PEFT
Safetensors
FatDPOv2LoRA / trainer_state.json
jeiku's picture
Upload 8 files
ca2e945 verified
raw
history blame
220 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.0,
"eval_steps": 500,
"global_step": 422,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.004739336492890996,
"grad_norm": 4.468888282775879,
"learning_rate": 2.3255813953488374e-07,
"logits/chosen": 0.17294234037399292,
"logits/rejected": 0.25382864475250244,
"logps/chosen": -432.88006591796875,
"logps/rejected": -432.22393798828125,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.009478672985781991,
"grad_norm": 4.5246357917785645,
"learning_rate": 4.651162790697675e-07,
"logits/chosen": -0.02501661702990532,
"logits/rejected": -0.052543528378009796,
"logps/chosen": -435.35772705078125,
"logps/rejected": -450.81500244140625,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 2
},
{
"epoch": 0.014218009478672985,
"grad_norm": 4.940310001373291,
"learning_rate": 6.976744186046513e-07,
"logits/chosen": -0.32847464084625244,
"logits/rejected": -0.23367789387702942,
"logps/chosen": -383.3726806640625,
"logps/rejected": -430.9427185058594,
"loss": 0.689,
"rewards/accuracies": 0.40625,
"rewards/chosen": 0.022322356700897217,
"rewards/margins": 0.011036252602934837,
"rewards/rejected": 0.011286105029284954,
"step": 3
},
{
"epoch": 0.018957345971563982,
"grad_norm": 4.610801696777344,
"learning_rate": 9.30232558139535e-07,
"logits/chosen": -0.053772248327732086,
"logits/rejected": 0.035066261887550354,
"logps/chosen": -511.0668640136719,
"logps/rejected": -549.9104614257812,
"loss": 0.6915,
"rewards/accuracies": 0.40625,
"rewards/chosen": 0.03445637226104736,
"rewards/margins": 0.005329892970621586,
"rewards/rejected": 0.029126476496458054,
"step": 4
},
{
"epoch": 0.023696682464454975,
"grad_norm": 5.364634990692139,
"learning_rate": 1.1627906976744188e-06,
"logits/chosen": -0.0910184308886528,
"logits/rejected": -0.18476980924606323,
"logps/chosen": -398.50714111328125,
"logps/rejected": -442.03125,
"loss": 0.7036,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.030631696805357933,
"rewards/margins": -0.017725910991430283,
"rewards/rejected": -0.012905789539217949,
"step": 5
},
{
"epoch": 0.02843601895734597,
"grad_norm": 4.0496826171875,
"learning_rate": 1.3953488372093025e-06,
"logits/chosen": -0.08958485722541809,
"logits/rejected": -0.16801398992538452,
"logps/chosen": -450.0301818847656,
"logps/rejected": -480.5152282714844,
"loss": 0.7015,
"rewards/accuracies": 0.28125,
"rewards/chosen": -0.02475794591009617,
"rewards/margins": -0.015588914975523949,
"rewards/rejected": -0.009169031865894794,
"step": 6
},
{
"epoch": 0.03317535545023697,
"grad_norm": 4.628786563873291,
"learning_rate": 1.6279069767441862e-06,
"logits/chosen": 0.3181225061416626,
"logits/rejected": 0.4123896062374115,
"logps/chosen": -390.36358642578125,
"logps/rejected": -464.38818359375,
"loss": 0.6947,
"rewards/accuracies": 0.4375,
"rewards/chosen": 0.021739648655056953,
"rewards/margins": -0.00039539369754493237,
"rewards/rejected": 0.022135045379400253,
"step": 7
},
{
"epoch": 0.037914691943127965,
"grad_norm": 9.313066482543945,
"learning_rate": 1.86046511627907e-06,
"logits/chosen": 0.05047750473022461,
"logits/rejected": 0.2502543330192566,
"logps/chosen": -357.781982421875,
"logps/rejected": -437.6925964355469,
"loss": 0.6953,
"rewards/accuracies": 0.4375,
"rewards/chosen": 0.04442901909351349,
"rewards/margins": -0.0012866007164120674,
"rewards/rejected": 0.04571562260389328,
"step": 8
},
{
"epoch": 0.04265402843601896,
"grad_norm": 5.039402961730957,
"learning_rate": 2.0930232558139536e-06,
"logits/chosen": -0.023771263659000397,
"logits/rejected": -0.10481404513120651,
"logps/chosen": -551.095458984375,
"logps/rejected": -534.9522705078125,
"loss": 0.7109,
"rewards/accuracies": 0.28125,
"rewards/chosen": -0.014494383707642555,
"rewards/margins": -0.032524872571229935,
"rewards/rejected": 0.01803048700094223,
"step": 9
},
{
"epoch": 0.04739336492890995,
"grad_norm": 5.4466633796691895,
"learning_rate": 2.3255813953488376e-06,
"logits/chosen": 0.3669472336769104,
"logits/rejected": 0.3625236749649048,
"logps/chosen": -443.6736755371094,
"logps/rejected": -452.95770263671875,
"loss": 0.6875,
"rewards/accuracies": 0.4375,
"rewards/chosen": 1.2372620403766632e-05,
"rewards/margins": 0.013320828787982464,
"rewards/rejected": -0.013308453373610973,
"step": 10
},
{
"epoch": 0.052132701421800945,
"grad_norm": 4.648890495300293,
"learning_rate": 2.558139534883721e-06,
"logits/chosen": -0.03190188109874725,
"logits/rejected": -0.17422696948051453,
"logps/chosen": -453.6244812011719,
"logps/rejected": -391.9805603027344,
"loss": 0.6937,
"rewards/accuracies": 0.40625,
"rewards/chosen": 0.011450385674834251,
"rewards/margins": 0.0010257705580443144,
"rewards/rejected": 0.010424615815281868,
"step": 11
},
{
"epoch": 0.05687203791469194,
"grad_norm": 4.735479354858398,
"learning_rate": 2.790697674418605e-06,
"logits/chosen": 0.19764292240142822,
"logits/rejected": 0.31575778126716614,
"logps/chosen": -536.1707763671875,
"logps/rejected": -618.8275146484375,
"loss": 0.7078,
"rewards/accuracies": 0.28125,
"rewards/chosen": -0.008671605959534645,
"rewards/margins": -0.026794325560331345,
"rewards/rejected": 0.01812272146344185,
"step": 12
},
{
"epoch": 0.061611374407582936,
"grad_norm": 4.4631147384643555,
"learning_rate": 3.0232558139534885e-06,
"logits/chosen": 0.054879866540431976,
"logits/rejected": 0.17054584622383118,
"logps/chosen": -529.5116577148438,
"logps/rejected": -521.1355590820312,
"loss": 0.6873,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.0014204876497387886,
"rewards/margins": 0.014372622594237328,
"rewards/rejected": -0.015793107450008392,
"step": 13
},
{
"epoch": 0.06635071090047394,
"grad_norm": 5.378948211669922,
"learning_rate": 3.2558139534883724e-06,
"logits/chosen": 0.06459379941225052,
"logits/rejected": 0.09842963516712189,
"logps/chosen": -496.6351013183594,
"logps/rejected": -505.83819580078125,
"loss": 0.6943,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.006574093829840422,
"rewards/margins": 0.0018039373680949211,
"rewards/rejected": -0.008378027006983757,
"step": 14
},
{
"epoch": 0.07109004739336493,
"grad_norm": 3.6396803855895996,
"learning_rate": 3.4883720930232564e-06,
"logits/chosen": 0.30802494287490845,
"logits/rejected": 0.6311283111572266,
"logps/chosen": -366.74993896484375,
"logps/rejected": -442.3916015625,
"loss": 0.682,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.011980474926531315,
"rewards/margins": 0.023720718920230865,
"rewards/rejected": -0.011740244925022125,
"step": 15
},
{
"epoch": 0.07582938388625593,
"grad_norm": 4.317215919494629,
"learning_rate": 3.72093023255814e-06,
"logits/chosen": 0.19539549946784973,
"logits/rejected": 0.1992233693599701,
"logps/chosen": -458.35693359375,
"logps/rejected": -509.75177001953125,
"loss": 0.6703,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.0014730808325111866,
"rewards/margins": 0.048042502254247665,
"rewards/rejected": -0.049515582621097565,
"step": 16
},
{
"epoch": 0.08056872037914692,
"grad_norm": 4.411427974700928,
"learning_rate": 3.953488372093024e-06,
"logits/chosen": 0.2250944823026657,
"logits/rejected": 0.23612278699874878,
"logps/chosen": -529.0587158203125,
"logps/rejected": -525.65380859375,
"loss": 0.6898,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.037429191172122955,
"rewards/margins": 0.009371446445584297,
"rewards/rejected": -0.0468006357550621,
"step": 17
},
{
"epoch": 0.08530805687203792,
"grad_norm": 5.560946941375732,
"learning_rate": 4.186046511627907e-06,
"logits/chosen": 0.3321930170059204,
"logits/rejected": 0.4437577724456787,
"logps/chosen": -366.59173583984375,
"logps/rejected": -427.55413818359375,
"loss": 0.6641,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0035821665078401566,
"rewards/margins": 0.0638398677110672,
"rewards/rejected": -0.060257695615291595,
"step": 18
},
{
"epoch": 0.09004739336492891,
"grad_norm": 5.310560703277588,
"learning_rate": 4.418604651162791e-06,
"logits/chosen": 0.4229128360748291,
"logits/rejected": 0.4650568962097168,
"logps/chosen": -397.71112060546875,
"logps/rejected": -421.1392822265625,
"loss": 0.6813,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0423818975687027,
"rewards/margins": 0.025886641815304756,
"rewards/rejected": -0.0682685375213623,
"step": 19
},
{
"epoch": 0.0947867298578199,
"grad_norm": 4.838053226470947,
"learning_rate": 4.651162790697675e-06,
"logits/chosen": 0.2030269056558609,
"logits/rejected": -0.20078778266906738,
"logps/chosen": -507.3893737792969,
"logps/rejected": -482.47882080078125,
"loss": 0.6876,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.08994510769844055,
"rewards/margins": 0.015940451994538307,
"rewards/rejected": -0.10588555783033371,
"step": 20
},
{
"epoch": 0.0995260663507109,
"grad_norm": 3.838012456893921,
"learning_rate": 4.883720930232559e-06,
"logits/chosen": 0.024930402636528015,
"logits/rejected": 0.07886520028114319,
"logps/chosen": -404.5774230957031,
"logps/rejected": -484.91766357421875,
"loss": 0.6642,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.043739788234233856,
"rewards/margins": 0.0630723387002945,
"rewards/rejected": -0.10681212693452835,
"step": 21
},
{
"epoch": 0.10426540284360189,
"grad_norm": 4.912895202636719,
"learning_rate": 5.116279069767442e-06,
"logits/chosen": -0.024710316210985184,
"logits/rejected": 0.08874442428350449,
"logps/chosen": -380.9420166015625,
"logps/rejected": -390.6964111328125,
"loss": 0.6743,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.06719113886356354,
"rewards/margins": 0.0431598462164402,
"rewards/rejected": -0.11035098135471344,
"step": 22
},
{
"epoch": 0.10900473933649289,
"grad_norm": 4.68002462387085,
"learning_rate": 5.348837209302326e-06,
"logits/chosen": 0.09284278750419617,
"logits/rejected": 0.36523595452308655,
"logps/chosen": -291.48138427734375,
"logps/rejected": -371.36328125,
"loss": 0.6673,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.08336787670850754,
"rewards/margins": 0.06293603777885437,
"rewards/rejected": -0.1463039219379425,
"step": 23
},
{
"epoch": 0.11374407582938388,
"grad_norm": 4.751553058624268,
"learning_rate": 5.58139534883721e-06,
"logits/chosen": 0.1513872593641281,
"logits/rejected": 0.3874552845954895,
"logps/chosen": -542.8883056640625,
"logps/rejected": -584.90869140625,
"loss": 0.6544,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.14591416716575623,
"rewards/margins": 0.09063325077295303,
"rewards/rejected": -0.23654744029045105,
"step": 24
},
{
"epoch": 0.11848341232227488,
"grad_norm": 5.642929553985596,
"learning_rate": 5.8139534883720935e-06,
"logits/chosen": 0.34986090660095215,
"logits/rejected": 0.1881510615348816,
"logps/chosen": -462.16815185546875,
"logps/rejected": -504.80389404296875,
"loss": 0.7215,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.19018429517745972,
"rewards/margins": -0.04456442594528198,
"rewards/rejected": -0.14561986923217773,
"step": 25
},
{
"epoch": 0.12322274881516587,
"grad_norm": 4.584853172302246,
"learning_rate": 6.046511627906977e-06,
"logits/chosen": 0.5207738876342773,
"logits/rejected": 0.3684951066970825,
"logps/chosen": -432.26153564453125,
"logps/rejected": -423.8664855957031,
"loss": 0.6668,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.17351892590522766,
"rewards/margins": 0.06544799357652664,
"rewards/rejected": -0.2389669120311737,
"step": 26
},
{
"epoch": 0.12796208530805686,
"grad_norm": 4.37154483795166,
"learning_rate": 6.279069767441861e-06,
"logits/chosen": 0.17182780802249908,
"logits/rejected": 0.043800633400678635,
"logps/chosen": -356.12115478515625,
"logps/rejected": -354.2303466796875,
"loss": 0.672,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.14720980823040009,
"rewards/margins": 0.05135134607553482,
"rewards/rejected": -0.1985611617565155,
"step": 27
},
{
"epoch": 0.13270142180094788,
"grad_norm": 4.520308017730713,
"learning_rate": 6.511627906976745e-06,
"logits/chosen": 0.09184260666370392,
"logits/rejected": 0.19092530012130737,
"logps/chosen": -400.747314453125,
"logps/rejected": -500.5384521484375,
"loss": 0.6762,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.1759171187877655,
"rewards/margins": 0.05190613120794296,
"rewards/rejected": -0.22782325744628906,
"step": 28
},
{
"epoch": 0.13744075829383887,
"grad_norm": 4.618706226348877,
"learning_rate": 6.744186046511628e-06,
"logits/chosen": -0.016063084825873375,
"logits/rejected": -0.07409578561782837,
"logps/chosen": -472.5158996582031,
"logps/rejected": -468.94647216796875,
"loss": 0.626,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.18945825099945068,
"rewards/margins": 0.16357678174972534,
"rewards/rejected": -0.353035032749176,
"step": 29
},
{
"epoch": 0.14218009478672985,
"grad_norm": 4.5725998878479,
"learning_rate": 6.976744186046513e-06,
"logits/chosen": 0.35617199540138245,
"logits/rejected": 0.371913880109787,
"logps/chosen": -473.22821044921875,
"logps/rejected": -441.5536193847656,
"loss": 0.6542,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.2541047930717468,
"rewards/margins": 0.09789780527353287,
"rewards/rejected": -0.3520025908946991,
"step": 30
},
{
"epoch": 0.14691943127962084,
"grad_norm": 4.441779613494873,
"learning_rate": 7.209302325581395e-06,
"logits/chosen": -0.3232620358467102,
"logits/rejected": -0.3168506920337677,
"logps/chosen": -440.8768005371094,
"logps/rejected": -491.36517333984375,
"loss": 0.6467,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.1957191824913025,
"rewards/margins": 0.12279976159334183,
"rewards/rejected": -0.3185189366340637,
"step": 31
},
{
"epoch": 0.15165876777251186,
"grad_norm": 4.8312201499938965,
"learning_rate": 7.44186046511628e-06,
"logits/chosen": 0.20703125,
"logits/rejected": 0.158937007188797,
"logps/chosen": -474.72247314453125,
"logps/rejected": -430.9871520996094,
"loss": 0.6253,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.2362278252840042,
"rewards/margins": 0.1699885129928589,
"rewards/rejected": -0.4062163233757019,
"step": 32
},
{
"epoch": 0.15639810426540285,
"grad_norm": 4.127583026885986,
"learning_rate": 7.674418604651164e-06,
"logits/chosen": 0.012601375579833984,
"logits/rejected": 0.16285735368728638,
"logps/chosen": -512.3396606445312,
"logps/rejected": -529.702392578125,
"loss": 0.6475,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.2504887282848358,
"rewards/margins": 0.11251004040241241,
"rewards/rejected": -0.3629987835884094,
"step": 33
},
{
"epoch": 0.16113744075829384,
"grad_norm": 4.508236885070801,
"learning_rate": 7.906976744186048e-06,
"logits/chosen": -0.10328027606010437,
"logits/rejected": -0.07670910656452179,
"logps/chosen": -533.043701171875,
"logps/rejected": -608.9775390625,
"loss": 0.633,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.265808641910553,
"rewards/margins": 0.15493102371692657,
"rewards/rejected": -0.42073971033096313,
"step": 34
},
{
"epoch": 0.16587677725118483,
"grad_norm": 3.833841323852539,
"learning_rate": 8.139534883720931e-06,
"logits/chosen": -0.00557418167591095,
"logits/rejected": 0.06319032609462738,
"logps/chosen": -440.6729431152344,
"logps/rejected": -457.2427978515625,
"loss": 0.6497,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.27264466881752014,
"rewards/margins": 0.13150610029697418,
"rewards/rejected": -0.40415075421333313,
"step": 35
},
{
"epoch": 0.17061611374407584,
"grad_norm": 4.1322197914123535,
"learning_rate": 8.372093023255815e-06,
"logits/chosen": 0.3982032537460327,
"logits/rejected": 0.033034421503543854,
"logps/chosen": -511.9226989746094,
"logps/rejected": -457.24462890625,
"loss": 0.666,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.4206628203392029,
"rewards/margins": 0.09507569670677185,
"rewards/rejected": -0.5157385468482971,
"step": 36
},
{
"epoch": 0.17535545023696683,
"grad_norm": 4.012040138244629,
"learning_rate": 8.604651162790698e-06,
"logits/chosen": -0.22452551126480103,
"logits/rejected": 0.13141538202762604,
"logps/chosen": -365.06890869140625,
"logps/rejected": -484.9878845214844,
"loss": 0.6399,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.3754124045372009,
"rewards/margins": 0.17148733139038086,
"rewards/rejected": -0.5468997359275818,
"step": 37
},
{
"epoch": 0.18009478672985782,
"grad_norm": 5.076535701751709,
"learning_rate": 8.837209302325582e-06,
"logits/chosen": 0.38980191946029663,
"logits/rejected": 0.29863467812538147,
"logps/chosen": -453.6428527832031,
"logps/rejected": -450.8316650390625,
"loss": 0.6328,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.4603796899318695,
"rewards/margins": 0.19886130094528198,
"rewards/rejected": -0.6592409610748291,
"step": 38
},
{
"epoch": 0.1848341232227488,
"grad_norm": 4.119342803955078,
"learning_rate": 9.069767441860465e-06,
"logits/chosen": 0.32120779156684875,
"logits/rejected": 0.2722249925136566,
"logps/chosen": -403.55084228515625,
"logps/rejected": -382.61187744140625,
"loss": 0.6389,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.5036495923995972,
"rewards/margins": 0.18385529518127441,
"rewards/rejected": -0.6875049471855164,
"step": 39
},
{
"epoch": 0.1895734597156398,
"grad_norm": 3.443678379058838,
"learning_rate": 9.30232558139535e-06,
"logits/chosen": -0.06186838448047638,
"logits/rejected": 0.4023992419242859,
"logps/chosen": -342.36395263671875,
"logps/rejected": -402.0476379394531,
"loss": 0.5339,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.3065139651298523,
"rewards/margins": 0.5137544274330139,
"rewards/rejected": -0.820268452167511,
"step": 40
},
{
"epoch": 0.1943127962085308,
"grad_norm": 4.469974994659424,
"learning_rate": 9.534883720930234e-06,
"logits/chosen": -0.039235230535268784,
"logits/rejected": -0.04238155111670494,
"logps/chosen": -412.5050354003906,
"logps/rejected": -429.1283264160156,
"loss": 0.6047,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.3574320673942566,
"rewards/margins": 0.24463370442390442,
"rewards/rejected": -0.6020657420158386,
"step": 41
},
{
"epoch": 0.1990521327014218,
"grad_norm": 4.837284564971924,
"learning_rate": 9.767441860465117e-06,
"logits/chosen": -0.07826237380504608,
"logits/rejected": 0.1768716275691986,
"logps/chosen": -434.6158752441406,
"logps/rejected": -486.9299011230469,
"loss": 0.6364,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.5381902456283569,
"rewards/margins": 0.2264844924211502,
"rewards/rejected": -0.7646747827529907,
"step": 42
},
{
"epoch": 0.2037914691943128,
"grad_norm": 5.0225067138671875,
"learning_rate": 1e-05,
"logits/chosen": 0.0798925831913948,
"logits/rejected": -0.17932507395744324,
"logps/chosen": -346.35858154296875,
"logps/rejected": -313.78546142578125,
"loss": 0.7154,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.5803533792495728,
"rewards/margins": 0.07696369290351868,
"rewards/rejected": -0.657317042350769,
"step": 43
},
{
"epoch": 0.20853080568720378,
"grad_norm": 4.39346981048584,
"learning_rate": 9.999828225438954e-06,
"logits/chosen": -0.09884198009967804,
"logits/rejected": -0.10861906409263611,
"logps/chosen": -421.8805847167969,
"logps/rejected": -475.72491455078125,
"loss": 0.6138,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.5504233837127686,
"rewards/margins": 0.2702760398387909,
"rewards/rejected": -0.8206994533538818,
"step": 44
},
{
"epoch": 0.2132701421800948,
"grad_norm": 4.667287349700928,
"learning_rate": 9.999312913558414e-06,
"logits/chosen": 0.11334254592657089,
"logits/rejected": 0.03371644765138626,
"logps/chosen": -425.48028564453125,
"logps/rejected": -387.76605224609375,
"loss": 0.6278,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.6281574964523315,
"rewards/margins": 0.3120865821838379,
"rewards/rejected": -0.9402441382408142,
"step": 45
},
{
"epoch": 0.21800947867298578,
"grad_norm": 4.5937418937683105,
"learning_rate": 9.998454099765368e-06,
"logits/chosen": 0.04098134487867355,
"logits/rejected": 0.05865456163883209,
"logps/chosen": -364.27978515625,
"logps/rejected": -396.471923828125,
"loss": 0.6359,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.5031190514564514,
"rewards/margins": 0.2547236979007721,
"rewards/rejected": -0.7578427791595459,
"step": 46
},
{
"epoch": 0.22274881516587677,
"grad_norm": 4.146316051483154,
"learning_rate": 9.997251843068763e-06,
"logits/chosen": 0.25591805577278137,
"logits/rejected": -0.025539308786392212,
"logps/chosen": -524.7024536132812,
"logps/rejected": -446.9095458984375,
"loss": 0.6168,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.7542736530303955,
"rewards/margins": 0.2909476161003113,
"rewards/rejected": -1.045221209526062,
"step": 47
},
{
"epoch": 0.22748815165876776,
"grad_norm": 3.7932281494140625,
"learning_rate": 9.995706226075445e-06,
"logits/chosen": 0.14000652730464935,
"logits/rejected": 0.14868205785751343,
"logps/chosen": -457.1429138183594,
"logps/rejected": -432.461181640625,
"loss": 0.5377,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.6045001745223999,
"rewards/margins": 0.5118697881698608,
"rewards/rejected": -1.1163699626922607,
"step": 48
},
{
"epoch": 0.23222748815165878,
"grad_norm": 8.021098136901855,
"learning_rate": 9.993817354984485e-06,
"logits/chosen": 0.5763025283813477,
"logits/rejected": 0.5917288661003113,
"logps/chosen": -491.5912170410156,
"logps/rejected": -451.4367370605469,
"loss": 0.5349,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.5605940818786621,
"rewards/margins": 0.4851032793521881,
"rewards/rejected": -1.0456973314285278,
"step": 49
},
{
"epoch": 0.23696682464454977,
"grad_norm": 3.2265753746032715,
"learning_rate": 9.991585359579886e-06,
"logits/chosen": 0.3617466390132904,
"logits/rejected": 0.3173240125179291,
"logps/chosen": -372.00970458984375,
"logps/rejected": -400.85772705078125,
"loss": 0.5642,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.5665906667709351,
"rewards/margins": 0.4627717435359955,
"rewards/rejected": -1.029362440109253,
"step": 50
},
{
"epoch": 0.24170616113744076,
"grad_norm": 5.218678951263428,
"learning_rate": 9.989010393221657e-06,
"logits/chosen": 0.4278137981891632,
"logits/rejected": 0.45443713665008545,
"logps/chosen": -380.2688293457031,
"logps/rejected": -458.2431640625,
"loss": 0.6472,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.6341795921325684,
"rewards/margins": 0.33292362093925476,
"rewards/rejected": -0.9671032428741455,
"step": 51
},
{
"epoch": 0.24644549763033174,
"grad_norm": 3.540945053100586,
"learning_rate": 9.986092632835287e-06,
"logits/chosen": 0.16431818902492523,
"logits/rejected": 0.29724952578544617,
"logps/chosen": -462.51519775390625,
"logps/rejected": -506.53662109375,
"loss": 0.5851,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.5587910413742065,
"rewards/margins": 0.4125136137008667,
"rewards/rejected": -0.9713046550750732,
"step": 52
},
{
"epoch": 0.25118483412322273,
"grad_norm": 3.1407723426818848,
"learning_rate": 9.982832278899582e-06,
"logits/chosen": -0.12014055997133255,
"logits/rejected": -0.07937571406364441,
"logps/chosen": -293.3476867675781,
"logps/rejected": -329.0513916015625,
"loss": 0.4514,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.43457847833633423,
"rewards/margins": 0.7354463338851929,
"rewards/rejected": -1.1700247526168823,
"step": 53
},
{
"epoch": 0.2559241706161137,
"grad_norm": 4.521864891052246,
"learning_rate": 9.979229555432884e-06,
"logits/chosen": 0.311644583940506,
"logits/rejected": 0.13228081166744232,
"logps/chosen": -477.2836608886719,
"logps/rejected": -484.87908935546875,
"loss": 0.566,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.6064110994338989,
"rewards/margins": 0.5102828145027161,
"rewards/rejected": -1.1166939735412598,
"step": 54
},
{
"epoch": 0.26066350710900477,
"grad_norm": 4.260916233062744,
"learning_rate": 9.97528470997769e-06,
"logits/chosen": -0.25045445561408997,
"logits/rejected": -0.012767120264470577,
"logps/chosen": -327.21075439453125,
"logps/rejected": -342.33819580078125,
"loss": 0.5444,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.7582281231880188,
"rewards/margins": 0.687684178352356,
"rewards/rejected": -1.44591224193573,
"step": 55
},
{
"epoch": 0.26540284360189575,
"grad_norm": 4.670787811279297,
"learning_rate": 9.970998013583643e-06,
"logits/chosen": -0.09858112782239914,
"logits/rejected": 0.059129491448402405,
"logps/chosen": -434.63409423828125,
"logps/rejected": -482.1939392089844,
"loss": 0.6477,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.7279840111732483,
"rewards/margins": 0.4684455394744873,
"rewards/rejected": -1.1964294910430908,
"step": 56
},
{
"epoch": 0.27014218009478674,
"grad_norm": 4.042267799377441,
"learning_rate": 9.966369760788895e-06,
"logits/chosen": 0.23529131710529327,
"logits/rejected": 0.20713797211647034,
"logps/chosen": -387.02764892578125,
"logps/rejected": -347.3026428222656,
"loss": 0.5679,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.6105923652648926,
"rewards/margins": 0.441728413105011,
"rewards/rejected": -1.0523207187652588,
"step": 57
},
{
"epoch": 0.27488151658767773,
"grad_norm": 4.473219394683838,
"learning_rate": 9.961400269599885e-06,
"logits/chosen": 0.11737806349992752,
"logits/rejected": 0.036520302295684814,
"logps/chosen": -529.5386962890625,
"logps/rejected": -520.997314453125,
"loss": 0.6037,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.1079742908477783,
"rewards/margins": 0.503280758857727,
"rewards/rejected": -1.6112549304962158,
"step": 58
},
{
"epoch": 0.2796208530805687,
"grad_norm": 4.352895736694336,
"learning_rate": 9.956089881469481e-06,
"logits/chosen": 0.04756692051887512,
"logits/rejected": 0.06536822766065598,
"logps/chosen": -573.615478515625,
"logps/rejected": -529.3450317382812,
"loss": 0.6585,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.8045705556869507,
"rewards/margins": 0.36565208435058594,
"rewards/rejected": -1.1702226400375366,
"step": 59
},
{
"epoch": 0.2843601895734597,
"grad_norm": 5.498625755310059,
"learning_rate": 9.950438961273517e-06,
"logits/chosen": 0.075594462454319,
"logits/rejected": -0.22279205918312073,
"logps/chosen": -511.9619140625,
"logps/rejected": -507.88232421875,
"loss": 0.6781,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.8025746941566467,
"rewards/margins": 0.23016151785850525,
"rewards/rejected": -1.0327363014221191,
"step": 60
},
{
"epoch": 0.2890995260663507,
"grad_norm": 5.152298450469971,
"learning_rate": 9.94444789728573e-06,
"logits/chosen": -0.11181878298521042,
"logits/rejected": -0.005520373582839966,
"logps/chosen": -480.442138671875,
"logps/rejected": -491.8740539550781,
"loss": 0.5376,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.5731409788131714,
"rewards/margins": 0.5521252155303955,
"rewards/rejected": -1.125266194343567,
"step": 61
},
{
"epoch": 0.2938388625592417,
"grad_norm": 3.373650550842285,
"learning_rate": 9.93811710115107e-06,
"logits/chosen": 0.43766140937805176,
"logits/rejected": 0.4306361675262451,
"logps/chosen": -407.8697509765625,
"logps/rejected": -409.0340576171875,
"loss": 0.6231,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.5725124478340149,
"rewards/margins": 0.3360479176044464,
"rewards/rejected": -0.9085603952407837,
"step": 62
},
{
"epoch": 0.2985781990521327,
"grad_norm": 7.151122093200684,
"learning_rate": 9.931447007857433e-06,
"logits/chosen": 0.1370643973350525,
"logits/rejected": -0.04412531852722168,
"logps/chosen": -422.349853515625,
"logps/rejected": -445.34442138671875,
"loss": 0.6487,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6347966194152832,
"rewards/margins": 0.35178446769714355,
"rewards/rejected": -0.9865810871124268,
"step": 63
},
{
"epoch": 0.3033175355450237,
"grad_norm": 4.557085990905762,
"learning_rate": 9.924438075705755e-06,
"logits/chosen": 0.3025519847869873,
"logits/rejected": 0.3690996766090393,
"logps/chosen": -429.2871398925781,
"logps/rejected": -478.50494384765625,
"loss": 0.5381,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.5606861114501953,
"rewards/margins": 0.6541247963905334,
"rewards/rejected": -1.2148109674453735,
"step": 64
},
{
"epoch": 0.3080568720379147,
"grad_norm": 4.507608890533447,
"learning_rate": 9.917090786278533e-06,
"logits/chosen": 0.14490962028503418,
"logits/rejected": 0.20545634627342224,
"logps/chosen": -404.56536865234375,
"logps/rejected": -478.4001159667969,
"loss": 0.5636,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.4651917815208435,
"rewards/margins": 0.5367893576622009,
"rewards/rejected": -1.0019811391830444,
"step": 65
},
{
"epoch": 0.3127962085308057,
"grad_norm": 3.593722105026245,
"learning_rate": 9.909405644406738e-06,
"logits/chosen": -0.40012237429618835,
"logits/rejected": -0.3070281147956848,
"logps/chosen": -450.7489929199219,
"logps/rejected": -488.6943054199219,
"loss": 0.5341,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.6263750791549683,
"rewards/margins": 0.5151941776275635,
"rewards/rejected": -1.1415691375732422,
"step": 66
},
{
"epoch": 0.3175355450236967,
"grad_norm": 3.789606809616089,
"learning_rate": 9.901383178135113e-06,
"logits/chosen": 0.3894610106945038,
"logits/rejected": 0.41446149349212646,
"logps/chosen": -353.05316162109375,
"logps/rejected": -367.498046875,
"loss": 0.5364,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.4287835657596588,
"rewards/margins": 0.5223977565765381,
"rewards/rejected": -0.9511812925338745,
"step": 67
},
{
"epoch": 0.3222748815165877,
"grad_norm": 3.219656467437744,
"learning_rate": 9.893023938685911e-06,
"logits/chosen": -0.24614711105823517,
"logits/rejected": -0.00788339227437973,
"logps/chosen": -466.84930419921875,
"logps/rejected": -469.96142578125,
"loss": 0.5237,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.5285323858261108,
"rewards/margins": 0.5399156808853149,
"rewards/rejected": -1.0684479475021362,
"step": 68
},
{
"epoch": 0.32701421800947866,
"grad_norm": 3.697284698486328,
"learning_rate": 9.884328500421005e-06,
"logits/chosen": 0.18675777316093445,
"logits/rejected": 0.45374056696891785,
"logps/chosen": -363.5547790527344,
"logps/rejected": -370.5771484375,
"loss": 0.5697,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.3627626299858093,
"rewards/margins": 0.4581056833267212,
"rewards/rejected": -0.8208683133125305,
"step": 69
},
{
"epoch": 0.33175355450236965,
"grad_norm": 4.949876308441162,
"learning_rate": 9.87529746080243e-06,
"logits/chosen": 0.0644780769944191,
"logits/rejected": 0.23735983669757843,
"logps/chosen": -365.99298095703125,
"logps/rejected": -462.49224853515625,
"loss": 0.6512,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.4640830159187317,
"rewards/margins": 0.4212184548377991,
"rewards/rejected": -0.8853014707565308,
"step": 70
},
{
"epoch": 0.33649289099526064,
"grad_norm": 3.680051565170288,
"learning_rate": 9.865931440351338e-06,
"logits/chosen": 0.41640418767929077,
"logits/rejected": 0.5758565664291382,
"logps/chosen": -414.5774230957031,
"logps/rejected": -456.61993408203125,
"loss": 0.5691,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.7768192887306213,
"rewards/margins": 0.48619693517684937,
"rewards/rejected": -1.2630162239074707,
"step": 71
},
{
"epoch": 0.3412322274881517,
"grad_norm": 4.522833824157715,
"learning_rate": 9.856231082605344e-06,
"logits/chosen": 0.10547494888305664,
"logits/rejected": 0.06724017858505249,
"logps/chosen": -442.7686462402344,
"logps/rejected": -453.6602783203125,
"loss": 0.6052,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.5551941990852356,
"rewards/margins": 0.4050966501235962,
"rewards/rejected": -0.9602909088134766,
"step": 72
},
{
"epoch": 0.3459715639810427,
"grad_norm": 4.556945323944092,
"learning_rate": 9.846197054074325e-06,
"logits/chosen": 0.17173755168914795,
"logits/rejected": 0.1122516393661499,
"logps/chosen": -542.6156005859375,
"logps/rejected": -547.9420776367188,
"loss": 0.5518,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.42728012800216675,
"rewards/margins": 0.4563668966293335,
"rewards/rejected": -0.8836470246315002,
"step": 73
},
{
"epoch": 0.35071090047393366,
"grad_norm": 4.758415699005127,
"learning_rate": 9.835830044194625e-06,
"logits/chosen": 0.2476077526807785,
"logits/rejected": 0.2559727430343628,
"logps/chosen": -467.78192138671875,
"logps/rejected": -531.5208129882812,
"loss": 0.638,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.7260033488273621,
"rewards/margins": 0.5160312056541443,
"rewards/rejected": -1.2420345544815063,
"step": 74
},
{
"epoch": 0.35545023696682465,
"grad_norm": 7.977559566497803,
"learning_rate": 9.825130765281668e-06,
"logits/chosen": 0.034104928374290466,
"logits/rejected": 0.12088382244110107,
"logps/chosen": -359.2759704589844,
"logps/rejected": -388.21124267578125,
"loss": 0.472,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.19367767870426178,
"rewards/margins": 0.8202424049377441,
"rewards/rejected": -1.0139200687408447,
"step": 75
},
{
"epoch": 0.36018957345971564,
"grad_norm": 5.986490726470947,
"learning_rate": 9.814099952481031e-06,
"logits/chosen": -0.1111285462975502,
"logits/rejected": 0.2162112295627594,
"logps/chosen": -320.3255920410156,
"logps/rejected": -356.7198181152344,
"loss": 0.4873,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.21107947826385498,
"rewards/margins": 0.7759298086166382,
"rewards/rejected": -0.9870092868804932,
"step": 76
},
{
"epoch": 0.36492890995260663,
"grad_norm": 4.860304355621338,
"learning_rate": 9.802738363717927e-06,
"logits/chosen": 0.24542132019996643,
"logits/rejected": 0.26582974195480347,
"logps/chosen": -474.4150085449219,
"logps/rejected": -491.0259704589844,
"loss": 0.6354,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.4190457761287689,
"rewards/margins": 0.2866881489753723,
"rewards/rejected": -0.7057338953018188,
"step": 77
},
{
"epoch": 0.3696682464454976,
"grad_norm": 3.6817736625671387,
"learning_rate": 9.791046779645121e-06,
"logits/chosen": 0.18554499745368958,
"logits/rejected": 0.08219505846500397,
"logps/chosen": -367.29046630859375,
"logps/rejected": -377.1193542480469,
"loss": 0.5124,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.5033748149871826,
"rewards/margins": 0.595119833946228,
"rewards/rejected": -1.0984946489334106,
"step": 78
},
{
"epoch": 0.3744075829383886,
"grad_norm": 4.450536251068115,
"learning_rate": 9.779026003589304e-06,
"logits/chosen": 0.15589359402656555,
"logits/rejected": -0.018097057938575745,
"logps/chosen": -354.592041015625,
"logps/rejected": -387.45611572265625,
"loss": 0.6397,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.6710065603256226,
"rewards/margins": 0.43026500940322876,
"rewards/rejected": -1.101271629333496,
"step": 79
},
{
"epoch": 0.3791469194312796,
"grad_norm": 4.346735954284668,
"learning_rate": 9.766676861495888e-06,
"logits/chosen": 0.4525200128555298,
"logits/rejected": 0.45285752415657043,
"logps/chosen": -460.7939147949219,
"logps/rejected": -497.489501953125,
"loss": 0.5764,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.3665574789047241,
"rewards/margins": 0.5861377716064453,
"rewards/rejected": -0.9526952505111694,
"step": 80
},
{
"epoch": 0.38388625592417064,
"grad_norm": 4.672888278961182,
"learning_rate": 9.754000201872258e-06,
"logits/chosen": 0.2648187279701233,
"logits/rejected": 0.12692871689796448,
"logps/chosen": -462.68035888671875,
"logps/rejected": -446.92498779296875,
"loss": 0.5777,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.3308072090148926,
"rewards/margins": 0.4868035316467285,
"rewards/rejected": -0.8176107406616211,
"step": 81
},
{
"epoch": 0.3886255924170616,
"grad_norm": 4.8653130531311035,
"learning_rate": 9.74099689572947e-06,
"logits/chosen": 0.6338875889778137,
"logits/rejected": 0.5265121459960938,
"logps/chosen": -543.588623046875,
"logps/rejected": -460.85968017578125,
"loss": 0.4477,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.7309427857398987,
"rewards/margins": 0.8499249219894409,
"rewards/rejected": -1.5808677673339844,
"step": 82
},
{
"epoch": 0.3933649289099526,
"grad_norm": 3.210146188735962,
"learning_rate": 9.727667836522408e-06,
"logits/chosen": 0.1440345197916031,
"logits/rejected": 0.17331324517726898,
"logps/chosen": -403.3349609375,
"logps/rejected": -424.3764953613281,
"loss": 0.4817,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.47448769211769104,
"rewards/margins": 0.8300908803939819,
"rewards/rejected": -1.3045785427093506,
"step": 83
},
{
"epoch": 0.3981042654028436,
"grad_norm": 3.6252570152282715,
"learning_rate": 9.714013940088388e-06,
"logits/chosen": 0.35549432039260864,
"logits/rejected": 0.06034126877784729,
"logps/chosen": -417.7995300292969,
"logps/rejected": -471.5025634765625,
"loss": 0.5118,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.4528375267982483,
"rewards/margins": 0.6505048274993896,
"rewards/rejected": -1.1033422946929932,
"step": 84
},
{
"epoch": 0.4028436018957346,
"grad_norm": 3.3232247829437256,
"learning_rate": 9.700036144584237e-06,
"logits/chosen": 0.2271973192691803,
"logits/rejected": 0.47027313709259033,
"logps/chosen": -413.29510498046875,
"logps/rejected": -434.47119140625,
"loss": 0.4943,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6590039134025574,
"rewards/margins": 0.5970410108566284,
"rewards/rejected": -1.256044864654541,
"step": 85
},
{
"epoch": 0.4075829383886256,
"grad_norm": 5.408909320831299,
"learning_rate": 9.68573541042183e-06,
"logits/chosen": 0.13957802951335907,
"logits/rejected": -0.01698121801018715,
"logps/chosen": -397.6392517089844,
"logps/rejected": -405.05560302734375,
"loss": 0.6102,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.647673487663269,
"rewards/margins": 0.43150627613067627,
"rewards/rejected": -1.0791796445846558,
"step": 86
},
{
"epoch": 0.41232227488151657,
"grad_norm": 3.922076940536499,
"learning_rate": 9.6711127202021e-06,
"logits/chosen": 0.09419327974319458,
"logits/rejected": -0.07248637080192566,
"logps/chosen": -391.400390625,
"logps/rejected": -396.4415283203125,
"loss": 0.5324,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.48413023352622986,
"rewards/margins": 0.5888237953186035,
"rewards/rejected": -1.0729540586471558,
"step": 87
},
{
"epoch": 0.41706161137440756,
"grad_norm": 4.851089000701904,
"learning_rate": 9.656169078647525e-06,
"logits/chosen": 0.2503357231616974,
"logits/rejected": 0.17676730453968048,
"logps/chosen": -513.9971923828125,
"logps/rejected": -513.199951171875,
"loss": 0.6425,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.8412836194038391,
"rewards/margins": 0.38562363386154175,
"rewards/rejected": -1.2269072532653809,
"step": 88
},
{
"epoch": 0.4218009478672986,
"grad_norm": 3.372457504272461,
"learning_rate": 9.640905512533092e-06,
"logits/chosen": -0.17032505571842194,
"logits/rejected": -0.10938653349876404,
"logps/chosen": -418.69781494140625,
"logps/rejected": -375.2335205078125,
"loss": 0.487,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6325968503952026,
"rewards/margins": 0.7536253929138184,
"rewards/rejected": -1.3862221240997314,
"step": 89
},
{
"epoch": 0.4265402843601896,
"grad_norm": 3.9925975799560547,
"learning_rate": 9.625323070615751e-06,
"logits/chosen": 0.13802289962768555,
"logits/rejected": 0.20489519834518433,
"logps/chosen": -473.04534912109375,
"logps/rejected": -501.421630859375,
"loss": 0.5283,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.5786250829696655,
"rewards/margins": 0.5097320079803467,
"rewards/rejected": -1.0883569717407227,
"step": 90
},
{
"epoch": 0.4312796208530806,
"grad_norm": 4.2329936027526855,
"learning_rate": 9.609422823562345e-06,
"logits/chosen": 0.15958569943904877,
"logits/rejected": 0.17509707808494568,
"logps/chosen": -400.16668701171875,
"logps/rejected": -358.54327392578125,
"loss": 0.6406,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.5397723317146301,
"rewards/margins": 0.3339753746986389,
"rewards/rejected": -0.873747706413269,
"step": 91
},
{
"epoch": 0.43601895734597157,
"grad_norm": 3.645078182220459,
"learning_rate": 9.593205863876062e-06,
"logits/chosen": 0.5184372663497925,
"logits/rejected": 0.299396812915802,
"logps/chosen": -413.80816650390625,
"logps/rejected": -438.11663818359375,
"loss": 0.5111,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.656336784362793,
"rewards/margins": 0.5980467796325684,
"rewards/rejected": -1.2543835639953613,
"step": 92
},
{
"epoch": 0.44075829383886256,
"grad_norm": 5.40891695022583,
"learning_rate": 9.576673305821353e-06,
"logits/chosen": 0.4078008830547333,
"logits/rejected": 0.4366263449192047,
"logps/chosen": -335.3667297363281,
"logps/rejected": -352.97711181640625,
"loss": 0.6841,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.5545145273208618,
"rewards/margins": 0.23330146074295044,
"rewards/rejected": -0.7878159284591675,
"step": 93
},
{
"epoch": 0.44549763033175355,
"grad_norm": 3.772813558578491,
"learning_rate": 9.55982628534738e-06,
"logits/chosen": 0.04203655570745468,
"logits/rejected": 0.07998668402433395,
"logps/chosen": -416.1581115722656,
"logps/rejected": -376.9602355957031,
"loss": 0.4464,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.26440906524658203,
"rewards/margins": 0.8176214098930359,
"rewards/rejected": -1.0820305347442627,
"step": 94
},
{
"epoch": 0.45023696682464454,
"grad_norm": 4.10103702545166,
"learning_rate": 9.54266596000996e-06,
"logits/chosen": 0.3479623794555664,
"logits/rejected": 0.24994038045406342,
"logps/chosen": -473.52728271484375,
"logps/rejected": -449.1643981933594,
"loss": 0.6106,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.8664776086807251,
"rewards/margins": 0.31524401903152466,
"rewards/rejected": -1.1817216873168945,
"step": 95
},
{
"epoch": 0.4549763033175355,
"grad_norm": 5.940450668334961,
"learning_rate": 9.525193508892034e-06,
"logits/chosen": 0.12402056902647018,
"logits/rejected": 0.3821167051792145,
"logps/chosen": -390.4453430175781,
"logps/rejected": -498.83343505859375,
"loss": 0.5119,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.408942312002182,
"rewards/margins": 0.7732877135276794,
"rewards/rejected": -1.182229995727539,
"step": 96
},
{
"epoch": 0.4597156398104265,
"grad_norm": 4.347564220428467,
"learning_rate": 9.507410132522652e-06,
"logits/chosen": 0.3013518452644348,
"logits/rejected": 0.24353615939617157,
"logps/chosen": -519.27490234375,
"logps/rejected": -445.6244812011719,
"loss": 0.5246,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.5490106344223022,
"rewards/margins": 0.48622170090675354,
"rewards/rejected": -1.035232424736023,
"step": 97
},
{
"epoch": 0.46445497630331756,
"grad_norm": 5.856849670410156,
"learning_rate": 9.489317052794482e-06,
"logits/chosen": 0.18041574954986572,
"logits/rejected": 0.12899483740329742,
"logps/chosen": -535.3862915039062,
"logps/rejected": -527.10986328125,
"loss": 0.6128,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.6298852562904358,
"rewards/margins": 0.45580825209617615,
"rewards/rejected": -1.0856934785842896,
"step": 98
},
{
"epoch": 0.46919431279620855,
"grad_norm": 6.874017715454102,
"learning_rate": 9.470915512879853e-06,
"logits/chosen": 0.17756867408752441,
"logits/rejected": 0.1246391087770462,
"logps/chosen": -489.64434814453125,
"logps/rejected": -523.1498413085938,
"loss": 0.6378,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.9944641590118408,
"rewards/margins": 0.40292513370513916,
"rewards/rejected": -1.39738929271698,
"step": 99
},
{
"epoch": 0.47393364928909953,
"grad_norm": 4.015621185302734,
"learning_rate": 9.452206777145343e-06,
"logits/chosen": -0.0221431702375412,
"logits/rejected": 0.039132505655288696,
"logps/chosen": -505.1900634765625,
"logps/rejected": -460.162353515625,
"loss": 0.5233,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.8876866102218628,
"rewards/margins": 0.621982991695404,
"rewards/rejected": -1.509669542312622,
"step": 100
},
{
"epoch": 0.4786729857819905,
"grad_norm": 3.8857150077819824,
"learning_rate": 9.4331921310649e-06,
"logits/chosen": 0.14884430170059204,
"logits/rejected": 0.24638631939888,
"logps/chosen": -409.65289306640625,
"logps/rejected": -438.5373229980469,
"loss": 0.5,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.9228386878967285,
"rewards/margins": 0.7441995739936829,
"rewards/rejected": -1.6670382022857666,
"step": 101
},
{
"epoch": 0.4834123222748815,
"grad_norm": 3.3094451427459717,
"learning_rate": 9.413872881131519e-06,
"logits/chosen": 0.12251371890306473,
"logits/rejected": 0.2717350125312805,
"logps/chosen": -373.90777587890625,
"logps/rejected": -377.60546875,
"loss": 0.4469,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.9558136463165283,
"rewards/margins": 0.9420158863067627,
"rewards/rejected": -1.8978296518325806,
"step": 102
},
{
"epoch": 0.4881516587677725,
"grad_norm": 2.9940288066864014,
"learning_rate": 9.394250354767467e-06,
"logits/chosen": 0.00997384637594223,
"logits/rejected": -0.056838229298591614,
"logps/chosen": -479.9316711425781,
"logps/rejected": -432.07196044921875,
"loss": 0.3949,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.5534849166870117,
"rewards/margins": 1.0289510488510132,
"rewards/rejected": -1.5824360847473145,
"step": 103
},
{
"epoch": 0.4928909952606635,
"grad_norm": 3.61464524269104,
"learning_rate": 9.374325900233087e-06,
"logits/chosen": 0.5703778266906738,
"logits/rejected": 0.3040878474712372,
"logps/chosen": -341.7536926269531,
"logps/rejected": -341.8878173828125,
"loss": 0.4715,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.46594947576522827,
"rewards/margins": 0.9190239906311035,
"rewards/rejected": -1.3849735260009766,
"step": 104
},
{
"epoch": 0.4976303317535545,
"grad_norm": 4.028224945068359,
"learning_rate": 9.354100886534152e-06,
"logits/chosen": 0.11834748834371567,
"logits/rejected": 0.14537352323532104,
"logps/chosen": -376.5423583984375,
"logps/rejected": -432.7532958984375,
"loss": 0.482,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.4749223589897156,
"rewards/margins": 0.890184223651886,
"rewards/rejected": -1.365106463432312,
"step": 105
},
{
"epoch": 0.5023696682464455,
"grad_norm": 3.813199281692505,
"learning_rate": 9.333576703327803e-06,
"logits/chosen": 0.23954571783542633,
"logits/rejected": 0.18846619129180908,
"logps/chosen": -429.6502685546875,
"logps/rejected": -413.8288269042969,
"loss": 0.5218,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.8689689636230469,
"rewards/margins": 0.7078925371170044,
"rewards/rejected": -1.5768615007400513,
"step": 106
},
{
"epoch": 0.5071090047393365,
"grad_norm": 3.947744131088257,
"learning_rate": 9.312754760827061e-06,
"logits/chosen": 0.2891814112663269,
"logits/rejected": 0.10350443422794342,
"logps/chosen": -481.8363342285156,
"logps/rejected": -498.7795104980469,
"loss": 0.5892,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.8580895662307739,
"rewards/margins": 0.4509773850440979,
"rewards/rejected": -1.309066891670227,
"step": 107
},
{
"epoch": 0.5118483412322274,
"grad_norm": 9.869499206542969,
"learning_rate": 9.291636489703943e-06,
"logits/chosen": 0.04037293419241905,
"logits/rejected": 0.007937178015708923,
"logps/chosen": -313.94744873046875,
"logps/rejected": -337.9429931640625,
"loss": 0.5211,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6449707746505737,
"rewards/margins": 0.8841655254364014,
"rewards/rejected": -1.529136300086975,
"step": 108
},
{
"epoch": 0.5165876777251185,
"grad_norm": 2.261894464492798,
"learning_rate": 9.270223340991147e-06,
"logits/chosen": 0.45323365926742554,
"logits/rejected": 0.31425660848617554,
"logps/chosen": -518.0477905273438,
"logps/rejected": -518.2217407226562,
"loss": 0.4742,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.0188068151474,
"rewards/margins": 0.7373921871185303,
"rewards/rejected": -1.7561988830566406,
"step": 109
},
{
"epoch": 0.5213270142180095,
"grad_norm": 5.321746826171875,
"learning_rate": 9.248516785982365e-06,
"logits/chosen": 0.05109311267733574,
"logits/rejected": -0.22819462418556213,
"logps/chosen": -456.70703125,
"logps/rejected": -432.2109680175781,
"loss": 0.5835,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.9423930644989014,
"rewards/margins": 0.6178849935531616,
"rewards/rejected": -1.560278058052063,
"step": 110
},
{
"epoch": 0.5260663507109005,
"grad_norm": 4.055011749267578,
"learning_rate": 9.226518316131176e-06,
"logits/chosen": 0.47002896666526794,
"logits/rejected": 0.16273677349090576,
"logps/chosen": -509.3221740722656,
"logps/rejected": -459.65460205078125,
"loss": 0.6222,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.0381011962890625,
"rewards/margins": 0.38067197799682617,
"rewards/rejected": -1.4187732934951782,
"step": 111
},
{
"epoch": 0.5308056872037915,
"grad_norm": 3.9691321849823,
"learning_rate": 9.204229442948585e-06,
"logits/chosen": -0.06951727718114853,
"logits/rejected": 0.031339049339294434,
"logps/chosen": -489.317138671875,
"logps/rejected": -497.0957336425781,
"loss": 0.5213,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.6243175864219666,
"rewards/margins": 0.8498878479003906,
"rewards/rejected": -1.4742053747177124,
"step": 112
},
{
"epoch": 0.5355450236966824,
"grad_norm": 3.802368640899658,
"learning_rate": 9.181651697899153e-06,
"logits/chosen": 0.04360884055495262,
"logits/rejected": 0.08763141930103302,
"logps/chosen": -426.64227294921875,
"logps/rejected": -385.061767578125,
"loss": 0.4656,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.930860698223114,
"rewards/margins": 0.9945526719093323,
"rewards/rejected": -1.9254133701324463,
"step": 113
},
{
"epoch": 0.5402843601895735,
"grad_norm": 4.00538969039917,
"learning_rate": 9.158786632295776e-06,
"logits/chosen": -0.1742032915353775,
"logits/rejected": -0.051782816648483276,
"logps/chosen": -377.5306091308594,
"logps/rejected": -421.15216064453125,
"loss": 0.4133,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.657263457775116,
"rewards/margins": 0.9514760375022888,
"rewards/rejected": -1.6087393760681152,
"step": 114
},
{
"epoch": 0.5450236966824644,
"grad_norm": 2.9875290393829346,
"learning_rate": 9.1356358171931e-06,
"logits/chosen": -0.22956418991088867,
"logits/rejected": -0.03947896510362625,
"logps/chosen": -410.5157470703125,
"logps/rejected": -540.3042602539062,
"loss": 0.4268,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.0504915714263916,
"rewards/margins": 1.2017786502838135,
"rewards/rejected": -2.252270221710205,
"step": 115
},
{
"epoch": 0.5497630331753555,
"grad_norm": 4.0371994972229,
"learning_rate": 9.112200843279565e-06,
"logits/chosen": 0.2015472799539566,
"logits/rejected": 0.17825239896774292,
"logps/chosen": -471.33587646484375,
"logps/rejected": -473.25396728515625,
"loss": 0.5356,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.7962719202041626,
"rewards/margins": 0.7986223697662354,
"rewards/rejected": -1.5948941707611084,
"step": 116
},
{
"epoch": 0.5545023696682464,
"grad_norm": 3.3941633701324463,
"learning_rate": 9.088483320768114e-06,
"logits/chosen": -0.3230453431606293,
"logits/rejected": 0.14231722056865692,
"logps/chosen": -340.79547119140625,
"logps/rejected": -417.64764404296875,
"loss": 0.3766,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.7756860256195068,
"rewards/margins": 1.1981326341629028,
"rewards/rejected": -1.9738185405731201,
"step": 117
},
{
"epoch": 0.5592417061611374,
"grad_norm": 3.545902729034424,
"learning_rate": 9.064484879285555e-06,
"logits/chosen": 0.16888631880283356,
"logits/rejected": 0.36799800395965576,
"logps/chosen": -433.0151062011719,
"logps/rejected": -458.9261474609375,
"loss": 0.4182,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6435146331787109,
"rewards/margins": 1.1051644086837769,
"rewards/rejected": -1.7486791610717773,
"step": 118
},
{
"epoch": 0.5639810426540285,
"grad_norm": 4.029533386230469,
"learning_rate": 9.040207167760587e-06,
"logits/chosen": 0.1876976191997528,
"logits/rejected": 0.3228274881839752,
"logps/chosen": -434.99456787109375,
"logps/rejected": -496.84783935546875,
"loss": 0.5507,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.1236001253128052,
"rewards/margins": 0.6825075149536133,
"rewards/rejected": -1.8061076402664185,
"step": 119
},
{
"epoch": 0.5687203791469194,
"grad_norm": 6.328832149505615,
"learning_rate": 9.015651854310505e-06,
"logits/chosen": 0.5520612001419067,
"logits/rejected": 0.27829062938690186,
"logps/chosen": -524.2474975585938,
"logps/rejected": -500.3922424316406,
"loss": 0.5677,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.6885144710540771,
"rewards/margins": 0.6914829015731812,
"rewards/rejected": -1.3799974918365479,
"step": 120
},
{
"epoch": 0.5734597156398105,
"grad_norm": 4.104017734527588,
"learning_rate": 8.99082062612659e-06,
"logits/chosen": -0.059406764805316925,
"logits/rejected": 0.06676661223173141,
"logps/chosen": -397.0128173828125,
"logps/rejected": -465.90789794921875,
"loss": 0.4868,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.9093624353408813,
"rewards/margins": 0.9140886068344116,
"rewards/rejected": -1.823451042175293,
"step": 121
},
{
"epoch": 0.5781990521327014,
"grad_norm": 6.223032474517822,
"learning_rate": 8.965715189358165e-06,
"logits/chosen": 0.4450455605983734,
"logits/rejected": 0.40869832038879395,
"logps/chosen": -459.8601989746094,
"logps/rejected": -437.349365234375,
"loss": 0.6882,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.1986238956451416,
"rewards/margins": 0.5860114097595215,
"rewards/rejected": -1.784635305404663,
"step": 122
},
{
"epoch": 0.5829383886255924,
"grad_norm": 3.415480613708496,
"learning_rate": 8.940337268995385e-06,
"logits/chosen": 0.1984516978263855,
"logits/rejected": 0.016264721751213074,
"logps/chosen": -404.98858642578125,
"logps/rejected": -388.9734191894531,
"loss": 0.4402,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.8752626776695251,
"rewards/margins": 1.1214944124221802,
"rewards/rejected": -1.9967570304870605,
"step": 123
},
{
"epoch": 0.5876777251184834,
"grad_norm": 5.029397010803223,
"learning_rate": 8.9146886087507e-06,
"logits/chosen": 0.2871186137199402,
"logits/rejected": 0.08225645869970322,
"logps/chosen": -431.2898254394531,
"logps/rejected": -411.8453063964844,
"loss": 0.598,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.2498584985733032,
"rewards/margins": 0.6237522959709167,
"rewards/rejected": -1.8736108541488647,
"step": 124
},
{
"epoch": 0.5924170616113744,
"grad_norm": 4.441456317901611,
"learning_rate": 8.888770970939058e-06,
"logits/chosen": 0.1208953708410263,
"logits/rejected": 0.29052025079727173,
"logps/chosen": -439.34161376953125,
"logps/rejected": -485.6873474121094,
"loss": 0.6003,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.4302982091903687,
"rewards/margins": 0.5271378755569458,
"rewards/rejected": -1.9574360847473145,
"step": 125
},
{
"epoch": 0.5971563981042654,
"grad_norm": 4.362582206726074,
"learning_rate": 8.862586136356794e-06,
"logits/chosen": 0.03537209331989288,
"logits/rejected": 0.15606579184532166,
"logps/chosen": -459.3066711425781,
"logps/rejected": -456.8779296875,
"loss": 0.569,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.2706105709075928,
"rewards/margins": 0.9128097891807556,
"rewards/rejected": -2.183420419692993,
"step": 126
},
{
"epoch": 0.6018957345971564,
"grad_norm": 4.520688056945801,
"learning_rate": 8.836135904159301e-06,
"logits/chosen": 0.5592459440231323,
"logits/rejected": 0.4826202690601349,
"logps/chosen": -376.41552734375,
"logps/rejected": -411.979248046875,
"loss": 0.5533,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.8318498134613037,
"rewards/margins": 0.8411949276924133,
"rewards/rejected": -1.6730448007583618,
"step": 127
},
{
"epoch": 0.6066350710900474,
"grad_norm": 3.8013787269592285,
"learning_rate": 8.809422091737388e-06,
"logits/chosen": 0.3936067521572113,
"logits/rejected": 0.41032177209854126,
"logps/chosen": -373.95660400390625,
"logps/rejected": -375.69482421875,
"loss": 0.4864,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.8840177655220032,
"rewards/margins": 1.040097951889038,
"rewards/rejected": -1.9241158962249756,
"step": 128
},
{
"epoch": 0.6113744075829384,
"grad_norm": 3.8055856227874756,
"learning_rate": 8.782446534592412e-06,
"logits/chosen": 0.2286073863506317,
"logits/rejected": 0.13278520107269287,
"logps/chosen": -407.1571350097656,
"logps/rejected": -483.6879577636719,
"loss": 0.4939,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.2148370742797852,
"rewards/margins": 0.8403700590133667,
"rewards/rejected": -2.0552072525024414,
"step": 129
},
{
"epoch": 0.6161137440758294,
"grad_norm": 5.259015083312988,
"learning_rate": 8.755211086210172e-06,
"logits/chosen": 0.014815211296081543,
"logits/rejected": -0.05246584862470627,
"logps/chosen": -524.09765625,
"logps/rejected": -512.9163208007812,
"loss": 0.5758,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.9052078723907471,
"rewards/margins": 0.7172214984893799,
"rewards/rejected": -1.6224294900894165,
"step": 130
},
{
"epoch": 0.6208530805687204,
"grad_norm": 3.168894052505493,
"learning_rate": 8.727717617933545e-06,
"logits/chosen": 0.14947810769081116,
"logits/rejected": 0.01599368453025818,
"logps/chosen": -430.6431579589844,
"logps/rejected": -431.24945068359375,
"loss": 0.3968,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.7660319805145264,
"rewards/margins": 1.2241827249526978,
"rewards/rejected": -1.9902147054672241,
"step": 131
},
{
"epoch": 0.6255924170616114,
"grad_norm": 2.955793857574463,
"learning_rate": 8.699968018833903e-06,
"logits/chosen": 0.14900101721286774,
"logits/rejected": 0.20070943236351013,
"logps/chosen": -505.23480224609375,
"logps/rejected": -445.3937683105469,
"loss": 0.4073,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6804631352424622,
"rewards/margins": 1.0907790660858154,
"rewards/rejected": -1.7712422609329224,
"step": 132
},
{
"epoch": 0.6303317535545023,
"grad_norm": 3.240260601043701,
"learning_rate": 8.671964195581336e-06,
"logits/chosen": 0.03239504247903824,
"logits/rejected": 0.16187995672225952,
"logps/chosen": -396.4449157714844,
"logps/rejected": -417.7769775390625,
"loss": 0.413,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.2402898073196411,
"rewards/margins": 1.028062105178833,
"rewards/rejected": -2.2683520317077637,
"step": 133
},
{
"epoch": 0.6350710900473934,
"grad_norm": 3.1110308170318604,
"learning_rate": 8.643708072313619e-06,
"logits/chosen": 0.13824498653411865,
"logits/rejected": 0.1213301569223404,
"logps/chosen": -312.6982421875,
"logps/rejected": -343.7561340332031,
"loss": 0.403,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.7657801508903503,
"rewards/margins": 1.0914647579193115,
"rewards/rejected": -1.8572447299957275,
"step": 134
},
{
"epoch": 0.6398104265402843,
"grad_norm": 4.998546600341797,
"learning_rate": 8.615201590504016e-06,
"logits/chosen": 0.2914201617240906,
"logits/rejected": 0.4888171851634979,
"logps/chosen": -459.3593444824219,
"logps/rejected": -461.03265380859375,
"loss": 0.5944,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.0463979244232178,
"rewards/margins": 0.7969912886619568,
"rewards/rejected": -1.8433892726898193,
"step": 135
},
{
"epoch": 0.6445497630331753,
"grad_norm": 4.108184337615967,
"learning_rate": 8.586446708827896e-06,
"logits/chosen": -0.007564418017864227,
"logits/rejected": -0.18753255903720856,
"logps/chosen": -378.04345703125,
"logps/rejected": -370.3771667480469,
"loss": 0.5053,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.8995454907417297,
"rewards/margins": 0.9091691374778748,
"rewards/rejected": -1.808714747428894,
"step": 136
},
{
"epoch": 0.6492890995260664,
"grad_norm": 5.3410868644714355,
"learning_rate": 8.557445403028122e-06,
"logits/chosen": 0.23545342683792114,
"logits/rejected": 0.17830106616020203,
"logps/chosen": -533.0670776367188,
"logps/rejected": -545.6083984375,
"loss": 0.5424,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.9214141368865967,
"rewards/margins": 0.982886016368866,
"rewards/rejected": -1.9043000936508179,
"step": 137
},
{
"epoch": 0.6540284360189573,
"grad_norm": 4.749429702758789,
"learning_rate": 8.52819966577933e-06,
"logits/chosen": -0.03985929116606712,
"logits/rejected": 0.04942977800965309,
"logps/chosen": -502.4162902832031,
"logps/rejected": -525.4476318359375,
"loss": 0.5552,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.466578483581543,
"rewards/margins": 0.7347130179405212,
"rewards/rejected": -2.201291561126709,
"step": 138
},
{
"epoch": 0.6587677725118484,
"grad_norm": 4.427668571472168,
"learning_rate": 8.498711506550984e-06,
"logits/chosen": 0.10724569857120514,
"logits/rejected": -0.042100317776203156,
"logps/chosen": -424.52679443359375,
"logps/rejected": -472.065673828125,
"loss": 0.5214,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.4550511837005615,
"rewards/margins": 1.1595511436462402,
"rewards/rejected": -2.6146023273468018,
"step": 139
},
{
"epoch": 0.6635071090047393,
"grad_norm": 3.2623369693756104,
"learning_rate": 8.468982951469334e-06,
"logits/chosen": 0.20815303921699524,
"logits/rejected": -0.07932104170322418,
"logps/chosen": -477.3085021972656,
"logps/rejected": -457.29376220703125,
"loss": 0.4059,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.5089505910873413,
"rewards/margins": 1.101161241531372,
"rewards/rejected": -2.610111713409424,
"step": 140
},
{
"epoch": 0.6682464454976303,
"grad_norm": 3.9398345947265625,
"learning_rate": 8.439016043178176e-06,
"logits/chosen": -0.0940229594707489,
"logits/rejected": -0.044844143092632294,
"logps/chosen": -409.72442626953125,
"logps/rejected": -431.5579833984375,
"loss": 0.5502,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.1987154483795166,
"rewards/margins": 1.0584969520568848,
"rewards/rejected": -2.2572121620178223,
"step": 141
},
{
"epoch": 0.6729857819905213,
"grad_norm": 3.487725019454956,
"learning_rate": 8.408812840698517e-06,
"logits/chosen": -0.07787841558456421,
"logits/rejected": 0.11167508363723755,
"logps/chosen": -436.72723388671875,
"logps/rejected": -441.3120422363281,
"loss": 0.4822,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.9513049125671387,
"rewards/margins": 0.7828549742698669,
"rewards/rejected": -1.7341598272323608,
"step": 142
},
{
"epoch": 0.6777251184834123,
"grad_norm": 4.247493743896484,
"learning_rate": 8.378375419287098e-06,
"logits/chosen": 0.11863535642623901,
"logits/rejected": 0.010739833116531372,
"logps/chosen": -420.63818359375,
"logps/rejected": -442.8919372558594,
"loss": 0.4843,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.8487941026687622,
"rewards/margins": 0.7895632386207581,
"rewards/rejected": -1.6383572816848755,
"step": 143
},
{
"epoch": 0.6824644549763034,
"grad_norm": 3.8593363761901855,
"learning_rate": 8.3477058702938e-06,
"logits/chosen": 0.14162854850292206,
"logits/rejected": 0.2802544832229614,
"logps/chosen": -383.4293518066406,
"logps/rejected": -507.2069091796875,
"loss": 0.495,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.2597415447235107,
"rewards/margins": 0.9453345537185669,
"rewards/rejected": -2.205075979232788,
"step": 144
},
{
"epoch": 0.6872037914691943,
"grad_norm": 5.035678386688232,
"learning_rate": 8.31680630101795e-06,
"logits/chosen": 0.09797637909650803,
"logits/rejected": -0.0621592253446579,
"logps/chosen": -503.2251281738281,
"logps/rejected": -537.2774047851562,
"loss": 0.5595,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.707611083984375,
"rewards/margins": 0.8160245418548584,
"rewards/rejected": -2.5236356258392334,
"step": 145
},
{
"epoch": 0.6919431279620853,
"grad_norm": 4.702147006988525,
"learning_rate": 8.285678834563524e-06,
"logits/chosen": 0.025235813111066818,
"logits/rejected": 0.17205679416656494,
"logps/chosen": -457.6421203613281,
"logps/rejected": -525.4061279296875,
"loss": 0.5071,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.136285424232483,
"rewards/margins": 1.050948143005371,
"rewards/rejected": -2.1872334480285645,
"step": 146
},
{
"epoch": 0.6966824644549763,
"grad_norm": 4.684539794921875,
"learning_rate": 8.25432560969328e-06,
"logits/chosen": 0.06013096123933792,
"logits/rejected": -0.04088731110095978,
"logps/chosen": -354.15350341796875,
"logps/rejected": -332.03009033203125,
"loss": 0.4914,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.3446829319000244,
"rewards/margins": 0.8111345767974854,
"rewards/rejected": -2.1558172702789307,
"step": 147
},
{
"epoch": 0.7014218009478673,
"grad_norm": 4.876381874084473,
"learning_rate": 8.22274878068179e-06,
"logits/chosen": 0.2597839832305908,
"logits/rejected": 0.029643792659044266,
"logps/chosen": -407.7253112792969,
"logps/rejected": -370.39093017578125,
"loss": 0.5041,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.9906654953956604,
"rewards/margins": 0.9019131660461426,
"rewards/rejected": -1.8925787210464478,
"step": 148
},
{
"epoch": 0.7061611374407583,
"grad_norm": 4.655327796936035,
"learning_rate": 8.190950517167438e-06,
"logits/chosen": -0.03012210875749588,
"logits/rejected": 0.011793004348874092,
"logps/chosen": -451.87530517578125,
"logps/rejected": -471.72772216796875,
"loss": 0.5981,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.4353324174880981,
"rewards/margins": 0.618436872959137,
"rewards/rejected": -2.05376935005188,
"step": 149
},
{
"epoch": 0.7109004739336493,
"grad_norm": 3.944150924682617,
"learning_rate": 8.15893300400332e-06,
"logits/chosen": 0.07667016983032227,
"logits/rejected": -0.02443986013531685,
"logps/chosen": -449.56829833984375,
"logps/rejected": -527.9904174804688,
"loss": 0.5454,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.1610910892486572,
"rewards/margins": 0.8665155172348022,
"rewards/rejected": -2.02760648727417,
"step": 150
},
{
"epoch": 0.7156398104265402,
"grad_norm": 3.960772752761841,
"learning_rate": 8.126698441107146e-06,
"logits/chosen": 0.08127880841493607,
"logits/rejected": 0.21806366741657257,
"logps/chosen": -358.4165954589844,
"logps/rejected": -442.06890869140625,
"loss": 0.4735,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.9019060134887695,
"rewards/margins": 1.199015498161316,
"rewards/rejected": -2.100921630859375,
"step": 151
},
{
"epoch": 0.7203791469194313,
"grad_norm": 3.778970718383789,
"learning_rate": 8.094249043310074e-06,
"logits/chosen": 0.38509824872016907,
"logits/rejected": 0.17644795775413513,
"logps/chosen": -482.83935546875,
"logps/rejected": -469.6008605957031,
"loss": 0.4956,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.3014028072357178,
"rewards/margins": 0.7829511165618896,
"rewards/rejected": -2.0843539237976074,
"step": 152
},
{
"epoch": 0.7251184834123223,
"grad_norm": 4.452014446258545,
"learning_rate": 8.061587040204528e-06,
"logits/chosen": 0.08200319856405258,
"logits/rejected": 0.11494339257478714,
"logps/chosen": -437.13232421875,
"logps/rejected": -410.8484191894531,
"loss": 0.5028,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.0245773792266846,
"rewards/margins": 1.0526708364486694,
"rewards/rejected": -2.0772480964660645,
"step": 153
},
{
"epoch": 0.7298578199052133,
"grad_norm": 5.873034954071045,
"learning_rate": 8.028714675991008e-06,
"logits/chosen": 0.11512620002031326,
"logits/rejected": 0.27921169996261597,
"logps/chosen": -414.84832763671875,
"logps/rejected": -395.1297607421875,
"loss": 0.6651,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.3350560665130615,
"rewards/margins": 0.6761168241500854,
"rewards/rejected": -2.0111730098724365,
"step": 154
},
{
"epoch": 0.7345971563981043,
"grad_norm": 3.5979249477386475,
"learning_rate": 7.995634209323886e-06,
"logits/chosen": 0.3013925850391388,
"logits/rejected": 0.07984738051891327,
"logps/chosen": -405.5412292480469,
"logps/rejected": -362.7952575683594,
"loss": 0.3824,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.0144352912902832,
"rewards/margins": 1.340194821357727,
"rewards/rejected": -2.3546299934387207,
"step": 155
},
{
"epoch": 0.7393364928909952,
"grad_norm": 4.739889144897461,
"learning_rate": 7.962347913156217e-06,
"logits/chosen": 0.3595688045024872,
"logits/rejected": 0.21113444864749908,
"logps/chosen": -439.633544921875,
"logps/rejected": -453.5863952636719,
"loss": 0.5315,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.0545451641082764,
"rewards/margins": 0.9417620301246643,
"rewards/rejected": -1.996307134628296,
"step": 156
},
{
"epoch": 0.7440758293838863,
"grad_norm": 3.8085782527923584,
"learning_rate": 7.92885807458357e-06,
"logits/chosen": 0.11922314763069153,
"logits/rejected": -0.1748376190662384,
"logps/chosen": -459.07763671875,
"logps/rejected": -382.17578125,
"loss": 0.4493,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.6546452045440674,
"rewards/margins": 1.0991568565368652,
"rewards/rejected": -1.753801941871643,
"step": 157
},
{
"epoch": 0.7488151658767772,
"grad_norm": 3.6355323791503906,
"learning_rate": 7.895166994686869e-06,
"logits/chosen": 0.022161336615681648,
"logits/rejected": 0.10442040860652924,
"logps/chosen": -426.6912841796875,
"logps/rejected": -411.6988525390625,
"loss": 0.5699,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.2272857427597046,
"rewards/margins": 0.8166685104370117,
"rewards/rejected": -2.0439541339874268,
"step": 158
},
{
"epoch": 0.7535545023696683,
"grad_norm": 4.030194282531738,
"learning_rate": 7.861276988374303e-06,
"logits/chosen": 0.031902965158224106,
"logits/rejected": -0.08190877735614777,
"logps/chosen": -392.4819030761719,
"logps/rejected": -432.68878173828125,
"loss": 0.5024,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.4378745555877686,
"rewards/margins": 1.0156277418136597,
"rewards/rejected": -2.4535024166107178,
"step": 159
},
{
"epoch": 0.7582938388625592,
"grad_norm": 3.8472750186920166,
"learning_rate": 7.827190384222249e-06,
"logits/chosen": 0.0007600486278533936,
"logits/rejected": 0.19025975465774536,
"logps/chosen": -417.6579895019531,
"logps/rejected": -449.7379150390625,
"loss": 0.5382,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.9642939567565918,
"rewards/margins": 1.006424903869629,
"rewards/rejected": -1.9707188606262207,
"step": 160
},
{
"epoch": 0.7630331753554502,
"grad_norm": 3.9590466022491455,
"learning_rate": 7.7929095243153e-06,
"logits/chosen": 0.13090184330940247,
"logits/rejected": 0.11003851890563965,
"logps/chosen": -411.51605224609375,
"logps/rejected": -465.15020751953125,
"loss": 0.5193,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.2225273847579956,
"rewards/margins": 0.8529220819473267,
"rewards/rejected": -2.0754494667053223,
"step": 161
},
{
"epoch": 0.7677725118483413,
"grad_norm": 4.315500259399414,
"learning_rate": 7.758436764085315e-06,
"logits/chosen": 0.024673327803611755,
"logits/rejected": 0.28246718645095825,
"logps/chosen": -484.12347412109375,
"logps/rejected": -489.55560302734375,
"loss": 0.5323,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.279213547706604,
"rewards/margins": 0.6366978883743286,
"rewards/rejected": -1.9159114360809326,
"step": 162
},
{
"epoch": 0.7725118483412322,
"grad_norm": 5.20791482925415,
"learning_rate": 7.723774472149601e-06,
"logits/chosen": -0.07025258243083954,
"logits/rejected": 0.05357924848794937,
"logps/chosen": -453.21856689453125,
"logps/rejected": -479.0498962402344,
"loss": 0.4967,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.7992047071456909,
"rewards/margins": 0.9488848447799683,
"rewards/rejected": -1.7480895519256592,
"step": 163
},
{
"epoch": 0.7772511848341233,
"grad_norm": 3.974989414215088,
"learning_rate": 7.68892503014815e-06,
"logits/chosen": -0.2237871140241623,
"logits/rejected": -0.042462363839149475,
"logps/chosen": -474.47930908203125,
"logps/rejected": -524.5669555664062,
"loss": 0.4879,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.8265652656555176,
"rewards/margins": 0.8583210110664368,
"rewards/rejected": -1.6848862171173096,
"step": 164
},
{
"epoch": 0.7819905213270142,
"grad_norm": 4.090900421142578,
"learning_rate": 7.653890832580003e-06,
"logits/chosen": 0.3308762013912201,
"logits/rejected": 0.18862006068229675,
"logps/chosen": -356.6459655761719,
"logps/rejected": -375.0365295410156,
"loss": 0.4406,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.9785811901092529,
"rewards/margins": 1.0347230434417725,
"rewards/rejected": -2.0133042335510254,
"step": 165
},
{
"epoch": 0.7867298578199052,
"grad_norm": 4.364016532897949,
"learning_rate": 7.61867428663872e-06,
"logits/chosen": -0.40676772594451904,
"logits/rejected": -0.011762067675590515,
"logps/chosen": -437.35137939453125,
"logps/rejected": -484.0832214355469,
"loss": 0.4485,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.8578428030014038,
"rewards/margins": 1.132391095161438,
"rewards/rejected": -1.9902338981628418,
"step": 166
},
{
"epoch": 0.7914691943127962,
"grad_norm": 3.145108222961426,
"learning_rate": 7.583277812046993e-06,
"logits/chosen": 0.23439201712608337,
"logits/rejected": 0.25310468673706055,
"logps/chosen": -469.490478515625,
"logps/rejected": -526.10693359375,
"loss": 0.3757,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.6878955960273743,
"rewards/margins": 1.2261947393417358,
"rewards/rejected": -1.9140903949737549,
"step": 167
},
{
"epoch": 0.7962085308056872,
"grad_norm": 5.070702075958252,
"learning_rate": 7.547703840890377e-06,
"logits/chosen": -0.17462563514709473,
"logits/rejected": 0.02190116047859192,
"logps/chosen": -328.0517272949219,
"logps/rejected": -350.8951721191406,
"loss": 0.461,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.6660545468330383,
"rewards/margins": 1.430792212486267,
"rewards/rejected": -2.09684681892395,
"step": 168
},
{
"epoch": 0.8009478672985783,
"grad_norm": 3.767350196838379,
"learning_rate": 7.511954817450181e-06,
"logits/chosen": -0.22375909984111786,
"logits/rejected": -0.385820597410202,
"logps/chosen": -455.30841064453125,
"logps/rejected": -472.73822021484375,
"loss": 0.4375,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.4691078662872314,
"rewards/margins": 1.1987533569335938,
"rewards/rejected": -2.667861223220825,
"step": 169
},
{
"epoch": 0.8056872037914692,
"grad_norm": 3.563715696334839,
"learning_rate": 7.476033198035532e-06,
"logits/chosen": -0.21226292848587036,
"logits/rejected": -0.0044862329959869385,
"logps/chosen": -328.20428466796875,
"logps/rejected": -420.17022705078125,
"loss": 0.4346,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.0093038082122803,
"rewards/margins": 1.093737244606018,
"rewards/rejected": -2.103041172027588,
"step": 170
},
{
"epoch": 0.8104265402843602,
"grad_norm": 3.6341946125030518,
"learning_rate": 7.439941450814591e-06,
"logits/chosen": -0.03815871477127075,
"logits/rejected": 0.10924215614795685,
"logps/chosen": -331.86993408203125,
"logps/rejected": -382.3277282714844,
"loss": 0.3771,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.846264123916626,
"rewards/margins": 1.4025492668151855,
"rewards/rejected": -2.2488136291503906,
"step": 171
},
{
"epoch": 0.8151658767772512,
"grad_norm": 4.456596851348877,
"learning_rate": 7.4036820556449704e-06,
"logits/chosen": -0.03349509835243225,
"logits/rejected": -0.18871161341667175,
"logps/chosen": -516.8641967773438,
"logps/rejected": -498.6783447265625,
"loss": 0.5287,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.3133604526519775,
"rewards/margins": 0.7690563797950745,
"rewards/rejected": -2.0824170112609863,
"step": 172
},
{
"epoch": 0.8199052132701422,
"grad_norm": 3.4334120750427246,
"learning_rate": 7.367257503903349e-06,
"logits/chosen": -0.13274812698364258,
"logits/rejected": -0.00023769401013851166,
"logps/chosen": -447.26861572265625,
"logps/rejected": -463.07232666015625,
"loss": 0.3816,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.9694781303405762,
"rewards/margins": 1.1989856958389282,
"rewards/rejected": -2.168463945388794,
"step": 173
},
{
"epoch": 0.8246445497630331,
"grad_norm": 2.757255792617798,
"learning_rate": 7.330670298314282e-06,
"logits/chosen": 0.38800767064094543,
"logits/rejected": 0.5887751579284668,
"logps/chosen": -386.99114990234375,
"logps/rejected": -446.8887023925781,
"loss": 0.383,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.8738099932670593,
"rewards/margins": 1.4890499114990234,
"rewards/rejected": -2.3628599643707275,
"step": 174
},
{
"epoch": 0.8293838862559242,
"grad_norm": 3.8897387981414795,
"learning_rate": 7.293922952778239e-06,
"logits/chosen": 0.17443230748176575,
"logits/rejected": -0.09194624423980713,
"logps/chosen": -379.7416076660156,
"logps/rejected": -319.03955078125,
"loss": 0.399,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.069253921508789,
"rewards/margins": 1.2073869705200195,
"rewards/rejected": -2.2766408920288086,
"step": 175
},
{
"epoch": 0.8341232227488151,
"grad_norm": 4.58038854598999,
"learning_rate": 7.257017992198879e-06,
"logits/chosen": -0.27831190824508667,
"logits/rejected": -0.11275164037942886,
"logps/chosen": -335.1506652832031,
"logps/rejected": -393.7295227050781,
"loss": 0.4676,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.8761013150215149,
"rewards/margins": 0.9996771812438965,
"rewards/rejected": -1.8757784366607666,
"step": 176
},
{
"epoch": 0.8388625592417062,
"grad_norm": 4.329710960388184,
"learning_rate": 7.219957952309568e-06,
"logits/chosen": 0.41356244683265686,
"logits/rejected": 0.5604581236839294,
"logps/chosen": -379.12786865234375,
"logps/rejected": -415.4062805175781,
"loss": 0.4857,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.2384291887283325,
"rewards/margins": 1.2731423377990723,
"rewards/rejected": -2.5115714073181152,
"step": 177
},
{
"epoch": 0.8436018957345972,
"grad_norm": 2.7617781162261963,
"learning_rate": 7.182745379499134e-06,
"logits/chosen": 0.45363107323646545,
"logits/rejected": 0.35966387391090393,
"logps/chosen": -350.72308349609375,
"logps/rejected": -379.54150390625,
"loss": 0.2672,
"rewards/accuracies": 0.84375,
"rewards/chosen": -0.8143477439880371,
"rewards/margins": 2.0239524841308594,
"rewards/rejected": -2.8383002281188965,
"step": 178
},
{
"epoch": 0.8483412322274881,
"grad_norm": 3.4556071758270264,
"learning_rate": 7.145382830636925e-06,
"logits/chosen": 0.22734937071800232,
"logits/rejected": -0.007221966981887817,
"logps/chosen": -451.68951416015625,
"logps/rejected": -407.8339538574219,
"loss": 0.4269,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.2716048955917358,
"rewards/margins": 1.0846253633499146,
"rewards/rejected": -2.3562304973602295,
"step": 179
},
{
"epoch": 0.8530805687203792,
"grad_norm": 3.9663004875183105,
"learning_rate": 7.107872872897112e-06,
"logits/chosen": 0.43388858437538147,
"logits/rejected": 0.3206801414489746,
"logps/chosen": -492.4407958984375,
"logps/rejected": -506.46466064453125,
"loss": 0.508,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.438705325126648,
"rewards/margins": 0.8778525590896606,
"rewards/rejected": -2.3165578842163086,
"step": 180
},
{
"epoch": 0.8578199052132701,
"grad_norm": 3.9396262168884277,
"learning_rate": 7.070218083582306e-06,
"logits/chosen": -0.01552341878414154,
"logits/rejected": -0.040410738438367844,
"logps/chosen": -434.32440185546875,
"logps/rejected": -409.5101013183594,
"loss": 0.478,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.0567607879638672,
"rewards/margins": 1.020298719406128,
"rewards/rejected": -2.077059268951416,
"step": 181
},
{
"epoch": 0.8625592417061612,
"grad_norm": 3.724102258682251,
"learning_rate": 7.032421049946467e-06,
"logits/chosen": -0.15967610478401184,
"logits/rejected": -0.3239976465702057,
"logps/chosen": -533.2361450195312,
"logps/rejected": -586.9674682617188,
"loss": 0.4023,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.5228052139282227,
"rewards/margins": 1.5015686750411987,
"rewards/rejected": -3.024374008178711,
"step": 182
},
{
"epoch": 0.8672985781990521,
"grad_norm": 3.7462353706359863,
"learning_rate": 6.994484369017142e-06,
"logits/chosen": 0.2605791985988617,
"logits/rejected": 0.2061198353767395,
"logps/chosen": -387.6251220703125,
"logps/rejected": -398.81976318359375,
"loss": 0.3708,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.1234267950057983,
"rewards/margins": 1.770387887954712,
"rewards/rejected": -2.8938148021698,
"step": 183
},
{
"epoch": 0.8720379146919431,
"grad_norm": 4.484975337982178,
"learning_rate": 6.9564106474170165e-06,
"logits/chosen": 0.14537742733955383,
"logits/rejected": 0.3496534824371338,
"logps/chosen": -483.0646667480469,
"logps/rejected": -531.8125610351562,
"loss": 0.5152,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.5899852514266968,
"rewards/margins": 0.8374561071395874,
"rewards/rejected": -2.427441358566284,
"step": 184
},
{
"epoch": 0.8767772511848341,
"grad_norm": 5.1885552406311035,
"learning_rate": 6.9182025011848156e-06,
"logits/chosen": 0.16828392446041107,
"logits/rejected": 0.11987558007240295,
"logps/chosen": -531.3955688476562,
"logps/rejected": -551.9052124023438,
"loss": 0.5603,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.5659953355789185,
"rewards/margins": 0.9932804107666016,
"rewards/rejected": -2.5592756271362305,
"step": 185
},
{
"epoch": 0.8815165876777251,
"grad_norm": 4.2418293952941895,
"learning_rate": 6.879862555595559e-06,
"logits/chosen": 0.2174081802368164,
"logits/rejected": 0.09607304632663727,
"logps/chosen": -375.9927062988281,
"logps/rejected": -363.47442626953125,
"loss": 0.4827,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.5877923965454102,
"rewards/margins": 1.0314569473266602,
"rewards/rejected": -2.6192493438720703,
"step": 186
},
{
"epoch": 0.8862559241706162,
"grad_norm": 4.966315269470215,
"learning_rate": 6.841393444980177e-06,
"logits/chosen": 0.13937367498874664,
"logits/rejected": 0.11896172165870667,
"logps/chosen": -465.7598876953125,
"logps/rejected": -472.2224426269531,
"loss": 0.5893,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.710962176322937,
"rewards/margins": 0.8563452959060669,
"rewards/rejected": -2.567307472229004,
"step": 187
},
{
"epoch": 0.8909952606635071,
"grad_norm": 4.166539192199707,
"learning_rate": 6.802797812544503e-06,
"logits/chosen": -0.12522928416728973,
"logits/rejected": 0.08225654065608978,
"logps/chosen": -495.0504150390625,
"logps/rejected": -538.3297729492188,
"loss": 0.4766,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.693655252456665,
"rewards/margins": 1.2182517051696777,
"rewards/rejected": -2.9119069576263428,
"step": 188
},
{
"epoch": 0.8957345971563981,
"grad_norm": 4.550651550292969,
"learning_rate": 6.764078310187667e-06,
"logits/chosen": 0.05809233337640762,
"logits/rejected": -0.03961513191461563,
"logps/chosen": -395.87408447265625,
"logps/rejected": -448.75543212890625,
"loss": 0.5111,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.5405490398406982,
"rewards/margins": 1.279184341430664,
"rewards/rejected": -2.819733142852783,
"step": 189
},
{
"epoch": 0.9004739336492891,
"grad_norm": 3.5397918224334717,
"learning_rate": 6.725237598319877e-06,
"logits/chosen": 0.2512727975845337,
"logits/rejected": 0.4075152277946472,
"logps/chosen": -414.4294128417969,
"logps/rejected": -459.6047058105469,
"loss": 0.3978,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.8530254364013672,
"rewards/margins": 1.6902107000350952,
"rewards/rejected": -3.543236255645752,
"step": 190
},
{
"epoch": 0.9052132701421801,
"grad_norm": 5.755289554595947,
"learning_rate": 6.686278345679626e-06,
"logits/chosen": 0.19829702377319336,
"logits/rejected": 0.030918434262275696,
"logps/chosen": -425.8815002441406,
"logps/rejected": -406.3869323730469,
"loss": 0.5094,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.4468215703964233,
"rewards/margins": 1.3219070434570312,
"rewards/rejected": -2.768728494644165,
"step": 191
},
{
"epoch": 0.909952606635071,
"grad_norm": 4.378885746002197,
"learning_rate": 6.647203229150323e-06,
"logits/chosen": 0.011551991105079651,
"logits/rejected": 0.19708117842674255,
"logps/chosen": -427.0396728515625,
"logps/rejected": -457.7227783203125,
"loss": 0.4901,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.6884351968765259,
"rewards/margins": 1.0523674488067627,
"rewards/rejected": -2.740802526473999,
"step": 192
},
{
"epoch": 0.9146919431279621,
"grad_norm": 3.92692232131958,
"learning_rate": 6.608014933576362e-06,
"logits/chosen": 0.24341322481632233,
"logits/rejected": 0.24133481085300446,
"logps/chosen": -467.0690002441406,
"logps/rejected": -515.750244140625,
"loss": 0.4556,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.0007805824279785,
"rewards/margins": 1.1552082300186157,
"rewards/rejected": -3.1559886932373047,
"step": 193
},
{
"epoch": 0.919431279620853,
"grad_norm": 5.567759037017822,
"learning_rate": 6.568716151578653e-06,
"logits/chosen": 0.18795177340507507,
"logits/rejected": 0.128329798579216,
"logps/chosen": -408.3992919921875,
"logps/rejected": -449.4272155761719,
"loss": 0.5582,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.641984224319458,
"rewards/margins": 1.3132719993591309,
"rewards/rejected": -2.955256462097168,
"step": 194
},
{
"epoch": 0.9241706161137441,
"grad_norm": 2.768008232116699,
"learning_rate": 6.529309583369606e-06,
"logits/chosen": 0.4808191657066345,
"logits/rejected": 0.31546342372894287,
"logps/chosen": -395.28741455078125,
"logps/rejected": -452.6786193847656,
"loss": 0.3497,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.0516895055770874,
"rewards/margins": 1.699798822402954,
"rewards/rejected": -2.751488208770752,
"step": 195
},
{
"epoch": 0.9289099526066351,
"grad_norm": 4.628823757171631,
"learning_rate": 6.489797936567604e-06,
"logits/chosen": -0.099003866314888,
"logits/rejected": 0.046685896813869476,
"logps/chosen": -515.1678466796875,
"logps/rejected": -565.892578125,
"loss": 0.521,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.4594172239303589,
"rewards/margins": 1.0597906112670898,
"rewards/rejected": -2.5192079544067383,
"step": 196
},
{
"epoch": 0.933649289099526,
"grad_norm": 3.6848068237304688,
"learning_rate": 6.45018392601096e-06,
"logits/chosen": 0.2267099916934967,
"logits/rejected": 0.1157350018620491,
"logps/chosen": -493.6409606933594,
"logps/rejected": -533.650390625,
"loss": 0.4944,
"rewards/accuracies": 0.46875,
"rewards/chosen": -1.6284621953964233,
"rewards/margins": 1.1400401592254639,
"rewards/rejected": -2.7685022354125977,
"step": 197
},
{
"epoch": 0.9383886255924171,
"grad_norm": 4.227545261383057,
"learning_rate": 6.410470273571386e-06,
"logits/chosen": -0.06747046858072281,
"logits/rejected": 0.018513869494199753,
"logps/chosen": -481.58990478515625,
"logps/rejected": -538.8641967773438,
"loss": 0.5573,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.0931174755096436,
"rewards/margins": 0.9540358781814575,
"rewards/rejected": -3.0471532344818115,
"step": 198
},
{
"epoch": 0.943127962085308,
"grad_norm": 5.139426231384277,
"learning_rate": 6.370659707966967e-06,
"logits/chosen": 0.326305627822876,
"logits/rejected": 0.3613607883453369,
"logps/chosen": -530.1773681640625,
"logps/rejected": -555.13720703125,
"loss": 0.6477,
"rewards/accuracies": 0.5,
"rewards/chosen": -1.7542710304260254,
"rewards/margins": 0.993602454662323,
"rewards/rejected": -2.7478737831115723,
"step": 199
},
{
"epoch": 0.9478672985781991,
"grad_norm": 4.221323490142822,
"learning_rate": 6.3307549645746756e-06,
"logits/chosen": 0.173638254404068,
"logits/rejected": 0.30959224700927734,
"logps/chosen": -390.31353759765625,
"logps/rejected": -444.4742431640625,
"loss": 0.4113,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.6390478610992432,
"rewards/margins": 1.65326988697052,
"rewards/rejected": -3.2923176288604736,
"step": 200
},
{
"epoch": 0.95260663507109,
"grad_norm": 4.31206750869751,
"learning_rate": 6.290758785242425e-06,
"logits/chosen": 0.06258545815944672,
"logits/rejected": 0.35492444038391113,
"logps/chosen": -358.9757080078125,
"logps/rejected": -418.6335754394531,
"loss": 0.4458,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.3634282350540161,
"rewards/margins": 1.3426121473312378,
"rewards/rejected": -2.706040143966675,
"step": 201
},
{
"epoch": 0.957345971563981,
"grad_norm": 5.077805995941162,
"learning_rate": 6.250673918100671e-06,
"logits/chosen": 0.0679941326379776,
"logits/rejected": 0.21719148755073547,
"logps/chosen": -397.947265625,
"logps/rejected": -410.0476379394531,
"loss": 0.5446,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.099858283996582,
"rewards/margins": 0.8603665828704834,
"rewards/rejected": -2.9602246284484863,
"step": 202
},
{
"epoch": 0.9620853080568721,
"grad_norm": 3.2156455516815186,
"learning_rate": 6.210503117373601e-06,
"logits/chosen": -0.03979983180761337,
"logits/rejected": -0.14512544870376587,
"logps/chosen": -471.88946533203125,
"logps/rejected": -456.22967529296875,
"loss": 0.3515,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.0477066040039062,
"rewards/margins": 1.6663435697555542,
"rewards/rejected": -3.714050054550171,
"step": 203
},
{
"epoch": 0.966824644549763,
"grad_norm": 4.370273113250732,
"learning_rate": 6.170249143189878e-06,
"logits/chosen": 0.10432646423578262,
"logits/rejected": -0.09444320946931839,
"logps/chosen": -398.3092041015625,
"logps/rejected": -394.1639709472656,
"loss": 0.4812,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.064572334289551,
"rewards/margins": 1.2048075199127197,
"rewards/rejected": -3.2693798542022705,
"step": 204
},
{
"epoch": 0.9715639810426541,
"grad_norm": 4.564243793487549,
"learning_rate": 6.129914761393001e-06,
"logits/chosen": -0.1701563596725464,
"logits/rejected": -0.0811741054058075,
"logps/chosen": -342.63153076171875,
"logps/rejected": -331.88641357421875,
"loss": 0.5135,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.414586067199707,
"rewards/margins": 1.498974323272705,
"rewards/rejected": -2.913560390472412,
"step": 205
},
{
"epoch": 0.976303317535545,
"grad_norm": 4.680627822875977,
"learning_rate": 6.089502743351259e-06,
"logits/chosen": 0.41326117515563965,
"logits/rejected": 0.31307393312454224,
"logps/chosen": -523.71484375,
"logps/rejected": -527.9585571289062,
"loss": 0.5034,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.9313371181488037,
"rewards/margins": 1.1199026107788086,
"rewards/rejected": -3.0512397289276123,
"step": 206
},
{
"epoch": 0.981042654028436,
"grad_norm": 7.876077175140381,
"learning_rate": 6.049015865767318e-06,
"logits/chosen": -0.13413012027740479,
"logits/rejected": 0.10655626654624939,
"logps/chosen": -447.656005859375,
"logps/rejected": -466.0491943359375,
"loss": 0.6883,
"rewards/accuracies": 0.5,
"rewards/chosen": -1.5786304473876953,
"rewards/margins": 0.6814507246017456,
"rewards/rejected": -2.2600812911987305,
"step": 207
},
{
"epoch": 0.985781990521327,
"grad_norm": 4.203638076782227,
"learning_rate": 6.0084569104874276e-06,
"logits/chosen": -0.014623567461967468,
"logits/rejected": 0.012134553864598274,
"logps/chosen": -424.0162353515625,
"logps/rejected": -511.604248046875,
"loss": 0.5633,
"rewards/accuracies": 0.5,
"rewards/chosen": -2.323808431625366,
"rewards/margins": 0.872652530670166,
"rewards/rejected": -3.1964609622955322,
"step": 208
},
{
"epoch": 0.990521327014218,
"grad_norm": 3.263500213623047,
"learning_rate": 5.967828664310283e-06,
"logits/chosen": 0.13446447253227234,
"logits/rejected": 0.23876450955867767,
"logps/chosen": -381.72882080078125,
"logps/rejected": -437.77508544921875,
"loss": 0.3952,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.151979923248291,
"rewards/margins": 1.6185417175292969,
"rewards/rejected": -2.770521640777588,
"step": 209
},
{
"epoch": 0.995260663507109,
"grad_norm": 4.874969959259033,
"learning_rate": 5.927133918795548e-06,
"logits/chosen": -0.3298684358596802,
"logits/rejected": -0.1773088425397873,
"logps/chosen": -498.17230224609375,
"logps/rejected": -499.8948669433594,
"loss": 0.4685,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.082857608795166,
"rewards/margins": 1.1343200206756592,
"rewards/rejected": -3.217177629470825,
"step": 210
},
{
"epoch": 1.0,
"grad_norm": 6.354842185974121,
"learning_rate": 5.8863754700720396e-06,
"logits/chosen": 0.06099194288253784,
"logits/rejected": 0.0313149094581604,
"logps/chosen": -570.9532470703125,
"logps/rejected": -558.2776489257812,
"loss": 0.6792,
"rewards/accuracies": 0.5208333134651184,
"rewards/chosen": -2.8509020805358887,
"rewards/margins": 0.9109289646148682,
"rewards/rejected": -3.7618308067321777,
"step": 211
},
{
"epoch": 1.004739336492891,
"grad_norm": 3.0235018730163574,
"learning_rate": 5.845556118645613e-06,
"logits/chosen": 0.2777591347694397,
"logits/rejected": 0.28722092509269714,
"logps/chosen": -455.98736572265625,
"logps/rejected": -512.40087890625,
"loss": 0.4001,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.6508451700210571,
"rewards/margins": 1.2029709815979004,
"rewards/rejected": -2.853816032409668,
"step": 212
},
{
"epoch": 1.009478672985782,
"grad_norm": 4.504083156585693,
"learning_rate": 5.804678669206738e-06,
"logits/chosen": 0.2876705229282379,
"logits/rejected": 0.2990463376045227,
"logps/chosen": -469.29486083984375,
"logps/rejected": -462.4917297363281,
"loss": 0.5315,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.7522639036178589,
"rewards/margins": 1.0054848194122314,
"rewards/rejected": -2.757749080657959,
"step": 213
},
{
"epoch": 1.014218009478673,
"grad_norm": 3.2557713985443115,
"learning_rate": 5.763745930437787e-06,
"logits/chosen": 0.2619165778160095,
"logits/rejected": 0.18662908673286438,
"logps/chosen": -522.1643676757812,
"logps/rejected": -524.9882202148438,
"loss": 0.2847,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.959651231765747,
"rewards/margins": 1.7540154457092285,
"rewards/rejected": -3.7136669158935547,
"step": 214
},
{
"epoch": 1.018957345971564,
"grad_norm": 3.222653388977051,
"learning_rate": 5.722760714820057e-06,
"logits/chosen": -0.08670102059841156,
"logits/rejected": -0.035274673253297806,
"logps/chosen": -492.94903564453125,
"logps/rejected": -469.5622863769531,
"loss": 0.4155,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.162580966949463,
"rewards/margins": 1.281327724456787,
"rewards/rejected": -3.44390869140625,
"step": 215
},
{
"epoch": 1.0236966824644549,
"grad_norm": 3.483283758163452,
"learning_rate": 5.681725838440515e-06,
"logits/chosen": -0.15460936725139618,
"logits/rejected": -0.12697185575962067,
"logps/chosen": -485.9268493652344,
"logps/rejected": -487.96917724609375,
"loss": 0.364,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.5690550804138184,
"rewards/margins": 1.441988468170166,
"rewards/rejected": -3.0110433101654053,
"step": 216
},
{
"epoch": 1.028436018957346,
"grad_norm": 3.2059872150421143,
"learning_rate": 5.640644120798312e-06,
"logits/chosen": 0.17371203005313873,
"logits/rejected": 0.09106084704399109,
"logps/chosen": -375.935546875,
"logps/rejected": -388.53076171875,
"loss": 0.332,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.4462567567825317,
"rewards/margins": 1.620455026626587,
"rewards/rejected": -3.066711902618408,
"step": 217
},
{
"epoch": 1.033175355450237,
"grad_norm": 3.8562071323394775,
"learning_rate": 5.599518384611052e-06,
"logits/chosen": 0.2562037706375122,
"logits/rejected": 0.18646736443042755,
"logps/chosen": -418.3273620605469,
"logps/rejected": -407.41162109375,
"loss": 0.2776,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.1467158794403076,
"rewards/margins": 2.1470839977264404,
"rewards/rejected": -3.293799877166748,
"step": 218
},
{
"epoch": 1.037914691943128,
"grad_norm": 2.9552791118621826,
"learning_rate": 5.558351455620852e-06,
"logits/chosen": 0.2555815577507019,
"logits/rejected": 0.17480355501174927,
"logps/chosen": -402.4105529785156,
"logps/rejected": -408.10552978515625,
"loss": 0.282,
"rewards/accuracies": 0.84375,
"rewards/chosen": -0.7671201825141907,
"rewards/margins": 1.6860198974609375,
"rewards/rejected": -2.4531400203704834,
"step": 219
},
{
"epoch": 1.042654028436019,
"grad_norm": 2.3003854751586914,
"learning_rate": 5.517146162400171e-06,
"logits/chosen": 0.08371573686599731,
"logits/rejected": 0.0009252801537513733,
"logps/chosen": -368.94744873046875,
"logps/rejected": -376.25372314453125,
"loss": 0.3108,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.2716785669326782,
"rewards/margins": 1.600062608718872,
"rewards/rejected": -2.8717408180236816,
"step": 220
},
{
"epoch": 1.04739336492891,
"grad_norm": 3.363819122314453,
"learning_rate": 5.475905336157473e-06,
"logits/chosen": 0.395662784576416,
"logits/rejected": 0.32149434089660645,
"logps/chosen": -459.91204833984375,
"logps/rejected": -506.9620361328125,
"loss": 0.3905,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.9222174882888794,
"rewards/margins": 1.5406379699707031,
"rewards/rejected": -3.462855815887451,
"step": 221
},
{
"epoch": 1.052132701421801,
"grad_norm": 3.3270413875579834,
"learning_rate": 5.434631810542688e-06,
"logits/chosen": 0.1943908929824829,
"logits/rejected": 0.13462212681770325,
"logps/chosen": -437.5592041015625,
"logps/rejected": -512.220947265625,
"loss": 0.3339,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.5035971403121948,
"rewards/margins": 1.7122483253479004,
"rewards/rejected": -3.2158455848693848,
"step": 222
},
{
"epoch": 1.0568720379146919,
"grad_norm": 4.051716327667236,
"learning_rate": 5.393328421452514e-06,
"logits/chosen": 0.09970849752426147,
"logits/rejected": 0.34582486748695374,
"logps/chosen": -404.18646240234375,
"logps/rejected": -461.6804504394531,
"loss": 0.4114,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.4395058155059814,
"rewards/margins": 1.6363253593444824,
"rewards/rejected": -3.075831413269043,
"step": 223
},
{
"epoch": 1.061611374407583,
"grad_norm": 3.6509335041046143,
"learning_rate": 5.351998006835563e-06,
"logits/chosen": 0.016243021935224533,
"logits/rejected": 0.12777027487754822,
"logps/chosen": -506.6263427734375,
"logps/rejected": -551.2019653320312,
"loss": 0.4227,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.7827990055084229,
"rewards/margins": 1.2781894207000732,
"rewards/rejected": -3.060988187789917,
"step": 224
},
{
"epoch": 1.066350710900474,
"grad_norm": 3.744976043701172,
"learning_rate": 5.3106434064973665e-06,
"logits/chosen": 0.16931653022766113,
"logits/rejected": 0.07942713797092438,
"logps/chosen": -528.5944213867188,
"logps/rejected": -571.782958984375,
"loss": 0.3732,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.5323777198791504,
"rewards/margins": 1.6922569274902344,
"rewards/rejected": -3.2246346473693848,
"step": 225
},
{
"epoch": 1.0710900473933649,
"grad_norm": 2.9027132987976074,
"learning_rate": 5.269267461905253e-06,
"logits/chosen": -0.03642325848340988,
"logits/rejected": 0.27906686067581177,
"logps/chosen": -315.7832336425781,
"logps/rejected": -381.1293029785156,
"loss": 0.3682,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.1626783609390259,
"rewards/margins": 1.8461182117462158,
"rewards/rejected": -3.0087969303131104,
"step": 226
},
{
"epoch": 1.0758293838862558,
"grad_norm": 4.7471208572387695,
"learning_rate": 5.227873015993108e-06,
"logits/chosen": -0.5098191499710083,
"logits/rejected": -0.4582579731941223,
"logps/chosen": -518.41455078125,
"logps/rejected": -516.23193359375,
"loss": 0.4351,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.9056390523910522,
"rewards/margins": 1.5796176195144653,
"rewards/rejected": -3.4852566719055176,
"step": 227
},
{
"epoch": 1.080568720379147,
"grad_norm": 3.349379062652588,
"learning_rate": 5.186462912966047e-06,
"logits/chosen": 0.2078540027141571,
"logits/rejected": 0.22412914037704468,
"logps/chosen": -354.3533935546875,
"logps/rejected": -354.4374084472656,
"loss": 0.3274,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.2849071025848389,
"rewards/margins": 1.9719514846801758,
"rewards/rejected": -3.2568585872650146,
"step": 228
},
{
"epoch": 1.085308056872038,
"grad_norm": 2.522663116455078,
"learning_rate": 5.145039998104975e-06,
"logits/chosen": -0.10203185677528381,
"logits/rejected": 0.12238292396068573,
"logps/chosen": -420.65643310546875,
"logps/rejected": -487.3912353515625,
"loss": 0.3761,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.6576614379882812,
"rewards/margins": 1.5825456380844116,
"rewards/rejected": -3.2402069568634033,
"step": 229
},
{
"epoch": 1.0900473933649288,
"grad_norm": 3.699955701828003,
"learning_rate": 5.1036071175710984e-06,
"logits/chosen": 0.3874337077140808,
"logits/rejected": 0.3852435052394867,
"logps/chosen": -425.8739013671875,
"logps/rejected": -397.9554748535156,
"loss": 0.4133,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.5295315980911255,
"rewards/margins": 1.3557639122009277,
"rewards/rejected": -2.885295867919922,
"step": 230
},
{
"epoch": 1.09478672985782,
"grad_norm": 2.6708872318267822,
"learning_rate": 5.062167118210367e-06,
"logits/chosen": 0.000867149792611599,
"logits/rejected": -0.034792158752679825,
"logps/chosen": -434.9617919921875,
"logps/rejected": -451.12725830078125,
"loss": 0.3438,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.3274954557418823,
"rewards/margins": 1.2974351644515991,
"rewards/rejected": -2.6249306201934814,
"step": 231
},
{
"epoch": 1.099526066350711,
"grad_norm": 4.831852436065674,
"learning_rate": 5.020722847357858e-06,
"logits/chosen": -0.03341293707489967,
"logits/rejected": -0.06871610134840012,
"logps/chosen": -563.6790161132812,
"logps/rejected": -461.8121337890625,
"loss": 0.4987,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.9696635007858276,
"rewards/margins": 1.1347928047180176,
"rewards/rejected": -3.1044564247131348,
"step": 232
},
{
"epoch": 1.1042654028436019,
"grad_norm": 3.7005701065063477,
"learning_rate": 4.979277152642145e-06,
"logits/chosen": 0.03712880611419678,
"logits/rejected": 0.08507107198238373,
"logps/chosen": -336.78387451171875,
"logps/rejected": -477.3684387207031,
"loss": 0.3123,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.632716178894043,
"rewards/margins": 2.3301315307617188,
"rewards/rejected": -3.9628477096557617,
"step": 233
},
{
"epoch": 1.1090047393364928,
"grad_norm": 3.4187428951263428,
"learning_rate": 4.937832881789635e-06,
"logits/chosen": -0.0916728526353836,
"logits/rejected": -0.02996966987848282,
"logps/chosen": -460.55352783203125,
"logps/rejected": -454.41326904296875,
"loss": 0.3771,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.8706817626953125,
"rewards/margins": 1.590428352355957,
"rewards/rejected": -3.4611101150512695,
"step": 234
},
{
"epoch": 1.113744075829384,
"grad_norm": 2.6118838787078857,
"learning_rate": 4.8963928824289015e-06,
"logits/chosen": 0.1105157807469368,
"logits/rejected": 0.5052557587623596,
"logps/chosen": -362.1018981933594,
"logps/rejected": -456.45074462890625,
"loss": 0.3053,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.652755856513977,
"rewards/margins": 1.88970148563385,
"rewards/rejected": -3.5424575805664062,
"step": 235
},
{
"epoch": 1.1184834123222749,
"grad_norm": 3.3146190643310547,
"learning_rate": 4.854960001895027e-06,
"logits/chosen": 0.19854995608329773,
"logits/rejected": 0.2642470598220825,
"logps/chosen": -455.2829895019531,
"logps/rejected": -508.4591064453125,
"loss": 0.3811,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.6405863761901855,
"rewards/margins": 1.4157403707504272,
"rewards/rejected": -3.0563266277313232,
"step": 236
},
{
"epoch": 1.1232227488151658,
"grad_norm": 2.650205612182617,
"learning_rate": 4.813537087033954e-06,
"logits/chosen": -0.1753956526517868,
"logits/rejected": -0.15638449788093567,
"logps/chosen": -425.37603759765625,
"logps/rejected": -428.813232421875,
"loss": 0.2871,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.447309970855713,
"rewards/margins": 1.9025884866714478,
"rewards/rejected": -3.349898338317871,
"step": 237
},
{
"epoch": 1.1279620853080567,
"grad_norm": 5.000858783721924,
"learning_rate": 4.772126984006892e-06,
"logits/chosen": 0.25283899903297424,
"logits/rejected": 0.16875059902668,
"logps/chosen": -455.42425537109375,
"logps/rejected": -470.0882873535156,
"loss": 0.4152,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.5532653331756592,
"rewards/margins": 1.7187979221343994,
"rewards/rejected": -3.2720632553100586,
"step": 238
},
{
"epoch": 1.132701421800948,
"grad_norm": 2.7232418060302734,
"learning_rate": 4.7307325380947495e-06,
"logits/chosen": 0.0400099977850914,
"logits/rejected": 0.13005331158638,
"logps/chosen": -358.5569152832031,
"logps/rejected": -500.84222412109375,
"loss": 0.3338,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.523422122001648,
"rewards/margins": 1.6523953676223755,
"rewards/rejected": -3.1758174896240234,
"step": 239
},
{
"epoch": 1.1374407582938388,
"grad_norm": 3.443223714828491,
"learning_rate": 4.689356593502635e-06,
"logits/chosen": 0.3939206600189209,
"logits/rejected": 0.2064402997493744,
"logps/chosen": -429.4366760253906,
"logps/rejected": -439.154296875,
"loss": 0.3795,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.792637825012207,
"rewards/margins": 1.6197113990783691,
"rewards/rejected": -3.412349224090576,
"step": 240
},
{
"epoch": 1.1421800947867298,
"grad_norm": 2.805168390274048,
"learning_rate": 4.648001993164438e-06,
"logits/chosen": 0.06738889217376709,
"logits/rejected": 0.238703191280365,
"logps/chosen": -490.4261169433594,
"logps/rejected": -511.69317626953125,
"loss": 0.2848,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.8254410028457642,
"rewards/margins": 2.290480375289917,
"rewards/rejected": -4.115921497344971,
"step": 241
},
{
"epoch": 1.146919431279621,
"grad_norm": 4.80793571472168,
"learning_rate": 4.606671578547488e-06,
"logits/chosen": 0.3581339418888092,
"logits/rejected": -0.04958381503820419,
"logps/chosen": -646.1534423828125,
"logps/rejected": -614.028564453125,
"loss": 0.467,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.5761334896087646,
"rewards/margins": 1.4988954067230225,
"rewards/rejected": -4.075028896331787,
"step": 242
},
{
"epoch": 1.1516587677725119,
"grad_norm": 3.446601152420044,
"learning_rate": 4.565368189457313e-06,
"logits/chosen": -0.129550501704216,
"logits/rejected": 0.18526773154735565,
"logps/chosen": -371.20709228515625,
"logps/rejected": -493.4662780761719,
"loss": 0.3148,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.3364325761795044,
"rewards/margins": 1.9403244256973267,
"rewards/rejected": -3.27675724029541,
"step": 243
},
{
"epoch": 1.1563981042654028,
"grad_norm": 3.778529167175293,
"learning_rate": 4.524094663842528e-06,
"logits/chosen": -0.27983617782592773,
"logits/rejected": -0.11589246988296509,
"logps/chosen": -510.7512512207031,
"logps/rejected": -511.7879333496094,
"loss": 0.4457,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.7793432474136353,
"rewards/margins": 1.390049695968628,
"rewards/rejected": -3.1693930625915527,
"step": 244
},
{
"epoch": 1.161137440758294,
"grad_norm": 3.1189873218536377,
"learning_rate": 4.482853837599831e-06,
"logits/chosen": -0.13306924700737,
"logits/rejected": -0.3103134334087372,
"logps/chosen": -494.39837646484375,
"logps/rejected": -465.83734130859375,
"loss": 0.3125,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.1813969612121582,
"rewards/margins": 1.8126945495605469,
"rewards/rejected": -2.994091510772705,
"step": 245
},
{
"epoch": 1.1658767772511849,
"grad_norm": 3.904042959213257,
"learning_rate": 4.441648544379149e-06,
"logits/chosen": 0.007268290966749191,
"logits/rejected": 0.07453174144029617,
"logps/chosen": -432.7102966308594,
"logps/rejected": -487.7734069824219,
"loss": 0.4032,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.6436212062835693,
"rewards/margins": 1.462231159210205,
"rewards/rejected": -3.1058521270751953,
"step": 246
},
{
"epoch": 1.1706161137440758,
"grad_norm": 3.9296388626098633,
"learning_rate": 4.400481615388948e-06,
"logits/chosen": 0.17288221418857574,
"logits/rejected": 0.1198100745677948,
"logps/chosen": -484.23211669921875,
"logps/rejected": -434.79248046875,
"loss": 0.468,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.9902846813201904,
"rewards/margins": 1.1091970205307007,
"rewards/rejected": -3.0994820594787598,
"step": 247
},
{
"epoch": 1.1753554502369667,
"grad_norm": 3.4638829231262207,
"learning_rate": 4.359355879201691e-06,
"logits/chosen": 0.0034363269805908203,
"logits/rejected": 0.07013356685638428,
"logps/chosen": -434.6395263671875,
"logps/rejected": -472.5997619628906,
"loss": 0.2568,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.535360336303711,
"rewards/margins": 2.1144678592681885,
"rewards/rejected": -3.6498279571533203,
"step": 248
},
{
"epoch": 1.180094786729858,
"grad_norm": 4.689034938812256,
"learning_rate": 4.318274161559487e-06,
"logits/chosen": -0.19193895161151886,
"logits/rejected": -0.24672627449035645,
"logps/chosen": -499.9013671875,
"logps/rejected": -477.8822021484375,
"loss": 0.5502,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.9828531742095947,
"rewards/margins": 0.8065874576568604,
"rewards/rejected": -2.789440631866455,
"step": 249
},
{
"epoch": 1.1848341232227488,
"grad_norm": 2.314577341079712,
"learning_rate": 4.277239285179944e-06,
"logits/chosen": 0.2183598279953003,
"logits/rejected": 0.3239366412162781,
"logps/chosen": -481.9766540527344,
"logps/rejected": -555.5982666015625,
"loss": 0.3188,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.300562858581543,
"rewards/margins": 1.5612691640853882,
"rewards/rejected": -2.8618321418762207,
"step": 250
},
{
"epoch": 1.1895734597156398,
"grad_norm": 3.666198968887329,
"learning_rate": 4.236254069562214e-06,
"logits/chosen": -0.1494772732257843,
"logits/rejected": -0.3153489828109741,
"logps/chosen": -548.8350219726562,
"logps/rejected": -581.28857421875,
"loss": 0.3582,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.901492953300476,
"rewards/margins": 1.6575918197631836,
"rewards/rejected": -3.559084892272949,
"step": 251
},
{
"epoch": 1.1943127962085307,
"grad_norm": 3.301483631134033,
"learning_rate": 4.195321330793264e-06,
"logits/chosen": 0.3761504590511322,
"logits/rejected": 0.12152983993291855,
"logps/chosen": -474.353759765625,
"logps/rejected": -493.3826904296875,
"loss": 0.3582,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.6552190780639648,
"rewards/margins": 1.3365132808685303,
"rewards/rejected": -2.991732358932495,
"step": 252
},
{
"epoch": 1.1990521327014219,
"grad_norm": 2.335012197494507,
"learning_rate": 4.154443881354388e-06,
"logits/chosen": -0.2192879021167755,
"logits/rejected": -0.32967904210090637,
"logps/chosen": -446.9404602050781,
"logps/rejected": -512.4267578125,
"loss": 0.3133,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.2855887413024902,
"rewards/margins": 1.6910157203674316,
"rewards/rejected": -2.976604461669922,
"step": 253
},
{
"epoch": 1.2037914691943128,
"grad_norm": 5.097269058227539,
"learning_rate": 4.113624529927963e-06,
"logits/chosen": 0.1676633656024933,
"logits/rejected": 0.22459819912910461,
"logps/chosen": -380.5593566894531,
"logps/rejected": -356.56219482421875,
"loss": 0.336,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.4843287467956543,
"rewards/margins": 1.580984354019165,
"rewards/rejected": -3.0653133392333984,
"step": 254
},
{
"epoch": 1.2085308056872037,
"grad_norm": 2.614943265914917,
"learning_rate": 4.072866081204453e-06,
"logits/chosen": 0.11476369202136993,
"logits/rejected": 0.1240728348493576,
"logps/chosen": -502.7626953125,
"logps/rejected": -518.6180419921875,
"loss": 0.2659,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.6948649883270264,
"rewards/margins": 1.8275578022003174,
"rewards/rejected": -3.5224225521087646,
"step": 255
},
{
"epoch": 1.2132701421800949,
"grad_norm": 3.184537649154663,
"learning_rate": 4.032171335689718e-06,
"logits/chosen": 0.15463709831237793,
"logits/rejected": -0.19407552480697632,
"logps/chosen": -424.0015869140625,
"logps/rejected": -435.4846496582031,
"loss": 0.3429,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.1620564460754395,
"rewards/margins": 1.5143887996673584,
"rewards/rejected": -3.6764450073242188,
"step": 256
},
{
"epoch": 1.2180094786729858,
"grad_norm": 5.3068437576293945,
"learning_rate": 3.991543089512574e-06,
"logits/chosen": -0.03403574973344803,
"logits/rejected": 0.20418822765350342,
"logps/chosen": -546.001220703125,
"logps/rejected": -661.1378784179688,
"loss": 0.4368,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.00406813621521,
"rewards/margins": 1.6435837745666504,
"rewards/rejected": -3.6476516723632812,
"step": 257
},
{
"epoch": 1.2227488151658767,
"grad_norm": 2.3214268684387207,
"learning_rate": 3.9509841342326835e-06,
"logits/chosen": 0.18236321210861206,
"logits/rejected": 0.012637853622436523,
"logps/chosen": -340.8662109375,
"logps/rejected": -393.6156005859375,
"loss": 0.242,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.4383803606033325,
"rewards/margins": 2.1315698623657227,
"rewards/rejected": -3.5699501037597656,
"step": 258
},
{
"epoch": 1.2274881516587677,
"grad_norm": 2.3699498176574707,
"learning_rate": 3.910497256648742e-06,
"logits/chosen": -0.17738784849643707,
"logits/rejected": 0.058183833956718445,
"logps/chosen": -337.60540771484375,
"logps/rejected": -408.6331481933594,
"loss": 0.2667,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.7594650983810425,
"rewards/margins": 2.0310773849487305,
"rewards/rejected": -3.7905426025390625,
"step": 259
},
{
"epoch": 1.2322274881516588,
"grad_norm": 4.384218692779541,
"learning_rate": 3.870085238607002e-06,
"logits/chosen": 0.002813771367073059,
"logits/rejected": -0.24638204276561737,
"logps/chosen": -433.4595031738281,
"logps/rejected": -462.3036193847656,
"loss": 0.3907,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.449141263961792,
"rewards/margins": 1.7780585289001465,
"rewards/rejected": -3.2271997928619385,
"step": 260
},
{
"epoch": 1.2369668246445498,
"grad_norm": 3.777799606323242,
"learning_rate": 3.829750856810123e-06,
"logits/chosen": 0.040886975824832916,
"logits/rejected": 0.15636533498764038,
"logps/chosen": -398.0843200683594,
"logps/rejected": -536.947998046875,
"loss": 0.3703,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.132894515991211,
"rewards/margins": 1.518797755241394,
"rewards/rejected": -3.6516923904418945,
"step": 261
},
{
"epoch": 1.2417061611374407,
"grad_norm": 3.2709872722625732,
"learning_rate": 3.7894968826263993e-06,
"logits/chosen": 0.35484781861305237,
"logits/rejected": -0.043566398322582245,
"logps/chosen": -551.0684814453125,
"logps/rejected": -534.43701171875,
"loss": 0.2318,
"rewards/accuracies": 0.90625,
"rewards/chosen": -1.4083945751190186,
"rewards/margins": 2.284555673599243,
"rewards/rejected": -3.6929502487182617,
"step": 262
},
{
"epoch": 1.2464454976303316,
"grad_norm": 3.492917060852051,
"learning_rate": 3.7493260818993295e-06,
"logits/chosen": -0.24769189953804016,
"logits/rejected": -0.033029571175575256,
"logps/chosen": -437.48638916015625,
"logps/rejected": -473.00543212890625,
"loss": 0.3564,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.5751187801361084,
"rewards/margins": 1.6503984928131104,
"rewards/rejected": -3.2255170345306396,
"step": 263
},
{
"epoch": 1.2511848341232228,
"grad_norm": 2.8335609436035156,
"learning_rate": 3.7092412147575763e-06,
"logits/chosen": 0.08972194790840149,
"logits/rejected": 0.0525023378431797,
"logps/chosen": -399.0003662109375,
"logps/rejected": -462.21539306640625,
"loss": 0.3082,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.152060031890869,
"rewards/margins": 2.0616841316223145,
"rewards/rejected": -4.213744163513184,
"step": 264
},
{
"epoch": 1.2559241706161137,
"grad_norm": 3.3234729766845703,
"learning_rate": 3.6692450354253244e-06,
"logits/chosen": -0.11747504025697708,
"logits/rejected": -0.09165113419294357,
"logps/chosen": -487.0151062011719,
"logps/rejected": -492.8780212402344,
"loss": 0.3621,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.7996101379394531,
"rewards/margins": 1.4663692712783813,
"rewards/rejected": -3.265979528427124,
"step": 265
},
{
"epoch": 1.2606635071090047,
"grad_norm": 2.617558240890503,
"learning_rate": 3.6293402920330345e-06,
"logits/chosen": 0.3675645589828491,
"logits/rejected": 0.2720622420310974,
"logps/chosen": -421.515869140625,
"logps/rejected": -430.33367919921875,
"loss": 0.369,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.132265567779541,
"rewards/margins": 1.4423668384552002,
"rewards/rejected": -3.5746326446533203,
"step": 266
},
{
"epoch": 1.2654028436018958,
"grad_norm": 2.5565578937530518,
"learning_rate": 3.589529726428615e-06,
"logits/chosen": 0.07273396104574203,
"logits/rejected": 0.060611553490161896,
"logps/chosen": -418.9151306152344,
"logps/rejected": -446.34698486328125,
"loss": 0.3304,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.6537911891937256,
"rewards/margins": 1.5491554737091064,
"rewards/rejected": -3.202946662902832,
"step": 267
},
{
"epoch": 1.2701421800947867,
"grad_norm": 3.8108534812927246,
"learning_rate": 3.54981607398904e-06,
"logits/chosen": 0.035040952265262604,
"logits/rejected": 0.022193975746631622,
"logps/chosen": -488.48675537109375,
"logps/rejected": -434.921875,
"loss": 0.3831,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.611822485923767,
"rewards/margins": 1.6759777069091797,
"rewards/rejected": -3.2878000736236572,
"step": 268
},
{
"epoch": 1.2748815165876777,
"grad_norm": 2.969774007797241,
"learning_rate": 3.5102020634323974e-06,
"logits/chosen": 0.2675975561141968,
"logits/rejected": 0.2470274567604065,
"logps/chosen": -481.4004821777344,
"logps/rejected": -485.03863525390625,
"loss": 0.432,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.6142653226852417,
"rewards/margins": 1.4393080472946167,
"rewards/rejected": -3.0535733699798584,
"step": 269
},
{
"epoch": 1.2796208530805688,
"grad_norm": 2.932152271270752,
"learning_rate": 3.470690416630395e-06,
"logits/chosen": -0.06690172851085663,
"logits/rejected": 0.149437814950943,
"logps/chosen": -383.09259033203125,
"logps/rejected": -452.948974609375,
"loss": 0.3637,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.385080099105835,
"rewards/margins": 1.3777412176132202,
"rewards/rejected": -2.7628211975097656,
"step": 270
},
{
"epoch": 1.2843601895734598,
"grad_norm": 3.806013584136963,
"learning_rate": 3.431283848421347e-06,
"logits/chosen": 0.1305023580789566,
"logits/rejected": -0.12427529692649841,
"logps/chosen": -432.52520751953125,
"logps/rejected": -456.1821594238281,
"loss": 0.3401,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.410179853439331,
"rewards/margins": 1.6679764986038208,
"rewards/rejected": -3.078155994415283,
"step": 271
},
{
"epoch": 1.2890995260663507,
"grad_norm": 3.743184804916382,
"learning_rate": 3.39198506642364e-06,
"logits/chosen": 0.18078383803367615,
"logits/rejected": 0.16558128595352173,
"logps/chosen": -404.4733581542969,
"logps/rejected": -463.1748046875,
"loss": 0.4068,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.8519794940948486,
"rewards/margins": 1.207563042640686,
"rewards/rejected": -3.059542655944824,
"step": 272
},
{
"epoch": 1.2938388625592416,
"grad_norm": 3.420830726623535,
"learning_rate": 3.352796770849679e-06,
"logits/chosen": -0.22084331512451172,
"logits/rejected": -0.03927930071949959,
"logps/chosen": -379.3165588378906,
"logps/rejected": -375.982666015625,
"loss": 0.2515,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.0934913158416748,
"rewards/margins": 2.5695981979370117,
"rewards/rejected": -3.6630892753601074,
"step": 273
},
{
"epoch": 1.2985781990521326,
"grad_norm": 2.730637788772583,
"learning_rate": 3.3137216543203747e-06,
"logits/chosen": 0.3634192943572998,
"logits/rejected": 0.02868695929646492,
"logps/chosen": -395.314697265625,
"logps/rejected": -386.64801025390625,
"loss": 0.3487,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.4456108808517456,
"rewards/margins": 1.609388828277588,
"rewards/rejected": -3.054999828338623,
"step": 274
},
{
"epoch": 1.3033175355450237,
"grad_norm": 5.403792858123779,
"learning_rate": 3.2747624016801245e-06,
"logits/chosen": -0.20156198740005493,
"logits/rejected": 0.0912720337510109,
"logps/chosen": -481.54888916015625,
"logps/rejected": -518.7413940429688,
"loss": 0.5796,
"rewards/accuracies": 0.53125,
"rewards/chosen": -2.1031124591827393,
"rewards/margins": 1.3092026710510254,
"rewards/rejected": -3.4123151302337646,
"step": 275
},
{
"epoch": 1.3080568720379147,
"grad_norm": 3.5099925994873047,
"learning_rate": 3.2359216898123343e-06,
"logits/chosen": -0.05037354677915573,
"logits/rejected": -0.1334826946258545,
"logps/chosen": -556.6934814453125,
"logps/rejected": -531.0350341796875,
"loss": 0.4337,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.9162890911102295,
"rewards/margins": 1.1660747528076172,
"rewards/rejected": -3.0823636054992676,
"step": 276
},
{
"epoch": 1.3127962085308056,
"grad_norm": 3.462001323699951,
"learning_rate": 3.197202187455498e-06,
"logits/chosen": -0.10171887278556824,
"logits/rejected": -0.15261708199977875,
"logps/chosen": -555.4314575195312,
"logps/rejected": -588.11767578125,
"loss": 0.3584,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.8054559230804443,
"rewards/margins": 1.69133722782135,
"rewards/rejected": -3.496793270111084,
"step": 277
},
{
"epoch": 1.3175355450236967,
"grad_norm": 2.394829511642456,
"learning_rate": 3.158606555019826e-06,
"logits/chosen": -0.19711580872535706,
"logits/rejected": -0.17292340099811554,
"logps/chosen": -590.0509033203125,
"logps/rejected": -577.7540893554688,
"loss": 0.2994,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.7370967864990234,
"rewards/margins": 1.9557228088378906,
"rewards/rejected": -3.692819833755493,
"step": 278
},
{
"epoch": 1.3222748815165877,
"grad_norm": 3.8144333362579346,
"learning_rate": 3.120137444404442e-06,
"logits/chosen": 0.1638478934764862,
"logits/rejected": 0.1452624797821045,
"logps/chosen": -334.2840576171875,
"logps/rejected": -392.561279296875,
"loss": 0.373,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.239801049232483,
"rewards/margins": 1.3264257907867432,
"rewards/rejected": -2.5662269592285156,
"step": 279
},
{
"epoch": 1.3270142180094786,
"grad_norm": 3.646484613418579,
"learning_rate": 3.081797498815185e-06,
"logits/chosen": -0.1268320083618164,
"logits/rejected": -0.17580845952033997,
"logps/chosen": -460.2949523925781,
"logps/rejected": -452.21783447265625,
"loss": 0.3658,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.822378158569336,
"rewards/margins": 1.5197012424468994,
"rewards/rejected": -3.3420791625976562,
"step": 280
},
{
"epoch": 1.3317535545023698,
"grad_norm": 3.2970638275146484,
"learning_rate": 3.0435893525829847e-06,
"logits/chosen": 0.13139204680919647,
"logits/rejected": 0.24725201725959778,
"logps/chosen": -373.4732971191406,
"logps/rejected": -456.06658935546875,
"loss": 0.2123,
"rewards/accuracies": 0.90625,
"rewards/chosen": -0.9837716221809387,
"rewards/margins": 2.4389843940734863,
"rewards/rejected": -3.4227561950683594,
"step": 281
},
{
"epoch": 1.3364928909952607,
"grad_norm": 2.8067758083343506,
"learning_rate": 3.005515630982858e-06,
"logits/chosen": 0.13529516756534576,
"logits/rejected": 0.13115084171295166,
"logps/chosen": -541.8878784179688,
"logps/rejected": -540.979736328125,
"loss": 0.3042,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.8211042881011963,
"rewards/margins": 1.9124205112457275,
"rewards/rejected": -3.733524799346924,
"step": 282
},
{
"epoch": 1.3412322274881516,
"grad_norm": 4.505223751068115,
"learning_rate": 2.9675789500535328e-06,
"logits/chosen": 0.28925737738609314,
"logits/rejected": -0.0881756916642189,
"logps/chosen": -447.83154296875,
"logps/rejected": -411.06842041015625,
"loss": 0.3877,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.7665940523147583,
"rewards/margins": 1.4863183498382568,
"rewards/rejected": -3.2529125213623047,
"step": 283
},
{
"epoch": 1.3459715639810428,
"grad_norm": 3.6730592250823975,
"learning_rate": 2.9297819164176964e-06,
"logits/chosen": -0.024140790104866028,
"logits/rejected": -0.05335076153278351,
"logps/chosen": -509.5386047363281,
"logps/rejected": -589.5108642578125,
"loss": 0.4042,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.5151710510253906,
"rewards/margins": 1.5130798816680908,
"rewards/rejected": -4.0282511711120605,
"step": 284
},
{
"epoch": 1.3507109004739337,
"grad_norm": 2.930845260620117,
"learning_rate": 2.8921271271028896e-06,
"logits/chosen": 0.4924960732460022,
"logits/rejected": 0.44390055537223816,
"logps/chosen": -551.4659423828125,
"logps/rejected": -540.36181640625,
"loss": 0.3399,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.004122495651245,
"rewards/margins": 1.6733444929122925,
"rewards/rejected": -3.677466869354248,
"step": 285
},
{
"epoch": 1.3554502369668247,
"grad_norm": 3.2506113052368164,
"learning_rate": 2.854617169363075e-06,
"logits/chosen": -0.24163638055324554,
"logits/rejected": 0.028947792947292328,
"logps/chosen": -525.6647338867188,
"logps/rejected": -594.4169311523438,
"loss": 0.2765,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.7295421361923218,
"rewards/margins": 2.143195390701294,
"rewards/rejected": -3.8727376461029053,
"step": 286
},
{
"epoch": 1.3601895734597156,
"grad_norm": 4.907602787017822,
"learning_rate": 2.817254620500868e-06,
"logits/chosen": -0.05144549906253815,
"logits/rejected": -0.03435233607888222,
"logps/chosen": -322.93670654296875,
"logps/rejected": -336.6121826171875,
"loss": 0.5032,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.0222997665405273,
"rewards/margins": 1.0926249027252197,
"rewards/rejected": -3.114924669265747,
"step": 287
},
{
"epoch": 1.3649289099526065,
"grad_norm": 2.1702044010162354,
"learning_rate": 2.7800420476904337e-06,
"logits/chosen": 0.17638254165649414,
"logits/rejected": 0.03706669062376022,
"logps/chosen": -482.89410400390625,
"logps/rejected": -495.45635986328125,
"loss": 0.2756,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.5554332733154297,
"rewards/margins": 1.9448996782302856,
"rewards/rejected": -3.500332832336426,
"step": 288
},
{
"epoch": 1.3696682464454977,
"grad_norm": 3.919252872467041,
"learning_rate": 2.7429820078011215e-06,
"logits/chosen": 0.1973332315683365,
"logits/rejected": 0.26703959703445435,
"logps/chosen": -465.7378845214844,
"logps/rejected": -464.8862609863281,
"loss": 0.4067,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.0374157428741455,
"rewards/margins": 1.585756540298462,
"rewards/rejected": -3.6231720447540283,
"step": 289
},
{
"epoch": 1.3744075829383886,
"grad_norm": 2.8202483654022217,
"learning_rate": 2.7060770472217635e-06,
"logits/chosen": 0.22294802963733673,
"logits/rejected": 0.4262656569480896,
"logps/chosen": -464.995849609375,
"logps/rejected": -498.488037109375,
"loss": 0.254,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.8536605834960938,
"rewards/margins": 2.032717704772949,
"rewards/rejected": -3.886378288269043,
"step": 290
},
{
"epoch": 1.3791469194312795,
"grad_norm": 3.604506015777588,
"learning_rate": 2.669329701685719e-06,
"logits/chosen": 0.27663156390190125,
"logits/rejected": 0.06393498927354813,
"logps/chosen": -639.620849609375,
"logps/rejected": -598.626220703125,
"loss": 0.3477,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.7442690134048462,
"rewards/margins": 1.770451307296753,
"rewards/rejected": -3.5147204399108887,
"step": 291
},
{
"epoch": 1.3838862559241707,
"grad_norm": 4.539163112640381,
"learning_rate": 2.632742496096651e-06,
"logits/chosen": -0.019270386546850204,
"logits/rejected": -0.1904076784849167,
"logps/chosen": -312.65716552734375,
"logps/rejected": -371.73785400390625,
"loss": 0.3459,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.4120981693267822,
"rewards/margins": 1.6321998834609985,
"rewards/rejected": -3.0442981719970703,
"step": 292
},
{
"epoch": 1.3886255924170616,
"grad_norm": 3.6314282417297363,
"learning_rate": 2.59631794435503e-06,
"logits/chosen": 0.3927982747554779,
"logits/rejected": 0.5379042625427246,
"logps/chosen": -445.4381103515625,
"logps/rejected": -451.68243408203125,
"loss": 0.4375,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.476792573928833,
"rewards/margins": 1.4225971698760986,
"rewards/rejected": -2.8993897438049316,
"step": 293
},
{
"epoch": 1.3933649289099526,
"grad_norm": 2.562574625015259,
"learning_rate": 2.560058549185412e-06,
"logits/chosen": -0.25260213017463684,
"logits/rejected": -0.2509850561618805,
"logps/chosen": -364.51385498046875,
"logps/rejected": -367.390869140625,
"loss": 0.2561,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.4405786991119385,
"rewards/margins": 2.285914182662964,
"rewards/rejected": -3.7264928817749023,
"step": 294
},
{
"epoch": 1.3981042654028437,
"grad_norm": 3.246412754058838,
"learning_rate": 2.523966801964468e-06,
"logits/chosen": 0.0612456277012825,
"logits/rejected": 0.02570994198322296,
"logps/chosen": -399.10467529296875,
"logps/rejected": -454.84991455078125,
"loss": 0.3405,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.6198558807373047,
"rewards/margins": 1.5225186347961426,
"rewards/rejected": -3.1423745155334473,
"step": 295
},
{
"epoch": 1.4028436018957346,
"grad_norm": 3.30670166015625,
"learning_rate": 2.488045182549819e-06,
"logits/chosen": 0.0521719716489315,
"logits/rejected": 0.15891367197036743,
"logps/chosen": -341.91900634765625,
"logps/rejected": -461.2604064941406,
"loss": 0.2899,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.4679073095321655,
"rewards/margins": 2.0705184936523438,
"rewards/rejected": -3.5384254455566406,
"step": 296
},
{
"epoch": 1.4075829383886256,
"grad_norm": 4.309845924377441,
"learning_rate": 2.4522961591096246e-06,
"logits/chosen": -0.23573561012744904,
"logits/rejected": -0.42021211981773376,
"logps/chosen": -423.3658752441406,
"logps/rejected": -412.6981201171875,
"loss": 0.2852,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.66180419921875,
"rewards/margins": 1.9530575275421143,
"rewards/rejected": -3.614861488342285,
"step": 297
},
{
"epoch": 1.4123222748815165,
"grad_norm": 2.6515750885009766,
"learning_rate": 2.4167221879530063e-06,
"logits/chosen": 0.01921391859650612,
"logits/rejected": -0.13795889914035797,
"logps/chosen": -400.0218505859375,
"logps/rejected": -504.5172424316406,
"loss": 0.3005,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.8166048526763916,
"rewards/margins": 2.0179452896118164,
"rewards/rejected": -3.834550380706787,
"step": 298
},
{
"epoch": 1.4170616113744074,
"grad_norm": 3.4206173419952393,
"learning_rate": 2.381325713361283e-06,
"logits/chosen": -0.13432744145393372,
"logits/rejected": 0.1736375093460083,
"logps/chosen": -385.46893310546875,
"logps/rejected": -490.6543273925781,
"loss": 0.318,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.8471072912216187,
"rewards/margins": 2.0031168460845947,
"rewards/rejected": -3.850224018096924,
"step": 299
},
{
"epoch": 1.4218009478672986,
"grad_norm": 4.789295673370361,
"learning_rate": 2.3461091674199998e-06,
"logits/chosen": 0.19233857095241547,
"logits/rejected": 0.11930015683174133,
"logps/chosen": -492.05926513671875,
"logps/rejected": -572.2630615234375,
"loss": 0.3674,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.9044451713562012,
"rewards/margins": 2.156116008758545,
"rewards/rejected": -4.060561180114746,
"step": 300
},
{
"epoch": 1.4265402843601895,
"grad_norm": 3.0746889114379883,
"learning_rate": 2.311074969851852e-06,
"logits/chosen": 0.2876991927623749,
"logits/rejected": 0.5138101577758789,
"logps/chosen": -518.71630859375,
"logps/rejected": -563.4703979492188,
"loss": 0.385,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.7743141651153564,
"rewards/margins": 1.297276258468628,
"rewards/rejected": -3.0715904235839844,
"step": 301
},
{
"epoch": 1.4312796208530805,
"grad_norm": 4.4295654296875,
"learning_rate": 2.276225527850401e-06,
"logits/chosen": 0.11437669396400452,
"logits/rejected": 0.037944670766592026,
"logps/chosen": -390.86199951171875,
"logps/rejected": -374.7806701660156,
"loss": 0.4342,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.7658305168151855,
"rewards/margins": 1.4489760398864746,
"rewards/rejected": -3.21480655670166,
"step": 302
},
{
"epoch": 1.4360189573459716,
"grad_norm": 3.8123362064361572,
"learning_rate": 2.2415632359146855e-06,
"logits/chosen": 0.03791449964046478,
"logits/rejected": 0.27941614389419556,
"logps/chosen": -424.6317138671875,
"logps/rejected": -474.95452880859375,
"loss": 0.4074,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.4859094619750977,
"rewards/margins": 1.23709237575531,
"rewards/rejected": -3.723001480102539,
"step": 303
},
{
"epoch": 1.4407582938388626,
"grad_norm": 3.9473981857299805,
"learning_rate": 2.2070904756847023e-06,
"logits/chosen": -0.010878009721636772,
"logits/rejected": 0.011582762002944946,
"logps/chosen": -429.3747863769531,
"logps/rejected": -443.65106201171875,
"loss": 0.2943,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.9164514541625977,
"rewards/margins": 2.2701520919799805,
"rewards/rejected": -4.186603546142578,
"step": 304
},
{
"epoch": 1.4454976303317535,
"grad_norm": 3.664656400680542,
"learning_rate": 2.1728096157777516e-06,
"logits/chosen": 0.11286801099777222,
"logits/rejected": 0.1344190537929535,
"logps/chosen": -492.8208923339844,
"logps/rejected": -498.2126159667969,
"loss": 0.3742,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.920929193496704,
"rewards/margins": 1.7695159912109375,
"rewards/rejected": -3.6904454231262207,
"step": 305
},
{
"epoch": 1.4502369668246446,
"grad_norm": 3.3644583225250244,
"learning_rate": 2.1387230116257004e-06,
"logits/chosen": 0.29534584283828735,
"logits/rejected": 0.16407078504562378,
"logps/chosen": -429.521728515625,
"logps/rejected": -413.7642822265625,
"loss": 0.4601,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.3842999935150146,
"rewards/margins": 1.3130927085876465,
"rewards/rejected": -3.697392702102661,
"step": 306
},
{
"epoch": 1.4549763033175356,
"grad_norm": 3.42510986328125,
"learning_rate": 2.104833005313131e-06,
"logits/chosen": 0.04616807401180267,
"logits/rejected": 0.15600745379924774,
"logps/chosen": -403.85400390625,
"logps/rejected": -450.656982421875,
"loss": 0.3924,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.6164143085479736,
"rewards/margins": 2.050086498260498,
"rewards/rejected": -4.666500568389893,
"step": 307
},
{
"epoch": 1.4597156398104265,
"grad_norm": 4.1046271324157715,
"learning_rate": 2.071141925416431e-06,
"logits/chosen": -0.1839747130870819,
"logits/rejected": -0.08555589616298676,
"logps/chosen": -362.833740234375,
"logps/rejected": -407.0841064453125,
"loss": 0.3899,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.9180059432983398,
"rewards/margins": 1.3669946193695068,
"rewards/rejected": -3.2850005626678467,
"step": 308
},
{
"epoch": 1.4644549763033177,
"grad_norm": 4.176200866699219,
"learning_rate": 2.0376520868437838e-06,
"logits/chosen": 0.3360610902309418,
"logits/rejected": 0.32926660776138306,
"logps/chosen": -462.2861328125,
"logps/rejected": -485.8418884277344,
"loss": 0.4298,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.327528715133667,
"rewards/margins": 1.478387475013733,
"rewards/rejected": -3.8059163093566895,
"step": 309
},
{
"epoch": 1.4691943127962086,
"grad_norm": 3.149693727493286,
"learning_rate": 2.0043657906761144e-06,
"logits/chosen": 0.013693787157535553,
"logits/rejected": 0.25990521907806396,
"logps/chosen": -371.7657165527344,
"logps/rejected": -410.69744873046875,
"loss": 0.3518,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.1581993103027344,
"rewards/margins": 2.091628313064575,
"rewards/rejected": -4.2498273849487305,
"step": 310
},
{
"epoch": 1.4739336492890995,
"grad_norm": 2.863271951675415,
"learning_rate": 1.971285324008994e-06,
"logits/chosen": 0.43576109409332275,
"logits/rejected": 0.3071168065071106,
"logps/chosen": -366.2781677246094,
"logps/rejected": -386.5714111328125,
"loss": 0.2796,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.3572580814361572,
"rewards/margins": 1.936234951019287,
"rewards/rejected": -3.2934930324554443,
"step": 311
},
{
"epoch": 1.4786729857819905,
"grad_norm": 3.9290127754211426,
"learning_rate": 1.938412959795472e-06,
"logits/chosen": 0.11994842439889908,
"logits/rejected": 0.14935952425003052,
"logps/chosen": -545.4857177734375,
"logps/rejected": -637.3936767578125,
"loss": 0.3632,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.8452956676483154,
"rewards/margins": 1.575514554977417,
"rewards/rejected": -4.420809745788574,
"step": 312
},
{
"epoch": 1.4834123222748814,
"grad_norm": 2.9018099308013916,
"learning_rate": 1.9057509566899268e-06,
"logits/chosen": 0.07368522882461548,
"logits/rejected": 0.16368550062179565,
"logps/chosen": -349.2361145019531,
"logps/rejected": -417.56201171875,
"loss": 0.2909,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.6519646644592285,
"rewards/margins": 1.7734508514404297,
"rewards/rejected": -3.425415515899658,
"step": 313
},
{
"epoch": 1.4881516587677726,
"grad_norm": 3.14565110206604,
"learning_rate": 1.873301558892855e-06,
"logits/chosen": 0.06656618416309357,
"logits/rejected": 0.12518161535263062,
"logps/chosen": -396.86651611328125,
"logps/rejected": -444.3167724609375,
"loss": 0.364,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.1594903469085693,
"rewards/margins": 1.6573652029037476,
"rewards/rejected": -3.8168554306030273,
"step": 314
},
{
"epoch": 1.4928909952606635,
"grad_norm": 3.205932855606079,
"learning_rate": 1.8410669959966804e-06,
"logits/chosen": 0.07552038878202438,
"logits/rejected": 0.12510335445404053,
"logps/chosen": -420.3112487792969,
"logps/rejected": -473.9259033203125,
"loss": 0.3091,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.9273097515106201,
"rewards/margins": 1.7265255451202393,
"rewards/rejected": -3.6538352966308594,
"step": 315
},
{
"epoch": 1.4976303317535544,
"grad_norm": 3.193234443664551,
"learning_rate": 1.8090494828325634e-06,
"logits/chosen": -0.24580290913581848,
"logits/rejected": -0.2829708456993103,
"logps/chosen": -416.750732421875,
"logps/rejected": -461.3362731933594,
"loss": 0.2596,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.465597152709961,
"rewards/margins": 2.6725990772247314,
"rewards/rejected": -4.1381964683532715,
"step": 316
},
{
"epoch": 1.5023696682464456,
"grad_norm": 2.840527057647705,
"learning_rate": 1.7772512193182096e-06,
"logits/chosen": 0.32139959931373596,
"logits/rejected": 0.3541162610054016,
"logps/chosen": -432.45733642578125,
"logps/rejected": -466.65545654296875,
"loss": 0.3187,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.2416555881500244,
"rewards/margins": 1.9411081075668335,
"rewards/rejected": -4.182764053344727,
"step": 317
},
{
"epoch": 1.5071090047393365,
"grad_norm": 1.9531362056732178,
"learning_rate": 1.745674390306722e-06,
"logits/chosen": 0.16892415285110474,
"logits/rejected": 0.2631104588508606,
"logps/chosen": -413.987548828125,
"logps/rejected": -437.920166015625,
"loss": 0.2867,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.9473943710327148,
"rewards/margins": 2.0647332668304443,
"rewards/rejected": -4.012127876281738,
"step": 318
},
{
"epoch": 1.5118483412322274,
"grad_norm": 2.907794237136841,
"learning_rate": 1.7143211654364761e-06,
"logits/chosen": 0.061596665531396866,
"logits/rejected": 0.03803150728344917,
"logps/chosen": -405.4927062988281,
"logps/rejected": -378.6966247558594,
"loss": 0.2711,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.5833370685577393,
"rewards/margins": 2.4727444648742676,
"rewards/rejected": -4.056081771850586,
"step": 319
},
{
"epoch": 1.5165876777251186,
"grad_norm": 5.282046794891357,
"learning_rate": 1.6831936989820507e-06,
"logits/chosen": 0.13025835156440735,
"logits/rejected": -0.02494555525481701,
"logps/chosen": -419.9814758300781,
"logps/rejected": -380.5059814453125,
"loss": 0.381,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.9292106628417969,
"rewards/margins": 1.795340895652771,
"rewards/rejected": -3.7245516777038574,
"step": 320
},
{
"epoch": 1.5213270142180095,
"grad_norm": 3.417613983154297,
"learning_rate": 1.6522941297061996e-06,
"logits/chosen": 0.20090161263942719,
"logits/rejected": 0.2483268678188324,
"logps/chosen": -479.95306396484375,
"logps/rejected": -536.9041137695312,
"loss": 0.3483,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.9097940921783447,
"rewards/margins": 1.6619257926940918,
"rewards/rejected": -3.5717201232910156,
"step": 321
},
{
"epoch": 1.5260663507109005,
"grad_norm": 4.454870700836182,
"learning_rate": 1.6216245807129005e-06,
"logits/chosen": 0.013529401272535324,
"logits/rejected": 0.07809659838676453,
"logps/chosen": -399.9565124511719,
"logps/rejected": -465.4268798828125,
"loss": 0.4229,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.1004106998443604,
"rewards/margins": 2.0842761993408203,
"rewards/rejected": -4.18468713760376,
"step": 322
},
{
"epoch": 1.5308056872037916,
"grad_norm": 3.5815956592559814,
"learning_rate": 1.5911871593014838e-06,
"logits/chosen": 0.4060709774494171,
"logits/rejected": 0.6513478755950928,
"logps/chosen": -345.8062438964844,
"logps/rejected": -354.0713806152344,
"loss": 0.4097,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.48841392993927,
"rewards/margins": 1.9469943046569824,
"rewards/rejected": -3.435408115386963,
"step": 323
},
{
"epoch": 1.5355450236966823,
"grad_norm": 2.94301176071167,
"learning_rate": 1.5609839568218248e-06,
"logits/chosen": 0.16032955050468445,
"logits/rejected": 0.19198182225227356,
"logps/chosen": -486.0198669433594,
"logps/rejected": -518.156494140625,
"loss": 0.2529,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.108952760696411,
"rewards/margins": 2.007476329803467,
"rewards/rejected": -4.116428852081299,
"step": 324
},
{
"epoch": 1.5402843601895735,
"grad_norm": 3.2696046829223633,
"learning_rate": 1.5310170485306675e-06,
"logits/chosen": 0.5102041959762573,
"logits/rejected": 0.533448338508606,
"logps/chosen": -525.1609497070312,
"logps/rejected": -560.2240600585938,
"loss": 0.2644,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.106445074081421,
"rewards/margins": 2.0785865783691406,
"rewards/rejected": -4.185031890869141,
"step": 325
},
{
"epoch": 1.5450236966824644,
"grad_norm": 2.56247878074646,
"learning_rate": 1.5012884934490168e-06,
"logits/chosen": 0.10596761107444763,
"logits/rejected": -0.22906875610351562,
"logps/chosen": -428.1797180175781,
"logps/rejected": -428.2669372558594,
"loss": 0.3297,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.578533411026001,
"rewards/margins": 1.8542442321777344,
"rewards/rejected": -3.4327776432037354,
"step": 326
},
{
"epoch": 1.5497630331753554,
"grad_norm": 3.698256731033325,
"learning_rate": 1.4718003342206722e-06,
"logits/chosen": -0.03314054012298584,
"logits/rejected": 0.09218861162662506,
"logps/chosen": -460.627685546875,
"logps/rejected": -545.28125,
"loss": 0.4134,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.307816743850708,
"rewards/margins": 1.7128374576568604,
"rewards/rejected": -4.020654201507568,
"step": 327
},
{
"epoch": 1.5545023696682465,
"grad_norm": 3.7442150115966797,
"learning_rate": 1.44255459697188e-06,
"logits/chosen": 0.434778094291687,
"logits/rejected": 0.3917992115020752,
"logps/chosen": -400.646484375,
"logps/rejected": -504.3604736328125,
"loss": 0.4126,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.231884002685547,
"rewards/margins": 1.6030129194259644,
"rewards/rejected": -3.8348968029022217,
"step": 328
},
{
"epoch": 1.5592417061611374,
"grad_norm": 2.8240530490875244,
"learning_rate": 1.413553291172106e-06,
"logits/chosen": 0.16396990418434143,
"logits/rejected": -0.013539992272853851,
"logps/chosen": -416.2026062011719,
"logps/rejected": -393.49652099609375,
"loss": 0.2875,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.2988057136535645,
"rewards/margins": 1.8778635263442993,
"rewards/rejected": -3.176669120788574,
"step": 329
},
{
"epoch": 1.5639810426540284,
"grad_norm": 3.9064056873321533,
"learning_rate": 1.3847984094959843e-06,
"logits/chosen": -0.3242209553718567,
"logits/rejected": -0.22871044278144836,
"logps/chosen": -350.4272155761719,
"logps/rejected": -461.5364990234375,
"loss": 0.3061,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.0482442378997803,
"rewards/margins": 2.088869571685791,
"rewards/rejected": -4.137113571166992,
"step": 330
},
{
"epoch": 1.5687203791469195,
"grad_norm": 2.855518341064453,
"learning_rate": 1.3562919276863846e-06,
"logits/chosen": -0.12499962747097015,
"logits/rejected": 0.03998437523841858,
"logps/chosen": -443.1326904296875,
"logps/rejected": -450.186767578125,
"loss": 0.3047,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.8554315567016602,
"rewards/margins": 1.7155771255493164,
"rewards/rejected": -3.5710082054138184,
"step": 331
},
{
"epoch": 1.5734597156398105,
"grad_norm": 2.2054951190948486,
"learning_rate": 1.3280358044186648e-06,
"logits/chosen": 0.0808248519897461,
"logits/rejected": 0.0358705073595047,
"logps/chosen": -460.5960693359375,
"logps/rejected": -517.7080078125,
"loss": 0.2432,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.286323070526123,
"rewards/margins": 2.1876108646392822,
"rewards/rejected": -4.473934173583984,
"step": 332
},
{
"epoch": 1.5781990521327014,
"grad_norm": 4.370570182800293,
"learning_rate": 1.3000319811660967e-06,
"logits/chosen": -0.05356225371360779,
"logits/rejected": -0.1574479043483734,
"logps/chosen": -510.5133361816406,
"logps/rejected": -494.0157470703125,
"loss": 0.4098,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.589010238647461,
"rewards/margins": 1.8569483757019043,
"rewards/rejected": -3.4459586143493652,
"step": 333
},
{
"epoch": 1.5829383886255926,
"grad_norm": 3.086522340774536,
"learning_rate": 1.2722823820664577e-06,
"logits/chosen": 0.25446993112564087,
"logits/rejected": 0.10102240741252899,
"logps/chosen": -556.471923828125,
"logps/rejected": -558.4532470703125,
"loss": 0.3263,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7523441314697266,
"rewards/margins": 1.519374966621399,
"rewards/rejected": -4.271718978881836,
"step": 334
},
{
"epoch": 1.5876777251184833,
"grad_norm": 6.325353145599365,
"learning_rate": 1.2447889137898295e-06,
"logits/chosen": 0.0681871697306633,
"logits/rejected": 0.07994731515645981,
"logps/chosen": -447.1323547363281,
"logps/rejected": -446.35675048828125,
"loss": 0.5166,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.193911075592041,
"rewards/margins": 1.1045620441436768,
"rewards/rejected": -3.2984731197357178,
"step": 335
},
{
"epoch": 1.5924170616113744,
"grad_norm": 4.334328651428223,
"learning_rate": 1.2175534654075887e-06,
"logits/chosen": -0.12054158747196198,
"logits/rejected": -0.25014710426330566,
"logps/chosen": -368.5024108886719,
"logps/rejected": -418.28466796875,
"loss": 0.3067,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.830749273300171,
"rewards/margins": 2.00579571723938,
"rewards/rejected": -3.836544990539551,
"step": 336
},
{
"epoch": 1.5971563981042654,
"grad_norm": 3.325434923171997,
"learning_rate": 1.190577908262614e-06,
"logits/chosen": 0.1261437088251114,
"logits/rejected": 0.15501365065574646,
"logps/chosen": -501.2939453125,
"logps/rejected": -559.4841918945312,
"loss": 0.3501,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.977611780166626,
"rewards/margins": 1.6551153659820557,
"rewards/rejected": -3.6327271461486816,
"step": 337
},
{
"epoch": 1.6018957345971563,
"grad_norm": 5.469727993011475,
"learning_rate": 1.1638640958407e-06,
"logits/chosen": 0.11842088401317596,
"logits/rejected": 0.2777405381202698,
"logps/chosen": -422.850341796875,
"logps/rejected": -439.694580078125,
"loss": 0.4922,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.9282774925231934,
"rewards/margins": 1.2889878749847412,
"rewards/rejected": -3.2172651290893555,
"step": 338
},
{
"epoch": 1.6066350710900474,
"grad_norm": 4.092334747314453,
"learning_rate": 1.1374138636432054e-06,
"logits/chosen": -0.13197553157806396,
"logits/rejected": -0.16363224387168884,
"logps/chosen": -526.3876953125,
"logps/rejected": -595.45947265625,
"loss": 0.4096,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.6760362386703491,
"rewards/margins": 1.171120524406433,
"rewards/rejected": -2.8471570014953613,
"step": 339
},
{
"epoch": 1.6113744075829384,
"grad_norm": 3.614046812057495,
"learning_rate": 1.1112290290609444e-06,
"logits/chosen": -0.2544490098953247,
"logits/rejected": -0.02984345145523548,
"logps/chosen": -474.2083740234375,
"logps/rejected": -504.4893493652344,
"loss": 0.3167,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.2328014373779297,
"rewards/margins": 1.6542941331863403,
"rewards/rejected": -3.8870954513549805,
"step": 340
},
{
"epoch": 1.6161137440758293,
"grad_norm": 3.6634645462036133,
"learning_rate": 1.085311391249299e-06,
"logits/chosen": 0.023314230144023895,
"logits/rejected": 0.059964969754219055,
"logps/chosen": -469.96905517578125,
"logps/rejected": -568.8093872070312,
"loss": 0.3357,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.085527181625366,
"rewards/margins": 1.6656991243362427,
"rewards/rejected": -3.7512261867523193,
"step": 341
},
{
"epoch": 1.6208530805687205,
"grad_norm": 3.1424379348754883,
"learning_rate": 1.0596627310046165e-06,
"logits/chosen": 0.31347164511680603,
"logits/rejected": 0.1853667050600052,
"logps/chosen": -552.0761108398438,
"logps/rejected": -505.68707275390625,
"loss": 0.3661,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.9287023544311523,
"rewards/margins": 1.3750154972076416,
"rewards/rejected": -3.303718090057373,
"step": 342
},
{
"epoch": 1.6255924170616114,
"grad_norm": 2.257039785385132,
"learning_rate": 1.0342848106418368e-06,
"logits/chosen": -0.1234850063920021,
"logits/rejected": 0.1165139451622963,
"logps/chosen": -371.6340637207031,
"logps/rejected": -417.32159423828125,
"loss": 0.2868,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.7357150316238403,
"rewards/margins": 1.8563973903656006,
"rewards/rejected": -3.5921125411987305,
"step": 343
},
{
"epoch": 1.6303317535545023,
"grad_norm": 3.4644458293914795,
"learning_rate": 1.0091793738734113e-06,
"logits/chosen": 0.21419385075569153,
"logits/rejected": -0.015899434685707092,
"logps/chosen": -417.99237060546875,
"logps/rejected": -396.4125061035156,
"loss": 0.336,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.6989412307739258,
"rewards/margins": 1.975459098815918,
"rewards/rejected": -3.6744003295898438,
"step": 344
},
{
"epoch": 1.6350710900473935,
"grad_norm": 3.3371543884277344,
"learning_rate": 9.843481456894949e-07,
"logits/chosen": 0.14167696237564087,
"logits/rejected": 0.13582485914230347,
"logps/chosen": -405.79290771484375,
"logps/rejected": -467.8961486816406,
"loss": 0.3739,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.369532823562622,
"rewards/margins": 1.6152913570404053,
"rewards/rejected": -3.9848241806030273,
"step": 345
},
{
"epoch": 1.6398104265402842,
"grad_norm": 3.4584085941314697,
"learning_rate": 9.59792832239415e-07,
"logits/chosen": 0.1842474341392517,
"logits/rejected": 0.36914995312690735,
"logps/chosen": -430.4465026855469,
"logps/rejected": -491.39923095703125,
"loss": 0.4111,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.37587308883667,
"rewards/margins": 1.4323196411132812,
"rewards/rejected": -3.808192729949951,
"step": 346
},
{
"epoch": 1.6445497630331753,
"grad_norm": 3.6146063804626465,
"learning_rate": 9.35515120714447e-07,
"logits/chosen": 0.21485333144664764,
"logits/rejected": 0.3836483955383301,
"logps/chosen": -385.3580017089844,
"logps/rejected": -413.75958251953125,
"loss": 0.3057,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.34371280670166,
"rewards/margins": 1.8871243000030518,
"rewards/rejected": -4.230837345123291,
"step": 347
},
{
"epoch": 1.6492890995260665,
"grad_norm": 2.6541919708251953,
"learning_rate": 9.115166792318858e-07,
"logits/chosen": -0.2523440718650818,
"logits/rejected": -0.3873641788959503,
"logps/chosen": -405.66363525390625,
"logps/rejected": -387.85675048828125,
"loss": 0.2713,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.7661632299423218,
"rewards/margins": 2.105201244354248,
"rewards/rejected": -3.8713645935058594,
"step": 348
},
{
"epoch": 1.6540284360189572,
"grad_norm": 2.7787058353424072,
"learning_rate": 8.877991567204353e-07,
"logits/chosen": 0.04583475738763809,
"logits/rejected": 0.22186128795146942,
"logps/chosen": -469.5939025878906,
"logps/rejected": -490.9723815917969,
"loss": 0.3186,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.797175407409668,
"rewards/margins": 1.9968000650405884,
"rewards/rejected": -3.793975830078125,
"step": 349
},
{
"epoch": 1.6587677725118484,
"grad_norm": 3.4025254249572754,
"learning_rate": 8.643641828069004e-07,
"logits/chosen": 0.1861867904663086,
"logits/rejected": 0.26269257068634033,
"logps/chosen": -388.3940734863281,
"logps/rejected": -477.13372802734375,
"loss": 0.3042,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.0583114624023438,
"rewards/margins": 2.277298927307129,
"rewards/rejected": -4.335610389709473,
"step": 350
},
{
"epoch": 1.6635071090047393,
"grad_norm": 2.7238893508911133,
"learning_rate": 8.412133677042239e-07,
"logits/chosen": 0.2202744483947754,
"logits/rejected": 0.14431580901145935,
"logps/chosen": -494.9244689941406,
"logps/rejected": -511.6275939941406,
"loss": 0.2694,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.7077330350875854,
"rewards/margins": 2.26680850982666,
"rewards/rejected": -3.974541425704956,
"step": 351
},
{
"epoch": 1.6682464454976302,
"grad_norm": 2.596195697784424,
"learning_rate": 8.183483021008498e-07,
"logits/chosen": 0.059778667986392975,
"logits/rejected": 0.16687698662281036,
"logps/chosen": -459.7418518066406,
"logps/rejected": -528.6278076171875,
"loss": 0.3188,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.31968092918396,
"rewards/margins": 1.6766258478164673,
"rewards/rejected": -3.996306896209717,
"step": 352
},
{
"epoch": 1.6729857819905214,
"grad_norm": 3.991448402404785,
"learning_rate": 7.957705570514163e-07,
"logits/chosen": 0.13453525304794312,
"logits/rejected": 0.25876981019973755,
"logps/chosen": -451.383056640625,
"logps/rejected": -449.3511657714844,
"loss": 0.4566,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.6044600009918213,
"rewards/margins": 1.0652509927749634,
"rewards/rejected": -2.669711112976074,
"step": 353
},
{
"epoch": 1.6777251184834123,
"grad_norm": 3.8690874576568604,
"learning_rate": 7.734816838688247e-07,
"logits/chosen": 0.13873904943466187,
"logits/rejected": -0.005100712180137634,
"logps/chosen": -468.6593017578125,
"logps/rejected": -474.06884765625,
"loss": 0.283,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.1768066883087158,
"rewards/margins": 2.2671823501586914,
"rewards/rejected": -3.4439892768859863,
"step": 354
},
{
"epoch": 1.6824644549763033,
"grad_norm": 3.6120941638946533,
"learning_rate": 7.51483214017637e-07,
"logits/chosen": 0.019669074565172195,
"logits/rejected": 0.13328783214092255,
"logps/chosen": -476.302001953125,
"logps/rejected": -512.1363525390625,
"loss": 0.3317,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.948580265045166,
"rewards/margins": 1.781925916671753,
"rewards/rejected": -3.73050594329834,
"step": 355
},
{
"epoch": 1.6872037914691944,
"grad_norm": 3.360703229904175,
"learning_rate": 7.297766590088529e-07,
"logits/chosen": 0.19215039908885956,
"logits/rejected": 0.12218157947063446,
"logps/chosen": -403.3554382324219,
"logps/rejected": -421.522705078125,
"loss": 0.3112,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.5849626064300537,
"rewards/margins": 1.5150251388549805,
"rewards/rejected": -3.0999879837036133,
"step": 356
},
{
"epoch": 1.6919431279620853,
"grad_norm": 4.554477691650391,
"learning_rate": 7.083635102960584e-07,
"logits/chosen": -0.1721559464931488,
"logits/rejected": 0.14503641426563263,
"logps/chosen": -373.804443359375,
"logps/rejected": -457.3600769042969,
"loss": 0.3669,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.9074959754943848,
"rewards/margins": 1.5698236227035522,
"rewards/rejected": -3.4773197174072266,
"step": 357
},
{
"epoch": 1.6966824644549763,
"grad_norm": 5.0574140548706055,
"learning_rate": 6.8724523917294e-07,
"logits/chosen": 0.19052550196647644,
"logits/rejected": 0.1522517204284668,
"logps/chosen": -394.22149658203125,
"logps/rejected": -469.25360107421875,
"loss": 0.3164,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.4775445461273193,
"rewards/margins": 1.945703148841858,
"rewards/rejected": -3.423247814178467,
"step": 358
},
{
"epoch": 1.7014218009478674,
"grad_norm": 3.6440603733062744,
"learning_rate": 6.664232966721995e-07,
"logits/chosen": -0.30218714475631714,
"logits/rejected": -0.21477612853050232,
"logps/chosen": -362.7777099609375,
"logps/rejected": -383.90869140625,
"loss": 0.2301,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.3972433805465698,
"rewards/margins": 2.2594199180603027,
"rewards/rejected": -3.656663179397583,
"step": 359
},
{
"epoch": 1.7061611374407581,
"grad_norm": 3.930510997772217,
"learning_rate": 6.458991134658487e-07,
"logits/chosen": -0.056913457810878754,
"logits/rejected": -0.14112192392349243,
"logps/chosen": -438.1789855957031,
"logps/rejected": -482.021240234375,
"loss": 0.2415,
"rewards/accuracies": 0.9375,
"rewards/chosen": -1.4954298734664917,
"rewards/margins": 2.081784248352051,
"rewards/rejected": -3.577214002609253,
"step": 360
},
{
"epoch": 1.7109004739336493,
"grad_norm": 3.474404811859131,
"learning_rate": 6.256740997669142e-07,
"logits/chosen": 0.305749773979187,
"logits/rejected": 0.372690349817276,
"logps/chosen": -465.25396728515625,
"logps/rejected": -520.6639404296875,
"loss": 0.3113,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.961679220199585,
"rewards/margins": 1.8247203826904297,
"rewards/rejected": -3.7863998413085938,
"step": 361
},
{
"epoch": 1.7156398104265402,
"grad_norm": 4.286850929260254,
"learning_rate": 6.057496452325345e-07,
"logits/chosen": 0.06337082386016846,
"logits/rejected": 0.03823350369930267,
"logps/chosen": -510.4189453125,
"logps/rejected": -534.214111328125,
"loss": 0.3706,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.2660622596740723,
"rewards/margins": 1.5975626707077026,
"rewards/rejected": -3.8636245727539062,
"step": 362
},
{
"epoch": 1.7203791469194312,
"grad_norm": 7.2974042892456055,
"learning_rate": 5.861271188684819e-07,
"logits/chosen": -0.47425615787506104,
"logits/rejected": -0.38017037510871887,
"logps/chosen": -471.4033508300781,
"logps/rejected": -448.49322509765625,
"loss": 0.4664,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.868884801864624,
"rewards/margins": 1.9620721340179443,
"rewards/rejected": -3.8309569358825684,
"step": 363
},
{
"epoch": 1.7251184834123223,
"grad_norm": 2.614375352859497,
"learning_rate": 5.668078689351008e-07,
"logits/chosen": 0.08667676150798798,
"logits/rejected": 0.026178985834121704,
"logps/chosen": -527.921142578125,
"logps/rejected": -473.9797058105469,
"loss": 0.3323,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.2779064178466797,
"rewards/margins": 1.8057652711868286,
"rewards/rejected": -3.0836715698242188,
"step": 364
},
{
"epoch": 1.7298578199052133,
"grad_norm": 4.204647064208984,
"learning_rate": 5.477932228546574e-07,
"logits/chosen": -0.001731783151626587,
"logits/rejected": -0.11266829073429108,
"logps/chosen": -507.29205322265625,
"logps/rejected": -497.1200256347656,
"loss": 0.3806,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.4385986328125,
"rewards/margins": 1.3973621129989624,
"rewards/rejected": -3.835960626602173,
"step": 365
},
{
"epoch": 1.7345971563981042,
"grad_norm": 4.324245452880859,
"learning_rate": 5.290844871201483e-07,
"logits/chosen": 0.09860187768936157,
"logits/rejected": 0.0050681233406066895,
"logps/chosen": -489.361083984375,
"logps/rejected": -526.6915893554688,
"loss": 0.4866,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.8216359615325928,
"rewards/margins": 1.4107340574264526,
"rewards/rejected": -3.232370138168335,
"step": 366
},
{
"epoch": 1.7393364928909953,
"grad_norm": 4.808085918426514,
"learning_rate": 5.106829472055203e-07,
"logits/chosen": 0.05839775502681732,
"logits/rejected": -0.02708090841770172,
"logps/chosen": -545.265869140625,
"logps/rejected": -521.0439453125,
"loss": 0.4935,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.1644363403320312,
"rewards/margins": 1.105941653251648,
"rewards/rejected": -3.2703781127929688,
"step": 367
},
{
"epoch": 1.7440758293838863,
"grad_norm": 4.010900020599365,
"learning_rate": 4.925898674773488e-07,
"logits/chosen": 0.7153588533401489,
"logits/rejected": 0.9175931215286255,
"logps/chosen": -495.9670104980469,
"logps/rejected": -522.7073364257812,
"loss": 0.4523,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.1978981494903564,
"rewards/margins": 1.3793647289276123,
"rewards/rejected": -3.5772628784179688,
"step": 368
},
{
"epoch": 1.7488151658767772,
"grad_norm": 4.280209064483643,
"learning_rate": 4.7480649110796693e-07,
"logits/chosen": -0.2100297212600708,
"logits/rejected": -0.035903140902519226,
"logps/chosen": -503.14190673828125,
"logps/rejected": -545.4959716796875,
"loss": 0.352,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.9035875797271729,
"rewards/margins": 2.3268721103668213,
"rewards/rejected": -4.230460166931152,
"step": 369
},
{
"epoch": 1.7535545023696684,
"grad_norm": 3.580965518951416,
"learning_rate": 4.5733403999004167e-07,
"logits/chosen": 0.03610946238040924,
"logits/rejected": 0.12372472882270813,
"logps/chosen": -351.2373352050781,
"logps/rejected": -424.42236328125,
"loss": 0.3629,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.7467509508132935,
"rewards/margins": 2.5998826026916504,
"rewards/rejected": -4.346633434295654,
"step": 370
},
{
"epoch": 1.758293838862559,
"grad_norm": 5.5822529792785645,
"learning_rate": 4.401737146526219e-07,
"logits/chosen": 0.11202114820480347,
"logits/rejected": 0.2623312473297119,
"logps/chosen": -404.4681701660156,
"logps/rejected": -468.0955810546875,
"loss": 0.4676,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.091884136199951,
"rewards/margins": 1.5518180131912231,
"rewards/rejected": -3.6437020301818848,
"step": 371
},
{
"epoch": 1.7630331753554502,
"grad_norm": 3.6275458335876465,
"learning_rate": 4.2332669417864734e-07,
"logits/chosen": 0.2248685210943222,
"logits/rejected": 0.5064141750335693,
"logps/chosen": -422.96673583984375,
"logps/rejected": -538.043701171875,
"loss": 0.343,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.6720224618911743,
"rewards/margins": 1.9135386943817139,
"rewards/rejected": -3.5855612754821777,
"step": 372
},
{
"epoch": 1.7677725118483414,
"grad_norm": 2.365963935852051,
"learning_rate": 4.067941361239386e-07,
"logits/chosen": 0.22563423216342926,
"logits/rejected": 0.27167263627052307,
"logps/chosen": -459.28155517578125,
"logps/rejected": -489.7861633300781,
"loss": 0.2706,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.141444206237793,
"rewards/margins": 2.016284465789795,
"rewards/rejected": -4.157728672027588,
"step": 373
},
{
"epoch": 1.772511848341232,
"grad_norm": 3.179948329925537,
"learning_rate": 3.905771764376554e-07,
"logits/chosen": -0.04605560749769211,
"logits/rejected": -0.13275614380836487,
"logps/chosen": -500.1363220214844,
"logps/rejected": -484.475341796875,
"loss": 0.3462,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.7449672222137451,
"rewards/margins": 1.7927236557006836,
"rewards/rejected": -3.5376906394958496,
"step": 374
},
{
"epoch": 1.7772511848341233,
"grad_norm": 3.8642663955688477,
"learning_rate": 3.746769293842506e-07,
"logits/chosen": 0.32506126165390015,
"logits/rejected": 0.25315555930137634,
"logps/chosen": -408.6797790527344,
"logps/rejected": -459.6020202636719,
"loss": 0.3154,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.6381193399429321,
"rewards/margins": 2.112222671508789,
"rewards/rejected": -3.7503418922424316,
"step": 375
},
{
"epoch": 1.7819905213270142,
"grad_norm": 1.845032811164856,
"learning_rate": 3.590944874669089e-07,
"logits/chosen": 0.16336679458618164,
"logits/rejected": -0.06345739960670471,
"logps/chosen": -456.1724853515625,
"logps/rejected": -412.36376953125,
"loss": 0.2595,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.056734561920166,
"rewards/margins": 2.2098302841186523,
"rewards/rejected": -4.26656436920166,
"step": 376
},
{
"epoch": 1.7867298578199051,
"grad_norm": 3.6954963207244873,
"learning_rate": 3.4383092135247543e-07,
"logits/chosen": -0.042060717940330505,
"logits/rejected": -0.28105732798576355,
"logps/chosen": -455.3412170410156,
"logps/rejected": -469.8854064941406,
"loss": 0.3753,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.4552671909332275,
"rewards/margins": 1.3413841724395752,
"rewards/rejected": -3.7966513633728027,
"step": 377
},
{
"epoch": 1.7914691943127963,
"grad_norm": 3.0183236598968506,
"learning_rate": 3.288872797979009e-07,
"logits/chosen": -0.005820997059345245,
"logits/rejected": 0.14580360054969788,
"logps/chosen": -479.91644287109375,
"logps/rejected": -515.1129760742188,
"loss": 0.285,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.6726536750793457,
"rewards/margins": 2.280189275741577,
"rewards/rejected": -3.952843189239502,
"step": 378
},
{
"epoch": 1.7962085308056872,
"grad_norm": 3.68314528465271,
"learning_rate": 3.1426458957817154e-07,
"logits/chosen": 0.29249462485313416,
"logits/rejected": 0.28890830278396606,
"logps/chosen": -384.9244384765625,
"logps/rejected": -381.51416015625,
"loss": 0.3259,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.2649922370910645,
"rewards/margins": 1.6742699146270752,
"rewards/rejected": -3.9392619132995605,
"step": 379
},
{
"epoch": 1.8009478672985781,
"grad_norm": 4.089728832244873,
"learning_rate": 2.999638554157636e-07,
"logits/chosen": 0.16874819993972778,
"logits/rejected": 0.33503326773643494,
"logps/chosen": -405.65863037109375,
"logps/rejected": -436.0166320800781,
"loss": 0.331,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.2233431339263916,
"rewards/margins": 1.7847846746444702,
"rewards/rejected": -3.0081279277801514,
"step": 380
},
{
"epoch": 1.8056872037914693,
"grad_norm": 3.578709363937378,
"learning_rate": 2.859860599116127e-07,
"logits/chosen": 0.0348060317337513,
"logits/rejected": -0.028133656829595566,
"logps/chosen": -424.2684020996094,
"logps/rejected": -405.78546142578125,
"loss": 0.2824,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.338637113571167,
"rewards/margins": 2.2470810413360596,
"rewards/rejected": -3.5857181549072266,
"step": 381
},
{
"epoch": 1.8104265402843602,
"grad_norm": 3.0040953159332275,
"learning_rate": 2.723321634775927e-07,
"logits/chosen": 0.14009594917297363,
"logits/rejected": 0.09075860679149628,
"logps/chosen": -466.3572692871094,
"logps/rejected": -451.00390625,
"loss": 0.34,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.8811650276184082,
"rewards/margins": 1.9055883884429932,
"rewards/rejected": -3.7867536544799805,
"step": 382
},
{
"epoch": 1.8151658767772512,
"grad_norm": 3.576838731765747,
"learning_rate": 2.590031042705304e-07,
"logits/chosen": 0.00659569725394249,
"logits/rejected": -0.039983510971069336,
"logps/chosen": -386.4693908691406,
"logps/rejected": -437.7843933105469,
"loss": 0.3241,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.694931983947754,
"rewards/margins": 2.0972633361816406,
"rewards/rejected": -3.7921950817108154,
"step": 383
},
{
"epoch": 1.8199052132701423,
"grad_norm": 3.0422523021698,
"learning_rate": 2.459997981277423e-07,
"logits/chosen": 0.3022235333919525,
"logits/rejected": 0.3308006227016449,
"logps/chosen": -515.8807373046875,
"logps/rejected": -566.62646484375,
"loss": 0.2557,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.9875129461288452,
"rewards/margins": 2.3573031425476074,
"rewards/rejected": -4.344816207885742,
"step": 384
},
{
"epoch": 1.824644549763033,
"grad_norm": 3.56526780128479,
"learning_rate": 2.3332313850411238e-07,
"logits/chosen": 0.15668299794197083,
"logits/rejected": 0.3108135461807251,
"logps/chosen": -376.9952697753906,
"logps/rejected": -416.5478515625,
"loss": 0.4554,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.9508757591247559,
"rewards/margins": 1.343916893005371,
"rewards/rejected": -3.294792652130127,
"step": 385
},
{
"epoch": 1.8293838862559242,
"grad_norm": 4.2836503982543945,
"learning_rate": 2.2097399641069662e-07,
"logits/chosen": -0.1609744131565094,
"logits/rejected": -0.16733995079994202,
"logps/chosen": -461.4750061035156,
"logps/rejected": -506.1349792480469,
"loss": 0.4442,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.8771613836288452,
"rewards/margins": 1.3958663940429688,
"rewards/rejected": -3.2730276584625244,
"step": 386
},
{
"epoch": 1.8341232227488151,
"grad_norm": 5.192793369293213,
"learning_rate": 2.0895322035487942e-07,
"logits/chosen": -0.011422211304306984,
"logits/rejected": -0.3574138581752777,
"logps/chosen": -494.67852783203125,
"logps/rejected": -470.2755126953125,
"loss": 0.4215,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.6957664489746094,
"rewards/margins": 1.5437688827514648,
"rewards/rejected": -3.2395355701446533,
"step": 387
},
{
"epoch": 1.838862559241706,
"grad_norm": 4.824533462524414,
"learning_rate": 1.972616362820745e-07,
"logits/chosen": 0.04909311980009079,
"logits/rejected": -0.057477012276649475,
"logps/chosen": -415.3355712890625,
"logps/rejected": -469.35955810546875,
"loss": 0.3899,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.789413332939148,
"rewards/margins": 1.886869192123413,
"rewards/rejected": -3.6762826442718506,
"step": 388
},
{
"epoch": 1.8436018957345972,
"grad_norm": 3.7289130687713623,
"learning_rate": 1.8590004751896872e-07,
"logits/chosen": 0.011208392679691315,
"logits/rejected": 0.10715402662754059,
"logps/chosen": -448.327880859375,
"logps/rejected": -458.47607421875,
"loss": 0.332,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.5432627201080322,
"rewards/margins": 1.664372444152832,
"rewards/rejected": -3.207634925842285,
"step": 389
},
{
"epoch": 1.8483412322274881,
"grad_norm": 2.420408010482788,
"learning_rate": 1.7486923471833283e-07,
"logits/chosen": 1.776963472366333e-05,
"logits/rejected": 0.13077270984649658,
"logps/chosen": -420.689697265625,
"logps/rejected": -484.0677490234375,
"loss": 0.2435,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.5387448072433472,
"rewards/margins": 2.5836400985717773,
"rewards/rejected": -4.122385025024414,
"step": 390
},
{
"epoch": 1.853080568720379,
"grad_norm": 4.320888042449951,
"learning_rate": 1.6416995580537665e-07,
"logits/chosen": -0.17437168955802917,
"logits/rejected": 0.09179955720901489,
"logps/chosen": -436.16131591796875,
"logps/rejected": -444.27105712890625,
"loss": 0.4105,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.3086187839508057,
"rewards/margins": 1.6180086135864258,
"rewards/rejected": -2.9266276359558105,
"step": 391
},
{
"epoch": 1.8578199052132702,
"grad_norm": 4.555395603179932,
"learning_rate": 1.5380294592567514e-07,
"logits/chosen": 0.3249743580818176,
"logits/rejected": -0.06773354858160019,
"logps/chosen": -475.42095947265625,
"logps/rejected": -426.3134765625,
"loss": 0.3782,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.2762033939361572,
"rewards/margins": 1.5152695178985596,
"rewards/rejected": -3.791472911834717,
"step": 392
},
{
"epoch": 1.8625592417061612,
"grad_norm": 4.829329967498779,
"learning_rate": 1.4376891739465826e-07,
"logits/chosen": 0.21649743616580963,
"logits/rejected": 0.0726737305521965,
"logps/chosen": -464.2098388671875,
"logps/rejected": -444.3758544921875,
"loss": 0.5695,
"rewards/accuracies": 0.46875,
"rewards/chosen": -2.082812786102295,
"rewards/margins": 1.1739708185195923,
"rewards/rejected": -3.2567837238311768,
"step": 393
},
{
"epoch": 1.867298578199052,
"grad_norm": 4.030221939086914,
"learning_rate": 1.3406855964866405e-07,
"logits/chosen": 0.029259920120239258,
"logits/rejected": 0.09799515455961227,
"logps/chosen": -440.05010986328125,
"logps/rejected": -463.4384765625,
"loss": 0.3544,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.6618742942810059,
"rewards/margins": 1.9635189771652222,
"rewards/rejected": -3.6253933906555176,
"step": 394
},
{
"epoch": 1.8720379146919433,
"grad_norm": 2.8961992263793945,
"learning_rate": 1.247025391975698e-07,
"logits/chosen": -0.30902552604675293,
"logits/rejected": -0.13309930264949799,
"logps/chosen": -462.56378173828125,
"logps/rejected": -469.8658447265625,
"loss": 0.3764,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.1713790893554688,
"rewards/margins": 1.4321413040161133,
"rewards/rejected": -3.603520631790161,
"step": 395
},
{
"epoch": 1.876777251184834,
"grad_norm": 3.45127010345459,
"learning_rate": 1.1567149957899592e-07,
"logits/chosen": 0.017819374799728394,
"logits/rejected": -0.011249087750911713,
"logps/chosen": -431.1867370605469,
"logps/rejected": -441.91229248046875,
"loss": 0.324,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.7995635271072388,
"rewards/margins": 1.8794021606445312,
"rewards/rejected": -3.6789655685424805,
"step": 396
},
{
"epoch": 1.8815165876777251,
"grad_norm": 3.5824666023254395,
"learning_rate": 1.0697606131408966e-07,
"logits/chosen": 0.0838717445731163,
"logits/rejected": 0.017135512083768845,
"logps/chosen": -521.8138427734375,
"logps/rejected": -532.1849365234375,
"loss": 0.3537,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.107038974761963,
"rewards/margins": 1.835722804069519,
"rewards/rejected": -3.9427616596221924,
"step": 397
},
{
"epoch": 1.8862559241706163,
"grad_norm": 3.6466410160064697,
"learning_rate": 9.861682186488697e-08,
"logits/chosen": 0.21266713738441467,
"logits/rejected": 0.2840477526187897,
"logps/chosen": -412.107666015625,
"logps/rejected": -420.9535827636719,
"loss": 0.3189,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.7778931856155396,
"rewards/margins": 1.6844923496246338,
"rewards/rejected": -3.462385416030884,
"step": 398
},
{
"epoch": 1.890995260663507,
"grad_norm": 4.38985013961792,
"learning_rate": 9.059435559326258e-08,
"logits/chosen": 0.14922155439853668,
"logits/rejected": 0.4349655210971832,
"logps/chosen": -356.972412109375,
"logps/rejected": -440.1178283691406,
"loss": 0.4094,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.0182883739471436,
"rewards/margins": 1.7074682712554932,
"rewards/rejected": -3.7257564067840576,
"step": 399
},
{
"epoch": 1.8957345971563981,
"grad_norm": 3.211653232574463,
"learning_rate": 8.290921372146654e-08,
"logits/chosen": -0.14845699071884155,
"logits/rejected": -0.0703466609120369,
"logps/chosen": -477.7309875488281,
"logps/rejected": -578.8424682617188,
"loss": 0.3564,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.0129945278167725,
"rewards/margins": 1.810671091079712,
"rewards/rejected": -3.8236656188964844,
"step": 400
},
{
"epoch": 1.900473933649289,
"grad_norm": 3.0175628662109375,
"learning_rate": 7.55619242942457e-08,
"logits/chosen": 0.2208501398563385,
"logits/rejected": 0.3406308889389038,
"logps/chosen": -396.18011474609375,
"logps/rejected": -454.838134765625,
"loss": 0.3354,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.469225525856018,
"rewards/margins": 1.550800085067749,
"rewards/rejected": -3.0200257301330566,
"step": 401
},
{
"epoch": 1.90521327014218,
"grad_norm": 3.1398770809173584,
"learning_rate": 6.855299214256817e-08,
"logits/chosen": -0.11644039303064346,
"logits/rejected": -0.09911998361349106,
"logps/chosen": -411.9039001464844,
"logps/rejected": -435.63165283203125,
"loss": 0.3259,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.821744680404663,
"rewards/margins": 1.5382413864135742,
"rewards/rejected": -3.3599863052368164,
"step": 402
},
{
"epoch": 1.9099526066350712,
"grad_norm": 3.548295497894287,
"learning_rate": 6.188289884893062e-08,
"logits/chosen": 0.13989268243312836,
"logits/rejected": 0.13753288984298706,
"logps/chosen": -420.6592712402344,
"logps/rejected": -420.6832275390625,
"loss": 0.3468,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.8524384498596191,
"rewards/margins": 2.089343309402466,
"rewards/rejected": -3.941781997680664,
"step": 403
},
{
"epoch": 1.914691943127962,
"grad_norm": 2.794528007507324,
"learning_rate": 5.5552102714271914e-08,
"logits/chosen": 0.07798265665769577,
"logits/rejected": 0.2044173628091812,
"logps/chosen": -465.3638916015625,
"logps/rejected": -485.63140869140625,
"loss": 0.3254,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.057124137878418,
"rewards/margins": 1.8226839303970337,
"rewards/rejected": -3.879808187484741,
"step": 404
},
{
"epoch": 1.919431279620853,
"grad_norm": 4.69295072555542,
"learning_rate": 4.9561038726483326e-08,
"logits/chosen": -0.28620800375938416,
"logits/rejected": -0.35853612422943115,
"logps/chosen": -459.1697692871094,
"logps/rejected": -483.31903076171875,
"loss": 0.3833,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.9887439012527466,
"rewards/margins": 1.8978276252746582,
"rewards/rejected": -3.8865716457366943,
"step": 405
},
{
"epoch": 1.9241706161137442,
"grad_norm": 4.612025260925293,
"learning_rate": 4.391011853051963e-08,
"logits/chosen": 0.013378657400608063,
"logits/rejected": 0.2598281502723694,
"logps/chosen": -445.204345703125,
"logps/rejected": -443.0850830078125,
"loss": 0.414,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.9544767141342163,
"rewards/margins": 1.2461696863174438,
"rewards/rejected": -3.20064640045166,
"step": 406
},
{
"epoch": 1.9289099526066351,
"grad_norm": 3.3378067016601562,
"learning_rate": 3.859973040011511e-08,
"logits/chosen": 0.12009446322917938,
"logits/rejected": -0.013997566886246204,
"logps/chosen": -415.43780517578125,
"logps/rejected": -425.35455322265625,
"loss": 0.283,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.4632946252822876,
"rewards/margins": 2.7979512214660645,
"rewards/rejected": -4.2612457275390625,
"step": 407
},
{
"epoch": 1.933649289099526,
"grad_norm": 2.524122953414917,
"learning_rate": 3.363023921110542e-08,
"logits/chosen": -0.16053307056427002,
"logits/rejected": -0.2901630699634552,
"logps/chosen": -484.54705810546875,
"logps/rejected": -468.14947509765625,
"loss": 0.3086,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.6832842826843262,
"rewards/margins": 1.9632474184036255,
"rewards/rejected": -3.646531581878662,
"step": 408
},
{
"epoch": 1.9383886255924172,
"grad_norm": 5.851565361022949,
"learning_rate": 2.9001986416358763e-08,
"logits/chosen": -0.3602317273616791,
"logits/rejected": -0.4473731815814972,
"logps/chosen": -341.96038818359375,
"logps/rejected": -375.1501159667969,
"loss": 0.5682,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.745245337486267,
"rewards/margins": 1.315721035003662,
"rewards/rejected": -3.0609664916992188,
"step": 409
},
{
"epoch": 1.943127962085308,
"grad_norm": 4.054327011108398,
"learning_rate": 2.4715290022310212e-08,
"logits/chosen": 0.09333401173353195,
"logits/rejected": 0.12337815761566162,
"logps/chosen": -407.29718017578125,
"logps/rejected": -444.52783203125,
"loss": 0.3384,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.8622604608535767,
"rewards/margins": 1.8048306703567505,
"rewards/rejected": -3.667090892791748,
"step": 410
},
{
"epoch": 1.947867298578199,
"grad_norm": 5.784478187561035,
"learning_rate": 2.0770444567118075e-08,
"logits/chosen": 0.12872330844402313,
"logits/rejected": -0.020948439836502075,
"logps/chosen": -689.5375366210938,
"logps/rejected": -653.1371459960938,
"loss": 0.5016,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.1100049018859863,
"rewards/margins": 1.1901695728302002,
"rewards/rejected": -3.3001744747161865,
"step": 411
},
{
"epoch": 1.95260663507109,
"grad_norm": 3.3096017837524414,
"learning_rate": 1.7167721100420086e-08,
"logits/chosen": -0.1594451665878296,
"logits/rejected": -0.20826411247253418,
"logps/chosen": -327.7048034667969,
"logps/rejected": -397.76702880859375,
"loss": 0.2559,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.474926471710205,
"rewards/margins": 2.178825855255127,
"rewards/rejected": -3.653752326965332,
"step": 412
},
{
"epoch": 1.957345971563981,
"grad_norm": 3.7895545959472656,
"learning_rate": 1.3907367164713304e-08,
"logits/chosen": -0.2774580717086792,
"logits/rejected": -0.025296274572610855,
"logps/chosen": -454.33074951171875,
"logps/rejected": -562.2850341796875,
"loss": 0.3308,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.998180866241455,
"rewards/margins": 1.9229161739349365,
"rewards/rejected": -3.9210968017578125,
"step": 413
},
{
"epoch": 1.962085308056872,
"grad_norm": 5.27892541885376,
"learning_rate": 1.0989606778344375e-08,
"logits/chosen": 0.028818532824516296,
"logits/rejected": -0.06880553811788559,
"logps/chosen": -479.7122802734375,
"logps/rejected": -466.845947265625,
"loss": 0.3081,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.92867112159729,
"rewards/margins": 2.0744426250457764,
"rewards/rejected": -4.003113746643066,
"step": 414
},
{
"epoch": 1.966824644549763,
"grad_norm": 4.537078857421875,
"learning_rate": 8.414640420116305e-09,
"logits/chosen": 0.08707749843597412,
"logits/rejected": -0.05305865406990051,
"logps/chosen": -447.78302001953125,
"logps/rejected": -430.87030029296875,
"loss": 0.367,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.6580579280853271,
"rewards/margins": 1.6174925565719604,
"rewards/rejected": -3.275550365447998,
"step": 415
},
{
"epoch": 1.971563981042654,
"grad_norm": 2.382775068283081,
"learning_rate": 6.182645015516131e-09,
"logits/chosen": 0.23202013969421387,
"logits/rejected": 0.24202974140644073,
"logps/chosen": -426.9042663574219,
"logps/rejected": -456.701904296875,
"loss": 0.2973,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.9331663846969604,
"rewards/margins": 1.9195029735565186,
"rewards/rejected": -3.8526692390441895,
"step": 416
},
{
"epoch": 1.9763033175355451,
"grad_norm": 4.986690521240234,
"learning_rate": 4.2937739245563216e-09,
"logits/chosen": 0.08618678152561188,
"logits/rejected": 0.05446825921535492,
"logps/chosen": -410.23797607421875,
"logps/rejected": -479.15032958984375,
"loss": 0.4803,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.676472783088684,
"rewards/margins": 1.6006869077682495,
"rewards/rejected": -3.2771594524383545,
"step": 417
},
{
"epoch": 1.981042654028436,
"grad_norm": 4.937625408172607,
"learning_rate": 2.7481569312381995e-09,
"logits/chosen": 0.2503683865070343,
"logits/rejected": 0.25900202989578247,
"logps/chosen": -392.8948974609375,
"logps/rejected": -445.6235046386719,
"loss": 0.2971,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.1311068534851074,
"rewards/margins": 2.716494083404541,
"rewards/rejected": -4.847600936889648,
"step": 418
},
{
"epoch": 1.985781990521327,
"grad_norm": 3.498244524002075,
"learning_rate": 1.5459002346324137e-09,
"logits/chosen": -0.018450409173965454,
"logits/rejected": -0.062519371509552,
"logps/chosen": -406.3209228515625,
"logps/rejected": -451.2742919921875,
"loss": 0.3417,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.9635508060455322,
"rewards/margins": 1.8704030513763428,
"rewards/rejected": -3.833953857421875,
"step": 419
},
{
"epoch": 1.9905213270142181,
"grad_norm": 2.944807767868042,
"learning_rate": 6.870864415875478e-10,
"logits/chosen": -0.22453802824020386,
"logits/rejected": 0.1273200809955597,
"logps/chosen": -398.6549987792969,
"logps/rejected": -451.1443176269531,
"loss": 0.3175,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.9729655981063843,
"rewards/margins": 1.8536114692687988,
"rewards/rejected": -3.8265771865844727,
"step": 420
},
{
"epoch": 1.9952606635071088,
"grad_norm": 2.9744060039520264,
"learning_rate": 1.7177456104688905e-10,
"logits/chosen": -0.29849839210510254,
"logits/rejected": -0.11612643301486969,
"logps/chosen": -422.93353271484375,
"logps/rejected": -584.3493041992188,
"loss": 0.2149,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.9120306968688965,
"rewards/margins": 2.681218385696411,
"rewards/rejected": -4.593249320983887,
"step": 421
},
{
"epoch": 2.0,
"grad_norm": 5.863999843597412,
"learning_rate": 0.0,
"logits/chosen": 0.13755494356155396,
"logits/rejected": 0.18007084727287292,
"logps/chosen": -402.926513671875,
"logps/rejected": -388.6621398925781,
"loss": 0.5596,
"rewards/accuracies": 0.6458333134651184,
"rewards/chosen": -2.5218539237976074,
"rewards/margins": 0.7689557671546936,
"rewards/rejected": -3.2908096313476562,
"step": 422
}
],
"logging_steps": 1,
"max_steps": 422,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}