{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9985693848354793,
  "eval_steps": 100,
  "global_step": 349,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 1482.8898632632786,
      "learning_rate": 1.4285714285714284e-09,
      "logits/chosen": -4.490396976470947,
      "logits/rejected": -4.787891387939453,
      "logps/chosen": -300.56573486328125,
      "logps/rejected": -263.39849853515625,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 1626.423546399318,
      "learning_rate": 1.4285714285714284e-08,
      "logits/chosen": -4.29224967956543,
      "logits/rejected": -4.469963550567627,
      "logps/chosen": -285.7841796875,
      "logps/rejected": -241.37472534179688,
      "loss": 0.7238,
      "rewards/accuracies": 0.4097222089767456,
      "rewards/chosen": 0.00654969597235322,
      "rewards/margins": -0.022695984691381454,
      "rewards/rejected": 0.029245682060718536,
      "step": 10
    },
    {
      "epoch": 0.06,
      "grad_norm": 1614.6738239528886,
      "learning_rate": 2.857142857142857e-08,
      "logits/chosen": -4.210062026977539,
      "logits/rejected": -4.502069473266602,
      "logps/chosen": -312.73968505859375,
      "logps/rejected": -247.76315307617188,
      "loss": 0.7059,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": 0.01807965151965618,
      "rewards/margins": 0.015493685379624367,
      "rewards/rejected": 0.0025859654415398836,
      "step": 20
    },
    {
      "epoch": 0.09,
      "grad_norm": 1444.832849324016,
      "learning_rate": 4.285714285714285e-08,
      "logits/chosen": -4.140606880187988,
      "logits/rejected": -4.359016418457031,
      "logps/chosen": -323.00701904296875,
      "logps/rejected": -274.5583801269531,
      "loss": 0.6754,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": 0.26756447553634644,
      "rewards/margins": 0.19910377264022827,
      "rewards/rejected": 0.06846068799495697,
      "step": 30
    },
    {
      "epoch": 0.11,
      "grad_norm": 1322.1199574632033,
      "learning_rate": 4.9968724814144246e-08,
      "logits/chosen": -4.296597480773926,
      "logits/rejected": -4.558178901672363,
      "logps/chosen": -277.67041015625,
      "logps/rejected": -227.9638214111328,
      "loss": 0.5887,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.46597638726234436,
      "rewards/margins": 0.30959779024124146,
      "rewards/rejected": 0.1563786268234253,
      "step": 40
    },
    {
      "epoch": 0.14,
      "grad_norm": 1226.7969075975607,
      "learning_rate": 4.971899263739325e-08,
      "logits/chosen": -4.339926719665527,
      "logits/rejected": -4.592724323272705,
      "logps/chosen": -270.9559020996094,
      "logps/rejected": -228.2990264892578,
      "loss": 0.5099,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.9494625329971313,
      "rewards/margins": 0.5203765034675598,
      "rewards/rejected": 0.4290861189365387,
      "step": 50
    },
    {
      "epoch": 0.17,
      "grad_norm": 929.2104534934765,
      "learning_rate": 4.9222026055025726e-08,
      "logits/chosen": -4.333067893981934,
      "logits/rejected": -4.564365386962891,
      "logps/chosen": -303.43310546875,
      "logps/rejected": -252.1361846923828,
      "loss": 0.4317,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 1.6878736019134521,
      "rewards/margins": 1.0692347288131714,
      "rewards/rejected": 0.6186389923095703,
      "step": 60
    },
    {
      "epoch": 0.2,
      "grad_norm": 1085.9024365118285,
      "learning_rate": 4.8482795627104736e-08,
      "logits/chosen": -4.332296371459961,
      "logits/rejected": -4.567011833190918,
      "logps/chosen": -289.8699645996094,
      "logps/rejected": -241.077880859375,
      "loss": 0.4089,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 1.8221668004989624,
      "rewards/margins": 1.204461693763733,
      "rewards/rejected": 0.6177049875259399,
      "step": 70
    },
    {
      "epoch": 0.23,
      "grad_norm": 978.9007166471121,
      "learning_rate": 4.750869498807735e-08,
      "logits/chosen": -4.28688383102417,
      "logits/rejected": -4.543013572692871,
      "logps/chosen": -322.56048583984375,
      "logps/rejected": -271.0233154296875,
      "loss": 0.3625,
      "rewards/accuracies": 0.875,
      "rewards/chosen": 2.1792995929718018,
      "rewards/margins": 1.4446518421173096,
      "rewards/rejected": 0.7346473932266235,
      "step": 80
    },
    {
      "epoch": 0.26,
      "grad_norm": 755.2675019751182,
      "learning_rate": 4.630946689712609e-08,
      "logits/chosen": -4.393240928649902,
      "logits/rejected": -4.636073112487793,
      "logps/chosen": -283.79290771484375,
      "logps/rejected": -234.5099334716797,
      "loss": 0.3436,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 2.397665500640869,
      "rewards/margins": 1.638594388961792,
      "rewards/rejected": 0.7590711116790771,
      "step": 90
    },
    {
      "epoch": 0.29,
      "grad_norm": 893.2972482207422,
      "learning_rate": 4.4897105793046024e-08,
      "logits/chosen": -4.302220344543457,
      "logits/rejected": -4.582569122314453,
      "logps/chosen": -270.37457275390625,
      "logps/rejected": -229.5946502685547,
      "loss": 0.3652,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 2.673680067062378,
      "rewards/margins": 2.020587921142578,
      "rewards/rejected": 0.6530919075012207,
      "step": 100
    },
    {
      "epoch": 0.29,
      "eval_logits/chosen": -4.696198463439941,
      "eval_logits/rejected": -4.924900054931641,
      "eval_logps/chosen": -403.4751892089844,
      "eval_logps/rejected": -519.0816650390625,
      "eval_loss": 1.7488452196121216,
      "eval_rewards/accuracies": 0.3515625,
      "eval_rewards/chosen": -2.2159295082092285,
      "eval_rewards/margins": -1.206247329711914,
      "eval_rewards/rejected": -1.009682059288025,
      "eval_runtime": 97.4797,
      "eval_samples_per_second": 20.517,
      "eval_steps_per_second": 0.328,
      "step": 100
    },
    {
      "epoch": 0.31,
      "grad_norm": 872.5882367590035,
      "learning_rate": 4.328573782827409e-08,
      "logits/chosen": -4.3329758644104,
      "logits/rejected": -4.617272853851318,
      "logps/chosen": -296.02618408203125,
      "logps/rejected": -242.73428344726562,
      "loss": 0.349,
      "rewards/accuracies": 0.875,
      "rewards/chosen": 2.7564496994018555,
      "rewards/margins": 2.076282024383545,
      "rewards/rejected": 0.6801677942276001,
      "step": 110
    },
    {
      "epoch": 0.34,
      "grad_norm": 1043.3393138533181,
      "learning_rate": 4.1491479581946166e-08,
      "logits/chosen": -4.3465352058410645,
      "logits/rejected": -4.575117111206055,
      "logps/chosen": -300.99859619140625,
      "logps/rejected": -265.1387023925781,
      "loss": 0.3567,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": 2.6209876537323,
      "rewards/margins": 1.951716423034668,
      "rewards/rejected": 0.6692714095115662,
      "step": 120
    },
    {
      "epoch": 0.37,
      "grad_norm": 837.2392359904699,
      "learning_rate": 3.953227686510564e-08,
      "logits/chosen": -4.319238185882568,
      "logits/rejected": -4.565236568450928,
      "logps/chosen": -288.36688232421875,
      "logps/rejected": -240.29177856445312,
      "loss": 0.3089,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 2.86527419090271,
      "rewards/margins": 2.168524980545044,
      "rewards/rejected": 0.6967490911483765,
      "step": 130
    },
    {
      "epoch": 0.4,
      "grad_norm": 1077.8316620275032,
      "learning_rate": 3.7427725230301354e-08,
      "logits/chosen": -4.310965538024902,
      "logits/rejected": -4.550887584686279,
      "logps/chosen": -295.83868408203125,
      "logps/rejected": -240.91311645507812,
      "loss": 0.3113,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 3.000100612640381,
      "rewards/margins": 2.261690855026245,
      "rewards/rejected": 0.7384099960327148,
      "step": 140
    },
    {
      "epoch": 0.43,
      "grad_norm": 805.166891537404,
      "learning_rate": 3.5198873980801955e-08,
      "logits/chosen": -4.3193488121032715,
      "logits/rejected": -4.625092506408691,
      "logps/chosen": -294.2840881347656,
      "logps/rejected": -245.18655395507812,
      "loss": 0.318,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 3.1946704387664795,
      "rewards/margins": 2.1761879920959473,
      "rewards/rejected": 1.0184824466705322,
      "step": 150
    },
    {
      "epoch": 0.46,
      "grad_norm": 711.9582620759091,
      "learning_rate": 3.2868015639687205e-08,
      "logits/chosen": -4.259413242340088,
      "logits/rejected": -4.450949668884277,
      "logps/chosen": -311.91326904296875,
      "logps/rejected": -270.4508972167969,
      "loss": 0.3344,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 3.2649455070495605,
      "rewards/margins": 2.298511266708374,
      "rewards/rejected": 0.966434121131897,
      "step": 160
    },
    {
      "epoch": 0.49,
      "grad_norm": 894.588925085696,
      "learning_rate": 3.0458462984504134e-08,
      "logits/chosen": -4.303664207458496,
      "logits/rejected": -4.500936031341553,
      "logps/chosen": -274.3706359863281,
      "logps/rejected": -235.48678588867188,
      "loss": 0.2963,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": 3.1083619594573975,
      "rewards/margins": 2.440159559249878,
      "rewards/rejected": 0.6682023406028748,
      "step": 170
    },
    {
      "epoch": 0.52,
      "grad_norm": 787.4594621489149,
      "learning_rate": 2.7994315877542628e-08,
      "logits/chosen": -4.386083602905273,
      "logits/rejected": -4.545100212097168,
      "logps/chosen": -263.3323059082031,
      "logps/rejected": -228.91329956054688,
      "loss": 0.329,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": 2.6448895931243896,
      "rewards/margins": 2.054766893386841,
      "rewards/rejected": 0.5901231169700623,
      "step": 180
    },
    {
      "epoch": 0.54,
      "grad_norm": 1024.0437070924656,
      "learning_rate": 2.5500220223847356e-08,
      "logits/chosen": -4.346238613128662,
      "logits/rejected": -4.559948444366455,
      "logps/chosen": -290.6450500488281,
      "logps/rejected": -252.0496063232422,
      "loss": 0.3385,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 3.277946949005127,
      "rewards/margins": 2.545849323272705,
      "rewards/rejected": 0.7320975661277771,
      "step": 190
    },
    {
      "epoch": 0.57,
      "grad_norm": 937.2977444102543,
      "learning_rate": 2.3001121467819626e-08,
      "logits/chosen": -4.290076732635498,
      "logits/rejected": -4.555208683013916,
      "logps/chosen": -317.6783447265625,
      "logps/rejected": -267.9818115234375,
      "loss": 0.312,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": 3.4544365406036377,
      "rewards/margins": 2.3504996299743652,
      "rewards/rejected": 1.1039369106292725,
      "step": 200
    },
    {
      "epoch": 0.57,
      "eval_logits/chosen": -4.7096052169799805,
      "eval_logits/rejected": -4.93910026550293,
      "eval_logps/chosen": -405.4331970214844,
      "eval_logps/rejected": -520.6950073242188,
      "eval_loss": 1.959585189819336,
      "eval_rewards/accuracies": 0.33984375,
      "eval_rewards/chosen": -3.194929599761963,
      "eval_rewards/margins": -1.3785794973373413,
      "eval_rewards/rejected": -1.8163501024246216,
      "eval_runtime": 97.467,
      "eval_samples_per_second": 20.52,
      "eval_steps_per_second": 0.328,
      "step": 200
    },
    {
      "epoch": 0.6,
      "grad_norm": 839.0813036142636,
      "learning_rate": 2.0522015093886612e-08,
      "logits/chosen": -4.317226409912109,
      "logits/rejected": -4.618459224700928,
      "logps/chosen": -294.03253173828125,
      "logps/rejected": -234.196533203125,
      "loss": 0.2938,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": 3.499882459640503,
      "rewards/margins": 2.7310853004455566,
      "rewards/rejected": 0.7687975168228149,
      "step": 210
    },
    {
      "epoch": 0.63,
      "grad_norm": 996.0408528817461,
      "learning_rate": 1.808769662668035e-08,
      "logits/chosen": -4.358494758605957,
      "logits/rejected": -4.704705238342285,
      "logps/chosen": -299.7410888671875,
      "logps/rejected": -235.092041015625,
      "loss": 0.2943,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 3.0801563262939453,
      "rewards/margins": 2.7920236587524414,
      "rewards/rejected": 0.28813308477401733,
      "step": 220
    },
    {
      "epoch": 0.66,
      "grad_norm": 781.9269603792492,
      "learning_rate": 1.5722513631174444e-08,
      "logits/chosen": -4.392434597015381,
      "logits/rejected": -4.748046398162842,
      "logps/chosen": -289.19927978515625,
      "logps/rejected": -234.593017578125,
      "loss": 0.3084,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": 2.994288921356201,
      "rewards/margins": 2.508836507797241,
      "rewards/rejected": 0.4854525625705719,
      "step": 230
    },
    {
      "epoch": 0.69,
      "grad_norm": 795.1407214961006,
      "learning_rate": 1.345012219322345e-08,
      "logits/chosen": -4.238420486450195,
      "logits/rejected": -4.548759460449219,
      "logps/chosen": -285.19232177734375,
      "logps/rejected": -240.72509765625,
      "loss": 0.3269,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": 2.8941564559936523,
      "rewards/margins": 2.3365750312805176,
      "rewards/rejected": 0.5575811266899109,
      "step": 240
    },
    {
      "epoch": 0.72,
      "grad_norm": 879.951904317175,
      "learning_rate": 1.1293250316137664e-08,
      "logits/chosen": -4.293431282043457,
      "logits/rejected": -4.5221476554870605,
      "logps/chosen": -294.2128601074219,
      "logps/rejected": -254.7281494140625,
      "loss": 0.3131,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": 3.089404821395874,
      "rewards/margins": 2.4765713214874268,
      "rewards/rejected": 0.6128337383270264,
      "step": 250
    },
    {
      "epoch": 0.74,
      "grad_norm": 1094.0422336259135,
      "learning_rate": 9.273470599753375e-09,
      "logits/chosen": -4.372658729553223,
      "logits/rejected": -4.525150775909424,
      "logps/chosen": -294.89874267578125,
      "logps/rejected": -263.60577392578125,
      "loss": 0.3409,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": 2.7965197563171387,
      "rewards/margins": 2.07285737991333,
      "rewards/rejected": 0.7236624956130981,
      "step": 260
    },
    {
      "epoch": 0.77,
      "grad_norm": 775.0184329679462,
      "learning_rate": 7.410984475616819e-09,
      "logits/chosen": -4.244694709777832,
      "logits/rejected": -4.5279622077941895,
      "logps/chosen": -282.0103454589844,
      "logps/rejected": -239.31399536132812,
      "loss": 0.2834,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": 3.1753017902374268,
      "rewards/margins": 2.7872982025146484,
      "rewards/rejected": 0.38800328969955444,
      "step": 270
    },
    {
      "epoch": 0.8,
      "grad_norm": 796.3667611732337,
      "learning_rate": 5.724420156318405e-09,
      "logits/chosen": -4.291975498199463,
      "logits/rejected": -4.621099472045898,
      "logps/chosen": -285.0769958496094,
      "logps/rejected": -232.0653839111328,
      "loss": 0.3144,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": 2.847846508026123,
      "rewards/margins": 2.3038220405578613,
      "rewards/rejected": 0.5440241098403931,
      "step": 280
    },
    {
      "epoch": 0.83,
      "grad_norm": 1101.8200803185337,
      "learning_rate": 4.230646319847259e-09,
      "logits/chosen": -4.424475193023682,
      "logits/rejected": -4.705140113830566,
      "logps/chosen": -295.53857421875,
      "logps/rejected": -232.05245971679688,
      "loss": 0.3121,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 2.7051048278808594,
      "rewards/margins": 2.4193003177642822,
      "rewards/rejected": 0.2858046591281891,
      "step": 290
    },
    {
      "epoch": 0.86,
      "grad_norm": 1165.1379404537536,
      "learning_rate": 2.944603392457931e-09,
      "logits/chosen": -4.335474491119385,
      "logits/rejected": -4.49271297454834,
      "logps/chosen": -292.6704406738281,
      "logps/rejected": -252.3407440185547,
      "loss": 0.2993,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": 3.1620662212371826,
      "rewards/margins": 2.2422008514404297,
      "rewards/rejected": 0.9198653101921082,
      "step": 300
    },
    {
      "epoch": 0.86,
      "eval_logits/chosen": -4.706819534301758,
      "eval_logits/rejected": -4.936363697052002,
      "eval_logps/chosen": -405.887939453125,
      "eval_logps/rejected": -521.2874755859375,
      "eval_loss": 1.9827619791030884,
      "eval_rewards/accuracies": 0.35546875,
      "eval_rewards/chosen": -3.4222922325134277,
      "eval_rewards/margins": -1.3097174167633057,
      "eval_rewards/rejected": -2.112574815750122,
      "eval_runtime": 97.4214,
      "eval_samples_per_second": 20.529,
      "eval_steps_per_second": 0.328,
      "step": 300
    },
    {
      "epoch": 0.89,
      "grad_norm": 1000.5341122694366,
      "learning_rate": 1.8791541175240787e-09,
      "logits/chosen": -4.3908562660217285,
      "logits/rejected": -4.607417583465576,
      "logps/chosen": -301.6288146972656,
      "logps/rejected": -246.2473907470703,
      "loss": 0.2942,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 3.3309147357940674,
      "rewards/margins": 2.577249765396118,
      "rewards/rejected": 0.753665030002594,
      "step": 310
    },
    {
      "epoch": 0.92,
      "grad_norm": 942.4369602130394,
      "learning_rate": 1.0449549049596136e-09,
      "logits/chosen": -4.230744361877441,
      "logits/rejected": -4.579751968383789,
      "logps/chosen": -287.4718322753906,
      "logps/rejected": -232.3734893798828,
      "loss": 0.3078,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 3.2846877574920654,
      "rewards/margins": 2.4689810276031494,
      "rewards/rejected": 0.8157066106796265,
      "step": 320
    },
    {
      "epoch": 0.94,
      "grad_norm": 1041.373917696444,
      "learning_rate": 4.5034924794443707e-10,
      "logits/chosen": -4.3931074142456055,
      "logits/rejected": -4.620477199554443,
      "logps/chosen": -280.0606994628906,
      "logps/rejected": -233.14346313476562,
      "loss": 0.2917,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": 3.2919604778289795,
      "rewards/margins": 2.5820116996765137,
      "rewards/rejected": 0.7099487781524658,
      "step": 330
    },
    {
      "epoch": 0.97,
      "grad_norm": 859.9709811153568,
      "learning_rate": 1.0128427297940723e-10,
      "logits/chosen": -4.328027248382568,
      "logits/rejected": -4.5136332511901855,
      "logps/chosen": -284.2748718261719,
      "logps/rejected": -244.03817749023438,
      "loss": 0.2952,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 3.204352855682373,
      "rewards/margins": 2.4567666053771973,
      "rewards/rejected": 0.7475861310958862,
      "step": 340
    },
    {
      "epoch": 1.0,
      "step": 349,
      "total_flos": 0.0,
      "train_loss": 0.3696309270011661,
      "train_runtime": 5591.2149,
      "train_samples_per_second": 7.991,
      "train_steps_per_second": 0.062
    }
  ],
  "logging_steps": 10,
  "max_steps": 349,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}