[
  {
    "entropy": 2.555550271272659,
    "epoch": 0.08,
    "grad_norm": 175.83624267578125,
    "learning_rate": 2.8571428571428573e-06,
    "logits/chosen": 0.41239441768472307,
    "logits/rejected": 0.4313717963372231,
    "logps/chosen": -157.81461162567138,
    "logps/rejected": -150.4321650505066,
    "loss": 0.6844395637512207,
    "mean_token_accuracy": 0.45540447160601616,
    "num_tokens": 63280.0,
    "rewards/accuracies": 0.55,
    "rewards/chosen": 0.09316946864128113,
    "rewards/margins": 0.030635011196136475,
    "rewards/rejected": 0.06253445744514466,
    "step": 5
  },
  {
    "entropy": 2.584770825505257,
    "epoch": 0.16,
    "grad_norm": 275.8374938964844,
    "learning_rate": 4.984280524733107e-06,
    "logits/chosen": 0.23018057449154425,
    "logits/rejected": 0.4137123508420314,
    "logps/chosen": -160.37082138061524,
    "logps/rejected": -154.9328401565552,
    "loss": 0.848213005065918,
    "mean_token_accuracy": 0.4679958797991276,
    "num_tokens": 128751.0,
    "rewards/accuracies": 0.4625,
    "rewards/chosen": 0.06873167753219604,
    "rewards/margins": -0.08100449442863464,
    "rewards/rejected": 0.1497361719608307,
    "step": 10
  },
  {
    "entropy": 2.485997956991196,
    "epoch": 0.24,
    "grad_norm": 273.0233154296875,
    "learning_rate": 4.809698831278217e-06,
    "logits/chosen": 0.1816923364201413,
    "logits/rejected": 0.23945612540369207,
    "logps/chosen": -151.38499155044556,
    "logps/rejected": -163.4970989227295,
    "loss": 0.8470439910888672,
    "mean_token_accuracy": 0.4733552895486355,
    "num_tokens": 194142.0,
    "rewards/accuracies": 0.45,
    "rewards/chosen": -0.32968448400497435,
    "rewards/margins": 0.06030769646167755,
    "rewards/rejected": -0.3899921804666519,
    "step": 15
  },
  {
    "entropy": 2.5318914532661436,
    "epoch": 0.32,
    "grad_norm": 234.37681579589844,
    "learning_rate": 4.454578706170075e-06,
    "logits/chosen": 0.07677000825058375,
    "logits/rejected": 0.18887137972641857,
    "logps/chosen": -135.0070219039917,
    "logps/rejected": -135.28493957519532,
    "loss": 0.9311827659606934,
    "mean_token_accuracy": 0.4595716767013073,
    "num_tokens": 254871.0,
    "rewards/accuracies": 0.5375,
    "rewards/chosen": -0.7181475728750228,
    "rewards/margins": 0.2766463726758957,
    "rewards/rejected": -0.9947939455509186,
    "step": 20
  },
  {
    "entropy": 2.462447080016136,
    "epoch": 0.4,
    "grad_norm": 250.9570770263672,
    "learning_rate": 3.946678240449515e-06,
    "logits/chosen": 0.20993061968672994,
    "logits/rejected": 0.20386893310256043,
    "logps/chosen": -145.39961681365966,
    "logps/rejected": -137.4315857887268,
    "loss": 0.9580174446105957,
    "mean_token_accuracy": 0.45824490338563917,
    "num_tokens": 317752.0,
    "rewards/accuracies": 0.58125,
    "rewards/chosen": -0.7164888560771943,
    "rewards/margins": 0.16717674732208251,
    "rewards/rejected": -0.8836656033992767,
    "step": 25
  },
  {
    "entropy": 2.4387545943260194,
    "epoch": 0.48,
    "grad_norm": 316.4103088378906,
    "learning_rate": 3.3256976548879183e-06,
    "logits/chosen": 0.21743581064338796,
    "logits/rejected": 0.2765391890690152,
    "logps/chosen": -148.74287595748902,
    "logps/rejected": -147.26203107833862,
    "loss": 0.9910862922668457,
    "mean_token_accuracy": 0.4532994642853737,
    "num_tokens": 387452.0,
    "rewards/accuracies": 0.5125,
    "rewards/chosen": -0.2023364007472992,
    "rewards/margins": 0.1426894724369049,
    "rewards/rejected": -0.3450258731842041,
    "step": 30
  },
  {
    "entropy": 2.469076004624367,
    "epoch": 0.56,
    "grad_norm": 256.6429748535156,
    "learning_rate": 2.6401761180929798e-06,
    "logits/chosen": 0.18299585651420763,
    "logits/rejected": 0.250682400907656,
    "logps/chosen": -178.9021216392517,
    "logps/rejected": -171.8313166618347,
    "loss": 0.8882085800170898,
    "mean_token_accuracy": 0.47426797077059746,
    "num_tokens": 454873.0,
    "rewards/accuracies": 0.53125,
    "rewards/chosen": 0.1502027690410614,
    "rewards/margins": 0.38659809827804564,
    "rewards/rejected": -0.23639532923698425,
    "step": 35
  },
  {
    "entropy": 2.4749947816133497,
    "epoch": 0.64,
    "grad_norm": 248.1305694580078,
    "learning_rate": 1.9436976651092143e-06,
    "logits/chosen": 0.27816104129128527,
    "logits/rejected": 0.37676442462284854,
    "logps/chosen": -157.81686353683472,
    "logps/rejected": -158.98215370178224,
    "loss": 1.0146113395690919,
    "mean_token_accuracy": 0.465312571823597,
    "num_tokens": 518527.0,
    "rewards/accuracies": 0.55625,
    "rewards/chosen": 0.05624905824661255,
    "rewards/margins": 0.09228863418102265,
    "rewards/rejected": -0.036039575934410095,
    "step": 40
  },
  {
    "entropy": 2.409931382536888,
    "epoch": 0.72,
    "grad_norm": 219.939697265625,
    "learning_rate": 1.2907027822369006e-06,
    "logits/chosen": 0.2784307162908892,
    "logits/rejected": 0.3215469827251049,
    "logps/chosen": -153.8018159866333,
    "logps/rejected": -159.60495262145997,
    "loss": 0.921458625793457,
    "mean_token_accuracy": 0.4758016161620617,
    "num_tokens": 586496.0,
    "rewards/accuracies": 0.525,
    "rewards/chosen": 0.1898827075958252,
    "rewards/margins": 0.46674283146858214,
    "rewards/rejected": -0.27686012387275694,
    "step": 45
  },
  {
    "entropy": 2.4379068195819853,
    "epoch": 0.8,
    "grad_norm": 235.46241760253906,
    "learning_rate": 7.322330470336314e-07,
    "logits/chosen": 0.28584471002719647,
    "logits/rejected": 0.3651064125555795,
    "logps/chosen": -163.8307611465454,
    "logps/rejected": -163.46687126159668,
    "loss": 0.979340648651123,
    "mean_token_accuracy": 0.4653128065168858,
    "num_tokens": 659345.0,
    "rewards/accuracies": 0.5,
    "rewards/chosen": 0.16986804008483886,
    "rewards/margins": 0.1913771450519562,
    "rewards/rejected": -0.02150910496711731,
    "step": 50
  },
  {
    "epoch": 0.8,
    "eval_entropy": 2.5069966416358946,
    "eval_logits/chosen": 0.29315065189355716,
    "eval_logits/rejected": 0.348470281243937,
    "eval_logps/chosen": -161.13176106262208,
    "eval_logps/rejected": -153.30834275054931,
    "eval_loss": 0.9624814391136169,
    "eval_mean_token_accuracy": 0.4618878421783447,
    "eval_num_tokens": 659345.0,
    "eval_rewards/accuracies": 0.531,
    "eval_rewards/chosen": 0.1580237216949463,
    "eval_rewards/margins": 0.2344533715248108,
    "eval_rewards/rejected": -0.0764296498298645,
    "eval_runtime": 203.9137,
    "eval_samples_per_second": 4.904,
    "eval_steps_per_second": 1.226,
    "step": 50
  },
  {
    "entropy": 2.47237910926342,
    "epoch": 0.88,
    "grad_norm": 254.37533569335938,
    "learning_rate": 3.119414452281158e-07,
    "logits/chosen": 0.31621869752899623,
    "logits/rejected": 0.438719071421892,
    "logps/chosen": -148.24002447128296,
    "logps/rejected": -142.998854637146,
    "loss": 0.9957857131958008,
    "mean_token_accuracy": 0.47029968202114103,
    "num_tokens": 731475.0,
    "rewards/accuracies": 0.59375,
    "rewards/chosen": 0.2521739423274994,
    "rewards/margins": 0.1846893012523651,
    "rewards/rejected": 0.06748464107513427,
    "step": 55
  },
  {
    "entropy": 2.4306333363056183,
    "epoch": 0.96,
    "grad_norm": 226.28863525390625,
    "learning_rate": 6.268021954544095e-08,
    "logits/chosen": 0.27655188516120244,
    "logits/rejected": 0.40121284602477314,
    "logps/chosen": -156.78616828918456,
    "logps/rejected": -163.7671283721924,
    "loss": 0.8905390739440918,
    "mean_token_accuracy": 0.4723482772707939,
    "num_tokens": 797757.0,
    "rewards/accuracies": 0.5125,
    "rewards/chosen": 0.2607013940811157,
    "rewards/margins": 0.30302430391311647,
    "rewards/rejected": -0.04232290983200073,
    "step": 60
  },
  {
    "epoch": 1.0,
    "eval_entropy": 2.5108700938224793,
    "eval_logits/chosen": 0.2960853689074952,
    "eval_logits/rejected": 0.35102944673757674,
    "eval_logps/chosen": -161.2097156829834,
    "eval_logps/rejected": -153.38722127532958,
    "eval_loss": 0.9581051468849182,
    "eval_mean_token_accuracy": 0.4621651805639267,
    "eval_num_tokens": 829784.0,
    "eval_rewards/accuracies": 0.532,
    "eval_rewards/chosen": 0.11904637241363525,
    "eval_rewards/margins": 0.23491560649871826,
    "eval_rewards/rejected": -0.11586923408508301,
    "eval_runtime": 204.736,
    "eval_samples_per_second": 4.884,
    "eval_steps_per_second": 1.221,
    "step": 63
  },
  {
    "loss": 0.20300202071666718,
    "grad_norm": 70.3385238647461,
    "learning_rate": 4.184239109116393e-06,
    "entropy": 2.4772405922412872,
    "num_tokens": 27180.0,
    "logits/chosen": 0.29999179148367466,
    "logits/rejected": 0.3366013686931765,
    "mean_token_accuracy": 0.4766918160021305,
    "rewards/chosen": 3.185014322400093,
    "rewards/rejected": -3.17814964056015,
    "rewards/accuracies": 0.84375,
    "rewards/margins": 6.363163933157921,
    "logps/chosen": -145.2724313735962,
    "logps/rejected": -168.38544821739197,
    "epoch": 1.032,
    "step": 65
  },
  {
    "loss": 0.16016263961791993,
    "grad_norm": 46.46247482299805,
    "learning_rate": 4.006586590948141e-06,
    "entropy": 2.5059203624725344,
    "num_tokens": 86611.0,
    "logits/chosen": 0.30248451255827935,
    "logits/rejected": 0.4369742989935694,
    "mean_token_accuracy": 0.48107142075896264,
    "rewards/chosen": 3.108510234951973,
    "rewards/rejected": -3.5679119408130644,
    "rewards/accuracies": 0.93125,
    "rewards/margins": 6.6764221549034115,
    "logps/chosen": -151.66757192611695,
    "logps/rejected": -152.82527809143068,
    "epoch": 1.112,
    "step": 70
  },
  {
    "loss": 0.1984729766845703,
    "grad_norm": 112.07616424560547,
    "learning_rate": 3.81608040719339e-06,
    "entropy": 2.5288314819335938,
    "num_tokens": 153056.0,
    "logits/chosen": 0.3381024272346584,
    "logits/rejected": 0.3919709001428062,
    "mean_token_accuracy": 0.47099064812064173,
    "rewards/chosen": 2.1314487934112547,
    "rewards/rejected": -3.7832603842020034,
    "rewards/accuracies": 0.88125,
    "rewards/margins": 5.91470921933651,
    "logps/chosen": -151.99142513275146,
    "logps/rejected": -158.68587474823,
    "epoch": 1.192,
    "step": 75
  },
  {
    "loss": 0.17381094694137572,
    "grad_norm": 47.97988510131836,
    "learning_rate": 3.6143458894413463e-06,
    "entropy": 2.5066609025001525,
    "num_tokens": 215920.0,
    "logits/chosen": 0.25372562736194504,
    "logits/rejected": 0.3085220085200435,
    "mean_token_accuracy": 0.47169380336999894,
    "rewards/chosen": 2.709118887782097,
    "rewards/rejected": -3.153385379910469,
    "rewards/accuracies": 0.9375,
    "rewards/margins": 5.862504243850708,
    "logps/chosen": -127.87930164337158,
    "logps/rejected": -152.39403762817383,
    "epoch": 1.272,
    "step": 80
  },
  {
    "loss": 0.17821238040924073,
    "grad_norm": 76.58686065673828,
    "learning_rate": 3.403104165467883e-06,
    "entropy": 2.42382150888443,
    "num_tokens": 287127.0,
    "logits/chosen": 0.27046587254706894,
    "logits/rejected": 0.3554108506103396,
    "mean_token_accuracy": 0.49741546660661695,
    "rewards/chosen": 3.4181615144014357,
    "rewards/rejected": -3.8993850499391556,
    "rewards/accuracies": 0.91875,
    "rewards/margins": 7.317546558380127,
    "logps/chosen": -135.95888204574584,
    "logps/rejected": -156.9404288291931,
    "epoch": 1.3519999999999999,
    "step": 85
  },
  {
    "loss": 0.23036706447601318,
    "grad_norm": 80.35875701904297,
    "learning_rate": 3.184157475180208e-06,
    "entropy": 2.5261152565479277,
    "num_tokens": 354068.0,
    "logits/chosen": 0.2524602257174569,
    "logits/rejected": 0.30296234191214516,
    "mean_token_accuracy": 0.46826066076755524,
    "rewards/chosen": 3.04692747592926,
    "rewards/rejected": -2.7697408348321915,
    "rewards/accuracies": 0.89375,
    "rewards/margins": 5.816668316721916,
    "logps/chosen": -156.0525354385376,
    "logps/rejected": -149.48046531677247,
    "epoch": 1.432,
    "step": 90
  },
  {
    "loss": 0.16570401191711426,
    "grad_norm": 53.557090759277344,
    "learning_rate": 2.9593737945414264e-06,
    "entropy": 2.4211883395910263,
    "num_tokens": 421426.0,
    "logits/chosen": 0.17738279076359578,
    "logits/rejected": 0.17275444698543332,
    "mean_token_accuracy": 0.4731420993804932,
    "rewards/chosen": 2.8382038921117783,
    "rewards/rejected": -3.731307566165924,
    "rewards/accuracies": 0.89375,
    "rewards/margins": 6.569511443376541,
    "logps/chosen": -135.7138165473938,
    "logps/rejected": -157.41348686218262,
    "epoch": 1.512,
    "step": 95
  },
  {
    "loss": 0.1649105429649353,
    "grad_norm": 127.4275131225586,
    "learning_rate": 2.730670898658255e-06,
    "entropy": 2.341481050848961,
    "num_tokens": 489806.0,
    "logits/chosen": -0.04766494180334789,
    "logits/rejected": 0.09169179261277503,
    "mean_token_accuracy": 0.4882186971604824,
    "rewards/chosen": 3.46675723195076,
    "rewards/rejected": -3.4928264021873474,
    "rewards/accuracies": 0.9125,
    "rewards/margins": 6.959583753347397,
    "logps/chosen": -153.73596954345703,
    "logps/rejected": -152.93934383392335,
    "epoch": 1.592,
    "step": 100
  },
  {
    "eval_loss": 1.181304931640625,
    "eval_runtime": 206.6722,
    "eval_samples_per_second": 4.839,
    "eval_steps_per_second": 1.21,
    "eval_entropy": 2.3993507494926454,
    "eval_num_tokens": 489806.0,
    "eval_logits/chosen": -0.03254038471881061,
    "eval_logits/rejected": 0.02311345798919562,
    "eval_mean_token_accuracy": 0.4609604501724243,
    "eval_rewards/chosen": -0.21504417657852173,
    "eval_rewards/rejected": -0.5070599136352539,
    "eval_rewards/accuracies": 0.516,
    "eval_rewards/margins": 0.29201573705673217,
    "eval_logps/chosen": -161.87789566040038,
    "eval_logps/rejected": -154.16960359954834,
    "epoch": 1.592,
    "step": 100
  },
  {
    "loss": 0.14087371826171874,
    "grad_norm": 88.25654602050781,
    "learning_rate": 2.5e-06,
    "entropy": 2.3751278162002563,
    "num_tokens": 557126.0,
    "logits/chosen": -0.06720577408088163,
    "logits/rejected": 0.009747601991268295,
    "mean_token_accuracy": 0.4826005406677723,
    "rewards/chosen": 4.133222645521164,
    "rewards/rejected": -4.368338066339493,
    "rewards/accuracies": 0.9375,
    "rewards/margins": 8.501560699939727,
    "logps/chosen": -163.05800094604493,
    "logps/rejected": -181.62881116867067,
    "epoch": 1.6720000000000002,
    "step": 105
  },
  {
    "loss": 0.23338742256164552,
    "grad_norm": 139.0855255126953,
    "learning_rate": 2.269329101341745e-06,
    "entropy": 2.292782101035118,
    "num_tokens": 621946.0,
    "logits/chosen": -0.1230251408579616,
    "logits/rejected": -0.04259267106090293,
    "mean_token_accuracy": 0.48731986358761786,
    "rewards/chosen": 3.1854370057582857,
    "rewards/rejected": -4.053957831859589,
    "rewards/accuracies": 0.8875,
    "rewards/margins": 7.239394760131836,
    "logps/chosen": -156.97511377334595,
    "logps/rejected": -170.4386715888977,
    "epoch": 1.752,
    "step": 110
  },
  {
    "loss": 0.24213719367980957,
    "grad_norm": 52.13258743286133,
    "learning_rate": 2.040626205458574e-06,
    "entropy": 2.2945144355297087,
    "num_tokens": 691164.0,
    "logits/chosen": -0.2037358547111102,
    "logits/rejected": -0.04959464698416915,
    "mean_token_accuracy": 0.48034665957093237,
    "rewards/chosen": 2.6465539157390596,
    "rewards/rejected": -3.8257445871829985,
    "rewards/accuracies": 0.89375,
    "rewards/margins": 6.472298520803451,
    "logps/chosen": -148.64922075271608,
    "logps/rejected": -163.68930807113648,
    "epoch": 1.8319999999999999,
    "step": 115
  },
  {
    "loss": 0.16733760833740235,
    "grad_norm": 51.697227478027344,
    "learning_rate": 1.8158425248197931e-06,
    "entropy": 2.318876361846924,
    "num_tokens": 754527.0,
    "logits/chosen": -0.16565151490742838,
    "logits/rejected": -0.09252247863799953,
    "mean_token_accuracy": 0.4730261914432049,
    "rewards/chosen": 2.554230135679245,
    "rewards/rejected": -3.980307912826538,
    "rewards/accuracies": 0.90625,
    "rewards/margins": 6.5345380246639255,
    "logps/chosen": -151.05957136154174,
    "logps/rejected": -158.38998804092407,
    "epoch": 1.912,
    "step": 120
  },
  {
    "loss": 0.17435957193374635,
    "grad_norm": 67.3870620727539,
    "learning_rate": 1.5968958345321178e-06,
    "entropy": 2.373268890380859,
    "num_tokens": 823843.0,
    "logits/chosen": -0.19111626640901203,
    "logits/rejected": -0.09068037643756863,
    "mean_token_accuracy": 0.473250250518322,
    "rewards/chosen": 2.5309324622154237,
    "rewards/rejected": -3.9597329974174498,
    "rewards/accuracies": 0.90625,
    "rewards/margins": 6.4906654238700865,
    "logps/chosen": -144.4349822998047,
    "logps/rejected": -157.0623517036438,
    "epoch": 1.992,
    "step": 125
  },
  {
    "loss": 0.08384315967559815,
    "grad_norm": 6.317872524261475,
    "learning_rate": 1.3856541105586545e-06,
    "entropy": 2.378200375371509,
    "num_tokens": 880621.0,
    "logits/chosen": -0.11781226721693444,
    "logits/rejected": -0.042003764714465604,
    "mean_token_accuracy": 0.49078691171275246,
    "rewards/chosen": 4.196374575297038,
    "rewards/rejected": -5.790639360745748,
    "rewards/accuracies": 0.9583333333333334,
    "rewards/margins": 9.98701392279731,
    "logps/chosen": -139.36247073279486,
    "logps/rejected": -160.00994406806097,
    "epoch": 2.064,
    "step": 130
  },
  {
    "loss": 0.0357888251543045,
    "grad_norm": 16.372520446777344,
    "learning_rate": 1.1839195928066101e-06,
    "entropy": 2.3539724469184877,
    "num_tokens": 947062.0,
    "logits/chosen": -0.1456372075386406,
    "logits/rejected": -0.059097048477018174,
    "mean_token_accuracy": 0.4876934640109539,
    "rewards/chosen": 3.8123921036720274,
    "rewards/rejected": -5.0069794178009035,
    "rewards/accuracies": 0.975,
    "rewards/margins": 8.819371503591537,
    "logps/chosen": -144.68906011581421,
    "logps/rejected": -159.5312638282776,
    "epoch": 2.144,
    "step": 135
  },
  {
    "loss": 0.0521980881690979,
    "grad_norm": 29.679048538208008,
    "learning_rate": 9.934134090518593e-07,
    "entropy": 2.3356225162744524,
    "num_tokens": 1009620.0,
    "logits/chosen": -0.11882373575890333,
    "logits/rejected": 0.026453084276548428,
    "mean_token_accuracy": 0.48251245468854903,
    "rewards/chosen": 3.927339309453964,
    "rewards/rejected": -5.288290357589721,
    "rewards/accuracies": 0.95625,
    "rewards/margins": 9.215629595518113,
    "logps/chosen": -139.31078090667725,
    "logps/rejected": -145.80226907730102,
    "epoch": 2.224,
    "step": 140
  },
  {
    "loss": 0.03299629390239715,
    "grad_norm": 7.268758773803711,
    "learning_rate": 8.157608908836071e-07,
    "entropy": 2.3158033907413484,
    "num_tokens": 1073365.0,
    "logits/chosen": -0.17821111338030823,
    "logits/rejected": -0.10465911260719843,
    "mean_token_accuracy": 0.48976081162691115,
    "rewards/chosen": 4.720877766609192,
    "rewards/rejected": -5.949343764781952,
    "rewards/accuracies": 0.98125,
    "rewards/margins": 10.670221543312072,
    "logps/chosen": -145.11655292510986,
    "logps/rejected": -164.69618825912477,
    "epoch": 2.304,
    "step": 145
  },
  {
    "loss": 0.045380374789237975,
    "grad_norm": 12.778029441833496,
    "learning_rate": 6.524777069483526e-07,
    "entropy": 2.3338232010602953,
    "num_tokens": 1145086.0,
    "logits/chosen": -0.24220225234879864,
    "logits/rejected": -0.18092035396691847,
    "mean_token_accuracy": 0.4738804578781128,
    "rewards/chosen": 4.283064597845078,
    "rewards/rejected": -6.082264894247055,
    "rewards/accuracies": 0.95625,
    "rewards/margins": 10.365329492092133,
    "logps/chosen": -149.46499347686768,
    "logps/rejected": -172.2482801437378,
    "epoch": 2.384,
    "step": 150
  },
  {
    "eval_loss": 1.1916899681091309,
    "eval_runtime": 206.7309,
    "eval_samples_per_second": 4.837,
    "eval_steps_per_second": 1.209,
    "eval_entropy": 2.3283381662368776,
    "eval_num_tokens": 1145086.0,
    "eval_logits/chosen": -0.24593508894178606,
    "eval_logits/rejected": -0.18872405137363696,
    "eval_mean_token_accuracy": 0.4613952031135559,
    "eval_rewards/chosen": -0.6765917901992797,
    "eval_rewards/rejected": -1.086084306716919,
    "eval_rewards/accuracies": 0.533,
    "eval_rewards/margins": 0.4094925169944763,
    "eval_logps/chosen": -162.80099214172364,
    "eval_logps/rejected": -155.3276520462036,
    "epoch": 2.384,
    "step": 150
  },
  {
    "loss": 0.03839964866638183,
    "grad_norm": 26.799619674682617,
    "learning_rate": 5.049569317994013e-07,
    "entropy": 2.218647238612175,
    "num_tokens": 1213916.0,
    "logits/chosen": -0.2961522508715828,
    "logits/rejected": -0.21192918570690328,
    "mean_token_accuracy": 0.49626709893345833,
    "rewards/chosen": 3.6871350944042205,
    "rewards/rejected": -6.713302159309388,
    "rewards/accuracies": 0.96875,
    "rewards/margins": 10.40043729543686,
    "logps/chosen": -138.59407653808594,
    "logps/rejected": -181.92112674713135,
    "epoch": 2.464,
    "step": 155
  },
  {
    "loss": 0.03359568119049072,
    "grad_norm": 6.236860275268555,
    "learning_rate": 3.7445716067596506e-07,
    "entropy": 2.3031299352645873,
    "num_tokens": 1277766.0,
    "logits/chosen": -0.2408471276221063,
    "logits/rejected": -0.05440791536781857,
    "mean_token_accuracy": 0.48194736018776896,
    "rewards/chosen": 4.138598304986954,
    "rewards/rejected": -5.22828494310379,
    "rewards/accuracies": 0.975,
    "rewards/margins": 9.36688324213028,
    "logps/chosen": -127.97663774490357,
    "logps/rejected": -142.3651433944702,
    "epoch": 2.544,
    "step": 160
  },
  {
    "loss": 0.03351616859436035,
    "grad_norm": 8.428302764892578,
    "learning_rate": 2.620917716123444e-07,
    "entropy": 2.2714572310447694,
    "num_tokens": 1344187.0,
    "logits/chosen": -0.30565182218347137,
    "logits/rejected": -0.262391552365428,
    "mean_token_accuracy": 0.4966980829834938,
    "rewards/chosen": 3.4794578850269318,
    "rewards/rejected": -5.821536821126938,
    "rewards/accuracies": 0.98125,
    "rewards/margins": 9.300994729995727,
    "logps/chosen": -143.35228729248047,
    "logps/rejected": -163.5957480430603,
    "epoch": 2.624,
    "step": 165
  },
  {
    "loss": 0.03787025213241577,
    "grad_norm": 19.33875846862793,
    "learning_rate": 1.6881942648911077e-07,
    "entropy": 2.319387698173523,
    "num_tokens": 1412024.0,
    "logits/chosen": -0.2771282747536266,
    "logits/rejected": -0.1903030416102311,
    "mean_token_accuracy": 0.483600977063179,
    "rewards/chosen": 3.7413900166749956,
    "rewards/rejected": -6.968815433979034,
    "rewards/accuracies": 0.96875,
    "rewards/margins": 10.71020541191101,
    "logps/chosen": -159.3884250640869,
    "logps/rejected": -202.81982517242432,
    "epoch": 2.7039999999999997,
    "step": 170
  },
  {
    "loss": 0.030017280578613283,
    "grad_norm": 8.236591339111328,
    "learning_rate": 9.54358920679524e-08,
    "entropy": 2.296932601928711,
    "num_tokens": 1479107.0,
    "logits/chosen": -0.35752807567610184,
    "logits/rejected": -0.2152659888252221,
    "mean_token_accuracy": 0.48842705860733987,
    "rewards/chosen": 5.196063238382339,
    "rewards/rejected": -5.935459566116333,
    "rewards/accuracies": 0.98125,
    "rewards/margins": 11.131522840261459,
    "logps/chosen": -162.68478546142578,
    "logps/rejected": -166.8199408531189,
    "epoch": 2.784,
    "step": 175
  },
  {
    "loss": 0.04067146480083465,
    "grad_norm": 4.00937557220459,
    "learning_rate": 4.256725079024554e-08,
    "entropy": 2.326069247722626,
    "num_tokens": 1546945.0,
    "logits/chosen": -0.2914971237007954,
    "logits/rejected": -0.18431776261809868,
    "mean_token_accuracy": 0.4746674746274948,
    "rewards/chosen": 4.784848666191101,
    "rewards/rejected": -5.148918730020523,
    "rewards/accuracies": 0.975,
    "rewards/margins": 9.933767330646514,
    "logps/chosen": -148.99251294136047,
    "logps/rejected": -147.76169719696045,
    "epoch": 2.864,
    "step": 180
  },
  {
    "loss": 0.03373933434486389,
    "grad_norm": 6.369950771331787,
    "learning_rate": 1.0664559262413831e-08,
    "entropy": 2.2894690483808517,
    "num_tokens": 1610498.0,
    "logits/chosen": -0.31424430259859776,
    "logits/rejected": -0.1595707523796258,
    "mean_token_accuracy": 0.4808969467878342,
    "rewards/chosen": 4.609399086236953,
    "rewards/rejected": -5.366133022308349,
    "rewards/accuracies": 0.96875,
    "rewards/margins": 9.97553214430809,
    "logps/chosen": -153.0141131401062,
    "logps/rejected": -162.40540452003478,
    "epoch": 2.944,
    "step": 185
  },
  {
    "eval_loss": 1.209433674812317,
    "eval_runtime": 207.0068,
    "eval_samples_per_second": 4.831,
    "eval_steps_per_second": 1.208,
    "eval_entropy": 2.320549825191498,
    "eval_num_tokens": 1659568.0,
    "eval_logits/chosen": -0.2871145028256165,
    "eval_logits/rejected": -0.23016112940510325,
    "eval_mean_token_accuracy": 0.4607901349067688,
    "eval_rewards/chosen": -0.8323136119842529,
    "eval_rewards/rejected": -1.2596224284172057,
    "eval_rewards/accuracies": 0.529,
    "eval_rewards/margins": 0.4273088164329529,
    "eval_logps/chosen": -163.11243586730956,
    "eval_logps/rejected": -155.6747285079956,
    "epoch": 3.0,
    "step": 189
  },
  {
    "train_runtime": 2676.0265,
    "train_samples_per_second": 2.242,
    "train_steps_per_second": 0.071,
    "total_flos": 9840760918671360.0,
    "train_loss": 0.07604808514080351,
    "epoch": 3.0,
    "step": 189
  }
]