diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,4199 @@
+{
+  "best_global_step": 2375,
+  "best_metric": 0.3474566638469696,
+  "best_model_checkpoint": "saves/prefix-tuning/llama-3-8b-instruct/train_wsc_1756729607/checkpoint-2375",
+  "epoch": 10.0,
+  "eval_steps": 125,
+  "global_step": 2490,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.020080321285140562,
+      "grad_norm": 208.77610778808594,
+      "learning_rate": 8.032128514056225e-07,
+      "loss": 11.8262,
+      "num_input_tokens_seen": 832,
+      "step": 5
+    },
+    {
+      "epoch": 0.040160642570281124,
+      "grad_norm": 188.124267578125,
+      "learning_rate": 1.8072289156626506e-06,
+      "loss": 10.3109,
+      "num_input_tokens_seen": 1760,
+      "step": 10
+    },
+    {
+      "epoch": 0.060240963855421686,
+      "grad_norm": 118.94561767578125,
+      "learning_rate": 2.811244979919679e-06,
+      "loss": 8.5244,
+      "num_input_tokens_seen": 2608,
+      "step": 15
+    },
+    {
+      "epoch": 0.08032128514056225,
+      "grad_norm": 93.1993637084961,
+      "learning_rate": 3.8152610441767074e-06,
+      "loss": 6.312,
+      "num_input_tokens_seen": 3536,
+      "step": 20
+    },
+    {
+      "epoch": 0.10040160642570281,
+      "grad_norm": 80.68607330322266,
+      "learning_rate": 4.819277108433735e-06,
+      "loss": 4.854,
+      "num_input_tokens_seen": 4496,
+      "step": 25
+    },
+    {
+      "epoch": 0.12048192771084337,
+      "grad_norm": 53.525691986083984,
+      "learning_rate": 5.823293172690764e-06,
+      "loss": 2.9588,
+      "num_input_tokens_seen": 5424,
+      "step": 30
+    },
+    {
+      "epoch": 0.14056224899598393,
+      "grad_norm": 74.08545684814453,
+      "learning_rate": 6.827309236947792e-06,
+      "loss": 1.8179,
+      "num_input_tokens_seen": 6304,
+      "step": 35
+    },
+    {
+      "epoch": 0.1606425702811245,
+      "grad_norm": 39.353179931640625,
+      "learning_rate": 7.83132530120482e-06,
+      "loss": 0.9304,
+      "num_input_tokens_seen": 7072,
+      "step": 40
+    },
+    {
+      "epoch": 0.18072289156626506,
+      "grad_norm": 34.21826934814453,
+      "learning_rate": 8.835341365461847e-06,
+      "loss": 0.7447,
+      "num_input_tokens_seen": 7856,
+      "step": 45
+    },
+    {
+      "epoch": 0.20080321285140562,
+      "grad_norm": 46.137203216552734,
+      "learning_rate": 9.839357429718876e-06,
+      "loss": 0.4663,
+      "num_input_tokens_seen": 8880,
+      "step": 50
+    },
+    {
+      "epoch": 0.22088353413654618,
+      "grad_norm": 21.007869720458984,
+      "learning_rate": 1.0843373493975904e-05,
+      "loss": 0.419,
+      "num_input_tokens_seen": 9680,
+      "step": 55
+    },
+    {
+      "epoch": 0.24096385542168675,
+      "grad_norm": 32.040428161621094,
+      "learning_rate": 1.1847389558232933e-05,
+      "loss": 0.3886,
+      "num_input_tokens_seen": 10576,
+      "step": 60
+    },
+    {
+      "epoch": 0.26104417670682734,
+      "grad_norm": 33.35049819946289,
+      "learning_rate": 1.285140562248996e-05,
+      "loss": 0.3719,
+      "num_input_tokens_seen": 11424,
+      "step": 65
+    },
+    {
+      "epoch": 0.28112449799196787,
+      "grad_norm": 27.942962646484375,
+      "learning_rate": 1.3855421686746989e-05,
+      "loss": 0.4068,
+      "num_input_tokens_seen": 12224,
+      "step": 70
+    },
+    {
+      "epoch": 0.30120481927710846,
+      "grad_norm": 48.38943862915039,
+      "learning_rate": 1.4859437751004016e-05,
+      "loss": 0.4348,
+      "num_input_tokens_seen": 13168,
+      "step": 75
+    },
+    {
+      "epoch": 0.321285140562249,
+      "grad_norm": 54.91352462768555,
+      "learning_rate": 1.5863453815261046e-05,
+      "loss": 0.4399,
+      "num_input_tokens_seen": 14080,
+      "step": 80
+    },
+    {
+      "epoch": 0.3413654618473896,
+      "grad_norm": 100.44646453857422,
+      "learning_rate": 1.6867469879518073e-05,
+      "loss": 0.873,
+      "num_input_tokens_seen": 15056,
+      "step": 85
+    },
+    {
+      "epoch": 0.3614457831325301,
+      "grad_norm": 26.445947647094727,
+      "learning_rate": 1.78714859437751e-05,
+      "loss": 0.5697,
+      "num_input_tokens_seen": 15904,
+      "step": 90
+    },
+    {
+      "epoch": 0.3815261044176707,
+      "grad_norm": 8.536154747009277,
+      "learning_rate": 1.8875502008032127e-05,
+      "loss": 0.318,
+      "num_input_tokens_seen": 16688,
+      "step": 95
+    },
+    {
+      "epoch": 0.40160642570281124,
+      "grad_norm": 0.5681027173995972,
+      "learning_rate": 1.9879518072289157e-05,
+      "loss": 0.0212,
+      "num_input_tokens_seen": 17552,
+      "step": 100
+    },
+    {
+      "epoch": 0.42168674698795183,
+      "grad_norm": 43.08428955078125,
+      "learning_rate": 2.0883534136546184e-05,
+      "loss": 1.5137,
+      "num_input_tokens_seen": 18400,
+      "step": 105
+    },
+    {
+      "epoch": 0.44176706827309237,
+      "grad_norm": 32.574432373046875,
+      "learning_rate": 2.1887550200803214e-05,
+      "loss": 0.9558,
+      "num_input_tokens_seen": 19456,
+      "step": 110
+    },
+    {
+      "epoch": 0.46184738955823296,
+      "grad_norm": 12.149736404418945,
+      "learning_rate": 2.289156626506024e-05,
+      "loss": 0.2867,
+      "num_input_tokens_seen": 20288,
+      "step": 115
+    },
+    {
+      "epoch": 0.4819277108433735,
+      "grad_norm": 59.057838439941406,
+      "learning_rate": 2.389558232931727e-05,
+      "loss": 0.7216,
+      "num_input_tokens_seen": 21328,
+      "step": 120
+    },
+    {
+      "epoch": 0.5020080321285141,
+      "grad_norm": 38.10276794433594,
+      "learning_rate": 2.48995983935743e-05,
+      "loss": 0.5349,
+      "num_input_tokens_seen": 22304,
+      "step": 125
+    },
+    {
+      "epoch": 0.5020080321285141,
+      "eval_loss": 1.2113524675369263,
+      "eval_runtime": 1.2329,
+      "eval_samples_per_second": 45.42,
+      "eval_steps_per_second": 22.71,
+      "num_input_tokens_seen": 22304,
+      "step": 125
+    },
+    {
+      "epoch": 0.5220883534136547,
+      "grad_norm": 51.98005676269531,
+      "learning_rate": 2.5903614457831325e-05,
+      "loss": 1.3225,
+      "num_input_tokens_seen": 23056,
+      "step": 130
+    },
+    {
+      "epoch": 0.5421686746987951,
+      "grad_norm": 16.78879165649414,
+      "learning_rate": 2.6907630522088356e-05,
+      "loss": 0.3651,
+      "num_input_tokens_seen": 23840,
+      "step": 135
+    },
+    {
+      "epoch": 0.5622489959839357,
+      "grad_norm": 7.14707088470459,
+      "learning_rate": 2.791164658634538e-05,
+      "loss": 0.5608,
+      "num_input_tokens_seen": 24832,
+      "step": 140
+    },
+    {
+      "epoch": 0.5823293172690763,
+      "grad_norm": 2.345181465148926,
+      "learning_rate": 2.891566265060241e-05,
+      "loss": 0.4016,
+      "num_input_tokens_seen": 25648,
+      "step": 145
+    },
+    {
+      "epoch": 0.6024096385542169,
+      "grad_norm": 13.92953109741211,
+      "learning_rate": 2.991967871485944e-05,
+      "loss": 0.4123,
+      "num_input_tokens_seen": 26496,
+      "step": 150
+    },
+    {
+      "epoch": 0.6224899598393574,
+      "grad_norm": 13.532402038574219,
+      "learning_rate": 3.092369477911647e-05,
+      "loss": 0.538,
+      "num_input_tokens_seen": 27392,
+      "step": 155
+    },
+    {
+      "epoch": 0.642570281124498,
+      "grad_norm": 11.975309371948242,
+      "learning_rate": 3.192771084337349e-05,
+      "loss": 0.3054,
+      "num_input_tokens_seen": 28272,
+      "step": 160
+    },
+    {
+      "epoch": 0.6626506024096386,
+      "grad_norm": 10.431339263916016,
+      "learning_rate": 3.2931726907630524e-05,
+      "loss": 0.4957,
+      "num_input_tokens_seen": 29184,
+      "step": 165
+    },
+    {
+      "epoch": 0.6827309236947792,
+      "grad_norm": 2.0895514488220215,
+      "learning_rate": 3.393574297188755e-05,
+      "loss": 0.3453,
+      "num_input_tokens_seen": 30128,
+      "step": 170
+    },
+    {
+      "epoch": 0.7028112449799196,
+      "grad_norm": 3.5245909690856934,
+      "learning_rate": 3.4939759036144585e-05,
+      "loss": 0.4688,
+      "num_input_tokens_seen": 30976,
+      "step": 175
+    },
+    {
+      "epoch": 0.7228915662650602,
+      "grad_norm": 5.512556552886963,
+      "learning_rate": 3.5943775100401605e-05,
+      "loss": 0.4289,
+      "num_input_tokens_seen": 31776,
+      "step": 180
+    },
+    {
+      "epoch": 0.7429718875502008,
+      "grad_norm": 3.1914124488830566,
+      "learning_rate": 3.694779116465863e-05,
+      "loss": 0.3237,
+      "num_input_tokens_seen": 32608,
+      "step": 185
+    },
+    {
+      "epoch": 0.7630522088353414,
+      "grad_norm": 12.990925788879395,
+      "learning_rate": 3.7951807228915666e-05,
+      "loss": 0.5221,
+      "num_input_tokens_seen": 33360,
+      "step": 190
+    },
+    {
+      "epoch": 0.7831325301204819,
+      "grad_norm": 0.20752322673797607,
+      "learning_rate": 3.895582329317269e-05,
+      "loss": 0.2575,
+      "num_input_tokens_seen": 34176,
+      "step": 195
+    },
+    {
+      "epoch": 0.8032128514056225,
+      "grad_norm": 3.7412514686584473,
+      "learning_rate": 3.995983935742972e-05,
+      "loss": 1.23,
+      "num_input_tokens_seen": 34992,
+      "step": 200
+    },
+    {
+      "epoch": 0.8232931726907631,
+      "grad_norm": 2.3068435192108154,
+      "learning_rate": 4.0963855421686746e-05,
+      "loss": 0.4066,
+      "num_input_tokens_seen": 35888,
+      "step": 205
+    },
+    {
+      "epoch": 0.8433734939759037,
+      "grad_norm": 2.8328449726104736,
+      "learning_rate": 4.196787148594378e-05,
+      "loss": 0.3846,
+      "num_input_tokens_seen": 36848,
+      "step": 210
+    },
+    {
+      "epoch": 0.8634538152610441,
+      "grad_norm": 2.105517625808716,
+      "learning_rate": 4.297188755020081e-05,
+      "loss": 0.3303,
+      "num_input_tokens_seen": 37888,
+      "step": 215
+    },
+    {
+      "epoch": 0.8835341365461847,
+      "grad_norm": 2.6428239345550537,
+      "learning_rate": 4.3975903614457834e-05,
+      "loss": 0.4356,
+      "num_input_tokens_seen": 38768,
+      "step": 220
+    },
+    {
+      "epoch": 0.9036144578313253,
+      "grad_norm": 6.220267295837402,
+      "learning_rate": 4.497991967871486e-05,
+      "loss": 0.4328,
+      "num_input_tokens_seen": 39488,
+      "step": 225
+    },
+    {
+      "epoch": 0.9236947791164659,
+      "grad_norm": 4.502399444580078,
+      "learning_rate": 4.598393574297189e-05,
+      "loss": 0.6742,
+      "num_input_tokens_seen": 40336,
+      "step": 230
+    },
+    {
+      "epoch": 0.9437751004016064,
+      "grad_norm": 4.788579940795898,
+      "learning_rate": 4.698795180722892e-05,
+      "loss": 0.3987,
+      "num_input_tokens_seen": 41328,
+      "step": 235
+    },
+    {
+      "epoch": 0.963855421686747,
+      "grad_norm": 3.492311954498291,
+      "learning_rate": 4.799196787148594e-05,
+      "loss": 0.3388,
+      "num_input_tokens_seen": 42176,
+      "step": 240
+    },
+    {
+      "epoch": 0.9839357429718876,
+      "grad_norm": 2.018967628479004,
+      "learning_rate": 4.8995983935742975e-05,
+      "loss": 0.8551,
+      "num_input_tokens_seen": 43312,
+      "step": 245
+    },
+    {
+      "epoch": 1.0040160642570282,
+      "grad_norm": 1.773940086364746,
+      "learning_rate": 5e-05,
+      "loss": 0.4306,
+      "num_input_tokens_seen": 44064,
+      "step": 250
+    },
+    {
+      "epoch": 1.0040160642570282,
+      "eval_loss": 0.4651975631713867,
+      "eval_runtime": 1.2232,
+      "eval_samples_per_second": 45.781,
+      "eval_steps_per_second": 22.891,
+      "num_input_tokens_seen": 44064,
+      "step": 250
+    },
+    {
+      "epoch": 1.0240963855421688,
+      "grad_norm": 13.694576263427734,
+      "learning_rate": 4.9999385864396127e-05,
+      "loss": 0.7501,
+      "num_input_tokens_seen": 44816,
+      "step": 255
+    },
+    {
+      "epoch": 1.0441767068273093,
+      "grad_norm": 7.658875942230225,
+      "learning_rate": 4.99975434877575e-05,
+      "loss": 1.1828,
+      "num_input_tokens_seen": 45776,
+      "step": 260
+    },
+    {
+      "epoch": 1.0642570281124497,
+      "grad_norm": 1.0638893842697144,
+      "learning_rate": 4.999447296060165e-05,
+      "loss": 0.7021,
+      "num_input_tokens_seen": 46592,
+      "step": 265
+    },
+    {
+      "epoch": 1.0843373493975903,
+      "grad_norm": 4.289158344268799,
+      "learning_rate": 4.999017443378618e-05,
+      "loss": 0.37,
+      "num_input_tokens_seen": 47536,
+      "step": 270
+    },
+    {
+      "epoch": 1.104417670682731,
+      "grad_norm": 3.189358711242676,
+      "learning_rate": 4.998464811850137e-05,
+      "loss": 0.3415,
+      "num_input_tokens_seen": 48320,
+      "step": 275
+    },
+    {
+      "epoch": 1.1244979919678715,
+      "grad_norm": 2.700920581817627,
+      "learning_rate": 4.997789428625975e-05,
+      "loss": 0.381,
+      "num_input_tokens_seen": 49216,
+      "step": 280
+    },
+    {
+      "epoch": 1.144578313253012,
+      "grad_norm": 2.0674901008605957,
+      "learning_rate": 4.996991326888286e-05,
+      "loss": 0.3487,
+      "num_input_tokens_seen": 50048,
+      "step": 285
+    },
+    {
+      "epoch": 1.1646586345381527,
+      "grad_norm": 2.0477442741394043,
+      "learning_rate": 4.996070545848484e-05,
+      "loss": 0.346,
+      "num_input_tokens_seen": 50832,
+      "step": 290
+    },
+    {
+      "epoch": 1.1847389558232932,
+      "grad_norm": 2.814277410507202,
+      "learning_rate": 4.995027130745321e-05,
+      "loss": 0.3439,
+      "num_input_tokens_seen": 51824,
+      "step": 295
+    },
+    {
+      "epoch": 1.2048192771084336,
+      "grad_norm": 1.5188792943954468,
+      "learning_rate": 4.9938611328426685e-05,
+      "loss": 0.5375,
+      "num_input_tokens_seen": 52608,
+      "step": 300
+    },
+    {
+      "epoch": 1.2248995983935742,
+      "grad_norm": 0.27745991945266724,
+      "learning_rate": 4.992572609426992e-05,
+      "loss": 0.3537,
+      "num_input_tokens_seen": 53440,
+      "step": 305
+    },
+    {
+      "epoch": 1.2449799196787148,
+      "grad_norm": 0.40650656819343567,
+      "learning_rate": 4.99116162380454e-05,
+      "loss": 0.3549,
+      "num_input_tokens_seen": 54320,
+      "step": 310
+    },
+    {
+      "epoch": 1.2650602409638554,
+      "grad_norm": 1.8287845849990845,
+      "learning_rate": 4.989628245298233e-05,
+      "loss": 0.3352,
+      "num_input_tokens_seen": 55072,
+      "step": 315
+    },
+    {
+      "epoch": 1.285140562248996,
+      "grad_norm": 0.5225471258163452,
+      "learning_rate": 4.987972549244257e-05,
+      "loss": 0.3695,
+      "num_input_tokens_seen": 56224,
+      "step": 320
+    },
+    {
+      "epoch": 1.3052208835341366,
+      "grad_norm": 1.361476182937622,
+      "learning_rate": 4.986194616988364e-05,
+      "loss": 0.281,
+      "num_input_tokens_seen": 56912,
+      "step": 325
+    },
+    {
+      "epoch": 1.3253012048192772,
+      "grad_norm": 1.0196197032928467,
+      "learning_rate": 4.984294535881875e-05,
+      "loss": 0.488,
+      "num_input_tokens_seen": 57648,
+      "step": 330
+    },
+    {
+      "epoch": 1.3453815261044177,
+      "grad_norm": 0.7983678579330444,
+      "learning_rate": 4.982272399277386e-05,
+      "loss": 0.3598,
+      "num_input_tokens_seen": 58608,
+      "step": 335
+    },
+    {
+      "epoch": 1.3654618473895583,
+      "grad_norm": 0.21441864967346191,
+      "learning_rate": 4.980128306524183e-05,
+      "loss": 0.3973,
+      "num_input_tokens_seen": 59424,
+      "step": 340
+    },
+    {
+      "epoch": 1.3855421686746987,
+      "grad_norm": 1.7802695035934448,
+      "learning_rate": 4.9778623629633635e-05,
+      "loss": 0.3078,
+      "num_input_tokens_seen": 60272,
+      "step": 345
+    },
+    {
+      "epoch": 1.4056224899598393,
+      "grad_norm": 2.1653060913085938,
+      "learning_rate": 4.975474679922655e-05,
+      "loss": 0.4871,
+      "num_input_tokens_seen": 61056,
+      "step": 350
+    },
+    {
+      "epoch": 1.4257028112449799,
+      "grad_norm": 1.056137204170227,
+      "learning_rate": 4.972965374710952e-05,
+      "loss": 0.283,
+      "num_input_tokens_seen": 61968,
+      "step": 355
+    },
+    {
+      "epoch": 1.4457831325301205,
+      "grad_norm": 0.806865394115448,
+      "learning_rate": 4.9703345706125485e-05,
+      "loss": 0.3467,
+      "num_input_tokens_seen": 62800,
+      "step": 360
+    },
+    {
+      "epoch": 1.465863453815261,
+      "grad_norm": 0.4291660189628601,
+      "learning_rate": 4.96758239688108e-05,
+      "loss": 0.4493,
+      "num_input_tokens_seen": 63824,
+      "step": 365
+    },
+    {
+      "epoch": 1.4859437751004017,
+      "grad_norm": 0.9207323789596558,
+      "learning_rate": 4.964708988733178e-05,
+      "loss": 0.3217,
+      "num_input_tokens_seen": 64800,
+      "step": 370
+    },
+    {
+      "epoch": 1.5060240963855422,
+      "grad_norm": 0.7873408198356628,
+      "learning_rate": 4.961714487341822e-05,
+      "loss": 0.3766,
+      "num_input_tokens_seen": 65808,
+      "step": 375
+    },
+    {
+      "epoch": 1.5060240963855422,
+      "eval_loss": 0.38677313923835754,
+      "eval_runtime": 1.2517,
+      "eval_samples_per_second": 44.74,
+      "eval_steps_per_second": 22.37,
+      "num_input_tokens_seen": 65808,
+      "step": 375
+    },
+    {
+      "epoch": 1.5261044176706826,
+      "grad_norm": 0.23284302651882172,
+      "learning_rate": 4.9585990398294043e-05,
+      "loss": 0.4091,
+      "num_input_tokens_seen": 66752,
+      "step": 380
+    },
+    {
+      "epoch": 1.5461847389558234,
+      "grad_norm": 0.18355536460876465,
+      "learning_rate": 4.9553627992605066e-05,
+      "loss": 0.3531,
+      "num_input_tokens_seen": 67632,
+      "step": 385
+    },
+    {
+      "epoch": 1.5662650602409638,
+      "grad_norm": 0.12290728837251663,
+      "learning_rate": 4.952005924634372e-05,
+      "loss": 0.3506,
+      "num_input_tokens_seen": 68400,
+      "step": 390
+    },
+    {
+      "epoch": 1.5863453815261044,
+      "grad_norm": 0.8354266881942749,
+      "learning_rate": 4.948528580877099e-05,
+      "loss": 0.3255,
+      "num_input_tokens_seen": 69408,
+      "step": 395
+    },
+    {
+      "epoch": 1.606425702811245,
+      "grad_norm": 0.3600679337978363,
+      "learning_rate": 4.944930938833535e-05,
+      "loss": 0.3689,
+      "num_input_tokens_seen": 70352,
+      "step": 400
+    },
+    {
+      "epoch": 1.6265060240963856,
+      "grad_norm": 0.2819594442844391,
+      "learning_rate": 4.9412131752588874e-05,
+      "loss": 0.374,
+      "num_input_tokens_seen": 71184,
+      "step": 405
+    },
+    {
+      "epoch": 1.6465863453815262,
+      "grad_norm": 0.9768486618995667,
+      "learning_rate": 4.937375472810033e-05,
+      "loss": 0.3785,
+      "num_input_tokens_seen": 72272,
+      "step": 410
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.11346471309661865,
+      "learning_rate": 4.9334180200365486e-05,
+      "loss": 0.3645,
+      "num_input_tokens_seen": 73136,
+      "step": 415
+    },
+    {
+      "epoch": 1.6867469879518073,
+      "grad_norm": 0.15687525272369385,
+      "learning_rate": 4.929341011371448e-05,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 73872,
+      "step": 420
+    },
+    {
+      "epoch": 1.7068273092369477,
+      "grad_norm": 0.31356531381607056,
+      "learning_rate": 4.9251446471216226e-05,
+      "loss": 0.3495,
+      "num_input_tokens_seen": 74784,
+      "step": 425
+    },
+    {
+      "epoch": 1.7269076305220885,
+      "grad_norm": 0.2755882441997528,
+      "learning_rate": 4.9208291334580104e-05,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 75664,
+      "step": 430
+    },
+    {
+      "epoch": 1.7469879518072289,
+      "grad_norm": 1.0474437475204468,
+      "learning_rate": 4.9163946824054574e-05,
+      "loss": 0.4005,
+      "num_input_tokens_seen": 76592,
+      "step": 435
+    },
+    {
+      "epoch": 1.7670682730923695,
+      "grad_norm": 0.9340880513191223,
+      "learning_rate": 4.911841511832305e-05,
+      "loss": 0.3454,
+      "num_input_tokens_seen": 77408,
+      "step": 440
+    },
+    {
+      "epoch": 1.78714859437751,
+      "grad_norm": 0.7397641539573669,
+      "learning_rate": 4.907169845439688e-05,
+      "loss": 0.3494,
+      "num_input_tokens_seen": 78272,
+      "step": 445
+    },
+    {
+      "epoch": 1.8072289156626506,
+      "grad_norm": 0.32739967107772827,
+      "learning_rate": 4.902379912750537e-05,
+      "loss": 0.3211,
+      "num_input_tokens_seen": 79200,
+      "step": 450
+    },
+    {
+      "epoch": 1.8273092369477912,
+      "grad_norm": 0.6281508207321167,
+      "learning_rate": 4.897471949098309e-05,
+      "loss": 0.3843,
+      "num_input_tokens_seen": 80112,
+      "step": 455
+    },
+    {
+      "epoch": 1.8473895582329316,
+      "grad_norm": 0.6183443069458008,
+      "learning_rate": 4.892446195615423e-05,
+      "loss": 0.3143,
+      "num_input_tokens_seen": 81168,
+      "step": 460
+    },
+    {
+      "epoch": 1.8674698795180724,
+      "grad_norm": 0.6530241966247559,
+      "learning_rate": 4.88730289922141e-05,
+      "loss": 0.3845,
+      "num_input_tokens_seen": 82112,
+      "step": 465
+    },
+    {
+      "epoch": 1.8875502008032128,
+      "grad_norm": 0.2493213266134262,
+      "learning_rate": 4.8820423126107845e-05,
+      "loss": 0.3467,
+      "num_input_tokens_seen": 83072,
+      "step": 470
+    },
+    {
+      "epoch": 1.9076305220883534,
+      "grad_norm": 0.7678804993629456,
+      "learning_rate": 4.87666469424063e-05,
+      "loss": 0.3683,
+      "num_input_tokens_seen": 83920,
+      "step": 475
+    },
+    {
+      "epoch": 1.927710843373494,
+      "grad_norm": 0.8148710131645203,
+      "learning_rate": 4.8711703083178986e-05,
+      "loss": 0.3512,
+      "num_input_tokens_seen": 84768,
+      "step": 480
+    },
+    {
+      "epoch": 1.9477911646586346,
+      "grad_norm": 0.8058724999427795,
+      "learning_rate": 4.865559424786432e-05,
+      "loss": 0.3478,
+      "num_input_tokens_seen": 85616,
+      "step": 485
+    },
+    {
+      "epoch": 1.9678714859437751,
+      "grad_norm": 0.1543155312538147,
+      "learning_rate": 4.859832319313697e-05,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 86400,
+      "step": 490
+    },
+    {
+      "epoch": 1.9879518072289155,
+      "grad_norm": 1.0387096405029297,
+      "learning_rate": 4.8539892732772455e-05,
+      "loss": 0.3753,
+      "num_input_tokens_seen": 87216,
+      "step": 495
+    },
+    {
+      "epoch": 2.0080321285140563,
+      "grad_norm": 0.2593827545642853,
+      "learning_rate": 4.848030573750885e-05,
+      "loss": 0.3159,
+      "num_input_tokens_seen": 88048,
+      "step": 500
+    },
+    {
+      "epoch": 2.0080321285140563,
+      "eval_loss": 0.3829512596130371,
+      "eval_runtime": 1.2272,
+      "eval_samples_per_second": 45.632,
+      "eval_steps_per_second": 22.816,
+      "num_input_tokens_seen": 88048,
+      "step": 500
+    },
+    {
+      "epoch": 2.0281124497991967,
+      "grad_norm": 0.6328459978103638,
+      "learning_rate": 4.841956513490577e-05,
+      "loss": 0.3501,
+      "num_input_tokens_seen": 88896,
+      "step": 505
+    },
+    {
+      "epoch": 2.0481927710843375,
+      "grad_norm": 0.17319746315479279,
+      "learning_rate": 4.8357673909200563e-05,
+      "loss": 0.3452,
+      "num_input_tokens_seen": 89744,
+      "step": 510
+    },
+    {
+      "epoch": 2.068273092369478,
+      "grad_norm": 0.13001570105552673,
+      "learning_rate": 4.8294635101161645e-05,
+      "loss": 0.3738,
+      "num_input_tokens_seen": 90528,
+      "step": 515
+    },
+    {
+      "epoch": 2.0883534136546187,
+      "grad_norm": 0.1256828010082245,
+      "learning_rate": 4.8230451807939135e-05,
+      "loss": 0.3347,
+      "num_input_tokens_seen": 91360,
+      "step": 520
+    },
+    {
+      "epoch": 2.108433734939759,
+      "grad_norm": 0.7363554835319519,
+      "learning_rate": 4.816512718291267e-05,
+      "loss": 0.346,
+      "num_input_tokens_seen": 92176,
+      "step": 525
+    },
+    {
+      "epoch": 2.1285140562248994,
+      "grad_norm": 0.19434113800525665,
+      "learning_rate": 4.80986644355365e-05,
+      "loss": 0.341,
+      "num_input_tokens_seen": 93104,
+      "step": 530
+    },
+    {
+      "epoch": 2.1485943775100402,
+      "grad_norm": 0.1456967294216156,
+      "learning_rate": 4.803106683118177e-05,
+      "loss": 0.3588,
+      "num_input_tokens_seen": 93984,
+      "step": 535
+    },
+    {
+      "epoch": 2.1686746987951806,
+      "grad_norm": 0.9451452493667603,
+      "learning_rate": 4.796233769097615e-05,
+      "loss": 0.3438,
+      "num_input_tokens_seen": 94896,
+      "step": 540
+    },
+    {
+      "epoch": 2.1887550200803214,
+      "grad_norm": 0.1497451663017273,
+      "learning_rate": 4.789248039164058e-05,
+      "loss": 0.375,
+      "num_input_tokens_seen": 95824,
+      "step": 545
+    },
+    {
+      "epoch": 2.208835341365462,
+      "grad_norm": 0.8181028366088867,
+      "learning_rate": 4.782149836532345e-05,
+      "loss": 0.3607,
+      "num_input_tokens_seen": 96688,
+      "step": 550
+    },
+    {
+      "epoch": 2.2289156626506026,
+      "grad_norm": 0.6427181959152222,
+      "learning_rate": 4.7749395099431924e-05,
+      "loss": 0.3312,
+      "num_input_tokens_seen": 97488,
+      "step": 555
+    },
+    {
+      "epoch": 2.248995983935743,
+      "grad_norm": 1.3496628999710083,
+      "learning_rate": 4.7676174136460625e-05,
+      "loss": 0.4083,
+      "num_input_tokens_seen": 98288,
+      "step": 560
+    },
+    {
+      "epoch": 2.2690763052208833,
+      "grad_norm": 0.2799893915653229,
+      "learning_rate": 4.760183907381757e-05,
+      "loss": 0.3447,
+      "num_input_tokens_seen": 99200,
+      "step": 565
+    },
+    {
+      "epoch": 2.289156626506024,
+      "grad_norm": 0.22829043865203857,
+      "learning_rate": 4.752639356364744e-05,
+      "loss": 0.3228,
+      "num_input_tokens_seen": 99984,
+      "step": 570
+    },
+    {
+      "epoch": 2.3092369477911645,
+      "grad_norm": 0.17503726482391357,
+      "learning_rate": 4.7449841312652166e-05,
+      "loss": 0.3781,
+      "num_input_tokens_seen": 100784,
+      "step": 575
+    },
+    {
+      "epoch": 2.3293172690763053,
+      "grad_norm": 0.8941397666931152,
+      "learning_rate": 4.737218608190878e-05,
+      "loss": 0.367,
+      "num_input_tokens_seen": 101584,
+      "step": 580
+    },
+    {
+      "epoch": 2.3493975903614457,
+      "grad_norm": 0.7553220987319946,
+      "learning_rate": 4.729343168668463e-05,
+      "loss": 0.3603,
+      "num_input_tokens_seen": 102480,
+      "step": 585
+    },
+    {
+      "epoch": 2.3694779116465865,
+      "grad_norm": 0.2362431138753891,
+      "learning_rate": 4.721358199624997e-05,
+      "loss": 0.3631,
+      "num_input_tokens_seen": 103408,
+      "step": 590
+    },
+    {
+      "epoch": 2.389558232931727,
+      "grad_norm": 0.2296190857887268,
+      "learning_rate": 4.713264093368783e-05,
+      "loss": 0.3911,
+      "num_input_tokens_seen": 104160,
+      "step": 595
+    },
+    {
+      "epoch": 2.4096385542168672,
+      "grad_norm": 0.16303616762161255,
+      "learning_rate": 4.705061247570128e-05,
+      "loss": 0.3406,
+      "num_input_tokens_seen": 105040,
+      "step": 600
+    },
+    {
+      "epoch": 2.429718875502008,
+      "grad_norm": 0.15550056099891663,
+      "learning_rate": 4.6967500652418034e-05,
+      "loss": 0.3582,
+      "num_input_tokens_seen": 105856,
+      "step": 605
+    },
+    {
+      "epoch": 2.4497991967871484,
+      "grad_norm": 0.9154328107833862,
+      "learning_rate": 4.6883309547192476e-05,
+      "loss": 0.3701,
+      "num_input_tokens_seen": 106928,
+      "step": 610
+    },
+    {
+      "epoch": 2.4698795180722892,
+      "grad_norm": 0.5483267903327942,
+      "learning_rate": 4.679804329640505e-05,
+      "loss": 0.3423,
+      "num_input_tokens_seen": 107808,
+      "step": 615
+    },
+    {
+      "epoch": 2.4899598393574296,
+      "grad_norm": 0.5274845957756042,
+      "learning_rate": 4.6711706089258955e-05,
+      "loss": 0.3104,
+      "num_input_tokens_seen": 108656,
+      "step": 620
+    },
+    {
+      "epoch": 2.5100401606425704,
+      "grad_norm": 0.23020295798778534,
+      "learning_rate": 4.6624302167574436e-05,
+      "loss": 0.3958,
+      "num_input_tokens_seen": 109696,
+      "step": 625
+    },
+    {
+      "epoch": 2.5100401606425704,
+      "eval_loss": 0.3675794303417206,
+      "eval_runtime": 1.213,
+      "eval_samples_per_second": 46.168,
+      "eval_steps_per_second": 23.084,
+      "num_input_tokens_seen": 109696,
+      "step": 625
+    },
+    {
+      "epoch": 2.5301204819277108,
+      "grad_norm": 0.13925790786743164,
+      "learning_rate": 4.653583582558031e-05,
+      "loss": 0.3587,
+      "num_input_tokens_seen": 110576,
+      "step": 630
+    },
+    {
+      "epoch": 2.550200803212851,
+      "grad_norm": 0.9564950466156006,
+      "learning_rate": 4.6446311409703006e-05,
+      "loss": 0.365,
+      "num_input_tokens_seen": 111440,
+      "step": 635
+    },
+    {
+      "epoch": 2.570281124497992,
+      "grad_norm": 0.2103191763162613,
+      "learning_rate": 4.635573331835302e-05,
+      "loss": 0.3339,
+      "num_input_tokens_seen": 112192,
+      "step": 640
+    },
+    {
+      "epoch": 2.5903614457831328,
+      "grad_norm": 0.230157271027565,
+      "learning_rate": 4.6264106001708824e-05,
+      "loss": 0.3631,
+      "num_input_tokens_seen": 113024,
+      "step": 645
+    },
+    {
+      "epoch": 2.610441767068273,
+      "grad_norm": 0.19389480352401733,
+      "learning_rate": 4.61714339614982e-05,
+      "loss": 0.3795,
+      "num_input_tokens_seen": 113952,
+      "step": 650
+    },
+    {
+      "epoch": 2.6305220883534135,
+      "grad_norm": 0.14889311790466309,
+      "learning_rate": 4.607772175077711e-05,
+      "loss": 0.3586,
+      "num_input_tokens_seen": 114928,
+      "step": 655
+    },
+    {
+      "epoch": 2.6506024096385543,
+      "grad_norm": 0.10515565425157547,
+      "learning_rate": 4.598297397370596e-05,
+      "loss": 0.3726,
+      "num_input_tokens_seen": 115728,
+      "step": 660
+    },
+    {
+      "epoch": 2.6706827309236947,
+      "grad_norm": 0.6394887566566467,
+      "learning_rate": 4.588719528532342e-05,
+      "loss": 0.3549,
+      "num_input_tokens_seen": 116544,
+      "step": 665
+    },
+    {
+      "epoch": 2.6907630522088355,
+      "grad_norm": 0.14106032252311707,
+      "learning_rate": 4.5790390391317675e-05,
+      "loss": 0.3379,
+      "num_input_tokens_seen": 117568,
+      "step": 670
+    },
+    {
+      "epoch": 2.710843373493976,
+      "grad_norm": 0.08746648579835892,
+      "learning_rate": 4.5692564047795316e-05,
+      "loss": 0.3688,
+      "num_input_tokens_seen": 118368,
+      "step": 675
+    },
+    {
+      "epoch": 2.7309236947791167,
+      "grad_norm": 0.5902343392372131,
+      "learning_rate": 4.5593721061047576e-05,
+      "loss": 0.3455,
+      "num_input_tokens_seen": 119120,
+      "step": 680
+    },
+    {
+      "epoch": 2.751004016064257,
+      "grad_norm": 0.57841557264328,
+      "learning_rate": 4.549386628731425e-05,
+      "loss": 0.3575,
+      "num_input_tokens_seen": 120064,
+      "step": 685
+    },
+    {
+      "epoch": 2.7710843373493974,
+      "grad_norm": 0.10715785622596741,
+      "learning_rate": 4.5393004632545064e-05,
+      "loss": 0.3721,
+      "num_input_tokens_seen": 120960,
+      "step": 690
+    },
+    {
+      "epoch": 2.791164658634538,
+      "grad_norm": 0.09298089146614075,
+      "learning_rate": 4.529114105215869e-05,
+      "loss": 0.3545,
+      "num_input_tokens_seen": 121760,
+      "step": 695
+    },
+    {
+      "epoch": 2.8112449799196786,
+      "grad_norm": 0.659833550453186,
+      "learning_rate": 4.518828055079925e-05,
+      "loss": 0.3675,
+      "num_input_tokens_seen": 122720,
+      "step": 700
+    },
+    {
+      "epoch": 2.8313253012048194,
+      "grad_norm": 0.600629985332489,
+      "learning_rate": 4.508442818209042e-05,
+      "loss": 0.3543,
+      "num_input_tokens_seen": 123712,
+      "step": 705
+    },
+    {
+      "epoch": 2.8514056224899598,
+      "grad_norm": 0.1273653358221054,
+      "learning_rate": 4.4979589048387186e-05,
+      "loss": 0.3561,
+      "num_input_tokens_seen": 124624,
+      "step": 710
+    },
+    {
+      "epoch": 2.8714859437751006,
+      "grad_norm": 0.5190867185592651,
+      "learning_rate": 4.487376830052511e-05,
+      "loss": 0.3474,
+      "num_input_tokens_seen": 125696,
+      "step": 715
+    },
+    {
+      "epoch": 2.891566265060241,
+      "grad_norm": 0.5262147784233093,
+      "learning_rate": 4.476697113756731e-05,
+      "loss": 0.2977,
+      "num_input_tokens_seen": 126480,
+      "step": 720
+    },
+    {
+      "epoch": 2.9116465863453813,
+      "grad_norm": 0.9929115772247314,
+      "learning_rate": 4.465920280654901e-05,
+      "loss": 0.3658,
+      "num_input_tokens_seen": 127312,
+      "step": 725
+    },
+    {
+      "epoch": 2.931726907630522,
+      "grad_norm": 0.26060280203819275,
+      "learning_rate": 4.4550468602219716e-05,
+      "loss": 0.3475,
+      "num_input_tokens_seen": 128352,
+      "step": 730
+    },
+    {
+      "epoch": 2.9518072289156625,
+      "grad_norm": 0.2218063920736313,
+      "learning_rate": 4.4440773866783136e-05,
+      "loss": 0.4262,
+      "num_input_tokens_seen": 129232,
+      "step": 735
+    },
+    {
+      "epoch": 2.9718875502008033,
+      "grad_norm": 0.6310736536979675,
+      "learning_rate": 4.433012398963468e-05,
+      "loss": 0.4037,
+      "num_input_tokens_seen": 130080,
+      "step": 740
+    },
+    {
+      "epoch": 2.9919678714859437,
+      "grad_norm": 0.5500450134277344,
+      "learning_rate": 4.421852440709666e-05,
+      "loss": 0.3459,
+      "num_input_tokens_seen": 130880,
+      "step": 745
+    },
+    {
+      "epoch": 3.0120481927710845,
+      "grad_norm": 0.5436108112335205,
+      "learning_rate": 4.4105980602151256e-05,
+      "loss": 0.3521,
+      "num_input_tokens_seen": 131872,
+      "step": 750
+    },
+    {
+      "epoch": 3.0120481927710845,
+      "eval_loss": 0.35073402523994446,
+      "eval_runtime": 1.2195,
+      "eval_samples_per_second": 45.921,
+      "eval_steps_per_second": 22.961,
+      "num_input_tokens_seen": 131872,
+      "step": 750
+    },
+    {
+      "epoch": 3.032128514056225,
+      "grad_norm": 0.5504011511802673,
+      "learning_rate": 4.399249810417108e-05,
+      "loss": 0.354,
+      "num_input_tokens_seen": 132656,
+      "step": 755
+    },
+    {
+      "epoch": 3.0522088353413657,
+      "grad_norm": 0.09489892423152924,
+      "learning_rate": 4.387808248864751e-05,
+      "loss": 0.3708,
+      "num_input_tokens_seen": 133472,
+      "step": 760
+    },
+    {
+      "epoch": 3.072289156626506,
+      "grad_norm": 0.1529596596956253,
+      "learning_rate": 4.376273937691681e-05,
+      "loss": 0.3463,
+      "num_input_tokens_seen": 134416,
+      "step": 765
+    },
+    {
+      "epoch": 3.0923694779116464,
+      "grad_norm": 0.08475496619939804,
+      "learning_rate": 4.364647443588389e-05,
+      "loss": 0.3485,
+      "num_input_tokens_seen": 135344,
+      "step": 770
+    },
+    {
+      "epoch": 3.112449799196787,
+      "grad_norm": 0.12930436432361603,
+      "learning_rate": 4.352929337774395e-05,
+      "loss": 0.3382,
+      "num_input_tokens_seen": 136240,
+      "step": 775
+    },
+    {
+      "epoch": 3.1325301204819276,
+      "grad_norm": 1.1815729141235352,
+      "learning_rate": 4.341120195970178e-05,
+      "loss": 0.3559,
+      "num_input_tokens_seen": 137120,
+      "step": 780
+    },
+    {
+      "epoch": 3.1526104417670684,
+      "grad_norm": 0.3863106071949005,
+      "learning_rate": 4.3292205983688905e-05,
+      "loss": 0.36,
+      "num_input_tokens_seen": 138112,
+      "step": 785
+    },
+    {
+      "epoch": 3.1726907630522088,
+      "grad_norm": 0.4967614412307739,
+      "learning_rate": 4.3172311296078595e-05,
+      "loss": 0.3472,
+      "num_input_tokens_seen": 138960,
+      "step": 790
+    },
+    {
+      "epoch": 3.1927710843373496,
+      "grad_norm": 0.13387857377529144,
+      "learning_rate": 4.305152378739855e-05,
+      "loss": 0.3646,
+      "num_input_tokens_seen": 140016,
+      "step": 795
+    },
+    {
+      "epoch": 3.21285140562249,
+      "grad_norm": 0.11857640743255615,
+      "learning_rate": 4.292984939204155e-05,
+      "loss": 0.3357,
+      "num_input_tokens_seen": 140768,
+      "step": 800
+    },
+    {
+      "epoch": 3.2329317269076308,
+      "grad_norm": 0.12806545197963715,
+      "learning_rate": 4.2807294087973834e-05,
+      "loss": 0.3444,
+      "num_input_tokens_seen": 141664,
+      "step": 805
+    },
+    {
+      "epoch": 3.253012048192771,
+      "grad_norm": 0.11283908039331436,
+      "learning_rate": 4.2683863896441475e-05,
+      "loss": 0.3541,
+      "num_input_tokens_seen": 142448,
+      "step": 810
+    },
+    {
+      "epoch": 3.2730923694779115,
+      "grad_norm": 0.06589579582214355,
+      "learning_rate": 4.255956488167449e-05,
+      "loss": 0.3619,
+      "num_input_tokens_seen": 143408,
+      "step": 815
+    },
+    {
+      "epoch": 3.2931726907630523,
+      "grad_norm": 0.6529943346977234,
+      "learning_rate": 4.2434403150588895e-05,
+      "loss": 0.3449,
+      "num_input_tokens_seen": 144256,
+      "step": 820
+    },
+    {
+      "epoch": 3.3132530120481927,
+      "grad_norm": 0.618222177028656,
+      "learning_rate": 4.230838485248674e-05,
+      "loss": 0.3504,
+      "num_input_tokens_seen": 145120,
+      "step": 825
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 0.566098690032959,
+      "learning_rate": 4.21815161787539e-05,
+      "loss": 0.3445,
+      "num_input_tokens_seen": 146080,
+      "step": 830
+    },
+    {
+      "epoch": 3.353413654618474,
+      "grad_norm": 0.13646817207336426,
+      "learning_rate": 4.205380336255594e-05,
+      "loss": 0.3418,
+      "num_input_tokens_seen": 146912,
+      "step": 835
+    },
+    {
+      "epoch": 3.3734939759036147,
+      "grad_norm": 0.21328403055667877,
+      "learning_rate": 4.192525267853188e-05,
+      "loss": 0.2934,
+      "num_input_tokens_seen": 147776,
+      "step": 840
+    },
+    {
+      "epoch": 3.393574297188755,
+      "grad_norm": 0.3965797424316406,
+      "learning_rate": 4.179587044248585e-05,
+      "loss": 0.2829,
+      "num_input_tokens_seen": 148768,
+      "step": 845
+    },
+    {
+      "epoch": 3.4136546184738954,
+      "grad_norm": 0.24930249154567719,
+      "learning_rate": 4.166566301107687e-05,
+      "loss": 0.5387,
+      "num_input_tokens_seen": 149728,
+      "step": 850
+    },
+    {
+      "epoch": 3.433734939759036,
+      "grad_norm": 0.7255300879478455,
+      "learning_rate": 4.153463678150651e-05,
+      "loss": 0.3639,
+      "num_input_tokens_seen": 150784,
+      "step": 855
+    },
+    {
+      "epoch": 3.4538152610441766,
+      "grad_norm": 0.11454490572214127,
+      "learning_rate": 4.140279819120457e-05,
+      "loss": 0.3721,
+      "num_input_tokens_seen": 151728,
+      "step": 860
+    },
+    {
+      "epoch": 3.4738955823293174,
+      "grad_norm": 0.7094687819480896,
+      "learning_rate": 4.127015371751284e-05,
+      "loss": 0.3656,
+      "num_input_tokens_seen": 152640,
+      "step": 865
+    },
+    {
+      "epoch": 3.4939759036144578,
+      "grad_norm": 0.5258346796035767,
+      "learning_rate": 4.1136709877366844e-05,
+      "loss": 0.3193,
+      "num_input_tokens_seen": 153424,
+      "step": 870
+    },
+    {
+      "epoch": 3.5140562248995986,
+      "grad_norm": 0.5166998505592346,
+      "learning_rate": 4.100247322697562e-05,
+      "loss": 0.3677,
+      "num_input_tokens_seen": 154416,
+      "step": 875
+    },
+    {
+      "epoch": 3.5140562248995986,
+      "eval_loss": 0.3535325825214386,
+      "eval_runtime": 1.2211,
+      "eval_samples_per_second": 45.861,
+      "eval_steps_per_second": 22.931,
+      "num_input_tokens_seen": 154416,
+      "step": 875
+    },
+    {
+      "epoch": 3.534136546184739,
+      "grad_norm": 0.49516451358795166,
+      "learning_rate": 4.08674503614997e-05,
+      "loss": 0.3907,
+      "num_input_tokens_seen": 155184,
+      "step": 880
+    },
+    {
+      "epoch": 3.5542168674698793,
+      "grad_norm": 0.0980529636144638,
+      "learning_rate": 4.0731647914727004e-05,
+      "loss": 0.3941,
+      "num_input_tokens_seen": 156000,
+      "step": 885
+    },
+    {
+      "epoch": 3.57429718875502,
+      "grad_norm": 0.5644952058792114,
+      "learning_rate": 4.059507255874694e-05,
+      "loss": 0.345,
+      "num_input_tokens_seen": 156976,
+      "step": 890
+    },
+    {
+      "epoch": 3.5943775100401605,
+      "grad_norm": 0.5101115703582764,
+      "learning_rate": 4.0457731003622606e-05,
+      "loss": 0.3331,
+      "num_input_tokens_seen": 157904,
+      "step": 895
+    },
+    {
+      "epoch": 3.6144578313253013,
+      "grad_norm": 0.4910569190979004,
+      "learning_rate": 4.0319629997061116e-05,
+      "loss": 0.3339,
+      "num_input_tokens_seen": 158864,
+      "step": 900
+    },
+    {
+      "epoch": 3.6345381526104417,
+      "grad_norm": 0.48415863513946533,
+      "learning_rate": 4.018077632408207e-05,
+      "loss": 0.2827,
+      "num_input_tokens_seen": 159744,
+      "step": 905
+    },
+    {
+      "epoch": 3.6546184738955825,
+      "grad_norm": 0.4711949825286865,
+      "learning_rate": 4.004117680668422e-05,
+      "loss": 0.3838,
+      "num_input_tokens_seen": 160608,
+      "step": 910
+    },
+    {
+      "epoch": 3.674698795180723,
+      "grad_norm": 0.935171902179718,
+      "learning_rate": 3.990083830351027e-05,
+      "loss": 0.3816,
+      "num_input_tokens_seen": 161488,
+      "step": 915
+    },
+    {
+      "epoch": 3.694779116465863,
+      "grad_norm": 0.48552215099334717,
+      "learning_rate": 3.975976770950994e-05,
+      "loss": 0.4066,
+      "num_input_tokens_seen": 162224,
+      "step": 920
+    },
+    {
+      "epoch": 3.714859437751004,
+      "grad_norm": 0.5080327391624451,
+      "learning_rate": 3.961797195560118e-05,
+      "loss": 0.3183,
+      "num_input_tokens_seen": 163056,
+      "step": 925
+    },
+    {
+      "epoch": 3.734939759036145,
+      "grad_norm": 0.606795072555542,
+      "learning_rate": 3.947545800832967e-05,
+      "loss": 0.3641,
+      "num_input_tokens_seen": 163856,
+      "step": 930
+    },
+    {
+      "epoch": 3.755020080321285,
+      "grad_norm": 0.5324833989143372,
+      "learning_rate": 3.9332232869526534e-05,
+      "loss": 0.3394,
+      "num_input_tokens_seen": 164768,
+      "step": 935
+    },
+    {
+      "epoch": 3.7751004016064256,
+      "grad_norm": 0.10697121173143387,
+      "learning_rate": 3.918830357596434e-05,
+      "loss": 0.3368,
+      "num_input_tokens_seen": 165600,
+      "step": 940
+    },
+    {
+      "epoch": 3.7951807228915664,
+      "grad_norm": 0.13268576562404633,
+      "learning_rate": 3.9043677199011364e-05,
+      "loss": 0.3511,
+      "num_input_tokens_seen": 166400,
+      "step": 945
+    },
+    {
+      "epoch": 3.8152610441767068,
+      "grad_norm": 0.12882153689861298,
+      "learning_rate": 3.889836084428422e-05,
+      "loss": 0.328,
+      "num_input_tokens_seen": 167296,
+      "step": 950
+    },
+    {
+      "epoch": 3.835341365461847,
+      "grad_norm": 0.14181359112262726,
+      "learning_rate": 3.8752361651298675e-05,
+      "loss": 0.369,
+      "num_input_tokens_seen": 168208,
+      "step": 955
+    },
+    {
+      "epoch": 3.855421686746988,
+      "grad_norm": 0.4742559492588043,
+      "learning_rate": 3.860568679311893e-05,
+      "loss": 0.3657,
+      "num_input_tokens_seen": 169056,
+      "step": 960
+    },
+    {
+      "epoch": 3.8755020080321287,
+      "grad_norm": 0.1299924999475479,
+      "learning_rate": 3.8458343476005196e-05,
+      "loss": 0.3849,
+      "num_input_tokens_seen": 169888,
+      "step": 965
+    },
+    {
+      "epoch": 3.895582329317269,
+      "grad_norm": 0.08048601448535919,
+      "learning_rate": 3.8310338939059644e-05,
+      "loss": 0.3541,
+      "num_input_tokens_seen": 170704,
+      "step": 970
+    },
+    {
+      "epoch": 3.9156626506024095,
+      "grad_norm": 0.6629543304443359,
+      "learning_rate": 3.8161680453870715e-05,
+      "loss": 0.3558,
+      "num_input_tokens_seen": 171600,
+      "step": 975
+    },
+    {
+      "epoch": 3.9357429718875503,
+      "grad_norm": 0.15418274700641632,
+      "learning_rate": 3.8012375324155904e-05,
+      "loss": 0.3131,
+      "num_input_tokens_seen": 172480,
+      "step": 980
+    },
+    {
+      "epoch": 3.9558232931726907,
+      "grad_norm": 0.4927317202091217,
+      "learning_rate": 3.7862430885402876e-05,
+      "loss": 0.3661,
+      "num_input_tokens_seen": 173504,
+      "step": 985
+    },
+    {
+      "epoch": 3.9759036144578315,
+      "grad_norm": 0.4768475890159607,
+      "learning_rate": 3.7711854504509135e-05,
+      "loss": 0.3373,
+      "num_input_tokens_seen": 174288,
+      "step": 990
+    },
+    {
+      "epoch": 3.995983935742972,
+      "grad_norm": 0.7225349545478821,
+      "learning_rate": 3.756065357941999e-05,
+      "loss": 0.3623,
+      "num_input_tokens_seen": 175104,
+      "step": 995
+    },
+    {
+      "epoch": 4.016064257028113,
+      "grad_norm": 0.7921448945999146,
+      "learning_rate": 3.740883553876515e-05,
+      "loss": 0.3426,
+      "num_input_tokens_seen": 176048,
+      "step": 1000
+    },
+    {
+      "epoch": 4.016064257028113,
+      "eval_loss": 0.35071006417274475,
+      "eval_runtime": 1.2197,
+      "eval_samples_per_second": 45.913,
+      "eval_steps_per_second": 22.957,
+      "num_input_tokens_seen": 176048,
+      "step": 1000
+    },
+    {
+      "epoch": 4.036144578313253,
+      "grad_norm": 0.11272845417261124,
+      "learning_rate": 3.725640784149375e-05,
+      "loss": 0.4204,
+      "num_input_tokens_seen": 176880,
+      "step": 1005
+    },
+    {
+      "epoch": 4.056224899598393,
+      "grad_norm": 0.08953544497489929,
+      "learning_rate": 3.710337797650787e-05,
+      "loss": 0.339,
+      "num_input_tokens_seen": 177680,
+      "step": 1010
+    },
+    {
+      "epoch": 4.076305220883534,
+      "grad_norm": 0.4361952543258667,
+      "learning_rate": 3.694975346229458e-05,
+      "loss": 0.3311,
+      "num_input_tokens_seen": 178608,
+      "step": 1015
+    },
+    {
+      "epoch": 4.096385542168675,
+      "grad_norm": 0.09541574120521545,
+      "learning_rate": 3.679554184655659e-05,
+      "loss": 0.3611,
+      "num_input_tokens_seen": 179600,
+      "step": 1020
+    },
+    {
+      "epoch": 4.116465863453815,
+      "grad_norm": 0.5854984521865845,
+      "learning_rate": 3.6640750705841405e-05,
+      "loss": 0.3403,
+      "num_input_tokens_seen": 180464,
+      "step": 1025
+    },
+    {
+      "epoch": 4.136546184738956,
+      "grad_norm": 0.10651904344558716,
+      "learning_rate": 3.6485387645169064e-05,
+      "loss": 0.3243,
+      "num_input_tokens_seen": 181344,
+      "step": 1030
+    },
+    {
+      "epoch": 4.156626506024097,
+      "grad_norm": 0.5942978262901306,
+      "learning_rate": 3.632946029765856e-05,
+      "loss": 0.3965,
+      "num_input_tokens_seen": 182080,
+      "step": 1035
+    },
+    {
+      "epoch": 4.176706827309237,
+      "grad_norm": 0.07312840223312378,
+      "learning_rate": 3.617297632415273e-05,
+      "loss": 0.3719,
+      "num_input_tokens_seen": 182848,
+      "step": 1040
+    },
+    {
+      "epoch": 4.196787148594377,
+      "grad_norm": 0.5075451135635376,
+      "learning_rate": 3.601594341284195e-05,
+      "loss": 0.3512,
+      "num_input_tokens_seen": 183840,
+      "step": 1045
+    },
+    {
+      "epoch": 4.216867469879518,
+      "grad_norm": 0.047960445284843445,
+      "learning_rate": 3.5858369278886354e-05,
+      "loss": 0.3388,
+      "num_input_tokens_seen": 184720,
+      "step": 1050
+    },
+    {
+      "epoch": 4.236947791164659,
+      "grad_norm": 0.08333683758974075,
+      "learning_rate": 3.5700261664036827e-05,
+      "loss": 0.3457,
+      "num_input_tokens_seen": 185504,
+      "step": 1055
+    },
+    {
+      "epoch": 4.257028112449799,
+      "grad_norm": 0.0653541311621666,
+      "learning_rate": 3.55416283362546e-05,
+      "loss": 0.3588,
+      "num_input_tokens_seen": 186272,
+      "step": 1060
+    },
+    {
+      "epoch": 4.27710843373494,
+      "grad_norm": 0.5113236308097839,
+      "learning_rate": 3.5382477089329646e-05,
+      "loss": 0.3579,
+      "num_input_tokens_seen": 187296,
+      "step": 1065
+    },
+    {
+      "epoch": 4.2971887550200805,
+      "grad_norm": 0.07462375611066818,
+      "learning_rate": 3.522281574249774e-05,
+      "loss": 0.348,
+      "num_input_tokens_seen": 188320,
+      "step": 1070
+    },
+    {
+      "epoch": 4.317269076305221,
+      "grad_norm": 0.11710739135742188,
+      "learning_rate": 3.5062652140056275e-05,
+      "loss": 0.3282,
+      "num_input_tokens_seen": 189248,
+      "step": 1075
+    },
+    {
+      "epoch": 4.337349397590361,
+      "grad_norm": 0.15031148493289948,
+      "learning_rate": 3.490199415097892e-05,
+      "loss": 0.3005,
+      "num_input_tokens_seen": 190432,
+      "step": 1080
+    },
+    {
+      "epoch": 4.357429718875502,
+      "grad_norm": 0.6503745913505554,
+      "learning_rate": 3.474084966852897e-05,
+      "loss": 0.4539,
+      "num_input_tokens_seen": 191296,
+      "step": 1085
+    },
+    {
+      "epoch": 4.377510040160643,
+      "grad_norm": 0.14889173209667206,
+      "learning_rate": 3.457922660987155e-05,
+      "loss": 0.3682,
+      "num_input_tokens_seen": 192368,
+      "step": 1090
+    },
+    {
+      "epoch": 4.397590361445783,
+      "grad_norm": 0.4457005560398102,
+      "learning_rate": 3.441713291568462e-05,
+      "loss": 0.3338,
+      "num_input_tokens_seen": 193232,
+      "step": 1095
+    },
+    {
+      "epoch": 4.417670682730924,
+      "grad_norm": 0.5409120321273804,
+      "learning_rate": 3.42545765497689e-05,
+      "loss": 0.3587,
+      "num_input_tokens_seen": 194128,
+      "step": 1100
+    },
+    {
+      "epoch": 4.437751004016064,
+      "grad_norm": 0.085002101957798,
+      "learning_rate": 3.409156549865654e-05,
+      "loss": 0.3609,
+      "num_input_tokens_seen": 194944,
+      "step": 1105
+    },
+    {
+      "epoch": 4.457831325301205,
+      "grad_norm": 0.49231743812561035,
+      "learning_rate": 3.392810777121876e-05,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 195840,
+      "step": 1110
+    },
+    {
+      "epoch": 4.477911646586345,
+      "grad_norm": 0.5549922585487366,
+      "learning_rate": 3.376421139827237e-05,
+      "loss": 0.3871,
+      "num_input_tokens_seen": 196640,
+      "step": 1115
+    },
+    {
+      "epoch": 4.497991967871486,
+      "grad_norm": 0.06657743453979492,
+      "learning_rate": 3.3599884432185225e-05,
+      "loss": 0.3481,
+      "num_input_tokens_seen": 197440,
+      "step": 1120
+    },
+    {
+      "epoch": 4.518072289156627,
+      "grad_norm": 0.13579045236110687,
+      "learning_rate": 3.343513494648055e-05,
+      "loss": 0.3393,
+      "num_input_tokens_seen": 198432,
+      "step": 1125
+    },
+    {
+      "epoch": 4.518072289156627,
+      "eval_loss": 0.3545505702495575,
+      "eval_runtime": 1.4272,
+      "eval_samples_per_second": 39.238,
+      "eval_steps_per_second": 19.619,
+      "num_input_tokens_seen": 198432,
+      "step": 1125
+    },
+    {
+      "epoch": 4.538152610441767,
+      "grad_norm": 0.4401510953903198,
+      "learning_rate": 3.326997103544035e-05,
+      "loss": 0.3349,
+      "num_input_tokens_seen": 199232,
+      "step": 1130
+    },
+    {
+      "epoch": 4.5582329317269075,
+      "grad_norm": 0.14849136769771576,
+      "learning_rate": 3.310440081370767e-05,
+      "loss": 0.3373,
+      "num_input_tokens_seen": 200144,
+      "step": 1135
+    },
+    {
+      "epoch": 4.578313253012048,
+      "grad_norm": 0.621387243270874,
+      "learning_rate": 3.2938432415887984e-05,
+      "loss": 0.3213,
+      "num_input_tokens_seen": 200896,
+      "step": 1140
+    },
+    {
+      "epoch": 4.598393574297189,
+      "grad_norm": 0.7517121434211731,
+      "learning_rate": 3.2772073996149435e-05,
+      "loss": 0.3475,
+      "num_input_tokens_seen": 201760,
+      "step": 1145
+    },
+    {
+      "epoch": 4.618473895582329,
+      "grad_norm": 0.41686856746673584,
+      "learning_rate": 3.260533372782234e-05,
+      "loss": 0.4032,
+      "num_input_tokens_seen": 202688,
+      "step": 1150
+    },
+    {
+      "epoch": 4.63855421686747,
+      "grad_norm": 0.6020703315734863,
+      "learning_rate": 3.24382198029975e-05,
+      "loss": 0.3564,
+      "num_input_tokens_seen": 203392,
+      "step": 1155
+    },
+    {
+      "epoch": 4.658634538152611,
+      "grad_norm": 0.40914788842201233,
+      "learning_rate": 3.227074043212383e-05,
+      "loss": 0.322,
+      "num_input_tokens_seen": 204080,
+      "step": 1160
+    },
+    {
+      "epoch": 4.678714859437751,
+      "grad_norm": 0.09926328808069229,
+      "learning_rate": 3.2102903843604885e-05,
+      "loss": 0.373,
+      "num_input_tokens_seen": 204816,
+      "step": 1165
+    },
+    {
+      "epoch": 4.698795180722891,
+      "grad_norm": 0.08548900485038757,
+      "learning_rate": 3.1934718283394646e-05,
+      "loss": 0.3587,
+      "num_input_tokens_seen": 205616,
+      "step": 1170
+    },
+    {
+      "epoch": 4.718875502008032,
+      "grad_norm": 0.11746017634868622,
+      "learning_rate": 3.1766192014592344e-05,
+      "loss": 0.3571,
+      "num_input_tokens_seen": 206512,
+      "step": 1175
+    },
+    {
+      "epoch": 4.738955823293173,
+      "grad_norm": 0.4761631190776825,
+      "learning_rate": 3.1597333317036545e-05,
+      "loss": 0.3507,
+      "num_input_tokens_seen": 207424,
+      "step": 1180
+    },
+    {
+      "epoch": 4.759036144578313,
+      "grad_norm": 0.5010347366333008,
+      "learning_rate": 3.142815048689828e-05,
+      "loss": 0.3575,
+      "num_input_tokens_seen": 208464,
+      "step": 1185
+    },
+    {
+      "epoch": 4.779116465863454,
+      "grad_norm": 0.07341606169939041,
+      "learning_rate": 3.125865183627354e-05,
+      "loss": 0.3579,
+      "num_input_tokens_seen": 209280,
+      "step": 1190
+    },
+    {
+      "epoch": 4.7991967871485945,
+      "grad_norm": 0.43029800057411194,
+      "learning_rate": 3.10888456927748e-05,
+      "loss": 0.3327,
+      "num_input_tokens_seen": 210080,
+      "step": 1195
+    },
+    {
+      "epoch": 4.8192771084337345,
+      "grad_norm": 0.49401140213012695,
+      "learning_rate": 3.091874039912195e-05,
+      "loss": 0.3619,
+      "num_input_tokens_seen": 210960,
+      "step": 1200
+    },
+    {
+      "epoch": 4.839357429718875,
+      "grad_norm": 0.07773241400718689,
+      "learning_rate": 3.074834431273236e-05,
+      "loss": 0.3488,
+      "num_input_tokens_seen": 211776,
+      "step": 1205
+    },
+    {
+      "epoch": 4.859437751004016,
+      "grad_norm": 0.4646616280078888,
+      "learning_rate": 3.057766580531031e-05,
+      "loss": 0.3542,
+      "num_input_tokens_seen": 212576,
+      "step": 1210
+    },
+    {
+      "epoch": 4.879518072289157,
+      "grad_norm": 0.4323027431964874,
+      "learning_rate": 3.0406713262435656e-05,
+      "loss": 0.3362,
+      "num_input_tokens_seen": 213360,
+      "step": 1215
+    },
+    {
+      "epoch": 4.899598393574297,
+      "grad_norm": 0.502923846244812,
+      "learning_rate": 3.0235495083151844e-05,
+      "loss": 0.3814,
+      "num_input_tokens_seen": 214304,
+      "step": 1220
+    },
+    {
+      "epoch": 4.919678714859438,
+      "grad_norm": 0.43908852338790894,
+      "learning_rate": 3.0064019679553274e-05,
+      "loss": 0.3492,
+      "num_input_tokens_seen": 215072,
+      "step": 1225
+    },
+    {
+      "epoch": 4.9397590361445785,
+      "grad_norm": 0.07500998675823212,
+      "learning_rate": 2.9892295476371988e-05,
+      "loss": 0.3542,
+      "num_input_tokens_seen": 215904,
+      "step": 1230
+    },
+    {
+      "epoch": 4.959839357429718,
+      "grad_norm": 0.10780856758356094,
+      "learning_rate": 2.9720330910563772e-05,
+      "loss": 0.3543,
+      "num_input_tokens_seen": 216864,
+      "step": 1235
+    },
+    {
+      "epoch": 4.979919678714859,
+      "grad_norm": 0.0808030292391777,
+      "learning_rate": 2.9548134430893604e-05,
+      "loss": 0.3387,
+      "num_input_tokens_seen": 217856,
+      "step": 1240
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.0589243620634079,
+      "learning_rate": 2.9375714497520623e-05,
+      "loss": 0.339,
+      "num_input_tokens_seen": 218864,
+      "step": 1245
+    },
+    {
+      "epoch": 5.020080321285141,
+      "grad_norm": 0.05416898429393768,
+      "learning_rate": 2.920307958158241e-05,
+      "loss": 0.3601,
+      "num_input_tokens_seen": 219680,
+      "step": 1250
+    },
+    {
+      "epoch": 5.020080321285141,
+      "eval_loss": 0.3591695725917816,
+      "eval_runtime": 1.2186,
+      "eval_samples_per_second": 45.953,
+      "eval_steps_per_second": 22.976,
+      "num_input_tokens_seen": 219680,
+      "step": 1250
+    },
+    {
+      "epoch": 5.040160642570281,
+      "grad_norm": 0.40490707755088806,
+      "learning_rate": 2.903023816477885e-05,
+      "loss": 0.3239,
+      "num_input_tokens_seen": 220560,
+      "step": 1255
+    },
+    {
+      "epoch": 5.0602409638554215,
+      "grad_norm": 0.12411481887102127,
+      "learning_rate": 2.885719873895536e-05,
+      "loss": 0.3419,
+      "num_input_tokens_seen": 221440,
+      "step": 1260
+    },
+    {
+      "epoch": 5.080321285140562,
+      "grad_norm": 0.5238391160964966,
+      "learning_rate": 2.868396980568572e-05,
+      "loss": 0.348,
+      "num_input_tokens_seen": 222304,
+      "step": 1265
+    },
+    {
+      "epoch": 5.100401606425703,
+      "grad_norm": 0.358173131942749,
+      "learning_rate": 2.8510559875854377e-05,
+      "loss": 0.2762,
+      "num_input_tokens_seen": 223248,
+      "step": 1270
+    },
+    {
+      "epoch": 5.120481927710843,
+      "grad_norm": 0.3846758306026459,
+      "learning_rate": 2.833697746923829e-05,
+      "loss": 0.2662,
+      "num_input_tokens_seen": 224000,
+      "step": 1275
+    },
+    {
+      "epoch": 5.140562248995984,
+      "grad_norm": 0.233295738697052,
+      "learning_rate": 2.816323111408835e-05,
+      "loss": 0.3421,
+      "num_input_tokens_seen": 224880,
+      "step": 1280
+    },
+    {
+      "epoch": 5.160642570281125,
+      "grad_norm": 0.7878521084785461,
+      "learning_rate": 2.7989329346710375e-05,
+      "loss": 0.4232,
+      "num_input_tokens_seen": 225776,
+      "step": 1285
+    },
+    {
+      "epoch": 5.180722891566265,
+      "grad_norm": 0.3618873655796051,
+      "learning_rate": 2.7815280711045717e-05,
+      "loss": 0.3838,
+      "num_input_tokens_seen": 226576,
+      "step": 1290
+    },
+    {
+      "epoch": 5.2008032128514055,
+      "grad_norm": 0.14807634055614471,
+      "learning_rate": 2.7641093758251497e-05,
+      "loss": 0.3104,
+      "num_input_tokens_seen": 227360,
+      "step": 1295
+    },
+    {
+      "epoch": 5.220883534136546,
+      "grad_norm": 0.3595626950263977,
+      "learning_rate": 2.7466777046280457e-05,
+      "loss": 0.3105,
+      "num_input_tokens_seen": 228112,
+      "step": 1300
+    },
+    {
+      "epoch": 5.240963855421687,
+      "grad_norm": 0.6282910108566284,
+      "learning_rate": 2.7292339139460556e-05,
+      "loss": 0.3474,
+      "num_input_tokens_seen": 228992,
+      "step": 1305
+    },
+    {
+      "epoch": 5.261044176706827,
+      "grad_norm": 0.357793390750885,
+      "learning_rate": 2.71177886080741e-05,
+      "loss": 0.3076,
+      "num_input_tokens_seen": 229872,
+      "step": 1310
+    },
+    {
+      "epoch": 5.281124497991968,
+      "grad_norm": 0.1209518164396286,
+      "learning_rate": 2.69431340279368e-05,
+      "loss": 0.4231,
+      "num_input_tokens_seen": 230720,
+      "step": 1315
+    },
+    {
+      "epoch": 5.301204819277109,
+      "grad_norm": 0.09653452038764954,
+      "learning_rate": 2.676838397997633e-05,
+      "loss": 0.3725,
+      "num_input_tokens_seen": 231568,
+      "step": 1320
+    },
+    {
+      "epoch": 5.321285140562249,
+      "grad_norm": 0.4242056608200073,
+      "learning_rate": 2.659354704981078e-05,
+      "loss": 0.3237,
+      "num_input_tokens_seen": 232368,
+      "step": 1325
+    },
+    {
+      "epoch": 5.341365461847389,
+      "grad_norm": 0.10253465920686722,
+      "learning_rate": 2.6418631827326857e-05,
+      "loss": 0.3534,
+      "num_input_tokens_seen": 233184,
+      "step": 1330
+    },
+    {
+      "epoch": 5.36144578313253,
+      "grad_norm": 0.08719359338283539,
+      "learning_rate": 2.6243646906257806e-05,
+      "loss": 0.338,
+      "num_input_tokens_seen": 233984,
+      "step": 1335
+    },
+    {
+      "epoch": 5.381526104417671,
+      "grad_norm": 0.5312795639038086,
+      "learning_rate": 2.606860088376126e-05,
+      "loss": 0.3687,
+      "num_input_tokens_seen": 234848,
+      "step": 1340
+    },
+    {
+      "epoch": 5.401606425702811,
+      "grad_norm": 0.47353821992874146,
+      "learning_rate": 2.5893502359996786e-05,
+      "loss": 0.3449,
+      "num_input_tokens_seen": 235760,
+      "step": 1345
+    },
+    {
+      "epoch": 5.421686746987952,
+      "grad_norm": 0.49804431200027466,
+      "learning_rate": 2.5718359937703408e-05,
+      "loss": 0.3504,
+      "num_input_tokens_seen": 236640,
+      "step": 1350
+    },
+    {
+      "epoch": 5.4417670682730925,
+      "grad_norm": 0.5166244506835938,
+      "learning_rate": 2.554318222177689e-05,
+      "loss": 0.3538,
+      "num_input_tokens_seen": 237616,
+      "step": 1355
+    },
+    {
+      "epoch": 5.461847389558233,
+      "grad_norm": 0.1073295921087265,
+      "learning_rate": 2.5367977818847034e-05,
+      "loss": 0.3354,
+      "num_input_tokens_seen": 238528,
+      "step": 1360
+    },
+    {
+      "epoch": 5.481927710843373,
+      "grad_norm": 0.49680086970329285,
+      "learning_rate": 2.519275533685477e-05,
+      "loss": 0.3354,
+      "num_input_tokens_seen": 239424,
+      "step": 1365
+    },
+    {
+      "epoch": 5.502008032128514,
+      "grad_norm": 0.11628666520118713,
+      "learning_rate": 2.5017523384629298e-05,
+      "loss": 0.354,
+      "num_input_tokens_seen": 240272,
+      "step": 1370
+    },
+    {
+      "epoch": 5.522088353413655,
+      "grad_norm": 0.11434385180473328,
+      "learning_rate": 2.484229057146507e-05,
+      "loss": 0.3422,
+      "num_input_tokens_seen": 241136,
+      "step": 1375
+    },
+    {
+      "epoch": 5.522088353413655,
+      "eval_loss": 0.35063984990119934,
+      "eval_runtime": 1.4257,
+      "eval_samples_per_second": 39.28,
+      "eval_steps_per_second": 19.64,
+      "num_input_tokens_seen": 241136,
+      "step": 1375
+    },
+    {
+      "epoch": 5.542168674698795,
+      "grad_norm": 0.5793833136558533,
+      "learning_rate": 2.466706550669886e-05,
+      "loss": 0.3574,
+      "num_input_tokens_seen": 241936,
+      "step": 1380
+    },
+    {
+      "epoch": 5.562248995983936,
+      "grad_norm": 0.5316616296768188,
+      "learning_rate": 2.449185679928672e-05,
+      "loss": 0.3748,
+      "num_input_tokens_seen": 242672,
+      "step": 1385
+    },
+    {
+      "epoch": 5.582329317269076,
+      "grad_norm": 0.10326164960861206,
+      "learning_rate": 2.431667305738112e-05,
+      "loss": 0.3507,
+      "num_input_tokens_seen": 243808,
+      "step": 1390
+    },
+    {
+      "epoch": 5.602409638554217,
+      "grad_norm": 0.46083107590675354,
+      "learning_rate": 2.414152288790787e-05,
+      "loss": 0.3506,
+      "num_input_tokens_seen": 244688,
+      "step": 1395
+    },
+    {
+      "epoch": 5.622489959839357,
+      "grad_norm": 0.42732155323028564,
+      "learning_rate": 2.3966414896143385e-05,
+      "loss": 0.3386,
+      "num_input_tokens_seen": 245696,
+      "step": 1400
+    },
+    {
+      "epoch": 5.642570281124498,
+      "grad_norm": 0.4096117317676544,
+      "learning_rate": 2.3791357685291863e-05,
+      "loss": 0.3298,
+      "num_input_tokens_seen": 246544,
+      "step": 1405
+    },
+    {
+      "epoch": 5.662650602409639,
+      "grad_norm": 0.39851030707359314,
+      "learning_rate": 2.361635985606256e-05,
+      "loss": 0.3413,
+      "num_input_tokens_seen": 247744,
+      "step": 1410
+    },
+    {
+      "epoch": 5.682730923694779,
+      "grad_norm": 0.385628879070282,
+      "learning_rate": 2.344143000624729e-05,
+      "loss": 0.3623,
+      "num_input_tokens_seen": 248480,
+      "step": 1415
+    },
+    {
+      "epoch": 5.7028112449799195,
+      "grad_norm": 0.38799935579299927,
+      "learning_rate": 2.3266576730297956e-05,
+      "loss": 0.3284,
+      "num_input_tokens_seen": 249312,
+      "step": 1420
+    },
+    {
+      "epoch": 5.72289156626506,
+      "grad_norm": 0.11895764619112015,
+      "learning_rate": 2.3091808618904352e-05,
+      "loss": 0.3679,
+      "num_input_tokens_seen": 250304,
+      "step": 1425
+    },
+    {
+      "epoch": 5.742971887550201,
+      "grad_norm": 0.3756169080734253,
+      "learning_rate": 2.2917134258572038e-05,
+      "loss": 0.3506,
+      "num_input_tokens_seen": 251216,
+      "step": 1430
+    },
+    {
+      "epoch": 5.763052208835341,
+      "grad_norm": 0.5232722163200378,
+      "learning_rate": 2.274256223120051e-05,
+      "loss": 0.3512,
+      "num_input_tokens_seen": 251952,
+      "step": 1435
+    },
+    {
+      "epoch": 5.783132530120482,
+      "grad_norm": 0.06907851248979568,
+      "learning_rate": 2.2568101113661577e-05,
+      "loss": 0.3292,
+      "num_input_tokens_seen": 253072,
+      "step": 1440
+    },
+    {
+      "epoch": 5.803212851405623,
+      "grad_norm": 0.08446797728538513,
+      "learning_rate": 2.239375947737793e-05,
+      "loss": 0.3499,
+      "num_input_tokens_seen": 253840,
+      "step": 1445
+    },
+    {
+      "epoch": 5.823293172690763,
+      "grad_norm": 0.06622636318206787,
+      "learning_rate": 2.221954588790206e-05,
+      "loss": 0.3647,
+      "num_input_tokens_seen": 254640,
+      "step": 1450
+    },
+    {
+      "epoch": 5.843373493975903,
+      "grad_norm": 0.03167250007390976,
+      "learning_rate": 2.2045468904495415e-05,
+      "loss": 0.3518,
+      "num_input_tokens_seen": 255456,
+      "step": 1455
+    },
+    {
+      "epoch": 5.863453815261044,
+      "grad_norm": 0.07110590487718582,
+      "learning_rate": 2.1871537079707833e-05,
+      "loss": 0.354,
+      "num_input_tokens_seen": 256304,
+      "step": 1460
+    },
+    {
+      "epoch": 5.883534136546185,
+      "grad_norm": 0.5281655788421631,
+      "learning_rate": 2.1697758958957448e-05,
+      "loss": 0.3385,
+      "num_input_tokens_seen": 257104,
+      "step": 1465
+    },
+    {
+      "epoch": 5.903614457831325,
+      "grad_norm": 0.5472500920295715,
+      "learning_rate": 2.1524143080110716e-05,
+      "loss": 0.3532,
+      "num_input_tokens_seen": 258080,
+      "step": 1470
+    },
+    {
+      "epoch": 5.923694779116466,
+      "grad_norm": 0.5764107704162598,
+      "learning_rate": 2.135069797306308e-05,
+      "loss": 0.3701,
+      "num_input_tokens_seen": 259056,
+      "step": 1475
+    },
+    {
+      "epoch": 5.943775100401607,
+      "grad_norm": 0.5297293663024902,
+      "learning_rate": 2.1177432159319754e-05,
+      "loss": 0.3721,
+      "num_input_tokens_seen": 260000,
+      "step": 1480
+    },
+    {
+      "epoch": 5.9638554216867465,
+      "grad_norm": 0.03520062938332558,
+      "learning_rate": 2.100435415157718e-05,
+      "loss": 0.3517,
+      "num_input_tokens_seen": 260768,
+      "step": 1485
+    },
+    {
+      "epoch": 5.983935742971887,
+      "grad_norm": 0.451326847076416,
+      "learning_rate": 2.083147245330468e-05,
+      "loss": 0.3572,
+      "num_input_tokens_seen": 261760,
+      "step": 1490
+    },
+    {
+      "epoch": 6.004016064257028,
+      "grad_norm": 0.4301218092441559,
+      "learning_rate": 2.0658795558326743e-05,
+      "loss": 0.3476,
+      "num_input_tokens_seen": 262752,
+      "step": 1495
+    },
+    {
+      "epoch": 6.024096385542169,
+      "grad_norm": 0.09133653342723846,
+      "learning_rate": 2.048633195040572e-05,
+      "loss": 0.3609,
+      "num_input_tokens_seen": 263616,
+      "step": 1500
+    },
+    {
+      "epoch": 6.024096385542169,
+      "eval_loss": 0.3502369225025177,
+      "eval_runtime": 1.2111,
+      "eval_samples_per_second": 46.238,
+      "eval_steps_per_second": 23.119,
+      "num_input_tokens_seen": 263616,
+      "step": 1500
+    },
+    {
+      "epoch": 6.044176706827309,
+      "grad_norm": 0.4608069062232971,
+      "learning_rate": 2.0314090102824963e-05,
+      "loss": 0.3669,
+      "num_input_tokens_seen": 264432,
+      "step": 1505
+    },
+    {
+      "epoch": 6.06425702811245,
+      "grad_norm": 0.08459752053022385,
+      "learning_rate": 2.014207847797256e-05,
+      "loss": 0.3542,
+      "num_input_tokens_seen": 265184,
+      "step": 1510
+    },
+    {
+      "epoch": 6.0843373493975905,
+      "grad_norm": 0.04717332869768143,
+      "learning_rate": 1.997030552692556e-05,
+      "loss": 0.3509,
+      "num_input_tokens_seen": 266064,
+      "step": 1515
+    },
+    {
+      "epoch": 6.104417670682731,
+      "grad_norm": 0.07227271795272827,
+      "learning_rate": 1.9798779689034757e-05,
+      "loss": 0.3483,
+      "num_input_tokens_seen": 266928,
+      "step": 1520
+    },
+    {
+      "epoch": 6.124497991967871,
+      "grad_norm": 0.4734782874584198,
+      "learning_rate": 1.9627509391510086e-05,
+      "loss": 0.3542,
+      "num_input_tokens_seen": 267824,
+      "step": 1525
+    },
+    {
+      "epoch": 6.144578313253012,
+      "grad_norm": 0.430178701877594,
+      "learning_rate": 1.9456503049006542e-05,
+      "loss": 0.3479,
+      "num_input_tokens_seen": 268608,
+      "step": 1530
+    },
+    {
+      "epoch": 6.164658634538153,
+      "grad_norm": 0.4436280131340027,
+      "learning_rate": 1.9285769063210812e-05,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 269696,
+      "step": 1535
+    },
+    {
+      "epoch": 6.184738955823293,
+      "grad_norm": 0.44335198402404785,
+      "learning_rate": 1.9115315822428437e-05,
+      "loss": 0.351,
+      "num_input_tokens_seen": 270704,
+      "step": 1540
+    },
+    {
+      "epoch": 6.204819277108434,
+      "grad_norm": 0.0994335189461708,
+      "learning_rate": 1.8945151701171755e-05,
+      "loss": 0.3447,
+      "num_input_tokens_seen": 271568,
+      "step": 1545
+    },
+    {
+      "epoch": 6.224899598393574,
+      "grad_norm": 0.42669767141342163,
+      "learning_rate": 1.877528505974838e-05,
+      "loss": 0.3386,
+      "num_input_tokens_seen": 272304,
+      "step": 1550
+    },
+    {
+      "epoch": 6.244979919678715,
+      "grad_norm": 0.091577909886837,
+      "learning_rate": 1.8605724243850502e-05,
+      "loss": 0.3302,
+      "num_input_tokens_seen": 273152,
+      "step": 1555
+    },
+    {
+      "epoch": 6.265060240963855,
+      "grad_norm": 0.5046164393424988,
+      "learning_rate": 1.8436477584144863e-05,
+      "loss": 0.3962,
+      "num_input_tokens_seen": 274112,
+      "step": 1560
+    },
+    {
+      "epoch": 6.285140562248996,
+      "grad_norm": 0.1121777817606926,
+      "learning_rate": 1.826755339586341e-05,
+      "loss": 0.3337,
+      "num_input_tokens_seen": 274944,
+      "step": 1565
+    },
+    {
+      "epoch": 6.305220883534137,
+      "grad_norm": 0.40517058968544006,
+      "learning_rate": 1.809895997839482e-05,
+      "loss": 0.3484,
+      "num_input_tokens_seen": 275712,
+      "step": 1570
+    },
+    {
+      "epoch": 6.325301204819277,
+      "grad_norm": 0.09261249750852585,
+      "learning_rate": 1.793070561487672e-05,
+      "loss": 0.3391,
+      "num_input_tokens_seen": 276560,
+      "step": 1575
+    },
+    {
+      "epoch": 6.3453815261044175,
+      "grad_norm": 0.49878165125846863,
+      "learning_rate": 1.7762798571788707e-05,
+      "loss": 0.3948,
+      "num_input_tokens_seen": 277456,
+      "step": 1580
+    },
+    {
+      "epoch": 6.365461847389558,
+      "grad_norm": 0.415039598941803,
+      "learning_rate": 1.759524709854626e-05,
+      "loss": 0.3246,
+      "num_input_tokens_seen": 278352,
+      "step": 1585
+    },
+    {
+      "epoch": 6.385542168674699,
+      "grad_norm": 0.11119506508111954,
+      "learning_rate": 1.742805942709538e-05,
+      "loss": 0.3468,
+      "num_input_tokens_seen": 279264,
+      "step": 1590
+    },
+    {
+      "epoch": 6.405622489959839,
+      "grad_norm": 0.10621926933526993,
+      "learning_rate": 1.7261243771508208e-05,
+      "loss": 0.3428,
+      "num_input_tokens_seen": 280144,
+      "step": 1595
+    },
+    {
+      "epoch": 6.42570281124498,
+      "grad_norm": 0.10251349955797195,
+      "learning_rate": 1.70948083275794e-05,
+      "loss": 0.3439,
+      "num_input_tokens_seen": 281008,
+      "step": 1600
+    },
+    {
+      "epoch": 6.445783132530121,
+      "grad_norm": 0.4213389456272125,
+      "learning_rate": 1.6928761272423522e-05,
+      "loss": 0.3717,
+      "num_input_tokens_seen": 281792,
+      "step": 1605
+    },
+    {
+      "epoch": 6.4658634538152615,
+      "grad_norm": 0.08103923499584198,
+      "learning_rate": 1.6763110764073235e-05,
+      "loss": 0.3517,
+      "num_input_tokens_seen": 282560,
+      "step": 1610
+    },
+    {
+      "epoch": 6.485943775100401,
+      "grad_norm": 0.46979421377182007,
+      "learning_rate": 1.6597864941078552e-05,
+      "loss": 0.3423,
+      "num_input_tokens_seen": 283440,
+      "step": 1615
+    },
+    {
+      "epoch": 6.506024096385542,
+      "grad_norm": 0.12765252590179443,
+      "learning_rate": 1.643303192210693e-05,
+      "loss": 0.358,
+      "num_input_tokens_seen": 284592,
+      "step": 1620
+    },
+    {
+      "epoch": 6.526104417670683,
+      "grad_norm": 0.06467333436012268,
+      "learning_rate": 1.626861980554441e-05,
+      "loss": 0.3457,
+      "num_input_tokens_seen": 285424,
+      "step": 1625
+    },
+    {
+      "epoch": 6.526104417670683,
+      "eval_loss": 0.3553968071937561,
+      "eval_runtime": 1.2158,
+      "eval_samples_per_second": 46.06,
+      "eval_steps_per_second": 23.03,
+      "num_input_tokens_seen": 285424,
+      "step": 1625
+    },
+    {
+      "epoch": 6.546184738955823,
+      "grad_norm": 0.41862520575523376,
+      "learning_rate": 1.6104636669097776e-05,
+      "loss": 0.3518,
+      "num_input_tokens_seen": 286272,
+      "step": 1630
+    },
+    {
+      "epoch": 6.566265060240964,
+      "grad_norm": 0.0668744370341301,
+      "learning_rate": 1.5941090569397616e-05,
+      "loss": 0.3512,
+      "num_input_tokens_seen": 287200,
+      "step": 1635
+    },
+    {
+      "epoch": 6.586345381526105,
+      "grad_norm": 0.08563435077667236,
+      "learning_rate": 1.5777989541602533e-05,
+      "loss": 0.348,
+      "num_input_tokens_seen": 288224,
+      "step": 1640
+    },
+    {
+      "epoch": 6.606425702811245,
+      "grad_norm": 0.44376978278160095,
+      "learning_rate": 1.561534159900441e-05,
+      "loss": 0.3353,
+      "num_input_tokens_seen": 289216,
+      "step": 1645
+    },
+    {
+      "epoch": 6.626506024096385,
+      "grad_norm": 0.12750263512134552,
+      "learning_rate": 1.5453154732634616e-05,
+      "loss": 0.3476,
+      "num_input_tokens_seen": 290080,
+      "step": 1650
+    },
+    {
+      "epoch": 6.646586345381526,
+      "grad_norm": 0.4618399739265442,
+      "learning_rate": 1.52914369108715e-05,
+      "loss": 0.351,
+      "num_input_tokens_seen": 290880,
+      "step": 1655
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 0.07694806158542633,
+      "learning_rate": 1.513019607904882e-05,
+      "loss": 0.3607,
+      "num_input_tokens_seen": 291728,
+      "step": 1660
+    },
+    {
+      "epoch": 6.686746987951807,
+      "grad_norm": 0.09036281704902649,
+      "learning_rate": 1.4969440159065439e-05,
+      "loss": 0.3573,
+      "num_input_tokens_seen": 292624,
+      "step": 1665
+    },
+    {
+      "epoch": 6.706827309236948,
+      "grad_norm": 0.08298249542713165,
+      "learning_rate": 1.4809177048996064e-05,
+      "loss": 0.3476,
+      "num_input_tokens_seen": 293488,
+      "step": 1670
+    },
+    {
+      "epoch": 6.7269076305220885,
+      "grad_norm": 0.4784527122974396,
+      "learning_rate": 1.464941462270325e-05,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 294400,
+      "step": 1675
+    },
+    {
+      "epoch": 6.746987951807229,
+      "grad_norm": 0.45142361521720886,
+      "learning_rate": 1.449016072945053e-05,
+      "loss": 0.357,
+      "num_input_tokens_seen": 295184,
+      "step": 1680
+    },
+    {
+      "epoch": 6.767068273092369,
+      "grad_norm": 0.4861927330493927,
+      "learning_rate": 1.4331423193516768e-05,
+      "loss": 0.3575,
+      "num_input_tokens_seen": 296176,
+      "step": 1685
+    },
+    {
+      "epoch": 6.78714859437751,
+      "grad_norm": 0.41130194067955017,
+      "learning_rate": 1.4173209813811788e-05,
+      "loss": 0.3358,
+      "num_input_tokens_seen": 297072,
+      "step": 1690
+    },
+    {
+      "epoch": 6.807228915662651,
+      "grad_norm": 0.40937137603759766,
+      "learning_rate": 1.4015528363493125e-05,
+      "loss": 0.3491,
+      "num_input_tokens_seen": 297856,
+      "step": 1695
+    },
+    {
+      "epoch": 6.827309236947791,
+      "grad_norm": 0.40058666467666626,
+      "learning_rate": 1.3858386589584187e-05,
+      "loss": 0.3253,
+      "num_input_tokens_seen": 298896,
+      "step": 1700
+    },
+    {
+      "epoch": 6.847389558232932,
+      "grad_norm": 0.11114007234573364,
+      "learning_rate": 1.3701792212593662e-05,
+      "loss": 0.3302,
+      "num_input_tokens_seen": 299712,
+      "step": 1705
+    },
+    {
+      "epoch": 6.867469879518072,
+      "grad_norm": 0.11977384239435196,
+      "learning_rate": 1.354575292613611e-05,
+      "loss": 0.3882,
+      "num_input_tokens_seen": 300720,
+      "step": 1710
+    },
+    {
+      "epoch": 6.887550200803213,
+      "grad_norm": 0.5076762437820435,
+      "learning_rate": 1.3390276396554052e-05,
+      "loss": 0.3658,
+      "num_input_tokens_seen": 301552,
+      "step": 1715
+    },
+    {
+      "epoch": 6.907630522088353,
+      "grad_norm": 0.07989758253097534,
+      "learning_rate": 1.3235370262541272e-05,
+      "loss": 0.3388,
+      "num_input_tokens_seen": 302352,
+      "step": 1720
+    },
+    {
+      "epoch": 6.927710843373494,
+      "grad_norm": 0.38454264402389526,
+      "learning_rate": 1.3081042134767554e-05,
+      "loss": 0.3335,
+      "num_input_tokens_seen": 303232,
+      "step": 1725
+    },
+    {
+      "epoch": 6.947791164658635,
+      "grad_norm": 0.07989054918289185,
+      "learning_rate": 1.292729959550473e-05,
+      "loss": 0.3262,
+      "num_input_tokens_seen": 304016,
+      "step": 1730
+    },
+    {
+      "epoch": 6.967871485943775,
+      "grad_norm": 0.3905481994152069,
+      "learning_rate": 1.277415019825417e-05,
+      "loss": 0.3396,
+      "num_input_tokens_seen": 304944,
+      "step": 1735
+    },
+    {
+      "epoch": 6.9879518072289155,
+      "grad_norm": 0.39736074209213257,
+      "learning_rate": 1.2621601467375684e-05,
+      "loss": 0.3422,
+      "num_input_tokens_seen": 305984,
+      "step": 1740
+    },
+    {
+      "epoch": 7.008032128514056,
+      "grad_norm": 0.1431533843278885,
+      "learning_rate": 1.2469660897717816e-05,
+      "loss": 0.3182,
+      "num_input_tokens_seen": 306992,
+      "step": 1745
+    },
+    {
+      "epoch": 7.028112449799197,
+      "grad_norm": 0.39490222930908203,
+      "learning_rate": 1.2318335954249669e-05,
+      "loss": 0.315,
+      "num_input_tokens_seen": 307792,
+      "step": 1750
+    },
+    {
+      "epoch": 7.028112449799197,
+      "eval_loss": 0.36506387591362,
+      "eval_runtime": 1.2174,
+      "eval_samples_per_second": 45.998,
+      "eval_steps_per_second": 22.999,
+      "num_input_tokens_seen": 307792,
+      "step": 1750
+    },
+    {
+      "epoch": 7.048192771084337,
+      "grad_norm": 0.13177447021007538,
+      "learning_rate": 1.2167634071694081e-05,
+      "loss": 0.3174,
+      "num_input_tokens_seen": 308624,
+      "step": 1755
+    },
+    {
+      "epoch": 7.068273092369478,
+      "grad_norm": 0.38232582807540894,
+      "learning_rate": 1.2017562654162357e-05,
+      "loss": 0.2887,
+      "num_input_tokens_seen": 309680,
+      "step": 1760
+    },
+    {
+      "epoch": 7.088353413654619,
+      "grad_norm": 0.19812007248401642,
+      "learning_rate": 1.1868129074790577e-05,
+      "loss": 0.3394,
+      "num_input_tokens_seen": 310544,
+      "step": 1765
+    },
+    {
+      "epoch": 7.108433734939759,
+      "grad_norm": 0.2020581066608429,
+      "learning_rate": 1.1719340675377252e-05,
+      "loss": 0.3113,
+      "num_input_tokens_seen": 311568,
+      "step": 1770
+    },
+    {
+      "epoch": 7.128514056224899,
+      "grad_norm": 0.16722743213176727,
+      "learning_rate": 1.1571204766022665e-05,
+      "loss": 0.4907,
+      "num_input_tokens_seen": 312432,
+      "step": 1775
+    },
+    {
+      "epoch": 7.14859437751004,
+      "grad_norm": 0.14364704489707947,
+      "learning_rate": 1.1423728624769695e-05,
+      "loss": 0.3627,
+      "num_input_tokens_seen": 313168,
+      "step": 1780
+    },
+    {
+      "epoch": 7.168674698795181,
+      "grad_norm": 0.11236248165369034,
+      "learning_rate": 1.1276919497246288e-05,
+      "loss": 0.3648,
+      "num_input_tokens_seen": 313968,
+      "step": 1785
+    },
+    {
+      "epoch": 7.188755020080321,
+      "grad_norm": 0.11342489719390869,
+      "learning_rate": 1.1130784596309409e-05,
+      "loss": 0.3585,
+      "num_input_tokens_seen": 314736,
+      "step": 1790
+    },
+    {
+      "epoch": 7.208835341365462,
+      "grad_norm": 0.09296204149723053,
+      "learning_rate": 1.098533110169071e-05,
+      "loss": 0.3485,
+      "num_input_tokens_seen": 315664,
+      "step": 1795
+    },
+    {
+      "epoch": 7.228915662650603,
+      "grad_norm": 0.1172434464097023,
+      "learning_rate": 1.084056615964377e-05,
+      "loss": 0.3442,
+      "num_input_tokens_seen": 316704,
+      "step": 1800
+    },
+    {
+      "epoch": 7.2489959839357425,
+      "grad_norm": 0.0936344638466835,
+      "learning_rate": 1.069649688259299e-05,
+      "loss": 0.388,
+      "num_input_tokens_seen": 317520,
+      "step": 1805
+    },
+    {
+      "epoch": 7.269076305220883,
+      "grad_norm": 0.4171283543109894,
+      "learning_rate": 1.0553130348784182e-05,
+      "loss": 0.3306,
+      "num_input_tokens_seen": 318496,
+      "step": 1810
+    },
+    {
+      "epoch": 7.289156626506024,
+      "grad_norm": 0.403551310300827,
+      "learning_rate": 1.0410473601936765e-05,
+      "loss": 0.3181,
+      "num_input_tokens_seen": 319344,
+      "step": 1815
+    },
+    {
+      "epoch": 7.309236947791165,
+      "grad_norm": 0.08753710985183716,
+      "learning_rate": 1.026853365089773e-05,
+      "loss": 0.3494,
+      "num_input_tokens_seen": 320224,
+      "step": 1820
+    },
+    {
+      "epoch": 7.329317269076305,
+      "grad_norm": 0.393200546503067,
+      "learning_rate": 1.0127317469297277e-05,
+      "loss": 0.3193,
+      "num_input_tokens_seen": 320976,
+      "step": 1825
+    },
+    {
+      "epoch": 7.349397590361446,
+      "grad_norm": 0.1304997056722641,
+      "learning_rate": 9.986831995206195e-06,
+      "loss": 0.3271,
+      "num_input_tokens_seen": 321808,
+      "step": 1830
+    },
+    {
+      "epoch": 7.3694779116465865,
+      "grad_norm": 0.37172961235046387,
+      "learning_rate": 9.847084130795028e-06,
+      "loss": 0.3504,
+      "num_input_tokens_seen": 322624,
+      "step": 1835
+    },
+    {
+      "epoch": 7.389558232931727,
+      "grad_norm": 0.3792823255062103,
+      "learning_rate": 9.708080741994868e-06,
+      "loss": 0.3165,
+      "num_input_tokens_seen": 323696,
+      "step": 1840
+    },
+    {
+      "epoch": 7.409638554216867,
+      "grad_norm": 0.365556925535202,
+      "learning_rate": 9.569828658160158e-06,
+      "loss": 0.3184,
+      "num_input_tokens_seen": 324496,
+      "step": 1845
+    },
+    {
+      "epoch": 7.429718875502008,
+      "grad_norm": 0.10406666994094849,
+      "learning_rate": 9.432334671733039e-06,
+      "loss": 0.3824,
+      "num_input_tokens_seen": 325328,
+      "step": 1850
+    },
+    {
+      "epoch": 7.449799196787149,
+      "grad_norm": 0.3579881191253662,
+      "learning_rate": 9.295605537909708e-06,
+      "loss": 0.3336,
+      "num_input_tokens_seen": 326304,
+      "step": 1855
+    },
+    {
+      "epoch": 7.469879518072289,
+      "grad_norm": 0.171662375330925,
+      "learning_rate": 9.159647974308494e-06,
+      "loss": 0.3148,
+      "num_input_tokens_seen": 327120,
+      "step": 1860
+    },
+    {
+      "epoch": 7.48995983935743,
+      "grad_norm": 0.5380930304527283,
+      "learning_rate": 9.024468660639826e-06,
+      "loss": 0.3811,
+      "num_input_tokens_seen": 328128,
+      "step": 1865
+    },
+    {
+      "epoch": 7.51004016064257,
+      "grad_norm": 0.11611666530370712,
+      "learning_rate": 8.890074238378074e-06,
+      "loss": 0.351,
+      "num_input_tokens_seen": 329008,
+      "step": 1870
+    },
+    {
+      "epoch": 7.530120481927711,
+      "grad_norm": 0.10266629606485367,
+      "learning_rate": 8.756471310435204e-06,
+      "loss": 0.3149,
+      "num_input_tokens_seen": 329840,
+      "step": 1875
+    },
+    {
+      "epoch": 7.530120481927711,
+      "eval_loss": 0.3625907897949219,
+      "eval_runtime": 1.2148,
+      "eval_samples_per_second": 46.1,
+      "eval_steps_per_second": 23.05,
+      "num_input_tokens_seen": 329840,
+      "step": 1875
+    },
+    {
+      "epoch": 7.550200803212851,
+      "grad_norm": 0.15210554003715515,
+      "learning_rate": 8.623666440836404e-06,
+      "loss": 0.3623,
+      "num_input_tokens_seen": 330624,
+      "step": 1880
+    },
+    {
+      "epoch": 7.570281124497992,
+      "grad_norm": 0.09614621847867966,
+      "learning_rate": 8.491666154397573e-06,
+      "loss": 0.3149,
+      "num_input_tokens_seen": 331440,
+      "step": 1885
+    },
+    {
+      "epoch": 7.590361445783133,
+      "grad_norm": 0.11491771787405014,
+      "learning_rate": 8.360476936404754e-06,
+      "loss": 0.3897,
+      "num_input_tokens_seen": 332192,
+      "step": 1890
+    },
+    {
+      "epoch": 7.610441767068274,
+      "grad_norm": 0.09532292187213898,
+      "learning_rate": 8.230105232295538e-06,
+      "loss": 0.3736,
+      "num_input_tokens_seen": 333168,
+      "step": 1895
+    },
+    {
+      "epoch": 7.6305220883534135,
+      "grad_norm": 0.5067830681800842,
+      "learning_rate": 8.100557447342327e-06,
+      "loss": 0.3618,
+      "num_input_tokens_seen": 334080,
+      "step": 1900
+    },
+    {
+      "epoch": 7.650602409638554,
+      "grad_norm": 0.13660845160484314,
+      "learning_rate": 7.971839946337698e-06,
+      "loss": 0.3533,
+      "num_input_tokens_seen": 335040,
+      "step": 1905
+    },
+    {
+      "epoch": 7.670682730923695,
+      "grad_norm": 0.41254377365112305,
+      "learning_rate": 7.843959053281663e-06,
+      "loss": 0.3532,
+      "num_input_tokens_seen": 335824,
+      "step": 1910
+    },
+    {
+      "epoch": 7.690763052208835,
+      "grad_norm": 0.07652189582586288,
+      "learning_rate": 7.71692105107098e-06,
+      "loss": 0.3362,
+      "num_input_tokens_seen": 336656,
+      "step": 1915
+    },
+    {
+      "epoch": 7.710843373493976,
+      "grad_norm": 0.0759321078658104,
+      "learning_rate": 7.590732181190482e-06,
+      "loss": 0.3608,
+      "num_input_tokens_seen": 337488,
+      "step": 1920
+    },
+    {
+      "epoch": 7.730923694779117,
+      "grad_norm": 0.0837617963552475,
+      "learning_rate": 7.465398643406366e-06,
+      "loss": 0.342,
+      "num_input_tokens_seen": 338400,
+      "step": 1925
+    },
+    {
+      "epoch": 7.7510040160642575,
+      "grad_norm": 0.09123997390270233,
+      "learning_rate": 7.340926595461687e-06,
+      "loss": 0.3573,
+      "num_input_tokens_seen": 339248,
+      "step": 1930
+    },
+    {
+      "epoch": 7.771084337349397,
+      "grad_norm": 0.09809573739767075,
+      "learning_rate": 7.217322152773742e-06,
+      "loss": 0.3539,
+      "num_input_tokens_seen": 340112,
+      "step": 1935
+    },
+    {
+      "epoch": 7.791164658634538,
+      "grad_norm": 0.4673003554344177,
+      "learning_rate": 7.094591388133659e-06,
+      "loss": 0.3471,
+      "num_input_tokens_seen": 340896,
+      "step": 1940
+    },
+    {
+      "epoch": 7.811244979919679,
+      "grad_norm": 0.46651169657707214,
+      "learning_rate": 6.972740331408015e-06,
+      "loss": 0.3599,
+      "num_input_tokens_seen": 341760,
+      "step": 1945
+    },
+    {
+      "epoch": 7.831325301204819,
+      "grad_norm": 0.09857732057571411,
+      "learning_rate": 6.851774969242589e-06,
+      "loss": 0.3382,
+      "num_input_tokens_seen": 342608,
+      "step": 1950
+    },
+    {
+      "epoch": 7.85140562248996,
+      "grad_norm": 0.14156574010849,
+      "learning_rate": 6.731701244768254e-06,
+      "loss": 0.338,
+      "num_input_tokens_seen": 343632,
+      "step": 1955
+    },
+    {
+      "epoch": 7.871485943775101,
+      "grad_norm": 0.08120275288820267,
+      "learning_rate": 6.612525057308949e-06,
+      "loss": 0.3473,
+      "num_input_tokens_seen": 344528,
+      "step": 1960
+    },
+    {
+      "epoch": 7.891566265060241,
+      "grad_norm": 0.4519568085670471,
+      "learning_rate": 6.494252262091857e-06,
+      "loss": 0.3505,
+      "num_input_tokens_seen": 345568,
+      "step": 1965
+    },
+    {
+      "epoch": 7.911646586345381,
+      "grad_norm": 0.10098995268344879,
+      "learning_rate": 6.3768886699597436e-06,
+      "loss": 0.3443,
+      "num_input_tokens_seen": 346496,
+      "step": 1970
+    },
+    {
+      "epoch": 7.931726907630522,
+      "grad_norm": 0.44386422634124756,
+      "learning_rate": 6.260440047085439e-06,
+      "loss": 0.3473,
+      "num_input_tokens_seen": 347360,
+      "step": 1975
+    },
+    {
+      "epoch": 7.951807228915663,
+      "grad_norm": 0.06753702461719513,
+      "learning_rate": 6.1449121146885894e-06,
+      "loss": 0.3445,
+      "num_input_tokens_seen": 348128,
+      "step": 1980
+    },
+    {
+      "epoch": 7.971887550200803,
+      "grad_norm": 0.42976370453834534,
+      "learning_rate": 6.030310548754506e-06,
+      "loss": 0.3509,
+      "num_input_tokens_seen": 348960,
+      "step": 1985
+    },
+    {
+      "epoch": 7.991967871485944,
+      "grad_norm": 0.4281879961490631,
+      "learning_rate": 5.9166409797553415e-06,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 349856,
+      "step": 1990
+    },
+    {
+      "epoch": 8.012048192771084,
+      "grad_norm": 0.4621274471282959,
+      "learning_rate": 5.803908992373449e-06,
+      "loss": 0.338,
+      "num_input_tokens_seen": 350784,
+      "step": 1995
+    },
+    {
+      "epoch": 8.032128514056225,
+      "grad_norm": 0.08658602088689804,
+      "learning_rate": 5.692120125226993e-06,
+      "loss": 0.3441,
+      "num_input_tokens_seen": 351552,
+      "step": 2000
+    },
+    {
+      "epoch": 8.032128514056225,
+      "eval_loss": 0.3484961986541748,
+      "eval_runtime": 1.2167,
+      "eval_samples_per_second": 46.027,
+      "eval_steps_per_second": 23.014,
+      "num_input_tokens_seen": 351552,
+      "step": 2000
+    },
+    {
+      "epoch": 8.052208835341366,
+      "grad_norm": 0.0772676169872284,
+      "learning_rate": 5.581279870597867e-06,
+      "loss": 0.3537,
+      "num_input_tokens_seen": 352592,
+      "step": 2005
+    },
+    {
+      "epoch": 8.072289156626505,
+      "grad_norm": 0.08750884979963303,
+      "learning_rate": 5.4713936741617845e-06,
+      "loss": 0.3441,
+      "num_input_tokens_seen": 353392,
+      "step": 2010
+    },
+    {
+      "epoch": 8.092369477911646,
+      "grad_norm": 0.4653143882751465,
+      "learning_rate": 5.3624669347208085e-06,
+      "loss": 0.3473,
+      "num_input_tokens_seen": 354176,
+      "step": 2015
+    },
+    {
+      "epoch": 8.112449799196787,
+      "grad_norm": 0.07544849812984467,
+      "learning_rate": 5.254505003938043e-06,
+      "loss": 0.335,
+      "num_input_tokens_seen": 355040,
+      "step": 2020
+    },
+    {
+      "epoch": 8.132530120481928,
+      "grad_norm": 0.4536716043949127,
+      "learning_rate": 5.147513186074751e-06,
+      "loss": 0.3445,
+      "num_input_tokens_seen": 355984,
+      "step": 2025
+    },
+    {
+      "epoch": 8.152610441767068,
+      "grad_norm": 0.06866519898176193,
+      "learning_rate": 5.041496737729687e-06,
+      "loss": 0.3443,
+      "num_input_tokens_seen": 356768,
+      "step": 2030
+    },
+    {
+      "epoch": 8.17269076305221,
+      "grad_norm": 0.07859829813241959,
+      "learning_rate": 4.936460867580889e-06,
+      "loss": 0.3604,
+      "num_input_tokens_seen": 357648,
+      "step": 2035
+    },
+    {
+      "epoch": 8.19277108433735,
+      "grad_norm": 0.48313626646995544,
+      "learning_rate": 4.832410736129778e-06,
+      "loss": 0.3509,
+      "num_input_tokens_seen": 358464,
+      "step": 2040
+    },
+    {
+      "epoch": 8.21285140562249,
+      "grad_norm": 0.4338780343532562,
+      "learning_rate": 4.729351455447573e-06,
+      "loss": 0.3421,
+      "num_input_tokens_seen": 359520,
+      "step": 2045
+    },
+    {
+      "epoch": 8.23293172690763,
+      "grad_norm": 0.4279603660106659,
+      "learning_rate": 4.627288088924156e-06,
+      "loss": 0.3447,
+      "num_input_tokens_seen": 360256,
+      "step": 2050
+    },
+    {
+      "epoch": 8.25301204819277,
+      "grad_norm": 0.09719569236040115,
+      "learning_rate": 4.526225651019309e-06,
+      "loss": 0.3479,
+      "num_input_tokens_seen": 361184,
+      "step": 2055
+    },
+    {
+      "epoch": 8.273092369477911,
+      "grad_norm": 0.11000480502843857,
+      "learning_rate": 4.4261691070163316e-06,
+      "loss": 0.3447,
+      "num_input_tokens_seen": 362064,
+      "step": 2060
+    },
+    {
+      "epoch": 8.293172690763052,
+      "grad_norm": 0.11856409162282944,
+      "learning_rate": 4.327123372778122e-06,
+      "loss": 0.3415,
+      "num_input_tokens_seen": 362928,
+      "step": 2065
+    },
+    {
+      "epoch": 8.313253012048193,
+      "grad_norm": 0.06177099421620369,
+      "learning_rate": 4.229093314505619e-06,
+      "loss": 0.336,
+      "num_input_tokens_seen": 363888,
+      "step": 2070
+    },
+    {
+      "epoch": 8.333333333333334,
+      "grad_norm": 0.07357407361268997,
+      "learning_rate": 4.132083748498744e-06,
+      "loss": 0.3572,
+      "num_input_tokens_seen": 364800,
+      "step": 2075
+    },
+    {
+      "epoch": 8.353413654618475,
+      "grad_norm": 0.1254434585571289,
+      "learning_rate": 4.036099440919763e-06,
+      "loss": 0.3479,
+      "num_input_tokens_seen": 365680,
+      "step": 2080
+    },
+    {
+      "epoch": 8.373493975903614,
+      "grad_norm": 0.12399672716856003,
+      "learning_rate": 3.9411451075591464e-06,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 366560,
+      "step": 2085
+    },
+    {
+      "epoch": 8.393574297188755,
+      "grad_norm": 0.4643438458442688,
+      "learning_rate": 3.847225413603839e-06,
+      "loss": 0.3449,
+      "num_input_tokens_seen": 367424,
+      "step": 2090
+    },
+    {
+      "epoch": 8.413654618473895,
+      "grad_norm": 0.09002748876810074,
+      "learning_rate": 3.754344973408064e-06,
+      "loss": 0.3538,
+      "num_input_tokens_seen": 368272,
+      "step": 2095
+    },
+    {
+      "epoch": 8.433734939759036,
+      "grad_norm": 0.06885527074337006,
+      "learning_rate": 3.6625083502666554e-06,
+      "loss": 0.3472,
+      "num_input_tokens_seen": 369040,
+      "step": 2100
+    },
+    {
+      "epoch": 8.453815261044177,
+      "grad_norm": 0.4609389007091522,
+      "learning_rate": 3.5717200561908026e-06,
+      "loss": 0.3411,
+      "num_input_tokens_seen": 369808,
+      "step": 2105
+    },
+    {
+      "epoch": 8.473895582329318,
+      "grad_norm": 0.09589619934558868,
+      "learning_rate": 3.481984551686429e-06,
+      "loss": 0.3383,
+      "num_input_tokens_seen": 370672,
+      "step": 2110
+    },
+    {
+      "epoch": 8.493975903614459,
+      "grad_norm": 0.11223804205656052,
+      "learning_rate": 3.3933062455349744e-06,
+      "loss": 0.3417,
+      "num_input_tokens_seen": 371520,
+      "step": 2115
+    },
+    {
+      "epoch": 8.514056224899598,
+      "grad_norm": 0.47606217861175537,
+      "learning_rate": 3.305689494576847e-06,
+      "loss": 0.36,
+      "num_input_tokens_seen": 372368,
+      "step": 2120
+    },
+    {
+      "epoch": 8.534136546184738,
+      "grad_norm": 0.5172960162162781,
+      "learning_rate": 3.2191386034973627e-06,
+      "loss": 0.3574,
+      "num_input_tokens_seen": 373424,
+      "step": 2125
+    },
+    {
+      "epoch": 8.534136546184738,
+      "eval_loss": 0.3515996038913727,
+      "eval_runtime": 1.2192,
+      "eval_samples_per_second": 45.93,
+      "eval_steps_per_second": 22.965,
+      "num_input_tokens_seen": 373424,
+      "step": 2125
+    },
+    {
+      "epoch": 8.55421686746988,
+      "grad_norm": 0.4503559172153473,
+      "learning_rate": 3.1336578246152103e-06,
+      "loss": 0.3443,
+      "num_input_tokens_seen": 374240,
+      "step": 2130
+    },
+    {
+      "epoch": 8.57429718875502,
+      "grad_norm": 0.11173038929700851,
+      "learning_rate": 3.049251357673577e-06,
+      "loss": 0.3383,
+      "num_input_tokens_seen": 375104,
+      "step": 2135
+    },
+    {
+      "epoch": 8.594377510040161,
+      "grad_norm": 0.09678234905004501,
+      "learning_rate": 2.9659233496337786e-06,
+      "loss": 0.3476,
+      "num_input_tokens_seen": 376080,
+      "step": 2140
+    },
+    {
+      "epoch": 8.614457831325302,
+      "grad_norm": 0.08732342720031738,
+      "learning_rate": 2.8836778944715454e-06,
+      "loss": 0.3415,
+      "num_input_tokens_seen": 376928,
+      "step": 2145
+    },
+    {
+      "epoch": 8.634538152610443,
+      "grad_norm": 0.12217875570058823,
+      "learning_rate": 2.802519032975859e-06,
+      "loss": 0.351,
+      "num_input_tokens_seen": 377856,
+      "step": 2150
+    },
+    {
+      "epoch": 8.654618473895582,
+      "grad_norm": 0.4865402579307556,
+      "learning_rate": 2.722450752550429e-06,
+      "loss": 0.3417,
+      "num_input_tokens_seen": 378784,
+      "step": 2155
+    },
+    {
+      "epoch": 8.674698795180722,
+      "grad_norm": 0.4950112998485565,
+      "learning_rate": 2.6434769870177985e-06,
+      "loss": 0.3604,
+      "num_input_tokens_seen": 379696,
+      "step": 2160
+    },
+    {
+      "epoch": 8.694779116465863,
+      "grad_norm": 0.07212843000888824,
+      "learning_rate": 2.5656016164260554e-06,
+      "loss": 0.3447,
+      "num_input_tokens_seen": 380512,
+      "step": 2165
+    },
+    {
+      "epoch": 8.714859437751004,
+      "grad_norm": 0.09282089024782181,
+      "learning_rate": 2.4888284668582285e-06,
+      "loss": 0.3445,
+      "num_input_tokens_seen": 381520,
+      "step": 2170
+    },
+    {
+      "epoch": 8.734939759036145,
+      "grad_norm": 0.08465081453323364,
+      "learning_rate": 2.4131613102442857e-06,
+      "loss": 0.3354,
+      "num_input_tokens_seen": 382480,
+      "step": 2175
+    },
+    {
+      "epoch": 8.755020080321286,
+      "grad_norm": 0.45736274123191833,
+      "learning_rate": 2.3386038641758063e-06,
+      "loss": 0.3383,
+      "num_input_tokens_seen": 383440,
+      "step": 2180
+    },
+    {
+      "epoch": 8.775100401606426,
+      "grad_norm": 0.46960991621017456,
+      "learning_rate": 2.265159791723373e-06,
+      "loss": 0.3508,
+      "num_input_tokens_seen": 384240,
+      "step": 2185
+    },
+    {
+      "epoch": 8.795180722891565,
+      "grad_norm": 0.4292510449886322,
+      "learning_rate": 2.1928327012565696e-06,
+      "loss": 0.3483,
+      "num_input_tokens_seen": 385120,
+      "step": 2190
+    },
+    {
+      "epoch": 8.815261044176706,
+      "grad_norm": 0.07858365774154663,
+      "learning_rate": 2.121626146266706e-06,
+      "loss": 0.3546,
+      "num_input_tokens_seen": 385984,
+      "step": 2195
+    },
+    {
+      "epoch": 8.835341365461847,
+      "grad_norm": 0.5046152472496033,
+      "learning_rate": 2.051543625192226e-06,
+      "loss": 0.3609,
+      "num_input_tokens_seen": 386896,
+      "step": 2200
+    },
+    {
+      "epoch": 8.855421686746988,
+      "grad_norm": 0.43499457836151123,
+      "learning_rate": 1.9825885812468524e-06,
+      "loss": 0.33,
+      "num_input_tokens_seen": 387776,
+      "step": 2205
+    },
+    {
+      "epoch": 8.875502008032129,
+      "grad_norm": 0.4260156750679016,
+      "learning_rate": 1.914764402250385e-06,
+      "loss": 0.3487,
+      "num_input_tokens_seen": 388704,
+      "step": 2210
+    },
+    {
+      "epoch": 8.89558232931727,
+      "grad_norm": 0.43153613805770874,
+      "learning_rate": 1.8480744204622757e-06,
+      "loss": 0.3512,
+      "num_input_tokens_seen": 389456,
+      "step": 2215
+    },
+    {
+      "epoch": 8.91566265060241,
+      "grad_norm": 0.4854128956794739,
+      "learning_rate": 1.7825219124179004e-06,
+      "loss": 0.3522,
+      "num_input_tokens_seen": 390304,
+      "step": 2220
+    },
+    {
+      "epoch": 8.93574297188755,
+      "grad_norm": 0.08903124928474426,
+      "learning_rate": 1.7181100987675862e-06,
+      "loss": 0.3356,
+      "num_input_tokens_seen": 391104,
+      "step": 2225
+    },
+    {
+      "epoch": 8.95582329317269,
+      "grad_norm": 0.49092555046081543,
+      "learning_rate": 1.6548421441183875e-06,
+      "loss": 0.3516,
+      "num_input_tokens_seen": 392112,
+      "step": 2230
+    },
+    {
+      "epoch": 8.975903614457831,
+      "grad_norm": 0.08050279319286346,
+      "learning_rate": 1.5927211568785878e-06,
+      "loss": 0.3449,
+      "num_input_tokens_seen": 392880,
+      "step": 2235
+    },
+    {
+      "epoch": 8.995983935742972,
+      "grad_norm": 0.4535515308380127,
+      "learning_rate": 1.5317501891049719e-06,
+      "loss": 0.3302,
+      "num_input_tokens_seen": 393728,
+      "step": 2240
+    },
+    {
+      "epoch": 9.016064257028113,
+      "grad_norm": 0.48401492834091187,
+      "learning_rate": 1.4719322363529242e-06,
+      "loss": 0.3487,
+      "num_input_tokens_seen": 394688,
+      "step": 2245
+    },
+    {
+      "epoch": 9.036144578313253,
+      "grad_norm": 0.4877987205982208,
+      "learning_rate": 1.4132702375291989e-06,
+      "loss": 0.3673,
+      "num_input_tokens_seen": 395616,
+      "step": 2250
+    },
+    {
+      "epoch": 9.036144578313253,
+      "eval_loss": 0.35450634360313416,
+      "eval_runtime": 1.2106,
+      "eval_samples_per_second": 46.258,
+      "eval_steps_per_second": 23.129,
+      "num_input_tokens_seen": 395616,
+      "step": 2250
+    },
+    {
+      "epoch": 9.056224899598394,
+      "grad_norm": 0.4849400520324707,
+      "learning_rate": 1.3557670747475714e-06,
+      "loss": 0.3455,
+      "num_input_tokens_seen": 396560,
+      "step": 2255
+    },
+    {
+      "epoch": 9.076305220883533,
+      "grad_norm": 0.4335307478904724,
+      "learning_rate": 1.2994255731871963e-06,
+      "loss": 0.3489,
+      "num_input_tokens_seen": 397456,
+      "step": 2260
+    },
+    {
+      "epoch": 9.096385542168674,
+      "grad_norm": 0.10791157931089401,
+      "learning_rate": 1.244248500953854e-06,
+      "loss": 0.3648,
+      "num_input_tokens_seen": 398448,
+      "step": 2265
+    },
+    {
+      "epoch": 9.116465863453815,
+      "grad_norm": 0.4895195960998535,
+      "learning_rate": 1.1902385689439022e-06,
+      "loss": 0.3544,
+      "num_input_tokens_seen": 399248,
+      "step": 2270
+    },
+    {
+      "epoch": 9.136546184738956,
+      "grad_norm": 0.45951637625694275,
+      "learning_rate": 1.137398430711123e-06,
+      "loss": 0.3574,
+      "num_input_tokens_seen": 400096,
+      "step": 2275
+    },
+    {
+      "epoch": 9.156626506024097,
+      "grad_norm": 0.44542670249938965,
+      "learning_rate": 1.085730682336325e-06,
+      "loss": 0.3477,
+      "num_input_tokens_seen": 401024,
+      "step": 2280
+    },
+    {
+      "epoch": 9.176706827309237,
+      "grad_norm": 0.10208380967378616,
+      "learning_rate": 1.0352378622998204e-06,
+      "loss": 0.3506,
+      "num_input_tokens_seen": 401856,
+      "step": 2285
+    },
+    {
+      "epoch": 9.196787148594378,
+      "grad_norm": 0.43247994780540466,
+      "learning_rate": 9.85922451356694e-07,
+      "loss": 0.3388,
+      "num_input_tokens_seen": 402736,
+      "step": 2290
+    },
+    {
+      "epoch": 9.216867469879517,
+      "grad_norm": 0.13525407016277313,
+      "learning_rate": 9.377868724149197e-07,
+      "loss": 0.3413,
+      "num_input_tokens_seen": 403696,
+      "step": 2295
+    },
+    {
+      "epoch": 9.236947791164658,
+      "grad_norm": 0.4610027074813843,
+      "learning_rate": 8.908334904163207e-07,
+      "loss": 0.3445,
+      "num_input_tokens_seen": 404480,
+      "step": 2300
+    },
+    {
+      "epoch": 9.257028112449799,
+      "grad_norm": 0.4158003330230713,
+      "learning_rate": 8.450646122203865e-07,
+      "loss": 0.3233,
+      "num_input_tokens_seen": 405536,
+      "step": 2305
+    },
+    {
+      "epoch": 9.27710843373494,
+      "grad_norm": 0.4854465425014496,
+      "learning_rate": 8.004824864909277e-07,
+      "loss": 0.3513,
+      "num_input_tokens_seen": 406368,
+      "step": 2310
+    },
+    {
+      "epoch": 9.29718875502008,
+      "grad_norm": 0.42681699991226196,
+      "learning_rate": 7.570893035856091e-07,
+      "loss": 0.3417,
+      "num_input_tokens_seen": 407184,
+      "step": 2315
+    },
+    {
+      "epoch": 9.317269076305221,
+      "grad_norm": 0.48750439286231995,
+      "learning_rate": 7.148871954483105e-07,
+      "loss": 0.3542,
+      "num_input_tokens_seen": 407904,
+      "step": 2320
+    },
+    {
+      "epoch": 9.337349397590362,
+      "grad_norm": 0.11087116599082947,
+      "learning_rate": 6.738782355044049e-07,
+      "loss": 0.3387,
+      "num_input_tokens_seen": 408736,
+      "step": 2325
+    },
+    {
+      "epoch": 9.357429718875501,
+      "grad_norm": 0.4181234538555145,
+      "learning_rate": 6.340644385588846e-07,
+      "loss": 0.33,
+      "num_input_tokens_seen": 409664,
+      "step": 2330
+    },
+    {
+      "epoch": 9.377510040160642,
+      "grad_norm": 0.1247372180223465,
+      "learning_rate": 5.954477606973679e-07,
+      "loss": 0.3643,
+      "num_input_tokens_seen": 410736,
+      "step": 2335
+    },
+    {
+      "epoch": 9.397590361445783,
+      "grad_norm": 0.09054780006408691,
+      "learning_rate": 5.580300991899989e-07,
+      "loss": 0.3612,
+      "num_input_tokens_seen": 411680,
+      "step": 2340
+    },
+    {
+      "epoch": 9.417670682730924,
+      "grad_norm": 0.44483572244644165,
+      "learning_rate": 5.218132923982267e-07,
+      "loss": 0.3417,
+      "num_input_tokens_seen": 412480,
+      "step": 2345
+    },
+    {
+      "epoch": 9.437751004016064,
+      "grad_norm": 0.07839091122150421,
+      "learning_rate": 4.867991196844918e-07,
+      "loss": 0.3352,
+      "num_input_tokens_seen": 413248,
+      "step": 2350
+    },
+    {
+      "epoch": 9.457831325301205,
+      "grad_norm": 0.12925294041633606,
+      "learning_rate": 4.5298930132480213e-07,
+      "loss": 0.3606,
+      "num_input_tokens_seen": 414080,
+      "step": 2355
+    },
+    {
+      "epoch": 9.477911646586346,
+      "grad_norm": 0.15520651638507843,
+      "learning_rate": 4.203854984242195e-07,
+      "loss": 0.3481,
+      "num_input_tokens_seen": 414928,
+      "step": 2360
+    },
+    {
+      "epoch": 9.497991967871485,
+      "grad_norm": 0.4360347092151642,
+      "learning_rate": 3.8898931283523344e-07,
+      "loss": 0.364,
+      "num_input_tokens_seen": 415728,
+      "step": 2365
+    },
+    {
+      "epoch": 9.518072289156626,
+      "grad_norm": 0.4386354088783264,
+      "learning_rate": 3.5880228707907417e-07,
+      "loss": 0.3336,
+      "num_input_tokens_seen": 416672,
+      "step": 2370
+    },
+    {
+      "epoch": 9.538152610441767,
+      "grad_norm": 0.48464497923851013,
+      "learning_rate": 3.2982590426993145e-07,
+      "loss": 0.3419,
+      "num_input_tokens_seen": 417520,
+      "step": 2375
+    },
+    {
+      "epoch": 9.538152610441767,
+      "eval_loss": 0.3474566638469696,
+      "eval_runtime": 1.2124,
+      "eval_samples_per_second": 46.19,
+      "eval_steps_per_second": 23.095,
+      "num_input_tokens_seen": 417520,
+      "step": 2375
+    },
+    {
+      "epoch": 9.558232931726907,
+      "grad_norm": 0.43182554841041565,
+      "learning_rate": 3.020615880420713e-07,
+      "loss": 0.3326,
+      "num_input_tokens_seen": 418400,
+      "step": 2380
+    },
+    {
+      "epoch": 9.578313253012048,
+      "grad_norm": 0.09252151101827621,
+      "learning_rate": 2.7551070247990305e-07,
+      "loss": 0.3449,
+      "num_input_tokens_seen": 419248,
+      "step": 2385
+    },
+    {
+      "epoch": 9.598393574297189,
+      "grad_norm": 0.44276589155197144,
+      "learning_rate": 2.501745520509552e-07,
+      "loss": 0.3481,
+      "num_input_tokens_seen": 420096,
+      "step": 2390
+    },
+    {
+      "epoch": 9.61847389558233,
+      "grad_norm": 0.07001478224992752,
+      "learning_rate": 2.2605438154179038e-07,
+      "loss": 0.3386,
+      "num_input_tokens_seen": 420848,
+      "step": 2395
+    },
+    {
+      "epoch": 9.638554216867469,
+      "grad_norm": 0.11785867065191269,
+      "learning_rate": 2.0315137599685174e-07,
+      "loss": 0.3294,
+      "num_input_tokens_seen": 421728,
+      "step": 2400
+    },
+    {
+      "epoch": 9.65863453815261,
+      "grad_norm": 0.08471374958753586,
+      "learning_rate": 1.814666606602261e-07,
+      "loss": 0.3604,
+      "num_input_tokens_seen": 422656,
+      "step": 2405
+    },
+    {
+      "epoch": 9.67871485943775,
+      "grad_norm": 0.4377971589565277,
+      "learning_rate": 1.6100130092037703e-07,
+      "loss": 0.3457,
+      "num_input_tokens_seen": 423600,
+      "step": 2410
+    },
+    {
+      "epoch": 9.698795180722891,
+      "grad_norm": 0.12307439744472504,
+      "learning_rate": 1.4175630225778947e-07,
+      "loss": 0.3447,
+      "num_input_tokens_seen": 424448,
+      "step": 2415
+    },
+    {
+      "epoch": 9.718875502008032,
+      "grad_norm": 0.11983578652143478,
+      "learning_rate": 1.237326101955677e-07,
+      "loss": 0.3544,
+      "num_input_tokens_seen": 425632,
+      "step": 2420
+    },
+    {
+      "epoch": 9.738955823293173,
+      "grad_norm": 0.1070198193192482,
+      "learning_rate": 1.0693111025300017e-07,
+      "loss": 0.3385,
+      "num_input_tokens_seen": 426576,
+      "step": 2425
+    },
+    {
+      "epoch": 9.759036144578314,
+      "grad_norm": 0.44310370087623596,
+      "learning_rate": 9.13526279020277e-08,
+      "loss": 0.3296,
+      "num_input_tokens_seen": 427376,
+      "step": 2430
+    },
+    {
+      "epoch": 9.779116465863455,
+      "grad_norm": 0.4734679162502289,
+      "learning_rate": 7.699792852670362e-08,
+      "loss": 0.3514,
+      "num_input_tokens_seen": 428256,
+      "step": 2435
+    },
+    {
+      "epoch": 9.799196787148594,
+      "grad_norm": 0.13541147112846375,
+      "learning_rate": 6.386771738558506e-08,
+      "loss": 0.3389,
+      "num_input_tokens_seen": 429216,
+      "step": 2440
+    },
+    {
+      "epoch": 9.819277108433734,
+      "grad_norm": 0.48132413625717163,
+      "learning_rate": 5.196263957708836e-08,
+      "loss": 0.3542,
+      "num_input_tokens_seen": 430208,
+      "step": 2445
+    },
+    {
+      "epoch": 9.839357429718875,
+      "grad_norm": 0.4347268044948578,
+      "learning_rate": 4.1283280007778366e-08,
+      "loss": 0.3292,
+      "num_input_tokens_seen": 430960,
+      "step": 2450
+    },
+    {
+      "epoch": 9.859437751004016,
+      "grad_norm": 0.43290168046951294,
+      "learning_rate": 3.1830163363655296e-08,
+      "loss": 0.355,
+      "num_input_tokens_seen": 431936,
+      "step": 2455
+    },
+    {
+      "epoch": 9.879518072289157,
+      "grad_norm": 0.10507988184690475,
+      "learning_rate": 2.3603754084358663e-08,
+      "loss": 0.3425,
+      "num_input_tokens_seen": 432912,
+      "step": 2460
+    },
+    {
+      "epoch": 9.899598393574298,
+      "grad_norm": 0.4859294593334198,
+      "learning_rate": 1.6604456340352235e-08,
+      "loss": 0.3449,
+      "num_input_tokens_seen": 433696,
+      "step": 2465
+    },
+    {
+      "epoch": 9.919678714859439,
+      "grad_norm": 0.1426219940185547,
+      "learning_rate": 1.0832614013073228e-08,
+      "loss": 0.3513,
+      "num_input_tokens_seen": 434528,
+      "step": 2470
+    },
+    {
+      "epoch": 9.939759036144578,
+      "grad_norm": 0.08813058584928513,
+      "learning_rate": 6.288510678031934e-09,
+      "loss": 0.3633,
+      "num_input_tokens_seen": 435280,
+      "step": 2475
+    },
+    {
+      "epoch": 9.959839357429718,
+      "grad_norm": 0.4306187033653259,
+      "learning_rate": 2.972369590878432e-09,
+      "loss": 0.3483,
+      "num_input_tokens_seen": 436096,
+      "step": 2480
+    },
+    {
+      "epoch": 9.97991967871486,
+      "grad_norm": 0.07171786576509476,
+      "learning_rate": 8.843536764419069e-10,
+      "loss": 0.3521,
+      "num_input_tokens_seen": 437008,
+      "step": 2485
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 0.0965154618024826,
+      "learning_rate": 2.4565520709285417e-11,
+      "loss": 0.3294,
+      "num_input_tokens_seen": 437760,
+      "step": 2490
+    },
+    {
+      "epoch": 10.0,
+      "num_input_tokens_seen": 437760,
+      "step": 2490,
+      "total_flos": 1.971213494648832e+16,
+      "train_loss": 0.4594599806580199,
+      "train_runtime": 258.0589,
+      "train_samples_per_second": 19.298,
+      "train_steps_per_second": 9.649
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2490,
+  "num_input_tokens_seen": 437760,
+  "num_train_epochs": 10,
+  "save_steps": 125,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.971213494648832e+16,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}