{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9984301412872841,
  "eval_steps": 100,
  "global_step": 477,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 0.400390625,
      "learning_rate": 1.0416666666666667e-07,
      "logits/chosen": -2.2547454833984375,
      "logits/rejected": -2.401865005493164,
      "logps/chosen": -53.759212493896484,
      "logps/rejected": -48.83185958862305,
      "loss": 0.6931,
      "pred_label": 0.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1,
      "use_label": 10.0
    },
    {
      "epoch": 0.02,
      "grad_norm": 0.4609375,
      "learning_rate": 1.0416666666666667e-06,
      "logits/chosen": -2.2421748638153076,
      "logits/rejected": -2.2769579887390137,
      "logps/chosen": -51.987098693847656,
      "logps/rejected": -64.96717071533203,
      "loss": 0.6929,
      "pred_label": 0.0,
      "rewards/accuracies": 0.25,
      "rewards/chosen": 0.0019227324519306421,
      "rewards/margins": 0.0004911368596367538,
      "rewards/rejected": 0.0014315954176709056,
      "step": 10,
      "use_label": 90.0
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.39453125,
      "learning_rate": 2.0833333333333334e-06,
      "logits/chosen": -2.2521612644195557,
      "logits/rejected": -2.255767822265625,
      "logps/chosen": -62.4937629699707,
      "logps/rejected": -72.63874816894531,
      "loss": 0.6919,
      "pred_label": 0.0,
      "rewards/accuracies": 0.2874999940395355,
      "rewards/chosen": 0.01600126549601555,
      "rewards/margins": 0.0011427802965044975,
      "rewards/rejected": 0.0148584870621562,
      "step": 20,
      "use_label": 242.0
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.51171875,
      "learning_rate": 3.125e-06,
      "logits/chosen": -2.3423426151275635,
      "logits/rejected": -2.3549609184265137,
      "logps/chosen": -79.10475158691406,
      "logps/rejected": -98.8157958984375,
      "loss": 0.6897,
      "pred_label": 0.0,
      "rewards/accuracies": 0.28125,
      "rewards/chosen": 0.03137165680527687,
      "rewards/margins": 0.0032712810207158327,
      "rewards/rejected": 0.028100375086069107,
      "step": 30,
      "use_label": 402.0
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.51953125,
      "learning_rate": 4.166666666666667e-06,
      "logits/chosen": -2.323338270187378,
      "logits/rejected": -2.3015079498291016,
      "logps/chosen": -82.85453796386719,
      "logps/rejected": -82.39984893798828,
      "loss": 0.6866,
      "pred_label": 0.0,
      "rewards/accuracies": 0.2874999940395355,
      "rewards/chosen": 0.03337595611810684,
      "rewards/margins": 0.011919925920665264,
      "rewards/rejected": 0.021456023678183556,
      "step": 40,
      "use_label": 562.0
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.671875,
      "learning_rate": 4.999731868769027e-06,
      "logits/chosen": -2.2404515743255615,
      "logits/rejected": -2.262972354888916,
      "logps/chosen": -67.89888000488281,
      "logps/rejected": -81.8695068359375,
      "loss": 0.6805,
      "pred_label": 0.0,
      "rewards/accuracies": 0.32499998807907104,
      "rewards/chosen": 0.009319942444562912,
      "rewards/margins": 0.030618786811828613,
      "rewards/rejected": -0.0212988443672657,
      "step": 50,
      "use_label": 722.0
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0234375,
      "learning_rate": 4.9903533134293035e-06,
      "logits/chosen": -2.2157275676727295,
      "logits/rejected": -2.155928134918213,
      "logps/chosen": -63.64031982421875,
      "logps/rejected": -73.28236389160156,
      "loss": 0.6752,
      "pred_label": 0.0,
      "rewards/accuracies": 0.3062500059604645,
      "rewards/chosen": -0.03914070501923561,
      "rewards/margins": 0.04399287328124046,
      "rewards/rejected": -0.08313358575105667,
      "step": 60,
      "use_label": 882.0
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.859375,
      "learning_rate": 4.967625656594782e-06,
      "logits/chosen": -2.114478588104248,
      "logits/rejected": -2.1126065254211426,
      "logps/chosen": -70.76527404785156,
      "logps/rejected": -83.94652557373047,
      "loss": 0.6712,
      "pred_label": 0.0,
      "rewards/accuracies": 0.25,
      "rewards/chosen": -0.15054164826869965,
      "rewards/margins": 0.030909737572073936,
      "rewards/rejected": -0.18145139515399933,
      "step": 70,
      "use_label": 1042.0
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.1640625,
      "learning_rate": 4.93167072587771e-06,
      "logits/chosen": -2.2166943550109863,
      "logits/rejected": -2.1609182357788086,
      "logps/chosen": -54.8065185546875,
      "logps/rejected": -69.45613861083984,
      "loss": 0.6589,
      "pred_label": 0.4749999940395355,
      "rewards/accuracies": 0.26249998807907104,
      "rewards/chosen": -0.06275613605976105,
      "rewards/margins": 0.10003063827753067,
      "rewards/rejected": -0.16278676688671112,
      "step": 80,
      "use_label": 1201.5250244140625
    },
    {
      "epoch": 0.19,
      "grad_norm": 1.8125,
      "learning_rate": 4.882681251368549e-06,
      "logits/chosen": -1.9692049026489258,
      "logits/rejected": -1.9792039394378662,
      "logps/chosen": -76.60871887207031,
      "logps/rejected": -96.53330993652344,
      "loss": 0.6564,
      "pred_label": 2.0999999046325684,
      "rewards/accuracies": 0.29374998807907104,
      "rewards/chosen": -0.18226662278175354,
      "rewards/margins": 0.09542477130889893,
      "rewards/rejected": -0.27769142389297485,
      "step": 90,
      "use_label": 1359.9000244140625
    },
    {
      "epoch": 0.21,
      "grad_norm": 2.171875,
      "learning_rate": 4.8209198325401815e-06,
      "logits/chosen": -1.9027693271636963,
      "logits/rejected": -1.8775581121444702,
      "logps/chosen": -92.94733428955078,
      "logps/rejected": -84.73824310302734,
      "loss": 0.6531,
      "pred_label": 4.0,
      "rewards/accuracies": 0.32499998807907104,
      "rewards/chosen": -0.12917451560497284,
      "rewards/margins": 0.07954015582799911,
      "rewards/rejected": -0.20871467888355255,
      "step": 100,
      "use_label": 1518.0
    },
    {
      "epoch": 0.21,
      "eval_logits/chosen": -1.7353737354278564,
      "eval_logits/rejected": -1.7198325395584106,
      "eval_logps/chosen": -80.33845520019531,
      "eval_logps/rejected": -106.64702606201172,
      "eval_loss": 0.6527961492538452,
      "eval_pred_label": 6.6875,
      "eval_rewards/accuracies": 0.36328125,
      "eval_rewards/chosen": -0.1642620712518692,
      "eval_rewards/margins": 0.13027876615524292,
      "eval_rewards/rejected": -0.2945408225059509,
      "eval_runtime": 125.2319,
      "eval_samples_per_second": 15.97,
      "eval_steps_per_second": 0.256,
      "eval_use_label": 1725.3125,
      "step": 100
    },
    {
      "epoch": 0.23,
      "grad_norm": 2.0,
      "learning_rate": 4.746717530629565e-06,
      "logits/chosen": -1.7974278926849365,
      "logits/rejected": -1.7697474956512451,
      "logps/chosen": -89.79286193847656,
      "logps/rejected": -113.6241455078125,
      "loss": 0.6479,
      "pred_label": 9.199999809265137,
      "rewards/accuracies": 0.36250001192092896,
      "rewards/chosen": -0.18692079186439514,
      "rewards/margins": 0.16341358423233032,
      "rewards/rejected": -0.3503343462944031,
      "step": 110,
      "use_label": 1928.800048828125
    },
    {
      "epoch": 0.25,
      "grad_norm": 2.890625,
      "learning_rate": 4.660472094042121e-06,
      "logits/chosen": -1.454304814338684,
      "logits/rejected": -1.3457725048065186,
      "logps/chosen": -109.3675537109375,
      "logps/rejected": -133.90725708007812,
      "loss": 0.6432,
      "pred_label": 14.949999809265137,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.3942197263240814,
      "rewards/margins": 0.21566259860992432,
      "rewards/rejected": -0.6098822951316833,
      "step": 120,
      "use_label": 2083.050048828125
    },
    {
      "epoch": 0.27,
      "grad_norm": 2.5625,
      "learning_rate": 4.5626458262912745e-06,
      "logits/chosen": -1.0859026908874512,
      "logits/rejected": -1.0426993370056152,
      "logps/chosen": -112.0394515991211,
      "logps/rejected": -139.61097717285156,
      "loss": 0.6391,
      "pred_label": 21.049999237060547,
      "rewards/accuracies": 0.36250001192092896,
      "rewards/chosen": -0.4626106321811676,
      "rewards/margins": 0.20503444969654083,
      "rewards/rejected": -0.6676451563835144,
      "step": 130,
      "use_label": 2236.949951171875
    },
    {
      "epoch": 0.29,
      "grad_norm": 2.484375,
      "learning_rate": 4.453763107901676e-06,
      "logits/chosen": -0.735418975353241,
      "logits/rejected": -0.8380192518234253,
      "logps/chosen": -138.07081604003906,
      "logps/rejected": -150.91665649414062,
      "loss": 0.6252,
      "pred_label": 31.399999618530273,
      "rewards/accuracies": 0.30000001192092896,
      "rewards/chosen": -0.5732325315475464,
      "rewards/margins": 0.1448771208524704,
      "rewards/rejected": -0.7181096076965332,
      "step": 140,
      "use_label": 2386.60009765625
    },
    {
      "epoch": 0.31,
      "grad_norm": 3.859375,
      "learning_rate": 4.33440758555951e-06,
      "logits/chosen": -0.48231878876686096,
      "logits/rejected": -0.43882569670677185,
      "logps/chosen": -117.69664001464844,
      "logps/rejected": -150.86083984375,
      "loss": 0.6219,
      "pred_label": 43.45000076293945,
      "rewards/accuracies": 0.32499998807907104,
      "rewards/chosen": -0.5254992246627808,
      "rewards/margins": 0.3047201633453369,
      "rewards/rejected": -0.8302194476127625,
      "step": 150,
      "use_label": 2534.550048828125
    },
    {
      "epoch": 0.33,
      "grad_norm": 2.890625,
      "learning_rate": 4.205219043576955e-06,
      "logits/chosen": -0.15186011791229248,
      "logits/rejected": -0.17336201667785645,
      "logps/chosen": -128.78500366210938,
      "logps/rejected": -159.26498413085938,
      "loss": 0.5982,
      "pred_label": 58.25,
      "rewards/accuracies": 0.2750000059604645,
      "rewards/chosen": -0.6445494294166565,
      "rewards/margins": 0.17397476732730865,
      "rewards/rejected": -0.818524181842804,
      "step": 160,
      "use_label": 2679.75
    },
    {
      "epoch": 0.36,
      "grad_norm": 3.328125,
      "learning_rate": 4.066889974440757e-06,
      "logits/chosen": 0.14322622120380402,
      "logits/rejected": 0.18100713193416595,
      "logps/chosen": -108.39127349853516,
      "logps/rejected": -140.55824279785156,
      "loss": 0.5938,
      "pred_label": 79.57499694824219,
      "rewards/accuracies": 0.3062500059604645,
      "rewards/chosen": -0.5288320779800415,
      "rewards/margins": 0.23454061150550842,
      "rewards/rejected": -0.7633727192878723,
      "step": 170,
      "use_label": 2818.425048828125
    },
    {
      "epoch": 0.38,
      "grad_norm": 3.0,
      "learning_rate": 3.92016186682789e-06,
      "logits/chosen": -0.20601686835289001,
      "logits/rejected": -0.09364790469408035,
      "logps/chosen": -105.94217681884766,
      "logps/rejected": -130.695556640625,
      "loss": 0.6262,
      "pred_label": 100.2750015258789,
      "rewards/accuracies": 0.35624998807907104,
      "rewards/chosen": -0.4315454065799713,
      "rewards/margins": 0.3125666677951813,
      "rewards/rejected": -0.7441121339797974,
      "step": 180,
      "use_label": 2957.72509765625
    },
    {
      "epoch": 0.4,
      "grad_norm": 2.734375,
      "learning_rate": 3.7658212309857576e-06,
      "logits/chosen": -0.34412023425102234,
      "logits/rejected": -0.07299783080816269,
      "logps/chosen": -107.5626449584961,
      "logps/rejected": -141.1322479248047,
      "loss": 0.6092,
      "pred_label": 121.05000305175781,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.48451024293899536,
      "rewards/margins": 0.28280580043792725,
      "rewards/rejected": -0.7673160433769226,
      "step": 190,
      "use_label": 3096.949951171875
    },
    {
      "epoch": 0.42,
      "grad_norm": 6.5,
      "learning_rate": 3.604695382782159e-06,
      "logits/chosen": 0.03128425031900406,
      "logits/rejected": 0.20205454528331757,
      "logps/chosen": -145.35342407226562,
      "logps/rejected": -162.05667114257812,
      "loss": 0.6041,
      "pred_label": 135.89999389648438,
      "rewards/accuracies": 0.30000001192092896,
      "rewards/chosen": -0.6367710828781128,
      "rewards/margins": 0.25234952569007874,
      "rewards/rejected": -0.8891205787658691,
      "step": 200,
      "use_label": 3242.10009765625
    },
    {
      "epoch": 0.42,
      "eval_logits/chosen": 0.886444091796875,
      "eval_logits/rejected": 0.9784458875656128,
      "eval_logps/chosen": -135.34742736816406,
      "eval_logps/rejected": -187.65963745117188,
      "eval_loss": 0.5936154723167419,
      "eval_pred_label": 167.40625,
      "eval_rewards/accuracies": 0.3515625,
      "eval_rewards/chosen": -0.7143516540527344,
      "eval_rewards/margins": 0.3903152644634247,
      "eval_rewards/rejected": -1.1046667098999023,
      "eval_runtime": 125.3006,
      "eval_samples_per_second": 15.962,
      "eval_steps_per_second": 0.255,
      "eval_use_label": 3420.59375,
      "step": 200
    },
    {
      "epoch": 0.44,
      "grad_norm": 3.796875,
      "learning_rate": 3.437648009023905e-06,
      "logits/chosen": 0.6729141473770142,
      "logits/rejected": 0.6579598188400269,
      "logps/chosen": -119.19351959228516,
      "logps/rejected": -159.00997924804688,
      "loss": 0.5936,
      "pred_label": 201.6999969482422,
      "rewards/accuracies": 0.3499999940395355,
      "rewards/chosen": -0.63218754529953,
      "rewards/margins": 0.3281194567680359,
      "rewards/rejected": -0.9603070020675659,
      "step": 210,
      "use_label": 3592.300048828125
    },
    {
      "epoch": 0.46,
      "grad_norm": 4.5,
      "learning_rate": 3.265574537815398e-06,
      "logits/chosen": 0.2854166626930237,
      "logits/rejected": 0.4488348066806793,
      "logps/chosen": -148.92379760742188,
      "logps/rejected": -161.19557189941406,
      "loss": 0.5938,
      "pred_label": 225.52499389648438,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.6929510235786438,
      "rewards/margins": 0.2553243637084961,
      "rewards/rejected": -0.9482753872871399,
      "step": 220,
      "use_label": 3728.47509765625
    },
    {
      "epoch": 0.48,
      "grad_norm": 2.28125,
      "learning_rate": 3.089397338773569e-06,
      "logits/chosen": 0.00020002425299026072,
      "logits/rejected": 0.1493436098098755,
      "logps/chosen": -103.05213928222656,
      "logps/rejected": -136.05099487304688,
      "loss": 0.597,
      "pred_label": 252.02499389648438,
      "rewards/accuracies": 0.30000001192092896,
      "rewards/chosen": -0.3861756920814514,
      "rewards/margins": 0.3467464745044708,
      "rewards/rejected": -0.7329221963882446,
      "step": 230,
      "use_label": 3861.97509765625
    },
    {
      "epoch": 0.5,
      "grad_norm": 3.125,
      "learning_rate": 2.9100607788275547e-06,
      "logits/chosen": 0.49308425188064575,
      "logits/rejected": 0.44487372040748596,
      "logps/chosen": -109.46275329589844,
      "logps/rejected": -153.8666534423828,
      "loss": 0.584,
      "pred_label": 275.375,
      "rewards/accuracies": 0.3687500059604645,
      "rewards/chosen": -0.40430212020874023,
      "rewards/margins": 0.3921273946762085,
      "rewards/rejected": -0.7964295148849487,
      "step": 240,
      "use_label": 3998.625
    },
    {
      "epoch": 0.52,
      "grad_norm": 2.0625,
      "learning_rate": 2.72852616010567e-06,
      "logits/chosen": 0.3891890347003937,
      "logits/rejected": 0.47166162729263306,
      "logps/chosen": -122.5915298461914,
      "logps/rejected": -153.12493896484375,
      "loss": 0.5998,
      "pred_label": 301.875,
      "rewards/accuracies": 0.36250001192092896,
      "rewards/chosen": -0.4919242262840271,
      "rewards/margins": 0.3470008671283722,
      "rewards/rejected": -0.8389250636100769,
      "step": 250,
      "use_label": 4132.125
    },
    {
      "epoch": 0.54,
      "grad_norm": 2.171875,
      "learning_rate": 2.5457665670441937e-06,
      "logits/chosen": 0.4214790463447571,
      "logits/rejected": 0.4202333092689514,
      "logps/chosen": -116.09378814697266,
      "logps/rejected": -156.8458251953125,
      "loss": 0.592,
      "pred_label": 326.3500061035156,
      "rewards/accuracies": 0.32499998807907104,
      "rewards/chosen": -0.4998815953731537,
      "rewards/margins": 0.324097216129303,
      "rewards/rejected": -0.8239787817001343,
      "step": 260,
      "use_label": 4267.64990234375
    },
    {
      "epoch": 0.57,
      "grad_norm": 3.40625,
      "learning_rate": 2.3627616503391813e-06,
      "logits/chosen": 0.9609361886978149,
      "logits/rejected": 0.8760908246040344,
      "logps/chosen": -142.81573486328125,
      "logps/rejected": -170.10379028320312,
      "loss": 0.5888,
      "pred_label": 343.04998779296875,
      "rewards/accuracies": 0.3375000059604645,
      "rewards/chosen": -0.698999285697937,
      "rewards/margins": 0.30828553438186646,
      "rewards/rejected": -1.0072848796844482,
      "step": 270,
      "use_label": 4410.9501953125
    },
    {
      "epoch": 0.59,
      "grad_norm": 2.234375,
      "learning_rate": 2.1804923757009885e-06,
      "logits/chosen": 1.1657536029815674,
      "logits/rejected": 1.3259608745574951,
      "logps/chosen": -131.4703826904297,
      "logps/rejected": -156.4979248046875,
      "loss": 0.6007,
      "pred_label": 361.5,
      "rewards/accuracies": 0.3375000059604645,
      "rewards/chosen": -0.6596485376358032,
      "rewards/margins": 0.27613458037376404,
      "rewards/rejected": -0.9357832074165344,
      "step": 280,
      "use_label": 4552.5
    },
    {
      "epoch": 0.61,
      "grad_norm": 2.875,
      "learning_rate": 1.9999357655598894e-06,
      "logits/chosen": 0.9594011306762695,
      "logits/rejected": 0.9126796722412109,
      "logps/chosen": -144.55104064941406,
      "logps/rejected": -183.51065063476562,
      "loss": 0.5899,
      "pred_label": 386.07501220703125,
      "rewards/accuracies": 0.3187499940395355,
      "rewards/chosen": -0.7800258994102478,
      "rewards/margins": 0.2914672791957855,
      "rewards/rejected": -1.0714932680130005,
      "step": 290,
      "use_label": 4687.9248046875
    },
    {
      "epoch": 0.63,
      "grad_norm": 3.1875,
      "learning_rate": 1.8220596619089576e-06,
      "logits/chosen": 1.2753574848175049,
      "logits/rejected": 1.1057071685791016,
      "logps/chosen": -165.4674072265625,
      "logps/rejected": -223.6466064453125,
      "loss": 0.5763,
      "pred_label": 409.9750061035156,
      "rewards/accuracies": 0.4124999940395355,
      "rewards/chosen": -0.8787307739257812,
      "rewards/margins": 0.41653138399124146,
      "rewards/rejected": -1.295262098312378,
      "step": 300,
      "use_label": 4824.02490234375
    },
    {
      "epoch": 0.63,
      "eval_logits/chosen": 1.6598718166351318,
      "eval_logits/rejected": 1.7526323795318604,
      "eval_logps/chosen": -143.2136993408203,
      "eval_logps/rejected": -200.36146545410156,
      "eval_loss": 0.5773172974586487,
      "eval_pred_label": 452.71875,
      "eval_rewards/accuracies": 0.3515625,
      "eval_rewards/chosen": -0.7930145263671875,
      "eval_rewards/margins": 0.4386705756187439,
      "eval_rewards/rejected": -1.2316851615905762,
      "eval_runtime": 125.3512,
      "eval_samples_per_second": 15.955,
      "eval_steps_per_second": 0.255,
      "eval_use_label": 4991.28125,
      "step": 300
    },
    {
      "epoch": 0.65,
      "grad_norm": 3.0625,
      "learning_rate": 1.647817538357072e-06,
      "logits/chosen": 1.3793504238128662,
      "logits/rejected": 1.4072078466415405,
      "logps/chosen": -126.9173583984375,
      "logps/rejected": -186.46255493164062,
      "loss": 0.5633,
      "pred_label": 494.42498779296875,
      "rewards/accuracies": 0.36250001192092896,
      "rewards/chosen": -0.7132076025009155,
      "rewards/margins": 0.4689141809940338,
      "rewards/rejected": -1.1821218729019165,
      "step": 310,
      "use_label": 5155.5751953125
    },
    {
      "epoch": 0.67,
      "grad_norm": 4.21875,
      "learning_rate": 1.4781433892011132e-06,
      "logits/chosen": 1.2615296840667725,
      "logits/rejected": 1.4717950820922852,
      "logps/chosen": -163.67529296875,
      "logps/rejected": -205.421142578125,
      "loss": 0.5761,
      "pred_label": 523.5250244140625,
      "rewards/accuracies": 0.3812499940395355,
      "rewards/chosen": -0.9060484766960144,
      "rewards/margins": 0.4762052893638611,
      "rewards/rejected": -1.382253885269165,
      "step": 320,
      "use_label": 5286.47509765625
    },
    {
      "epoch": 0.69,
      "grad_norm": 2.359375,
      "learning_rate": 1.3139467229135999e-06,
      "logits/chosen": 1.4169238805770874,
      "logits/rejected": 1.4296729564666748,
      "logps/chosen": -150.2149200439453,
      "logps/rejected": -186.73570251464844,
      "loss": 0.5799,
      "pred_label": 550.125,
      "rewards/accuracies": 0.33125001192092896,
      "rewards/chosen": -0.8010675311088562,
      "rewards/margins": 0.3802093267440796,
      "rewards/rejected": -1.1812770366668701,
      "step": 330,
      "use_label": 5419.875
    },
    {
      "epoch": 0.71,
      "grad_norm": 3.15625,
      "learning_rate": 1.1561076868822756e-06,
      "logits/chosen": 0.9984269142150879,
      "logits/rejected": 0.9373771548271179,
      "logps/chosen": -161.85842895507812,
      "logps/rejected": -182.74703979492188,
      "loss": 0.5933,
      "pred_label": 567.2000122070312,
      "rewards/accuracies": 0.32499998807907104,
      "rewards/chosen": -0.771192193031311,
      "rewards/margins": 0.2911759614944458,
      "rewards/rejected": -1.0623681545257568,
      "step": 340,
      "use_label": 5562.7998046875
    },
    {
      "epoch": 0.73,
      "grad_norm": 2.59375,
      "learning_rate": 1.0054723495346484e-06,
      "logits/chosen": 0.83796626329422,
      "logits/rejected": 0.8520887494087219,
      "logps/chosen": -176.03054809570312,
      "logps/rejected": -217.10214233398438,
      "loss": 0.5863,
      "pred_label": 598.5750122070312,
      "rewards/accuracies": 0.36250001192092896,
      "rewards/chosen": -0.86613929271698,
      "rewards/margins": 0.4522012174129486,
      "rewards/rejected": -1.318340539932251,
      "step": 350,
      "use_label": 5691.4248046875
    },
    {
      "epoch": 0.75,
      "grad_norm": 2.234375,
      "learning_rate": 8.628481651367876e-07,
      "logits/chosen": 0.7010875940322876,
      "logits/rejected": 0.8413160443305969,
      "logps/chosen": -126.9655532836914,
      "logps/rejected": -182.5807342529297,
      "loss": 0.5885,
      "pred_label": 629.0,
      "rewards/accuracies": 0.39375001192092896,
      "rewards/chosen": -0.6332792043685913,
      "rewards/margins": 0.47024598717689514,
      "rewards/rejected": -1.103525161743164,
      "step": 360,
      "use_label": 5821.0
    },
    {
      "epoch": 0.77,
      "grad_norm": 2.4375,
      "learning_rate": 7.289996455765749e-07,
      "logits/chosen": 0.8454801440238953,
      "logits/rejected": 0.9659041166305542,
      "logps/chosen": -120.26502990722656,
      "logps/rejected": -170.44923400878906,
      "loss": 0.585,
      "pred_label": 655.625,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.5483022928237915,
      "rewards/margins": 0.4770358204841614,
      "rewards/rejected": -1.0253381729125977,
      "step": 370,
      "use_label": 5954.375
    },
    {
      "epoch": 0.8,
      "grad_norm": 3.5625,
      "learning_rate": 6.046442623320145e-07,
      "logits/chosen": 0.7346574664115906,
      "logits/rejected": 0.7028430104255676,
      "logps/chosen": -131.0785675048828,
      "logps/rejected": -188.57435607910156,
      "loss": 0.589,
      "pred_label": 685.5250244140625,
      "rewards/accuracies": 0.3187499940395355,
      "rewards/chosen": -0.6524317264556885,
      "rewards/margins": 0.3696710765361786,
      "rewards/rejected": -1.0221028327941895,
      "step": 380,
      "use_label": 6084.47509765625
    },
    {
      "epoch": 0.82,
      "grad_norm": 3.3125,
      "learning_rate": 4.904486005914027e-07,
      "logits/chosen": 1.1143369674682617,
      "logits/rejected": 0.8643951416015625,
      "logps/chosen": -179.11276245117188,
      "logps/rejected": -220.11068725585938,
      "loss": 0.5727,
      "pred_label": 717.625,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.8629444241523743,
      "rewards/margins": 0.508463442325592,
      "rewards/rejected": -1.3714077472686768,
      "step": 390,
      "use_label": 6212.375
    },
    {
      "epoch": 0.84,
      "grad_norm": 3.25,
      "learning_rate": 3.8702478614051353e-07,
      "logits/chosen": 0.8043449521064758,
      "logits/rejected": 0.9917415380477905,
      "logps/chosen": -130.07017517089844,
      "logps/rejected": -163.469970703125,
      "loss": 0.5836,
      "pred_label": 747.5750122070312,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.5757918953895569,
      "rewards/margins": 0.42427974939346313,
      "rewards/rejected": -1.0000715255737305,
      "step": 400,
      "use_label": 6342.4248046875
    },
    {
      "epoch": 0.84,
      "eval_logits/chosen": 1.75760817527771,
      "eval_logits/rejected": 1.8499951362609863,
      "eval_logps/chosen": -130.3719482421875,
      "eval_logps/rejected": -190.7267303466797,
      "eval_loss": 0.5768851041793823,
      "eval_pred_label": 782.8125,
      "eval_rewards/accuracies": 0.37109375,
      "eval_rewards/chosen": -0.6645968556404114,
      "eval_rewards/margins": 0.4707409739494324,
      "eval_rewards/rejected": -1.1353378295898438,
      "eval_runtime": 147.391,
      "eval_samples_per_second": 13.569,
      "eval_steps_per_second": 0.217,
      "eval_use_label": 6517.1875,
      "step": 400
    },
    {
      "epoch": 0.86,
      "grad_norm": 3.390625,
      "learning_rate": 2.9492720416985004e-07,
      "logits/chosen": 1.1002473831176758,
      "logits/rejected": 1.1428117752075195,
      "logps/chosen": -126.85247802734375,
      "logps/rejected": -170.77365112304688,
      "loss": 0.5838,
      "pred_label": 822.75,
      "rewards/accuracies": 0.3687500059604645,
      "rewards/chosen": -0.6542948484420776,
      "rewards/margins": 0.4562492370605469,
      "rewards/rejected": -1.1105440855026245,
      "step": 410,
      "use_label": 6683.25
    },
    {
      "epoch": 0.88,
      "grad_norm": 2.609375,
      "learning_rate": 2.1464952759020857e-07,
      "logits/chosen": 1.3246395587921143,
      "logits/rejected": 1.2824434041976929,
      "logps/chosen": -122.80003356933594,
      "logps/rejected": -138.56423950195312,
      "loss": 0.5822,
      "pred_label": 846.4249877929688,
      "rewards/accuracies": 0.26249998807907104,
      "rewards/chosen": -0.6186091303825378,
      "rewards/margins": 0.25320303440093994,
      "rewards/rejected": -0.8718121647834778,
      "step": 420,
      "use_label": 6819.5751953125
    },
    {
      "epoch": 0.9,
      "grad_norm": 2.09375,
      "learning_rate": 1.4662207078575685e-07,
      "logits/chosen": 1.270193099975586,
      "logits/rejected": 1.253873348236084,
      "logps/chosen": -171.46336364746094,
      "logps/rejected": -207.75607299804688,
      "loss": 0.564,
      "pred_label": 873.5499877929688,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": -0.7219651341438293,
      "rewards/margins": 0.5620242357254028,
      "rewards/rejected": -1.283989429473877,
      "step": 430,
      "use_label": 6952.4501953125
    },
    {
      "epoch": 0.92,
      "grad_norm": 2.578125,
      "learning_rate": 9.120948298936422e-08,
      "logits/chosen": 1.221411943435669,
      "logits/rejected": 1.397247552871704,
      "logps/chosen": -136.4575653076172,
      "logps/rejected": -193.40870666503906,
      "loss": 0.5736,
      "pred_label": 905.5750122070312,
      "rewards/accuracies": 0.36250001192092896,
      "rewards/chosen": -0.6955360770225525,
      "rewards/margins": 0.4879213869571686,
      "rewards/rejected": -1.183457612991333,
      "step": 440,
      "use_label": 7080.4248046875
    },
    {
      "epoch": 0.94,
      "grad_norm": 4.78125,
      "learning_rate": 4.870879364444109e-08,
      "logits/chosen": 1.6054052114486694,
      "logits/rejected": 1.3484258651733398,
      "logps/chosen": -148.17161560058594,
      "logps/rejected": -205.789306640625,
      "loss": 0.583,
      "pred_label": 930.4749755859375,
      "rewards/accuracies": 0.35624998807907104,
      "rewards/chosen": -0.7591380476951599,
      "rewards/margins": 0.4158584177494049,
      "rewards/rejected": -1.1749964952468872,
      "step": 450,
      "use_label": 7215.52490234375
    },
    {
      "epoch": 0.96,
      "grad_norm": 2.875,
      "learning_rate": 1.93478202307823e-08,
      "logits/chosen": 1.4640157222747803,
      "logits/rejected": 1.4903802871704102,
      "logps/chosen": -96.6323471069336,
      "logps/rejected": -150.8868865966797,
      "loss": 0.5814,
      "pred_label": 961.25,
      "rewards/accuracies": 0.32499998807907104,
      "rewards/chosen": -0.5051247477531433,
      "rewards/margins": 0.3702928125858307,
      "rewards/rejected": -0.8754175901412964,
      "step": 460,
      "use_label": 7344.75
    },
    {
      "epoch": 0.98,
      "grad_norm": 2.75,
      "learning_rate": 3.283947088983663e-09,
      "logits/chosen": 1.464422345161438,
      "logits/rejected": 1.2297132015228271,
      "logps/chosen": -130.30838012695312,
      "logps/rejected": -166.67605590820312,
      "loss": 0.5822,
      "pred_label": 982.8499755859375,
      "rewards/accuracies": 0.30000001192092896,
      "rewards/chosen": -0.6297920346260071,
      "rewards/margins": 0.34639838337898254,
      "rewards/rejected": -0.9761903882026672,
      "step": 470,
      "use_label": 7483.14990234375
    },
    {
      "epoch": 1.0,
      "step": 477,
      "total_flos": 0.0,
      "train_loss": 0.6110695428068533,
      "train_runtime": 9999.3279,
      "train_samples_per_second": 6.114,
      "train_steps_per_second": 0.048
    }
  ],
  "logging_steps": 10,
  "max_steps": 477,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}