{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 355,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 2.1560881180741176,
      "learning_rate": 1.3888888888888888e-07,
      "logits/chosen": -2.804708957672119,
      "logits/rejected": -2.8150453567504883,
      "logps/chosen": -217.97438049316406,
      "logps/rejected": -216.58865356445312,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/margins_max": 0.0,
      "rewards/margins_min": 0.0,
      "rewards/margins_std": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 10.20672067373361,
      "learning_rate": 1.3888888888888892e-06,
      "logits/chosen": -2.8844592571258545,
      "logits/rejected": -2.79921555519104,
      "logps/chosen": -366.6466064453125,
      "logps/rejected": -275.4883728027344,
      "loss": 0.6927,
      "rewards/accuracies": 0.5138888955116272,
      "rewards/chosen": 0.0008819872164167464,
      "rewards/margins": 0.0012480419827625155,
      "rewards/margins_max": 0.005487875547260046,
      "rewards/margins_min": -0.0020528410095721483,
      "rewards/margins_std": 0.003371911123394966,
      "rewards/rejected": -0.00036605476634576917,
      "step": 10
    },
    {
      "epoch": 0.06,
      "grad_norm": 2.007156324220208,
      "learning_rate": 2.7777777777777783e-06,
      "logits/chosen": -2.7444252967834473,
      "logits/rejected": -2.6909613609313965,
      "logps/chosen": -328.3125915527344,
      "logps/rejected": -217.0628204345703,
      "loss": 0.6867,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 0.008345814421772957,
      "rewards/margins": 0.01153595745563507,
      "rewards/margins_max": 0.026045221835374832,
      "rewards/margins_min": 0.0008641376043669879,
      "rewards/margins_std": 0.011482259258627892,
      "rewards/rejected": -0.003190141636878252,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 2.158164459848784,
      "learning_rate": 4.166666666666667e-06,
      "logits/chosen": -2.825068712234497,
      "logits/rejected": -2.7449638843536377,
      "logps/chosen": -325.515869140625,
      "logps/rejected": -234.0052490234375,
      "loss": 0.6655,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": 0.0376933328807354,
      "rewards/margins": 0.0479232482612133,
      "rewards/margins_max": 0.11641480773687363,
      "rewards/margins_min": 0.0012140764156356454,
      "rewards/margins_std": 0.05392361804842949,
      "rewards/rejected": -0.010229915380477905,
      "step": 30
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.7675389406697828,
      "learning_rate": 4.998060489154965e-06,
      "logits/chosen": -2.7818994522094727,
      "logits/rejected": -2.7307980060577393,
      "logps/chosen": -276.9420471191406,
      "logps/rejected": -230.1887969970703,
      "loss": 0.6326,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 0.08895337581634521,
      "rewards/margins": 0.11473529040813446,
      "rewards/margins_max": 0.2537270188331604,
      "rewards/margins_min": 0.01772489957511425,
      "rewards/margins_std": 0.10599502176046371,
      "rewards/rejected": -0.025781899690628052,
      "step": 40
    },
    {
      "epoch": 0.14,
      "grad_norm": 2.021631325803066,
      "learning_rate": 4.976275538042932e-06,
      "logits/chosen": -2.7466044425964355,
      "logits/rejected": -2.6687169075012207,
      "logps/chosen": -304.30511474609375,
      "logps/rejected": -243.2073211669922,
      "loss": 0.589,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.13480214774608612,
      "rewards/margins": 0.22513285279273987,
      "rewards/margins_max": 0.46415281295776367,
      "rewards/margins_min": 0.0476544126868248,
      "rewards/margins_std": 0.18787814676761627,
      "rewards/rejected": -0.09033068269491196,
      "step": 50
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.9422427549351036,
      "learning_rate": 4.93049306999712e-06,
      "logits/chosen": -2.6410093307495117,
      "logits/rejected": -2.60492205619812,
      "logps/chosen": -327.55523681640625,
      "logps/rejected": -283.3590393066406,
      "loss": 0.5384,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.17225147783756256,
      "rewards/margins": 0.3661409616470337,
      "rewards/margins_max": 0.6910489201545715,
      "rewards/margins_min": 0.1382865607738495,
      "rewards/margins_std": 0.2512677311897278,
      "rewards/rejected": -0.19388948380947113,
      "step": 60
    },
    {
      "epoch": 0.2,
      "grad_norm": 1.7976861864205922,
      "learning_rate": 4.861156761634014e-06,
      "logits/chosen": -2.649125576019287,
      "logits/rejected": -2.5835797786712646,
      "logps/chosen": -343.18719482421875,
      "logps/rejected": -268.24822998046875,
      "loss": 0.4917,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.19151589274406433,
      "rewards/margins": 0.5065498352050781,
      "rewards/margins_max": 0.9112545251846313,
      "rewards/margins_min": 0.11504770815372467,
      "rewards/margins_std": 0.35929277539253235,
      "rewards/rejected": -0.3150339722633362,
      "step": 70
    },
    {
      "epoch": 0.23,
      "grad_norm": 3.0636121326493675,
      "learning_rate": 4.7689385491773934e-06,
      "logits/chosen": -2.6317451000213623,
      "logits/rejected": -2.5868935585021973,
      "logps/chosen": -316.9160461425781,
      "logps/rejected": -335.57354736328125,
      "loss": 0.456,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 0.1493242084980011,
      "rewards/margins": 0.6301286220550537,
      "rewards/margins_max": 1.0892175436019897,
      "rewards/margins_min": 0.18430864810943604,
      "rewards/margins_std": 0.39657825231552124,
      "rewards/rejected": -0.480804443359375,
      "step": 80
    },
    {
      "epoch": 0.25,
      "grad_norm": 4.331028594376021,
      "learning_rate": 4.654732116743193e-06,
      "logits/chosen": -2.5327608585357666,
      "logits/rejected": -2.4845287799835205,
      "logps/chosen": -329.42620849609375,
      "logps/rejected": -258.33453369140625,
      "loss": 0.4162,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 0.09532937407493591,
      "rewards/margins": 0.688909649848938,
      "rewards/margins_max": 1.1923458576202393,
      "rewards/margins_min": 0.23197071254253387,
      "rewards/margins_std": 0.4409102499485016,
      "rewards/rejected": -0.5935802459716797,
      "step": 90
    },
    {
      "epoch": 0.28,
      "grad_norm": 4.295716285529308,
      "learning_rate": 4.5196442356717526e-06,
      "logits/chosen": -2.578570604324341,
      "logits/rejected": -2.512251377105713,
      "logps/chosen": -341.71697998046875,
      "logps/rejected": -342.2934875488281,
      "loss": 0.3843,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 0.04628221318125725,
      "rewards/margins": 0.8309832811355591,
      "rewards/margins_max": 1.4554868936538696,
      "rewards/margins_min": 0.29576951265335083,
      "rewards/margins_std": 0.5407849550247192,
      "rewards/rejected": -0.7847011089324951,
      "step": 100
    },
    {
      "epoch": 0.28,
      "eval_logits/chosen": -2.500088691711426,
      "eval_logits/rejected": -2.47070574760437,
      "eval_logps/chosen": -366.80853271484375,
      "eval_logps/rejected": -353.1435546875,
      "eval_loss": 0.6753506064414978,
      "eval_rewards/accuracies": 0.6190000176429749,
      "eval_rewards/chosen": -0.8221511244773865,
      "eval_rewards/margins": 0.1234951838850975,
      "eval_rewards/margins_max": 0.8275310397148132,
      "eval_rewards/margins_min": -0.5849089622497559,
      "eval_rewards/margins_std": 0.4660783112049103,
      "eval_rewards/rejected": -0.9456462860107422,
      "eval_runtime": 429.3559,
      "eval_samples_per_second": 4.658,
      "eval_steps_per_second": 0.291,
      "step": 100
    },
    {
      "epoch": 0.31,
      "grad_norm": 3.6576814376248663,
      "learning_rate": 4.364984038837727e-06,
      "logits/chosen": -2.486511707305908,
      "logits/rejected": -2.4051737785339355,
      "logps/chosen": -395.6208801269531,
      "logps/rejected": -412.60369873046875,
      "loss": 0.2943,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -0.04258393868803978,
      "rewards/margins": 1.210031509399414,
      "rewards/margins_max": 1.81746506690979,
      "rewards/margins_min": 0.560971736907959,
      "rewards/margins_std": 0.5801268815994263,
      "rewards/rejected": -1.2526154518127441,
      "step": 110
    },
    {
      "epoch": 0.34,
      "grad_norm": 4.445946405194985,
      "learning_rate": 4.192250333880045e-06,
      "logits/chosen": -2.3081154823303223,
      "logits/rejected": -2.2312095165252686,
      "logps/chosen": -416.22509765625,
      "logps/rejected": -439.51318359375,
      "loss": 0.262,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.21479575335979462,
      "rewards/margins": 1.514960527420044,
      "rewards/margins_max": 2.5221104621887207,
      "rewards/margins_min": 0.6053955554962158,
      "rewards/margins_std": 0.8898022770881653,
      "rewards/rejected": -1.7297561168670654,
      "step": 120
    },
    {
      "epoch": 0.37,
      "grad_norm": 5.213009355206989,
      "learning_rate": 4.0031170782990214e-06,
      "logits/chosen": -2.059696912765503,
      "logits/rejected": -1.9685862064361572,
      "logps/chosen": -424.81292724609375,
      "logps/rejected": -523.0667114257812,
      "loss": 0.1912,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.1958635300397873,
      "rewards/margins": 2.025548219680786,
      "rewards/margins_max": 3.0465281009674072,
      "rewards/margins_min": 0.8110028505325317,
      "rewards/margins_std": 0.9938048124313354,
      "rewards/rejected": -2.221411943435669,
      "step": 130
    },
    {
      "epoch": 0.39,
      "grad_norm": 13.540271738049956,
      "learning_rate": 3.7994171571810756e-06,
      "logits/chosen": -1.8918193578720093,
      "logits/rejected": -1.8057823181152344,
      "logps/chosen": -401.6873474121094,
      "logps/rejected": -562.7242431640625,
      "loss": 0.179,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.28990477323532104,
      "rewards/margins": 2.6347408294677734,
      "rewards/margins_max": 3.8110084533691406,
      "rewards/margins_min": 1.1343365907669067,
      "rewards/margins_std": 1.2381587028503418,
      "rewards/rejected": -2.9246459007263184,
      "step": 140
    },
    {
      "epoch": 0.42,
      "grad_norm": 7.258937456966438,
      "learning_rate": 3.5831246207606597e-06,
      "logits/chosen": -1.8240772485733032,
      "logits/rejected": -1.6942760944366455,
      "logps/chosen": -383.18658447265625,
      "logps/rejected": -533.005859375,
      "loss": 0.1794,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.6033543348312378,
      "rewards/margins": 2.5889029502868652,
      "rewards/margins_max": 3.9328532218933105,
      "rewards/margins_min": 0.870226263999939,
      "rewards/margins_std": 1.4039599895477295,
      "rewards/rejected": -3.1922574043273926,
      "step": 150
    },
    {
      "epoch": 0.45,
      "grad_norm": 3.520516887340939,
      "learning_rate": 3.3563355539546795e-06,
      "logits/chosen": -1.6218408346176147,
      "logits/rejected": -1.5248044729232788,
      "logps/chosen": -405.44110107421875,
      "logps/rejected": -626.1888427734375,
      "loss": 0.1195,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -0.5993050336837769,
      "rewards/margins": 3.3186564445495605,
      "rewards/margins_max": 4.61026668548584,
      "rewards/margins_min": 1.6127792596817017,
      "rewards/margins_std": 1.3561828136444092,
      "rewards/rejected": -3.917961835861206,
      "step": 160
    },
    {
      "epoch": 0.48,
      "grad_norm": 4.463403878446236,
      "learning_rate": 3.121247763262235e-06,
      "logits/chosen": -1.5979077816009521,
      "logits/rejected": -1.4862514734268188,
      "logps/chosen": -569.4307861328125,
      "logps/rejected": -859.6317138671875,
      "loss": 0.1156,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -1.9402544498443604,
      "rewards/margins": 3.6933372020721436,
      "rewards/margins_max": 5.283831596374512,
      "rewards/margins_min": 1.5569112300872803,
      "rewards/margins_std": 1.7149642705917358,
      "rewards/rejected": -5.6335906982421875,
      "step": 170
    },
    {
      "epoch": 0.51,
      "grad_norm": 28.94814953201888,
      "learning_rate": 2.8801394778833475e-06,
      "logits/chosen": -1.4478727579116821,
      "logits/rejected": -1.3353261947631836,
      "logps/chosen": -677.9209594726562,
      "logps/rejected": -972.3616943359375,
      "loss": 0.1518,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": -3.2994747161865234,
      "rewards/margins": 3.4859726428985596,
      "rewards/margins_max": 5.324586868286133,
      "rewards/margins_min": 1.277085781097412,
      "rewards/margins_std": 1.7865108251571655,
      "rewards/rejected": -6.785447597503662,
      "step": 180
    },
    {
      "epoch": 0.54,
      "grad_norm": 1.6663635534583416,
      "learning_rate": 2.6353472714635443e-06,
      "logits/chosen": -1.4343066215515137,
      "logits/rejected": -1.2333465814590454,
      "logps/chosen": -710.7890014648438,
      "logps/rejected": -980.9490966796875,
      "loss": 0.1188,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -3.5141806602478027,
      "rewards/margins": 3.9161250591278076,
      "rewards/margins_max": 5.42164421081543,
      "rewards/margins_min": 1.7522633075714111,
      "rewards/margins_std": 1.6805788278579712,
      "rewards/rejected": -7.430306434631348,
      "step": 190
    },
    {
      "epoch": 0.56,
      "grad_norm": 3.6276038453521187,
      "learning_rate": 2.3892434184240536e-06,
      "logits/chosen": -1.5112391710281372,
      "logits/rejected": -1.3403419256210327,
      "logps/chosen": -812.4998779296875,
      "logps/rejected": -1101.1927490234375,
      "loss": 0.1345,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -4.135049343109131,
      "rewards/margins": 4.216553688049316,
      "rewards/margins_max": 6.037824630737305,
      "rewards/margins_min": 1.9130373001098633,
      "rewards/margins_std": 1.9190574884414673,
      "rewards/rejected": -8.351602554321289,
      "step": 200
    },
    {
      "epoch": 0.56,
      "eval_logits/chosen": -1.4037277698516846,
      "eval_logits/rejected": -1.349421739578247,
      "eval_logps/chosen": -952.8915405273438,
      "eval_logps/rejected": -992.8120727539062,
      "eval_loss": 0.9115286469459534,
      "eval_rewards/accuracies": 0.6230000257492065,
      "eval_rewards/chosen": -6.682980537414551,
      "eval_rewards/margins": 0.659350574016571,
      "eval_rewards/margins_max": 4.129995822906494,
      "eval_rewards/margins_min": -2.920621633529663,
      "eval_rewards/margins_std": 2.361518383026123,
      "eval_rewards/rejected": -7.342331886291504,
      "eval_runtime": 428.8033,
      "eval_samples_per_second": 4.664,
      "eval_steps_per_second": 0.292,
      "step": 200
    },
    {
      "epoch": 0.59,
      "grad_norm": 10.435489666521553,
      "learning_rate": 2.1442129043167877e-06,
      "logits/chosen": -1.39103102684021,
      "logits/rejected": -1.270121455192566,
      "logps/chosen": -762.1986694335938,
      "logps/rejected": -1084.1378173828125,
      "loss": 0.1299,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -4.0687384605407715,
      "rewards/margins": 4.200930118560791,
      "rewards/margins_max": 6.173401355743408,
      "rewards/margins_min": 1.691877007484436,
      "rewards/margins_std": 2.071268320083618,
      "rewards/rejected": -8.269669532775879,
      "step": 210
    },
    {
      "epoch": 0.62,
      "grad_norm": 5.123588367601374,
      "learning_rate": 1.9026303129961049e-06,
      "logits/chosen": -1.5539300441741943,
      "logits/rejected": -1.3544563055038452,
      "logps/chosen": -777.8438720703125,
      "logps/rejected": -1124.8797607421875,
      "loss": 0.0627,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -3.7083256244659424,
      "rewards/margins": 4.7932538986206055,
      "rewards/margins_max": 6.416988372802734,
      "rewards/margins_min": 2.6955573558807373,
      "rewards/margins_std": 1.6872974634170532,
      "rewards/rejected": -8.501578330993652,
      "step": 220
    },
    {
      "epoch": 0.65,
      "grad_norm": 8.383127396811519,
      "learning_rate": 1.66683681459314e-06,
      "logits/chosen": -1.5905059576034546,
      "logits/rejected": -1.3816958665847778,
      "logps/chosen": -772.0947265625,
      "logps/rejected": -1100.8663330078125,
      "loss": 0.0918,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -3.6928131580352783,
      "rewards/margins": 4.681826591491699,
      "rewards/margins_max": 6.284341335296631,
      "rewards/margins_min": 2.5316035747528076,
      "rewards/margins_std": 1.6845054626464844,
      "rewards/rejected": -8.374638557434082,
      "step": 230
    },
    {
      "epoch": 0.68,
      "grad_norm": 4.5829488188390926,
      "learning_rate": 1.4391174773015836e-06,
      "logits/chosen": -1.5085848569869995,
      "logits/rejected": -1.3704946041107178,
      "logps/chosen": -735.5416870117188,
      "logps/rejected": -1119.05029296875,
      "loss": 0.131,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -3.923596143722534,
      "rewards/margins": 4.414877414703369,
      "rewards/margins_max": 6.264102935791016,
      "rewards/margins_min": 2.1018009185791016,
      "rewards/margins_std": 1.8714433908462524,
      "rewards/rejected": -8.338472366333008,
      "step": 240
    },
    {
      "epoch": 0.7,
      "grad_norm": 7.92768529410158,
      "learning_rate": 1.2216791228457778e-06,
      "logits/chosen": -1.5116816759109497,
      "logits/rejected": -1.3140555620193481,
      "logps/chosen": -774.6199951171875,
      "logps/rejected": -1130.3253173828125,
      "loss": 0.0892,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -4.107260704040527,
      "rewards/margins": 4.6648783683776855,
      "rewards/margins_max": 6.308585166931152,
      "rewards/margins_min": 2.563100814819336,
      "rewards/margins_std": 1.706753134727478,
      "rewards/rejected": -8.772139549255371,
      "step": 250
    },
    {
      "epoch": 0.73,
      "grad_norm": 5.013402333461111,
      "learning_rate": 1.0166289402331391e-06,
      "logits/chosen": -1.6035802364349365,
      "logits/rejected": -1.4094022512435913,
      "logps/chosen": -718.6414794921875,
      "logps/rejected": -1079.9404296875,
      "loss": 0.077,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -3.621960163116455,
      "rewards/margins": 4.5798468589782715,
      "rewards/margins_max": 6.199366569519043,
      "rewards/margins_min": 2.420649290084839,
      "rewards/margins_std": 1.7149250507354736,
      "rewards/rejected": -8.201807022094727,
      "step": 260
    },
    {
      "epoch": 0.76,
      "grad_norm": 16.373360821165523,
      "learning_rate": 8.259540650444736e-07,
      "logits/chosen": -1.555152177810669,
      "logits/rejected": -1.3883273601531982,
      "logps/chosen": -743.9249877929688,
      "logps/rejected": -1082.8458251953125,
      "loss": 0.0929,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -3.660247802734375,
      "rewards/margins": 4.5421624183654785,
      "rewards/margins_max": 6.060577392578125,
      "rewards/margins_min": 2.4010677337646484,
      "rewards/margins_std": 1.6667591333389282,
      "rewards/rejected": -8.202409744262695,
      "step": 270
    },
    {
      "epoch": 0.79,
      "grad_norm": 8.099720759346864,
      "learning_rate": 6.515023221586722e-07,
      "logits/chosen": -1.4892373085021973,
      "logits/rejected": -1.3762261867523193,
      "logps/chosen": -695.7076416015625,
      "logps/rejected": -1113.3646240234375,
      "loss": 0.0897,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -3.6505286693573,
      "rewards/margins": 4.7476959228515625,
      "rewards/margins_max": 6.505625247955322,
      "rewards/margins_min": 2.4996237754821777,
      "rewards/margins_std": 1.8462820053100586,
      "rewards/rejected": -8.398224830627441,
      "step": 280
    },
    {
      "epoch": 0.82,
      "grad_norm": 18.72504361278716,
      "learning_rate": 4.949643185335288e-07,
      "logits/chosen": -1.5151816606521606,
      "logits/rejected": -1.3798022270202637,
      "logps/chosen": -686.1121826171875,
      "logps/rejected": -1038.52587890625,
      "loss": 0.0905,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -3.451200485229492,
      "rewards/margins": 4.272095680236816,
      "rewards/margins_max": 6.312056064605713,
      "rewards/margins_min": 1.573736310005188,
      "rewards/margins_std": 2.1745951175689697,
      "rewards/rejected": -7.72329568862915,
      "step": 290
    },
    {
      "epoch": 0.85,
      "grad_norm": 9.116977457145628,
      "learning_rate": 3.578570595810274e-07,
      "logits/chosen": -1.5946242809295654,
      "logits/rejected": -1.4539456367492676,
      "logps/chosen": -693.6813354492188,
      "logps/rejected": -1122.047119140625,
      "loss": 0.0467,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -3.300471782684326,
      "rewards/margins": 5.025434970855713,
      "rewards/margins_max": 6.420098781585693,
      "rewards/margins_min": 2.8854596614837646,
      "rewards/margins_std": 1.5815579891204834,
      "rewards/rejected": -8.325906753540039,
      "step": 300
    },
    {
      "epoch": 0.85,
      "eval_logits/chosen": -1.4909454584121704,
      "eval_logits/rejected": -1.4320358037948608,
      "eval_logps/chosen": -888.703857421875,
      "eval_logps/rejected": -956.687744140625,
      "eval_loss": 0.9370979070663452,
      "eval_rewards/accuracies": 0.6449999809265137,
      "eval_rewards/chosen": -6.041103839874268,
      "eval_rewards/margins": 0.9399833083152771,
      "eval_rewards/margins_max": 4.743701457977295,
      "eval_rewards/margins_min": -3.0368335247039795,
      "eval_rewards/margins_std": 2.654953718185425,
      "eval_rewards/rejected": -6.981087684631348,
      "eval_runtime": 428.6815,
      "eval_samples_per_second": 4.665,
      "eval_steps_per_second": 0.292,
      "step": 300
    },
    {
      "epoch": 0.87,
      "grad_norm": 17.025809050618562,
      "learning_rate": 2.4150924791035037e-07,
      "logits/chosen": -1.5680278539657593,
      "logits/rejected": -1.3463890552520752,
      "logps/chosen": -778.5208740234375,
      "logps/rejected": -1089.8924560546875,
      "loss": 0.1036,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -4.222385406494141,
      "rewards/margins": 4.524166584014893,
      "rewards/margins_max": 6.391068458557129,
      "rewards/margins_min": 2.287738084793091,
      "rewards/margins_std": 1.8728872537612915,
      "rewards/rejected": -8.746552467346191,
      "step": 310
    },
    {
      "epoch": 0.9,
      "grad_norm": 6.614213043020825,
      "learning_rate": 1.4704840690808658e-07,
      "logits/chosen": -1.5289008617401123,
      "logits/rejected": -1.3871232271194458,
      "logps/chosen": -733.2238159179688,
      "logps/rejected": -1093.3983154296875,
      "loss": 0.0976,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -3.8258163928985596,
      "rewards/margins": 4.487856388092041,
      "rewards/margins_max": 6.392783164978027,
      "rewards/margins_min": 1.399547815322876,
      "rewards/margins_std": 2.305619955062866,
      "rewards/rejected": -8.313672065734863,
      "step": 320
    },
    {
      "epoch": 0.93,
      "grad_norm": 9.60113340680737,
      "learning_rate": 7.538995394063996e-08,
      "logits/chosen": -1.629804253578186,
      "logits/rejected": -1.444265604019165,
      "logps/chosen": -782.9251708984375,
      "logps/rejected": -1121.7064208984375,
      "loss": 0.1186,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -3.826475143432617,
      "rewards/margins": 4.706925392150879,
      "rewards/margins_max": 6.423586368560791,
      "rewards/margins_min": 2.4785187244415283,
      "rewards/margins_std": 1.876365065574646,
      "rewards/rejected": -8.53339958190918,
      "step": 330
    },
    {
      "epoch": 0.96,
      "grad_norm": 18.972390442985663,
      "learning_rate": 2.722832907015971e-08,
      "logits/chosen": -1.5121577978134155,
      "logits/rejected": -1.355576753616333,
      "logps/chosen": -770.6697387695312,
      "logps/rejected": -1123.355712890625,
      "loss": 0.0955,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -4.256497383117676,
      "rewards/margins": 4.456811428070068,
      "rewards/margins_max": 6.2082414627075195,
      "rewards/margins_min": 2.056246280670166,
      "rewards/margins_std": 1.880099892616272,
      "rewards/rejected": -8.713308334350586,
      "step": 340
    },
    {
      "epoch": 0.99,
      "grad_norm": 26.738917827064355,
      "learning_rate": 3.030265255329623e-09,
      "logits/chosen": -1.5162473917007446,
      "logits/rejected": -1.3702640533447266,
      "logps/chosen": -736.2762451171875,
      "logps/rejected": -1147.2891845703125,
      "loss": 0.1122,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -3.7701263427734375,
      "rewards/margins": 4.834524631500244,
      "rewards/margins_max": 6.308111667633057,
      "rewards/margins_min": 2.398153781890869,
      "rewards/margins_std": 1.7776079177856445,
      "rewards/rejected": -8.604650497436523,
      "step": 350
    },
    {
      "epoch": 1.0,
      "step": 355,
      "total_flos": 0.0,
      "train_loss": 0.24729210323011372,
      "train_runtime": 4030.3896,
      "train_samples_per_second": 1.409,
      "train_steps_per_second": 0.088
    }
  ],
  "logging_steps": 10,
  "max_steps": 355,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}