diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,10043 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 6250,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0008,
+      "grad_norm": 0.7526811361312866,
+      "learning_rate": 4.999992104320636e-05,
+      "loss": 1.0428,
+      "num_input_tokens_seen": 26624,
+      "step": 5
+    },
+    {
+      "epoch": 0.0016,
+      "grad_norm": 0.5957724452018738,
+      "learning_rate": 4.999968417332415e-05,
+      "loss": 1.1061,
+      "num_input_tokens_seen": 51408,
+      "step": 10
+    },
+    {
+      "epoch": 0.0024,
+      "grad_norm": 0.8826403021812439,
+      "learning_rate": 4.999928939184958e-05,
+      "loss": 1.0426,
+      "num_input_tokens_seen": 75040,
+      "step": 15
+    },
+    {
+      "epoch": 0.0032,
+      "grad_norm": 0.6897421479225159,
+      "learning_rate": 4.9998736701276295e-05,
+      "loss": 1.1472,
+      "num_input_tokens_seen": 98848,
+      "step": 20
+    },
+    {
+      "epoch": 0.004,
+      "grad_norm": 0.7432862520217896,
+      "learning_rate": 4.9998026105095405e-05,
+      "loss": 0.9781,
+      "num_input_tokens_seen": 123104,
+      "step": 25
+    },
+    {
+      "epoch": 0.0048,
+      "grad_norm": 0.9830231666564941,
+      "learning_rate": 4.999715760779541e-05,
+      "loss": 0.9889,
+      "num_input_tokens_seen": 152144,
+      "step": 30
+    },
+    {
+      "epoch": 0.0056,
+      "grad_norm": 0.9177331924438477,
+      "learning_rate": 4.999613121486222e-05,
+      "loss": 0.9345,
+      "num_input_tokens_seen": 177216,
+      "step": 35
+    },
+    {
+      "epoch": 0.0064,
+      "grad_norm": 0.6646199822425842,
+      "learning_rate": 4.999494693277907e-05,
+      "loss": 0.8539,
+      "num_input_tokens_seen": 203152,
+      "step": 40
+    },
+    {
+      "epoch": 0.0072,
+      "grad_norm": 0.5822590589523315,
+      "learning_rate": 4.999360476902656e-05,
+      "loss": 0.9183,
+      "num_input_tokens_seen": 233568,
+      "step": 45
+    },
+    {
+      "epoch": 0.008,
+      "grad_norm": 0.7686595916748047,
+      "learning_rate": 4.99921047320825e-05,
+      "loss": 0.873,
+      "num_input_tokens_seen": 264752,
+      "step": 50
+    },
+    {
+      "epoch": 0.0088,
+      "grad_norm": 0.729837954044342,
+      "learning_rate": 4.9990446831421955e-05,
+      "loss": 0.8676,
+      "num_input_tokens_seen": 291040,
+      "step": 55
+    },
+    {
+      "epoch": 0.0096,
+      "grad_norm": 0.9523835778236389,
+      "learning_rate": 4.998863107751711e-05,
+      "loss": 0.9004,
+      "num_input_tokens_seen": 321760,
+      "step": 60
+    },
+    {
+      "epoch": 0.0104,
+      "grad_norm": 0.6720367670059204,
+      "learning_rate": 4.9986657481837277e-05,
+      "loss": 0.8536,
+      "num_input_tokens_seen": 347168,
+      "step": 65
+    },
+    {
+      "epoch": 0.0112,
+      "grad_norm": 0.4336840808391571,
+      "learning_rate": 4.998452605684874e-05,
+      "loss": 0.8027,
+      "num_input_tokens_seen": 373888,
+      "step": 70
+    },
+    {
+      "epoch": 0.012,
+      "grad_norm": 0.808559238910675,
+      "learning_rate": 4.998223681601473e-05,
+      "loss": 0.8075,
+      "num_input_tokens_seen": 398752,
+      "step": 75
+    },
+    {
+      "epoch": 0.0128,
+      "grad_norm": 0.5663979053497314,
+      "learning_rate": 4.997978977379536e-05,
+      "loss": 0.7919,
+      "num_input_tokens_seen": 421344,
+      "step": 80
+    },
+    {
+      "epoch": 0.0136,
+      "grad_norm": 0.5677878260612488,
+      "learning_rate": 4.9977184945647473e-05,
+      "loss": 0.7512,
+      "num_input_tokens_seen": 451296,
+      "step": 85
+    },
+    {
+      "epoch": 0.0144,
+      "grad_norm": 0.674132227897644,
+      "learning_rate": 4.997442234802456e-05,
+      "loss": 0.7713,
+      "num_input_tokens_seen": 482416,
+      "step": 90
+    },
+    {
+      "epoch": 0.0152,
+      "grad_norm": 0.5088427662849426,
+      "learning_rate": 4.997150199837671e-05,
+      "loss": 0.7965,
+      "num_input_tokens_seen": 513008,
+      "step": 95
+    },
+    {
+      "epoch": 0.016,
+      "grad_norm": 0.6657032370567322,
+      "learning_rate": 4.996842391515044e-05,
+      "loss": 0.8623,
+      "num_input_tokens_seen": 537984,
+      "step": 100
+    },
+    {
+      "epoch": 0.0168,
+      "grad_norm": 0.6862130761146545,
+      "learning_rate": 4.996518811778858e-05,
+      "loss": 0.7797,
+      "num_input_tokens_seen": 564528,
+      "step": 105
+    },
+    {
+      "epoch": 0.0176,
+      "grad_norm": 0.6449868083000183,
+      "learning_rate": 4.99617946267302e-05,
+      "loss": 0.7732,
+      "num_input_tokens_seen": 588608,
+      "step": 110
+    },
+    {
+      "epoch": 0.0184,
+      "grad_norm": 0.5512914657592773,
+      "learning_rate": 4.9958243463410414e-05,
+      "loss": 0.7478,
+      "num_input_tokens_seen": 620752,
+      "step": 115
+    },
+    {
+      "epoch": 0.0192,
+      "grad_norm": 0.7411808371543884,
+      "learning_rate": 4.995453465026032e-05,
+      "loss": 0.7194,
+      "num_input_tokens_seen": 649200,
+      "step": 120
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.9926447868347168,
+      "learning_rate": 4.995066821070679e-05,
+      "loss": 0.7506,
+      "num_input_tokens_seen": 679200,
+      "step": 125
+    },
+    {
+      "epoch": 0.0208,
+      "grad_norm": 0.7455246448516846,
+      "learning_rate": 4.9946644169172355e-05,
+      "loss": 0.6886,
+      "num_input_tokens_seen": 702144,
+      "step": 130
+    },
+    {
+      "epoch": 0.0216,
+      "grad_norm": 0.5429201126098633,
+      "learning_rate": 4.9942462551075056e-05,
+      "loss": 0.8481,
+      "num_input_tokens_seen": 730128,
+      "step": 135
+    },
+    {
+      "epoch": 0.0224,
+      "grad_norm": 0.49708107113838196,
+      "learning_rate": 4.993812338282826e-05,
+      "loss": 0.7999,
+      "num_input_tokens_seen": 757248,
+      "step": 140
+    },
+    {
+      "epoch": 0.0232,
+      "grad_norm": 0.6150819063186646,
+      "learning_rate": 4.993362669184051e-05,
+      "loss": 0.7877,
+      "num_input_tokens_seen": 786096,
+      "step": 145
+    },
+    {
+      "epoch": 0.024,
+      "grad_norm": 0.6751016974449158,
+      "learning_rate": 4.992897250651535e-05,
+      "loss": 0.9312,
+      "num_input_tokens_seen": 814192,
+      "step": 150
+    },
+    {
+      "epoch": 0.0248,
+      "grad_norm": 0.6245042085647583,
+      "learning_rate": 4.992416085625115e-05,
+      "loss": 0.8767,
+      "num_input_tokens_seen": 840144,
+      "step": 155
+    },
+    {
+      "epoch": 0.0256,
+      "grad_norm": 0.5385093688964844,
+      "learning_rate": 4.9919191771440905e-05,
+      "loss": 0.8646,
+      "num_input_tokens_seen": 870368,
+      "step": 160
+    },
+    {
+      "epoch": 0.0264,
+      "grad_norm": 0.5674800276756287,
+      "learning_rate": 4.991406528347206e-05,
+      "loss": 0.7159,
+      "num_input_tokens_seen": 897296,
+      "step": 165
+    },
+    {
+      "epoch": 0.0272,
+      "grad_norm": 0.3683065176010132,
+      "learning_rate": 4.990878142472628e-05,
+      "loss": 0.7573,
+      "num_input_tokens_seen": 924576,
+      "step": 170
+    },
+    {
+      "epoch": 0.028,
+      "grad_norm": 0.5186157822608948,
+      "learning_rate": 4.990334022857932e-05,
+      "loss": 0.8083,
+      "num_input_tokens_seen": 945856,
+      "step": 175
+    },
+    {
+      "epoch": 0.0288,
+      "grad_norm": 0.5749205350875854,
+      "learning_rate": 4.9897741729400705e-05,
+      "loss": 0.7074,
+      "num_input_tokens_seen": 970416,
+      "step": 180
+    },
+    {
+      "epoch": 0.0296,
+      "grad_norm": 0.5518808960914612,
+      "learning_rate": 4.9891985962553606e-05,
+      "loss": 0.7709,
+      "num_input_tokens_seen": 994288,
+      "step": 185
+    },
+    {
+      "epoch": 0.0304,
+      "grad_norm": 0.5436009764671326,
+      "learning_rate": 4.988607296439458e-05,
+      "loss": 0.7,
+      "num_input_tokens_seen": 1029872,
+      "step": 190
+    },
+    {
+      "epoch": 0.0312,
+      "grad_norm": 0.4599573612213135,
+      "learning_rate": 4.988000277227334e-05,
+      "loss": 0.8251,
+      "num_input_tokens_seen": 1061072,
+      "step": 195
+    },
+    {
+      "epoch": 0.032,
+      "grad_norm": 0.4540678858757019,
+      "learning_rate": 4.987377542453251e-05,
+      "loss": 0.6707,
+      "num_input_tokens_seen": 1089312,
+      "step": 200
+    },
+    {
+      "epoch": 0.0328,
+      "grad_norm": 0.5874660611152649,
+      "learning_rate": 4.98673909605074e-05,
+      "loss": 0.7264,
+      "num_input_tokens_seen": 1114272,
+      "step": 205
+    },
+    {
+      "epoch": 0.0336,
+      "grad_norm": 0.6792047619819641,
+      "learning_rate": 4.9860849420525766e-05,
+      "loss": 0.7906,
+      "num_input_tokens_seen": 1139808,
+      "step": 210
+    },
+    {
+      "epoch": 0.0344,
+      "grad_norm": 0.5740874409675598,
+      "learning_rate": 4.985415084590752e-05,
+      "loss": 0.8062,
+      "num_input_tokens_seen": 1163072,
+      "step": 215
+    },
+    {
+      "epoch": 0.0352,
+      "grad_norm": 0.5089894533157349,
+      "learning_rate": 4.9847295278964514e-05,
+      "loss": 0.7432,
+      "num_input_tokens_seen": 1193936,
+      "step": 220
+    },
+    {
+      "epoch": 0.036,
+      "grad_norm": 0.7231270670890808,
+      "learning_rate": 4.984028276300021e-05,
+      "loss": 0.7586,
+      "num_input_tokens_seen": 1219696,
+      "step": 225
+    },
+    {
+      "epoch": 0.0368,
+      "grad_norm": 0.6494696140289307,
+      "learning_rate": 4.98331133423095e-05,
+      "loss": 0.7532,
+      "num_input_tokens_seen": 1248096,
+      "step": 230
+    },
+    {
+      "epoch": 0.0376,
+      "grad_norm": 0.6063010692596436,
+      "learning_rate": 4.9825787062178315e-05,
+      "loss": 0.786,
+      "num_input_tokens_seen": 1276624,
+      "step": 235
+    },
+    {
+      "epoch": 0.0384,
+      "grad_norm": 0.8775933384895325,
+      "learning_rate": 4.981830396888344e-05,
+      "loss": 0.7947,
+      "num_input_tokens_seen": 1303472,
+      "step": 240
+    },
+    {
+      "epoch": 0.0392,
+      "grad_norm": 0.7558068633079529,
+      "learning_rate": 4.981066410969215e-05,
+      "loss": 0.6988,
+      "num_input_tokens_seen": 1326816,
+      "step": 245
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.8880596160888672,
+      "learning_rate": 4.980286753286195e-05,
+      "loss": 0.7078,
+      "num_input_tokens_seen": 1351008,
+      "step": 250
+    },
+    {
+      "epoch": 0.0408,
+      "grad_norm": 0.661428689956665,
+      "learning_rate": 4.979491428764026e-05,
+      "loss": 0.7491,
+      "num_input_tokens_seen": 1374656,
+      "step": 255
+    },
+    {
+      "epoch": 0.0416,
+      "grad_norm": 0.7624719738960266,
+      "learning_rate": 4.9786804424264085e-05,
+      "loss": 0.75,
+      "num_input_tokens_seen": 1399264,
+      "step": 260
+    },
+    {
+      "epoch": 0.0424,
+      "grad_norm": 0.6995192170143127,
+      "learning_rate": 4.977853799395976e-05,
+      "loss": 0.798,
+      "num_input_tokens_seen": 1422304,
+      "step": 265
+    },
+    {
+      "epoch": 0.0432,
+      "grad_norm": 0.5227561593055725,
+      "learning_rate": 4.977011504894252e-05,
+      "loss": 0.8814,
+      "num_input_tokens_seen": 1447184,
+      "step": 270
+    },
+    {
+      "epoch": 0.044,
+      "grad_norm": 0.7046292424201965,
+      "learning_rate": 4.976153564241628e-05,
+      "loss": 0.7203,
+      "num_input_tokens_seen": 1474304,
+      "step": 275
+    },
+    {
+      "epoch": 0.0448,
+      "grad_norm": 0.7567644119262695,
+      "learning_rate": 4.975279982857324e-05,
+      "loss": 0.6936,
+      "num_input_tokens_seen": 1500896,
+      "step": 280
+    },
+    {
+      "epoch": 0.0456,
+      "grad_norm": 0.6787880063056946,
+      "learning_rate": 4.9743907662593524e-05,
+      "loss": 0.7872,
+      "num_input_tokens_seen": 1528688,
+      "step": 285
+    },
+    {
+      "epoch": 0.0464,
+      "grad_norm": 0.5113949775695801,
+      "learning_rate": 4.9734859200644905e-05,
+      "loss": 0.7517,
+      "num_input_tokens_seen": 1561328,
+      "step": 290
+    },
+    {
+      "epoch": 0.0472,
+      "grad_norm": 0.7206217050552368,
+      "learning_rate": 4.972565449988239e-05,
+      "loss": 0.6726,
+      "num_input_tokens_seen": 1589088,
+      "step": 295
+    },
+    {
+      "epoch": 0.048,
+      "grad_norm": 0.602922797203064,
+      "learning_rate": 4.971629361844785e-05,
+      "loss": 0.7259,
+      "num_input_tokens_seen": 1615712,
+      "step": 300
+    },
+    {
+      "epoch": 0.0488,
+      "grad_norm": 0.7673738598823547,
+      "learning_rate": 4.9706776615469716e-05,
+      "loss": 0.8337,
+      "num_input_tokens_seen": 1638640,
+      "step": 305
+    },
+    {
+      "epoch": 0.0496,
+      "grad_norm": 0.7302682995796204,
+      "learning_rate": 4.9697103551062556e-05,
+      "loss": 0.731,
+      "num_input_tokens_seen": 1664304,
+      "step": 310
+    },
+    {
+      "epoch": 0.0504,
+      "grad_norm": 0.45416679978370667,
+      "learning_rate": 4.968727448632669e-05,
+      "loss": 0.7285,
+      "num_input_tokens_seen": 1697648,
+      "step": 315
+    },
+    {
+      "epoch": 0.0512,
+      "grad_norm": 0.5968911051750183,
+      "learning_rate": 4.967728948334784e-05,
+      "loss": 0.723,
+      "num_input_tokens_seen": 1726608,
+      "step": 320
+    },
+    {
+      "epoch": 0.052,
+      "grad_norm": 0.6134063601493835,
+      "learning_rate": 4.96671486051967e-05,
+      "loss": 0.7918,
+      "num_input_tokens_seen": 1750912,
+      "step": 325
+    },
+    {
+      "epoch": 0.0528,
+      "grad_norm": 0.5388225317001343,
+      "learning_rate": 4.965685191592859e-05,
+      "loss": 0.6448,
+      "num_input_tokens_seen": 1782912,
+      "step": 330
+    },
+    {
+      "epoch": 0.0536,
+      "grad_norm": 0.6615162491798401,
+      "learning_rate": 4.964639948058297e-05,
+      "loss": 0.7874,
+      "num_input_tokens_seen": 1804704,
+      "step": 335
+    },
+    {
+      "epoch": 0.0544,
+      "grad_norm": 0.8656606078147888,
+      "learning_rate": 4.963579136518312e-05,
+      "loss": 0.7025,
+      "num_input_tokens_seen": 1827248,
+      "step": 340
+    },
+    {
+      "epoch": 0.0552,
+      "grad_norm": 0.7784980535507202,
+      "learning_rate": 4.962502763673565e-05,
+      "loss": 0.6676,
+      "num_input_tokens_seen": 1854304,
+      "step": 345
+    },
+    {
+      "epoch": 0.056,
+      "grad_norm": 0.847607433795929,
+      "learning_rate": 4.9614108363230135e-05,
+      "loss": 0.7774,
+      "num_input_tokens_seen": 1878768,
+      "step": 350
+    },
+    {
+      "epoch": 0.0568,
+      "grad_norm": 0.5412896275520325,
+      "learning_rate": 4.9603033613638626e-05,
+      "loss": 0.7641,
+      "num_input_tokens_seen": 1905744,
+      "step": 355
+    },
+    {
+      "epoch": 0.0576,
+      "grad_norm": 0.5192331671714783,
+      "learning_rate": 4.959180345791528e-05,
+      "loss": 0.7169,
+      "num_input_tokens_seen": 1931392,
+      "step": 360
+    },
+    {
+      "epoch": 0.0584,
+      "grad_norm": 0.7992143630981445,
+      "learning_rate": 4.958041796699583e-05,
+      "loss": 0.7033,
+      "num_input_tokens_seen": 1954304,
+      "step": 365
+    },
+    {
+      "epoch": 0.0592,
+      "grad_norm": 0.49692437052726746,
+      "learning_rate": 4.956887721279726e-05,
+      "loss": 0.6569,
+      "num_input_tokens_seen": 1987264,
+      "step": 370
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7032391428947449,
+      "learning_rate": 4.9557181268217227e-05,
+      "loss": 0.7809,
+      "num_input_tokens_seen": 2010160,
+      "step": 375
+    },
+    {
+      "epoch": 0.0608,
+      "grad_norm": 0.780989944934845,
+      "learning_rate": 4.9545330207133664e-05,
+      "loss": 0.811,
+      "num_input_tokens_seen": 2038880,
+      "step": 380
+    },
+    {
+      "epoch": 0.0616,
+      "grad_norm": 0.819433867931366,
+      "learning_rate": 4.953332410440435e-05,
+      "loss": 0.825,
+      "num_input_tokens_seen": 2065344,
+      "step": 385
+    },
+    {
+      "epoch": 0.0624,
+      "grad_norm": 0.7076752781867981,
+      "learning_rate": 4.952116303586631e-05,
+      "loss": 0.7479,
+      "num_input_tokens_seen": 2092064,
+      "step": 390
+    },
+    {
+      "epoch": 0.0632,
+      "grad_norm": 0.6264218688011169,
+      "learning_rate": 4.9508847078335495e-05,
+      "loss": 0.7246,
+      "num_input_tokens_seen": 2119360,
+      "step": 395
+    },
+    {
+      "epoch": 0.064,
+      "grad_norm": 0.5829480290412903,
+      "learning_rate": 4.949637630960617e-05,
+      "loss": 0.6956,
+      "num_input_tokens_seen": 2146560,
+      "step": 400
+    },
+    {
+      "epoch": 0.0648,
+      "grad_norm": 0.5653419494628906,
+      "learning_rate": 4.94837508084505e-05,
+      "loss": 0.7315,
+      "num_input_tokens_seen": 2169232,
+      "step": 405
+    },
+    {
+      "epoch": 0.0656,
+      "grad_norm": 1.0192047357559204,
+      "learning_rate": 4.947097065461801e-05,
+      "loss": 0.7075,
+      "num_input_tokens_seen": 2192224,
+      "step": 410
+    },
+    {
+      "epoch": 0.0664,
+      "grad_norm": 0.7392141819000244,
+      "learning_rate": 4.945803592883509e-05,
+      "loss": 0.811,
+      "num_input_tokens_seen": 2216784,
+      "step": 415
+    },
+    {
+      "epoch": 0.0672,
+      "grad_norm": 0.6470807194709778,
+      "learning_rate": 4.9444946712804494e-05,
+      "loss": 0.7835,
+      "num_input_tokens_seen": 2243120,
+      "step": 420
+    },
+    {
+      "epoch": 0.068,
+      "grad_norm": 0.5305742025375366,
+      "learning_rate": 4.943170308920484e-05,
+      "loss": 0.7211,
+      "num_input_tokens_seen": 2270896,
+      "step": 425
+    },
+    {
+      "epoch": 0.0688,
+      "grad_norm": 0.8647666573524475,
+      "learning_rate": 4.941830514169004e-05,
+      "loss": 0.72,
+      "num_input_tokens_seen": 2298528,
+      "step": 430
+    },
+    {
+      "epoch": 0.0696,
+      "grad_norm": 0.6244668364524841,
+      "learning_rate": 4.9404752954888824e-05,
+      "loss": 0.7206,
+      "num_input_tokens_seen": 2328080,
+      "step": 435
+    },
+    {
+      "epoch": 0.0704,
+      "grad_norm": 0.6552883386611938,
+      "learning_rate": 4.939104661440415e-05,
+      "loss": 0.8018,
+      "num_input_tokens_seen": 2355776,
+      "step": 440
+    },
+    {
+      "epoch": 0.0712,
+      "grad_norm": 0.8276055455207825,
+      "learning_rate": 4.937718620681273e-05,
+      "loss": 0.8267,
+      "num_input_tokens_seen": 2379056,
+      "step": 445
+    },
+    {
+      "epoch": 0.072,
+      "grad_norm": 0.6930189728736877,
+      "learning_rate": 4.9363171819664434e-05,
+      "loss": 0.8961,
+      "num_input_tokens_seen": 2401664,
+      "step": 450
+    },
+    {
+      "epoch": 0.0728,
+      "grad_norm": 0.7441433668136597,
+      "learning_rate": 4.934900354148173e-05,
+      "loss": 0.6942,
+      "num_input_tokens_seen": 2427456,
+      "step": 455
+    },
+    {
+      "epoch": 0.0736,
+      "grad_norm": 0.5929616093635559,
+      "learning_rate": 4.933468146175918e-05,
+      "loss": 0.7874,
+      "num_input_tokens_seen": 2450752,
+      "step": 460
+    },
+    {
+      "epoch": 0.0744,
+      "grad_norm": 0.5789006948471069,
+      "learning_rate": 4.9320205670962814e-05,
+      "loss": 0.7162,
+      "num_input_tokens_seen": 2473856,
+      "step": 465
+    },
+    {
+      "epoch": 0.0752,
+      "grad_norm": 0.6359069347381592,
+      "learning_rate": 4.9305576260529607e-05,
+      "loss": 0.7434,
+      "num_input_tokens_seen": 2502928,
+      "step": 470
+    },
+    {
+      "epoch": 0.076,
+      "grad_norm": 0.6155191659927368,
+      "learning_rate": 4.929079332286685e-05,
+      "loss": 0.6932,
+      "num_input_tokens_seen": 2536144,
+      "step": 475
+    },
+    {
+      "epoch": 0.0768,
+      "grad_norm": 0.6511387228965759,
+      "learning_rate": 4.927585695135162e-05,
+      "loss": 0.8053,
+      "num_input_tokens_seen": 2562688,
+      "step": 480
+    },
+    {
+      "epoch": 0.0776,
+      "grad_norm": 0.5791414976119995,
+      "learning_rate": 4.926076724033016e-05,
+      "loss": 0.7482,
+      "num_input_tokens_seen": 2594480,
+      "step": 485
+    },
+    {
+      "epoch": 0.0784,
+      "grad_norm": 0.5258495807647705,
+      "learning_rate": 4.9245524285117274e-05,
+      "loss": 0.7075,
+      "num_input_tokens_seen": 2624736,
+      "step": 490
+    },
+    {
+      "epoch": 0.0792,
+      "grad_norm": 0.5191717743873596,
+      "learning_rate": 4.923012818199576e-05,
+      "loss": 0.6132,
+      "num_input_tokens_seen": 2648880,
+      "step": 495
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.8281647562980652,
+      "learning_rate": 4.9214579028215776e-05,
+      "loss": 0.6679,
+      "num_input_tokens_seen": 2675888,
+      "step": 500
+    },
+    {
+      "epoch": 0.0808,
+      "grad_norm": 0.588010847568512,
+      "learning_rate": 4.919887692199423e-05,
+      "loss": 0.7016,
+      "num_input_tokens_seen": 2699392,
+      "step": 505
+    },
+    {
+      "epoch": 0.0816,
+      "grad_norm": 0.8409311771392822,
+      "learning_rate": 4.918302196251415e-05,
+      "loss": 0.7216,
+      "num_input_tokens_seen": 2726432,
+      "step": 510
+    },
+    {
+      "epoch": 0.0824,
+      "grad_norm": 0.6029579639434814,
+      "learning_rate": 4.9167014249924075e-05,
+      "loss": 0.6602,
+      "num_input_tokens_seen": 2756336,
+      "step": 515
+    },
+    {
+      "epoch": 0.0832,
+      "grad_norm": 0.7269614934921265,
+      "learning_rate": 4.9150853885337426e-05,
+      "loss": 0.6956,
+      "num_input_tokens_seen": 2781648,
+      "step": 520
+    },
+    {
+      "epoch": 0.084,
+      "grad_norm": 0.5419861674308777,
+      "learning_rate": 4.913454097083185e-05,
+      "loss": 0.6427,
+      "num_input_tokens_seen": 2810336,
+      "step": 525
+    },
+    {
+      "epoch": 0.0848,
+      "grad_norm": 0.9006750583648682,
+      "learning_rate": 4.911807560944858e-05,
+      "loss": 0.8328,
+      "num_input_tokens_seen": 2836432,
+      "step": 530
+    },
+    {
+      "epoch": 0.0856,
+      "grad_norm": 0.7180121541023254,
+      "learning_rate": 4.9101457905191774e-05,
+      "loss": 0.8104,
+      "num_input_tokens_seen": 2863616,
+      "step": 535
+    },
+    {
+      "epoch": 0.0864,
+      "grad_norm": 0.6724479794502258,
+      "learning_rate": 4.9084687963027894e-05,
+      "loss": 0.6858,
+      "num_input_tokens_seen": 2891264,
+      "step": 540
+    },
+    {
+      "epoch": 0.0872,
+      "grad_norm": 0.7073305249214172,
+      "learning_rate": 4.906776588888502e-05,
+      "loss": 0.7271,
+      "num_input_tokens_seen": 2916256,
+      "step": 545
+    },
+    {
+      "epoch": 0.088,
+      "grad_norm": 0.7945154309272766,
+      "learning_rate": 4.905069178965215e-05,
+      "loss": 0.7527,
+      "num_input_tokens_seen": 2944112,
+      "step": 550
+    },
+    {
+      "epoch": 0.0888,
+      "grad_norm": 0.5791934728622437,
+      "learning_rate": 4.903346577317859e-05,
+      "loss": 0.7341,
+      "num_input_tokens_seen": 2972512,
+      "step": 555
+    },
+    {
+      "epoch": 0.0896,
+      "grad_norm": 0.8222031593322754,
+      "learning_rate": 4.90160879482732e-05,
+      "loss": 0.7339,
+      "num_input_tokens_seen": 2997168,
+      "step": 560
+    },
+    {
+      "epoch": 0.0904,
+      "grad_norm": 0.6719418168067932,
+      "learning_rate": 4.89985584247038e-05,
+      "loss": 0.768,
+      "num_input_tokens_seen": 3020880,
+      "step": 565
+    },
+    {
+      "epoch": 0.0912,
+      "grad_norm": 0.7740746140480042,
+      "learning_rate": 4.898087731319636e-05,
+      "loss": 0.7014,
+      "num_input_tokens_seen": 3044224,
+      "step": 570
+    },
+    {
+      "epoch": 0.092,
+      "grad_norm": 0.5642164945602417,
+      "learning_rate": 4.89630447254344e-05,
+      "loss": 0.6203,
+      "num_input_tokens_seen": 3071680,
+      "step": 575
+    },
+    {
+      "epoch": 0.0928,
+      "grad_norm": 1.4719825983047485,
+      "learning_rate": 4.894506077405824e-05,
+      "loss": 0.7461,
+      "num_input_tokens_seen": 3099088,
+      "step": 580
+    },
+    {
+      "epoch": 0.0936,
+      "grad_norm": 0.6961272954940796,
+      "learning_rate": 4.892692557266429e-05,
+      "loss": 0.7357,
+      "num_input_tokens_seen": 3127728,
+      "step": 585
+    },
+    {
+      "epoch": 0.0944,
+      "grad_norm": 0.686820924282074,
+      "learning_rate": 4.8908639235804324e-05,
+      "loss": 0.7819,
+      "num_input_tokens_seen": 3154336,
+      "step": 590
+    },
+    {
+      "epoch": 0.0952,
+      "grad_norm": 0.7145109176635742,
+      "learning_rate": 4.8890201878984796e-05,
+      "loss": 0.7121,
+      "num_input_tokens_seen": 3178768,
+      "step": 595
+    },
+    {
+      "epoch": 0.096,
+      "grad_norm": 0.6159213781356812,
+      "learning_rate": 4.887161361866608e-05,
+      "loss": 0.6698,
+      "num_input_tokens_seen": 3211968,
+      "step": 600
+    },
+    {
+      "epoch": 0.0968,
+      "grad_norm": 0.8054212927818298,
+      "learning_rate": 4.885287457226172e-05,
+      "loss": 0.7606,
+      "num_input_tokens_seen": 3238272,
+      "step": 605
+    },
+    {
+      "epoch": 0.0976,
+      "grad_norm": 1.1526386737823486,
+      "learning_rate": 4.8833984858137715e-05,
+      "loss": 0.7694,
+      "num_input_tokens_seen": 3270208,
+      "step": 610
+    },
+    {
+      "epoch": 0.0984,
+      "grad_norm": 0.5728780031204224,
+      "learning_rate": 4.8814944595611776e-05,
+      "loss": 0.7227,
+      "num_input_tokens_seen": 3296192,
+      "step": 615
+    },
+    {
+      "epoch": 0.0992,
+      "grad_norm": 0.6360820531845093,
+      "learning_rate": 4.8795753904952534e-05,
+      "loss": 0.7275,
+      "num_input_tokens_seen": 3321232,
+      "step": 620
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.6169213056564331,
+      "learning_rate": 4.877641290737884e-05,
+      "loss": 0.6746,
+      "num_input_tokens_seen": 3343984,
+      "step": 625
+    },
+    {
+      "epoch": 0.1008,
+      "grad_norm": 0.8000876307487488,
+      "learning_rate": 4.8756921725058934e-05,
+      "loss": 0.8223,
+      "num_input_tokens_seen": 3367824,
+      "step": 630
+    },
+    {
+      "epoch": 0.1016,
+      "grad_norm": 0.5983218550682068,
+      "learning_rate": 4.8737280481109724e-05,
+      "loss": 0.8487,
+      "num_input_tokens_seen": 3394800,
+      "step": 635
+    },
+    {
+      "epoch": 0.1024,
+      "grad_norm": 0.9402346014976501,
+      "learning_rate": 4.871748929959598e-05,
+      "loss": 0.7441,
+      "num_input_tokens_seen": 3421360,
+      "step": 640
+    },
+    {
+      "epoch": 0.1032,
+      "grad_norm": 0.6266387104988098,
+      "learning_rate": 4.869754830552956e-05,
+      "loss": 0.7631,
+      "num_input_tokens_seen": 3449584,
+      "step": 645
+    },
+    {
+      "epoch": 0.104,
+      "grad_norm": 0.7829232215881348,
+      "learning_rate": 4.867745762486861e-05,
+      "loss": 0.7793,
+      "num_input_tokens_seen": 3477168,
+      "step": 650
+    },
+    {
+      "epoch": 0.1048,
+      "grad_norm": 0.7125943303108215,
+      "learning_rate": 4.86572173845168e-05,
+      "loss": 0.7415,
+      "num_input_tokens_seen": 3505056,
+      "step": 655
+    },
+    {
+      "epoch": 0.1056,
+      "grad_norm": 0.6520003080368042,
+      "learning_rate": 4.863682771232248e-05,
+      "loss": 0.7157,
+      "num_input_tokens_seen": 3534576,
+      "step": 660
+    },
+    {
+      "epoch": 0.1064,
+      "grad_norm": 0.5907071828842163,
+      "learning_rate": 4.861628873707792e-05,
+      "loss": 0.7287,
+      "num_input_tokens_seen": 3560688,
+      "step": 665
+    },
+    {
+      "epoch": 0.1072,
+      "grad_norm": 0.8829016089439392,
+      "learning_rate": 4.859560058851844e-05,
+      "loss": 0.7351,
+      "num_input_tokens_seen": 3586176,
+      "step": 670
+    },
+    {
+      "epoch": 0.108,
+      "grad_norm": 0.917322039604187,
+      "learning_rate": 4.8574763397321614e-05,
+      "loss": 0.6213,
+      "num_input_tokens_seen": 3615472,
+      "step": 675
+    },
+    {
+      "epoch": 0.1088,
+      "grad_norm": 0.6344768404960632,
+      "learning_rate": 4.855377729510648e-05,
+      "loss": 0.729,
+      "num_input_tokens_seen": 3638256,
+      "step": 680
+    },
+    {
+      "epoch": 0.1096,
+      "grad_norm": 0.7305799722671509,
+      "learning_rate": 4.8532642414432674e-05,
+      "loss": 0.7242,
+      "num_input_tokens_seen": 3667824,
+      "step": 685
+    },
+    {
+      "epoch": 0.1104,
+      "grad_norm": 0.7569555044174194,
+      "learning_rate": 4.851135888879958e-05,
+      "loss": 0.7831,
+      "num_input_tokens_seen": 3695408,
+      "step": 690
+    },
+    {
+      "epoch": 0.1112,
+      "grad_norm": 0.7566932439804077,
+      "learning_rate": 4.8489926852645505e-05,
+      "loss": 0.7181,
+      "num_input_tokens_seen": 3719888,
+      "step": 695
+    },
+    {
+      "epoch": 0.112,
+      "grad_norm": 0.7932357788085938,
+      "learning_rate": 4.846834644134686e-05,
+      "loss": 0.7961,
+      "num_input_tokens_seen": 3744512,
+      "step": 700
+    },
+    {
+      "epoch": 0.1128,
+      "grad_norm": 0.708210825920105,
+      "learning_rate": 4.844661779121722e-05,
+      "loss": 0.8362,
+      "num_input_tokens_seen": 3771968,
+      "step": 705
+    },
+    {
+      "epoch": 0.1136,
+      "grad_norm": 0.7361094951629639,
+      "learning_rate": 4.8424741039506575e-05,
+      "loss": 0.7645,
+      "num_input_tokens_seen": 3801680,
+      "step": 710
+    },
+    {
+      "epoch": 0.1144,
+      "grad_norm": 0.48908814787864685,
+      "learning_rate": 4.840271632440038e-05,
+      "loss": 0.7042,
+      "num_input_tokens_seen": 3833952,
+      "step": 715
+    },
+    {
+      "epoch": 0.1152,
+      "grad_norm": 0.6167788505554199,
+      "learning_rate": 4.8380543785018677e-05,
+      "loss": 0.7476,
+      "num_input_tokens_seen": 3860144,
+      "step": 720
+    },
+    {
+      "epoch": 0.116,
+      "grad_norm": 0.68650883436203,
+      "learning_rate": 4.8358223561415304e-05,
+      "loss": 0.7415,
+      "num_input_tokens_seen": 3890304,
+      "step": 725
+    },
+    {
+      "epoch": 0.1168,
+      "grad_norm": 0.7059746384620667,
+      "learning_rate": 4.833575579457691e-05,
+      "loss": 0.6717,
+      "num_input_tokens_seen": 3914560,
+      "step": 730
+    },
+    {
+      "epoch": 0.1176,
+      "grad_norm": 0.8362336158752441,
+      "learning_rate": 4.8313140626422125e-05,
+      "loss": 0.7545,
+      "num_input_tokens_seen": 3940128,
+      "step": 735
+    },
+    {
+      "epoch": 0.1184,
+      "grad_norm": 0.5400592684745789,
+      "learning_rate": 4.829037819980065e-05,
+      "loss": 0.7809,
+      "num_input_tokens_seen": 3970608,
+      "step": 740
+    },
+    {
+      "epoch": 0.1192,
+      "grad_norm": 1.0431326627731323,
+      "learning_rate": 4.8267468658492335e-05,
+      "loss": 0.7904,
+      "num_input_tokens_seen": 3996960,
+      "step": 745
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.5358662605285645,
+      "learning_rate": 4.8244412147206284e-05,
+      "loss": 0.7688,
+      "num_input_tokens_seen": 4021488,
+      "step": 750
+    },
+    {
+      "epoch": 0.1208,
+      "grad_norm": 0.8147661685943604,
+      "learning_rate": 4.822120881157998e-05,
+      "loss": 0.7819,
+      "num_input_tokens_seen": 4047136,
+      "step": 755
+    },
+    {
+      "epoch": 0.1216,
+      "grad_norm": 0.6247139573097229,
+      "learning_rate": 4.819785879817827e-05,
+      "loss": 0.6757,
+      "num_input_tokens_seen": 4072256,
+      "step": 760
+    },
+    {
+      "epoch": 0.1224,
+      "grad_norm": 0.8849884271621704,
+      "learning_rate": 4.817436225449255e-05,
+      "loss": 0.8952,
+      "num_input_tokens_seen": 4095328,
+      "step": 765
+    },
+    {
+      "epoch": 0.1232,
+      "grad_norm": 0.8693557977676392,
+      "learning_rate": 4.8150719328939755e-05,
+      "loss": 0.6998,
+      "num_input_tokens_seen": 4118896,
+      "step": 770
+    },
+    {
+      "epoch": 0.124,
+      "grad_norm": 0.9492819905281067,
+      "learning_rate": 4.812693017086145e-05,
+      "loss": 0.7675,
+      "num_input_tokens_seen": 4144576,
+      "step": 775
+    },
+    {
+      "epoch": 0.1248,
+      "grad_norm": 0.8479375243186951,
+      "learning_rate": 4.810299493052289e-05,
+      "loss": 0.7332,
+      "num_input_tokens_seen": 4172448,
+      "step": 780
+    },
+    {
+      "epoch": 0.1256,
+      "grad_norm": 0.7956748008728027,
+      "learning_rate": 4.8078913759112066e-05,
+      "loss": 0.6942,
+      "num_input_tokens_seen": 4196032,
+      "step": 785
+    },
+    {
+      "epoch": 0.1264,
+      "grad_norm": 0.6426162123680115,
+      "learning_rate": 4.805468680873874e-05,
+      "loss": 0.7536,
+      "num_input_tokens_seen": 4224320,
+      "step": 790
+    },
+    {
+      "epoch": 0.1272,
+      "grad_norm": 0.6501713991165161,
+      "learning_rate": 4.803031423243349e-05,
+      "loss": 0.6722,
+      "num_input_tokens_seen": 4252752,
+      "step": 795
+    },
+    {
+      "epoch": 0.128,
+      "grad_norm": 0.773551881313324,
+      "learning_rate": 4.800579618414676e-05,
+      "loss": 0.7651,
+      "num_input_tokens_seen": 4278480,
+      "step": 800
+    },
+    {
+      "epoch": 0.1288,
+      "grad_norm": 0.6473078727722168,
+      "learning_rate": 4.7981132818747876e-05,
+      "loss": 0.6626,
+      "num_input_tokens_seen": 4305920,
+      "step": 805
+    },
+    {
+      "epoch": 0.1296,
+      "grad_norm": 0.5944277048110962,
+      "learning_rate": 4.795632429202405e-05,
+      "loss": 0.8511,
+      "num_input_tokens_seen": 4330448,
+      "step": 810
+    },
+    {
+      "epoch": 0.1304,
+      "grad_norm": 0.6878964900970459,
+      "learning_rate": 4.793137076067942e-05,
+      "loss": 0.7524,
+      "num_input_tokens_seen": 4356880,
+      "step": 815
+    },
+    {
+      "epoch": 0.1312,
+      "grad_norm": 0.9247101545333862,
+      "learning_rate": 4.790627238233405e-05,
+      "loss": 0.8498,
+      "num_input_tokens_seen": 4383744,
+      "step": 820
+    },
+    {
+      "epoch": 0.132,
+      "grad_norm": 0.9401747584342957,
+      "learning_rate": 4.788102931552294e-05,
+      "loss": 0.647,
+      "num_input_tokens_seen": 4411120,
+      "step": 825
+    },
+    {
+      "epoch": 0.1328,
+      "grad_norm": 0.765521764755249,
+      "learning_rate": 4.7855641719695023e-05,
+      "loss": 0.7766,
+      "num_input_tokens_seen": 4435920,
+      "step": 830
+    },
+    {
+      "epoch": 0.1336,
+      "grad_norm": 0.8985924124717712,
+      "learning_rate": 4.783010975521216e-05,
+      "loss": 0.7426,
+      "num_input_tokens_seen": 4462768,
+      "step": 835
+    },
+    {
+      "epoch": 0.1344,
+      "grad_norm": 0.8223104476928711,
+      "learning_rate": 4.78044335833481e-05,
+      "loss": 0.6919,
+      "num_input_tokens_seen": 4493232,
+      "step": 840
+    },
+    {
+      "epoch": 0.1352,
+      "grad_norm": 0.6721159219741821,
+      "learning_rate": 4.7778613366287505e-05,
+      "loss": 0.7221,
+      "num_input_tokens_seen": 4520048,
+      "step": 845
+    },
+    {
+      "epoch": 0.136,
+      "grad_norm": 0.6136463284492493,
+      "learning_rate": 4.775264926712489e-05,
+      "loss": 0.7344,
+      "num_input_tokens_seen": 4545984,
+      "step": 850
+    },
+    {
+      "epoch": 0.1368,
+      "grad_norm": 0.7662776708602905,
+      "learning_rate": 4.772654144986364e-05,
+      "loss": 0.6648,
+      "num_input_tokens_seen": 4577296,
+      "step": 855
+    },
+    {
+      "epoch": 0.1376,
+      "grad_norm": 0.8455452919006348,
+      "learning_rate": 4.7700290079414896e-05,
+      "loss": 0.7513,
+      "num_input_tokens_seen": 4602272,
+      "step": 860
+    },
+    {
+      "epoch": 0.1384,
+      "grad_norm": 0.8613116145133972,
+      "learning_rate": 4.767389532159659e-05,
+      "loss": 0.7792,
+      "num_input_tokens_seen": 4631008,
+      "step": 865
+    },
+    {
+      "epoch": 0.1392,
+      "grad_norm": 0.5791791677474976,
+      "learning_rate": 4.764735734313236e-05,
+      "loss": 0.7529,
+      "num_input_tokens_seen": 4660112,
+      "step": 870
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.8197824954986572,
+      "learning_rate": 4.762067631165049e-05,
+      "loss": 0.6728,
+      "num_input_tokens_seen": 4689504,
+      "step": 875
+    },
+    {
+      "epoch": 0.1408,
+      "grad_norm": 0.9617213606834412,
+      "learning_rate": 4.759385239568289e-05,
+      "loss": 0.6935,
+      "num_input_tokens_seen": 4715312,
+      "step": 880
+    },
+    {
+      "epoch": 0.1416,
+      "grad_norm": 0.7933773398399353,
+      "learning_rate": 4.756688576466398e-05,
+      "loss": 0.8062,
+      "num_input_tokens_seen": 4735936,
+      "step": 885
+    },
+    {
+      "epoch": 0.1424,
+      "grad_norm": 0.975724458694458,
+      "learning_rate": 4.753977658892967e-05,
+      "loss": 0.7149,
+      "num_input_tokens_seen": 4760256,
+      "step": 890
+    },
+    {
+      "epoch": 0.1432,
+      "grad_norm": 0.755167543888092,
+      "learning_rate": 4.751252503971624e-05,
+      "loss": 0.7062,
+      "num_input_tokens_seen": 4789264,
+      "step": 895
+    },
+    {
+      "epoch": 0.144,
+      "grad_norm": 0.84686279296875,
+      "learning_rate": 4.7485131289159276e-05,
+      "loss": 0.837,
+      "num_input_tokens_seen": 4815344,
+      "step": 900
+    },
+    {
+      "epoch": 0.1448,
+      "grad_norm": 0.9440627098083496,
+      "learning_rate": 4.745759551029261e-05,
+      "loss": 0.6907,
+      "num_input_tokens_seen": 4840528,
+      "step": 905
+    },
+    {
+      "epoch": 0.1456,
+      "grad_norm": 0.7293935418128967,
+      "learning_rate": 4.742991787704719e-05,
+      "loss": 0.7192,
+      "num_input_tokens_seen": 4868032,
+      "step": 910
+    },
+    {
+      "epoch": 0.1464,
+      "grad_norm": 0.6401370763778687,
+      "learning_rate": 4.7402098564249974e-05,
+      "loss": 0.7223,
+      "num_input_tokens_seen": 4893376,
+      "step": 915
+    },
+    {
+      "epoch": 0.1472,
+      "grad_norm": 0.8882667422294617,
+      "learning_rate": 4.737413774762287e-05,
+      "loss": 0.6847,
+      "num_input_tokens_seen": 4918288,
+      "step": 920
+    },
+    {
+      "epoch": 0.148,
+      "grad_norm": 0.7663973569869995,
+      "learning_rate": 4.73460356037816e-05,
+      "loss": 0.7273,
+      "num_input_tokens_seen": 4944688,
+      "step": 925
+    },
+    {
+      "epoch": 0.1488,
+      "grad_norm": 0.7966660857200623,
+      "learning_rate": 4.731779231023456e-05,
+      "loss": 0.7087,
+      "num_input_tokens_seen": 4969744,
+      "step": 930
+    },
+    {
+      "epoch": 0.1496,
+      "grad_norm": 1.3590271472930908,
+      "learning_rate": 4.728940804538176e-05,
+      "loss": 0.7771,
+      "num_input_tokens_seen": 4997072,
+      "step": 935
+    },
+    {
+      "epoch": 0.1504,
+      "grad_norm": 0.8245935440063477,
+      "learning_rate": 4.7260882988513624e-05,
+      "loss": 0.7598,
+      "num_input_tokens_seen": 5024672,
+      "step": 940
+    },
+    {
+      "epoch": 0.1512,
+      "grad_norm": 0.6972777247428894,
+      "learning_rate": 4.723221731980993e-05,
+      "loss": 0.7961,
+      "num_input_tokens_seen": 5051952,
+      "step": 945
+    },
+    {
+      "epoch": 0.152,
+      "grad_norm": 0.7691385746002197,
+      "learning_rate": 4.720341122033862e-05,
+      "loss": 0.773,
+      "num_input_tokens_seen": 5074528,
+      "step": 950
+    },
+    {
+      "epoch": 0.1528,
+      "grad_norm": 1.4361584186553955,
+      "learning_rate": 4.717446487205466e-05,
+      "loss": 0.7216,
+      "num_input_tokens_seen": 5099840,
+      "step": 955
+    },
+    {
+      "epoch": 0.1536,
+      "grad_norm": 0.9640448689460754,
+      "learning_rate": 4.714537845779894e-05,
+      "loss": 0.6569,
+      "num_input_tokens_seen": 5122848,
+      "step": 960
+    },
+    {
+      "epoch": 0.1544,
+      "grad_norm": 0.8036720156669617,
+      "learning_rate": 4.7116152161297045e-05,
+      "loss": 0.7994,
+      "num_input_tokens_seen": 5152320,
+      "step": 965
+    },
+    {
+      "epoch": 0.1552,
+      "grad_norm": 0.7904760241508484,
+      "learning_rate": 4.708678616715815e-05,
+      "loss": 0.7259,
+      "num_input_tokens_seen": 5178816,
+      "step": 970
+    },
+    {
+      "epoch": 0.156,
+      "grad_norm": 0.7007213830947876,
+      "learning_rate": 4.7057280660873835e-05,
+      "loss": 0.747,
+      "num_input_tokens_seen": 5208112,
+      "step": 975
+    },
+    {
+      "epoch": 0.1568,
+      "grad_norm": 0.8959905505180359,
+      "learning_rate": 4.702763582881692e-05,
+      "loss": 0.8487,
+      "num_input_tokens_seen": 5231200,
+      "step": 980
+    },
+    {
+      "epoch": 0.1576,
+      "grad_norm": 0.8828222155570984,
+      "learning_rate": 4.699785185824026e-05,
+      "loss": 0.7654,
+      "num_input_tokens_seen": 5257312,
+      "step": 985
+    },
+    {
+      "epoch": 0.1584,
+      "grad_norm": 1.0120501518249512,
+      "learning_rate": 4.696792893727562e-05,
+      "loss": 0.7748,
+      "num_input_tokens_seen": 5280288,
+      "step": 990
+    },
+    {
+      "epoch": 0.1592,
+      "grad_norm": 0.8728295564651489,
+      "learning_rate": 4.693786725493242e-05,
+      "loss": 0.6957,
+      "num_input_tokens_seen": 5308272,
+      "step": 995
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 1.252682089805603,
+      "learning_rate": 4.690766700109659e-05,
+      "loss": 0.7418,
+      "num_input_tokens_seen": 5335568,
+      "step": 1000
+    },
+    {
+      "epoch": 0.1608,
+      "grad_norm": 1.0933836698532104,
+      "learning_rate": 4.6877328366529346e-05,
+      "loss": 0.8225,
+      "num_input_tokens_seen": 5361872,
+      "step": 1005
+    },
+    {
+      "epoch": 0.1616,
+      "grad_norm": 0.8167702555656433,
+      "learning_rate": 4.684685154286599e-05,
+      "loss": 0.8552,
+      "num_input_tokens_seen": 5387456,
+      "step": 1010
+    },
+    {
+      "epoch": 0.1624,
+      "grad_norm": 0.6640987396240234,
+      "learning_rate": 4.681623672261469e-05,
+      "loss": 0.6654,
+      "num_input_tokens_seen": 5411472,
+      "step": 1015
+    },
+    {
+      "epoch": 0.1632,
+      "grad_norm": 0.9311304688453674,
+      "learning_rate": 4.678548409915532e-05,
+      "loss": 0.7339,
+      "num_input_tokens_seen": 5439648,
+      "step": 1020
+    },
+    {
+      "epoch": 0.164,
+      "grad_norm": 0.9733570218086243,
+      "learning_rate": 4.675459386673815e-05,
+      "loss": 0.7324,
+      "num_input_tokens_seen": 5468416,
+      "step": 1025
+    },
+    {
+      "epoch": 0.1648,
+      "grad_norm": 0.6521669030189514,
+      "learning_rate": 4.6723566220482664e-05,
+      "loss": 0.7065,
+      "num_input_tokens_seen": 5498800,
+      "step": 1030
+    },
+    {
+      "epoch": 0.1656,
+      "grad_norm": 0.6702024340629578,
+      "learning_rate": 4.669240135637635e-05,
+      "loss": 0.6822,
+      "num_input_tokens_seen": 5527856,
+      "step": 1035
+    },
+    {
+      "epoch": 0.1664,
+      "grad_norm": 1.3757870197296143,
+      "learning_rate": 4.666109947127343e-05,
+      "loss": 0.7554,
+      "num_input_tokens_seen": 5550848,
+      "step": 1040
+    },
+    {
+      "epoch": 0.1672,
+      "grad_norm": 0.7441242933273315,
+      "learning_rate": 4.662966076289362e-05,
+      "loss": 0.6784,
+      "num_input_tokens_seen": 5581552,
+      "step": 1045
+    },
+    {
+      "epoch": 0.168,
+      "grad_norm": 0.7709234356880188,
+      "learning_rate": 4.659808542982088e-05,
+      "loss": 0.8294,
+      "num_input_tokens_seen": 5604288,
+      "step": 1050
+    },
+    {
+      "epoch": 0.1688,
+      "grad_norm": 0.5358073115348816,
+      "learning_rate": 4.6566373671502196e-05,
+      "loss": 0.6633,
+      "num_input_tokens_seen": 5630336,
+      "step": 1055
+    },
+    {
+      "epoch": 0.1696,
+      "grad_norm": 0.5856006741523743,
+      "learning_rate": 4.653452568824625e-05,
+      "loss": 0.6684,
+      "num_input_tokens_seen": 5662480,
+      "step": 1060
+    },
+    {
+      "epoch": 0.1704,
+      "grad_norm": 0.7003797292709351,
+      "learning_rate": 4.650254168122222e-05,
+      "loss": 0.7109,
+      "num_input_tokens_seen": 5687376,
+      "step": 1065
+    },
+    {
+      "epoch": 0.1712,
+      "grad_norm": 0.7874431014060974,
+      "learning_rate": 4.647042185245847e-05,
+      "loss": 0.8036,
+      "num_input_tokens_seen": 5714896,
+      "step": 1070
+    },
+    {
+      "epoch": 0.172,
+      "grad_norm": 0.6988087296485901,
+      "learning_rate": 4.643816640484131e-05,
+      "loss": 0.6575,
+      "num_input_tokens_seen": 5740192,
+      "step": 1075
+    },
+    {
+      "epoch": 0.1728,
+      "grad_norm": 0.982477068901062,
+      "learning_rate": 4.640577554211366e-05,
+      "loss": 0.7477,
+      "num_input_tokens_seen": 5768656,
+      "step": 1080
+    },
+    {
+      "epoch": 0.1736,
+      "grad_norm": 1.1265698671340942,
+      "learning_rate": 4.6373249468873833e-05,
+      "loss": 0.7555,
+      "num_input_tokens_seen": 5794576,
+      "step": 1085
+    },
+    {
+      "epoch": 0.1744,
+      "grad_norm": 0.6747913360595703,
+      "learning_rate": 4.634058839057417e-05,
+      "loss": 0.6695,
+      "num_input_tokens_seen": 5823296,
+      "step": 1090
+    },
+    {
+      "epoch": 0.1752,
+      "grad_norm": 0.8027223348617554,
+      "learning_rate": 4.63077925135198e-05,
+      "loss": 0.6948,
+      "num_input_tokens_seen": 5846928,
+      "step": 1095
+    },
+    {
+      "epoch": 0.176,
+      "grad_norm": 0.7862293720245361,
+      "learning_rate": 4.6274862044867304e-05,
+      "loss": 0.7728,
+      "num_input_tokens_seen": 5871968,
+      "step": 1100
+    },
+    {
+      "epoch": 0.1768,
+      "grad_norm": 0.7790197134017944,
+      "learning_rate": 4.624179719262342e-05,
+      "loss": 0.765,
+      "num_input_tokens_seen": 5900304,
+      "step": 1105
+    },
+    {
+      "epoch": 0.1776,
+      "grad_norm": 0.8996221423149109,
+      "learning_rate": 4.6208598165643715e-05,
+      "loss": 0.6515,
+      "num_input_tokens_seen": 5925792,
+      "step": 1110
+    },
+    {
+      "epoch": 0.1784,
+      "grad_norm": 0.7972677946090698,
+      "learning_rate": 4.61752651736313e-05,
+      "loss": 0.75,
+      "num_input_tokens_seen": 5950672,
+      "step": 1115
+    },
+    {
+      "epoch": 0.1792,
+      "grad_norm": 0.6896753907203674,
+      "learning_rate": 4.614179842713547e-05,
+      "loss": 0.6592,
+      "num_input_tokens_seen": 5985552,
+      "step": 1120
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.823128342628479,
+      "learning_rate": 4.610819813755038e-05,
+      "loss": 0.8463,
+      "num_input_tokens_seen": 6009904,
+      "step": 1125
+    },
+    {
+      "epoch": 0.1808,
+      "grad_norm": 0.8550837635993958,
+      "learning_rate": 4.607446451711372e-05,
+      "loss": 0.7349,
+      "num_input_tokens_seen": 6034160,
+      "step": 1130
+    },
+    {
+      "epoch": 0.1816,
+      "grad_norm": 0.8120406270027161,
+      "learning_rate": 4.604059777890537e-05,
+      "loss": 0.6396,
+      "num_input_tokens_seen": 6056544,
+      "step": 1135
+    },
+    {
+      "epoch": 0.1824,
+      "grad_norm": 0.6196752786636353,
+      "learning_rate": 4.6006598136846056e-05,
+      "loss": 0.6164,
+      "num_input_tokens_seen": 6083920,
+      "step": 1140
+    },
+    {
+      "epoch": 0.1832,
+      "grad_norm": 0.6641353368759155,
+      "learning_rate": 4.5972465805695996e-05,
+      "loss": 0.6775,
+      "num_input_tokens_seen": 6111520,
+      "step": 1145
+    },
+    {
+      "epoch": 0.184,
+      "grad_norm": 0.7323867082595825,
+      "learning_rate": 4.593820100105355e-05,
+      "loss": 0.6295,
+      "num_input_tokens_seen": 6141056,
+      "step": 1150
+    },
+    {
+      "epoch": 0.1848,
+      "grad_norm": 0.6919586658477783,
+      "learning_rate": 4.590380393935383e-05,
+      "loss": 0.7429,
+      "num_input_tokens_seen": 6163408,
+      "step": 1155
+    },
+    {
+      "epoch": 0.1856,
+      "grad_norm": 0.9530206322669983,
+      "learning_rate": 4.5869274837867394e-05,
+      "loss": 0.7516,
+      "num_input_tokens_seen": 6188816,
+      "step": 1160
+    },
+    {
+      "epoch": 0.1864,
+      "grad_norm": 0.9966915845870972,
+      "learning_rate": 4.583461391469879e-05,
+      "loss": 0.7524,
+      "num_input_tokens_seen": 6216800,
+      "step": 1165
+    },
+    {
+      "epoch": 0.1872,
+      "grad_norm": 1.096708059310913,
+      "learning_rate": 4.579982138878527e-05,
+      "loss": 0.7337,
+      "num_input_tokens_seen": 6245888,
+      "step": 1170
+    },
+    {
+      "epoch": 0.188,
+      "grad_norm": 0.8707526326179504,
+      "learning_rate": 4.5764897479895317e-05,
+      "loss": 0.7891,
+      "num_input_tokens_seen": 6275120,
+      "step": 1175
+    },
+    {
+      "epoch": 0.1888,
+      "grad_norm": 0.7489879727363586,
+      "learning_rate": 4.5729842408627334e-05,
+      "loss": 0.79,
+      "num_input_tokens_seen": 6299760,
+      "step": 1180
+    },
+    {
+      "epoch": 0.1896,
+      "grad_norm": 0.7835171222686768,
+      "learning_rate": 4.5694656396408195e-05,
+      "loss": 0.7506,
+      "num_input_tokens_seen": 6326720,
+      "step": 1185
+    },
+    {
+      "epoch": 0.1904,
+      "grad_norm": 0.7588552832603455,
+      "learning_rate": 4.565933966549189e-05,
+      "loss": 0.6294,
+      "num_input_tokens_seen": 6353728,
+      "step": 1190
+    },
+    {
+      "epoch": 0.1912,
+      "grad_norm": 0.6706573367118835,
+      "learning_rate": 4.5623892438958074e-05,
+      "loss": 0.7564,
+      "num_input_tokens_seen": 6379536,
+      "step": 1195
+    },
+    {
+      "epoch": 0.192,
+      "grad_norm": 0.7340586185455322,
+      "learning_rate": 4.558831494071069e-05,
+      "loss": 0.7683,
+      "num_input_tokens_seen": 6407152,
+      "step": 1200
+    },
+    {
+      "epoch": 0.1928,
+      "grad_norm": 0.735789954662323,
+      "learning_rate": 4.555260739547657e-05,
+      "loss": 0.7701,
+      "num_input_tokens_seen": 6434480,
+      "step": 1205
+    },
+    {
+      "epoch": 0.1936,
+      "grad_norm": 0.8325262069702148,
+      "learning_rate": 4.5516770028803954e-05,
+      "loss": 0.694,
+      "num_input_tokens_seen": 6463424,
+      "step": 1210
+    },
+    {
+      "epoch": 0.1944,
+      "grad_norm": 0.7930346727371216,
+      "learning_rate": 4.548080306706114e-05,
+      "loss": 0.7322,
+      "num_input_tokens_seen": 6487136,
+      "step": 1215
+    },
+    {
+      "epoch": 0.1952,
+      "grad_norm": 0.7683930397033691,
+      "learning_rate": 4.5444706737435014e-05,
+      "loss": 0.7616,
+      "num_input_tokens_seen": 6513120,
+      "step": 1220
+    },
+    {
+      "epoch": 0.196,
+      "grad_norm": 0.600136399269104,
+      "learning_rate": 4.5408481267929605e-05,
+      "loss": 0.6743,
+      "num_input_tokens_seen": 6543040,
+      "step": 1225
+    },
+    {
+      "epoch": 0.1968,
+      "grad_norm": 0.9069085121154785,
+      "learning_rate": 4.5372126887364655e-05,
+      "loss": 0.7377,
+      "num_input_tokens_seen": 6572432,
+      "step": 1230
+    },
+    {
+      "epoch": 0.1976,
+      "grad_norm": 0.9226580262184143,
+      "learning_rate": 4.533564382537421e-05,
+      "loss": 0.7766,
+      "num_input_tokens_seen": 6593136,
+      "step": 1235
+    },
+    {
+      "epoch": 0.1984,
+      "grad_norm": 0.7376300096511841,
+      "learning_rate": 4.529903231240511e-05,
+      "loss": 0.7873,
+      "num_input_tokens_seen": 6621024,
+      "step": 1240
+    },
+    {
+      "epoch": 0.1992,
+      "grad_norm": 0.6371731162071228,
+      "learning_rate": 4.5262292579715556e-05,
+      "loss": 0.7096,
+      "num_input_tokens_seen": 6646480,
+      "step": 1245
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.8643271327018738,
+      "learning_rate": 4.522542485937369e-05,
+      "loss": 0.8187,
+      "num_input_tokens_seen": 6674032,
+      "step": 1250
+    },
+    {
+      "epoch": 0.2008,
+      "grad_norm": 0.8012109398841858,
+      "learning_rate": 4.518842938425605e-05,
+      "loss": 0.772,
+      "num_input_tokens_seen": 6700112,
+      "step": 1255
+    },
+    {
+      "epoch": 0.2016,
+      "grad_norm": 0.7719143033027649,
+      "learning_rate": 4.5151306388046175e-05,
+      "loss": 0.6796,
+      "num_input_tokens_seen": 6727008,
+      "step": 1260
+    },
+    {
+      "epoch": 0.2024,
+      "grad_norm": 0.8668113946914673,
+      "learning_rate": 4.511405610523309e-05,
+      "loss": 0.7177,
+      "num_input_tokens_seen": 6752768,
+      "step": 1265
+    },
+    {
+      "epoch": 0.2032,
+      "grad_norm": 0.8964220285415649,
+      "learning_rate": 4.5076678771109815e-05,
+      "loss": 0.7078,
+      "num_input_tokens_seen": 6778112,
+      "step": 1270
+    },
+    {
+      "epoch": 0.204,
+      "grad_norm": 0.7097613215446472,
+      "learning_rate": 4.503917462177192e-05,
+      "loss": 0.6496,
+      "num_input_tokens_seen": 6804432,
+      "step": 1275
+    },
+    {
+      "epoch": 0.2048,
+      "grad_norm": 0.842675507068634,
+      "learning_rate": 4.5001543894115975e-05,
+      "loss": 0.6802,
+      "num_input_tokens_seen": 6829824,
+      "step": 1280
+    },
+    {
+      "epoch": 0.2056,
+      "grad_norm": 0.7390193343162537,
+      "learning_rate": 4.496378682583813e-05,
+      "loss": 0.7187,
+      "num_input_tokens_seen": 6858480,
+      "step": 1285
+    },
+    {
+      "epoch": 0.2064,
+      "grad_norm": 0.5758505463600159,
+      "learning_rate": 4.492590365543253e-05,
+      "loss": 0.6198,
+      "num_input_tokens_seen": 6886960,
+      "step": 1290
+    },
+    {
+      "epoch": 0.2072,
+      "grad_norm": 0.9554662108421326,
+      "learning_rate": 4.488789462218987e-05,
+      "loss": 0.6105,
+      "num_input_tokens_seen": 6912560,
+      "step": 1295
+    },
+    {
+      "epoch": 0.208,
+      "grad_norm": 0.9423254728317261,
+      "learning_rate": 4.484975996619589e-05,
+      "loss": 0.7671,
+      "num_input_tokens_seen": 6938912,
+      "step": 1300
+    },
+    {
+      "epoch": 0.2088,
+      "grad_norm": 0.7120509743690491,
+      "learning_rate": 4.481149992832977e-05,
+      "loss": 0.6833,
+      "num_input_tokens_seen": 6967616,
+      "step": 1305
+    },
+    {
+      "epoch": 0.2096,
+      "grad_norm": 0.9409400224685669,
+      "learning_rate": 4.477311475026271e-05,
+      "loss": 0.7547,
+      "num_input_tokens_seen": 6993872,
+      "step": 1310
+    },
+    {
+      "epoch": 0.2104,
+      "grad_norm": 0.8102442026138306,
+      "learning_rate": 4.473460467445637e-05,
+      "loss": 0.7479,
+      "num_input_tokens_seen": 7020784,
+      "step": 1315
+    },
+    {
+      "epoch": 0.2112,
+      "grad_norm": 0.787486732006073,
+      "learning_rate": 4.46959699441613e-05,
+      "loss": 0.761,
+      "num_input_tokens_seen": 7045024,
+      "step": 1320
+    },
+    {
+      "epoch": 0.212,
+      "grad_norm": 0.8877683877944946,
+      "learning_rate": 4.465721080341547e-05,
+      "loss": 0.7612,
+      "num_input_tokens_seen": 7072448,
+      "step": 1325
+    },
+    {
+      "epoch": 0.2128,
+      "grad_norm": 0.7483372688293457,
+      "learning_rate": 4.461832749704268e-05,
+      "loss": 0.6792,
+      "num_input_tokens_seen": 7097776,
+      "step": 1330
+    },
+    {
+      "epoch": 0.2136,
+      "grad_norm": 0.7852973341941833,
+      "learning_rate": 4.457932027065102e-05,
+      "loss": 0.7357,
+      "num_input_tokens_seen": 7123568,
+      "step": 1335
+    },
+    {
+      "epoch": 0.2144,
+      "grad_norm": 0.7306565642356873,
+      "learning_rate": 4.4540189370631315e-05,
+      "loss": 0.6676,
+      "num_input_tokens_seen": 7151728,
+      "step": 1340
+    },
+    {
+      "epoch": 0.2152,
+      "grad_norm": 0.7990534901618958,
+      "learning_rate": 4.4500935044155626e-05,
+      "loss": 0.7394,
+      "num_input_tokens_seen": 7181664,
+      "step": 1345
+    },
+    {
+      "epoch": 0.216,
+      "grad_norm": 1.287644863128662,
+      "learning_rate": 4.4461557539175594e-05,
+      "loss": 0.8017,
+      "num_input_tokens_seen": 7210336,
+      "step": 1350
+    },
+    {
+      "epoch": 0.2168,
+      "grad_norm": 0.7476962208747864,
+      "learning_rate": 4.4422057104420946e-05,
+      "loss": 0.6533,
+      "num_input_tokens_seen": 7240992,
+      "step": 1355
+    },
+    {
+      "epoch": 0.2176,
+      "grad_norm": 0.8233410120010376,
+      "learning_rate": 4.4382433989397895e-05,
+      "loss": 0.7029,
+      "num_input_tokens_seen": 7268048,
+      "step": 1360
+    },
+    {
+      "epoch": 0.2184,
+      "grad_norm": 0.609846293926239,
+      "learning_rate": 4.434268844438758e-05,
+      "loss": 0.7096,
+      "num_input_tokens_seen": 7297616,
+      "step": 1365
+    },
+    {
+      "epoch": 0.2192,
+      "grad_norm": 1.010886549949646,
+      "learning_rate": 4.4302820720444456e-05,
+      "loss": 0.8103,
+      "num_input_tokens_seen": 7326912,
+      "step": 1370
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.7681688070297241,
+      "learning_rate": 4.426283106939474e-05,
+      "loss": 0.6238,
+      "num_input_tokens_seen": 7355136,
+      "step": 1375
+    },
+    {
+      "epoch": 0.2208,
+      "grad_norm": 0.7759270071983337,
+      "learning_rate": 4.422271974383479e-05,
+      "loss": 0.6625,
+      "num_input_tokens_seen": 7377584,
+      "step": 1380
+    },
+    {
+      "epoch": 0.2216,
+      "grad_norm": 0.831362783908844,
+      "learning_rate": 4.418248699712955e-05,
+      "loss": 0.6831,
+      "num_input_tokens_seen": 7405552,
+      "step": 1385
+    },
+    {
+      "epoch": 0.2224,
+      "grad_norm": 0.7530121207237244,
+      "learning_rate": 4.414213308341092e-05,
+      "loss": 0.7664,
+      "num_input_tokens_seen": 7430960,
+      "step": 1390
+    },
+    {
+      "epoch": 0.2232,
+      "grad_norm": 0.8572810292243958,
+      "learning_rate": 4.410165825757613e-05,
+      "loss": 0.7273,
+      "num_input_tokens_seen": 7457136,
+      "step": 1395
+    },
+    {
+      "epoch": 0.224,
+      "grad_norm": 0.7553160190582275,
+      "learning_rate": 4.40610627752862e-05,
+      "loss": 0.6607,
+      "num_input_tokens_seen": 7482208,
+      "step": 1400
+    },
+    {
+      "epoch": 0.2248,
+      "grad_norm": 0.6897515058517456,
+      "learning_rate": 4.4020346892964246e-05,
+      "loss": 0.731,
+      "num_input_tokens_seen": 7515760,
+      "step": 1405
+    },
+    {
+      "epoch": 0.2256,
+      "grad_norm": 0.7974178791046143,
+      "learning_rate": 4.3979510867793917e-05,
+      "loss": 0.7258,
+      "num_input_tokens_seen": 7542944,
+      "step": 1410
+    },
+    {
+      "epoch": 0.2264,
+      "grad_norm": 0.8745766282081604,
+      "learning_rate": 4.393855495771774e-05,
+      "loss": 0.6566,
+      "num_input_tokens_seen": 7573760,
+      "step": 1415
+    },
+    {
+      "epoch": 0.2272,
+      "grad_norm": 0.749857485294342,
+      "learning_rate": 4.38974794214355e-05,
+      "loss": 0.7433,
+      "num_input_tokens_seen": 7606592,
+      "step": 1420
+    },
+    {
+      "epoch": 0.228,
+      "grad_norm": 0.7722298502922058,
+      "learning_rate": 4.3856284518402594e-05,
+      "loss": 0.7452,
+      "num_input_tokens_seen": 7628672,
+      "step": 1425
+    },
+    {
+      "epoch": 0.2288,
+      "grad_norm": 0.8768362998962402,
+      "learning_rate": 4.381497050882845e-05,
+      "loss": 0.7077,
+      "num_input_tokens_seen": 7658528,
+      "step": 1430
+    },
+    {
+      "epoch": 0.2296,
+      "grad_norm": 0.7979273796081543,
+      "learning_rate": 4.377353765367479e-05,
+      "loss": 0.6274,
+      "num_input_tokens_seen": 7685248,
+      "step": 1435
+    },
+    {
+      "epoch": 0.2304,
+      "grad_norm": 0.988314151763916,
+      "learning_rate": 4.3731986214654035e-05,
+      "loss": 0.6845,
+      "num_input_tokens_seen": 7713616,
+      "step": 1440
+    },
+    {
+      "epoch": 0.2312,
+      "grad_norm": 0.7991346120834351,
+      "learning_rate": 4.3690316454227674e-05,
+      "loss": 0.7115,
+      "num_input_tokens_seen": 7740304,
+      "step": 1445
+    },
+    {
+      "epoch": 0.232,
+      "grad_norm": 1.072383999824524,
+      "learning_rate": 4.3648528635604556e-05,
+      "loss": 0.7209,
+      "num_input_tokens_seen": 7766848,
+      "step": 1450
+    },
+    {
+      "epoch": 0.2328,
+      "grad_norm": 1.357325792312622,
+      "learning_rate": 4.360662302273924e-05,
+      "loss": 0.8239,
+      "num_input_tokens_seen": 7791888,
+      "step": 1455
+    },
+    {
+      "epoch": 0.2336,
+      "grad_norm": 0.6083495020866394,
+      "learning_rate": 4.3564599880330385e-05,
+      "loss": 0.6199,
+      "num_input_tokens_seen": 7822448,
+      "step": 1460
+    },
+    {
+      "epoch": 0.2344,
+      "grad_norm": 0.7359746098518372,
+      "learning_rate": 4.352245947381898e-05,
+      "loss": 0.7481,
+      "num_input_tokens_seen": 7848464,
+      "step": 1465
+    },
+    {
+      "epoch": 0.2352,
+      "grad_norm": 0.9160847067832947,
+      "learning_rate": 4.348020206938672e-05,
+      "loss": 0.7235,
+      "num_input_tokens_seen": 7877216,
+      "step": 1470
+    },
+    {
+      "epoch": 0.236,
+      "grad_norm": 0.7445215582847595,
+      "learning_rate": 4.343782793395435e-05,
+      "loss": 0.7345,
+      "num_input_tokens_seen": 7904368,
+      "step": 1475
+    },
+    {
+      "epoch": 0.2368,
+      "grad_norm": 0.8324536681175232,
+      "learning_rate": 4.3395337335179945e-05,
+      "loss": 0.7532,
+      "num_input_tokens_seen": 7931520,
+      "step": 1480
+    },
+    {
+      "epoch": 0.2376,
+      "grad_norm": 1.0249683856964111,
+      "learning_rate": 4.335273054145722e-05,
+      "loss": 0.6902,
+      "num_input_tokens_seen": 7953296,
+      "step": 1485
+    },
+    {
+      "epoch": 0.2384,
+      "grad_norm": 0.6565669775009155,
+      "learning_rate": 4.3310007821913836e-05,
+      "loss": 0.7329,
+      "num_input_tokens_seen": 7978832,
+      "step": 1490
+    },
+    {
+      "epoch": 0.2392,
+      "grad_norm": 0.8256237506866455,
+      "learning_rate": 4.32671694464097e-05,
+      "loss": 0.6693,
+      "num_input_tokens_seen": 8004992,
+      "step": 1495
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.9722650051116943,
+      "learning_rate": 4.3224215685535294e-05,
+      "loss": 0.7418,
+      "num_input_tokens_seen": 8027824,
+      "step": 1500
+    },
+    {
+      "epoch": 0.2408,
+      "grad_norm": 0.599818229675293,
+      "learning_rate": 4.31811468106099e-05,
+      "loss": 0.6157,
+      "num_input_tokens_seen": 8058528,
+      "step": 1505
+    },
+    {
+      "epoch": 0.2416,
+      "grad_norm": 1.0976861715316772,
+      "learning_rate": 4.3137963093679945e-05,
+      "loss": 0.6302,
+      "num_input_tokens_seen": 8081984,
+      "step": 1510
+    },
+    {
+      "epoch": 0.2424,
+      "grad_norm": 0.5699600577354431,
+      "learning_rate": 4.309466480751726e-05,
+      "loss": 0.628,
+      "num_input_tokens_seen": 8113216,
+      "step": 1515
+    },
+    {
+      "epoch": 0.2432,
+      "grad_norm": 0.8899049758911133,
+      "learning_rate": 4.305125222561736e-05,
+      "loss": 0.635,
+      "num_input_tokens_seen": 8142080,
+      "step": 1520
+    },
+    {
+      "epoch": 0.244,
+      "grad_norm": 0.9494242072105408,
+      "learning_rate": 4.3007725622197674e-05,
+      "loss": 0.8114,
+      "num_input_tokens_seen": 8171008,
+      "step": 1525
+    },
+    {
+      "epoch": 0.2448,
+      "grad_norm": 0.9237959384918213,
+      "learning_rate": 4.296408527219592e-05,
+      "loss": 0.6678,
+      "num_input_tokens_seen": 8197696,
+      "step": 1530
+    },
+    {
+      "epoch": 0.2456,
+      "grad_norm": 0.8756378889083862,
+      "learning_rate": 4.292033145126825e-05,
+      "loss": 0.8364,
+      "num_input_tokens_seen": 8225552,
+      "step": 1535
+    },
+    {
+      "epoch": 0.2464,
+      "grad_norm": 0.9631836414337158,
+      "learning_rate": 4.287646443578758e-05,
+      "loss": 0.7312,
+      "num_input_tokens_seen": 8257120,
+      "step": 1540
+    },
+    {
+      "epoch": 0.2472,
+      "grad_norm": 0.920713484287262,
+      "learning_rate": 4.283248450284182e-05,
+      "loss": 0.8067,
+      "num_input_tokens_seen": 8282400,
+      "step": 1545
+    },
+    {
+      "epoch": 0.248,
+      "grad_norm": 1.0773414373397827,
+      "learning_rate": 4.2788391930232136e-05,
+      "loss": 0.7109,
+      "num_input_tokens_seen": 8309568,
+      "step": 1550
+    },
+    {
+      "epoch": 0.2488,
+      "grad_norm": 0.5621623396873474,
+      "learning_rate": 4.2744186996471174e-05,
+      "loss": 0.6543,
+      "num_input_tokens_seen": 8338864,
+      "step": 1555
+    },
+    {
+      "epoch": 0.2496,
+      "grad_norm": 0.8737258315086365,
+      "learning_rate": 4.269986998078132e-05,
+      "loss": 0.7401,
+      "num_input_tokens_seen": 8364592,
+      "step": 1560
+    },
+    {
+      "epoch": 0.2504,
+      "grad_norm": 0.8454060554504395,
+      "learning_rate": 4.265544116309294e-05,
+      "loss": 0.7538,
+      "num_input_tokens_seen": 8391120,
+      "step": 1565
+    },
+    {
+      "epoch": 0.2512,
+      "grad_norm": 0.8107228875160217,
+      "learning_rate": 4.261090082404258e-05,
+      "loss": 0.7705,
+      "num_input_tokens_seen": 8418320,
+      "step": 1570
+    },
+    {
+      "epoch": 0.252,
+      "grad_norm": 0.7339603304862976,
+      "learning_rate": 4.256624924497123e-05,
+      "loss": 0.6846,
+      "num_input_tokens_seen": 8446640,
+      "step": 1575
+    },
+    {
+      "epoch": 0.2528,
+      "grad_norm": 1.0036543607711792,
+      "learning_rate": 4.252148670792254e-05,
+      "loss": 0.8502,
+      "num_input_tokens_seen": 8470416,
+      "step": 1580
+    },
+    {
+      "epoch": 0.2536,
+      "grad_norm": 0.8186982870101929,
+      "learning_rate": 4.2476613495641026e-05,
+      "loss": 0.6987,
+      "num_input_tokens_seen": 8498160,
+      "step": 1585
+    },
+    {
+      "epoch": 0.2544,
+      "grad_norm": 0.9724487066268921,
+      "learning_rate": 4.2431629891570266e-05,
+      "loss": 0.6461,
+      "num_input_tokens_seen": 8525904,
+      "step": 1590
+    },
+    {
+      "epoch": 0.2552,
+      "grad_norm": 0.5958553552627563,
+      "learning_rate": 4.238653617985118e-05,
+      "loss": 0.7143,
+      "num_input_tokens_seen": 8551872,
+      "step": 1595
+    },
+    {
+      "epoch": 0.256,
+      "grad_norm": 1.0192784070968628,
+      "learning_rate": 4.234133264532012e-05,
+      "loss": 0.7215,
+      "num_input_tokens_seen": 8583440,
+      "step": 1600
+    },
+    {
+      "epoch": 0.2568,
+      "grad_norm": 0.7806874513626099,
+      "learning_rate": 4.229601957350722e-05,
+      "loss": 0.8008,
+      "num_input_tokens_seen": 8609632,
+      "step": 1605
+    },
+    {
+      "epoch": 0.2576,
+      "grad_norm": 1.086475133895874,
+      "learning_rate": 4.225059725063444e-05,
+      "loss": 0.6612,
+      "num_input_tokens_seen": 8633888,
+      "step": 1610
+    },
+    {
+      "epoch": 0.2584,
+      "grad_norm": 0.6213988065719604,
+      "learning_rate": 4.2205065963613864e-05,
+      "loss": 0.7544,
+      "num_input_tokens_seen": 8660288,
+      "step": 1615
+    },
+    {
+      "epoch": 0.2592,
+      "grad_norm": 1.0608100891113281,
+      "learning_rate": 4.2159426000045854e-05,
+      "loss": 0.7569,
+      "num_input_tokens_seen": 8689184,
+      "step": 1620
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.7601464986801147,
+      "learning_rate": 4.211367764821722e-05,
+      "loss": 0.8161,
+      "num_input_tokens_seen": 8713504,
+      "step": 1625
+    },
+    {
+      "epoch": 0.2608,
+      "grad_norm": 0.9310168623924255,
+      "learning_rate": 4.206782119709942e-05,
+      "loss": 0.8283,
+      "num_input_tokens_seen": 8741088,
+      "step": 1630
+    },
+    {
+      "epoch": 0.2616,
+      "grad_norm": 0.6408126354217529,
+      "learning_rate": 4.20218569363467e-05,
+      "loss": 0.5745,
+      "num_input_tokens_seen": 8767456,
+      "step": 1635
+    },
+    {
+      "epoch": 0.2624,
+      "grad_norm": 1.1697090864181519,
+      "learning_rate": 4.197578515629435e-05,
+      "loss": 0.7525,
+      "num_input_tokens_seen": 8791952,
+      "step": 1640
+    },
+    {
+      "epoch": 0.2632,
+      "grad_norm": 0.9160236716270447,
+      "learning_rate": 4.192960614795675e-05,
+      "loss": 0.7991,
+      "num_input_tokens_seen": 8816080,
+      "step": 1645
+    },
+    {
+      "epoch": 0.264,
+      "grad_norm": 1.0530091524124146,
+      "learning_rate": 4.188332020302561e-05,
+      "loss": 0.7297,
+      "num_input_tokens_seen": 8841536,
+      "step": 1650
+    },
+    {
+      "epoch": 0.2648,
+      "grad_norm": 0.8888834118843079,
+      "learning_rate": 4.183692761386813e-05,
+      "loss": 0.6276,
+      "num_input_tokens_seen": 8869872,
+      "step": 1655
+    },
+    {
+      "epoch": 0.2656,
+      "grad_norm": 0.6144154667854309,
+      "learning_rate": 4.179042867352511e-05,
+      "loss": 0.7127,
+      "num_input_tokens_seen": 8893152,
+      "step": 1660
+    },
+    {
+      "epoch": 0.2664,
+      "grad_norm": 0.814166247844696,
+      "learning_rate": 4.174382367570912e-05,
+      "loss": 0.7712,
+      "num_input_tokens_seen": 8923040,
+      "step": 1665
+    },
+    {
+      "epoch": 0.2672,
+      "grad_norm": 0.8960988521575928,
+      "learning_rate": 4.169711291480266e-05,
+      "loss": 0.8388,
+      "num_input_tokens_seen": 8945856,
+      "step": 1670
+    },
+    {
+      "epoch": 0.268,
+      "grad_norm": 0.8164514303207397,
+      "learning_rate": 4.165029668585629e-05,
+      "loss": 0.7538,
+      "num_input_tokens_seen": 8971664,
+      "step": 1675
+    },
+    {
+      "epoch": 0.2688,
+      "grad_norm": 0.8044324517250061,
+      "learning_rate": 4.160337528458676e-05,
+      "loss": 0.708,
+      "num_input_tokens_seen": 8996064,
+      "step": 1680
+    },
+    {
+      "epoch": 0.2696,
+      "grad_norm": 0.7704948782920837,
+      "learning_rate": 4.155634900737513e-05,
+      "loss": 0.668,
+      "num_input_tokens_seen": 9022416,
+      "step": 1685
+    },
+    {
+      "epoch": 0.2704,
+      "grad_norm": 0.8315603137016296,
+      "learning_rate": 4.150921815126493e-05,
+      "loss": 0.752,
+      "num_input_tokens_seen": 9052480,
+      "step": 1690
+    },
+    {
+      "epoch": 0.2712,
+      "grad_norm": 0.7516652345657349,
+      "learning_rate": 4.1461983013960245e-05,
+      "loss": 0.6534,
+      "num_input_tokens_seen": 9079760,
+      "step": 1695
+    },
+    {
+      "epoch": 0.272,
+      "grad_norm": 0.7449467182159424,
+      "learning_rate": 4.1414643893823914e-05,
+      "loss": 0.6808,
+      "num_input_tokens_seen": 9109424,
+      "step": 1700
+    },
+    {
+      "epoch": 0.2728,
+      "grad_norm": 0.6889111995697021,
+      "learning_rate": 4.136720108987552e-05,
+      "loss": 0.7627,
+      "num_input_tokens_seen": 9132128,
+      "step": 1705
+    },
+    {
+      "epoch": 0.2736,
+      "grad_norm": 0.9195050597190857,
+      "learning_rate": 4.131965490178959e-05,
+      "loss": 0.6527,
+      "num_input_tokens_seen": 9160960,
+      "step": 1710
+    },
+    {
+      "epoch": 0.2744,
+      "grad_norm": 0.9934877157211304,
+      "learning_rate": 4.1272005629893714e-05,
+      "loss": 0.7102,
+      "num_input_tokens_seen": 9190992,
+      "step": 1715
+    },
+    {
+      "epoch": 0.2752,
+      "grad_norm": 0.8816946148872375,
+      "learning_rate": 4.122425357516658e-05,
+      "loss": 0.67,
+      "num_input_tokens_seen": 9218320,
+      "step": 1720
+    },
+    {
+      "epoch": 0.276,
+      "grad_norm": 0.7904371619224548,
+      "learning_rate": 4.1176399039236116e-05,
+      "loss": 0.7159,
+      "num_input_tokens_seen": 9246304,
+      "step": 1725
+    },
+    {
+      "epoch": 0.2768,
+      "grad_norm": 0.795921266078949,
+      "learning_rate": 4.112844232437757e-05,
+      "loss": 0.8248,
+      "num_input_tokens_seen": 9271856,
+      "step": 1730
+    },
+    {
+      "epoch": 0.2776,
+      "grad_norm": 0.8109453320503235,
+      "learning_rate": 4.108038373351163e-05,
+      "loss": 0.7264,
+      "num_input_tokens_seen": 9297152,
+      "step": 1735
+    },
+    {
+      "epoch": 0.2784,
+      "grad_norm": 0.8012672066688538,
+      "learning_rate": 4.1032223570202474e-05,
+      "loss": 0.7368,
+      "num_input_tokens_seen": 9326896,
+      "step": 1740
+    },
+    {
+      "epoch": 0.2792,
+      "grad_norm": 0.8711723685264587,
+      "learning_rate": 4.0983962138655873e-05,
+      "loss": 0.6245,
+      "num_input_tokens_seen": 9351680,
+      "step": 1745
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 1.034636378288269,
+      "learning_rate": 4.093559974371725e-05,
+      "loss": 0.8033,
+      "num_input_tokens_seen": 9374896,
+      "step": 1750
+    },
+    {
+      "epoch": 0.2808,
+      "grad_norm": 0.8999419808387756,
+      "learning_rate": 4.088713669086977e-05,
+      "loss": 0.6803,
+      "num_input_tokens_seen": 9400592,
+      "step": 1755
+    },
+    {
+      "epoch": 0.2816,
+      "grad_norm": 0.5961094498634338,
+      "learning_rate": 4.083857328623243e-05,
+      "loss": 0.7384,
+      "num_input_tokens_seen": 9429280,
+      "step": 1760
+    },
+    {
+      "epoch": 0.2824,
+      "grad_norm": 1.194028377532959,
+      "learning_rate": 4.078990983655807e-05,
+      "loss": 0.8149,
+      "num_input_tokens_seen": 9454736,
+      "step": 1765
+    },
+    {
+      "epoch": 0.2832,
+      "grad_norm": 0.904292643070221,
+      "learning_rate": 4.0741146649231504e-05,
+      "loss": 0.7243,
+      "num_input_tokens_seen": 9479648,
+      "step": 1770
+    },
+    {
+      "epoch": 0.284,
+      "grad_norm": 0.8501243591308594,
+      "learning_rate": 4.0692284032267516e-05,
+      "loss": 0.7639,
+      "num_input_tokens_seen": 9504432,
+      "step": 1775
+    },
+    {
+      "epoch": 0.2848,
+      "grad_norm": 1.0718458890914917,
+      "learning_rate": 4.064332229430895e-05,
+      "loss": 0.6857,
+      "num_input_tokens_seen": 9528880,
+      "step": 1780
+    },
+    {
+      "epoch": 0.2856,
+      "grad_norm": 0.7065584063529968,
+      "learning_rate": 4.059426174462476e-05,
+      "loss": 0.69,
+      "num_input_tokens_seen": 9557360,
+      "step": 1785
+    },
+    {
+      "epoch": 0.2864,
+      "grad_norm": 1.0800750255584717,
+      "learning_rate": 4.054510269310803e-05,
+      "loss": 0.704,
+      "num_input_tokens_seen": 9580608,
+      "step": 1790
+    },
+    {
+      "epoch": 0.2872,
+      "grad_norm": 0.5907096862792969,
+      "learning_rate": 4.0495845450274064e-05,
+      "loss": 0.8015,
+      "num_input_tokens_seen": 9611376,
+      "step": 1795
+    },
+    {
+      "epoch": 0.288,
+      "grad_norm": 0.9455146789550781,
+      "learning_rate": 4.044649032725836e-05,
+      "loss": 0.7382,
+      "num_input_tokens_seen": 9640784,
+      "step": 1800
+    },
+    {
+      "epoch": 0.2888,
+      "grad_norm": 1.0408939123153687,
+      "learning_rate": 4.039703763581472e-05,
+      "loss": 0.7299,
+      "num_input_tokens_seen": 9667120,
+      "step": 1805
+    },
+    {
+      "epoch": 0.2896,
+      "grad_norm": 0.8098856806755066,
+      "learning_rate": 4.0347487688313194e-05,
+      "loss": 0.6402,
+      "num_input_tokens_seen": 9696832,
+      "step": 1810
+    },
+    {
+      "epoch": 0.2904,
+      "grad_norm": 0.695599377155304,
+      "learning_rate": 4.02978407977382e-05,
+      "loss": 0.711,
+      "num_input_tokens_seen": 9722080,
+      "step": 1815
+    },
+    {
+      "epoch": 0.2912,
+      "grad_norm": 0.6605217456817627,
+      "learning_rate": 4.024809727768648e-05,
+      "loss": 0.6587,
+      "num_input_tokens_seen": 9748096,
+      "step": 1820
+    },
+    {
+      "epoch": 0.292,
+      "grad_norm": 0.9249849915504456,
+      "learning_rate": 4.019825744236514e-05,
+      "loss": 0.6656,
+      "num_input_tokens_seen": 9774128,
+      "step": 1825
+    },
+    {
+      "epoch": 0.2928,
+      "grad_norm": 0.8226694464683533,
+      "learning_rate": 4.0148321606589656e-05,
+      "loss": 0.7143,
+      "num_input_tokens_seen": 9805488,
+      "step": 1830
+    },
+    {
+      "epoch": 0.2936,
+      "grad_norm": 1.0425550937652588,
+      "learning_rate": 4.009829008578192e-05,
+      "loss": 0.6735,
+      "num_input_tokens_seen": 9828480,
+      "step": 1835
+    },
+    {
+      "epoch": 0.2944,
+      "grad_norm": 0.6911535263061523,
+      "learning_rate": 4.0048163195968214e-05,
+      "loss": 0.7395,
+      "num_input_tokens_seen": 9863648,
+      "step": 1840
+    },
+    {
+      "epoch": 0.2952,
+      "grad_norm": 0.8600900769233704,
+      "learning_rate": 3.999794125377721e-05,
+      "loss": 0.729,
+      "num_input_tokens_seen": 9893184,
+      "step": 1845
+    },
+    {
+      "epoch": 0.296,
+      "grad_norm": 1.009696364402771,
+      "learning_rate": 3.9947624576437975e-05,
+      "loss": 0.6565,
+      "num_input_tokens_seen": 9922464,
+      "step": 1850
+    },
+    {
+      "epoch": 0.2968,
+      "grad_norm": 0.916327178478241,
+      "learning_rate": 3.9897213481778006e-05,
+      "loss": 0.691,
+      "num_input_tokens_seen": 9948384,
+      "step": 1855
+    },
+    {
+      "epoch": 0.2976,
+      "grad_norm": 0.9392701387405396,
+      "learning_rate": 3.984670828822118e-05,
+      "loss": 0.7408,
+      "num_input_tokens_seen": 9973760,
+      "step": 1860
+    },
+    {
+      "epoch": 0.2984,
+      "grad_norm": 0.9044517278671265,
+      "learning_rate": 3.979610931478574e-05,
+      "loss": 0.761,
+      "num_input_tokens_seen": 10001648,
+      "step": 1865
+    },
+    {
+      "epoch": 0.2992,
+      "grad_norm": 0.9471223950386047,
+      "learning_rate": 3.97454168810823e-05,
+      "loss": 0.8524,
+      "num_input_tokens_seen": 10024912,
+      "step": 1870
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 1.0985262393951416,
+      "learning_rate": 3.969463130731183e-05,
+      "loss": 0.7221,
+      "num_input_tokens_seen": 10049872,
+      "step": 1875
+    },
+    {
+      "epoch": 0.3008,
+      "grad_norm": 0.8284273147583008,
+      "learning_rate": 3.964375291426361e-05,
+      "loss": 0.7708,
+      "num_input_tokens_seen": 10073568,
+      "step": 1880
+    },
+    {
+      "epoch": 0.3016,
+      "grad_norm": 0.7012784481048584,
+      "learning_rate": 3.959278202331322e-05,
+      "loss": 0.6842,
+      "num_input_tokens_seen": 10098448,
+      "step": 1885
+    },
+    {
+      "epoch": 0.3024,
+      "grad_norm": 1.1056398153305054,
+      "learning_rate": 3.954171895642052e-05,
+      "loss": 0.772,
+      "num_input_tokens_seen": 10123168,
+      "step": 1890
+    },
+    {
+      "epoch": 0.3032,
+      "grad_norm": 1.0128076076507568,
+      "learning_rate": 3.949056403612758e-05,
+      "loss": 0.6993,
+      "num_input_tokens_seen": 10149440,
+      "step": 1895
+    },
+    {
+      "epoch": 0.304,
+      "grad_norm": 0.7793564796447754,
+      "learning_rate": 3.943931758555669e-05,
+      "loss": 0.7672,
+      "num_input_tokens_seen": 10174496,
+      "step": 1900
+    },
+    {
+      "epoch": 0.3048,
+      "grad_norm": 0.909677267074585,
+      "learning_rate": 3.938797992840828e-05,
+      "loss": 0.6716,
+      "num_input_tokens_seen": 10199648,
+      "step": 1905
+    },
+    {
+      "epoch": 0.3056,
+      "grad_norm": 0.8851680159568787,
+      "learning_rate": 3.933655138895889e-05,
+      "loss": 0.7062,
+      "num_input_tokens_seen": 10221840,
+      "step": 1910
+    },
+    {
+      "epoch": 0.3064,
+      "grad_norm": 0.9452556371688843,
+      "learning_rate": 3.928503229205913e-05,
+      "loss": 0.6748,
+      "num_input_tokens_seen": 10247504,
+      "step": 1915
+    },
+    {
+      "epoch": 0.3072,
+      "grad_norm": 0.8891339302062988,
+      "learning_rate": 3.9233422963131616e-05,
+      "loss": 0.6331,
+      "num_input_tokens_seen": 10277984,
+      "step": 1920
+    },
+    {
+      "epoch": 0.308,
+      "grad_norm": 0.9662081599235535,
+      "learning_rate": 3.9181723728168916e-05,
+      "loss": 0.779,
+      "num_input_tokens_seen": 10300400,
+      "step": 1925
+    },
+    {
+      "epoch": 0.3088,
+      "grad_norm": 0.9517924785614014,
+      "learning_rate": 3.91299349137315e-05,
+      "loss": 0.722,
+      "num_input_tokens_seen": 10326672,
+      "step": 1930
+    },
+    {
+      "epoch": 0.3096,
+      "grad_norm": 0.755901575088501,
+      "learning_rate": 3.907805684694566e-05,
+      "loss": 0.6321,
+      "num_input_tokens_seen": 10356864,
+      "step": 1935
+    },
+    {
+      "epoch": 0.3104,
+      "grad_norm": 0.8272456526756287,
+      "learning_rate": 3.902608985550147e-05,
+      "loss": 0.6077,
+      "num_input_tokens_seen": 10388032,
+      "step": 1940
+    },
+    {
+      "epoch": 0.3112,
+      "grad_norm": 1.138036847114563,
+      "learning_rate": 3.897403426765069e-05,
+      "loss": 0.6726,
+      "num_input_tokens_seen": 10417152,
+      "step": 1945
+    },
+    {
+      "epoch": 0.312,
+      "grad_norm": 0.8155280351638794,
+      "learning_rate": 3.8921890412204705e-05,
+      "loss": 0.741,
+      "num_input_tokens_seen": 10448128,
+      "step": 1950
+    },
+    {
+      "epoch": 0.3128,
+      "grad_norm": 0.7004032135009766,
+      "learning_rate": 3.886965861853244e-05,
+      "loss": 0.6555,
+      "num_input_tokens_seen": 10474960,
+      "step": 1955
+    },
+    {
+      "epoch": 0.3136,
+      "grad_norm": 0.9554680585861206,
+      "learning_rate": 3.881733921655829e-05,
+      "loss": 0.75,
+      "num_input_tokens_seen": 10502848,
+      "step": 1960
+    },
+    {
+      "epoch": 0.3144,
+      "grad_norm": 0.8525771498680115,
+      "learning_rate": 3.876493253676004e-05,
+      "loss": 0.7042,
+      "num_input_tokens_seen": 10532640,
+      "step": 1965
+    },
+    {
+      "epoch": 0.3152,
+      "grad_norm": 0.8739621043205261,
+      "learning_rate": 3.871243891016676e-05,
+      "loss": 0.6188,
+      "num_input_tokens_seen": 10560096,
+      "step": 1970
+    },
+    {
+      "epoch": 0.316,
+      "grad_norm": 0.9146223068237305,
+      "learning_rate": 3.865985866835673e-05,
+      "loss": 0.8165,
+      "num_input_tokens_seen": 10585520,
+      "step": 1975
+    },
+    {
+      "epoch": 0.3168,
+      "grad_norm": 1.1149648427963257,
+      "learning_rate": 3.8607192143455326e-05,
+      "loss": 0.7437,
+      "num_input_tokens_seen": 10614560,
+      "step": 1980
+    },
+    {
+      "epoch": 0.3176,
+      "grad_norm": 0.9382626414299011,
+      "learning_rate": 3.8554439668132946e-05,
+      "loss": 0.7758,
+      "num_input_tokens_seen": 10637344,
+      "step": 1985
+    },
+    {
+      "epoch": 0.3184,
+      "grad_norm": 0.9469596743583679,
+      "learning_rate": 3.85016015756029e-05,
+      "loss": 0.7593,
+      "num_input_tokens_seen": 10663440,
+      "step": 1990
+    },
+    {
+      "epoch": 0.3192,
+      "grad_norm": 0.8701977133750916,
+      "learning_rate": 3.844867819961928e-05,
+      "loss": 0.6535,
+      "num_input_tokens_seen": 10693392,
+      "step": 1995
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.7110251188278198,
+      "learning_rate": 3.8395669874474915e-05,
+      "loss": 0.8263,
+      "num_input_tokens_seen": 10719232,
+      "step": 2000
+    },
+    {
+      "epoch": 0.3208,
+      "grad_norm": 0.8518005609512329,
+      "learning_rate": 3.8342576934999184e-05,
+      "loss": 0.7992,
+      "num_input_tokens_seen": 10746560,
+      "step": 2005
+    },
+    {
+      "epoch": 0.3216,
+      "grad_norm": 0.9604689478874207,
+      "learning_rate": 3.828939971655595e-05,
+      "loss": 0.7513,
+      "num_input_tokens_seen": 10768512,
+      "step": 2010
+    },
+    {
+      "epoch": 0.3224,
+      "grad_norm": 0.8639784455299377,
+      "learning_rate": 3.8236138555041434e-05,
+      "loss": 0.6775,
+      "num_input_tokens_seen": 10803648,
+      "step": 2015
+    },
+    {
+      "epoch": 0.3232,
+      "grad_norm": 0.8527281880378723,
+      "learning_rate": 3.8182793786882065e-05,
+      "loss": 0.7856,
+      "num_input_tokens_seen": 10830640,
+      "step": 2020
+    },
+    {
+      "epoch": 0.324,
+      "grad_norm": 0.7717742919921875,
+      "learning_rate": 3.81293657490324e-05,
+      "loss": 0.6793,
+      "num_input_tokens_seen": 10860272,
+      "step": 2025
+    },
+    {
+      "epoch": 0.3248,
+      "grad_norm": 0.6685821413993835,
+      "learning_rate": 3.8075854778972955e-05,
+      "loss": 0.6546,
+      "num_input_tokens_seen": 10887664,
+      "step": 2030
+    },
+    {
+      "epoch": 0.3256,
+      "grad_norm": 0.9813340306282043,
+      "learning_rate": 3.802226121470811e-05,
+      "loss": 0.6673,
+      "num_input_tokens_seen": 10912000,
+      "step": 2035
+    },
+    {
+      "epoch": 0.3264,
+      "grad_norm": 1.0419212579727173,
+      "learning_rate": 3.796858539476394e-05,
+      "loss": 0.6933,
+      "num_input_tokens_seen": 10936704,
+      "step": 2040
+    },
+    {
+      "epoch": 0.3272,
+      "grad_norm": 0.851434588432312,
+      "learning_rate": 3.7914827658186103e-05,
+      "loss": 0.6593,
+      "num_input_tokens_seen": 10960464,
+      "step": 2045
+    },
+    {
+      "epoch": 0.328,
+      "grad_norm": 0.7272098660469055,
+      "learning_rate": 3.786098834453766e-05,
+      "loss": 0.6246,
+      "num_input_tokens_seen": 10989680,
+      "step": 2050
+    },
+    {
+      "epoch": 0.3288,
+      "grad_norm": 0.7740225791931152,
+      "learning_rate": 3.780706779389701e-05,
+      "loss": 0.7029,
+      "num_input_tokens_seen": 11014928,
+      "step": 2055
+    },
+    {
+      "epoch": 0.3296,
+      "grad_norm": 0.963455080986023,
+      "learning_rate": 3.775306634685562e-05,
+      "loss": 0.7331,
+      "num_input_tokens_seen": 11041920,
+      "step": 2060
+    },
+    {
+      "epoch": 0.3304,
+      "grad_norm": 0.7765479683876038,
+      "learning_rate": 3.7698984344515997e-05,
+      "loss": 0.6624,
+      "num_input_tokens_seen": 11070304,
+      "step": 2065
+    },
+    {
+      "epoch": 0.3312,
+      "grad_norm": 0.8283601999282837,
+      "learning_rate": 3.764482212848948e-05,
+      "loss": 0.7505,
+      "num_input_tokens_seen": 11099520,
+      "step": 2070
+    },
+    {
+      "epoch": 0.332,
+      "grad_norm": 0.5610854029655457,
+      "learning_rate": 3.759058004089402e-05,
+      "loss": 0.6908,
+      "num_input_tokens_seen": 11129008,
+      "step": 2075
+    },
+    {
+      "epoch": 0.3328,
+      "grad_norm": 0.8462053537368774,
+      "learning_rate": 3.753625842435216e-05,
+      "loss": 0.7062,
+      "num_input_tokens_seen": 11151600,
+      "step": 2080
+    },
+    {
+      "epoch": 0.3336,
+      "grad_norm": 0.8926122188568115,
+      "learning_rate": 3.748185762198873e-05,
+      "loss": 0.7177,
+      "num_input_tokens_seen": 11176784,
+      "step": 2085
+    },
+    {
+      "epoch": 0.3344,
+      "grad_norm": 0.6711943745613098,
+      "learning_rate": 3.742737797742878e-05,
+      "loss": 0.7504,
+      "num_input_tokens_seen": 11205008,
+      "step": 2090
+    },
+    {
+      "epoch": 0.3352,
+      "grad_norm": 1.014253854751587,
+      "learning_rate": 3.7372819834795335e-05,
+      "loss": 0.7144,
+      "num_input_tokens_seen": 11229872,
+      "step": 2095
+    },
+    {
+      "epoch": 0.336,
+      "grad_norm": 0.7249706983566284,
+      "learning_rate": 3.731818353870729e-05,
+      "loss": 0.6876,
+      "num_input_tokens_seen": 11253296,
+      "step": 2100
+    },
+    {
+      "epoch": 0.3368,
+      "grad_norm": 0.8249915838241577,
+      "learning_rate": 3.726346943427719e-05,
+      "loss": 0.7102,
+      "num_input_tokens_seen": 11279408,
+      "step": 2105
+    },
+    {
+      "epoch": 0.3376,
+      "grad_norm": 1.027541995048523,
+      "learning_rate": 3.720867786710904e-05,
+      "loss": 0.7708,
+      "num_input_tokens_seen": 11304176,
+      "step": 2110
+    },
+    {
+      "epoch": 0.3384,
+      "grad_norm": 0.7004812955856323,
+      "learning_rate": 3.7153809183296176e-05,
+      "loss": 0.5882,
+      "num_input_tokens_seen": 11330944,
+      "step": 2115
+    },
+    {
+      "epoch": 0.3392,
+      "grad_norm": 1.1122636795043945,
+      "learning_rate": 3.7098863729419e-05,
+      "loss": 0.6127,
+      "num_input_tokens_seen": 11354064,
+      "step": 2120
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.925553560256958,
+      "learning_rate": 3.704384185254288e-05,
+      "loss": 0.7732,
+      "num_input_tokens_seen": 11376288,
+      "step": 2125
+    },
+    {
+      "epoch": 0.3408,
+      "grad_norm": 0.6940233707427979,
+      "learning_rate": 3.6988743900215894e-05,
+      "loss": 0.7334,
+      "num_input_tokens_seen": 11405472,
+      "step": 2130
+    },
+    {
+      "epoch": 0.3416,
+      "grad_norm": 0.7634669542312622,
+      "learning_rate": 3.693357022046665e-05,
+      "loss": 0.8137,
+      "num_input_tokens_seen": 11431552,
+      "step": 2135
+    },
+    {
+      "epoch": 0.3424,
+      "grad_norm": 0.804530680179596,
+      "learning_rate": 3.68783211618021e-05,
+      "loss": 0.6987,
+      "num_input_tokens_seen": 11459152,
+      "step": 2140
+    },
+    {
+      "epoch": 0.3432,
+      "grad_norm": 1.1058536767959595,
+      "learning_rate": 3.682299707320532e-05,
+      "loss": 0.6614,
+      "num_input_tokens_seen": 11487552,
+      "step": 2145
+    },
+    {
+      "epoch": 0.344,
+      "grad_norm": 0.6808910369873047,
+      "learning_rate": 3.6767598304133324e-05,
+      "loss": 0.688,
+      "num_input_tokens_seen": 11515792,
+      "step": 2150
+    },
+    {
+      "epoch": 0.3448,
+      "grad_norm": 1.0619826316833496,
+      "learning_rate": 3.671212520451484e-05,
+      "loss": 0.7897,
+      "num_input_tokens_seen": 11541280,
+      "step": 2155
+    },
+    {
+      "epoch": 0.3456,
+      "grad_norm": 0.8404290080070496,
+      "learning_rate": 3.665657812474812e-05,
+      "loss": 0.7086,
+      "num_input_tokens_seen": 11569440,
+      "step": 2160
+    },
+    {
+      "epoch": 0.3464,
+      "grad_norm": 1.316372036933899,
+      "learning_rate": 3.660095741569871e-05,
+      "loss": 0.7421,
+      "num_input_tokens_seen": 11597792,
+      "step": 2165
+    },
+    {
+      "epoch": 0.3472,
+      "grad_norm": 0.7798539400100708,
+      "learning_rate": 3.654526342869724e-05,
+      "loss": 0.6954,
+      "num_input_tokens_seen": 11622864,
+      "step": 2170
+    },
+    {
+      "epoch": 0.348,
+      "grad_norm": 0.7101672887802124,
+      "learning_rate": 3.6489496515537204e-05,
+      "loss": 0.6764,
+      "num_input_tokens_seen": 11651280,
+      "step": 2175
+    },
+    {
+      "epoch": 0.3488,
+      "grad_norm": 0.8456715941429138,
+      "learning_rate": 3.643365702847272e-05,
+      "loss": 0.705,
+      "num_input_tokens_seen": 11680048,
+      "step": 2180
+    },
+    {
+      "epoch": 0.3496,
+      "grad_norm": 0.9790185689926147,
+      "learning_rate": 3.6377745320216346e-05,
+      "loss": 0.7433,
+      "num_input_tokens_seen": 11702144,
+      "step": 2185
+    },
+    {
+      "epoch": 0.3504,
+      "grad_norm": 0.9205552935600281,
+      "learning_rate": 3.632176174393682e-05,
+      "loss": 0.653,
+      "num_input_tokens_seen": 11728816,
+      "step": 2190
+    },
+    {
+      "epoch": 0.3512,
+      "grad_norm": 0.8499376177787781,
+      "learning_rate": 3.626570665325684e-05,
+      "loss": 0.6381,
+      "num_input_tokens_seen": 11756688,
+      "step": 2195
+    },
+    {
+      "epoch": 0.352,
+      "grad_norm": 0.7778225541114807,
+      "learning_rate": 3.6209580402250815e-05,
+      "loss": 0.7347,
+      "num_input_tokens_seen": 11781664,
+      "step": 2200
+    },
+    {
+      "epoch": 0.3528,
+      "grad_norm": 0.8913766145706177,
+      "learning_rate": 3.615338334544265e-05,
+      "loss": 0.8036,
+      "num_input_tokens_seen": 11808352,
+      "step": 2205
+    },
+    {
+      "epoch": 0.3536,
+      "grad_norm": 1.0191758871078491,
+      "learning_rate": 3.6097115837803505e-05,
+      "loss": 0.7486,
+      "num_input_tokens_seen": 11836400,
+      "step": 2210
+    },
+    {
+      "epoch": 0.3544,
+      "grad_norm": 0.7858436703681946,
+      "learning_rate": 3.604077823474954e-05,
+      "loss": 0.7885,
+      "num_input_tokens_seen": 11862608,
+      "step": 2215
+    },
+    {
+      "epoch": 0.3552,
+      "grad_norm": 0.6349871158599854,
+      "learning_rate": 3.5984370892139666e-05,
+      "loss": 0.7005,
+      "num_input_tokens_seen": 11886528,
+      "step": 2220
+    },
+    {
+      "epoch": 0.356,
+      "grad_norm": 0.8877844214439392,
+      "learning_rate": 3.592789416627332e-05,
+      "loss": 0.607,
+      "num_input_tokens_seen": 11915040,
+      "step": 2225
+    },
+    {
+      "epoch": 0.3568,
+      "grad_norm": 1.1504970788955688,
+      "learning_rate": 3.5871348413888204e-05,
+      "loss": 0.6723,
+      "num_input_tokens_seen": 11942768,
+      "step": 2230
+    },
+    {
+      "epoch": 0.3576,
+      "grad_norm": 0.7394466400146484,
+      "learning_rate": 3.581473399215802e-05,
+      "loss": 0.7302,
+      "num_input_tokens_seen": 11978464,
+      "step": 2235
+    },
+    {
+      "epoch": 0.3584,
+      "grad_norm": 1.1570250988006592,
+      "learning_rate": 3.575805125869022e-05,
+      "loss": 0.6867,
+      "num_input_tokens_seen": 12001392,
+      "step": 2240
+    },
+    {
+      "epoch": 0.3592,
+      "grad_norm": 0.8141620755195618,
+      "learning_rate": 3.5701300571523755e-05,
+      "loss": 0.7346,
+      "num_input_tokens_seen": 12030352,
+      "step": 2245
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.8653257489204407,
+      "learning_rate": 3.564448228912682e-05,
+      "loss": 0.6381,
+      "num_input_tokens_seen": 12062384,
+      "step": 2250
+    },
+    {
+      "epoch": 0.3608,
+      "grad_norm": 0.8065868020057678,
+      "learning_rate": 3.558759677039455e-05,
+      "loss": 0.7679,
+      "num_input_tokens_seen": 12089408,
+      "step": 2255
+    },
+    {
+      "epoch": 0.3616,
+      "grad_norm": 0.7610428929328918,
+      "learning_rate": 3.5530644374646815e-05,
+      "loss": 0.668,
+      "num_input_tokens_seen": 12114656,
+      "step": 2260
+    },
+    {
+      "epoch": 0.3624,
+      "grad_norm": 0.8063391447067261,
+      "learning_rate": 3.547362546162588e-05,
+      "loss": 0.7454,
+      "num_input_tokens_seen": 12144832,
+      "step": 2265
+    },
+    {
+      "epoch": 0.3632,
+      "grad_norm": 1.0300970077514648,
+      "learning_rate": 3.54165403914942e-05,
+      "loss": 0.7513,
+      "num_input_tokens_seen": 12170096,
+      "step": 2270
+    },
+    {
+      "epoch": 0.364,
+      "grad_norm": 1.1293412446975708,
+      "learning_rate": 3.535938952483211e-05,
+      "loss": 0.7881,
+      "num_input_tokens_seen": 12191104,
+      "step": 2275
+    },
+    {
+      "epoch": 0.3648,
+      "grad_norm": 0.8911874294281006,
+      "learning_rate": 3.5302173222635524e-05,
+      "loss": 0.7253,
+      "num_input_tokens_seen": 12214416,
+      "step": 2280
+    },
+    {
+      "epoch": 0.3656,
+      "grad_norm": 1.0665303468704224,
+      "learning_rate": 3.5244891846313736e-05,
+      "loss": 0.8122,
+      "num_input_tokens_seen": 12241344,
+      "step": 2285
+    },
+    {
+      "epoch": 0.3664,
+      "grad_norm": 0.6204916834831238,
+      "learning_rate": 3.5187545757687015e-05,
+      "loss": 0.6188,
+      "num_input_tokens_seen": 12269376,
+      "step": 2290
+    },
+    {
+      "epoch": 0.3672,
+      "grad_norm": 0.7871102094650269,
+      "learning_rate": 3.5130135318984456e-05,
+      "loss": 0.7138,
+      "num_input_tokens_seen": 12294960,
+      "step": 2295
+    },
+    {
+      "epoch": 0.368,
+      "grad_norm": 0.7584692239761353,
+      "learning_rate": 3.507266089284157e-05,
+      "loss": 0.7425,
+      "num_input_tokens_seen": 12318864,
+      "step": 2300
+    },
+    {
+      "epoch": 0.3688,
+      "grad_norm": 0.6678550839424133,
+      "learning_rate": 3.501512284229807e-05,
+      "loss": 0.7238,
+      "num_input_tokens_seen": 12345520,
+      "step": 2305
+    },
+    {
+      "epoch": 0.3696,
+      "grad_norm": 0.9825206398963928,
+      "learning_rate": 3.495752153079557e-05,
+      "loss": 0.684,
+      "num_input_tokens_seen": 12369600,
+      "step": 2310
+    },
+    {
+      "epoch": 0.3704,
+      "grad_norm": 0.8038123250007629,
+      "learning_rate": 3.489985732217525e-05,
+      "loss": 0.707,
+      "num_input_tokens_seen": 12394400,
+      "step": 2315
+    },
+    {
+      "epoch": 0.3712,
+      "grad_norm": 1.158873438835144,
+      "learning_rate": 3.484213058067559e-05,
+      "loss": 0.5843,
+      "num_input_tokens_seen": 12420848,
+      "step": 2320
+    },
+    {
+      "epoch": 0.372,
+      "grad_norm": 0.8114385604858398,
+      "learning_rate": 3.4784341670930065e-05,
+      "loss": 0.7014,
+      "num_input_tokens_seen": 12446192,
+      "step": 2325
+    },
+    {
+      "epoch": 0.3728,
+      "grad_norm": 0.8132364749908447,
+      "learning_rate": 3.4726490957964834e-05,
+      "loss": 0.777,
+      "num_input_tokens_seen": 12472960,
+      "step": 2330
+    },
+    {
+      "epoch": 0.3736,
+      "grad_norm": 0.7918152213096619,
+      "learning_rate": 3.466857880719645e-05,
+      "loss": 0.6856,
+      "num_input_tokens_seen": 12504256,
+      "step": 2335
+    },
+    {
+      "epoch": 0.3744,
+      "grad_norm": 0.8399984240531921,
+      "learning_rate": 3.461060558442952e-05,
+      "loss": 0.7742,
+      "num_input_tokens_seen": 12529872,
+      "step": 2340
+    },
+    {
+      "epoch": 0.3752,
+      "grad_norm": 1.0398231744766235,
+      "learning_rate": 3.455257165585444e-05,
+      "loss": 0.6815,
+      "num_input_tokens_seen": 12552368,
+      "step": 2345
+    },
+    {
+      "epoch": 0.376,
+      "grad_norm": 0.9708042144775391,
+      "learning_rate": 3.4494477388045035e-05,
+      "loss": 0.677,
+      "num_input_tokens_seen": 12576720,
+      "step": 2350
+    },
+    {
+      "epoch": 0.3768,
+      "grad_norm": 0.8928380012512207,
+      "learning_rate": 3.443632314795627e-05,
+      "loss": 0.6239,
+      "num_input_tokens_seen": 12606096,
+      "step": 2355
+    },
+    {
+      "epoch": 0.3776,
+      "grad_norm": 1.3437156677246094,
+      "learning_rate": 3.437810930292195e-05,
+      "loss": 0.7379,
+      "num_input_tokens_seen": 12631376,
+      "step": 2360
+    },
+    {
+      "epoch": 0.3784,
+      "grad_norm": 0.9309334754943848,
+      "learning_rate": 3.4319836220652335e-05,
+      "loss": 0.7315,
+      "num_input_tokens_seen": 12662096,
+      "step": 2365
+    },
+    {
+      "epoch": 0.3792,
+      "grad_norm": 1.4636520147323608,
+      "learning_rate": 3.4261504269231904e-05,
+      "loss": 0.7738,
+      "num_input_tokens_seen": 12691696,
+      "step": 2370
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.8436228632926941,
+      "learning_rate": 3.4203113817116957e-05,
+      "loss": 0.7307,
+      "num_input_tokens_seen": 12718368,
+      "step": 2375
+    },
+    {
+      "epoch": 0.3808,
+      "grad_norm": 0.877709150314331,
+      "learning_rate": 3.414466523313332e-05,
+      "loss": 0.7119,
+      "num_input_tokens_seen": 12743664,
+      "step": 2380
+    },
+    {
+      "epoch": 0.3816,
+      "grad_norm": 1.2288016080856323,
+      "learning_rate": 3.408615888647402e-05,
+      "loss": 0.781,
+      "num_input_tokens_seen": 12775088,
+      "step": 2385
+    },
+    {
+      "epoch": 0.3824,
+      "grad_norm": 0.8335594534873962,
+      "learning_rate": 3.402759514669694e-05,
+      "loss": 0.6256,
+      "num_input_tokens_seen": 12802576,
+      "step": 2390
+    },
+    {
+      "epoch": 0.3832,
+      "grad_norm": 1.0417710542678833,
+      "learning_rate": 3.3968974383722495e-05,
+      "loss": 0.7672,
+      "num_input_tokens_seen": 12831280,
+      "step": 2395
+    },
+    {
+      "epoch": 0.384,
+      "grad_norm": 1.1079373359680176,
+      "learning_rate": 3.3910296967831266e-05,
+      "loss": 0.7665,
+      "num_input_tokens_seen": 12853744,
+      "step": 2400
+    },
+    {
+      "epoch": 0.3848,
+      "grad_norm": 0.870614230632782,
+      "learning_rate": 3.3851563269661726e-05,
+      "loss": 0.6321,
+      "num_input_tokens_seen": 12883408,
+      "step": 2405
+    },
+    {
+      "epoch": 0.3856,
+      "grad_norm": 1.090280294418335,
+      "learning_rate": 3.379277366020782e-05,
+      "loss": 0.7086,
+      "num_input_tokens_seen": 12914592,
+      "step": 2410
+    },
+    {
+      "epoch": 0.3864,
+      "grad_norm": 0.8816367983818054,
+      "learning_rate": 3.373392851081668e-05,
+      "loss": 0.7712,
+      "num_input_tokens_seen": 12936832,
+      "step": 2415
+    },
+    {
+      "epoch": 0.3872,
+      "grad_norm": 0.8722823858261108,
+      "learning_rate": 3.367502819318624e-05,
+      "loss": 0.6844,
+      "num_input_tokens_seen": 12962864,
+      "step": 2420
+    },
+    {
+      "epoch": 0.388,
+      "grad_norm": 0.9704541563987732,
+      "learning_rate": 3.3616073079362926e-05,
+      "loss": 0.6877,
+      "num_input_tokens_seen": 12992560,
+      "step": 2425
+    },
+    {
+      "epoch": 0.3888,
+      "grad_norm": 0.8094004988670349,
+      "learning_rate": 3.355706354173928e-05,
+      "loss": 0.8139,
+      "num_input_tokens_seen": 13015440,
+      "step": 2430
+    },
+    {
+      "epoch": 0.3896,
+      "grad_norm": 0.8286037445068359,
+      "learning_rate": 3.349799995305162e-05,
+      "loss": 0.6696,
+      "num_input_tokens_seen": 13039008,
+      "step": 2435
+    },
+    {
+      "epoch": 0.3904,
+      "grad_norm": 0.985637366771698,
+      "learning_rate": 3.343888268637765e-05,
+      "loss": 0.7001,
+      "num_input_tokens_seen": 13067648,
+      "step": 2440
+    },
+    {
+      "epoch": 0.3912,
+      "grad_norm": 0.8938013911247253,
+      "learning_rate": 3.337971211513417e-05,
+      "loss": 0.8036,
+      "num_input_tokens_seen": 13090064,
+      "step": 2445
+    },
+    {
+      "epoch": 0.392,
+      "grad_norm": 0.7293727397918701,
+      "learning_rate": 3.332048861307467e-05,
+      "loss": 0.7405,
+      "num_input_tokens_seen": 13119856,
+      "step": 2450
+    },
+    {
+      "epoch": 0.3928,
+      "grad_norm": 0.5999038219451904,
+      "learning_rate": 3.3261212554286975e-05,
+      "loss": 0.6975,
+      "num_input_tokens_seen": 13148288,
+      "step": 2455
+    },
+    {
+      "epoch": 0.3936,
+      "grad_norm": 0.8091318607330322,
+      "learning_rate": 3.320188431319088e-05,
+      "loss": 0.6809,
+      "num_input_tokens_seen": 13175616,
+      "step": 2460
+    },
+    {
+      "epoch": 0.3944,
+      "grad_norm": 1.0293824672698975,
+      "learning_rate": 3.3142504264535804e-05,
+      "loss": 0.7749,
+      "num_input_tokens_seen": 13199280,
+      "step": 2465
+    },
+    {
+      "epoch": 0.3952,
+      "grad_norm": 0.793485701084137,
+      "learning_rate": 3.3083072783398416e-05,
+      "loss": 0.6873,
+      "num_input_tokens_seen": 13224640,
+      "step": 2470
+    },
+    {
+      "epoch": 0.396,
+      "grad_norm": 0.8636240363121033,
+      "learning_rate": 3.302359024518024e-05,
+      "loss": 0.7554,
+      "num_input_tokens_seen": 13250448,
+      "step": 2475
+    },
+    {
+      "epoch": 0.3968,
+      "grad_norm": 0.9471914172172546,
+      "learning_rate": 3.296405702560532e-05,
+      "loss": 0.7112,
+      "num_input_tokens_seen": 13273472,
+      "step": 2480
+    },
+    {
+      "epoch": 0.3976,
+      "grad_norm": 1.1579172611236572,
+      "learning_rate": 3.2904473500717824e-05,
+      "loss": 0.8207,
+      "num_input_tokens_seen": 13300608,
+      "step": 2485
+    },
+    {
+      "epoch": 0.3984,
+      "grad_norm": 1.022197961807251,
+      "learning_rate": 3.2844840046879686e-05,
+      "loss": 0.693,
+      "num_input_tokens_seen": 13326976,
+      "step": 2490
+    },
+    {
+      "epoch": 0.3992,
+      "grad_norm": 0.7574387788772583,
+      "learning_rate": 3.278515704076821e-05,
+      "loss": 0.6826,
+      "num_input_tokens_seen": 13358528,
+      "step": 2495
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.7097072005271912,
+      "learning_rate": 3.272542485937369e-05,
+      "loss": 0.6714,
+      "num_input_tokens_seen": 13384096,
+      "step": 2500
+    },
+    {
+      "epoch": 0.4008,
+      "grad_norm": 0.8780053853988647,
+      "learning_rate": 3.2665643879997056e-05,
+      "loss": 0.7387,
+      "num_input_tokens_seen": 13417120,
+      "step": 2505
+    },
+    {
+      "epoch": 0.4016,
+      "grad_norm": 0.8968010544776917,
+      "learning_rate": 3.260581448024745e-05,
+      "loss": 0.6875,
+      "num_input_tokens_seen": 13444832,
+      "step": 2510
+    },
+    {
+      "epoch": 0.4024,
+      "grad_norm": 0.9647771716117859,
+      "learning_rate": 3.25459370380399e-05,
+      "loss": 0.834,
+      "num_input_tokens_seen": 13472304,
+      "step": 2515
+    },
+    {
+      "epoch": 0.4032,
+      "grad_norm": 0.9738301038742065,
+      "learning_rate": 3.248601193159287e-05,
+      "loss": 0.7144,
+      "num_input_tokens_seen": 13495984,
+      "step": 2520
+    },
+    {
+      "epoch": 0.404,
+      "grad_norm": 1.03775155544281,
+      "learning_rate": 3.2426039539425876e-05,
+      "loss": 0.7171,
+      "num_input_tokens_seen": 13523360,
+      "step": 2525
+    },
+    {
+      "epoch": 0.4048,
+      "grad_norm": 1.3964909315109253,
+      "learning_rate": 3.236602024035716e-05,
+      "loss": 0.7197,
+      "num_input_tokens_seen": 13550016,
+      "step": 2530
+    },
+    {
+      "epoch": 0.4056,
+      "grad_norm": 1.0805152654647827,
+      "learning_rate": 3.230595441350125e-05,
+      "loss": 0.7997,
+      "num_input_tokens_seen": 13575088,
+      "step": 2535
+    },
+    {
+      "epoch": 0.4064,
+      "grad_norm": 0.9613687992095947,
+      "learning_rate": 3.2245842438266526e-05,
+      "loss": 0.7847,
+      "num_input_tokens_seen": 13600832,
+      "step": 2540
+    },
+    {
+      "epoch": 0.4072,
+      "grad_norm": 0.9843304753303528,
+      "learning_rate": 3.2185684694352916e-05,
+      "loss": 0.7213,
+      "num_input_tokens_seen": 13627328,
+      "step": 2545
+    },
+    {
+      "epoch": 0.408,
+      "grad_norm": 0.7906083464622498,
+      "learning_rate": 3.21254815617494e-05,
+      "loss": 0.633,
+      "num_input_tokens_seen": 13651664,
+      "step": 2550
+    },
+    {
+      "epoch": 0.4088,
+      "grad_norm": 0.788149893283844,
+      "learning_rate": 3.206523342073172e-05,
+      "loss": 0.7512,
+      "num_input_tokens_seen": 13677248,
+      "step": 2555
+    },
+    {
+      "epoch": 0.4096,
+      "grad_norm": 0.7680060863494873,
+      "learning_rate": 3.2004940651859844e-05,
+      "loss": 0.703,
+      "num_input_tokens_seen": 13705904,
+      "step": 2560
+    },
+    {
+      "epoch": 0.4104,
+      "grad_norm": 0.8078610301017761,
+      "learning_rate": 3.194460363597569e-05,
+      "loss": 0.7212,
+      "num_input_tokens_seen": 13731520,
+      "step": 2565
+    },
+    {
+      "epoch": 0.4112,
+      "grad_norm": 1.2152231931686401,
+      "learning_rate": 3.1884222754200625e-05,
+      "loss": 0.7009,
+      "num_input_tokens_seen": 13753840,
+      "step": 2570
+    },
+    {
+      "epoch": 0.412,
+      "grad_norm": 0.8687548637390137,
+      "learning_rate": 3.1823798387933134e-05,
+      "loss": 0.718,
+      "num_input_tokens_seen": 13777504,
+      "step": 2575
+    },
+    {
+      "epoch": 0.4128,
+      "grad_norm": 1.1128169298171997,
+      "learning_rate": 3.176333091884635e-05,
+      "loss": 0.6796,
+      "num_input_tokens_seen": 13805392,
+      "step": 2580
+    },
+    {
+      "epoch": 0.4136,
+      "grad_norm": 0.6620244383811951,
+      "learning_rate": 3.170282072888566e-05,
+      "loss": 0.6632,
+      "num_input_tokens_seen": 13835600,
+      "step": 2585
+    },
+    {
+      "epoch": 0.4144,
+      "grad_norm": 1.0803226232528687,
+      "learning_rate": 3.1642268200266317e-05,
+      "loss": 0.743,
+      "num_input_tokens_seen": 13862528,
+      "step": 2590
+    },
+    {
+      "epoch": 0.4152,
+      "grad_norm": 0.8314620852470398,
+      "learning_rate": 3.1581673715471006e-05,
+      "loss": 0.7091,
+      "num_input_tokens_seen": 13890272,
+      "step": 2595
+    },
+    {
+      "epoch": 0.416,
+      "grad_norm": 1.0047166347503662,
+      "learning_rate": 3.152103765724743e-05,
+      "loss": 0.8011,
+      "num_input_tokens_seen": 13913328,
+      "step": 2600
+    },
+    {
+      "epoch": 0.4168,
+      "grad_norm": 0.9856431484222412,
+      "learning_rate": 3.1460360408605866e-05,
+      "loss": 0.7569,
+      "num_input_tokens_seen": 13943040,
+      "step": 2605
+    },
+    {
+      "epoch": 0.4176,
+      "grad_norm": 0.8467027544975281,
+      "learning_rate": 3.139964235281682e-05,
+      "loss": 0.6976,
+      "num_input_tokens_seen": 13971872,
+      "step": 2610
+    },
+    {
+      "epoch": 0.4184,
+      "grad_norm": 1.2195795774459839,
+      "learning_rate": 3.1338883873408516e-05,
+      "loss": 0.7039,
+      "num_input_tokens_seen": 13997456,
+      "step": 2615
+    },
+    {
+      "epoch": 0.4192,
+      "grad_norm": 0.832929253578186,
+      "learning_rate": 3.127808535416454e-05,
+      "loss": 0.7153,
+      "num_input_tokens_seen": 14024656,
+      "step": 2620
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.8261767625808716,
+      "learning_rate": 3.121724717912138e-05,
+      "loss": 0.7317,
+      "num_input_tokens_seen": 14053680,
+      "step": 2625
+    },
+    {
+      "epoch": 0.4208,
+      "grad_norm": 0.8690986633300781,
+      "learning_rate": 3.1156369732566006e-05,
+      "loss": 0.6991,
+      "num_input_tokens_seen": 14080096,
+      "step": 2630
+    },
+    {
+      "epoch": 0.4216,
+      "grad_norm": 1.041561484336853,
+      "learning_rate": 3.1095453399033466e-05,
+      "loss": 0.7442,
+      "num_input_tokens_seen": 14108080,
+      "step": 2635
+    },
+    {
+      "epoch": 0.4224,
+      "grad_norm": 1.1139183044433594,
+      "learning_rate": 3.103449856330443e-05,
+      "loss": 0.7026,
+      "num_input_tokens_seen": 14132448,
+      "step": 2640
+    },
+    {
+      "epoch": 0.4232,
+      "grad_norm": 0.9388411045074463,
+      "learning_rate": 3.0973505610402765e-05,
+      "loss": 0.6425,
+      "num_input_tokens_seen": 14157312,
+      "step": 2645
+    },
+    {
+      "epoch": 0.424,
+      "grad_norm": 0.8923696279525757,
+      "learning_rate": 3.091247492559312e-05,
+      "loss": 0.7421,
+      "num_input_tokens_seen": 14184288,
+      "step": 2650
+    },
+    {
+      "epoch": 0.4248,
+      "grad_norm": 0.9683478474617004,
+      "learning_rate": 3.085140689437846e-05,
+      "loss": 0.7044,
+      "num_input_tokens_seen": 14207920,
+      "step": 2655
+    },
+    {
+      "epoch": 0.4256,
+      "grad_norm": 0.7942652106285095,
+      "learning_rate": 3.0790301902497666e-05,
+      "loss": 0.6892,
+      "num_input_tokens_seen": 14235504,
+      "step": 2660
+    },
+    {
+      "epoch": 0.4264,
+      "grad_norm": 0.9955897331237793,
+      "learning_rate": 3.072916033592307e-05,
+      "loss": 0.6595,
+      "num_input_tokens_seen": 14259280,
+      "step": 2665
+    },
+    {
+      "epoch": 0.4272,
+      "grad_norm": 0.9912785291671753,
+      "learning_rate": 3.0667982580858044e-05,
+      "loss": 0.6948,
+      "num_input_tokens_seen": 14286592,
+      "step": 2670
+    },
+    {
+      "epoch": 0.428,
+      "grad_norm": 1.352742314338684,
+      "learning_rate": 3.0606769023734536e-05,
+      "loss": 0.7009,
+      "num_input_tokens_seen": 14309280,
+      "step": 2675
+    },
+    {
+      "epoch": 0.4288,
+      "grad_norm": 1.183185338973999,
+      "learning_rate": 3.054552005121064e-05,
+      "loss": 0.6814,
+      "num_input_tokens_seen": 14335984,
+      "step": 2680
+    },
+    {
+      "epoch": 0.4296,
+      "grad_norm": 1.2679824829101562,
+      "learning_rate": 3.0484236050168153e-05,
+      "loss": 0.7468,
+      "num_input_tokens_seen": 14361024,
+      "step": 2685
+    },
+    {
+      "epoch": 0.4304,
+      "grad_norm": 1.1353107690811157,
+      "learning_rate": 3.0422917407710137e-05,
+      "loss": 0.629,
+      "num_input_tokens_seen": 14391440,
+      "step": 2690
+    },
+    {
+      "epoch": 0.4312,
+      "grad_norm": 1.1603094339370728,
+      "learning_rate": 3.0361564511158457e-05,
+      "loss": 0.7106,
+      "num_input_tokens_seen": 14417952,
+      "step": 2695
+    },
+    {
+      "epoch": 0.432,
+      "grad_norm": 0.9477285146713257,
+      "learning_rate": 3.0300177748051373e-05,
+      "loss": 0.7136,
+      "num_input_tokens_seen": 14446752,
+      "step": 2700
+    },
+    {
+      "epoch": 0.4328,
+      "grad_norm": 0.9295204281806946,
+      "learning_rate": 3.0238757506141012e-05,
+      "loss": 0.6269,
+      "num_input_tokens_seen": 14475280,
+      "step": 2705
+    },
+    {
+      "epoch": 0.4336,
+      "grad_norm": 0.8617603182792664,
+      "learning_rate": 3.0177304173391037e-05,
+      "loss": 0.6517,
+      "num_input_tokens_seen": 14498112,
+      "step": 2710
+    },
+    {
+      "epoch": 0.4344,
+      "grad_norm": 0.962295413017273,
+      "learning_rate": 3.0115818137974067e-05,
+      "loss": 0.6903,
+      "num_input_tokens_seen": 14525664,
+      "step": 2715
+    },
+    {
+      "epoch": 0.4352,
+      "grad_norm": 0.7317754030227661,
+      "learning_rate": 3.005429978826934e-05,
+      "loss": 0.7302,
+      "num_input_tokens_seen": 14551536,
+      "step": 2720
+    },
+    {
+      "epoch": 0.436,
+      "grad_norm": 0.9604383111000061,
+      "learning_rate": 2.9992749512860173e-05,
+      "loss": 0.7126,
+      "num_input_tokens_seen": 14574560,
+      "step": 2725
+    },
+    {
+      "epoch": 0.4368,
+      "grad_norm": 0.9363977313041687,
+      "learning_rate": 2.9931167700531578e-05,
+      "loss": 0.6701,
+      "num_input_tokens_seen": 14602384,
+      "step": 2730
+    },
+    {
+      "epoch": 0.4376,
+      "grad_norm": 1.0513427257537842,
+      "learning_rate": 2.9869554740267724e-05,
+      "loss": 0.5816,
+      "num_input_tokens_seen": 14633728,
+      "step": 2735
+    },
+    {
+      "epoch": 0.4384,
+      "grad_norm": 1.0142287015914917,
+      "learning_rate": 2.9807911021249573e-05,
+      "loss": 0.7965,
+      "num_input_tokens_seen": 14662752,
+      "step": 2740
+    },
+    {
+      "epoch": 0.4392,
+      "grad_norm": 0.8593106269836426,
+      "learning_rate": 2.9746236932852355e-05,
+      "loss": 0.6396,
+      "num_input_tokens_seen": 14690896,
+      "step": 2745
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.912413477897644,
+      "learning_rate": 2.9684532864643122e-05,
+      "loss": 0.6914,
+      "num_input_tokens_seen": 14717680,
+      "step": 2750
+    },
+    {
+      "epoch": 0.4408,
+      "grad_norm": 1.1753630638122559,
+      "learning_rate": 2.9622799206378305e-05,
+      "loss": 0.7188,
+      "num_input_tokens_seen": 14744176,
+      "step": 2755
+    },
+    {
+      "epoch": 0.4416,
+      "grad_norm": 1.0383411645889282,
+      "learning_rate": 2.956103634800126e-05,
+      "loss": 0.6936,
+      "num_input_tokens_seen": 14772464,
+      "step": 2760
+    },
+    {
+      "epoch": 0.4424,
+      "grad_norm": 0.8875827789306641,
+      "learning_rate": 2.949924467963975e-05,
+      "loss": 0.709,
+      "num_input_tokens_seen": 14800896,
+      "step": 2765
+    },
+    {
+      "epoch": 0.4432,
+      "grad_norm": 1.0359493494033813,
+      "learning_rate": 2.943742459160354e-05,
+      "loss": 0.6361,
+      "num_input_tokens_seen": 14826624,
+      "step": 2770
+    },
+    {
+      "epoch": 0.444,
+      "grad_norm": 0.7070389986038208,
+      "learning_rate": 2.9375576474381905e-05,
+      "loss": 0.6062,
+      "num_input_tokens_seen": 14859392,
+      "step": 2775
+    },
+    {
+      "epoch": 0.4448,
+      "grad_norm": 1.0716419219970703,
+      "learning_rate": 2.9313700718641167e-05,
+      "loss": 0.7882,
+      "num_input_tokens_seen": 14882336,
+      "step": 2780
+    },
+    {
+      "epoch": 0.4456,
+      "grad_norm": 0.8054667115211487,
+      "learning_rate": 2.925179771522223e-05,
+      "loss": 0.7978,
+      "num_input_tokens_seen": 14911312,
+      "step": 2785
+    },
+    {
+      "epoch": 0.4464,
+      "grad_norm": 0.9502078294754028,
+      "learning_rate": 2.9189867855138103e-05,
+      "loss": 0.6835,
+      "num_input_tokens_seen": 14938400,
+      "step": 2790
+    },
+    {
+      "epoch": 0.4472,
+      "grad_norm": 0.8377355933189392,
+      "learning_rate": 2.912791152957145e-05,
+      "loss": 0.6566,
+      "num_input_tokens_seen": 14965424,
+      "step": 2795
+    },
+    {
+      "epoch": 0.448,
+      "grad_norm": 0.8674115538597107,
+      "learning_rate": 2.9065929129872094e-05,
+      "loss": 0.6616,
+      "num_input_tokens_seen": 14994368,
+      "step": 2800
+    },
+    {
+      "epoch": 0.4488,
+      "grad_norm": 0.9967759251594543,
+      "learning_rate": 2.900392104755455e-05,
+      "loss": 0.8051,
+      "num_input_tokens_seen": 15018480,
+      "step": 2805
+    },
+    {
+      "epoch": 0.4496,
+      "grad_norm": 1.04585862159729,
+      "learning_rate": 2.894188767429557e-05,
+      "loss": 0.6961,
+      "num_input_tokens_seen": 15045840,
+      "step": 2810
+    },
+    {
+      "epoch": 0.4504,
+      "grad_norm": 0.8065064549446106,
+      "learning_rate": 2.8879829401931652e-05,
+      "loss": 0.6898,
+      "num_input_tokens_seen": 15070832,
+      "step": 2815
+    },
+    {
+      "epoch": 0.4512,
+      "grad_norm": 0.7077392935752869,
+      "learning_rate": 2.881774662245658e-05,
+      "loss": 0.6789,
+      "num_input_tokens_seen": 15097008,
+      "step": 2820
+    },
+    {
+      "epoch": 0.452,
+      "grad_norm": 1.068467378616333,
+      "learning_rate": 2.875563972801893e-05,
+      "loss": 0.6684,
+      "num_input_tokens_seen": 15120080,
+      "step": 2825
+    },
+    {
+      "epoch": 0.4528,
+      "grad_norm": 0.7860395312309265,
+      "learning_rate": 2.8693509110919598e-05,
+      "loss": 0.6577,
+      "num_input_tokens_seen": 15144976,
+      "step": 2830
+    },
+    {
+      "epoch": 0.4536,
+      "grad_norm": 0.86238032579422,
+      "learning_rate": 2.863135516360932e-05,
+      "loss": 0.7893,
+      "num_input_tokens_seen": 15174640,
+      "step": 2835
+    },
+    {
+      "epoch": 0.4544,
+      "grad_norm": 0.8910583257675171,
+      "learning_rate": 2.856917827868622e-05,
+      "loss": 0.7377,
+      "num_input_tokens_seen": 15198128,
+      "step": 2840
+    },
+    {
+      "epoch": 0.4552,
+      "grad_norm": 0.9576541781425476,
+      "learning_rate": 2.8506978848893302e-05,
+      "loss": 0.821,
+      "num_input_tokens_seen": 15222224,
+      "step": 2845
+    },
+    {
+      "epoch": 0.456,
+      "grad_norm": 1.114388108253479,
+      "learning_rate": 2.844475726711595e-05,
+      "loss": 0.695,
+      "num_input_tokens_seen": 15246640,
+      "step": 2850
+    },
+    {
+      "epoch": 0.4568,
+      "grad_norm": 0.8997412323951721,
+      "learning_rate": 2.8382513926379504e-05,
+      "loss": 0.7175,
+      "num_input_tokens_seen": 15277728,
+      "step": 2855
+    },
+    {
+      "epoch": 0.4576,
+      "grad_norm": 1.1595414876937866,
+      "learning_rate": 2.832024921984674e-05,
+      "loss": 0.6505,
+      "num_input_tokens_seen": 15307040,
+      "step": 2860
+    },
+    {
+      "epoch": 0.4584,
+      "grad_norm": 0.7592776417732239,
+      "learning_rate": 2.825796354081537e-05,
+      "loss": 0.6686,
+      "num_input_tokens_seen": 15334176,
+      "step": 2865
+    },
+    {
+      "epoch": 0.4592,
+      "grad_norm": 1.0087366104125977,
+      "learning_rate": 2.8195657282715594e-05,
+      "loss": 0.6365,
+      "num_input_tokens_seen": 15360496,
+      "step": 2870
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.9191427826881409,
+      "learning_rate": 2.8133330839107608e-05,
+      "loss": 0.6518,
+      "num_input_tokens_seen": 15381328,
+      "step": 2875
+    },
+    {
+      "epoch": 0.4608,
+      "grad_norm": 1.0468344688415527,
+      "learning_rate": 2.8070984603679107e-05,
+      "loss": 0.6262,
+      "num_input_tokens_seen": 15409936,
+      "step": 2880
+    },
+    {
+      "epoch": 0.4616,
+      "grad_norm": 1.1070493459701538,
+      "learning_rate": 2.800861897024279e-05,
+      "loss": 0.6684,
+      "num_input_tokens_seen": 15436848,
+      "step": 2885
+    },
+    {
+      "epoch": 0.4624,
+      "grad_norm": 1.3349978923797607,
+      "learning_rate": 2.79462343327339e-05,
+      "loss": 0.6978,
+      "num_input_tokens_seen": 15463328,
+      "step": 2890
+    },
+    {
+      "epoch": 0.4632,
+      "grad_norm": 0.7566163539886475,
+      "learning_rate": 2.7883831085207707e-05,
+      "loss": 0.7062,
+      "num_input_tokens_seen": 15489232,
+      "step": 2895
+    },
+    {
+      "epoch": 0.464,
+      "grad_norm": 0.7610609531402588,
+      "learning_rate": 2.782140962183704e-05,
+      "loss": 0.6642,
+      "num_input_tokens_seen": 15516224,
+      "step": 2900
+    },
+    {
+      "epoch": 0.4648,
+      "grad_norm": 0.7749585509300232,
+      "learning_rate": 2.7758970336909795e-05,
+      "loss": 0.6287,
+      "num_input_tokens_seen": 15545584,
+      "step": 2905
+    },
+    {
+      "epoch": 0.4656,
+      "grad_norm": 1.0202007293701172,
+      "learning_rate": 2.769651362482642e-05,
+      "loss": 0.6672,
+      "num_input_tokens_seen": 15571216,
+      "step": 2910
+    },
+    {
+      "epoch": 0.4664,
+      "grad_norm": 0.7980359792709351,
+      "learning_rate": 2.763403988009746e-05,
+      "loss": 0.737,
+      "num_input_tokens_seen": 15597744,
+      "step": 2915
+    },
+    {
+      "epoch": 0.4672,
+      "grad_norm": 0.942456841468811,
+      "learning_rate": 2.7571549497341042e-05,
+      "loss": 0.7715,
+      "num_input_tokens_seen": 15622496,
+      "step": 2920
+    },
+    {
+      "epoch": 0.468,
+      "grad_norm": 0.7782229781150818,
+      "learning_rate": 2.7509042871280372e-05,
+      "loss": 0.7435,
+      "num_input_tokens_seen": 15647344,
+      "step": 2925
+    },
+    {
+      "epoch": 0.4688,
+      "grad_norm": 1.0889509916305542,
+      "learning_rate": 2.744652039674129e-05,
+      "loss": 0.6946,
+      "num_input_tokens_seen": 15672672,
+      "step": 2930
+    },
+    {
+      "epoch": 0.4696,
+      "grad_norm": 1.0606461763381958,
+      "learning_rate": 2.7383982468649714e-05,
+      "loss": 0.7523,
+      "num_input_tokens_seen": 15696144,
+      "step": 2935
+    },
+    {
+      "epoch": 0.4704,
+      "grad_norm": 0.942613959312439,
+      "learning_rate": 2.73214294820292e-05,
+      "loss": 0.7218,
+      "num_input_tokens_seen": 15723600,
+      "step": 2940
+    },
+    {
+      "epoch": 0.4712,
+      "grad_norm": 0.8650354743003845,
+      "learning_rate": 2.7258861831998388e-05,
+      "loss": 0.6736,
+      "num_input_tokens_seen": 15749680,
+      "step": 2945
+    },
+    {
+      "epoch": 0.472,
+      "grad_norm": 0.8573226928710938,
+      "learning_rate": 2.7196279913768584e-05,
+      "loss": 0.6314,
+      "num_input_tokens_seen": 15776768,
+      "step": 2950
+    },
+    {
+      "epoch": 0.4728,
+      "grad_norm": 0.9692303538322449,
+      "learning_rate": 2.713368412264118e-05,
+      "loss": 0.7035,
+      "num_input_tokens_seen": 15801376,
+      "step": 2955
+    },
+    {
+      "epoch": 0.4736,
+      "grad_norm": 1.2111790180206299,
+      "learning_rate": 2.707107485400521e-05,
+      "loss": 0.6785,
+      "num_input_tokens_seen": 15828416,
+      "step": 2960
+    },
+    {
+      "epoch": 0.4744,
+      "grad_norm": 1.0816082954406738,
+      "learning_rate": 2.7008452503334858e-05,
+      "loss": 0.7672,
+      "num_input_tokens_seen": 15852720,
+      "step": 2965
+    },
+    {
+      "epoch": 0.4752,
+      "grad_norm": 0.980903148651123,
+      "learning_rate": 2.6945817466186912e-05,
+      "loss": 0.7723,
+      "num_input_tokens_seen": 15880624,
+      "step": 2970
+    },
+    {
+      "epoch": 0.476,
+      "grad_norm": 1.012623906135559,
+      "learning_rate": 2.6883170138198323e-05,
+      "loss": 0.6105,
+      "num_input_tokens_seen": 15912176,
+      "step": 2975
+    },
+    {
+      "epoch": 0.4768,
+      "grad_norm": 1.069486141204834,
+      "learning_rate": 2.6820510915083648e-05,
+      "loss": 0.6941,
+      "num_input_tokens_seen": 15944384,
+      "step": 2980
+    },
+    {
+      "epoch": 0.4776,
+      "grad_norm": 0.8411433100700378,
+      "learning_rate": 2.6757840192632598e-05,
+      "loss": 0.6669,
+      "num_input_tokens_seen": 15969680,
+      "step": 2985
+    },
+    {
+      "epoch": 0.4784,
+      "grad_norm": 0.6902319192886353,
+      "learning_rate": 2.6695158366707522e-05,
+      "loss": 0.6814,
+      "num_input_tokens_seen": 15997264,
+      "step": 2990
+    },
+    {
+      "epoch": 0.4792,
+      "grad_norm": 0.7844128012657166,
+      "learning_rate": 2.6632465833240893e-05,
+      "loss": 0.5641,
+      "num_input_tokens_seen": 16029664,
+      "step": 2995
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 1.1021162271499634,
+      "learning_rate": 2.656976298823284e-05,
+      "loss": 0.7797,
+      "num_input_tokens_seen": 16054864,
+      "step": 3000
+    },
+    {
+      "epoch": 0.4808,
+      "grad_norm": 1.2080223560333252,
+      "learning_rate": 2.650705022774859e-05,
+      "loss": 0.6778,
+      "num_input_tokens_seen": 16079552,
+      "step": 3005
+    },
+    {
+      "epoch": 0.4816,
+      "grad_norm": 1.1015046834945679,
+      "learning_rate": 2.6444327947916036e-05,
+      "loss": 0.6806,
+      "num_input_tokens_seen": 16105632,
+      "step": 3010
+    },
+    {
+      "epoch": 0.4824,
+      "grad_norm": 1.0269590616226196,
+      "learning_rate": 2.638159654492318e-05,
+      "loss": 0.7589,
+      "num_input_tokens_seen": 16134688,
+      "step": 3015
+    },
+    {
+      "epoch": 0.4832,
+      "grad_norm": 0.8565163612365723,
+      "learning_rate": 2.6318856415015664e-05,
+      "loss": 0.677,
+      "num_input_tokens_seen": 16163152,
+      "step": 3020
+    },
+    {
+      "epoch": 0.484,
+      "grad_norm": 0.8519279956817627,
+      "learning_rate": 2.6256107954494242e-05,
+      "loss": 0.6136,
+      "num_input_tokens_seen": 16189248,
+      "step": 3025
+    },
+    {
+      "epoch": 0.4848,
+      "grad_norm": 0.9466794729232788,
+      "learning_rate": 2.6193351559712292e-05,
+      "loss": 0.6369,
+      "num_input_tokens_seen": 16214832,
+      "step": 3030
+    },
+    {
+      "epoch": 0.4856,
+      "grad_norm": 0.9867402911186218,
+      "learning_rate": 2.6130587627073315e-05,
+      "loss": 0.7202,
+      "num_input_tokens_seen": 16244736,
+      "step": 3035
+    },
+    {
+      "epoch": 0.4864,
+      "grad_norm": 0.8893384337425232,
+      "learning_rate": 2.606781655302843e-05,
+      "loss": 0.7057,
+      "num_input_tokens_seen": 16272064,
+      "step": 3040
+    },
+    {
+      "epoch": 0.4872,
+      "grad_norm": 1.2341115474700928,
+      "learning_rate": 2.6005038734073833e-05,
+      "loss": 0.6715,
+      "num_input_tokens_seen": 16301344,
+      "step": 3045
+    },
+    {
+      "epoch": 0.488,
+      "grad_norm": 1.0158292055130005,
+      "learning_rate": 2.594225456674837e-05,
+      "loss": 0.7479,
+      "num_input_tokens_seen": 16325872,
+      "step": 3050
+    },
+    {
+      "epoch": 0.4888,
+      "grad_norm": 0.9316710233688354,
+      "learning_rate": 2.5879464447630946e-05,
+      "loss": 0.6581,
+      "num_input_tokens_seen": 16352272,
+      "step": 3055
+    },
+    {
+      "epoch": 0.4896,
+      "grad_norm": 0.9104299545288086,
+      "learning_rate": 2.5816668773338098e-05,
+      "loss": 0.691,
+      "num_input_tokens_seen": 16380464,
+      "step": 3060
+    },
+    {
+      "epoch": 0.4904,
+      "grad_norm": 0.8835129737854004,
+      "learning_rate": 2.575386794052142e-05,
+      "loss": 0.6606,
+      "num_input_tokens_seen": 16408736,
+      "step": 3065
+    },
+    {
+      "epoch": 0.4912,
+      "grad_norm": 0.869504451751709,
+      "learning_rate": 2.569106234586511e-05,
+      "loss": 0.729,
+      "num_input_tokens_seen": 16436352,
+      "step": 3070
+    },
+    {
+      "epoch": 0.492,
+      "grad_norm": 1.0879722833633423,
+      "learning_rate": 2.562825238608344e-05,
+      "loss": 0.7137,
+      "num_input_tokens_seen": 16464624,
+      "step": 3075
+    },
+    {
+      "epoch": 0.4928,
+      "grad_norm": 0.9328833818435669,
+      "learning_rate": 2.5565438457918244e-05,
+      "loss": 0.7238,
+      "num_input_tokens_seen": 16496720,
+      "step": 3080
+    },
+    {
+      "epoch": 0.4936,
+      "grad_norm": 0.7433749437332153,
+      "learning_rate": 2.5502620958136443e-05,
+      "loss": 0.7019,
+      "num_input_tokens_seen": 16524208,
+      "step": 3085
+    },
+    {
+      "epoch": 0.4944,
+      "grad_norm": 0.7768478989601135,
+      "learning_rate": 2.5439800283527494e-05,
+      "loss": 0.5851,
+      "num_input_tokens_seen": 16552192,
+      "step": 3090
+    },
+    {
+      "epoch": 0.4952,
+      "grad_norm": 1.139993667602539,
+      "learning_rate": 2.537697683090093e-05,
+      "loss": 0.7357,
+      "num_input_tokens_seen": 16578144,
+      "step": 3095
+    },
+    {
+      "epoch": 0.496,
+      "grad_norm": 0.9104892611503601,
+      "learning_rate": 2.531415099708382e-05,
+      "loss": 0.6254,
+      "num_input_tokens_seen": 16608288,
+      "step": 3100
+    },
+    {
+      "epoch": 0.4968,
+      "grad_norm": 0.6912931799888611,
+      "learning_rate": 2.5251323178918268e-05,
+      "loss": 0.7284,
+      "num_input_tokens_seen": 16636176,
+      "step": 3105
+    },
+    {
+      "epoch": 0.4976,
+      "grad_norm": 0.8370018601417542,
+      "learning_rate": 2.518849377325893e-05,
+      "loss": 0.8136,
+      "num_input_tokens_seen": 16659168,
+      "step": 3110
+    },
+    {
+      "epoch": 0.4984,
+      "grad_norm": 1.049914836883545,
+      "learning_rate": 2.5125663176970476e-05,
+      "loss": 0.7334,
+      "num_input_tokens_seen": 16687344,
+      "step": 3115
+    },
+    {
+      "epoch": 0.4992,
+      "grad_norm": 1.0298138856887817,
+      "learning_rate": 2.5062831786925102e-05,
+      "loss": 0.7599,
+      "num_input_tokens_seen": 16714496,
+      "step": 3120
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.7912611961364746,
+      "learning_rate": 2.5e-05,
+      "loss": 0.6517,
+      "num_input_tokens_seen": 16742528,
+      "step": 3125
+    },
+    {
+      "epoch": 0.5008,
+      "grad_norm": 1.0854965448379517,
+      "learning_rate": 2.4937168213074907e-05,
+      "loss": 0.6778,
+      "num_input_tokens_seen": 16771248,
+      "step": 3130
+    },
+    {
+      "epoch": 0.5016,
+      "grad_norm": 0.9674849510192871,
+      "learning_rate": 2.4874336823029526e-05,
+      "loss": 0.6847,
+      "num_input_tokens_seen": 16799136,
+      "step": 3135
+    },
+    {
+      "epoch": 0.5024,
+      "grad_norm": 0.8434900641441345,
+      "learning_rate": 2.481150622674108e-05,
+      "loss": 0.6638,
+      "num_input_tokens_seen": 16825648,
+      "step": 3140
+    },
+    {
+      "epoch": 0.5032,
+      "grad_norm": 0.8714620471000671,
+      "learning_rate": 2.4748676821081738e-05,
+      "loss": 0.7139,
+      "num_input_tokens_seen": 16852240,
+      "step": 3145
+    },
+    {
+      "epoch": 0.504,
+      "grad_norm": 0.8312164545059204,
+      "learning_rate": 2.4685849002916183e-05,
+      "loss": 0.7507,
+      "num_input_tokens_seen": 16878624,
+      "step": 3150
+    },
+    {
+      "epoch": 0.5048,
+      "grad_norm": 1.1353472471237183,
+      "learning_rate": 2.4623023169099073e-05,
+      "loss": 0.6951,
+      "num_input_tokens_seen": 16906864,
+      "step": 3155
+    },
+    {
+      "epoch": 0.5056,
+      "grad_norm": 1.1486365795135498,
+      "learning_rate": 2.4560199716472508e-05,
+      "loss": 0.733,
+      "num_input_tokens_seen": 16930080,
+      "step": 3160
+    },
+    {
+      "epoch": 0.5064,
+      "grad_norm": 0.9651095867156982,
+      "learning_rate": 2.449737904186357e-05,
+      "loss": 0.7517,
+      "num_input_tokens_seen": 16952240,
+      "step": 3165
+    },
+    {
+      "epoch": 0.5072,
+      "grad_norm": 0.8921483755111694,
+      "learning_rate": 2.4434561542081762e-05,
+      "loss": 0.7472,
+      "num_input_tokens_seen": 16985408,
+      "step": 3170
+    },
+    {
+      "epoch": 0.508,
+      "grad_norm": 1.0625066757202148,
+      "learning_rate": 2.4371747613916566e-05,
+      "loss": 0.7514,
+      "num_input_tokens_seen": 17013776,
+      "step": 3175
+    },
+    {
+      "epoch": 0.5088,
+      "grad_norm": 1.10313081741333,
+      "learning_rate": 2.4308937654134893e-05,
+      "loss": 0.7633,
+      "num_input_tokens_seen": 17039120,
+      "step": 3180
+    },
+    {
+      "epoch": 0.5096,
+      "grad_norm": 1.115670084953308,
+      "learning_rate": 2.4246132059478578e-05,
+      "loss": 0.6606,
+      "num_input_tokens_seen": 17065296,
+      "step": 3185
+    },
+    {
+      "epoch": 0.5104,
+      "grad_norm": 1.0417555570602417,
+      "learning_rate": 2.418333122666191e-05,
+      "loss": 0.764,
+      "num_input_tokens_seen": 17089264,
+      "step": 3190
+    },
+    {
+      "epoch": 0.5112,
+      "grad_norm": 0.9926926493644714,
+      "learning_rate": 2.412053555236906e-05,
+      "loss": 0.751,
+      "num_input_tokens_seen": 17117488,
+      "step": 3195
+    },
+    {
+      "epoch": 0.512,
+      "grad_norm": 1.1716359853744507,
+      "learning_rate": 2.4057745433251635e-05,
+      "loss": 0.7067,
+      "num_input_tokens_seen": 17141232,
+      "step": 3200
+    },
+    {
+      "epoch": 0.5128,
+      "grad_norm": 1.0248827934265137,
+      "learning_rate": 2.3994961265926166e-05,
+      "loss": 0.6432,
+      "num_input_tokens_seen": 17171632,
+      "step": 3205
+    },
+    {
+      "epoch": 0.5136,
+      "grad_norm": 0.8832619190216064,
+      "learning_rate": 2.3932183446971583e-05,
+      "loss": 0.6373,
+      "num_input_tokens_seen": 17198640,
+      "step": 3210
+    },
+    {
+      "epoch": 0.5144,
+      "grad_norm": 0.8581608533859253,
+      "learning_rate": 2.3869412372926687e-05,
+      "loss": 0.7347,
+      "num_input_tokens_seen": 17228240,
+      "step": 3215
+    },
+    {
+      "epoch": 0.5152,
+      "grad_norm": 1.0683339834213257,
+      "learning_rate": 2.3806648440287714e-05,
+      "loss": 0.6789,
+      "num_input_tokens_seen": 17259392,
+      "step": 3220
+    },
+    {
+      "epoch": 0.516,
+      "grad_norm": 0.9491643309593201,
+      "learning_rate": 2.3743892045505764e-05,
+      "loss": 0.7548,
+      "num_input_tokens_seen": 17287808,
+      "step": 3225
+    },
+    {
+      "epoch": 0.5168,
+      "grad_norm": 0.8620545864105225,
+      "learning_rate": 2.368114358498434e-05,
+      "loss": 0.7792,
+      "num_input_tokens_seen": 17311520,
+      "step": 3230
+    },
+    {
+      "epoch": 0.5176,
+      "grad_norm": 1.0956013202667236,
+      "learning_rate": 2.361840345507683e-05,
+      "loss": 0.6575,
+      "num_input_tokens_seen": 17340816,
+      "step": 3235
+    },
+    {
+      "epoch": 0.5184,
+      "grad_norm": 1.0029551982879639,
+      "learning_rate": 2.355567205208397e-05,
+      "loss": 0.6414,
+      "num_input_tokens_seen": 17363408,
+      "step": 3240
+    },
+    {
+      "epoch": 0.5192,
+      "grad_norm": 1.0472480058670044,
+      "learning_rate": 2.3492949772251414e-05,
+      "loss": 0.7161,
+      "num_input_tokens_seen": 17393248,
+      "step": 3245
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.8757247924804688,
+      "learning_rate": 2.3430237011767167e-05,
+      "loss": 0.6957,
+      "num_input_tokens_seen": 17425232,
+      "step": 3250
+    },
+    {
+      "epoch": 0.5208,
+      "grad_norm": 1.0374081134796143,
+      "learning_rate": 2.3367534166759102e-05,
+      "loss": 0.7615,
+      "num_input_tokens_seen": 17446864,
+      "step": 3255
+    },
+    {
+      "epoch": 0.5216,
+      "grad_norm": 1.0572500228881836,
+      "learning_rate": 2.3304841633292487e-05,
+      "loss": 0.6994,
+      "num_input_tokens_seen": 17470896,
+      "step": 3260
+    },
+    {
+      "epoch": 0.5224,
+      "grad_norm": 1.0209540128707886,
+      "learning_rate": 2.3242159807367408e-05,
+      "loss": 0.7116,
+      "num_input_tokens_seen": 17501488,
+      "step": 3265
+    },
+    {
+      "epoch": 0.5232,
+      "grad_norm": 1.0854222774505615,
+      "learning_rate": 2.3179489084916358e-05,
+      "loss": 0.7583,
+      "num_input_tokens_seen": 17526032,
+      "step": 3270
+    },
+    {
+      "epoch": 0.524,
+      "grad_norm": 1.0327874422073364,
+      "learning_rate": 2.3116829861801686e-05,
+      "loss": 0.7302,
+      "num_input_tokens_seen": 17550144,
+      "step": 3275
+    },
+    {
+      "epoch": 0.5248,
+      "grad_norm": 1.186990737915039,
+      "learning_rate": 2.3054182533813087e-05,
+      "loss": 0.6794,
+      "num_input_tokens_seen": 17575600,
+      "step": 3280
+    },
+    {
+      "epoch": 0.5256,
+      "grad_norm": 1.000475287437439,
+      "learning_rate": 2.2991547496665148e-05,
+      "loss": 0.7294,
+      "num_input_tokens_seen": 17601408,
+      "step": 3285
+    },
+    {
+      "epoch": 0.5264,
+      "grad_norm": 1.1120193004608154,
+      "learning_rate": 2.2928925145994794e-05,
+      "loss": 0.6333,
+      "num_input_tokens_seen": 17624752,
+      "step": 3290
+    },
+    {
+      "epoch": 0.5272,
+      "grad_norm": 0.7765217423439026,
+      "learning_rate": 2.286631587735883e-05,
+      "loss": 0.7779,
+      "num_input_tokens_seen": 17651040,
+      "step": 3295
+    },
+    {
+      "epoch": 0.528,
+      "grad_norm": 0.9403998255729675,
+      "learning_rate": 2.280372008623142e-05,
+      "loss": 0.7035,
+      "num_input_tokens_seen": 17678288,
+      "step": 3300
+    },
+    {
+      "epoch": 0.5288,
+      "grad_norm": 1.019305944442749,
+      "learning_rate": 2.2741138168001608e-05,
+      "loss": 0.719,
+      "num_input_tokens_seen": 17702816,
+      "step": 3305
+    },
+    {
+      "epoch": 0.5296,
+      "grad_norm": 1.0804177522659302,
+      "learning_rate": 2.267857051797081e-05,
+      "loss": 0.7134,
+      "num_input_tokens_seen": 17728848,
+      "step": 3310
+    },
+    {
+      "epoch": 0.5304,
+      "grad_norm": 0.7340876460075378,
+      "learning_rate": 2.2616017531350288e-05,
+      "loss": 0.6916,
+      "num_input_tokens_seen": 17756240,
+      "step": 3315
+    },
+    {
+      "epoch": 0.5312,
+      "grad_norm": 0.9618902802467346,
+      "learning_rate": 2.255347960325871e-05,
+      "loss": 0.6389,
+      "num_input_tokens_seen": 17781104,
+      "step": 3320
+    },
+    {
+      "epoch": 0.532,
+      "grad_norm": 0.9528229832649231,
+      "learning_rate": 2.2490957128719624e-05,
+      "loss": 0.6648,
+      "num_input_tokens_seen": 17808816,
+      "step": 3325
+    },
+    {
+      "epoch": 0.5328,
+      "grad_norm": 1.043328881263733,
+      "learning_rate": 2.2428450502658967e-05,
+      "loss": 0.6683,
+      "num_input_tokens_seen": 17834496,
+      "step": 3330
+    },
+    {
+      "epoch": 0.5336,
+      "grad_norm": 0.8162310719490051,
+      "learning_rate": 2.2365960119902545e-05,
+      "loss": 0.7686,
+      "num_input_tokens_seen": 17862880,
+      "step": 3335
+    },
+    {
+      "epoch": 0.5344,
+      "grad_norm": 0.8925397396087646,
+      "learning_rate": 2.2303486375173585e-05,
+      "loss": 0.7073,
+      "num_input_tokens_seen": 17890064,
+      "step": 3340
+    },
+    {
+      "epoch": 0.5352,
+      "grad_norm": 1.0610705614089966,
+      "learning_rate": 2.224102966309021e-05,
+      "loss": 0.642,
+      "num_input_tokens_seen": 17918144,
+      "step": 3345
+    },
+    {
+      "epoch": 0.536,
+      "grad_norm": 1.1452162265777588,
+      "learning_rate": 2.217859037816296e-05,
+      "loss": 0.7078,
+      "num_input_tokens_seen": 17945344,
+      "step": 3350
+    },
+    {
+      "epoch": 0.5368,
+      "grad_norm": 0.8698946833610535,
+      "learning_rate": 2.2116168914792292e-05,
+      "loss": 0.7437,
+      "num_input_tokens_seen": 17970096,
+      "step": 3355
+    },
+    {
+      "epoch": 0.5376,
+      "grad_norm": 1.1551156044006348,
+      "learning_rate": 2.205376566726611e-05,
+      "loss": 0.7606,
+      "num_input_tokens_seen": 17997328,
+      "step": 3360
+    },
+    {
+      "epoch": 0.5384,
+      "grad_norm": 1.3479046821594238,
+      "learning_rate": 2.1991381029757215e-05,
+      "loss": 0.6824,
+      "num_input_tokens_seen": 18022464,
+      "step": 3365
+    },
+    {
+      "epoch": 0.5392,
+      "grad_norm": 0.9218052625656128,
+      "learning_rate": 2.19290153963209e-05,
+      "loss": 0.7262,
+      "num_input_tokens_seen": 18052176,
+      "step": 3370
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 1.302252173423767,
+      "learning_rate": 2.186666916089239e-05,
+      "loss": 0.7491,
+      "num_input_tokens_seen": 18079008,
+      "step": 3375
+    },
+    {
+      "epoch": 0.5408,
+      "grad_norm": 1.4532941579818726,
+      "learning_rate": 2.1804342717284415e-05,
+      "loss": 0.6246,
+      "num_input_tokens_seen": 18102784,
+      "step": 3380
+    },
+    {
+      "epoch": 0.5416,
+      "grad_norm": 0.7572783827781677,
+      "learning_rate": 2.174203645918464e-05,
+      "loss": 0.6712,
+      "num_input_tokens_seen": 18130688,
+      "step": 3385
+    },
+    {
+      "epoch": 0.5424,
+      "grad_norm": 1.0954492092132568,
+      "learning_rate": 2.1679750780153267e-05,
+      "loss": 0.7238,
+      "num_input_tokens_seen": 18159200,
+      "step": 3390
+    },
+    {
+      "epoch": 0.5432,
+      "grad_norm": 1.1352787017822266,
+      "learning_rate": 2.1617486073620498e-05,
+      "loss": 0.663,
+      "num_input_tokens_seen": 18188736,
+      "step": 3395
+    },
+    {
+      "epoch": 0.544,
+      "grad_norm": 1.012987732887268,
+      "learning_rate": 2.155524273288405e-05,
+      "loss": 0.6928,
+      "num_input_tokens_seen": 18217856,
+      "step": 3400
+    },
+    {
+      "epoch": 0.5448,
+      "grad_norm": 0.8638446927070618,
+      "learning_rate": 2.1493021151106703e-05,
+      "loss": 0.7373,
+      "num_input_tokens_seen": 18247616,
+      "step": 3405
+    },
+    {
+      "epoch": 0.5456,
+      "grad_norm": 1.2647075653076172,
+      "learning_rate": 2.1430821721313782e-05,
+      "loss": 0.7593,
+      "num_input_tokens_seen": 18274416,
+      "step": 3410
+    },
+    {
+      "epoch": 0.5464,
+      "grad_norm": 0.8533580899238586,
+      "learning_rate": 2.1368644836390684e-05,
+      "loss": 0.6718,
+      "num_input_tokens_seen": 18298720,
+      "step": 3415
+    },
+    {
+      "epoch": 0.5472,
+      "grad_norm": 0.8091197609901428,
+      "learning_rate": 2.130649088908041e-05,
+      "loss": 0.7303,
+      "num_input_tokens_seen": 18326160,
+      "step": 3420
+    },
+    {
+      "epoch": 0.548,
+      "grad_norm": 0.886374294757843,
+      "learning_rate": 2.1244360271981073e-05,
+      "loss": 0.74,
+      "num_input_tokens_seen": 18351344,
+      "step": 3425
+    },
+    {
+      "epoch": 0.5488,
+      "grad_norm": 0.8284346461296082,
+      "learning_rate": 2.1182253377543425e-05,
+      "loss": 0.6448,
+      "num_input_tokens_seen": 18374752,
+      "step": 3430
+    },
+    {
+      "epoch": 0.5496,
+      "grad_norm": 0.9252715706825256,
+      "learning_rate": 2.112017059806835e-05,
+      "loss": 0.6759,
+      "num_input_tokens_seen": 18402432,
+      "step": 3435
+    },
+    {
+      "epoch": 0.5504,
+      "grad_norm": 1.0463306903839111,
+      "learning_rate": 2.1058112325704436e-05,
+      "loss": 0.7327,
+      "num_input_tokens_seen": 18428656,
+      "step": 3440
+    },
+    {
+      "epoch": 0.5512,
+      "grad_norm": 0.999754011631012,
+      "learning_rate": 2.0996078952445452e-05,
+      "loss": 0.7214,
+      "num_input_tokens_seen": 18451744,
+      "step": 3445
+    },
+    {
+      "epoch": 0.552,
+      "grad_norm": 0.759445071220398,
+      "learning_rate": 2.0934070870127912e-05,
+      "loss": 0.7192,
+      "num_input_tokens_seen": 18476960,
+      "step": 3450
+    },
+    {
+      "epoch": 0.5528,
+      "grad_norm": 0.757986843585968,
+      "learning_rate": 2.0872088470428553e-05,
+      "loss": 0.6481,
+      "num_input_tokens_seen": 18507280,
+      "step": 3455
+    },
+    {
+      "epoch": 0.5536,
+      "grad_norm": 1.555863857269287,
+      "learning_rate": 2.08101321448619e-05,
+      "loss": 0.6906,
+      "num_input_tokens_seen": 18531264,
+      "step": 3460
+    },
+    {
+      "epoch": 0.5544,
+      "grad_norm": 0.9648059606552124,
+      "learning_rate": 2.0748202284777777e-05,
+      "loss": 0.6691,
+      "num_input_tokens_seen": 18559552,
+      "step": 3465
+    },
+    {
+      "epoch": 0.5552,
+      "grad_norm": 0.7556829452514648,
+      "learning_rate": 2.0686299281358835e-05,
+      "loss": 0.743,
+      "num_input_tokens_seen": 18587408,
+      "step": 3470
+    },
+    {
+      "epoch": 0.556,
+      "grad_norm": 1.2063502073287964,
+      "learning_rate": 2.0624423525618098e-05,
+      "loss": 0.6896,
+      "num_input_tokens_seen": 18616384,
+      "step": 3475
+    },
+    {
+      "epoch": 0.5568,
+      "grad_norm": 1.0923407077789307,
+      "learning_rate": 2.056257540839647e-05,
+      "loss": 0.7799,
+      "num_input_tokens_seen": 18640432,
+      "step": 3480
+    },
+    {
+      "epoch": 0.5576,
+      "grad_norm": 0.9768779873847961,
+      "learning_rate": 2.050075532036026e-05,
+      "loss": 0.6796,
+      "num_input_tokens_seen": 18661696,
+      "step": 3485
+    },
+    {
+      "epoch": 0.5584,
+      "grad_norm": 1.0903396606445312,
+      "learning_rate": 2.0438963651998747e-05,
+      "loss": 0.6601,
+      "num_input_tokens_seen": 18689280,
+      "step": 3490
+    },
+    {
+      "epoch": 0.5592,
+      "grad_norm": 1.0859960317611694,
+      "learning_rate": 2.037720079362169e-05,
+      "loss": 0.7247,
+      "num_input_tokens_seen": 18713776,
+      "step": 3495
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.979040801525116,
+      "learning_rate": 2.031546713535688e-05,
+      "loss": 0.7113,
+      "num_input_tokens_seen": 18739632,
+      "step": 3500
+    },
+    {
+      "epoch": 0.5608,
+      "grad_norm": 0.9411745667457581,
+      "learning_rate": 2.0253763067147657e-05,
+      "loss": 0.604,
+      "num_input_tokens_seen": 18767504,
+      "step": 3505
+    },
+    {
+      "epoch": 0.5616,
+      "grad_norm": 1.0369850397109985,
+      "learning_rate": 2.0192088978750433e-05,
+      "loss": 0.6292,
+      "num_input_tokens_seen": 18794976,
+      "step": 3510
+    },
+    {
+      "epoch": 0.5624,
+      "grad_norm": 1.033288598060608,
+      "learning_rate": 2.0130445259732285e-05,
+      "loss": 0.7227,
+      "num_input_tokens_seen": 18823456,
+      "step": 3515
+    },
+    {
+      "epoch": 0.5632,
+      "grad_norm": 1.0590053796768188,
+      "learning_rate": 2.0068832299468428e-05,
+      "loss": 0.6536,
+      "num_input_tokens_seen": 18851104,
+      "step": 3520
+    },
+    {
+      "epoch": 0.564,
+      "grad_norm": 0.8934491276741028,
+      "learning_rate": 2.000725048713983e-05,
+      "loss": 0.6729,
+      "num_input_tokens_seen": 18882096,
+      "step": 3525
+    },
+    {
+      "epoch": 0.5648,
+      "grad_norm": 1.3166091442108154,
+      "learning_rate": 1.994570021173067e-05,
+      "loss": 0.8178,
+      "num_input_tokens_seen": 18903520,
+      "step": 3530
+    },
+    {
+      "epoch": 0.5656,
+      "grad_norm": 0.7813395857810974,
+      "learning_rate": 1.988418186202594e-05,
+      "loss": 0.6575,
+      "num_input_tokens_seen": 18937200,
+      "step": 3535
+    },
+    {
+      "epoch": 0.5664,
+      "grad_norm": 1.1987985372543335,
+      "learning_rate": 1.9822695826608972e-05,
+      "loss": 0.7789,
+      "num_input_tokens_seen": 18965424,
+      "step": 3540
+    },
+    {
+      "epoch": 0.5672,
+      "grad_norm": 0.9326161742210388,
+      "learning_rate": 1.9761242493858987e-05,
+      "loss": 0.6699,
+      "num_input_tokens_seen": 18989456,
+      "step": 3545
+    },
+    {
+      "epoch": 0.568,
+      "grad_norm": 0.736609160900116,
+      "learning_rate": 1.969982225194864e-05,
+      "loss": 0.7095,
+      "num_input_tokens_seen": 19017808,
+      "step": 3550
+    },
+    {
+      "epoch": 0.5688,
+      "grad_norm": 1.0842061042785645,
+      "learning_rate": 1.9638435488841546e-05,
+      "loss": 0.7191,
+      "num_input_tokens_seen": 19046496,
+      "step": 3555
+    },
+    {
+      "epoch": 0.5696,
+      "grad_norm": 0.9748028516769409,
+      "learning_rate": 1.957708259228987e-05,
+      "loss": 0.7016,
+      "num_input_tokens_seen": 19072128,
+      "step": 3560
+    },
+    {
+      "epoch": 0.5704,
+      "grad_norm": 1.0534452199935913,
+      "learning_rate": 1.951576394983185e-05,
+      "loss": 0.6903,
+      "num_input_tokens_seen": 19096528,
+      "step": 3565
+    },
+    {
+      "epoch": 0.5712,
+      "grad_norm": 0.860016405582428,
+      "learning_rate": 1.945447994878937e-05,
+      "loss": 0.6144,
+      "num_input_tokens_seen": 19126240,
+      "step": 3570
+    },
+    {
+      "epoch": 0.572,
+      "grad_norm": 0.9540638327598572,
+      "learning_rate": 1.9393230976265473e-05,
+      "loss": 0.6755,
+      "num_input_tokens_seen": 19152752,
+      "step": 3575
+    },
+    {
+      "epoch": 0.5728,
+      "grad_norm": 0.8391373157501221,
+      "learning_rate": 1.9332017419141962e-05,
+      "loss": 0.748,
+      "num_input_tokens_seen": 19179296,
+      "step": 3580
+    },
+    {
+      "epoch": 0.5736,
+      "grad_norm": 1.1639857292175293,
+      "learning_rate": 1.9270839664076936e-05,
+      "loss": 0.7011,
+      "num_input_tokens_seen": 19205616,
+      "step": 3585
+    },
+    {
+      "epoch": 0.5744,
+      "grad_norm": 0.8739202618598938,
+      "learning_rate": 1.920969809750234e-05,
+      "loss": 0.6672,
+      "num_input_tokens_seen": 19231440,
+      "step": 3590
+    },
+    {
+      "epoch": 0.5752,
+      "grad_norm": 0.8280954360961914,
+      "learning_rate": 1.914859310562154e-05,
+      "loss": 0.7261,
+      "num_input_tokens_seen": 19258288,
+      "step": 3595
+    },
+    {
+      "epoch": 0.576,
+      "grad_norm": 0.671859622001648,
+      "learning_rate": 1.908752507440689e-05,
+      "loss": 0.7838,
+      "num_input_tokens_seen": 19284464,
+      "step": 3600
+    },
+    {
+      "epoch": 0.5768,
+      "grad_norm": 0.7985883951187134,
+      "learning_rate": 1.9026494389597238e-05,
+      "loss": 0.6574,
+      "num_input_tokens_seen": 19312272,
+      "step": 3605
+    },
+    {
+      "epoch": 0.5776,
+      "grad_norm": 1.1776115894317627,
+      "learning_rate": 1.8965501436695577e-05,
+      "loss": 0.7692,
+      "num_input_tokens_seen": 19335408,
+      "step": 3610
+    },
+    {
+      "epoch": 0.5784,
+      "grad_norm": 0.7614251971244812,
+      "learning_rate": 1.890454660096654e-05,
+      "loss": 0.7165,
+      "num_input_tokens_seen": 19360768,
+      "step": 3615
+    },
+    {
+      "epoch": 0.5792,
+      "grad_norm": 1.0146969556808472,
+      "learning_rate": 1.8843630267434e-05,
+      "loss": 0.8187,
+      "num_input_tokens_seen": 19386016,
+      "step": 3620
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.8127625584602356,
+      "learning_rate": 1.8782752820878634e-05,
+      "loss": 0.6587,
+      "num_input_tokens_seen": 19408560,
+      "step": 3625
+    },
+    {
+      "epoch": 0.5808,
+      "grad_norm": 1.102415680885315,
+      "learning_rate": 1.872191464583547e-05,
+      "loss": 0.6268,
+      "num_input_tokens_seen": 19436592,
+      "step": 3630
+    },
+    {
+      "epoch": 0.5816,
+      "grad_norm": 0.8009477853775024,
+      "learning_rate": 1.866111612659149e-05,
+      "loss": 0.6977,
+      "num_input_tokens_seen": 19463440,
+      "step": 3635
+    },
+    {
+      "epoch": 0.5824,
+      "grad_norm": 0.9613442420959473,
+      "learning_rate": 1.8600357647183185e-05,
+      "loss": 0.6292,
+      "num_input_tokens_seen": 19493360,
+      "step": 3640
+    },
+    {
+      "epoch": 0.5832,
+      "grad_norm": 1.1276973485946655,
+      "learning_rate": 1.8539639591394133e-05,
+      "loss": 0.6547,
+      "num_input_tokens_seen": 19521392,
+      "step": 3645
+    },
+    {
+      "epoch": 0.584,
+      "grad_norm": 1.2128732204437256,
+      "learning_rate": 1.8478962342752583e-05,
+      "loss": 0.6717,
+      "num_input_tokens_seen": 19550336,
+      "step": 3650
+    },
+    {
+      "epoch": 0.5848,
+      "grad_norm": 1.1931806802749634,
+      "learning_rate": 1.8418326284528996e-05,
+      "loss": 0.7065,
+      "num_input_tokens_seen": 19575776,
+      "step": 3655
+    },
+    {
+      "epoch": 0.5856,
+      "grad_norm": 0.921335756778717,
+      "learning_rate": 1.8357731799733686e-05,
+      "loss": 0.7029,
+      "num_input_tokens_seen": 19598128,
+      "step": 3660
+    },
+    {
+      "epoch": 0.5864,
+      "grad_norm": 0.8000009655952454,
+      "learning_rate": 1.8297179271114346e-05,
+      "loss": 0.7311,
+      "num_input_tokens_seen": 19625648,
+      "step": 3665
+    },
+    {
+      "epoch": 0.5872,
+      "grad_norm": 1.0933367013931274,
+      "learning_rate": 1.8236669081153657e-05,
+      "loss": 0.7296,
+      "num_input_tokens_seen": 19649952,
+      "step": 3670
+    },
+    {
+      "epoch": 0.588,
+      "grad_norm": 0.8331469297409058,
+      "learning_rate": 1.817620161206687e-05,
+      "loss": 0.7534,
+      "num_input_tokens_seen": 19677680,
+      "step": 3675
+    },
+    {
+      "epoch": 0.5888,
+      "grad_norm": 1.3450491428375244,
+      "learning_rate": 1.811577724579938e-05,
+      "loss": 0.6995,
+      "num_input_tokens_seen": 19711904,
+      "step": 3680
+    },
+    {
+      "epoch": 0.5896,
+      "grad_norm": 1.0697826147079468,
+      "learning_rate": 1.8055396364024317e-05,
+      "loss": 0.7517,
+      "num_input_tokens_seen": 19734272,
+      "step": 3685
+    },
+    {
+      "epoch": 0.5904,
+      "grad_norm": 0.9218750596046448,
+      "learning_rate": 1.7995059348140165e-05,
+      "loss": 0.7048,
+      "num_input_tokens_seen": 19761136,
+      "step": 3690
+    },
+    {
+      "epoch": 0.5912,
+      "grad_norm": 0.7037175297737122,
+      "learning_rate": 1.7934766579268292e-05,
+      "loss": 0.6385,
+      "num_input_tokens_seen": 19784880,
+      "step": 3695
+    },
+    {
+      "epoch": 0.592,
+      "grad_norm": 0.9812880754470825,
+      "learning_rate": 1.7874518438250597e-05,
+      "loss": 0.8177,
+      "num_input_tokens_seen": 19811456,
+      "step": 3700
+    },
+    {
+      "epoch": 0.5928,
+      "grad_norm": 1.0128806829452515,
+      "learning_rate": 1.7814315305647093e-05,
+      "loss": 0.7373,
+      "num_input_tokens_seen": 19839168,
+      "step": 3705
+    },
+    {
+      "epoch": 0.5936,
+      "grad_norm": 0.8506542444229126,
+      "learning_rate": 1.7754157561733476e-05,
+      "loss": 0.723,
+      "num_input_tokens_seen": 19865584,
+      "step": 3710
+    },
+    {
+      "epoch": 0.5944,
+      "grad_norm": 0.8591898679733276,
+      "learning_rate": 1.7694045586498752e-05,
+      "loss": 0.6315,
+      "num_input_tokens_seen": 19893232,
+      "step": 3715
+    },
+    {
+      "epoch": 0.5952,
+      "grad_norm": 0.9761216640472412,
+      "learning_rate": 1.7633979759642844e-05,
+      "loss": 0.6184,
+      "num_input_tokens_seen": 19918512,
+      "step": 3720
+    },
+    {
+      "epoch": 0.596,
+      "grad_norm": 0.9515823721885681,
+      "learning_rate": 1.7573960460574133e-05,
+      "loss": 0.682,
+      "num_input_tokens_seen": 19944544,
+      "step": 3725
+    },
+    {
+      "epoch": 0.5968,
+      "grad_norm": 1.2393804788589478,
+      "learning_rate": 1.7513988068407146e-05,
+      "loss": 0.6738,
+      "num_input_tokens_seen": 19971104,
+      "step": 3730
+    },
+    {
+      "epoch": 0.5976,
+      "grad_norm": 1.2483285665512085,
+      "learning_rate": 1.74540629619601e-05,
+      "loss": 0.6895,
+      "num_input_tokens_seen": 19996352,
+      "step": 3735
+    },
+    {
+      "epoch": 0.5984,
+      "grad_norm": 1.162599802017212,
+      "learning_rate": 1.7394185519752545e-05,
+      "loss": 0.7436,
+      "num_input_tokens_seen": 20021744,
+      "step": 3740
+    },
+    {
+      "epoch": 0.5992,
+      "grad_norm": 0.8526731133460999,
+      "learning_rate": 1.7334356120002957e-05,
+      "loss": 0.7612,
+      "num_input_tokens_seen": 20046560,
+      "step": 3745
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 1.1033904552459717,
+      "learning_rate": 1.7274575140626318e-05,
+      "loss": 0.7139,
+      "num_input_tokens_seen": 20072560,
+      "step": 3750
+    },
+    {
+      "epoch": 0.6008,
+      "grad_norm": 0.9515017867088318,
+      "learning_rate": 1.7214842959231794e-05,
+      "loss": 0.6556,
+      "num_input_tokens_seen": 20103488,
+      "step": 3755
+    },
+    {
+      "epoch": 0.6016,
+      "grad_norm": 1.1830626726150513,
+      "learning_rate": 1.7155159953120313e-05,
+      "loss": 0.6884,
+      "num_input_tokens_seen": 20130784,
+      "step": 3760
+    },
+    {
+      "epoch": 0.6024,
+      "grad_norm": 1.1456624269485474,
+      "learning_rate": 1.7095526499282172e-05,
+      "loss": 0.7729,
+      "num_input_tokens_seen": 20158720,
+      "step": 3765
+    },
+    {
+      "epoch": 0.6032,
+      "grad_norm": 0.8993046879768372,
+      "learning_rate": 1.703594297439469e-05,
+      "loss": 0.7427,
+      "num_input_tokens_seen": 20180736,
+      "step": 3770
+    },
+    {
+      "epoch": 0.604,
+      "grad_norm": 1.0378142595291138,
+      "learning_rate": 1.6976409754819767e-05,
+      "loss": 0.6831,
+      "num_input_tokens_seen": 20203744,
+      "step": 3775
+    },
+    {
+      "epoch": 0.6048,
+      "grad_norm": 1.0006003379821777,
+      "learning_rate": 1.6916927216601593e-05,
+      "loss": 0.6098,
+      "num_input_tokens_seen": 20232784,
+      "step": 3780
+    },
+    {
+      "epoch": 0.6056,
+      "grad_norm": 0.9714456796646118,
+      "learning_rate": 1.6857495735464195e-05,
+      "loss": 0.812,
+      "num_input_tokens_seen": 20262256,
+      "step": 3785
+    },
+    {
+      "epoch": 0.6064,
+      "grad_norm": 1.226090431213379,
+      "learning_rate": 1.6798115686809125e-05,
+      "loss": 0.6337,
+      "num_input_tokens_seen": 20290720,
+      "step": 3790
+    },
+    {
+      "epoch": 0.6072,
+      "grad_norm": 1.2579602003097534,
+      "learning_rate": 1.6738787445713037e-05,
+      "loss": 0.7105,
+      "num_input_tokens_seen": 20314368,
+      "step": 3795
+    },
+    {
+      "epoch": 0.608,
+      "grad_norm": 0.9636203050613403,
+      "learning_rate": 1.6679511386925337e-05,
+      "loss": 0.7776,
+      "num_input_tokens_seen": 20337648,
+      "step": 3800
+    },
+    {
+      "epoch": 0.6088,
+      "grad_norm": 0.8365712761878967,
+      "learning_rate": 1.662028788486583e-05,
+      "loss": 0.6626,
+      "num_input_tokens_seen": 20367344,
+      "step": 3805
+    },
+    {
+      "epoch": 0.6096,
+      "grad_norm": 1.340468168258667,
+      "learning_rate": 1.656111731362236e-05,
+      "loss": 0.6983,
+      "num_input_tokens_seen": 20391616,
+      "step": 3810
+    },
+    {
+      "epoch": 0.6104,
+      "grad_norm": 0.878955066204071,
+      "learning_rate": 1.650200004694839e-05,
+      "loss": 0.6948,
+      "num_input_tokens_seen": 20419520,
+      "step": 3815
+    },
+    {
+      "epoch": 0.6112,
+      "grad_norm": 0.9543726444244385,
+      "learning_rate": 1.644293645826072e-05,
+      "loss": 0.8154,
+      "num_input_tokens_seen": 20446048,
+      "step": 3820
+    },
+    {
+      "epoch": 0.612,
+      "grad_norm": 1.2340530157089233,
+      "learning_rate": 1.6383926920637077e-05,
+      "loss": 0.7234,
+      "num_input_tokens_seen": 20472960,
+      "step": 3825
+    },
+    {
+      "epoch": 0.6128,
+      "grad_norm": 0.8794097900390625,
+      "learning_rate": 1.6324971806813767e-05,
+      "loss": 0.668,
+      "num_input_tokens_seen": 20502816,
+      "step": 3830
+    },
+    {
+      "epoch": 0.6136,
+      "grad_norm": 1.0433683395385742,
+      "learning_rate": 1.6266071489183327e-05,
+      "loss": 0.6936,
+      "num_input_tokens_seen": 20529056,
+      "step": 3835
+    },
+    {
+      "epoch": 0.6144,
+      "grad_norm": 1.3372730016708374,
+      "learning_rate": 1.620722633979219e-05,
+      "loss": 0.7988,
+      "num_input_tokens_seen": 20555392,
+      "step": 3840
+    },
+    {
+      "epoch": 0.6152,
+      "grad_norm": 1.0201383829116821,
+      "learning_rate": 1.614843673033828e-05,
+      "loss": 0.7752,
+      "num_input_tokens_seen": 20583888,
+      "step": 3845
+    },
+    {
+      "epoch": 0.616,
+      "grad_norm": 0.7360509634017944,
+      "learning_rate": 1.6089703032168733e-05,
+      "loss": 0.6784,
+      "num_input_tokens_seen": 20612112,
+      "step": 3850
+    },
+    {
+      "epoch": 0.6168,
+      "grad_norm": 0.8650053143501282,
+      "learning_rate": 1.603102561627751e-05,
+      "loss": 0.6483,
+      "num_input_tokens_seen": 20639296,
+      "step": 3855
+    },
+    {
+      "epoch": 0.6176,
+      "grad_norm": 0.9596717953681946,
+      "learning_rate": 1.5972404853303062e-05,
+      "loss": 0.6876,
+      "num_input_tokens_seen": 20663680,
+      "step": 3860
+    },
+    {
+      "epoch": 0.6184,
+      "grad_norm": 0.9903700947761536,
+      "learning_rate": 1.5913841113525992e-05,
+      "loss": 0.7651,
+      "num_input_tokens_seen": 20690592,
+      "step": 3865
+    },
+    {
+      "epoch": 0.6192,
+      "grad_norm": 1.0361056327819824,
+      "learning_rate": 1.585533476686669e-05,
+      "loss": 0.6692,
+      "num_input_tokens_seen": 20716944,
+      "step": 3870
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 1.3285175561904907,
+      "learning_rate": 1.5796886182883053e-05,
+      "loss": 0.708,
+      "num_input_tokens_seen": 20742128,
+      "step": 3875
+    },
+    {
+      "epoch": 0.6208,
+      "grad_norm": 0.9742456078529358,
+      "learning_rate": 1.5738495730768105e-05,
+      "loss": 0.6734,
+      "num_input_tokens_seen": 20769344,
+      "step": 3880
+    },
+    {
+      "epoch": 0.6216,
+      "grad_norm": 0.9866467118263245,
+      "learning_rate": 1.5680163779347667e-05,
+      "loss": 0.7442,
+      "num_input_tokens_seen": 20793920,
+      "step": 3885
+    },
+    {
+      "epoch": 0.6224,
+      "grad_norm": 1.2530503273010254,
+      "learning_rate": 1.562189069707807e-05,
+      "loss": 0.786,
+      "num_input_tokens_seen": 20819616,
+      "step": 3890
+    },
+    {
+      "epoch": 0.6232,
+      "grad_norm": 1.0180388689041138,
+      "learning_rate": 1.556367685204374e-05,
+      "loss": 0.6565,
+      "num_input_tokens_seen": 20843056,
+      "step": 3895
+    },
+    {
+      "epoch": 0.624,
+      "grad_norm": 1.1570924520492554,
+      "learning_rate": 1.5505522611954975e-05,
+      "loss": 0.8403,
+      "num_input_tokens_seen": 20870320,
+      "step": 3900
+    },
+    {
+      "epoch": 0.6248,
+      "grad_norm": 0.9555189609527588,
+      "learning_rate": 1.5447428344145563e-05,
+      "loss": 0.717,
+      "num_input_tokens_seen": 20894448,
+      "step": 3905
+    },
+    {
+      "epoch": 0.6256,
+      "grad_norm": 0.7047298550605774,
+      "learning_rate": 1.538939441557048e-05,
+      "loss": 0.6563,
+      "num_input_tokens_seen": 20926800,
+      "step": 3910
+    },
+    {
+      "epoch": 0.6264,
+      "grad_norm": 1.1212091445922852,
+      "learning_rate": 1.5331421192803565e-05,
+      "loss": 0.7742,
+      "num_input_tokens_seen": 20954016,
+      "step": 3915
+    },
+    {
+      "epoch": 0.6272,
+      "grad_norm": 0.9030645489692688,
+      "learning_rate": 1.5273509042035172e-05,
+      "loss": 0.6654,
+      "num_input_tokens_seen": 20982512,
+      "step": 3920
+    },
+    {
+      "epoch": 0.628,
+      "grad_norm": 0.9414677619934082,
+      "learning_rate": 1.521565832906994e-05,
+      "loss": 0.6737,
+      "num_input_tokens_seen": 21008768,
+      "step": 3925
+    },
+    {
+      "epoch": 0.6288,
+      "grad_norm": 1.1415228843688965,
+      "learning_rate": 1.515786941932441e-05,
+      "loss": 0.7259,
+      "num_input_tokens_seen": 21038144,
+      "step": 3930
+    },
+    {
+      "epoch": 0.6296,
+      "grad_norm": 1.0087826251983643,
+      "learning_rate": 1.5100142677824753e-05,
+      "loss": 0.6793,
+      "num_input_tokens_seen": 21065120,
+      "step": 3935
+    },
+    {
+      "epoch": 0.6304,
+      "grad_norm": 1.3329883813858032,
+      "learning_rate": 1.5042478469204435e-05,
+      "loss": 0.6934,
+      "num_input_tokens_seen": 21091296,
+      "step": 3940
+    },
+    {
+      "epoch": 0.6312,
+      "grad_norm": 0.9850925803184509,
+      "learning_rate": 1.4984877157701932e-05,
+      "loss": 0.7746,
+      "num_input_tokens_seen": 21117568,
+      "step": 3945
+    },
+    {
+      "epoch": 0.632,
+      "grad_norm": 0.8925058245658875,
+      "learning_rate": 1.4927339107158437e-05,
+      "loss": 0.6311,
+      "num_input_tokens_seen": 21146640,
+      "step": 3950
+    },
+    {
+      "epoch": 0.6328,
+      "grad_norm": 1.2707431316375732,
+      "learning_rate": 1.486986468101555e-05,
+      "loss": 0.7614,
+      "num_input_tokens_seen": 21169680,
+      "step": 3955
+    },
+    {
+      "epoch": 0.6336,
+      "grad_norm": 1.0344791412353516,
+      "learning_rate": 1.4812454242312979e-05,
+      "loss": 0.7291,
+      "num_input_tokens_seen": 21195360,
+      "step": 3960
+    },
+    {
+      "epoch": 0.6344,
+      "grad_norm": 0.8999541997909546,
+      "learning_rate": 1.4755108153686275e-05,
+      "loss": 0.7421,
+      "num_input_tokens_seen": 21218896,
+      "step": 3965
+    },
+    {
+      "epoch": 0.6352,
+      "grad_norm": 1.3539083003997803,
+      "learning_rate": 1.4697826777364477e-05,
+      "loss": 0.7831,
+      "num_input_tokens_seen": 21244080,
+      "step": 3970
+    },
+    {
+      "epoch": 0.636,
+      "grad_norm": 0.9629884362220764,
+      "learning_rate": 1.4640610475167898e-05,
+      "loss": 0.6907,
+      "num_input_tokens_seen": 21271024,
+      "step": 3975
+    },
+    {
+      "epoch": 0.6368,
+      "grad_norm": 0.9040243625640869,
+      "learning_rate": 1.4583459608505801e-05,
+      "loss": 0.7001,
+      "num_input_tokens_seen": 21298992,
+      "step": 3980
+    },
+    {
+      "epoch": 0.6376,
+      "grad_norm": 1.122290849685669,
+      "learning_rate": 1.4526374538374132e-05,
+      "loss": 0.6729,
+      "num_input_tokens_seen": 21324032,
+      "step": 3985
+    },
+    {
+      "epoch": 0.6384,
+      "grad_norm": 0.8082273602485657,
+      "learning_rate": 1.4469355625353198e-05,
+      "loss": 0.6636,
+      "num_input_tokens_seen": 21354256,
+      "step": 3990
+    },
+    {
+      "epoch": 0.6392,
+      "grad_norm": 0.7639278173446655,
+      "learning_rate": 1.4412403229605454e-05,
+      "loss": 0.6349,
+      "num_input_tokens_seen": 21382144,
+      "step": 3995
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 1.0013383626937866,
+      "learning_rate": 1.4355517710873184e-05,
+      "loss": 0.6892,
+      "num_input_tokens_seen": 21408112,
+      "step": 4000
+    },
+    {
+      "epoch": 0.6408,
+      "grad_norm": 0.918889582157135,
+      "learning_rate": 1.4298699428476236e-05,
+      "loss": 0.6602,
+      "num_input_tokens_seen": 21438800,
+      "step": 4005
+    },
+    {
+      "epoch": 0.6416,
+      "grad_norm": 0.9448719620704651,
+      "learning_rate": 1.4241948741309782e-05,
+      "loss": 0.6613,
+      "num_input_tokens_seen": 21466464,
+      "step": 4010
+    },
+    {
+      "epoch": 0.6424,
+      "grad_norm": 1.1950000524520874,
+      "learning_rate": 1.418526600784198e-05,
+      "loss": 0.6821,
+      "num_input_tokens_seen": 21496864,
+      "step": 4015
+    },
+    {
+      "epoch": 0.6432,
+      "grad_norm": 1.0359631776809692,
+      "learning_rate": 1.412865158611179e-05,
+      "loss": 0.6698,
+      "num_input_tokens_seen": 21523456,
+      "step": 4020
+    },
+    {
+      "epoch": 0.644,
+      "grad_norm": 0.9636697769165039,
+      "learning_rate": 1.4072105833726684e-05,
+      "loss": 0.5917,
+      "num_input_tokens_seen": 21554320,
+      "step": 4025
+    },
+    {
+      "epoch": 0.6448,
+      "grad_norm": 0.8568287491798401,
+      "learning_rate": 1.401562910786034e-05,
+      "loss": 0.7332,
+      "num_input_tokens_seen": 21584496,
+      "step": 4030
+    },
+    {
+      "epoch": 0.6456,
+      "grad_norm": 0.7950714230537415,
+      "learning_rate": 1.3959221765250469e-05,
+      "loss": 0.6826,
+      "num_input_tokens_seen": 21615104,
+      "step": 4035
+    },
+    {
+      "epoch": 0.6464,
+      "grad_norm": 0.9343571662902832,
+      "learning_rate": 1.3902884162196508e-05,
+      "loss": 0.7349,
+      "num_input_tokens_seen": 21642144,
+      "step": 4040
+    },
+    {
+      "epoch": 0.6472,
+      "grad_norm": 0.8434100151062012,
+      "learning_rate": 1.3846616654557362e-05,
+      "loss": 0.6341,
+      "num_input_tokens_seen": 21671408,
+      "step": 4045
+    },
+    {
+      "epoch": 0.648,
+      "grad_norm": 1.1461455821990967,
+      "learning_rate": 1.3790419597749199e-05,
+      "loss": 0.7531,
+      "num_input_tokens_seen": 21698880,
+      "step": 4050
+    },
+    {
+      "epoch": 0.6488,
+      "grad_norm": 1.261234164237976,
+      "learning_rate": 1.3734293346743168e-05,
+      "loss": 0.6284,
+      "num_input_tokens_seen": 21727280,
+      "step": 4055
+    },
+    {
+      "epoch": 0.6496,
+      "grad_norm": 1.0802944898605347,
+      "learning_rate": 1.367823825606319e-05,
+      "loss": 0.7148,
+      "num_input_tokens_seen": 21751824,
+      "step": 4060
+    },
+    {
+      "epoch": 0.6504,
+      "grad_norm": 1.1353379487991333,
+      "learning_rate": 1.3622254679783663e-05,
+      "loss": 0.7022,
+      "num_input_tokens_seen": 21782080,
+      "step": 4065
+    },
+    {
+      "epoch": 0.6512,
+      "grad_norm": 1.0912383794784546,
+      "learning_rate": 1.3566342971527291e-05,
+      "loss": 0.7308,
+      "num_input_tokens_seen": 21809376,
+      "step": 4070
+    },
+    {
+      "epoch": 0.652,
+      "grad_norm": 1.0142539739608765,
+      "learning_rate": 1.3510503484462805e-05,
+      "loss": 0.7338,
+      "num_input_tokens_seen": 21836240,
+      "step": 4075
+    },
+    {
+      "epoch": 0.6528,
+      "grad_norm": 1.0957950353622437,
+      "learning_rate": 1.3454736571302763e-05,
+      "loss": 0.6941,
+      "num_input_tokens_seen": 21862768,
+      "step": 4080
+    },
+    {
+      "epoch": 0.6536,
+      "grad_norm": 0.9035006761550903,
+      "learning_rate": 1.3399042584301298e-05,
+      "loss": 0.7218,
+      "num_input_tokens_seen": 21890304,
+      "step": 4085
+    },
+    {
+      "epoch": 0.6544,
+      "grad_norm": 1.0284723043441772,
+      "learning_rate": 1.3343421875251888e-05,
+      "loss": 0.8144,
+      "num_input_tokens_seen": 21912192,
+      "step": 4090
+    },
+    {
+      "epoch": 0.6552,
+      "grad_norm": 1.0489941835403442,
+      "learning_rate": 1.3287874795485167e-05,
+      "loss": 0.8236,
+      "num_input_tokens_seen": 21939984,
+      "step": 4095
+    },
+    {
+      "epoch": 0.656,
+      "grad_norm": 0.9491598606109619,
+      "learning_rate": 1.3232401695866687e-05,
+      "loss": 0.6512,
+      "num_input_tokens_seen": 21967168,
+      "step": 4100
+    },
+    {
+      "epoch": 0.6568,
+      "grad_norm": 1.0019705295562744,
+      "learning_rate": 1.3177002926794685e-05,
+      "loss": 0.7271,
+      "num_input_tokens_seen": 21999904,
+      "step": 4105
+    },
+    {
+      "epoch": 0.6576,
+      "grad_norm": 1.0153288841247559,
+      "learning_rate": 1.3121678838197909e-05,
+      "loss": 0.635,
+      "num_input_tokens_seen": 22028272,
+      "step": 4110
+    },
+    {
+      "epoch": 0.6584,
+      "grad_norm": 0.8823714852333069,
+      "learning_rate": 1.3066429779533351e-05,
+      "loss": 0.6708,
+      "num_input_tokens_seen": 22052224,
+      "step": 4115
+    },
+    {
+      "epoch": 0.6592,
+      "grad_norm": 0.8675245642662048,
+      "learning_rate": 1.3011256099784103e-05,
+      "loss": 0.5916,
+      "num_input_tokens_seen": 22081360,
+      "step": 4120
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 1.1484395265579224,
+      "learning_rate": 1.2956158147457115e-05,
+      "loss": 0.7399,
+      "num_input_tokens_seen": 22112080,
+      "step": 4125
+    },
+    {
+      "epoch": 0.6608,
+      "grad_norm": 1.2655909061431885,
+      "learning_rate": 1.2901136270580993e-05,
+      "loss": 0.6543,
+      "num_input_tokens_seen": 22139792,
+      "step": 4130
+    },
+    {
+      "epoch": 0.6616,
+      "grad_norm": 1.2049787044525146,
+      "learning_rate": 1.2846190816703835e-05,
+      "loss": 0.6808,
+      "num_input_tokens_seen": 22163136,
+      "step": 4135
+    },
+    {
+      "epoch": 0.6624,
+      "grad_norm": 0.7781268358230591,
+      "learning_rate": 1.279132213289096e-05,
+      "loss": 0.6939,
+      "num_input_tokens_seen": 22192464,
+      "step": 4140
+    },
+    {
+      "epoch": 0.6632,
+      "grad_norm": 1.1952673196792603,
+      "learning_rate": 1.273653056572282e-05,
+      "loss": 0.6628,
+      "num_input_tokens_seen": 22219424,
+      "step": 4145
+    },
+    {
+      "epoch": 0.664,
+      "grad_norm": 1.0534050464630127,
+      "learning_rate": 1.2681816461292715e-05,
+      "loss": 0.686,
+      "num_input_tokens_seen": 22244496,
+      "step": 4150
+    },
+    {
+      "epoch": 0.6648,
+      "grad_norm": 1.18624746799469,
+      "learning_rate": 1.2627180165204671e-05,
+      "loss": 0.7278,
+      "num_input_tokens_seen": 22271600,
+      "step": 4155
+    },
+    {
+      "epoch": 0.6656,
+      "grad_norm": 0.8680349588394165,
+      "learning_rate": 1.257262202257124e-05,
+      "loss": 0.6531,
+      "num_input_tokens_seen": 22298080,
+      "step": 4160
+    },
+    {
+      "epoch": 0.6664,
+      "grad_norm": 0.90425705909729,
+      "learning_rate": 1.251814237801128e-05,
+      "loss": 0.5756,
+      "num_input_tokens_seen": 22324832,
+      "step": 4165
+    },
+    {
+      "epoch": 0.6672,
+      "grad_norm": 1.0510259866714478,
+      "learning_rate": 1.246374157564785e-05,
+      "loss": 0.647,
+      "num_input_tokens_seen": 22353728,
+      "step": 4170
+    },
+    {
+      "epoch": 0.668,
+      "grad_norm": 1.1430630683898926,
+      "learning_rate": 1.2409419959105981e-05,
+      "loss": 0.7024,
+      "num_input_tokens_seen": 22374880,
+      "step": 4175
+    },
+    {
+      "epoch": 0.6688,
+      "grad_norm": 0.8265404105186462,
+      "learning_rate": 1.2355177871510538e-05,
+      "loss": 0.7661,
+      "num_input_tokens_seen": 22402288,
+      "step": 4180
+    },
+    {
+      "epoch": 0.6696,
+      "grad_norm": 0.8584622144699097,
+      "learning_rate": 1.2301015655484006e-05,
+      "loss": 0.6462,
+      "num_input_tokens_seen": 22430240,
+      "step": 4185
+    },
+    {
+      "epoch": 0.6704,
+      "grad_norm": 1.0526131391525269,
+      "learning_rate": 1.2246933653144385e-05,
+      "loss": 0.6487,
+      "num_input_tokens_seen": 22454800,
+      "step": 4190
+    },
+    {
+      "epoch": 0.6712,
+      "grad_norm": 1.1912094354629517,
+      "learning_rate": 1.2192932206103e-05,
+      "loss": 0.7369,
+      "num_input_tokens_seen": 22482528,
+      "step": 4195
+    },
+    {
+      "epoch": 0.672,
+      "grad_norm": 0.8804354071617126,
+      "learning_rate": 1.2139011655462337e-05,
+      "loss": 0.6942,
+      "num_input_tokens_seen": 22508976,
+      "step": 4200
+    },
+    {
+      "epoch": 0.6728,
+      "grad_norm": 0.9333446025848389,
+      "learning_rate": 1.2085172341813911e-05,
+      "loss": 0.7691,
+      "num_input_tokens_seen": 22538976,
+      "step": 4205
+    },
+    {
+      "epoch": 0.6736,
+      "grad_norm": 0.8501102328300476,
+      "learning_rate": 1.2031414605236066e-05,
+      "loss": 0.5865,
+      "num_input_tokens_seen": 22566368,
+      "step": 4210
+    },
+    {
+      "epoch": 0.6744,
+      "grad_norm": 0.89410001039505,
+      "learning_rate": 1.1977738785291895e-05,
+      "loss": 0.6916,
+      "num_input_tokens_seen": 22592656,
+      "step": 4215
+    },
+    {
+      "epoch": 0.6752,
+      "grad_norm": 0.9672756195068359,
+      "learning_rate": 1.1924145221027047e-05,
+      "loss": 0.7436,
+      "num_input_tokens_seen": 22619872,
+      "step": 4220
+    },
+    {
+      "epoch": 0.676,
+      "grad_norm": 0.8680210113525391,
+      "learning_rate": 1.1870634250967605e-05,
+      "loss": 0.6728,
+      "num_input_tokens_seen": 22650320,
+      "step": 4225
+    },
+    {
+      "epoch": 0.6768,
+      "grad_norm": 1.0170356035232544,
+      "learning_rate": 1.1817206213117946e-05,
+      "loss": 0.728,
+      "num_input_tokens_seen": 22676896,
+      "step": 4230
+    },
+    {
+      "epoch": 0.6776,
+      "grad_norm": 1.0950289964675903,
+      "learning_rate": 1.1763861444958573e-05,
+      "loss": 0.6696,
+      "num_input_tokens_seen": 22702352,
+      "step": 4235
+    },
+    {
+      "epoch": 0.6784,
+      "grad_norm": 1.2183908224105835,
+      "learning_rate": 1.1710600283444047e-05,
+      "loss": 0.7827,
+      "num_input_tokens_seen": 22728288,
+      "step": 4240
+    },
+    {
+      "epoch": 0.6792,
+      "grad_norm": 0.9134287238121033,
+      "learning_rate": 1.1657423065000811e-05,
+      "loss": 0.7166,
+      "num_input_tokens_seen": 22757632,
+      "step": 4245
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 1.0772439241409302,
+      "learning_rate": 1.1604330125525079e-05,
+      "loss": 0.7143,
+      "num_input_tokens_seen": 22783440,
+      "step": 4250
+    },
+    {
+      "epoch": 0.6808,
+      "grad_norm": 1.003915786743164,
+      "learning_rate": 1.155132180038072e-05,
+      "loss": 0.82,
+      "num_input_tokens_seen": 22809616,
+      "step": 4255
+    },
+    {
+      "epoch": 0.6816,
+      "grad_norm": 0.9822829961776733,
+      "learning_rate": 1.1498398424397106e-05,
+      "loss": 0.7297,
+      "num_input_tokens_seen": 22835792,
+      "step": 4260
+    },
+    {
+      "epoch": 0.6824,
+      "grad_norm": 0.855888307094574,
+      "learning_rate": 1.1445560331867053e-05,
+      "loss": 0.6956,
+      "num_input_tokens_seen": 22864560,
+      "step": 4265
+    },
+    {
+      "epoch": 0.6832,
+      "grad_norm": 0.850237250328064,
+      "learning_rate": 1.1392807856544683e-05,
+      "loss": 0.7157,
+      "num_input_tokens_seen": 22892912,
+      "step": 4270
+    },
+    {
+      "epoch": 0.684,
+      "grad_norm": 1.4631861448287964,
+      "learning_rate": 1.1340141331643276e-05,
+      "loss": 0.753,
+      "num_input_tokens_seen": 22912640,
+      "step": 4275
+    },
+    {
+      "epoch": 0.6848,
+      "grad_norm": 1.1579556465148926,
+      "learning_rate": 1.1287561089833248e-05,
+      "loss": 0.7247,
+      "num_input_tokens_seen": 22937072,
+      "step": 4280
+    },
+    {
+      "epoch": 0.6856,
+      "grad_norm": 1.3090944290161133,
+      "learning_rate": 1.1235067463239967e-05,
+      "loss": 0.7671,
+      "num_input_tokens_seen": 22961104,
+      "step": 4285
+    },
+    {
+      "epoch": 0.6864,
+      "grad_norm": 0.9593985676765442,
+      "learning_rate": 1.1182660783441718e-05,
+      "loss": 0.6771,
+      "num_input_tokens_seen": 22983744,
+      "step": 4290
+    },
+    {
+      "epoch": 0.6872,
+      "grad_norm": 0.6641414165496826,
+      "learning_rate": 1.1130341381467569e-05,
+      "loss": 0.7179,
+      "num_input_tokens_seen": 23010208,
+      "step": 4295
+    },
+    {
+      "epoch": 0.688,
+      "grad_norm": 0.6973831057548523,
+      "learning_rate": 1.107810958779531e-05,
+      "loss": 0.5792,
+      "num_input_tokens_seen": 23043392,
+      "step": 4300
+    },
+    {
+      "epoch": 0.6888,
+      "grad_norm": 0.9109097123146057,
+      "learning_rate": 1.1025965732349316e-05,
+      "loss": 0.6574,
+      "num_input_tokens_seen": 23074928,
+      "step": 4305
+    },
+    {
+      "epoch": 0.6896,
+      "grad_norm": 0.8845970034599304,
+      "learning_rate": 1.0973910144498534e-05,
+      "loss": 0.6544,
+      "num_input_tokens_seen": 23105728,
+      "step": 4310
+    },
+    {
+      "epoch": 0.6904,
+      "grad_norm": 1.181096076965332,
+      "learning_rate": 1.0921943153054343e-05,
+      "loss": 0.6503,
+      "num_input_tokens_seen": 23132768,
+      "step": 4315
+    },
+    {
+      "epoch": 0.6912,
+      "grad_norm": 0.785658061504364,
+      "learning_rate": 1.0870065086268505e-05,
+      "loss": 0.6502,
+      "num_input_tokens_seen": 23160080,
+      "step": 4320
+    },
+    {
+      "epoch": 0.692,
+      "grad_norm": 1.1847856044769287,
+      "learning_rate": 1.0818276271831093e-05,
+      "loss": 0.7114,
+      "num_input_tokens_seen": 23187696,
+      "step": 4325
+    },
+    {
+      "epoch": 0.6928,
+      "grad_norm": 1.1500554084777832,
+      "learning_rate": 1.0766577036868395e-05,
+      "loss": 0.6546,
+      "num_input_tokens_seen": 23211904,
+      "step": 4330
+    },
+    {
+      "epoch": 0.6936,
+      "grad_norm": 0.9399601221084595,
+      "learning_rate": 1.0714967707940875e-05,
+      "loss": 0.6965,
+      "num_input_tokens_seen": 23238144,
+      "step": 4335
+    },
+    {
+      "epoch": 0.6944,
+      "grad_norm": 0.928415060043335,
+      "learning_rate": 1.0663448611041113e-05,
+      "loss": 0.6232,
+      "num_input_tokens_seen": 23267104,
+      "step": 4340
+    },
+    {
+      "epoch": 0.6952,
+      "grad_norm": 1.2702407836914062,
+      "learning_rate": 1.0612020071591722e-05,
+      "loss": 0.6686,
+      "num_input_tokens_seen": 23298976,
+      "step": 4345
+    },
+    {
+      "epoch": 0.696,
+      "grad_norm": 1.1251682043075562,
+      "learning_rate": 1.0560682414443315e-05,
+      "loss": 0.6975,
+      "num_input_tokens_seen": 23329552,
+      "step": 4350
+    },
+    {
+      "epoch": 0.6968,
+      "grad_norm": 0.8569183945655823,
+      "learning_rate": 1.0509435963872422e-05,
+      "loss": 0.7017,
+      "num_input_tokens_seen": 23359664,
+      "step": 4355
+    },
+    {
+      "epoch": 0.6976,
+      "grad_norm": 0.9474520683288574,
+      "learning_rate": 1.0458281043579482e-05,
+      "loss": 0.6856,
+      "num_input_tokens_seen": 23386320,
+      "step": 4360
+    },
+    {
+      "epoch": 0.6984,
+      "grad_norm": 1.3578598499298096,
+      "learning_rate": 1.0407217976686775e-05,
+      "loss": 0.6739,
+      "num_input_tokens_seen": 23416512,
+      "step": 4365
+    },
+    {
+      "epoch": 0.6992,
+      "grad_norm": 1.1748965978622437,
+      "learning_rate": 1.0356247085736386e-05,
+      "loss": 0.6803,
+      "num_input_tokens_seen": 23439904,
+      "step": 4370
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.7888918519020081,
+      "learning_rate": 1.0305368692688174e-05,
+      "loss": 0.7111,
+      "num_input_tokens_seen": 23466256,
+      "step": 4375
+    },
+    {
+      "epoch": 0.7008,
+      "grad_norm": 1.137215256690979,
+      "learning_rate": 1.0254583118917698e-05,
+      "loss": 0.762,
+      "num_input_tokens_seen": 23493536,
+      "step": 4380
+    },
+    {
+      "epoch": 0.7016,
+      "grad_norm": 1.337811827659607,
+      "learning_rate": 1.020389068521426e-05,
+      "loss": 0.6206,
+      "num_input_tokens_seen": 23515632,
+      "step": 4385
+    },
+    {
+      "epoch": 0.7024,
+      "grad_norm": 1.0022634267807007,
+      "learning_rate": 1.0153291711778826e-05,
+      "loss": 0.6711,
+      "num_input_tokens_seen": 23541152,
+      "step": 4390
+    },
+    {
+      "epoch": 0.7032,
+      "grad_norm": 1.339572548866272,
+      "learning_rate": 1.0102786518221997e-05,
+      "loss": 0.7606,
+      "num_input_tokens_seen": 23565424,
+      "step": 4395
+    },
+    {
+      "epoch": 0.704,
+      "grad_norm": 0.9340786337852478,
+      "learning_rate": 1.0052375423562038e-05,
+      "loss": 0.7306,
+      "num_input_tokens_seen": 23592256,
+      "step": 4400
+    },
+    {
+      "epoch": 0.7048,
+      "grad_norm": 1.1283642053604126,
+      "learning_rate": 1.0002058746222806e-05,
+      "loss": 0.6121,
+      "num_input_tokens_seen": 23622240,
+      "step": 4405
+    },
+    {
+      "epoch": 0.7056,
+      "grad_norm": 0.7319700717926025,
+      "learning_rate": 9.951836804031794e-06,
+      "loss": 0.6273,
+      "num_input_tokens_seen": 23650544,
+      "step": 4410
+    },
+    {
+      "epoch": 0.7064,
+      "grad_norm": 1.1631896495819092,
+      "learning_rate": 9.90170991421808e-06,
+      "loss": 0.7261,
+      "num_input_tokens_seen": 23676016,
+      "step": 4415
+    },
+    {
+      "epoch": 0.7072,
+      "grad_norm": 0.7371265292167664,
+      "learning_rate": 9.851678393410343e-06,
+      "loss": 0.8013,
+      "num_input_tokens_seen": 23701232,
+      "step": 4420
+    },
+    {
+      "epoch": 0.708,
+      "grad_norm": 0.8485360741615295,
+      "learning_rate": 9.801742557634872e-06,
+      "loss": 0.7189,
+      "num_input_tokens_seen": 23731984,
+      "step": 4425
+    },
+    {
+      "epoch": 0.7088,
+      "grad_norm": 0.958996057510376,
+      "learning_rate": 9.751902722313527e-06,
+      "loss": 0.7397,
+      "num_input_tokens_seen": 23756560,
+      "step": 4430
+    },
+    {
+      "epoch": 0.7096,
+      "grad_norm": 0.9431763887405396,
+      "learning_rate": 9.702159202261801e-06,
+      "loss": 0.625,
+      "num_input_tokens_seen": 23785504,
+      "step": 4435
+    },
+    {
+      "epoch": 0.7104,
+      "grad_norm": 1.2089407444000244,
+      "learning_rate": 9.652512311686809e-06,
+      "loss": 0.7958,
+      "num_input_tokens_seen": 23811840,
+      "step": 4440
+    },
+    {
+      "epoch": 0.7112,
+      "grad_norm": 1.212649941444397,
+      "learning_rate": 9.602962364185286e-06,
+      "loss": 0.7092,
+      "num_input_tokens_seen": 23834688,
+      "step": 4445
+    },
+    {
+      "epoch": 0.712,
+      "grad_norm": 0.8656122088432312,
+      "learning_rate": 9.553509672741645e-06,
+      "loss": 0.6516,
+      "num_input_tokens_seen": 23858736,
+      "step": 4450
+    },
+    {
+      "epoch": 0.7128,
+      "grad_norm": 0.8871904611587524,
+      "learning_rate": 9.504154549725943e-06,
+      "loss": 0.6276,
+      "num_input_tokens_seen": 23883696,
+      "step": 4455
+    },
+    {
+      "epoch": 0.7136,
+      "grad_norm": 0.8539274334907532,
+      "learning_rate": 9.454897306891972e-06,
+      "loss": 0.6741,
+      "num_input_tokens_seen": 23909904,
+      "step": 4460
+    },
+    {
+      "epoch": 0.7144,
+      "grad_norm": 1.0730023384094238,
+      "learning_rate": 9.405738255375244e-06,
+      "loss": 0.7054,
+      "num_input_tokens_seen": 23933056,
+      "step": 4465
+    },
+    {
+      "epoch": 0.7152,
+      "grad_norm": 1.2047233581542969,
+      "learning_rate": 9.356677705691058e-06,
+      "loss": 0.7593,
+      "num_input_tokens_seen": 23957440,
+      "step": 4470
+    },
+    {
+      "epoch": 0.716,
+      "grad_norm": 0.8580662608146667,
+      "learning_rate": 9.307715967732491e-06,
+      "loss": 0.6264,
+      "num_input_tokens_seen": 23985088,
+      "step": 4475
+    },
+    {
+      "epoch": 0.7168,
+      "grad_norm": 0.884903073310852,
+      "learning_rate": 9.258853350768499e-06,
+      "loss": 0.6596,
+      "num_input_tokens_seen": 24010448,
+      "step": 4480
+    },
+    {
+      "epoch": 0.7176,
+      "grad_norm": 0.9185977578163147,
+      "learning_rate": 9.210090163441929e-06,
+      "loss": 0.7053,
+      "num_input_tokens_seen": 24035040,
+      "step": 4485
+    },
+    {
+      "epoch": 0.7184,
+      "grad_norm": 1.3897747993469238,
+      "learning_rate": 9.161426713767574e-06,
+      "loss": 0.6769,
+      "num_input_tokens_seen": 24057872,
+      "step": 4490
+    },
+    {
+      "epoch": 0.7192,
+      "grad_norm": 1.220688819885254,
+      "learning_rate": 9.112863309130235e-06,
+      "loss": 0.7486,
+      "num_input_tokens_seen": 24077920,
+      "step": 4495
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.9036649465560913,
+      "learning_rate": 9.064400256282757e-06,
+      "loss": 0.765,
+      "num_input_tokens_seen": 24104320,
+      "step": 4500
+    },
+    {
+      "epoch": 0.7208,
+      "grad_norm": 0.7980133891105652,
+      "learning_rate": 9.016037861344129e-06,
+      "loss": 0.653,
+      "num_input_tokens_seen": 24134144,
+      "step": 4505
+    },
+    {
+      "epoch": 0.7216,
+      "grad_norm": 0.7849147915840149,
+      "learning_rate": 8.967776429797528e-06,
+      "loss": 0.6412,
+      "num_input_tokens_seen": 24164576,
+      "step": 4510
+    },
+    {
+      "epoch": 0.7224,
+      "grad_norm": 0.8543937802314758,
+      "learning_rate": 8.919616266488373e-06,
+      "loss": 0.7113,
+      "num_input_tokens_seen": 24192736,
+      "step": 4515
+    },
+    {
+      "epoch": 0.7232,
+      "grad_norm": 0.9191213250160217,
+      "learning_rate": 8.871557675622441e-06,
+      "loss": 0.8171,
+      "num_input_tokens_seen": 24218064,
+      "step": 4520
+    },
+    {
+      "epoch": 0.724,
+      "grad_norm": 1.1177440881729126,
+      "learning_rate": 8.8236009607639e-06,
+      "loss": 0.7845,
+      "num_input_tokens_seen": 24244832,
+      "step": 4525
+    },
+    {
+      "epoch": 0.7248,
+      "grad_norm": 0.899111807346344,
+      "learning_rate": 8.775746424833427e-06,
+      "loss": 0.7025,
+      "num_input_tokens_seen": 24272848,
+      "step": 4530
+    },
+    {
+      "epoch": 0.7256,
+      "grad_norm": 1.1424217224121094,
+      "learning_rate": 8.727994370106288e-06,
+      "loss": 0.868,
+      "num_input_tokens_seen": 24298240,
+      "step": 4535
+    },
+    {
+      "epoch": 0.7264,
+      "grad_norm": 0.9559049010276794,
+      "learning_rate": 8.680345098210408e-06,
+      "loss": 0.6285,
+      "num_input_tokens_seen": 24327776,
+      "step": 4540
+    },
+    {
+      "epoch": 0.7272,
+      "grad_norm": 0.9032924771308899,
+      "learning_rate": 8.632798910124492e-06,
+      "loss": 0.6583,
+      "num_input_tokens_seen": 24355424,
+      "step": 4545
+    },
+    {
+      "epoch": 0.728,
+      "grad_norm": 1.056780457496643,
+      "learning_rate": 8.585356106176094e-06,
+      "loss": 0.756,
+      "num_input_tokens_seen": 24381104,
+      "step": 4550
+    },
+    {
+      "epoch": 0.7288,
+      "grad_norm": 1.2001997232437134,
+      "learning_rate": 8.538016986039754e-06,
+      "loss": 0.7739,
+      "num_input_tokens_seen": 24403760,
+      "step": 4555
+    },
+    {
+      "epoch": 0.7296,
+      "grad_norm": 1.1103582382202148,
+      "learning_rate": 8.49078184873508e-06,
+      "loss": 0.6998,
+      "num_input_tokens_seen": 24431280,
+      "step": 4560
+    },
+    {
+      "epoch": 0.7304,
+      "grad_norm": 1.0271393060684204,
+      "learning_rate": 8.443650992624877e-06,
+      "loss": 0.723,
+      "num_input_tokens_seen": 24459120,
+      "step": 4565
+    },
+    {
+      "epoch": 0.7312,
+      "grad_norm": 0.7871257066726685,
+      "learning_rate": 8.39662471541325e-06,
+      "loss": 0.7225,
+      "num_input_tokens_seen": 24485152,
+      "step": 4570
+    },
+    {
+      "epoch": 0.732,
+      "grad_norm": 0.8319628238677979,
+      "learning_rate": 8.34970331414371e-06,
+      "loss": 0.5801,
+      "num_input_tokens_seen": 24512416,
+      "step": 4575
+    },
+    {
+      "epoch": 0.7328,
+      "grad_norm": 0.7009981274604797,
+      "learning_rate": 8.302887085197341e-06,
+      "loss": 0.6043,
+      "num_input_tokens_seen": 24543328,
+      "step": 4580
+    },
+    {
+      "epoch": 0.7336,
+      "grad_norm": 1.0223398208618164,
+      "learning_rate": 8.256176324290885e-06,
+      "loss": 0.6533,
+      "num_input_tokens_seen": 24566000,
+      "step": 4585
+    },
+    {
+      "epoch": 0.7344,
+      "grad_norm": 1.127424955368042,
+      "learning_rate": 8.209571326474896e-06,
+      "loss": 0.6906,
+      "num_input_tokens_seen": 24594080,
+      "step": 4590
+    },
+    {
+      "epoch": 0.7352,
+      "grad_norm": 0.9771124124526978,
+      "learning_rate": 8.163072386131876e-06,
+      "loss": 0.6661,
+      "num_input_tokens_seen": 24621424,
+      "step": 4595
+    },
+    {
+      "epoch": 0.736,
+      "grad_norm": 0.859312117099762,
+      "learning_rate": 8.116679796974392e-06,
+      "loss": 0.6663,
+      "num_input_tokens_seen": 24644288,
+      "step": 4600
+    },
+    {
+      "epoch": 0.7368,
+      "grad_norm": 1.3819899559020996,
+      "learning_rate": 8.070393852043251e-06,
+      "loss": 0.7064,
+      "num_input_tokens_seen": 24674048,
+      "step": 4605
+    },
+    {
+      "epoch": 0.7376,
+      "grad_norm": 1.034734845161438,
+      "learning_rate": 8.024214843705646e-06,
+      "loss": 0.6837,
+      "num_input_tokens_seen": 24696320,
+      "step": 4610
+    },
+    {
+      "epoch": 0.7384,
+      "grad_norm": 0.9610295295715332,
+      "learning_rate": 7.978143063653298e-06,
+      "loss": 0.5342,
+      "num_input_tokens_seen": 24729280,
+      "step": 4615
+    },
+    {
+      "epoch": 0.7392,
+      "grad_norm": 1.166585922241211,
+      "learning_rate": 7.93217880290059e-06,
+      "loss": 0.6907,
+      "num_input_tokens_seen": 24758080,
+      "step": 4620
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 1.1341148614883423,
+      "learning_rate": 7.886322351782783e-06,
+      "loss": 0.6856,
+      "num_input_tokens_seen": 24787472,
+      "step": 4625
+    },
+    {
+      "epoch": 0.7408,
+      "grad_norm": 0.9481520056724548,
+      "learning_rate": 7.840573999954153e-06,
+      "loss": 0.713,
+      "num_input_tokens_seen": 24815936,
+      "step": 4630
+    },
+    {
+      "epoch": 0.7416,
+      "grad_norm": 1.2403899431228638,
+      "learning_rate": 7.79493403638614e-06,
+      "loss": 0.7692,
+      "num_input_tokens_seen": 24840096,
+      "step": 4635
+    },
+    {
+      "epoch": 0.7424,
+      "grad_norm": 0.9576728343963623,
+      "learning_rate": 7.749402749365572e-06,
+      "loss": 0.7177,
+      "num_input_tokens_seen": 24866480,
+      "step": 4640
+    },
+    {
+      "epoch": 0.7432,
+      "grad_norm": 1.0239994525909424,
+      "learning_rate": 7.703980426492791e-06,
+      "loss": 0.7124,
+      "num_input_tokens_seen": 24889456,
+      "step": 4645
+    },
+    {
+      "epoch": 0.744,
+      "grad_norm": 1.0492584705352783,
+      "learning_rate": 7.658667354679883e-06,
+      "loss": 0.7038,
+      "num_input_tokens_seen": 24913824,
+      "step": 4650
+    },
+    {
+      "epoch": 0.7448,
+      "grad_norm": 1.1247596740722656,
+      "learning_rate": 7.613463820148831e-06,
+      "loss": 0.6662,
+      "num_input_tokens_seen": 24940880,
+      "step": 4655
+    },
+    {
+      "epoch": 0.7456,
+      "grad_norm": 1.2390748262405396,
+      "learning_rate": 7.568370108429732e-06,
+      "loss": 0.7949,
+      "num_input_tokens_seen": 24965696,
+      "step": 4660
+    },
+    {
+      "epoch": 0.7464,
+      "grad_norm": 0.7792567610740662,
+      "learning_rate": 7.523386504358984e-06,
+      "loss": 0.7146,
+      "num_input_tokens_seen": 24992096,
+      "step": 4665
+    },
+    {
+      "epoch": 0.7472,
+      "grad_norm": 0.9417341351509094,
+      "learning_rate": 7.478513292077463e-06,
+      "loss": 0.669,
+      "num_input_tokens_seen": 25024320,
+      "step": 4670
+    },
+    {
+      "epoch": 0.748,
+      "grad_norm": 1.226563572883606,
+      "learning_rate": 7.433750755028773e-06,
+      "loss": 0.7789,
+      "num_input_tokens_seen": 25049152,
+      "step": 4675
+    },
+    {
+      "epoch": 0.7488,
+      "grad_norm": 0.8685075640678406,
+      "learning_rate": 7.389099175957429e-06,
+      "loss": 0.6992,
+      "num_input_tokens_seen": 25077328,
+      "step": 4680
+    },
+    {
+      "epoch": 0.7496,
+      "grad_norm": 0.7221574187278748,
+      "learning_rate": 7.344558836907067e-06,
+      "loss": 0.6421,
+      "num_input_tokens_seen": 25105008,
+      "step": 4685
+    },
+    {
+      "epoch": 0.7504,
+      "grad_norm": 0.8624604940414429,
+      "learning_rate": 7.300130019218687e-06,
+      "loss": 0.7656,
+      "num_input_tokens_seen": 25131392,
+      "step": 4690
+    },
+    {
+      "epoch": 0.7512,
+      "grad_norm": 1.0061527490615845,
+      "learning_rate": 7.255813003528833e-06,
+      "loss": 0.6506,
+      "num_input_tokens_seen": 25159984,
+      "step": 4695
+    },
+    {
+      "epoch": 0.752,
+      "grad_norm": 1.0879433155059814,
+      "learning_rate": 7.211608069767867e-06,
+      "loss": 0.6253,
+      "num_input_tokens_seen": 25188192,
+      "step": 4700
+    },
+    {
+      "epoch": 0.7528,
+      "grad_norm": 1.2521827220916748,
+      "learning_rate": 7.1675154971581785e-06,
+      "loss": 0.6776,
+      "num_input_tokens_seen": 25215360,
+      "step": 4705
+    },
+    {
+      "epoch": 0.7536,
+      "grad_norm": 1.4772545099258423,
+      "learning_rate": 7.123535564212422e-06,
+      "loss": 0.8286,
+      "num_input_tokens_seen": 25240384,
+      "step": 4710
+    },
+    {
+      "epoch": 0.7544,
+      "grad_norm": 1.0587224960327148,
+      "learning_rate": 7.079668548731758e-06,
+      "loss": 0.6152,
+      "num_input_tokens_seen": 25263104,
+      "step": 4715
+    },
+    {
+      "epoch": 0.7552,
+      "grad_norm": 1.562467336654663,
+      "learning_rate": 7.035914727804085e-06,
+      "loss": 0.7227,
+      "num_input_tokens_seen": 25288176,
+      "step": 4720
+    },
+    {
+      "epoch": 0.756,
+      "grad_norm": 1.3081474304199219,
+      "learning_rate": 6.992274377802327e-06,
+      "loss": 0.6808,
+      "num_input_tokens_seen": 25313536,
+      "step": 4725
+    },
+    {
+      "epoch": 0.7568,
+      "grad_norm": 1.1339465379714966,
+      "learning_rate": 6.94874777438265e-06,
+      "loss": 0.7039,
+      "num_input_tokens_seen": 25339744,
+      "step": 4730
+    },
+    {
+      "epoch": 0.7576,
+      "grad_norm": 0.8575751185417175,
+      "learning_rate": 6.905335192482737e-06,
+      "loss": 0.7081,
+      "num_input_tokens_seen": 25367440,
+      "step": 4735
+    },
+    {
+      "epoch": 0.7584,
+      "grad_norm": 0.9493206143379211,
+      "learning_rate": 6.862036906320058e-06,
+      "loss": 0.6139,
+      "num_input_tokens_seen": 25395952,
+      "step": 4740
+    },
+    {
+      "epoch": 0.7592,
+      "grad_norm": 1.0198074579238892,
+      "learning_rate": 6.818853189390104e-06,
+      "loss": 0.8142,
+      "num_input_tokens_seen": 25421744,
+      "step": 4745
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.8722714185714722,
+      "learning_rate": 6.775784314464717e-06,
+      "loss": 0.6538,
+      "num_input_tokens_seen": 25448944,
+      "step": 4750
+    },
+    {
+      "epoch": 0.7608,
+      "grad_norm": 0.750995397567749,
+      "learning_rate": 6.732830553590305e-06,
+      "loss": 0.6409,
+      "num_input_tokens_seen": 25476640,
+      "step": 4755
+    },
+    {
+      "epoch": 0.7616,
+      "grad_norm": 1.152645230293274,
+      "learning_rate": 6.689992178086174e-06,
+      "loss": 0.6814,
+      "num_input_tokens_seen": 25503328,
+      "step": 4760
+    },
+    {
+      "epoch": 0.7624,
+      "grad_norm": 0.9856323599815369,
+      "learning_rate": 6.647269458542793e-06,
+      "loss": 0.739,
+      "num_input_tokens_seen": 25530384,
+      "step": 4765
+    },
+    {
+      "epoch": 0.7632,
+      "grad_norm": 1.0248849391937256,
+      "learning_rate": 6.604662664820063e-06,
+      "loss": 0.6775,
+      "num_input_tokens_seen": 25558880,
+      "step": 4770
+    },
+    {
+      "epoch": 0.764,
+      "grad_norm": 1.1603997945785522,
+      "learning_rate": 6.562172066045655e-06,
+      "loss": 0.8087,
+      "num_input_tokens_seen": 25584016,
+      "step": 4775
+    },
+    {
+      "epoch": 0.7648,
+      "grad_norm": 1.0015392303466797,
+      "learning_rate": 6.519797930613289e-06,
+      "loss": 0.6836,
+      "num_input_tokens_seen": 25611712,
+      "step": 4780
+    },
+    {
+      "epoch": 0.7656,
+      "grad_norm": 0.929892897605896,
+      "learning_rate": 6.4775405261810364e-06,
+      "loss": 0.7174,
+      "num_input_tokens_seen": 25640928,
+      "step": 4785
+    },
+    {
+      "epoch": 0.7664,
+      "grad_norm": 1.0972721576690674,
+      "learning_rate": 6.435400119669618e-06,
+      "loss": 0.6151,
+      "num_input_tokens_seen": 25667376,
+      "step": 4790
+    },
+    {
+      "epoch": 0.7672,
+      "grad_norm": 1.1308437585830688,
+      "learning_rate": 6.3933769772607535e-06,
+      "loss": 0.7291,
+      "num_input_tokens_seen": 25697136,
+      "step": 4795
+    },
+    {
+      "epoch": 0.768,
+      "grad_norm": 0.9035334587097168,
+      "learning_rate": 6.3514713643954475e-06,
+      "loss": 0.7215,
+      "num_input_tokens_seen": 25718912,
+      "step": 4800
+    },
+    {
+      "epoch": 0.7688,
+      "grad_norm": 0.7563897371292114,
+      "learning_rate": 6.309683545772327e-06,
+      "loss": 0.7092,
+      "num_input_tokens_seen": 25746400,
+      "step": 4805
+    },
+    {
+      "epoch": 0.7696,
+      "grad_norm": 0.8460260629653931,
+      "learning_rate": 6.268013785345969e-06,
+      "loss": 0.6675,
+      "num_input_tokens_seen": 25772192,
+      "step": 4810
+    },
+    {
+      "epoch": 0.7704,
+      "grad_norm": 0.8550633788108826,
+      "learning_rate": 6.226462346325221e-06,
+      "loss": 0.7428,
+      "num_input_tokens_seen": 25802256,
+      "step": 4815
+    },
+    {
+      "epoch": 0.7712,
+      "grad_norm": 1.0745741128921509,
+      "learning_rate": 6.185029491171554e-06,
+      "loss": 0.7039,
+      "num_input_tokens_seen": 25829952,
+      "step": 4820
+    },
+    {
+      "epoch": 0.772,
+      "grad_norm": 0.9149506688117981,
+      "learning_rate": 6.143715481597404e-06,
+      "loss": 0.6733,
+      "num_input_tokens_seen": 25854752,
+      "step": 4825
+    },
+    {
+      "epoch": 0.7728,
+      "grad_norm": 1.1725239753723145,
+      "learning_rate": 6.102520578564508e-06,
+      "loss": 0.7872,
+      "num_input_tokens_seen": 25881264,
+      "step": 4830
+    },
+    {
+      "epoch": 0.7736,
+      "grad_norm": 1.1597212553024292,
+      "learning_rate": 6.061445042282271e-06,
+      "loss": 0.7586,
+      "num_input_tokens_seen": 25906064,
+      "step": 4835
+    },
+    {
+      "epoch": 0.7744,
+      "grad_norm": 0.9395809173583984,
+      "learning_rate": 6.020489132206089e-06,
+      "loss": 0.7269,
+      "num_input_tokens_seen": 25931280,
+      "step": 4840
+    },
+    {
+      "epoch": 0.7752,
+      "grad_norm": 0.8174002170562744,
+      "learning_rate": 5.979653107035757e-06,
+      "loss": 0.6304,
+      "num_input_tokens_seen": 25958880,
+      "step": 4845
+    },
+    {
+      "epoch": 0.776,
+      "grad_norm": 0.9226968884468079,
+      "learning_rate": 5.9389372247138e-06,
+      "loss": 0.6855,
+      "num_input_tokens_seen": 25984528,
+      "step": 4850
+    },
+    {
+      "epoch": 0.7768,
+      "grad_norm": 1.0765284299850464,
+      "learning_rate": 5.898341742423865e-06,
+      "loss": 0.7141,
+      "num_input_tokens_seen": 26014272,
+      "step": 4855
+    },
+    {
+      "epoch": 0.7776,
+      "grad_norm": 0.7940208315849304,
+      "learning_rate": 5.857866916589089e-06,
+      "loss": 0.668,
+      "num_input_tokens_seen": 26045888,
+      "step": 4860
+    },
+    {
+      "epoch": 0.7784,
+      "grad_norm": 0.9069024920463562,
+      "learning_rate": 5.81751300287045e-06,
+      "loss": 0.7404,
+      "num_input_tokens_seen": 26069232,
+      "step": 4865
+    },
+    {
+      "epoch": 0.7792,
+      "grad_norm": 1.2687326669692993,
+      "learning_rate": 5.777280256165218e-06,
+      "loss": 0.633,
+      "num_input_tokens_seen": 26095936,
+      "step": 4870
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 1.0579140186309814,
+      "learning_rate": 5.737168930605272e-06,
+      "loss": 0.6365,
+      "num_input_tokens_seen": 26121184,
+      "step": 4875
+    },
+    {
+      "epoch": 0.7808,
+      "grad_norm": 0.8767179846763611,
+      "learning_rate": 5.6971792795555505e-06,
+      "loss": 0.6427,
+      "num_input_tokens_seen": 26147504,
+      "step": 4880
+    },
+    {
+      "epoch": 0.7816,
+      "grad_norm": 0.9713358283042908,
+      "learning_rate": 5.6573115556124325e-06,
+      "loss": 0.6509,
+      "num_input_tokens_seen": 26174208,
+      "step": 4885
+    },
+    {
+      "epoch": 0.7824,
+      "grad_norm": 0.7532449960708618,
+      "learning_rate": 5.617566010602113e-06,
+      "loss": 0.7382,
+      "num_input_tokens_seen": 26200112,
+      "step": 4890
+    },
+    {
+      "epoch": 0.7832,
+      "grad_norm": 1.191658854484558,
+      "learning_rate": 5.577942895579064e-06,
+      "loss": 0.7537,
+      "num_input_tokens_seen": 26227952,
+      "step": 4895
+    },
+    {
+      "epoch": 0.784,
+      "grad_norm": 0.9605312943458557,
+      "learning_rate": 5.538442460824417e-06,
+      "loss": 0.673,
+      "num_input_tokens_seen": 26259392,
+      "step": 4900
+    },
+    {
+      "epoch": 0.7848,
+      "grad_norm": 1.0074589252471924,
+      "learning_rate": 5.499064955844382e-06,
+      "loss": 0.6684,
+      "num_input_tokens_seen": 26285456,
+      "step": 4905
+    },
+    {
+      "epoch": 0.7856,
+      "grad_norm": 0.8559053540229797,
+      "learning_rate": 5.4598106293686916e-06,
+      "loss": 0.7051,
+      "num_input_tokens_seen": 26316544,
+      "step": 4910
+    },
+    {
+      "epoch": 0.7864,
+      "grad_norm": 1.0885223150253296,
+      "learning_rate": 5.420679729348993e-06,
+      "loss": 0.6481,
+      "num_input_tokens_seen": 26341840,
+      "step": 4915
+    },
+    {
+      "epoch": 0.7872,
+      "grad_norm": 1.0094472169876099,
+      "learning_rate": 5.381672502957324e-06,
+      "loss": 0.7953,
+      "num_input_tokens_seen": 26371008,
+      "step": 4920
+    },
+    {
+      "epoch": 0.788,
+      "grad_norm": 1.1841135025024414,
+      "learning_rate": 5.342789196584527e-06,
+      "loss": 0.6966,
+      "num_input_tokens_seen": 26400048,
+      "step": 4925
+    },
+    {
+      "epoch": 0.7888,
+      "grad_norm": 0.9742115139961243,
+      "learning_rate": 5.304030055838705e-06,
+      "loss": 0.6804,
+      "num_input_tokens_seen": 26425408,
+      "step": 4930
+    },
+    {
+      "epoch": 0.7896,
+      "grad_norm": 1.327427625656128,
+      "learning_rate": 5.26539532554364e-06,
+      "loss": 0.6282,
+      "num_input_tokens_seen": 26452352,
+      "step": 4935
+    },
+    {
+      "epoch": 0.7904,
+      "grad_norm": 0.9497706890106201,
+      "learning_rate": 5.226885249737293e-06,
+      "loss": 0.588,
+      "num_input_tokens_seen": 26479456,
+      "step": 4940
+    },
+    {
+      "epoch": 0.7912,
+      "grad_norm": 1.0179611444473267,
+      "learning_rate": 5.1885000716702355e-06,
+      "loss": 0.738,
+      "num_input_tokens_seen": 26504912,
+      "step": 4945
+    },
+    {
+      "epoch": 0.792,
+      "grad_norm": 1.0633511543273926,
+      "learning_rate": 5.150240033804116e-06,
+      "loss": 0.661,
+      "num_input_tokens_seen": 26528320,
+      "step": 4950
+    },
+    {
+      "epoch": 0.7928,
+      "grad_norm": 0.887589693069458,
+      "learning_rate": 5.112105377810128e-06,
+      "loss": 0.8033,
+      "num_input_tokens_seen": 26553984,
+      "step": 4955
+    },
+    {
+      "epoch": 0.7936,
+      "grad_norm": 1.0450173616409302,
+      "learning_rate": 5.074096344567475e-06,
+      "loss": 0.6161,
+      "num_input_tokens_seen": 26582768,
+      "step": 4960
+    },
+    {
+      "epoch": 0.7944,
+      "grad_norm": 1.2726836204528809,
+      "learning_rate": 5.036213174161877e-06,
+      "loss": 0.7286,
+      "num_input_tokens_seen": 26610272,
+      "step": 4965
+    },
+    {
+      "epoch": 0.7952,
+      "grad_norm": 1.0096362829208374,
+      "learning_rate": 4.998456105884025e-06,
+      "loss": 0.7065,
+      "num_input_tokens_seen": 26636352,
+      "step": 4970
+    },
+    {
+      "epoch": 0.796,
+      "grad_norm": 1.1659733057022095,
+      "learning_rate": 4.960825378228082e-06,
+      "loss": 0.6842,
+      "num_input_tokens_seen": 26667824,
+      "step": 4975
+    },
+    {
+      "epoch": 0.7968,
+      "grad_norm": 0.9078534245491028,
+      "learning_rate": 4.9233212288901845e-06,
+      "loss": 0.7069,
+      "num_input_tokens_seen": 26698272,
+      "step": 4980
+    },
+    {
+      "epoch": 0.7976,
+      "grad_norm": 0.8440881967544556,
+      "learning_rate": 4.885943894766909e-06,
+      "loss": 0.5942,
+      "num_input_tokens_seen": 26725984,
+      "step": 4985
+    },
+    {
+      "epoch": 0.7984,
+      "grad_norm": 1.0457020998001099,
+      "learning_rate": 4.848693611953825e-06,
+      "loss": 0.8419,
+      "num_input_tokens_seen": 26751360,
+      "step": 4990
+    },
+    {
+      "epoch": 0.7992,
+      "grad_norm": 0.947726309299469,
+      "learning_rate": 4.811570615743952e-06,
+      "loss": 0.5888,
+      "num_input_tokens_seen": 26782672,
+      "step": 4995
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 1.1138640642166138,
+      "learning_rate": 4.7745751406263165e-06,
+      "loss": 0.6504,
+      "num_input_tokens_seen": 26809120,
+      "step": 5000
+    },
+    {
+      "epoch": 0.8008,
+      "grad_norm": 0.9211150407791138,
+      "learning_rate": 4.737707420284451e-06,
+      "loss": 0.6603,
+      "num_input_tokens_seen": 26839552,
+      "step": 5005
+    },
+    {
+      "epoch": 0.8016,
+      "grad_norm": 1.2926892042160034,
+      "learning_rate": 4.700967687594901e-06,
+      "loss": 0.627,
+      "num_input_tokens_seen": 26864416,
+      "step": 5010
+    },
+    {
+      "epoch": 0.8024,
+      "grad_norm": 0.9436898827552795,
+      "learning_rate": 4.664356174625795e-06,
+      "loss": 0.6509,
+      "num_input_tokens_seen": 26890368,
+      "step": 5015
+    },
+    {
+      "epoch": 0.8032,
+      "grad_norm": 0.8215711712837219,
+      "learning_rate": 4.627873112635345e-06,
+      "loss": 0.6673,
+      "num_input_tokens_seen": 26916064,
+      "step": 5020
+    },
+    {
+      "epoch": 0.804,
+      "grad_norm": 0.9311307072639465,
+      "learning_rate": 4.591518732070402e-06,
+      "loss": 0.7972,
+      "num_input_tokens_seen": 26940528,
+      "step": 5025
+    },
+    {
+      "epoch": 0.8048,
+      "grad_norm": 1.1058831214904785,
+      "learning_rate": 4.5552932625649944e-06,
+      "loss": 0.6977,
+      "num_input_tokens_seen": 26965296,
+      "step": 5030
+    },
+    {
+      "epoch": 0.8056,
+      "grad_norm": 1.2519973516464233,
+      "learning_rate": 4.5191969329388625e-06,
+      "loss": 0.8094,
+      "num_input_tokens_seen": 26988240,
+      "step": 5035
+    },
+    {
+      "epoch": 0.8064,
+      "grad_norm": 0.9225587248802185,
+      "learning_rate": 4.483229971196054e-06,
+      "loss": 0.7441,
+      "num_input_tokens_seen": 27015632,
+      "step": 5040
+    },
+    {
+      "epoch": 0.8072,
+      "grad_norm": 0.8636001944541931,
+      "learning_rate": 4.44739260452344e-06,
+      "loss": 0.6618,
+      "num_input_tokens_seen": 27040528,
+      "step": 5045
+    },
+    {
+      "epoch": 0.808,
+      "grad_norm": 1.094529151916504,
+      "learning_rate": 4.411685059289314e-06,
+      "loss": 0.7527,
+      "num_input_tokens_seen": 27066560,
+      "step": 5050
+    },
+    {
+      "epoch": 0.8088,
+      "grad_norm": 0.9794814586639404,
+      "learning_rate": 4.376107561041937e-06,
+      "loss": 0.7844,
+      "num_input_tokens_seen": 27089408,
+      "step": 5055
+    },
+    {
+      "epoch": 0.8096,
+      "grad_norm": 0.8608947396278381,
+      "learning_rate": 4.340660334508115e-06,
+      "loss": 0.7598,
+      "num_input_tokens_seen": 27114832,
+      "step": 5060
+    },
+    {
+      "epoch": 0.8104,
+      "grad_norm": 0.8481171727180481,
+      "learning_rate": 4.305343603591802e-06,
+      "loss": 0.6645,
+      "num_input_tokens_seen": 27140320,
+      "step": 5065
+    },
+    {
+      "epoch": 0.8112,
+      "grad_norm": 0.9115588068962097,
+      "learning_rate": 4.270157591372667e-06,
+      "loss": 0.7065,
+      "num_input_tokens_seen": 27171200,
+      "step": 5070
+    },
+    {
+      "epoch": 0.812,
+      "grad_norm": 1.4465726613998413,
+      "learning_rate": 4.235102520104681e-06,
+      "loss": 0.7481,
+      "num_input_tokens_seen": 27195056,
+      "step": 5075
+    },
+    {
+      "epoch": 0.8128,
+      "grad_norm": 1.1063685417175293,
+      "learning_rate": 4.200178611214736e-06,
+      "loss": 0.6154,
+      "num_input_tokens_seen": 27220816,
+      "step": 5080
+    },
+    {
+      "epoch": 0.8136,
+      "grad_norm": 1.3894023895263672,
+      "learning_rate": 4.165386085301212e-06,
+      "loss": 0.6661,
+      "num_input_tokens_seen": 27246400,
+      "step": 5085
+    },
+    {
+      "epoch": 0.8144,
+      "grad_norm": 1.0136696100234985,
+      "learning_rate": 4.130725162132612e-06,
+      "loss": 0.7043,
+      "num_input_tokens_seen": 27269936,
+      "step": 5090
+    },
+    {
+      "epoch": 0.8152,
+      "grad_norm": 1.36388099193573,
+      "learning_rate": 4.096196060646168e-06,
+      "loss": 0.8173,
+      "num_input_tokens_seen": 27293488,
+      "step": 5095
+    },
+    {
+      "epoch": 0.816,
+      "grad_norm": 0.9864152669906616,
+      "learning_rate": 4.061798998946459e-06,
+      "loss": 0.7154,
+      "num_input_tokens_seen": 27318592,
+      "step": 5100
+    },
+    {
+      "epoch": 0.8168,
+      "grad_norm": 0.9855313301086426,
+      "learning_rate": 4.027534194304005e-06,
+      "loss": 0.6336,
+      "num_input_tokens_seen": 27343616,
+      "step": 5105
+    },
+    {
+      "epoch": 0.8176,
+      "grad_norm": 0.9487363696098328,
+      "learning_rate": 3.99340186315395e-06,
+      "loss": 0.7355,
+      "num_input_tokens_seen": 27369216,
+      "step": 5110
+    },
+    {
+      "epoch": 0.8184,
+      "grad_norm": 0.8174062371253967,
+      "learning_rate": 3.959402221094635e-06,
+      "loss": 0.6034,
+      "num_input_tokens_seen": 27398704,
+      "step": 5115
+    },
+    {
+      "epoch": 0.8192,
+      "grad_norm": 1.17109215259552,
+      "learning_rate": 3.925535482886286e-06,
+      "loss": 0.7962,
+      "num_input_tokens_seen": 27424176,
+      "step": 5120
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.9322016835212708,
+      "learning_rate": 3.891801862449629e-06,
+      "loss": 0.7289,
+      "num_input_tokens_seen": 27452656,
+      "step": 5125
+    },
+    {
+      "epoch": 0.8208,
+      "grad_norm": 0.8374980688095093,
+      "learning_rate": 3.858201572864537e-06,
+      "loss": 0.6644,
+      "num_input_tokens_seen": 27478656,
+      "step": 5130
+    },
+    {
+      "epoch": 0.8216,
+      "grad_norm": 0.8750137686729431,
+      "learning_rate": 3.824734826368703e-06,
+      "loss": 0.7519,
+      "num_input_tokens_seen": 27507184,
+      "step": 5135
+    },
+    {
+      "epoch": 0.8224,
+      "grad_norm": 0.935739278793335,
+      "learning_rate": 3.7914018343562895e-06,
+      "loss": 0.7611,
+      "num_input_tokens_seen": 27536112,
+      "step": 5140
+    },
+    {
+      "epoch": 0.8232,
+      "grad_norm": 1.0115796327590942,
+      "learning_rate": 3.75820280737659e-06,
+      "loss": 0.6713,
+      "num_input_tokens_seen": 27563728,
+      "step": 5145
+    },
+    {
+      "epoch": 0.824,
+      "grad_norm": 1.1365923881530762,
+      "learning_rate": 3.725137955132707e-06,
+      "loss": 0.6522,
+      "num_input_tokens_seen": 27587120,
+      "step": 5150
+    },
+    {
+      "epoch": 0.8248,
+      "grad_norm": 1.2187998294830322,
+      "learning_rate": 3.692207486480209e-06,
+      "loss": 0.7707,
+      "num_input_tokens_seen": 27608240,
+      "step": 5155
+    },
+    {
+      "epoch": 0.8256,
+      "grad_norm": 1.2192776203155518,
+      "learning_rate": 3.6594116094258337e-06,
+      "loss": 0.6148,
+      "num_input_tokens_seen": 27637840,
+      "step": 5160
+    },
+    {
+      "epoch": 0.8264,
+      "grad_norm": 0.8527853488922119,
+      "learning_rate": 3.626750531126169e-06,
+      "loss": 0.6884,
+      "num_input_tokens_seen": 27662144,
+      "step": 5165
+    },
+    {
+      "epoch": 0.8272,
+      "grad_norm": 1.293915033340454,
+      "learning_rate": 3.594224457886336e-06,
+      "loss": 0.6954,
+      "num_input_tokens_seen": 27691184,
+      "step": 5170
+    },
+    {
+      "epoch": 0.828,
+      "grad_norm": 0.9655611515045166,
+      "learning_rate": 3.561833595158698e-06,
+      "loss": 0.6736,
+      "num_input_tokens_seen": 27719376,
+      "step": 5175
+    },
+    {
+      "epoch": 0.8288,
+      "grad_norm": 0.8014153838157654,
+      "learning_rate": 3.529578147541532e-06,
+      "loss": 0.7575,
+      "num_input_tokens_seen": 27749248,
+      "step": 5180
+    },
+    {
+      "epoch": 0.8296,
+      "grad_norm": 0.8693012595176697,
+      "learning_rate": 3.4974583187777852e-06,
+      "loss": 0.6607,
+      "num_input_tokens_seen": 27779136,
+      "step": 5185
+    },
+    {
+      "epoch": 0.8304,
+      "grad_norm": 1.2928762435913086,
+      "learning_rate": 3.4654743117537524e-06,
+      "loss": 0.844,
+      "num_input_tokens_seen": 27802592,
+      "step": 5190
+    },
+    {
+      "epoch": 0.8312,
+      "grad_norm": 0.9536007046699524,
+      "learning_rate": 3.433626328497805e-06,
+      "loss": 0.6596,
+      "num_input_tokens_seen": 27826864,
+      "step": 5195
+    },
+    {
+      "epoch": 0.832,
+      "grad_norm": 1.01909601688385,
+      "learning_rate": 3.4019145701791184e-06,
+      "loss": 0.7834,
+      "num_input_tokens_seen": 27851680,
+      "step": 5200
+    },
+    {
+      "epoch": 0.8328,
+      "grad_norm": 0.8944137096405029,
+      "learning_rate": 3.3703392371063845e-06,
+      "loss": 0.6874,
+      "num_input_tokens_seen": 27880208,
+      "step": 5205
+    },
+    {
+      "epoch": 0.8336,
+      "grad_norm": 1.2721880674362183,
+      "learning_rate": 3.338900528726571e-06,
+      "loss": 0.6468,
+      "num_input_tokens_seen": 27907392,
+      "step": 5210
+    },
+    {
+      "epoch": 0.8344,
+      "grad_norm": 1.3406318426132202,
+      "learning_rate": 3.3075986436236493e-06,
+      "loss": 0.6675,
+      "num_input_tokens_seen": 27934560,
+      "step": 5215
+    },
+    {
+      "epoch": 0.8352,
+      "grad_norm": 0.9489244818687439,
+      "learning_rate": 3.2764337795173435e-06,
+      "loss": 0.6704,
+      "num_input_tokens_seen": 27963248,
+      "step": 5220
+    },
+    {
+      "epoch": 0.836,
+      "grad_norm": 0.9695754647254944,
+      "learning_rate": 3.245406133261858e-06,
+      "loss": 0.6903,
+      "num_input_tokens_seen": 27989872,
+      "step": 5225
+    },
+    {
+      "epoch": 0.8368,
+      "grad_norm": 0.7802848219871521,
+      "learning_rate": 3.2145159008446807e-06,
+      "loss": 0.7441,
+      "num_input_tokens_seen": 28012208,
+      "step": 5230
+    },
+    {
+      "epoch": 0.8376,
+      "grad_norm": 1.2693451642990112,
+      "learning_rate": 3.1837632773853098e-06,
+      "loss": 0.6636,
+      "num_input_tokens_seen": 28041200,
+      "step": 5235
+    },
+    {
+      "epoch": 0.8384,
+      "grad_norm": 1.1437602043151855,
+      "learning_rate": 3.15314845713402e-06,
+      "loss": 0.7246,
+      "num_input_tokens_seen": 28068272,
+      "step": 5240
+    },
+    {
+      "epoch": 0.8392,
+      "grad_norm": 0.8950793743133545,
+      "learning_rate": 3.122671633470664e-06,
+      "loss": 0.6583,
+      "num_input_tokens_seen": 28092768,
+      "step": 5245
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 1.010606288909912,
+      "learning_rate": 3.0923329989034132e-06,
+      "loss": 0.6823,
+      "num_input_tokens_seen": 28122944,
+      "step": 5250
+    },
+    {
+      "epoch": 0.8408,
+      "grad_norm": 1.0807864665985107,
+      "learning_rate": 3.062132745067581e-06,
+      "loss": 0.7467,
+      "num_input_tokens_seen": 28151424,
+      "step": 5255
+    },
+    {
+      "epoch": 0.8416,
+      "grad_norm": 1.1845099925994873,
+      "learning_rate": 3.0320710627243813e-06,
+      "loss": 0.7541,
+      "num_input_tokens_seen": 28176752,
+      "step": 5260
+    },
+    {
+      "epoch": 0.8424,
+      "grad_norm": 0.7790080904960632,
+      "learning_rate": 3.002148141759739e-06,
+      "loss": 0.6829,
+      "num_input_tokens_seen": 28205456,
+      "step": 5265
+    },
+    {
+      "epoch": 0.8432,
+      "grad_norm": 1.364443302154541,
+      "learning_rate": 2.97236417118309e-06,
+      "loss": 0.692,
+      "num_input_tokens_seen": 28230304,
+      "step": 5270
+    },
+    {
+      "epoch": 0.844,
+      "grad_norm": 1.1289632320404053,
+      "learning_rate": 2.942719339126171e-06,
+      "loss": 0.7628,
+      "num_input_tokens_seen": 28255536,
+      "step": 5275
+    },
+    {
+      "epoch": 0.8448,
+      "grad_norm": 1.3984529972076416,
+      "learning_rate": 2.9132138328418573e-06,
+      "loss": 0.6972,
+      "num_input_tokens_seen": 28279600,
+      "step": 5280
+    },
+    {
+      "epoch": 0.8456,
+      "grad_norm": 1.0336651802062988,
+      "learning_rate": 2.8838478387029606e-06,
+      "loss": 0.6801,
+      "num_input_tokens_seen": 28304688,
+      "step": 5285
+    },
+    {
+      "epoch": 0.8464,
+      "grad_norm": 0.8985416293144226,
+      "learning_rate": 2.8546215422010638e-06,
+      "loss": 0.6697,
+      "num_input_tokens_seen": 28331584,
+      "step": 5290
+    },
+    {
+      "epoch": 0.8472,
+      "grad_norm": 1.3255959749221802,
+      "learning_rate": 2.8255351279453446e-06,
+      "loss": 0.6816,
+      "num_input_tokens_seen": 28360256,
+      "step": 5295
+    },
+    {
+      "epoch": 0.848,
+      "grad_norm": 0.9332570433616638,
+      "learning_rate": 2.7965887796613884e-06,
+      "loss": 0.7763,
+      "num_input_tokens_seen": 28385168,
+      "step": 5300
+    },
+    {
+      "epoch": 0.8488,
+      "grad_norm": 1.041506290435791,
+      "learning_rate": 2.767782680190073e-06,
+      "loss": 0.7837,
+      "num_input_tokens_seen": 28407248,
+      "step": 5305
+    },
+    {
+      "epoch": 0.8496,
+      "grad_norm": 1.075434923171997,
+      "learning_rate": 2.739117011486378e-06,
+      "loss": 0.6975,
+      "num_input_tokens_seen": 28434304,
+      "step": 5310
+    },
+    {
+      "epoch": 0.8504,
+      "grad_norm": 1.0231233835220337,
+      "learning_rate": 2.710591954618247e-06,
+      "loss": 0.7119,
+      "num_input_tokens_seen": 28465424,
+      "step": 5315
+    },
+    {
+      "epoch": 0.8512,
+      "grad_norm": 1.0655337572097778,
+      "learning_rate": 2.6822076897654452e-06,
+      "loss": 0.7644,
+      "num_input_tokens_seen": 28494416,
+      "step": 5320
+    },
+    {
+      "epoch": 0.852,
+      "grad_norm": 1.177964448928833,
+      "learning_rate": 2.6539643962184057e-06,
+      "loss": 0.644,
+      "num_input_tokens_seen": 28519552,
+      "step": 5325
+    },
+    {
+      "epoch": 0.8528,
+      "grad_norm": 1.0055785179138184,
+      "learning_rate": 2.6258622523771287e-06,
+      "loss": 0.7378,
+      "num_input_tokens_seen": 28545632,
+      "step": 5330
+    },
+    {
+      "epoch": 0.8536,
+      "grad_norm": 1.2302112579345703,
+      "learning_rate": 2.5979014357500248e-06,
+      "loss": 0.7267,
+      "num_input_tokens_seen": 28571440,
+      "step": 5335
+    },
+    {
+      "epoch": 0.8544,
+      "grad_norm": 1.2346506118774414,
+      "learning_rate": 2.570082122952816e-06,
+      "loss": 0.6015,
+      "num_input_tokens_seen": 28599472,
+      "step": 5340
+    },
+    {
+      "epoch": 0.8552,
+      "grad_norm": 0.8938889503479004,
+      "learning_rate": 2.5424044897073895e-06,
+      "loss": 0.6327,
+      "num_input_tokens_seen": 28629136,
+      "step": 5345
+    },
+    {
+      "epoch": 0.856,
+      "grad_norm": 0.8958183526992798,
+      "learning_rate": 2.514868710840723e-06,
+      "loss": 0.739,
+      "num_input_tokens_seen": 28648928,
+      "step": 5350
+    },
+    {
+      "epoch": 0.8568,
+      "grad_norm": 1.0484833717346191,
+      "learning_rate": 2.4874749602837697e-06,
+      "loss": 0.7279,
+      "num_input_tokens_seen": 28675056,
+      "step": 5355
+    },
+    {
+      "epoch": 0.8576,
+      "grad_norm": 0.887750506401062,
+      "learning_rate": 2.4602234110703364e-06,
+      "loss": 0.726,
+      "num_input_tokens_seen": 28698416,
+      "step": 5360
+    },
+    {
+      "epoch": 0.8584,
+      "grad_norm": 0.7482147216796875,
+      "learning_rate": 2.43311423533602e-06,
+      "loss": 0.7046,
+      "num_input_tokens_seen": 28729856,
+      "step": 5365
+    },
+    {
+      "epoch": 0.8592,
+      "grad_norm": 1.003188967704773,
+      "learning_rate": 2.406147604317119e-06,
+      "loss": 0.6922,
+      "num_input_tokens_seen": 28757360,
+      "step": 5370
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 1.5129293203353882,
+      "learning_rate": 2.379323688349516e-06,
+      "loss": 0.6664,
+      "num_input_tokens_seen": 28780624,
+      "step": 5375
+    },
+    {
+      "epoch": 0.8608,
+      "grad_norm": 0.9118067622184753,
+      "learning_rate": 2.3526426568676483e-06,
+      "loss": 0.6532,
+      "num_input_tokens_seen": 28805616,
+      "step": 5380
+    },
+    {
+      "epoch": 0.8616,
+      "grad_norm": 1.007717251777649,
+      "learning_rate": 2.326104678403415e-06,
+      "loss": 0.6678,
+      "num_input_tokens_seen": 28833504,
+      "step": 5385
+    },
+    {
+      "epoch": 0.8624,
+      "grad_norm": 0.9576060771942139,
+      "learning_rate": 2.299709920585108e-06,
+      "loss": 0.6152,
+      "num_input_tokens_seen": 28862704,
+      "step": 5390
+    },
+    {
+      "epoch": 0.8632,
+      "grad_norm": 1.1482651233673096,
+      "learning_rate": 2.2734585501363673e-06,
+      "loss": 0.7131,
+      "num_input_tokens_seen": 28886224,
+      "step": 5395
+    },
+    {
+      "epoch": 0.864,
+      "grad_norm": 1.0614038705825806,
+      "learning_rate": 2.2473507328751086e-06,
+      "loss": 0.735,
+      "num_input_tokens_seen": 28911760,
+      "step": 5400
+    },
+    {
+      "epoch": 0.8648,
+      "grad_norm": 1.0017966032028198,
+      "learning_rate": 2.2213866337125022e-06,
+      "loss": 0.6706,
+      "num_input_tokens_seen": 28941360,
+      "step": 5405
+    },
+    {
+      "epoch": 0.8656,
+      "grad_norm": 0.963431179523468,
+      "learning_rate": 2.1955664166519036e-06,
+      "loss": 0.7683,
+      "num_input_tokens_seen": 28965568,
+      "step": 5410
+    },
+    {
+      "epoch": 0.8664,
+      "grad_norm": 1.1716382503509521,
+      "learning_rate": 2.1698902447878477e-06,
+      "loss": 0.623,
+      "num_input_tokens_seen": 28994432,
+      "step": 5415
+    },
+    {
+      "epoch": 0.8672,
+      "grad_norm": 1.2640058994293213,
+      "learning_rate": 2.1443582803049755e-06,
+      "loss": 0.7774,
+      "num_input_tokens_seen": 29016560,
+      "step": 5420
+    },
+    {
+      "epoch": 0.868,
+      "grad_norm": 0.9828547239303589,
+      "learning_rate": 2.118970684477062e-06,
+      "loss": 0.6332,
+      "num_input_tokens_seen": 29043920,
+      "step": 5425
+    },
+    {
+      "epoch": 0.8688,
+      "grad_norm": 0.9180524349212646,
+      "learning_rate": 2.093727617665955e-06,
+      "loss": 0.6658,
+      "num_input_tokens_seen": 29073840,
+      "step": 5430
+    },
+    {
+      "epoch": 0.8696,
+      "grad_norm": 1.1137315034866333,
+      "learning_rate": 2.068629239320588e-06,
+      "loss": 0.7078,
+      "num_input_tokens_seen": 29102752,
+      "step": 5435
+    },
+    {
+      "epoch": 0.8704,
+      "grad_norm": 1.1765251159667969,
+      "learning_rate": 2.043675707975959e-06,
+      "loss": 0.7049,
+      "num_input_tokens_seen": 29126576,
+      "step": 5440
+    },
+    {
+      "epoch": 0.8712,
+      "grad_norm": 0.9444310665130615,
+      "learning_rate": 2.0188671812521292e-06,
+      "loss": 0.7931,
+      "num_input_tokens_seen": 29153120,
+      "step": 5445
+    },
+    {
+      "epoch": 0.872,
+      "grad_norm": 0.914959192276001,
+      "learning_rate": 1.9942038158532407e-06,
+      "loss": 0.8394,
+      "num_input_tokens_seen": 29182192,
+      "step": 5450
+    },
+    {
+      "epoch": 0.8728,
+      "grad_norm": 1.21523916721344,
+      "learning_rate": 1.969685767566512e-06,
+      "loss": 0.6915,
+      "num_input_tokens_seen": 29206368,
+      "step": 5455
+    },
+    {
+      "epoch": 0.8736,
+      "grad_norm": 0.8198549151420593,
+      "learning_rate": 1.9453131912612694e-06,
+      "loss": 0.6627,
+      "num_input_tokens_seen": 29235984,
+      "step": 5460
+    },
+    {
+      "epoch": 0.8744,
+      "grad_norm": 0.9284049868583679,
+      "learning_rate": 1.921086240887937e-06,
+      "loss": 0.6671,
+      "num_input_tokens_seen": 29260672,
+      "step": 5465
+    },
+    {
+      "epoch": 0.8752,
+      "grad_norm": 0.9254517555236816,
+      "learning_rate": 1.8970050694771064e-06,
+      "loss": 0.665,
+      "num_input_tokens_seen": 29287792,
+      "step": 5470
+    },
+    {
+      "epoch": 0.876,
+      "grad_norm": 0.8524356484413147,
+      "learning_rate": 1.8730698291385518e-06,
+      "loss": 0.663,
+      "num_input_tokens_seen": 29314656,
+      "step": 5475
+    },
+    {
+      "epoch": 0.8768,
+      "grad_norm": 0.8038460612297058,
+      "learning_rate": 1.8492806710602496e-06,
+      "loss": 0.7054,
+      "num_input_tokens_seen": 29338976,
+      "step": 5480
+    },
+    {
+      "epoch": 0.8776,
+      "grad_norm": 0.8914519548416138,
+      "learning_rate": 1.8256377455074525e-06,
+      "loss": 0.6905,
+      "num_input_tokens_seen": 29364912,
+      "step": 5485
+    },
+    {
+      "epoch": 0.8784,
+      "grad_norm": 0.8928874135017395,
+      "learning_rate": 1.802141201821736e-06,
+      "loss": 0.7641,
+      "num_input_tokens_seen": 29392960,
+      "step": 5490
+    },
+    {
+      "epoch": 0.8792,
+      "grad_norm": 0.7842042446136475,
+      "learning_rate": 1.7787911884200314e-06,
+      "loss": 0.6918,
+      "num_input_tokens_seen": 29416848,
+      "step": 5495
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 1.335999846458435,
+      "learning_rate": 1.7555878527937164e-06,
+      "loss": 0.6281,
+      "num_input_tokens_seen": 29445856,
+      "step": 5500
+    },
+    {
+      "epoch": 0.8808,
+      "grad_norm": 0.981039822101593,
+      "learning_rate": 1.7325313415076705e-06,
+      "loss": 0.7199,
+      "num_input_tokens_seen": 29474400,
+      "step": 5505
+    },
+    {
+      "epoch": 0.8816,
+      "grad_norm": 1.003205418586731,
+      "learning_rate": 1.7096218001993513e-06,
+      "loss": 0.7022,
+      "num_input_tokens_seen": 29501312,
+      "step": 5510
+    },
+    {
+      "epoch": 0.8824,
+      "grad_norm": 0.804409921169281,
+      "learning_rate": 1.686859373577876e-06,
+      "loss": 0.635,
+      "num_input_tokens_seen": 29530160,
+      "step": 5515
+    },
+    {
+      "epoch": 0.8832,
+      "grad_norm": 0.6613283157348633,
+      "learning_rate": 1.6642442054230934e-06,
+      "loss": 0.6752,
+      "num_input_tokens_seen": 29557168,
+      "step": 5520
+    },
+    {
+      "epoch": 0.884,
+      "grad_norm": 1.0624265670776367,
+      "learning_rate": 1.6417764385846996e-06,
+      "loss": 0.7824,
+      "num_input_tokens_seen": 29584832,
+      "step": 5525
+    },
+    {
+      "epoch": 0.8848,
+      "grad_norm": 1.0792144536972046,
+      "learning_rate": 1.6194562149813242e-06,
+      "loss": 0.6823,
+      "num_input_tokens_seen": 29609504,
+      "step": 5530
+    },
+    {
+      "epoch": 0.8856,
+      "grad_norm": 1.0641052722930908,
+      "learning_rate": 1.5972836755996285e-06,
+      "loss": 0.6777,
+      "num_input_tokens_seen": 29636768,
+      "step": 5535
+    },
+    {
+      "epoch": 0.8864,
+      "grad_norm": 1.2235264778137207,
+      "learning_rate": 1.5752589604934255e-06,
+      "loss": 0.7372,
+      "num_input_tokens_seen": 29660496,
+      "step": 5540
+    },
+    {
+      "epoch": 0.8872,
+      "grad_norm": 0.9904717206954956,
+      "learning_rate": 1.5533822087827805e-06,
+      "loss": 0.7126,
+      "num_input_tokens_seen": 29686928,
+      "step": 5545
+    },
+    {
+      "epoch": 0.888,
+      "grad_norm": 1.2209672927856445,
+      "learning_rate": 1.5316535586531483e-06,
+      "loss": 0.6564,
+      "num_input_tokens_seen": 29714800,
+      "step": 5550
+    },
+    {
+      "epoch": 0.8888,
+      "grad_norm": 1.2465245723724365,
+      "learning_rate": 1.5100731473544933e-06,
+      "loss": 0.8006,
+      "num_input_tokens_seen": 29741808,
+      "step": 5555
+    },
+    {
+      "epoch": 0.8896,
+      "grad_norm": 1.0243083238601685,
+      "learning_rate": 1.4886411112004255e-06,
+      "loss": 0.7322,
+      "num_input_tokens_seen": 29763088,
+      "step": 5560
+    },
+    {
+      "epoch": 0.8904,
+      "grad_norm": 0.9879063367843628,
+      "learning_rate": 1.4673575855673277e-06,
+      "loss": 0.7243,
+      "num_input_tokens_seen": 29791520,
+      "step": 5565
+    },
+    {
+      "epoch": 0.8912,
+      "grad_norm": 0.8593109250068665,
+      "learning_rate": 1.4462227048935183e-06,
+      "loss": 0.6955,
+      "num_input_tokens_seen": 29817600,
+      "step": 5570
+    },
+    {
+      "epoch": 0.892,
+      "grad_norm": 0.9585959911346436,
+      "learning_rate": 1.425236602678387e-06,
+      "loss": 0.6658,
+      "num_input_tokens_seen": 29843136,
+      "step": 5575
+    },
+    {
+      "epoch": 0.8928,
+      "grad_norm": 1.2681745290756226,
+      "learning_rate": 1.4043994114815661e-06,
+      "loss": 0.7943,
+      "num_input_tokens_seen": 29864864,
+      "step": 5580
+    },
+    {
+      "epoch": 0.8936,
+      "grad_norm": 0.8207817673683167,
+      "learning_rate": 1.38371126292208e-06,
+      "loss": 0.7734,
+      "num_input_tokens_seen": 29890416,
+      "step": 5585
+    },
+    {
+      "epoch": 0.8944,
+      "grad_norm": 1.0749483108520508,
+      "learning_rate": 1.3631722876775138e-06,
+      "loss": 0.7008,
+      "num_input_tokens_seen": 29916400,
+      "step": 5590
+    },
+    {
+      "epoch": 0.8952,
+      "grad_norm": 1.1091291904449463,
+      "learning_rate": 1.3427826154832042e-06,
+      "loss": 0.6434,
+      "num_input_tokens_seen": 29944304,
+      "step": 5595
+    },
+    {
+      "epoch": 0.896,
+      "grad_norm": 1.3164221048355103,
+      "learning_rate": 1.3225423751313942e-06,
+      "loss": 0.6952,
+      "num_input_tokens_seen": 29967648,
+      "step": 5600
+    },
+    {
+      "epoch": 0.8968,
+      "grad_norm": 1.2402032613754272,
+      "learning_rate": 1.3024516944704496e-06,
+      "loss": 0.6331,
+      "num_input_tokens_seen": 29989312,
+      "step": 5605
+    },
+    {
+      "epoch": 0.8976,
+      "grad_norm": 1.1296372413635254,
+      "learning_rate": 1.2825107004040272e-06,
+      "loss": 0.7894,
+      "num_input_tokens_seen": 30012384,
+      "step": 5610
+    },
+    {
+      "epoch": 0.8984,
+      "grad_norm": 1.109983205795288,
+      "learning_rate": 1.2627195188902791e-06,
+      "loss": 0.6819,
+      "num_input_tokens_seen": 30042656,
+      "step": 5615
+    },
+    {
+      "epoch": 0.8992,
+      "grad_norm": 1.0760643482208252,
+      "learning_rate": 1.2430782749410673e-06,
+      "loss": 0.8208,
+      "num_input_tokens_seen": 30068464,
+      "step": 5620
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 1.1865812540054321,
+      "learning_rate": 1.2235870926211619e-06,
+      "loss": 0.7203,
+      "num_input_tokens_seen": 30093216,
+      "step": 5625
+    },
+    {
+      "epoch": 0.9008,
+      "grad_norm": 1.0090526342391968,
+      "learning_rate": 1.2042460950474648e-06,
+      "loss": 0.7368,
+      "num_input_tokens_seen": 30117488,
+      "step": 5630
+    },
+    {
+      "epoch": 0.9016,
+      "grad_norm": 0.8406433463096619,
+      "learning_rate": 1.1850554043882328e-06,
+      "loss": 0.7681,
+      "num_input_tokens_seen": 30144016,
+      "step": 5635
+    },
+    {
+      "epoch": 0.9024,
+      "grad_norm": 1.057853102684021,
+      "learning_rate": 1.1660151418622922e-06,
+      "loss": 0.6962,
+      "num_input_tokens_seen": 30177184,
+      "step": 5640
+    },
+    {
+      "epoch": 0.9032,
+      "grad_norm": 0.9097471237182617,
+      "learning_rate": 1.1471254277382881e-06,
+      "loss": 0.7056,
+      "num_input_tokens_seen": 30206048,
+      "step": 5645
+    },
+    {
+      "epoch": 0.904,
+      "grad_norm": 1.0608327388763428,
+      "learning_rate": 1.1283863813339263e-06,
+      "loss": 0.7089,
+      "num_input_tokens_seen": 30229936,
+      "step": 5650
+    },
+    {
+      "epoch": 0.9048,
+      "grad_norm": 1.1075376272201538,
+      "learning_rate": 1.1097981210152043e-06,
+      "loss": 0.7794,
+      "num_input_tokens_seen": 30257760,
+      "step": 5655
+    },
+    {
+      "epoch": 0.9056,
+      "grad_norm": 0.9509792327880859,
+      "learning_rate": 1.0913607641956841e-06,
+      "loss": 0.77,
+      "num_input_tokens_seen": 30286464,
+      "step": 5660
+    },
+    {
+      "epoch": 0.9064,
+      "grad_norm": 0.981621265411377,
+      "learning_rate": 1.0730744273357213e-06,
+      "loss": 0.708,
+      "num_input_tokens_seen": 30317040,
+      "step": 5665
+    },
+    {
+      "epoch": 0.9072,
+      "grad_norm": 0.8428750038146973,
+      "learning_rate": 1.0549392259417646e-06,
+      "loss": 0.6423,
+      "num_input_tokens_seen": 30342672,
+      "step": 5670
+    },
+    {
+      "epoch": 0.908,
+      "grad_norm": 0.9921499490737915,
+      "learning_rate": 1.0369552745656013e-06,
+      "loss": 0.8176,
+      "num_input_tokens_seen": 30369952,
+      "step": 5675
+    },
+    {
+      "epoch": 0.9088,
+      "grad_norm": 0.9219129681587219,
+      "learning_rate": 1.0191226868036418e-06,
+      "loss": 0.6924,
+      "num_input_tokens_seen": 30400800,
+      "step": 5680
+    },
+    {
+      "epoch": 0.9096,
+      "grad_norm": 1.451660394668579,
+      "learning_rate": 1.001441575296208e-06,
+      "loss": 0.576,
+      "num_input_tokens_seen": 30436240,
+      "step": 5685
+    },
+    {
+      "epoch": 0.9104,
+      "grad_norm": 0.9631555676460266,
+      "learning_rate": 9.839120517267985e-07,
+      "loss": 0.6206,
+      "num_input_tokens_seen": 30465232,
+      "step": 5690
+    },
+    {
+      "epoch": 0.9112,
+      "grad_norm": 1.125351071357727,
+      "learning_rate": 9.665342268214166e-07,
+      "loss": 0.7424,
+      "num_input_tokens_seen": 30489776,
+      "step": 5695
+    },
+    {
+      "epoch": 0.912,
+      "grad_norm": 1.0404316186904907,
+      "learning_rate": 9.493082103478517e-07,
+      "loss": 0.6868,
+      "num_input_tokens_seen": 30514592,
+      "step": 5700
+    },
+    {
+      "epoch": 0.9128,
+      "grad_norm": 0.9020105004310608,
+      "learning_rate": 9.322341111149852e-07,
+      "loss": 0.7017,
+      "num_input_tokens_seen": 30544112,
+      "step": 5705
+    },
+    {
+      "epoch": 0.9136,
+      "grad_norm": 1.0047924518585205,
+      "learning_rate": 9.153120369721046e-07,
+      "loss": 0.6429,
+      "num_input_tokens_seen": 30577440,
+      "step": 5710
+    },
+    {
+      "epoch": 0.9144,
+      "grad_norm": 1.087547779083252,
+      "learning_rate": 8.985420948082329e-07,
+      "loss": 0.6507,
+      "num_input_tokens_seen": 30602704,
+      "step": 5715
+    },
+    {
+      "epoch": 0.9152,
+      "grad_norm": 0.9647343754768372,
+      "learning_rate": 8.819243905514308e-07,
+      "loss": 0.7508,
+      "num_input_tokens_seen": 30627360,
+      "step": 5720
+    },
+    {
+      "epoch": 0.916,
+      "grad_norm": 0.9557966589927673,
+      "learning_rate": 8.65459029168153e-07,
+      "loss": 0.6506,
+      "num_input_tokens_seen": 30657216,
+      "step": 5725
+    },
+    {
+      "epoch": 0.9168,
+      "grad_norm": 0.991233229637146,
+      "learning_rate": 8.491461146625773e-07,
+      "loss": 0.6426,
+      "num_input_tokens_seen": 30683472,
+      "step": 5730
+    },
+    {
+      "epoch": 0.9176,
+      "grad_norm": 0.8405401706695557,
+      "learning_rate": 8.329857500759292e-07,
+      "loss": 0.6227,
+      "num_input_tokens_seen": 30707392,
+      "step": 5735
+    },
+    {
+      "epoch": 0.9184,
+      "grad_norm": 1.2772481441497803,
+      "learning_rate": 8.169780374858577e-07,
+      "loss": 0.839,
+      "num_input_tokens_seen": 30732160,
+      "step": 5740
+    },
+    {
+      "epoch": 0.9192,
+      "grad_norm": 0.8571876883506775,
+      "learning_rate": 8.011230780057749e-07,
+      "loss": 0.6817,
+      "num_input_tokens_seen": 30760336,
+      "step": 5745
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 1.000186562538147,
+      "learning_rate": 7.854209717842231e-07,
+      "loss": 0.7016,
+      "num_input_tokens_seen": 30788800,
+      "step": 5750
+    },
+    {
+      "epoch": 0.9208,
+      "grad_norm": 1.0842921733856201,
+      "learning_rate": 7.698718180042392e-07,
+      "loss": 0.702,
+      "num_input_tokens_seen": 30813632,
+      "step": 5755
+    },
+    {
+      "epoch": 0.9216,
+      "grad_norm": 1.0399980545043945,
+      "learning_rate": 7.544757148827297e-07,
+      "loss": 0.7203,
+      "num_input_tokens_seen": 30840816,
+      "step": 5760
+    },
+    {
+      "epoch": 0.9224,
+      "grad_norm": 1.1896955966949463,
+      "learning_rate": 7.392327596698473e-07,
+      "loss": 0.6873,
+      "num_input_tokens_seen": 30861664,
+      "step": 5765
+    },
+    {
+      "epoch": 0.9232,
+      "grad_norm": 1.3494455814361572,
+      "learning_rate": 7.241430486483819e-07,
+      "loss": 0.6975,
+      "num_input_tokens_seen": 30886064,
+      "step": 5770
+    },
+    {
+      "epoch": 0.924,
+      "grad_norm": 1.1318798065185547,
+      "learning_rate": 7.092066771331507e-07,
+      "loss": 0.6058,
+      "num_input_tokens_seen": 30910608,
+      "step": 5775
+    },
+    {
+      "epoch": 0.9248,
+      "grad_norm": 1.230055332183838,
+      "learning_rate": 6.944237394703984e-07,
+      "loss": 0.8128,
+      "num_input_tokens_seen": 30935008,
+      "step": 5780
+    },
+    {
+      "epoch": 0.9256,
+      "grad_norm": 1.0150400400161743,
+      "learning_rate": 6.797943290371839e-07,
+      "loss": 0.7329,
+      "num_input_tokens_seen": 30959792,
+      "step": 5785
+    },
+    {
+      "epoch": 0.9264,
+      "grad_norm": 0.9498345255851746,
+      "learning_rate": 6.653185382408194e-07,
+      "loss": 0.673,
+      "num_input_tokens_seen": 30985856,
+      "step": 5790
+    },
+    {
+      "epoch": 0.9272,
+      "grad_norm": 1.047587275505066,
+      "learning_rate": 6.509964585182687e-07,
+      "loss": 0.7395,
+      "num_input_tokens_seen": 31013888,
+      "step": 5795
+    },
+    {
+      "epoch": 0.928,
+      "grad_norm": 0.9112536907196045,
+      "learning_rate": 6.368281803355691e-07,
+      "loss": 0.753,
+      "num_input_tokens_seen": 31038352,
+      "step": 5800
+    },
+    {
+      "epoch": 0.9288,
+      "grad_norm": 0.9714504480361938,
+      "learning_rate": 6.228137931872713e-07,
+      "loss": 0.7573,
+      "num_input_tokens_seen": 31066624,
+      "step": 5805
+    },
+    {
+      "epoch": 0.9296,
+      "grad_norm": 1.1300855875015259,
+      "learning_rate": 6.089533855958507e-07,
+      "loss": 0.759,
+      "num_input_tokens_seen": 31093184,
+      "step": 5810
+    },
+    {
+      "epoch": 0.9304,
+      "grad_norm": 1.0004905462265015,
+      "learning_rate": 5.95247045111183e-07,
+      "loss": 0.7482,
+      "num_input_tokens_seen": 31118352,
+      "step": 5815
+    },
+    {
+      "epoch": 0.9312,
+      "grad_norm": 0.8432052731513977,
+      "learning_rate": 5.816948583099613e-07,
+      "loss": 0.6295,
+      "num_input_tokens_seen": 31145616,
+      "step": 5820
+    },
+    {
+      "epoch": 0.932,
+      "grad_norm": 0.923195481300354,
+      "learning_rate": 5.68296910795163e-07,
+      "loss": 0.7596,
+      "num_input_tokens_seen": 31167088,
+      "step": 5825
+    },
+    {
+      "epoch": 0.9328,
+      "grad_norm": 1.0346819162368774,
+      "learning_rate": 5.550532871955061e-07,
+      "loss": 0.689,
+      "num_input_tokens_seen": 31192672,
+      "step": 5830
+    },
+    {
+      "epoch": 0.9336,
+      "grad_norm": 0.8770239949226379,
+      "learning_rate": 5.419640711649188e-07,
+      "loss": 0.6387,
+      "num_input_tokens_seen": 31224016,
+      "step": 5835
+    },
+    {
+      "epoch": 0.9344,
+      "grad_norm": 1.2829992771148682,
+      "learning_rate": 5.290293453819955e-07,
+      "loss": 0.7316,
+      "num_input_tokens_seen": 31247008,
+      "step": 5840
+    },
+    {
+      "epoch": 0.9352,
+      "grad_norm": 0.9933931231498718,
+      "learning_rate": 5.162491915495005e-07,
+      "loss": 0.7255,
+      "num_input_tokens_seen": 31273232,
+      "step": 5845
+    },
+    {
+      "epoch": 0.936,
+      "grad_norm": 0.8756529092788696,
+      "learning_rate": 5.036236903938285e-07,
+      "loss": 0.7188,
+      "num_input_tokens_seen": 31299504,
+      "step": 5850
+    },
+    {
+      "epoch": 0.9368,
+      "grad_norm": 0.85035240650177,
+      "learning_rate": 4.911529216645088e-07,
+      "loss": 0.6763,
+      "num_input_tokens_seen": 31325792,
+      "step": 5855
+    },
+    {
+      "epoch": 0.9376,
+      "grad_norm": 0.9068401455879211,
+      "learning_rate": 4.788369641336943e-07,
+      "loss": 0.6109,
+      "num_input_tokens_seen": 31351216,
+      "step": 5860
+    },
+    {
+      "epoch": 0.9384,
+      "grad_norm": 1.079689860343933,
+      "learning_rate": 4.666758955956613e-07,
+      "loss": 0.7778,
+      "num_input_tokens_seen": 31376464,
+      "step": 5865
+    },
+    {
+      "epoch": 0.9392,
+      "grad_norm": 0.964074969291687,
+      "learning_rate": 4.546697928663357e-07,
+      "loss": 0.6315,
+      "num_input_tokens_seen": 31408832,
+      "step": 5870
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 1.1026054620742798,
+      "learning_rate": 4.4281873178278475e-07,
+      "loss": 0.7918,
+      "num_input_tokens_seen": 31432352,
+      "step": 5875
+    },
+    {
+      "epoch": 0.9408,
+      "grad_norm": 0.914069652557373,
+      "learning_rate": 4.311227872027479e-07,
+      "loss": 0.6983,
+      "num_input_tokens_seen": 31457392,
+      "step": 5880
+    },
+    {
+      "epoch": 0.9416,
+      "grad_norm": 0.989115834236145,
+      "learning_rate": 4.1958203300417054e-07,
+      "loss": 0.7233,
+      "num_input_tokens_seen": 31482704,
+      "step": 5885
+    },
+    {
+      "epoch": 0.9424,
+      "grad_norm": 1.034597396850586,
+      "learning_rate": 4.0819654208472947e-07,
+      "loss": 0.6402,
+      "num_input_tokens_seen": 31512368,
+      "step": 5890
+    },
+    {
+      "epoch": 0.9432,
+      "grad_norm": 1.3309321403503418,
+      "learning_rate": 3.9696638636137206e-07,
+      "loss": 0.6942,
+      "num_input_tokens_seen": 31539040,
+      "step": 5895
+    },
+    {
+      "epoch": 0.944,
+      "grad_norm": 1.2237857580184937,
+      "learning_rate": 3.8589163676986674e-07,
+      "loss": 0.7119,
+      "num_input_tokens_seen": 31563712,
+      "step": 5900
+    },
+    {
+      "epoch": 0.9448,
+      "grad_norm": 1.0268304347991943,
+      "learning_rate": 3.7497236326434757e-07,
+      "loss": 0.6575,
+      "num_input_tokens_seen": 31587760,
+      "step": 5905
+    },
+    {
+      "epoch": 0.9456,
+      "grad_norm": 1.0060738325119019,
+      "learning_rate": 3.6420863481688437e-07,
+      "loss": 0.704,
+      "num_input_tokens_seen": 31612976,
+      "step": 5910
+    },
+    {
+      "epoch": 0.9464,
+      "grad_norm": 0.9219969511032104,
+      "learning_rate": 3.536005194170328e-07,
+      "loss": 0.7876,
+      "num_input_tokens_seen": 31639472,
+      "step": 5915
+    },
+    {
+      "epoch": 0.9472,
+      "grad_norm": 0.8863883018493652,
+      "learning_rate": 3.431480840714152e-07,
+      "loss": 0.7033,
+      "num_input_tokens_seen": 31670768,
+      "step": 5920
+    },
+    {
+      "epoch": 0.948,
+      "grad_norm": 0.9394556879997253,
+      "learning_rate": 3.328513948032991e-07,
+      "loss": 0.7095,
+      "num_input_tokens_seen": 31696624,
+      "step": 5925
+    },
+    {
+      "epoch": 0.9488,
+      "grad_norm": 0.8008967638015747,
+      "learning_rate": 3.227105166521638e-07,
+      "loss": 0.6629,
+      "num_input_tokens_seen": 31723840,
+      "step": 5930
+    },
+    {
+      "epoch": 0.9496,
+      "grad_norm": 0.9910029172897339,
+      "learning_rate": 3.127255136733093e-07,
+      "loss": 0.591,
+      "num_input_tokens_seen": 31752736,
+      "step": 5935
+    },
+    {
+      "epoch": 0.9504,
+      "grad_norm": 0.9355325698852539,
+      "learning_rate": 3.0289644893744527e-07,
+      "loss": 0.6641,
+      "num_input_tokens_seen": 31777760,
+      "step": 5940
+    },
+    {
+      "epoch": 0.9512,
+      "grad_norm": 0.9911002516746521,
+      "learning_rate": 2.9322338453028066e-07,
+      "loss": 0.6156,
+      "num_input_tokens_seen": 31805264,
+      "step": 5945
+    },
+    {
+      "epoch": 0.952,
+      "grad_norm": 1.4127229452133179,
+      "learning_rate": 2.8370638155215123e-07,
+      "loss": 0.7834,
+      "num_input_tokens_seen": 31828656,
+      "step": 5950
+    },
+    {
+      "epoch": 0.9528,
+      "grad_norm": 1.0222047567367554,
+      "learning_rate": 2.743455001176176e-07,
+      "loss": 0.6998,
+      "num_input_tokens_seen": 31855424,
+      "step": 5955
+    },
+    {
+      "epoch": 0.9536,
+      "grad_norm": 0.9893736839294434,
+      "learning_rate": 2.6514079935509584e-07,
+      "loss": 0.7458,
+      "num_input_tokens_seen": 31879168,
+      "step": 5960
+    },
+    {
+      "epoch": 0.9544,
+      "grad_norm": 0.8317204713821411,
+      "learning_rate": 2.560923374064772e-07,
+      "loss": 0.7061,
+      "num_input_tokens_seen": 31903824,
+      "step": 5965
+    },
+    {
+      "epoch": 0.9552,
+      "grad_norm": 1.3220785856246948,
+      "learning_rate": 2.472001714267674e-07,
+      "loss": 0.8603,
+      "num_input_tokens_seen": 31927184,
+      "step": 5970
+    },
+    {
+      "epoch": 0.956,
+      "grad_norm": 0.8110103607177734,
+      "learning_rate": 2.384643575837203e-07,
+      "loss": 0.6273,
+      "num_input_tokens_seen": 31955104,
+      "step": 5975
+    },
+    {
+      "epoch": 0.9568,
+      "grad_norm": 0.6332679390907288,
+      "learning_rate": 2.298849510574824e-07,
+      "loss": 0.714,
+      "num_input_tokens_seen": 31985888,
+      "step": 5980
+    },
+    {
+      "epoch": 0.9576,
+      "grad_norm": 0.9290034174919128,
+      "learning_rate": 2.2146200604024613e-07,
+      "loss": 0.6899,
+      "num_input_tokens_seen": 32013520,
+      "step": 5985
+    },
+    {
+      "epoch": 0.9584,
+      "grad_norm": 1.0509424209594727,
+      "learning_rate": 2.1319557573591108e-07,
+      "loss": 0.677,
+      "num_input_tokens_seen": 32038880,
+      "step": 5990
+    },
+    {
+      "epoch": 0.9592,
+      "grad_norm": 1.0169018507003784,
+      "learning_rate": 2.050857123597455e-07,
+      "loss": 0.7033,
+      "num_input_tokens_seen": 32062160,
+      "step": 5995
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 1.053408145904541,
+      "learning_rate": 1.9713246713805588e-07,
+      "loss": 0.6431,
+      "num_input_tokens_seen": 32085712,
+      "step": 6000
+    },
+    {
+      "epoch": 0.9608,
+      "grad_norm": 1.1077343225479126,
+      "learning_rate": 1.8933589030785682e-07,
+      "loss": 0.683,
+      "num_input_tokens_seen": 32115232,
+      "step": 6005
+    },
+    {
+      "epoch": 0.9616,
+      "grad_norm": 1.2601428031921387,
+      "learning_rate": 1.8169603111656552e-07,
+      "loss": 0.751,
+      "num_input_tokens_seen": 32142992,
+      "step": 6010
+    },
+    {
+      "epoch": 0.9624,
+      "grad_norm": 0.7372344136238098,
+      "learning_rate": 1.7421293782168835e-07,
+      "loss": 0.5808,
+      "num_input_tokens_seen": 32176176,
+      "step": 6015
+    },
+    {
+      "epoch": 0.9632,
+      "grad_norm": 0.856760561466217,
+      "learning_rate": 1.6688665769050703e-07,
+      "loss": 0.6852,
+      "num_input_tokens_seen": 32204992,
+      "step": 6020
+    },
+    {
+      "epoch": 0.964,
+      "grad_norm": 1.110574722290039,
+      "learning_rate": 1.5971723699979013e-07,
+      "loss": 0.6778,
+      "num_input_tokens_seen": 32232960,
+      "step": 6025
+    },
+    {
+      "epoch": 0.9648,
+      "grad_norm": 1.084190845489502,
+      "learning_rate": 1.5270472103549315e-07,
+      "loss": 0.7036,
+      "num_input_tokens_seen": 32262672,
+      "step": 6030
+    },
+    {
+      "epoch": 0.9656,
+      "grad_norm": 0.9454313516616821,
+      "learning_rate": 1.4584915409248112e-07,
+      "loss": 0.655,
+      "num_input_tokens_seen": 32285504,
+      "step": 6035
+    },
+    {
+      "epoch": 0.9664,
+      "grad_norm": 0.9206419587135315,
+      "learning_rate": 1.3915057947423705e-07,
+      "loss": 0.7324,
+      "num_input_tokens_seen": 32312288,
+      "step": 6040
+    },
+    {
+      "epoch": 0.9672,
+      "grad_norm": 0.9567137956619263,
+      "learning_rate": 1.3260903949260107e-07,
+      "loss": 0.7166,
+      "num_input_tokens_seen": 32339424,
+      "step": 6045
+    },
+    {
+      "epoch": 0.968,
+      "grad_norm": 1.0180697441101074,
+      "learning_rate": 1.2622457546749567e-07,
+      "loss": 0.7,
+      "num_input_tokens_seen": 32362848,
+      "step": 6050
+    },
+    {
+      "epoch": 0.9688,
+      "grad_norm": 1.2073848247528076,
+      "learning_rate": 1.1999722772666476e-07,
+      "loss": 0.7519,
+      "num_input_tokens_seen": 32393264,
+      "step": 6055
+    },
+    {
+      "epoch": 0.9696,
+      "grad_norm": 1.020180583000183,
+      "learning_rate": 1.1392703560542117e-07,
+      "loss": 0.7524,
+      "num_input_tokens_seen": 32418464,
+      "step": 6060
+    },
+    {
+      "epoch": 0.9704,
+      "grad_norm": 1.0939137935638428,
+      "learning_rate": 1.080140374463967e-07,
+      "loss": 0.5829,
+      "num_input_tokens_seen": 32449248,
+      "step": 6065
+    },
+    {
+      "epoch": 0.9712,
+      "grad_norm": 1.2307384014129639,
+      "learning_rate": 1.0225827059930083e-07,
+      "loss": 0.7017,
+      "num_input_tokens_seen": 32477312,
+      "step": 6070
+    },
+    {
+      "epoch": 0.972,
+      "grad_norm": 1.0141756534576416,
+      "learning_rate": 9.665977142068738e-08,
+      "loss": 0.6852,
+      "num_input_tokens_seen": 32505024,
+      "step": 6075
+    },
+    {
+      "epoch": 0.9728,
+      "grad_norm": 1.0366077423095703,
+      "learning_rate": 9.121857527372158e-08,
+      "loss": 0.733,
+      "num_input_tokens_seen": 32530080,
+      "step": 6080
+    },
+    {
+      "epoch": 0.9736,
+      "grad_norm": 0.9831274151802063,
+      "learning_rate": 8.593471652794949e-08,
+      "loss": 0.6721,
+      "num_input_tokens_seen": 32557488,
+      "step": 6085
+    },
+    {
+      "epoch": 0.9744,
+      "grad_norm": 1.0170478820800781,
+      "learning_rate": 8.080822855909831e-08,
+      "loss": 0.6572,
+      "num_input_tokens_seen": 32589072,
+      "step": 6090
+    },
+    {
+      "epoch": 0.9752,
+      "grad_norm": 1.0840100049972534,
+      "learning_rate": 7.583914374885426e-08,
+      "loss": 0.7535,
+      "num_input_tokens_seen": 32613296,
+      "step": 6095
+    },
+    {
+      "epoch": 0.976,
+      "grad_norm": 1.1899126768112183,
+      "learning_rate": 7.102749348465165e-08,
+      "loss": 0.639,
+      "num_input_tokens_seen": 32642512,
+      "step": 6100
+    },
+    {
+      "epoch": 0.9768,
+      "grad_norm": 1.0756986141204834,
+      "learning_rate": 6.637330815949527e-08,
+      "loss": 0.7558,
+      "num_input_tokens_seen": 32666064,
+      "step": 6105
+    },
+    {
+      "epoch": 0.9776,
+      "grad_norm": 0.9403240084648132,
+      "learning_rate": 6.187661717174386e-08,
+      "loss": 0.7228,
+      "num_input_tokens_seen": 32690016,
+      "step": 6110
+    },
+    {
+      "epoch": 0.9784,
+      "grad_norm": 0.9194949269294739,
+      "learning_rate": 5.753744892494639e-08,
+      "loss": 0.7079,
+      "num_input_tokens_seen": 32716240,
+      "step": 6115
+    },
+    {
+      "epoch": 0.9792,
+      "grad_norm": 0.9947624206542969,
+      "learning_rate": 5.335583082764495e-08,
+      "loss": 0.7692,
+      "num_input_tokens_seen": 32741648,
+      "step": 6120
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 1.2828369140625,
+      "learning_rate": 4.9331789293211026e-08,
+      "loss": 0.6285,
+      "num_input_tokens_seen": 32770224,
+      "step": 6125
+    },
+    {
+      "epoch": 0.9808,
+      "grad_norm": 1.0066205263137817,
+      "learning_rate": 4.546534973968175e-08,
+      "loss": 0.7464,
+      "num_input_tokens_seen": 32798864,
+      "step": 6130
+    },
+    {
+      "epoch": 0.9816,
+      "grad_norm": 1.3146965503692627,
+      "learning_rate": 4.1756536589585004e-08,
+      "loss": 0.6632,
+      "num_input_tokens_seen": 32829136,
+      "step": 6135
+    },
+    {
+      "epoch": 0.9824,
+      "grad_norm": 0.8514100909233093,
+      "learning_rate": 3.820537326980622e-08,
+      "loss": 0.7378,
+      "num_input_tokens_seen": 32858976,
+      "step": 6140
+    },
+    {
+      "epoch": 0.9832,
+      "grad_norm": 0.9065835475921631,
+      "learning_rate": 3.481188221142184e-08,
+      "loss": 0.7125,
+      "num_input_tokens_seen": 32886208,
+      "step": 6145
+    },
+    {
+      "epoch": 0.984,
+      "grad_norm": 1.2251099348068237,
+      "learning_rate": 3.157608484956332e-08,
+      "loss": 0.7723,
+      "num_input_tokens_seen": 32912960,
+      "step": 6150
+    },
+    {
+      "epoch": 0.9848,
+      "grad_norm": 1.0463021993637085,
+      "learning_rate": 2.8498001623286642e-08,
+      "loss": 0.6472,
+      "num_input_tokens_seen": 32941072,
+      "step": 6155
+    },
+    {
+      "epoch": 0.9856,
+      "grad_norm": 1.001555323600769,
+      "learning_rate": 2.557765197543638e-08,
+      "loss": 0.7462,
+      "num_input_tokens_seen": 32968304,
+      "step": 6160
+    },
+    {
+      "epoch": 0.9864,
+      "grad_norm": 1.0437195301055908,
+      "learning_rate": 2.281505435253184e-08,
+      "loss": 0.7079,
+      "num_input_tokens_seen": 32993008,
+      "step": 6165
+    },
+    {
+      "epoch": 0.9872,
+      "grad_norm": 1.1048009395599365,
+      "learning_rate": 2.0210226204639414e-08,
+      "loss": 0.7058,
+      "num_input_tokens_seen": 33018128,
+      "step": 6170
+    },
+    {
+      "epoch": 0.988,
+      "grad_norm": 1.1593177318572998,
+      "learning_rate": 1.7763183985269883e-08,
+      "loss": 0.5942,
+      "num_input_tokens_seen": 33047968,
+      "step": 6175
+    },
+    {
+      "epoch": 0.9888,
+      "grad_norm": 1.3036433458328247,
+      "learning_rate": 1.5473943151270153e-08,
+      "loss": 0.7738,
+      "num_input_tokens_seen": 33072560,
+      "step": 6180
+    },
+    {
+      "epoch": 0.9896,
+      "grad_norm": 1.0337815284729004,
+      "learning_rate": 1.3342518162728912e-08,
+      "loss": 0.781,
+      "num_input_tokens_seen": 33096928,
+      "step": 6185
+    },
+    {
+      "epoch": 0.9904,
+      "grad_norm": 1.0598255395889282,
+      "learning_rate": 1.136892248288779e-08,
+      "loss": 0.6607,
+      "num_input_tokens_seen": 33123808,
+      "step": 6190
+    },
+    {
+      "epoch": 0.9912,
+      "grad_norm": 1.658722162246704,
+      "learning_rate": 9.553168578049775e-09,
+      "loss": 0.7506,
+      "num_input_tokens_seen": 33148688,
+      "step": 6195
+    },
+    {
+      "epoch": 0.992,
+      "grad_norm": 0.9460881352424622,
+      "learning_rate": 7.895267917501504e-09,
+      "loss": 0.6521,
+      "num_input_tokens_seen": 33176624,
+      "step": 6200
+    },
+    {
+      "epoch": 0.9928,
+      "grad_norm": 0.9642547369003296,
+      "learning_rate": 6.395230973443856e-09,
+      "loss": 0.709,
+      "num_input_tokens_seen": 33198160,
+      "step": 6205
+    },
+    {
+      "epoch": 0.9936,
+      "grad_norm": 1.1588774919509888,
+      "learning_rate": 5.053067220925356e-09,
+      "loss": 0.6685,
+      "num_input_tokens_seen": 33226336,
+      "step": 6210
+    },
+    {
+      "epoch": 0.9944,
+      "grad_norm": 0.7818155288696289,
+      "learning_rate": 3.868785137786657e-09,
+      "loss": 0.6672,
+      "num_input_tokens_seen": 33245824,
+      "step": 6215
+    },
+    {
+      "epoch": 0.9952,
+      "grad_norm": 0.8517420887947083,
+      "learning_rate": 2.842392204591149e-09,
+      "loss": 0.7053,
+      "num_input_tokens_seen": 33274176,
+      "step": 6220
+    },
+    {
+      "epoch": 0.996,
+      "grad_norm": 1.1486226320266724,
+      "learning_rate": 1.973894904597207e-09,
+      "loss": 0.7184,
+      "num_input_tokens_seen": 33302528,
+      "step": 6225
+    },
+    {
+      "epoch": 0.9968,
+      "grad_norm": 0.9121309518814087,
+      "learning_rate": 1.2632987237054528e-09,
+      "loss": 0.7092,
+      "num_input_tokens_seen": 33330384,
+      "step": 6230
+    },
+    {
+      "epoch": 0.9976,
+      "grad_norm": 1.1165632009506226,
+      "learning_rate": 7.106081504254514e-10,
+      "loss": 0.6142,
+      "num_input_tokens_seen": 33357968,
+      "step": 6235
+    },
+    {
+      "epoch": 0.9984,
+      "grad_norm": 1.0336644649505615,
+      "learning_rate": 3.158266758562789e-10,
+      "loss": 0.7147,
+      "num_input_tokens_seen": 33381536,
+      "step": 6240
+    },
+    {
+      "epoch": 0.9992,
+      "grad_norm": 1.1494331359863281,
+      "learning_rate": 7.89567936476665e-11,
+      "loss": 0.7005,
+      "num_input_tokens_seen": 33409984,
+      "step": 6245
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.0414918661117554,
+      "learning_rate": 0.0,
+      "loss": 0.7516,
+      "num_input_tokens_seen": 33437856,
+      "step": 6250
+    },
+    {
+      "epoch": 1.0,
+      "num_input_tokens_seen": 33437856,
+      "step": 6250,
+      "total_flos": 7.1914395644928e+16,
+      "train_loss": 0.7151971128082275,
+      "train_runtime": 36754.4929,
+      "train_samples_per_second": 2.721,
+      "train_steps_per_second": 0.17
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 6250,
+  "num_input_tokens_seen": 33437856,
+  "num_train_epochs": 1,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.1914395644928e+16,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}