8482142857142856,
+      "grad_norm": 1.6080964191946114,
+      "learning_rate": 4.657436035408217e-06,
+      "loss": 0.2733,
+      "mean_token_accuracy": 0.9042270183563232,
+      "num_tokens": 719589662.0,
+      "step": 1914
+    },
+    {
+      "epoch": 2.849702380952381,
+      "grad_norm": 1.8231538128357967,
+      "learning_rate": 4.652253994584948e-06,
+      "loss": 0.2871,
+      "mean_token_accuracy": 0.8945656418800354,
+      "num_tokens": 719966635.0,
+      "step": 1915
+    },
+    {
+      "epoch": 2.8511904761904763,
+      "grad_norm": 1.860055819565278,
+      "learning_rate": 4.647072329078721e-06,
+      "loss": 0.3066,
+      "mean_token_accuracy": 0.8944454193115234,
+      "num_tokens": 720330073.0,
+      "step": 1916
+    },
+    {
+      "epoch": 2.852678571428571,
+      "grad_norm": 1.8560566526835716,
+      "learning_rate": 4.641891044482025e-06,
+      "loss": 0.3002,
+      "mean_token_accuracy": 0.8951916694641113,
+      "num_tokens": 720707169.0,
+      "step": 1917
+    },
+    {
+      "epoch": 2.8541666666666665,
+      "grad_norm": 1.8058202474569578,
+      "learning_rate": 4.636710146386948e-06,
+      "loss": 0.2473,
+      "mean_token_accuracy": 0.9108901023864746,
+      "num_tokens": 721083966.0,
+      "step": 1918
+    },
+    {
+      "epoch": 2.855654761904762,
+      "grad_norm": 1.7512709756063127,
+      "learning_rate": 4.631529640385151e-06,
+      "loss": 0.2947,
+      "mean_token_accuracy": 0.9000921845436096,
+      "num_tokens": 721461902.0,
+      "step": 1919
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 1.6962635384590525,
+      "learning_rate": 4.626349532067879e-06,
+      "loss": 0.2909,
+      "mean_token_accuracy": 0.9014671444892883,
+      "num_tokens": 721830036.0,
+      "step": 1920
+    },
+    {
+      "epoch": 2.8586309523809526,
+      "grad_norm": 1.8189714616621582,
+      "learning_rate": 4.6211698270259444e-06,
+      "loss": 0.2893,
+      "mean_token_accuracy": 0.8971878290176392,
+      "num_tokens": 722207320.0,
+      "step": 1921
+    },
+    {
+      "epoch": 2.8601190476190474,
+      "grad_norm": 1.659766448005871,
+      "learning_rate": 4.615990530849726e-06,
+      "loss": 0.288,
+      "mean_token_accuracy": 0.9008640646934509,
+      "num_tokens": 722587973.0,
+      "step": 1922
+    },
+    {
+      "epoch": 2.861607142857143,
+      "grad_norm": 1.6545682099427206,
+      "learning_rate": 4.6108116491291575e-06,
+      "loss": 0.2733,
+      "mean_token_accuracy": 0.9046619534492493,
+      "num_tokens": 722966483.0,
+      "step": 1923
+    },
+    {
+      "epoch": 2.863095238095238,
+      "grad_norm": 1.7259167677940046,
+      "learning_rate": 4.605633187453732e-06,
+      "loss": 0.295,
+      "mean_token_accuracy": 0.8974402546882629,
+      "num_tokens": 723344687.0,
+      "step": 1924
+    },
+    {
+      "epoch": 2.8645833333333335,
+      "grad_norm": 1.720440794518523,
+      "learning_rate": 4.600455151412482e-06,
+      "loss": 0.2853,
+      "mean_token_accuracy": 0.9008135795593262,
+      "num_tokens": 723723455.0,
+      "step": 1925
+    },
+    {
+      "epoch": 2.866071428571429,
+      "grad_norm": 1.6849295539537392,
+      "learning_rate": 4.595277546593984e-06,
+      "loss": 0.2535,
+      "mean_token_accuracy": 0.9076992869377136,
+      "num_tokens": 724101349.0,
+      "step": 1926
+    },
+    {
+      "epoch": 2.8675595238095237,
+      "grad_norm": 1.647213981831955,
+      "learning_rate": 4.590100378586352e-06,
+      "loss": 0.2883,
+      "mean_token_accuracy": 0.8995354175567627,
+      "num_tokens": 724480869.0,
+      "step": 1927
+    },
+    {
+      "epoch": 2.869047619047619,
+      "grad_norm": 1.9031209296169793,
+      "learning_rate": 4.584923652977224e-06,
+      "loss": 0.3078,
+      "mean_token_accuracy": 0.8912495374679565,
+      "num_tokens": 724858526.0,
+      "step": 1928
+    },
+    {
+      "epoch": 2.8705357142857144,
+      "grad_norm": 1.7354254726896905,
+      "learning_rate": 4.579747375353763e-06,
+      "loss": 0.2802,
+      "mean_token_accuracy": 0.9034203290939331,
+      "num_tokens": 725236260.0,
+      "step": 1929
+    },
+    {
+      "epoch": 2.8720238095238093,
+      "grad_norm": 1.7774044031276965,
+      "learning_rate": 4.5745715513026474e-06,
+      "loss": 0.3003,
+      "mean_token_accuracy": 0.8947597742080688,
+      "num_tokens": 725614613.0,
+      "step": 1930
+    },
+    {
+      "epoch": 2.8735119047619047,
+      "grad_norm": 1.7271192819612078,
+      "learning_rate": 4.569396186410063e-06,
+      "loss": 0.2667,
+      "mean_token_accuracy": 0.9057944416999817,
+      "num_tokens": 725981070.0,
+      "step": 1931
+    },
+    {
+      "epoch": 2.875,
+      "grad_norm": 1.9054420492061368,
+      "learning_rate": 4.564221286261709e-06,
+      "loss": 0.2878,
+      "mean_token_accuracy": 0.9036948084831238,
+      "num_tokens": 726359172.0,
+      "step": 1932
+    },
+    {
+      "epoch": 2.8764880952380953,
+      "grad_norm": 1.8184286638785758,
+      "learning_rate": 4.559046856442775e-06,
+      "loss": 0.32,
+      "mean_token_accuracy": 0.8893020749092102,
+      "num_tokens": 726737186.0,
+      "step": 1933
+    },
+    {
+      "epoch": 2.8779761904761907,
+      "grad_norm": 1.6785692966388417,
+      "learning_rate": 4.553872902537944e-06,
+      "loss": 0.2599,
+      "mean_token_accuracy": 0.9107905030250549,
+      "num_tokens": 727114451.0,
+      "step": 1934
+    },
+    {
+      "epoch": 2.8794642857142856,
+      "grad_norm": 1.8248039350513916,
+      "learning_rate": 4.548699430131388e-06,
+      "loss": 0.2898,
+      "mean_token_accuracy": 0.9016985893249512,
+      "num_tokens": 727491450.0,
+      "step": 1935
+    },
+    {
+      "epoch": 2.880952380952381,
+      "grad_norm": 1.6324631660333775,
+      "learning_rate": 4.5435264448067595e-06,
+      "loss": 0.2711,
+      "mean_token_accuracy": 0.90768963098526,
+      "num_tokens": 727857878.0,
+      "step": 1936
+    },
+    {
+      "epoch": 2.8824404761904763,
+      "grad_norm": 1.7659759708888025,
+      "learning_rate": 4.538353952147181e-06,
+      "loss": 0.2712,
+      "mean_token_accuracy": 0.9098491668701172,
+      "num_tokens": 728236314.0,
+      "step": 1937
+    },
+    {
+      "epoch": 2.883928571428571,
+      "grad_norm": 1.766872165593601,
+      "learning_rate": 4.533181957735247e-06,
+      "loss": 0.2619,
+      "mean_token_accuracy": 0.9051488041877747,
+      "num_tokens": 728613930.0,
+      "step": 1938
+    },
+    {
+      "epoch": 2.8854166666666665,
+      "grad_norm": 1.8276922619479636,
+      "learning_rate": 4.528010467153015e-06,
+      "loss": 0.3153,
+      "mean_token_accuracy": 0.8894700407981873,
+      "num_tokens": 728992492.0,
+      "step": 1939
+    },
+    {
+      "epoch": 2.886904761904762,
+      "grad_norm": 1.9747509309239126,
+      "learning_rate": 4.522839485981994e-06,
+      "loss": 0.3128,
+      "mean_token_accuracy": 0.8932432532310486,
+      "num_tokens": 729369993.0,
+      "step": 1940
+    },
+    {
+      "epoch": 2.888392857142857,
+      "grad_norm": 1.7821210226875468,
+      "learning_rate": 4.51766901980315e-06,
+      "loss": 0.2907,
+      "mean_token_accuracy": 0.8967875242233276,
+      "num_tokens": 729737188.0,
+      "step": 1941
+    },
+    {
+      "epoch": 2.8898809523809526,
+      "grad_norm": 1.8252146962409774,
+      "learning_rate": 4.512499074196888e-06,
+      "loss": 0.3255,
+      "mean_token_accuracy": 0.8887887001037598,
+      "num_tokens": 730116018.0,
+      "step": 1942
+    },
+    {
+      "epoch": 2.8913690476190474,
+      "grad_norm": 1.7473587640626254,
+      "learning_rate": 4.5073296547430506e-06,
+      "loss": 0.2837,
+      "mean_token_accuracy": 0.8972108960151672,
+      "num_tokens": 730493605.0,
+      "step": 1943
+    },
+    {
+      "epoch": 2.892857142857143,
+      "grad_norm": 1.851443621389755,
+      "learning_rate": 4.502160767020918e-06,
+      "loss": 0.3051,
+      "mean_token_accuracy": 0.8890276551246643,
+      "num_tokens": 730865982.0,
+      "step": 1944
+    },
+    {
+      "epoch": 2.894345238095238,
+      "grad_norm": 1.6272624408230791,
+      "learning_rate": 4.49699241660919e-06,
+      "loss": 0.2549,
+      "mean_token_accuracy": 0.9085449576377869,
+      "num_tokens": 731244246.0,
+      "step": 1945
+    },
+    {
+      "epoch": 2.8958333333333335,
+      "grad_norm": 1.8709343457539427,
+      "learning_rate": 4.4918246090859905e-06,
+      "loss": 0.3088,
+      "mean_token_accuracy": 0.8943824172019958,
+      "num_tokens": 731622247.0,
+      "step": 1946
+    },
+    {
+      "epoch": 2.897321428571429,
+      "grad_norm": 1.7424951034234446,
+      "learning_rate": 4.486657350028859e-06,
+      "loss": 0.302,
+      "mean_token_accuracy": 0.8948343992233276,
+      "num_tokens": 732001109.0,
+      "step": 1947
+    },
+    {
+      "epoch": 2.8988095238095237,
+      "grad_norm": 1.8825566227386992,
+      "learning_rate": 4.4814906450147376e-06,
+      "loss": 0.2668,
+      "mean_token_accuracy": 0.9060908555984497,
+      "num_tokens": 732360931.0,
+      "step": 1948
+    },
+    {
+      "epoch": 2.900297619047619,
+      "grad_norm": 1.812320782801374,
+      "learning_rate": 4.476324499619974e-06,
+      "loss": 0.3154,
+      "mean_token_accuracy": 0.8868579268455505,
+      "num_tokens": 732739398.0,
+      "step": 1949
+    },
+    {
+      "epoch": 2.9017857142857144,
+      "grad_norm": 1.7200152269761566,
+      "learning_rate": 4.471158919420312e-06,
+      "loss": 0.3043,
+      "mean_token_accuracy": 0.8966609835624695,
+      "num_tokens": 733119010.0,
+      "step": 1950
+    },
+    {
+      "epoch": 2.9032738095238093,
+      "grad_norm": 1.9023828130132856,
+      "learning_rate": 4.4659939099908845e-06,
+      "loss": 0.3138,
+      "mean_token_accuracy": 0.8891192674636841,
+      "num_tokens": 733496885.0,
+      "step": 1951
+    },
+    {
+      "epoch": 2.9047619047619047,
+      "grad_norm": 1.7111902705433992,
+      "learning_rate": 4.460829476906208e-06,
+      "loss": 0.2811,
+      "mean_token_accuracy": 0.9018703699111938,
+      "num_tokens": 733866840.0,
+      "step": 1952
+    },
+    {
+      "epoch": 2.90625,
+      "grad_norm": 1.5930454559847957,
+      "learning_rate": 4.4556656257401786e-06,
+      "loss": 0.2792,
+      "mean_token_accuracy": 0.9049389958381653,
+      "num_tokens": 734246548.0,
+      "step": 1953
+    },
+    {
+      "epoch": 2.9077380952380953,
+      "grad_norm": 1.7562449413698926,
+      "learning_rate": 4.4505023620660625e-06,
+      "loss": 0.286,
+      "mean_token_accuracy": 0.8983035683631897,
+      "num_tokens": 734625026.0,
+      "step": 1954
+    },
+    {
+      "epoch": 2.9092261904761907,
+      "grad_norm": 1.8605346945598484,
+      "learning_rate": 4.445339691456491e-06,
+      "loss": 0.2998,
+      "mean_token_accuracy": 0.8965129256248474,
+      "num_tokens": 735003755.0,
+      "step": 1955
+    },
+    {
+      "epoch": 2.9107142857142856,
+      "grad_norm": 1.9303367947882384,
+      "learning_rate": 4.4401776194834615e-06,
+      "loss": 0.2702,
+      "mean_token_accuracy": 0.9069682955741882,
+      "num_tokens": 735381946.0,
+      "step": 1956
+    },
+    {
+      "epoch": 2.912202380952381,
+      "grad_norm": 1.7715415839999755,
+      "learning_rate": 4.4350161517183175e-06,
+      "loss": 0.2843,
+      "mean_token_accuracy": 0.8971412777900696,
+      "num_tokens": 735759208.0,
+      "step": 1957
+    },
+    {
+      "epoch": 2.9136904761904763,
+      "grad_norm": 1.7134539573825902,
+      "learning_rate": 4.429855293731753e-06,
+      "loss": 0.292,
+      "mean_token_accuracy": 0.9004191160202026,
+      "num_tokens": 736138587.0,
+      "step": 1958
+    },
+    {
+      "epoch": 2.915178571428571,
+      "grad_norm": 1.7439709202050009,
+      "learning_rate": 4.424695051093807e-06,
+      "loss": 0.2632,
+      "mean_token_accuracy": 0.9081304669380188,
+      "num_tokens": 736516039.0,
+      "step": 1959
+    },
+    {
+      "epoch": 2.9166666666666665,
+      "grad_norm": 1.793316715644253,
+      "learning_rate": 4.4195354293738484e-06,
+      "loss": 0.34,
+      "mean_token_accuracy": 0.8836826682090759,
+      "num_tokens": 736894927.0,
+      "step": 1960
+    },
+    {
+      "epoch": 2.918154761904762,
+      "grad_norm": 1.9372051531831838,
+      "learning_rate": 4.414376434140584e-06,
+      "loss": 0.3317,
+      "mean_token_accuracy": 0.8841872811317444,
+      "num_tokens": 737269681.0,
+      "step": 1961
+    },
+    {
+      "epoch": 2.919642857142857,
+      "grad_norm": 1.9399972505101666,
+      "learning_rate": 4.409218070962036e-06,
+      "loss": 0.2868,
+      "mean_token_accuracy": 0.9015487432479858,
+      "num_tokens": 737646379.0,
+      "step": 1962
+    },
+    {
+      "epoch": 2.9211309523809526,
+      "grad_norm": 1.7774957092024564,
+      "learning_rate": 4.40406034540555e-06,
+      "loss": 0.2959,
+      "mean_token_accuracy": 0.897647500038147,
+      "num_tokens": 738023444.0,
+      "step": 1963
+    },
+    {
+      "epoch": 2.9226190476190474,
+      "grad_norm": 1.8061092167746082,
+      "learning_rate": 4.398903263037783e-06,
+      "loss": 0.2925,
+      "mean_token_accuracy": 0.9005966186523438,
+      "num_tokens": 738380013.0,
+      "step": 1964
+    },
+    {
+      "epoch": 2.924107142857143,
+      "grad_norm": 1.735824974106,
+      "learning_rate": 4.393746829424693e-06,
+      "loss": 0.2578,
+      "mean_token_accuracy": 0.9104001522064209,
+      "num_tokens": 738758024.0,
+      "step": 1965
+    },
+    {
+      "epoch": 2.925595238095238,
+      "grad_norm": 1.7297490855994402,
+      "learning_rate": 4.3885910501315435e-06,
+      "loss": 0.2892,
+      "mean_token_accuracy": 0.8950415849685669,
+      "num_tokens": 739139089.0,
+      "step": 1966
+    },
+    {
+      "epoch": 2.9270833333333335,
+      "grad_norm": 1.7630655176365382,
+      "learning_rate": 4.38343593072289e-06,
+      "loss": 0.3204,
+      "mean_token_accuracy": 0.8859832882881165,
+      "num_tokens": 739518422.0,
+      "step": 1967
+    },
+    {
+      "epoch": 2.928571428571429,
+      "grad_norm": 1.8274577373023382,
+      "learning_rate": 4.3782814767625755e-06,
+      "loss": 0.2662,
+      "mean_token_accuracy": 0.9055770039558411,
+      "num_tokens": 739895622.0,
+      "step": 1968
+    },
+    {
+      "epoch": 2.9300595238095237,
+      "grad_norm": 1.7123714381233859,
+      "learning_rate": 4.373127693813722e-06,
+      "loss": 0.2993,
+      "mean_token_accuracy": 0.8909000158309937,
+      "num_tokens": 740274923.0,
+      "step": 1969
+    },
+    {
+      "epoch": 2.931547619047619,
+      "grad_norm": 1.859541719501866,
+      "learning_rate": 4.367974587438733e-06,
+      "loss": 0.3396,
+      "mean_token_accuracy": 0.8900694847106934,
+      "num_tokens": 740653915.0,
+      "step": 1970
+    },
+    {
+      "epoch": 2.9330357142857144,
+      "grad_norm": 1.7176367217277497,
+      "learning_rate": 4.36282216319928e-06,
+      "loss": 0.294,
+      "mean_token_accuracy": 0.902297854423523,
+      "num_tokens": 741032594.0,
+      "step": 1971
+    },
+    {
+      "epoch": 2.9345238095238093,
+      "grad_norm": 1.8146469624729664,
+      "learning_rate": 4.357670426656293e-06,
+      "loss": 0.2659,
+      "mean_token_accuracy": 0.9061084985733032,
+      "num_tokens": 741409996.0,
+      "step": 1972
+    },
+    {
+      "epoch": 2.9360119047619047,
+      "grad_norm": 1.7339762344405105,
+      "learning_rate": 4.352519383369966e-06,
+      "loss": 0.292,
+      "mean_token_accuracy": 0.897413432598114,
+      "num_tokens": 741787974.0,
+      "step": 1973
+    },
+    {
+      "epoch": 2.9375,
+      "grad_norm": 1.6532233717054063,
+      "learning_rate": 4.347369038899744e-06,
+      "loss": 0.3008,
+      "mean_token_accuracy": 0.8907099366188049,
+      "num_tokens": 742167097.0,
+      "step": 1974
+    },
+    {
+      "epoch": 2.9389880952380953,
+      "grad_norm": 1.8785739740849152,
+      "learning_rate": 4.342219398804311e-06,
+      "loss": 0.3147,
+      "mean_token_accuracy": 0.8916872143745422,
+      "num_tokens": 742544768.0,
+      "step": 1975
+    },
+    {
+      "epoch": 2.9404761904761907,
+      "grad_norm": 1.7624506798073833,
+      "learning_rate": 4.337070468641604e-06,
+      "loss": 0.2781,
+      "mean_token_accuracy": 0.8969314694404602,
+      "num_tokens": 742924163.0,
+      "step": 1976
+    },
+    {
+      "epoch": 2.9419642857142856,
+      "grad_norm": 1.6743862697204301,
+      "learning_rate": 4.331922253968781e-06,
+      "loss": 0.2758,
+      "mean_token_accuracy": 0.9063072204589844,
+      "num_tokens": 743303097.0,
+      "step": 1977
+    },
+    {
+      "epoch": 2.943452380952381,
+      "grad_norm": 1.8415947490959323,
+      "learning_rate": 4.326774760342234e-06,
+      "loss": 0.3085,
+      "mean_token_accuracy": 0.8880035281181335,
+      "num_tokens": 743681675.0,
+      "step": 1978
+    },
+    {
+      "epoch": 2.9449404761904763,
+      "grad_norm": 1.8183687162690387,
+      "learning_rate": 4.321627993317576e-06,
+      "loss": 0.3056,
+      "mean_token_accuracy": 0.8891053795814514,
+      "num_tokens": 744048639.0,
+      "step": 1979
+    },
+    {
+      "epoch": 2.946428571428571,
+      "grad_norm": 1.9829554606458708,
+      "learning_rate": 4.316481958449634e-06,
+      "loss": 0.3059,
+      "mean_token_accuracy": 0.892644464969635,
+      "num_tokens": 744424355.0,
+      "step": 1980
+    },
+    {
+      "epoch": 2.9479166666666665,
+      "grad_norm": 1.7392090274671421,
+      "learning_rate": 4.311336661292447e-06,
+      "loss": 0.2819,
+      "mean_token_accuracy": 0.904319703578949,
+      "num_tokens": 744803285.0,
+      "step": 1981
+    },
+    {
+      "epoch": 2.949404761904762,
+      "grad_norm": 1.746543582667389,
+      "learning_rate": 4.306192107399257e-06,
+      "loss": 0.2801,
+      "mean_token_accuracy": 0.8986444473266602,
+      "num_tokens": 745154089.0,
+      "step": 1982
+    },
+    {
+      "epoch": 2.950892857142857,
+      "grad_norm": 1.7742201407799127,
+      "learning_rate": 4.3010483023225045e-06,
+      "loss": 0.3319,
+      "mean_token_accuracy": 0.8841135501861572,
+      "num_tokens": 745518030.0,
+      "step": 1983
+    },
+    {
+      "epoch": 2.9523809523809526,
+      "grad_norm": 1.8682098706523882,
+      "learning_rate": 4.295905251613817e-06,
+      "loss": 0.3018,
+      "mean_token_accuracy": 0.8947604298591614,
+      "num_tokens": 745896126.0,
+      "step": 1984
+    },
+    {
+      "epoch": 2.9538690476190474,
+      "grad_norm": 1.7614875804810324,
+      "learning_rate": 4.290762960824018e-06,
+      "loss": 0.2894,
+      "mean_token_accuracy": 0.9000506401062012,
+      "num_tokens": 746275010.0,
+      "step": 1985
+    },
+    {
+      "epoch": 2.955357142857143,
+      "grad_norm": 1.6885043914607991,
+      "learning_rate": 4.285621435503101e-06,
+      "loss": 0.2693,
+      "mean_token_accuracy": 0.9035341739654541,
+      "num_tokens": 746653676.0,
+      "step": 1986
+    },
+    {
+      "epoch": 2.956845238095238,
+      "grad_norm": 1.7309190253925781,
+      "learning_rate": 4.280480681200237e-06,
+      "loss": 0.3262,
+      "mean_token_accuracy": 0.8892326354980469,
+      "num_tokens": 747034251.0,
+      "step": 1987
+    },
+    {
+      "epoch": 2.9583333333333335,
+      "grad_norm": 1.915080397048904,
+      "learning_rate": 4.275340703463767e-06,
+      "loss": 0.3385,
+      "mean_token_accuracy": 0.8814331293106079,
+      "num_tokens": 747412749.0,
+      "step": 1988
+    },
+    {
+      "epoch": 2.959821428571429,
+      "grad_norm": 1.8308206902111952,
+      "learning_rate": 4.270201507841191e-06,
+      "loss": 0.3313,
+      "mean_token_accuracy": 0.879142165184021,
+      "num_tokens": 747790282.0,
+      "step": 1989
+    },
+    {
+      "epoch": 2.9613095238095237,
+      "grad_norm": 1.7325922311790787,
+      "learning_rate": 4.265063099879162e-06,
+      "loss": 0.2818,
+      "mean_token_accuracy": 0.9017789363861084,
+      "num_tokens": 748168948.0,
+      "step": 1990
+    },
+    {
+      "epoch": 2.962797619047619,
+      "grad_norm": 1.7561107713161779,
+      "learning_rate": 4.259925485123492e-06,
+      "loss": 0.2859,
+      "mean_token_accuracy": 0.8960176706314087,
+      "num_tokens": 748547337.0,
+      "step": 1991
+    },
+    {
+      "epoch": 2.9642857142857144,
+      "grad_norm": 1.670428393922109,
+      "learning_rate": 4.254788669119127e-06,
+      "loss": 0.2814,
+      "mean_token_accuracy": 0.8967117071151733,
+      "num_tokens": 748926199.0,
+      "step": 1992
+    },
+    {
+      "epoch": 2.9657738095238093,
+      "grad_norm": 1.7595692308928905,
+      "learning_rate": 4.2496526574101594e-06,
+      "loss": 0.2544,
+      "mean_token_accuracy": 0.911347508430481,
+      "num_tokens": 749303519.0,
+      "step": 1993
+    },
+    {
+      "epoch": 2.9672619047619047,
+      "grad_norm": 1.706461201300987,
+      "learning_rate": 4.244517455539808e-06,
+      "loss": 0.2856,
+      "mean_token_accuracy": 0.9016900658607483,
+      "num_tokens": 749682871.0,
+      "step": 1994
+    },
+    {
+      "epoch": 2.96875,
+      "grad_norm": 1.7265743010784953,
+      "learning_rate": 4.239383069050417e-06,
+      "loss": 0.2899,
+      "mean_token_accuracy": 0.9024329781532288,
+      "num_tokens": 750062253.0,
+      "step": 1995
+    },
+    {
+      "epoch": 2.9702380952380953,
+      "grad_norm": 1.778865135135174,
+      "learning_rate": 4.234249503483455e-06,
+      "loss": 0.3434,
+      "mean_token_accuracy": 0.8815294504165649,
+      "num_tokens": 750441987.0,
+      "step": 1996
+    },
+    {
+      "epoch": 2.9717261904761907,
+      "grad_norm": 1.8669699321544502,
+      "learning_rate": 4.229116764379501e-06,
+      "loss": 0.3247,
+      "mean_token_accuracy": 0.8865083456039429,
+      "num_tokens": 750817552.0,
+      "step": 1997
+    },
+    {
+      "epoch": 2.9732142857142856,
+      "grad_norm": 1.7968744634414493,
+      "learning_rate": 4.223984857278242e-06,
+      "loss": 0.2969,
+      "mean_token_accuracy": 0.8986260294914246,
+      "num_tokens": 751195536.0,
+      "step": 1998
+    },
+    {
+      "epoch": 2.974702380952381,
+      "grad_norm": 1.761329113585825,
+      "learning_rate": 4.218853787718472e-06,
+      "loss": 0.3308,
+      "mean_token_accuracy": 0.8874468803405762,
+      "num_tokens": 751575474.0,
+      "step": 1999
+    },
+    {
+      "epoch": 2.9761904761904763,
+      "grad_norm": 1.7342238984618414,
+      "learning_rate": 4.213723561238074e-06,
+      "loss": 0.2734,
+      "mean_token_accuracy": 0.9040635824203491,
+      "num_tokens": 751953269.0,
+      "step": 2000
+    },
+    {
+      "epoch": 2.977678571428571,
+      "grad_norm": 1.7704940978616839,
+      "learning_rate": 4.208594183374026e-06,
+      "loss": 0.3121,
+      "mean_token_accuracy": 0.8928819298744202,
+      "num_tokens": 752315461.0,
+      "step": 2001
+    },
+    {
+      "epoch": 2.9791666666666665,
+      "grad_norm": 1.7383618842249018,
+      "learning_rate": 4.203465659662388e-06,
+      "loss": 0.3067,
+      "mean_token_accuracy": 0.8935613036155701,
+      "num_tokens": 752676165.0,
+      "step": 2002
+    },
+    {
+      "epoch": 2.980654761904762,
+      "grad_norm": 1.8233271404082767,
+      "learning_rate": 4.198337995638299e-06,
+      "loss": 0.2814,
+      "mean_token_accuracy": 0.9032906889915466,
+      "num_tokens": 753053375.0,
+      "step": 2003
+    },
+    {
+      "epoch": 2.982142857142857,
+      "grad_norm": 1.7863952954920848,
+      "learning_rate": 4.193211196835967e-06,
+      "loss": 0.2836,
+      "mean_token_accuracy": 0.9027765393257141,
+      "num_tokens": 753431397.0,
+      "step": 2004
+    },
+    {
+      "epoch": 2.9836309523809526,
+      "grad_norm": 1.74475949883536,
+      "learning_rate": 4.188085268788673e-06,
+      "loss": 0.3186,
+      "mean_token_accuracy": 0.8850882053375244,
+      "num_tokens": 753809453.0,
+      "step": 2005
+    },
+    {
+      "epoch": 2.9851190476190474,
+      "grad_norm": 1.8564729568159533,
+      "learning_rate": 4.182960217028752e-06,
+      "loss": 0.2911,
+      "mean_token_accuracy": 0.8956653475761414,
+      "num_tokens": 754176443.0,
+      "step": 2006
+    },
+    {
+      "epoch": 2.986607142857143,
+      "grad_norm": 1.7341652736483906,
+      "learning_rate": 4.177836047087596e-06,
+      "loss": 0.3078,
+      "mean_token_accuracy": 0.8920948505401611,
+      "num_tokens": 754556059.0,
+      "step": 2007
+    },
+    {
+      "epoch": 2.988095238095238,
+      "grad_norm": 1.7462295106400243,
+      "learning_rate": 4.172712764495645e-06,
+      "loss": 0.2883,
+      "mean_token_accuracy": 0.8985545635223389,
+      "num_tokens": 754933677.0,
+      "step": 2008
+    },
+    {
+      "epoch": 2.9895833333333335,
+      "grad_norm": 1.699901457908299,
+      "learning_rate": 4.1675903747823795e-06,
+      "loss": 0.29,
+      "mean_token_accuracy": 0.8954553604125977,
+      "num_tokens": 755312832.0,
+      "step": 2009
+    },
+    {
+      "epoch": 2.991071428571429,
+      "grad_norm": 1.914565983846665,
+      "learning_rate": 4.162468883476319e-06,
+      "loss": 0.3034,
+      "mean_token_accuracy": 0.8924578428268433,
+      "num_tokens": 755685480.0,
+      "step": 2010
+    },
+    {
+      "epoch": 2.9925595238095237,
+      "grad_norm": 1.771646453115979,
+      "learning_rate": 4.157348296105012e-06,
+      "loss": 0.3062,
+      "mean_token_accuracy": 0.8925783038139343,
+      "num_tokens": 756064280.0,
+      "step": 2011
+    },
+    {
+      "epoch": 2.994047619047619,
+      "grad_norm": 1.8565935775543823,
+      "learning_rate": 4.152228618195031e-06,
+      "loss": 0.3279,
+      "mean_token_accuracy": 0.885180652141571,
+      "num_tokens": 756441460.0,
+      "step": 2012
+    },
+    {
+      "epoch": 2.9955357142857144,
+      "grad_norm": 1.9355966887610936,
+      "learning_rate": 4.147109855271966e-06,
+      "loss": 0.2494,
+      "mean_token_accuracy": 0.909260630607605,
+      "num_tokens": 756817484.0,
+      "step": 2013
+    },
+    {
+      "epoch": 2.9970238095238093,
+      "grad_norm": 1.801266313624046,
+      "learning_rate": 4.141992012860425e-06,
+      "loss": 0.2786,
+      "mean_token_accuracy": 0.9034545421600342,
+      "num_tokens": 757195629.0,
+      "step": 2014
+    },
+    {
+      "epoch": 2.9985119047619047,
+      "grad_norm": 1.846317679093046,
+      "learning_rate": 4.136875096484016e-06,
+      "loss": 0.3129,
+      "mean_token_accuracy": 0.8934651017189026,
+      "num_tokens": 757574166.0,
+      "step": 2015
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.7821720787544482,
+      "learning_rate": 4.131759111665349e-06,
+      "loss": 0.3137,
+      "mean_token_accuracy": 0.8886114954948425,
+      "num_tokens": 757952340.0,
+      "step": 2016
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 3360,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}