{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9989071038251366,
  "eval_steps": 400,
  "global_step": 457,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01092896174863388,
      "grad_norm": 56.07717009412548,
      "learning_rate": 9.782608695652174e-08,
      "logits/chosen": -1.0135059356689453,
      "logits/rejected": -1.0081793069839478,
      "logps/chosen": -0.28040584921836853,
      "logps/rejected": -0.28580397367477417,
      "loss": 4.1045,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -2.804058313369751,
      "rewards/margins": 0.05398142337799072,
      "rewards/rejected": -2.8580400943756104,
      "semantic_entropy": 0.7516118884086609,
      "step": 5
    },
    {
      "epoch": 0.02185792349726776,
      "grad_norm": 72.03538939992451,
      "learning_rate": 1.9565217391304347e-07,
      "logits/chosen": -1.060633659362793,
      "logits/rejected": -1.009230375289917,
      "logps/chosen": -0.25705546140670776,
      "logps/rejected": -0.270809531211853,
      "loss": 4.0645,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -2.570554494857788,
      "rewards/margins": 0.13754098117351532,
      "rewards/rejected": -2.7080955505371094,
      "semantic_entropy": 0.709734320640564,
      "step": 10
    },
    {
      "epoch": 0.03278688524590164,
      "grad_norm": 60.80949908739426,
      "learning_rate": 2.9347826086956523e-07,
      "logits/chosen": -1.0096722841262817,
      "logits/rejected": -0.9631622433662415,
      "logps/chosen": -0.2670869827270508,
      "logps/rejected": -0.2734836935997009,
      "loss": 4.0535,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -2.670869827270508,
      "rewards/margins": 0.06396698206663132,
      "rewards/rejected": -2.734837055206299,
      "semantic_entropy": 0.7273667454719543,
      "step": 15
    },
    {
      "epoch": 0.04371584699453552,
      "grad_norm": 72.87577064643818,
      "learning_rate": 3.9130434782608694e-07,
      "logits/chosen": -0.9465762376785278,
      "logits/rejected": -0.8959148526191711,
      "logps/chosen": -0.27229827642440796,
      "logps/rejected": -0.28428345918655396,
      "loss": 4.0997,
      "rewards/accuracies": 0.5687500238418579,
      "rewards/chosen": -2.7229831218719482,
      "rewards/margins": 0.1198514848947525,
      "rewards/rejected": -2.842834711074829,
      "semantic_entropy": 0.743862509727478,
      "step": 20
    },
    {
      "epoch": 0.0546448087431694,
      "grad_norm": 35.73788589652073,
      "learning_rate": 4.891304347826087e-07,
      "logits/chosen": -0.9495905637741089,
      "logits/rejected": -0.8757770657539368,
      "logps/chosen": -0.2737862467765808,
      "logps/rejected": -0.2946169972419739,
      "loss": 4.0619,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": -2.7378625869750977,
      "rewards/margins": 0.2083074301481247,
      "rewards/rejected": -2.9461700916290283,
      "semantic_entropy": 0.7538469433784485,
      "step": 25
    },
    {
      "epoch": 0.06557377049180328,
      "grad_norm": 56.199447993934285,
      "learning_rate": 5.869565217391305e-07,
      "logits/chosen": -1.0543675422668457,
      "logits/rejected": -0.9891208410263062,
      "logps/chosen": -0.2664986550807953,
      "logps/rejected": -0.28541743755340576,
      "loss": 4.0445,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -2.6649868488311768,
      "rewards/margins": 0.18918730318546295,
      "rewards/rejected": -2.8541738986968994,
      "semantic_entropy": 0.7244771122932434,
      "step": 30
    },
    {
      "epoch": 0.07650273224043716,
      "grad_norm": 64.82660227272935,
      "learning_rate": 6.847826086956522e-07,
      "logits/chosen": -0.9974210858345032,
      "logits/rejected": -0.9332624673843384,
      "logps/chosen": -0.2572976052761078,
      "logps/rejected": -0.2904777228832245,
      "loss": 3.9891,
      "rewards/accuracies": 0.518750011920929,
      "rewards/chosen": -2.5729756355285645,
      "rewards/margins": 0.33180171251296997,
      "rewards/rejected": -2.9047772884368896,
      "semantic_entropy": 0.7203728556632996,
      "step": 35
    },
    {
      "epoch": 0.08743169398907104,
      "grad_norm": 74.43759886551268,
      "learning_rate": 7.826086956521739e-07,
      "logits/chosen": -0.9570473432540894,
      "logits/rejected": -0.8978286981582642,
      "logps/chosen": -0.28474289178848267,
      "logps/rejected": -0.31194889545440674,
      "loss": 4.1025,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -2.847428798675537,
      "rewards/margins": 0.27205973863601685,
      "rewards/rejected": -3.1194887161254883,
      "semantic_entropy": 0.7648034691810608,
      "step": 40
    },
    {
      "epoch": 0.09836065573770492,
      "grad_norm": 34.176663547437016,
      "learning_rate": 8.804347826086956e-07,
      "logits/chosen": -1.021121859550476,
      "logits/rejected": -0.9401710629463196,
      "logps/chosen": -0.2921110689640045,
      "logps/rejected": -0.33298516273498535,
      "loss": 4.0055,
      "rewards/accuracies": 0.518750011920929,
      "rewards/chosen": -2.9211106300354004,
      "rewards/margins": 0.4087409973144531,
      "rewards/rejected": -3.3298516273498535,
      "semantic_entropy": 0.7668289542198181,
      "step": 45
    },
    {
      "epoch": 0.1092896174863388,
      "grad_norm": 80.21226502338544,
      "learning_rate": 8.997896779678812e-07,
      "logits/chosen": -0.955331027507782,
      "logits/rejected": -0.8773272633552551,
      "logps/chosen": -0.3032563030719757,
      "logps/rejected": -0.3285720944404602,
      "loss": 4.0066,
      "rewards/accuracies": 0.4937500059604645,
      "rewards/chosen": -3.0325629711151123,
      "rewards/margins": 0.2531580626964569,
      "rewards/rejected": -3.2857213020324707,
      "semantic_entropy": 0.765364944934845,
      "step": 50
    },
    {
      "epoch": 0.12021857923497267,
      "grad_norm": 43.70683163065837,
      "learning_rate": 8.989355816502526e-07,
      "logits/chosen": -0.9907596707344055,
      "logits/rejected": -0.8809078931808472,
      "logps/chosen": -0.2901116907596588,
      "logps/rejected": -0.3407103419303894,
      "loss": 3.896,
      "rewards/accuracies": 0.581250011920929,
      "rewards/chosen": -2.9011168479919434,
      "rewards/margins": 0.5059866905212402,
      "rewards/rejected": -3.4071037769317627,
      "semantic_entropy": 0.7688600420951843,
      "step": 55
    },
    {
      "epoch": 0.13114754098360656,
      "grad_norm": 40.17800097293932,
      "learning_rate": 8.974258123725301e-07,
      "logits/chosen": -1.0390875339508057,
      "logits/rejected": -0.995086669921875,
      "logps/chosen": -0.2762259840965271,
      "logps/rejected": -0.34549298882484436,
      "loss": 3.8357,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -2.7622599601745605,
      "rewards/margins": 0.6926698684692383,
      "rewards/rejected": -3.454929828643799,
      "semantic_entropy": 0.7664283514022827,
      "step": 60
    },
    {
      "epoch": 0.14207650273224043,
      "grad_norm": 59.513778490950706,
      "learning_rate": 8.952625751612967e-07,
      "logits/chosen": -1.0262492895126343,
      "logits/rejected": -0.9614121317863464,
      "logps/chosen": -0.324962317943573,
      "logps/rejected": -0.35770028829574585,
      "loss": 3.9978,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -3.2496230602264404,
      "rewards/margins": 0.3273797631263733,
      "rewards/rejected": -3.577003002166748,
      "semantic_entropy": 0.8130875825881958,
      "step": 65
    },
    {
      "epoch": 0.15300546448087432,
      "grad_norm": 207.7015350833011,
      "learning_rate": 8.924490294367533e-07,
      "logits/chosen": -0.9874998331069946,
      "logits/rejected": -0.9699292182922363,
      "logps/chosen": -0.31067711114883423,
      "logps/rejected": -0.364091157913208,
      "loss": 3.8828,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": -3.1067709922790527,
      "rewards/margins": 0.534140408039093,
      "rewards/rejected": -3.64091157913208,
      "semantic_entropy": 0.7709770798683167,
      "step": 70
    },
    {
      "epoch": 0.16393442622950818,
      "grad_norm": 44.352990988039096,
      "learning_rate": 8.889892843983686e-07,
      "logits/chosen": -0.9768747091293335,
      "logits/rejected": -0.9624359011650085,
      "logps/chosen": -0.3188754916191101,
      "logps/rejected": -0.39044779539108276,
      "loss": 3.9066,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -3.1887545585632324,
      "rewards/margins": 0.715722918510437,
      "rewards/rejected": -3.904478073120117,
      "semantic_entropy": 0.8010842204093933,
      "step": 75
    },
    {
      "epoch": 0.17486338797814208,
      "grad_norm": 96.76821204915929,
      "learning_rate": 8.848883930233695e-07,
      "logits/chosen": -0.9784610867500305,
      "logits/rejected": -0.9192510843276978,
      "logps/chosen": -0.332768976688385,
      "logps/rejected": -0.36657267808914185,
      "loss": 3.8709,
      "rewards/accuracies": 0.606249988079071,
      "rewards/chosen": -3.3276901245117188,
      "rewards/margins": 0.3380367159843445,
      "rewards/rejected": -3.665726900100708,
      "semantic_entropy": 0.7843155860900879,
      "step": 80
    },
    {
      "epoch": 0.18579234972677597,
      "grad_norm": 75.68920123621399,
      "learning_rate": 8.801523446868397e-07,
      "logits/chosen": -0.9815126657485962,
      "logits/rejected": -0.9339221119880676,
      "logps/chosen": -0.3239663243293762,
      "logps/rejected": -0.4012156128883362,
      "loss": 3.832,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -3.2396633625030518,
      "rewards/margins": 0.7724922895431519,
      "rewards/rejected": -4.012156009674072,
      "semantic_entropy": 0.8030881881713867,
      "step": 85
    },
    {
      "epoch": 0.19672131147540983,
      "grad_norm": 52.172085418459005,
      "learning_rate": 8.74788056414203e-07,
      "logits/chosen": -1.024124264717102,
      "logits/rejected": -0.9470185041427612,
      "logps/chosen": -0.3651575446128845,
      "logps/rejected": -0.40893810987472534,
      "loss": 3.8895,
      "rewards/accuracies": 0.5687500238418579,
      "rewards/chosen": -3.651575803756714,
      "rewards/margins": 0.43780526518821716,
      "rewards/rejected": -4.089381217956543,
      "semantic_entropy": 0.8523303270339966,
      "step": 90
    },
    {
      "epoch": 0.20765027322404372,
      "grad_norm": 55.856723424209804,
      "learning_rate": 8.688033627788689e-07,
      "logits/chosen": -0.9588603973388672,
      "logits/rejected": -0.9630708694458008,
      "logps/chosen": -0.35158681869506836,
      "logps/rejected": -0.3956569731235504,
      "loss": 3.7738,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": -3.5158684253692627,
      "rewards/margins": 0.4407011568546295,
      "rewards/rejected": -3.956569194793701,
      "semantic_entropy": 0.8362972140312195,
      "step": 95
    },
    {
      "epoch": 0.2185792349726776,
      "grad_norm": 133.22842224119972,
      "learning_rate": 8.622070044597933e-07,
      "logits/chosen": -0.9748640060424805,
      "logits/rejected": -0.929074764251709,
      "logps/chosen": -0.4111453890800476,
      "logps/rejected": -0.49381446838378906,
      "loss": 3.8025,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -4.111453533172607,
      "rewards/margins": 0.8266904950141907,
      "rewards/rejected": -4.938143730163574,
      "semantic_entropy": 0.8799806833267212,
      "step": 100
    },
    {
      "epoch": 0.22950819672131148,
      "grad_norm": 63.064394179489504,
      "learning_rate": 8.550086154756677e-07,
      "logits/chosen": -1.0247594118118286,
      "logits/rejected": -0.9960424304008484,
      "logps/chosen": -0.3660830855369568,
      "logps/rejected": -0.44689756631851196,
      "loss": 3.7174,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -3.6608314514160156,
      "rewards/margins": 0.8081439733505249,
      "rewards/rejected": -4.468975067138672,
      "semantic_entropy": 0.8604310154914856,
      "step": 105
    },
    {
      "epoch": 0.24043715846994534,
      "grad_norm": 63.6155571920592,
      "learning_rate": 8.472187091143803e-07,
      "logits/chosen": -1.0404878854751587,
      "logits/rejected": -1.019997000694275,
      "logps/chosen": -0.40719980001449585,
      "logps/rejected": -0.5280216336250305,
      "loss": 3.7079,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -4.07199764251709,
      "rewards/margins": 1.2082184553146362,
      "rewards/rejected": -5.280216217041016,
      "semantic_entropy": 0.903515636920929,
      "step": 110
    },
    {
      "epoch": 0.25136612021857924,
      "grad_norm": 68.78614353875135,
      "learning_rate": 8.388486625782994e-07,
      "logits/chosen": -1.0444709062576294,
      "logits/rejected": -1.0056098699569702,
      "logps/chosen": -0.40871691703796387,
      "logps/rejected": -0.5650209188461304,
      "loss": 3.6489,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -4.087169170379639,
      "rewards/margins": 1.563040018081665,
      "rewards/rejected": -5.650209426879883,
      "semantic_entropy": 0.8942676782608032,
      "step": 115
    },
    {
      "epoch": 0.26229508196721313,
      "grad_norm": 53.32792414437455,
      "learning_rate": 8.299107003678049e-07,
      "logits/chosen": -1.0611729621887207,
      "logits/rejected": -0.9919866323471069,
      "logps/chosen": -0.42750421166419983,
      "logps/rejected": -0.5232139825820923,
      "loss": 3.5791,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -4.2750420570373535,
      "rewards/margins": 0.9570980072021484,
      "rewards/rejected": -5.232140064239502,
      "semantic_entropy": 0.9266929626464844,
      "step": 120
    },
    {
      "epoch": 0.273224043715847,
      "grad_norm": 121.11035694148153,
      "learning_rate": 8.204178764273363e-07,
      "logits/chosen": -0.9674042463302612,
      "logits/rejected": -0.9337356686592102,
      "logps/chosen": -0.4846254885196686,
      "logps/rejected": -0.6142443418502808,
      "loss": 3.5767,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -4.846255302429199,
      "rewards/margins": 1.2961883544921875,
      "rewards/rejected": -6.1424431800842285,
      "semantic_entropy": 0.92425936460495,
      "step": 125
    },
    {
      "epoch": 0.28415300546448086,
      "grad_norm": 64.53667052645014,
      "learning_rate": 8.103840550800328e-07,
      "logits/chosen": -0.9902731776237488,
      "logits/rejected": -0.987317681312561,
      "logps/chosen": -0.5054492950439453,
      "logps/rejected": -0.7366899251937866,
      "loss": 3.5275,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -5.054492473602295,
      "rewards/margins": 2.3124070167541504,
      "rewards/rejected": -7.366899013519287,
      "semantic_entropy": 0.9140235185623169,
      "step": 130
    },
    {
      "epoch": 0.29508196721311475,
      "grad_norm": 59.7508648355179,
      "learning_rate": 7.998238907788109e-07,
      "logits/chosen": -1.0142452716827393,
      "logits/rejected": -0.9352253079414368,
      "logps/chosen": -0.5328418612480164,
      "logps/rejected": -0.72784823179245,
      "loss": 3.3724,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": -5.328418254852295,
      "rewards/margins": 1.9500631093978882,
      "rewards/rejected": -7.278481960296631,
      "semantic_entropy": 0.9628833532333374,
      "step": 135
    },
    {
      "epoch": 0.30601092896174864,
      "grad_norm": 78.9540175820449,
      "learning_rate": 7.887528067034528e-07,
      "logits/chosen": -1.017315149307251,
      "logits/rejected": -0.9636947512626648,
      "logps/chosen": -0.6016183495521545,
      "logps/rejected": -0.8618828058242798,
      "loss": 3.0483,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -6.016183853149414,
      "rewards/margins": 2.602644681930542,
      "rewards/rejected": -8.618828773498535,
      "semantic_entropy": 0.9850225448608398,
      "step": 140
    },
    {
      "epoch": 0.31693989071038253,
      "grad_norm": 99.62302257315461,
      "learning_rate": 7.77186972234965e-07,
      "logits/chosen": -0.9608656764030457,
      "logits/rejected": -0.9462074041366577,
      "logps/chosen": -0.6700846552848816,
      "logps/rejected": -0.8206523060798645,
      "loss": 3.2001,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -6.7008466720581055,
      "rewards/margins": 1.505676031112671,
      "rewards/rejected": -8.206521987915039,
      "semantic_entropy": 0.9759008288383484,
      "step": 145
    },
    {
      "epoch": 0.32786885245901637,
      "grad_norm": 89.43625982319897,
      "learning_rate": 7.651432793401065e-07,
      "logits/chosen": -1.010655164718628,
      "logits/rejected": -0.9779534339904785,
      "logps/chosen": -0.8374897241592407,
      "logps/rejected": -1.1644561290740967,
      "loss": 3.0851,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -8.374897956848145,
      "rewards/margins": 3.269664764404297,
      "rewards/rejected": -11.644561767578125,
      "semantic_entropy": 0.9569616317749023,
      "step": 150
    },
    {
      "epoch": 0.33879781420765026,
      "grad_norm": 89.48638184015236,
      "learning_rate": 7.526393179005737e-07,
      "logits/chosen": -0.9409087896347046,
      "logits/rejected": -0.8860008120536804,
      "logps/chosen": -0.9283156394958496,
      "logps/rejected": -1.1858878135681152,
      "loss": 2.8999,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -9.283157348632812,
      "rewards/margins": 2.575721263885498,
      "rewards/rejected": -11.858877182006836,
      "semantic_entropy": 0.8997370004653931,
      "step": 155
    },
    {
      "epoch": 0.34972677595628415,
      "grad_norm": 106.63607424638896,
      "learning_rate": 7.396933500228808e-07,
      "logits/chosen": -0.9566726684570312,
      "logits/rejected": -0.9040724039077759,
      "logps/chosen": -1.1487252712249756,
      "logps/rejected": -1.4483044147491455,
      "loss": 2.801,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -11.487253189086914,
      "rewards/margins": 2.9957919120788574,
      "rewards/rejected": -14.48304557800293,
      "semantic_entropy": 0.8422845005989075,
      "step": 160
    },
    {
      "epoch": 0.36065573770491804,
      "grad_norm": 173.29746360828278,
      "learning_rate": 7.263242833664484e-07,
      "logits/chosen": -0.911471962928772,
      "logits/rejected": -0.8458994030952454,
      "logps/chosen": -1.2985560894012451,
      "logps/rejected": -1.5788644552230835,
      "loss": 2.8804,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -12.985560417175293,
      "rewards/margins": 2.8030834197998047,
      "rewards/rejected": -15.788644790649414,
      "semantic_entropy": 0.802132785320282,
      "step": 165
    },
    {
      "epoch": 0.37158469945355194,
      "grad_norm": 100.11131084617674,
      "learning_rate": 7.1255164352886e-07,
      "logits/chosen": -0.9127036929130554,
      "logits/rejected": -0.8471270799636841,
      "logps/chosen": -1.3799912929534912,
      "logps/rejected": -1.7123008966445923,
      "loss": 2.8315,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -13.79991340637207,
      "rewards/margins": 3.3230953216552734,
      "rewards/rejected": -17.12300682067871,
      "semantic_entropy": 0.7457509636878967,
      "step": 170
    },
    {
      "epoch": 0.3825136612021858,
      "grad_norm": 111.52780471006011,
      "learning_rate": 6.98395545528617e-07,
      "logits/chosen": -0.9586197137832642,
      "logits/rejected": -0.9303746223449707,
      "logps/chosen": -1.5328519344329834,
      "logps/rejected": -1.8097922801971436,
      "loss": 2.5002,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -15.328518867492676,
      "rewards/margins": 2.769404411315918,
      "rewards/rejected": -18.097923278808594,
      "semantic_entropy": 0.6919700503349304,
      "step": 175
    },
    {
      "epoch": 0.39344262295081966,
      "grad_norm": 112.60716502155312,
      "learning_rate": 6.838766644270385e-07,
      "logits/chosen": -0.9389745593070984,
      "logits/rejected": -0.9307812452316284,
      "logps/chosen": -1.5702974796295166,
      "logps/rejected": -2.0943307876586914,
      "loss": 2.4165,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -15.702975273132324,
      "rewards/margins": 5.24033260345459,
      "rewards/rejected": -20.943307876586914,
      "semantic_entropy": 0.6626415252685547,
      "step": 180
    },
    {
      "epoch": 0.40437158469945356,
      "grad_norm": 178.966427933753,
      "learning_rate": 6.690162051322191e-07,
      "logits/chosen": -0.9021973609924316,
      "logits/rejected": -0.9323358535766602,
      "logps/chosen": -1.6979179382324219,
      "logps/rejected": -2.232943058013916,
      "loss": 2.2909,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -16.979177474975586,
      "rewards/margins": 5.350252151489258,
      "rewards/rejected": -22.329431533813477,
      "semantic_entropy": 0.6147797107696533,
      "step": 185
    },
    {
      "epoch": 0.41530054644808745,
      "grad_norm": 146.482478913679,
      "learning_rate": 6.538358714291389e-07,
      "logits/chosen": -0.9412842988967896,
      "logits/rejected": -0.8962854146957397,
      "logps/chosen": -1.7343536615371704,
      "logps/rejected": -2.2405014038085938,
      "loss": 2.3785,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -17.343536376953125,
      "rewards/margins": 5.061476707458496,
      "rewards/rejected": -22.405012130737305,
      "semantic_entropy": 0.6030288934707642,
      "step": 190
    },
    {
      "epoch": 0.4262295081967213,
      "grad_norm": 150.31719448352644,
      "learning_rate": 6.383578342811638e-07,
      "logits/chosen": -0.9521581530570984,
      "logits/rejected": -0.9250186085700989,
      "logps/chosen": -1.9811210632324219,
      "logps/rejected": -2.550245523452759,
      "loss": 2.3547,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -19.81121253967285,
      "rewards/margins": 5.691242694854736,
      "rewards/rejected": -25.502452850341797,
      "semantic_entropy": 0.5285295844078064,
      "step": 195
    },
    {
      "epoch": 0.4371584699453552,
      "grad_norm": 142.7731322709338,
      "learning_rate": 6.226046994492272e-07,
      "logits/chosen": -0.9036038517951965,
      "logits/rejected": -0.7988227605819702,
      "logps/chosen": -1.8621279001235962,
      "logps/rejected": -2.401458263397217,
      "loss": 2.1514,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -18.621280670166016,
      "rewards/margins": 5.393301486968994,
      "rewards/rejected": -24.01457977294922,
      "semantic_entropy": 0.5528938174247742,
      "step": 200
    },
    {
      "epoch": 0.44808743169398907,
      "grad_norm": 109.73924196519458,
      "learning_rate": 6.065994744759878e-07,
      "logits/chosen": -0.8110615015029907,
      "logits/rejected": -0.8245576024055481,
      "logps/chosen": -1.8828060626983643,
      "logps/rejected": -2.4408397674560547,
      "loss": 2.3483,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -18.828060150146484,
      "rewards/margins": 5.58033561706543,
      "rewards/rejected": -24.408397674560547,
      "semantic_entropy": 0.5373791456222534,
      "step": 205
    },
    {
      "epoch": 0.45901639344262296,
      "grad_norm": 168.41087843878026,
      "learning_rate": 5.903655350831845e-07,
      "logits/chosen": -0.8903123736381531,
      "logits/rejected": -0.8840487599372864,
      "logps/chosen": -1.9938600063323975,
      "logps/rejected": -2.660202980041504,
      "loss": 2.0737,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -19.938602447509766,
      "rewards/margins": 6.663429260253906,
      "rewards/rejected": -26.602031707763672,
      "semantic_entropy": 0.5236510634422302,
      "step": 210
    },
    {
      "epoch": 0.46994535519125685,
      "grad_norm": 180.98552250261446,
      "learning_rate": 5.73926591031261e-07,
      "logits/chosen": -0.8701471090316772,
      "logits/rejected": -0.7955431938171387,
      "logps/chosen": -2.0766396522521973,
      "logps/rejected": -2.737276077270508,
      "loss": 2.2802,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -20.76639747619629,
      "rewards/margins": 6.606365203857422,
      "rewards/rejected": -27.37276268005371,
      "semantic_entropy": 0.4769493639469147,
      "step": 215
    },
    {
      "epoch": 0.4808743169398907,
      "grad_norm": 113.13891029235478,
      "learning_rate": 5.573066514911273e-07,
      "logits/chosen": -0.9388381838798523,
      "logits/rejected": -0.916387677192688,
      "logps/chosen": -2.1589951515197754,
      "logps/rejected": -2.7171058654785156,
      "loss": 2.1685,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -21.589946746826172,
      "rewards/margins": 5.581110954284668,
      "rewards/rejected": -27.171056747436523,
      "semantic_entropy": 0.46796584129333496,
      "step": 220
    },
    {
      "epoch": 0.4918032786885246,
      "grad_norm": 104.5369543115771,
      "learning_rate": 5.405299899786287e-07,
      "logits/chosen": -0.9614135026931763,
      "logits/rejected": -0.9081883430480957,
      "logps/chosen": -2.1882309913635254,
      "logps/rejected": -2.7919082641601562,
      "loss": 2.2561,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -21.882308959960938,
      "rewards/margins": 6.036771297454834,
      "rewards/rejected": -27.919086456298828,
      "semantic_entropy": 0.45887380838394165,
      "step": 225
    },
    {
      "epoch": 0.5027322404371585,
      "grad_norm": 122.89716590246043,
      "learning_rate": 5.236211089029395e-07,
      "logits/chosen": -0.8663997650146484,
      "logits/rejected": -0.874904453754425,
      "logps/chosen": -2.1786980628967285,
      "logps/rejected": -2.7758872509002686,
      "loss": 2.3111,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -21.78697967529297,
      "rewards/margins": 5.971897125244141,
      "rewards/rejected": -27.758874893188477,
      "semantic_entropy": 0.45373231172561646,
      "step": 230
    },
    {
      "epoch": 0.5136612021857924,
      "grad_norm": 163.16841178090678,
      "learning_rate": 5.066047037806549e-07,
      "logits/chosen": -0.8985625505447388,
      "logits/rejected": -0.8524802923202515,
      "logps/chosen": -2.2341809272766113,
      "logps/rejected": -2.762202262878418,
      "loss": 1.952,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -22.34181022644043,
      "rewards/margins": 5.280209541320801,
      "rewards/rejected": -27.622018814086914,
      "semantic_entropy": 0.4423500895500183,
      "step": 235
    },
    {
      "epoch": 0.5245901639344263,
      "grad_norm": 126.28807692275711,
      "learning_rate": 4.895056271678488e-07,
      "logits/chosen": -0.9690831303596497,
      "logits/rejected": -0.9802172780036926,
      "logps/chosen": -2.1863203048706055,
      "logps/rejected": -2.8334527015686035,
      "loss": 1.9608,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -21.863203048706055,
      "rewards/margins": 6.4713263511657715,
      "rewards/rejected": -28.33452796936035,
      "semantic_entropy": 0.4405677318572998,
      "step": 240
    },
    {
      "epoch": 0.5355191256830601,
      "grad_norm": 203.4242664926994,
      "learning_rate": 4.723488523627758e-07,
      "logits/chosen": -0.9145323038101196,
      "logits/rejected": -0.8553255796432495,
      "logps/chosen": -2.0377883911132812,
      "logps/rejected": -2.7260096073150635,
      "loss": 2.0939,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -20.377883911132812,
      "rewards/margins": 6.882212162017822,
      "rewards/rejected": -27.260095596313477,
      "semantic_entropy": 0.49542418122291565,
      "step": 245
    },
    {
      "epoch": 0.546448087431694,
      "grad_norm": 99.66072304681889,
      "learning_rate": 4.55159436932227e-07,
      "logits/chosen": -0.9824727773666382,
      "logits/rejected": -0.9378935694694519,
      "logps/chosen": -2.12608003616333,
      "logps/rejected": -2.6889073848724365,
      "loss": 2.0268,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -21.26080322265625,
      "rewards/margins": 5.628273010253906,
      "rewards/rejected": -26.889074325561523,
      "semantic_entropy": 0.46515989303588867,
      "step": 250
    },
    {
      "epoch": 0.5573770491803278,
      "grad_norm": 126.02875800766768,
      "learning_rate": 4.3796248611481193e-07,
      "logits/chosen": -0.982226550579071,
      "logits/rejected": -0.947576642036438,
      "logps/chosen": -2.2678322792053223,
      "logps/rejected": -2.873493194580078,
      "loss": 2.1411,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -22.67832374572754,
      "rewards/margins": 6.056609153747559,
      "rewards/rejected": -28.734933853149414,
      "semantic_entropy": 0.4328394830226898,
      "step": 255
    },
    {
      "epoch": 0.5683060109289617,
      "grad_norm": 97.8896168726136,
      "learning_rate": 4.2078311615461545e-07,
      "logits/chosen": -0.961346447467804,
      "logits/rejected": -0.9446791410446167,
      "logps/chosen": -2.171950578689575,
      "logps/rejected": -2.801100492477417,
      "loss": 1.9527,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -21.719501495361328,
      "rewards/margins": 6.291501998901367,
      "rewards/rejected": -28.011005401611328,
      "semantic_entropy": 0.45725923776626587,
      "step": 260
    },
    {
      "epoch": 0.5792349726775956,
      "grad_norm": 92.04238840519959,
      "learning_rate": 4.0364641761878056e-07,
      "logits/chosen": -0.940658688545227,
      "logits/rejected": -0.8965815305709839,
      "logps/chosen": -2.2573418617248535,
      "logps/rejected": -2.931023597717285,
      "loss": 1.9009,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -22.573415756225586,
      "rewards/margins": 6.736820220947266,
      "rewards/rejected": -29.31023597717285,
      "semantic_entropy": 0.4323546886444092,
      "step": 265
    },
    {
      "epoch": 0.5901639344262295,
      "grad_norm": 132.08002962791124,
      "learning_rate": 3.8657741875259324e-07,
      "logits/chosen": -0.9581838846206665,
      "logits/rejected": -0.9301847219467163,
      "logps/chosen": -2.3147079944610596,
      "logps/rejected": -3.0581116676330566,
      "loss": 1.7921,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -23.147079467773438,
      "rewards/margins": 7.4340338706970215,
      "rewards/rejected": -30.581111907958984,
      "semantic_entropy": 0.41289639472961426,
      "step": 270
    },
    {
      "epoch": 0.6010928961748634,
      "grad_norm": 114.39315473375268,
      "learning_rate": 3.6960104892558897e-07,
      "logits/chosen": -0.9345510601997375,
      "logits/rejected": -0.862471878528595,
      "logps/chosen": -2.36122989654541,
      "logps/rejected": -2.888824701309204,
      "loss": 2.0653,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -23.6122989654541,
      "rewards/margins": 5.275951385498047,
      "rewards/rejected": -28.88825035095215,
      "semantic_entropy": 0.40508565306663513,
      "step": 275
    },
    {
      "epoch": 0.6120218579234973,
      "grad_norm": 124.01203860773359,
      "learning_rate": 3.5274210222206623e-07,
      "logits/chosen": -0.9259498715400696,
      "logits/rejected": -0.9285569190979004,
      "logps/chosen": -2.254788637161255,
      "logps/rejected": -2.976097583770752,
      "loss": 2.0752,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -22.54788589477539,
      "rewards/margins": 7.213089942932129,
      "rewards/rejected": -29.760974884033203,
      "semantic_entropy": 0.44933634996414185,
      "step": 280
    },
    {
      "epoch": 0.6229508196721312,
      "grad_norm": 117.08336782700206,
      "learning_rate": 3.360252012291866e-07,
      "logits/chosen": -1.0664575099945068,
      "logits/rejected": -1.0288488864898682,
      "logps/chosen": -2.2899527549743652,
      "logps/rejected": -3.0293755531311035,
      "loss": 1.654,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -22.899526596069336,
      "rewards/margins": 7.394228458404541,
      "rewards/rejected": -30.29375648498535,
      "semantic_entropy": 0.4348418712615967,
      "step": 285
    },
    {
      "epoch": 0.6338797814207651,
      "grad_norm": 121.62024337688516,
      "learning_rate": 3.1947476107554687e-07,
      "logits/chosen": -0.9870282411575317,
      "logits/rejected": -0.9533792734146118,
      "logps/chosen": -2.296736240386963,
      "logps/rejected": -2.971813917160034,
      "loss": 1.9371,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -22.967363357543945,
      "rewards/margins": 6.750775337219238,
      "rewards/rejected": -29.7181396484375,
      "semantic_entropy": 0.4140281081199646,
      "step": 290
    },
    {
      "epoch": 0.644808743169399,
      "grad_norm": 104.51223192918087,
      "learning_rate": 3.0311495377274497e-07,
      "logits/chosen": -0.9511557817459106,
      "logits/rejected": -0.9528511166572571,
      "logps/chosen": -2.3220698833465576,
      "logps/rejected": -3.1148197650909424,
      "loss": 1.5373,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -23.2206974029541,
      "rewards/margins": 7.927499294281006,
      "rewards/rejected": -31.148197174072266,
      "semantic_entropy": 0.41551512479782104,
      "step": 295
    },
    {
      "epoch": 0.6557377049180327,
      "grad_norm": 128.04282767271587,
      "learning_rate": 2.8696967291202045e-07,
      "logits/chosen": -1.0152642726898193,
      "logits/rejected": -0.9992272257804871,
      "logps/chosen": -2.2516703605651855,
      "logps/rejected": -2.9054460525512695,
      "loss": 1.8182,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -22.51670265197754,
      "rewards/margins": 6.537758827209473,
      "rewards/rejected": -29.054462432861328,
      "semantic_entropy": 0.4529096186161041,
      "step": 300
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 130.0472804365796,
      "learning_rate": 2.710624987675287e-07,
      "logits/chosen": -0.9901336431503296,
      "logits/rejected": -0.9319698214530945,
      "logps/chosen": -2.382580518722534,
      "logps/rejected": -3.083923816680908,
      "loss": 1.8928,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -23.8258056640625,
      "rewards/margins": 7.013431549072266,
      "rewards/rejected": -30.8392333984375,
      "semantic_entropy": 0.4096023142337799,
      "step": 305
    },
    {
      "epoch": 0.6775956284153005,
      "grad_norm": 135.85405045383933,
      "learning_rate": 2.554166638572174e-07,
      "logits/chosen": -0.9127476811408997,
      "logits/rejected": -0.9295948147773743,
      "logps/chosen": -2.382692575454712,
      "logps/rejected": -3.132550001144409,
      "loss": 1.9167,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -23.826923370361328,
      "rewards/margins": 7.498571872711182,
      "rewards/rejected": -31.32549476623535,
      "semantic_entropy": 0.3914796710014343,
      "step": 310
    },
    {
      "epoch": 0.6885245901639344,
      "grad_norm": 128.2044181966222,
      "learning_rate": 2.4005501901160237e-07,
      "logits/chosen": -0.9594842791557312,
      "logits/rejected": -0.9629005193710327,
      "logps/chosen": -2.5815389156341553,
      "logps/rejected": -3.1138715744018555,
      "loss": 1.7964,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -25.815387725830078,
      "rewards/margins": 5.323328495025635,
      "rewards/rejected": -31.138717651367188,
      "semantic_entropy": 0.3702000677585602,
      "step": 315
    },
    {
      "epoch": 0.6994535519125683,
      "grad_norm": 119.46232342615077,
      "learning_rate": 2.250000000000001e-07,
      "logits/chosen": -1.0557501316070557,
      "logits/rejected": -0.9983032941818237,
      "logps/chosen": -2.5179853439331055,
      "logps/rejected": -3.2003746032714844,
      "loss": 1.7829,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -25.179855346679688,
      "rewards/margins": 6.823893070220947,
      "rewards/rejected": -32.003746032714844,
      "semantic_entropy": 0.3816224932670593,
      "step": 320
    },
    {
      "epoch": 0.7103825136612022,
      "grad_norm": 168.54223657148586,
      "learning_rate": 2.1027359476295932e-07,
      "logits/chosen": -1.0253689289093018,
      "logits/rejected": -1.0082193613052368,
      "logps/chosen": -2.3298420906066895,
      "logps/rejected": -2.9450843334198,
      "loss": 1.7987,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -23.29842185974121,
      "rewards/margins": 6.152422904968262,
      "rewards/rejected": -29.450841903686523,
      "semantic_entropy": 0.43097060918807983,
      "step": 325
    },
    {
      "epoch": 0.7213114754098361,
      "grad_norm": 116.75076161131324,
      "learning_rate": 1.9589731129874957e-07,
      "logits/chosen": -0.9325264096260071,
      "logits/rejected": -0.9063482284545898,
      "logps/chosen": -2.3737525939941406,
      "logps/rejected": -3.1893439292907715,
      "loss": 1.7401,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -23.73752212524414,
      "rewards/margins": 8.15591812133789,
      "rewards/rejected": -31.893442153930664,
      "semantic_entropy": 0.42511922121047974,
      "step": 330
    },
    {
      "epoch": 0.73224043715847,
      "grad_norm": 112.9807304911019,
      "learning_rate": 1.8189214625080817e-07,
      "logits/chosen": -0.9031323194503784,
      "logits/rejected": -0.8634198904037476,
      "logps/chosen": -2.4427285194396973,
      "logps/rejected": -3.1281769275665283,
      "loss": 1.7125,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -24.427288055419922,
      "rewards/margins": 6.854477882385254,
      "rewards/rejected": -31.28176498413086,
      "semantic_entropy": 0.4148736000061035,
      "step": 335
    },
    {
      "epoch": 0.7431693989071039,
      "grad_norm": 144.29558828539845,
      "learning_rate": 1.6827855424202287e-07,
      "logits/chosen": -0.9443111419677734,
      "logits/rejected": -0.8933967351913452,
      "logps/chosen": -2.487252712249756,
      "logps/rejected": -3.187016248703003,
      "loss": 1.935,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -24.872528076171875,
      "rewards/margins": 6.997633457183838,
      "rewards/rejected": -31.870159149169922,
      "semantic_entropy": 0.3925535976886749,
      "step": 340
    },
    {
      "epoch": 0.7540983606557377,
      "grad_norm": 124.60190547977228,
      "learning_rate": 1.550764180006395e-07,
      "logits/chosen": -1.0004165172576904,
      "logits/rejected": -0.9962259531021118,
      "logps/chosen": -2.543126106262207,
      "logps/rejected": -3.2367870807647705,
      "loss": 1.9263,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -25.43126106262207,
      "rewards/margins": 6.936606407165527,
      "rewards/rejected": -32.36787033081055,
      "semantic_entropy": 0.38035836815834045,
      "step": 345
    },
    {
      "epoch": 0.7650273224043715,
      "grad_norm": 113.81092904036393,
      "learning_rate": 1.4230501932142528e-07,
      "logits/chosen": -0.9861739277839661,
      "logits/rejected": -0.9617244601249695,
      "logps/chosen": -2.442134141921997,
      "logps/rejected": -3.0121092796325684,
      "loss": 1.8226,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -24.421342849731445,
      "rewards/margins": 5.6997480392456055,
      "rewards/rejected": -30.12109375,
      "semantic_entropy": 0.3996952772140503,
      "step": 350
    },
    {
      "epoch": 0.7759562841530054,
      "grad_norm": 95.58347606179626,
      "learning_rate": 1.2998301090449732e-07,
      "logits/chosen": -0.9443106651306152,
      "logits/rejected": -0.9029051065444946,
      "logps/chosen": -2.4110779762268066,
      "logps/rejected": -3.094087600708008,
      "loss": 1.757,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -24.110782623291016,
      "rewards/margins": 6.8300933837890625,
      "rewards/rejected": -30.940876007080078,
      "semantic_entropy": 0.4104097783565521,
      "step": 355
    },
    {
      "epoch": 0.7868852459016393,
      "grad_norm": 113.47773960736876,
      "learning_rate": 1.181283891129494e-07,
      "logits/chosen": -0.8974916338920593,
      "logits/rejected": -0.9033212661743164,
      "logps/chosen": -2.495509386062622,
      "logps/rejected": -3.1737542152404785,
      "loss": 1.8246,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -24.955089569091797,
      "rewards/margins": 6.782454490661621,
      "rewards/rejected": -31.7375431060791,
      "semantic_entropy": 0.397562712430954,
      "step": 360
    },
    {
      "epoch": 0.7978142076502732,
      "grad_norm": 115.18767750795475,
      "learning_rate": 1.0675846768906093e-07,
      "logits/chosen": -0.997270405292511,
      "logits/rejected": -0.9995688199996948,
      "logps/chosen": -2.479579210281372,
      "logps/rejected": -3.164058208465576,
      "loss": 1.913,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -24.795791625976562,
      "rewards/margins": 6.84478759765625,
      "rewards/rejected": -31.640579223632812,
      "semantic_entropy": 0.38974660634994507,
      "step": 365
    },
    {
      "epoch": 0.8087431693989071,
      "grad_norm": 132.48592611233022,
      "learning_rate": 9.588985246747924e-08,
      "logits/chosen": -1.0140767097473145,
      "logits/rejected": -0.9754125475883484,
      "logps/chosen": -2.5327906608581543,
      "logps/rejected": -3.2368435859680176,
      "loss": 1.7973,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -25.32790184020996,
      "rewards/margins": 7.040532112121582,
      "rewards/rejected": -32.36843490600586,
      "semantic_entropy": 0.37877336144447327,
      "step": 370
    },
    {
      "epoch": 0.819672131147541,
      "grad_norm": 94.56211622988587,
      "learning_rate": 8.553841712230287e-08,
      "logits/chosen": -1.0174239873886108,
      "logits/rejected": -0.9506980180740356,
      "logps/chosen": -2.4886462688446045,
      "logps/rejected": -3.2155399322509766,
      "loss": 1.7098,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -24.88646125793457,
      "rewards/margins": 7.268939018249512,
      "rewards/rejected": -32.155399322509766,
      "semantic_entropy": 0.3927253186702728,
      "step": 375
    },
    {
      "epoch": 0.8306010928961749,
      "grad_norm": 112.32547052410033,
      "learning_rate": 7.571927998349143e-08,
      "logits/chosen": -0.878455638885498,
      "logits/rejected": -0.895898163318634,
      "logps/chosen": -2.6072468757629395,
      "logps/rejected": -3.277003765106201,
      "loss": 1.7202,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": -26.072467803955078,
      "rewards/margins": 6.69757080078125,
      "rewards/rejected": -32.77003860473633,
      "semantic_entropy": 0.3782675564289093,
      "step": 380
    },
    {
      "epoch": 0.8415300546448088,
      "grad_norm": 124.71355945194576,
      "learning_rate": 6.644678195645878e-08,
      "logits/chosen": -0.9814226031303406,
      "logits/rejected": -0.9645193815231323,
      "logps/chosen": -2.661468982696533,
      "logps/rejected": -3.464935779571533,
      "loss": 2.1272,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -26.61469078063965,
      "rewards/margins": 8.034662246704102,
      "rewards/rejected": -34.649356842041016,
      "semantic_entropy": 0.34283363819122314,
      "step": 385
    },
    {
      "epoch": 0.8524590163934426,
      "grad_norm": 139.6935787868564,
      "learning_rate": 5.77344655770996e-08,
      "logits/chosen": -0.9859868884086609,
      "logits/rejected": -0.961321234703064,
      "logps/chosen": -2.5931482315063477,
      "logps/rejected": -3.4174599647521973,
      "loss": 1.7084,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -25.93148422241211,
      "rewards/margins": 8.243112564086914,
      "rewards/rejected": -34.174598693847656,
      "semantic_entropy": 0.35839056968688965,
      "step": 390
    },
    {
      "epoch": 0.8633879781420765,
      "grad_norm": 151.19395191677748,
      "learning_rate": 4.9595055232838666e-08,
      "logits/chosen": -0.9353212118148804,
      "logits/rejected": -0.9160464406013489,
      "logps/chosen": -2.742602825164795,
      "logps/rejected": -3.4214999675750732,
      "loss": 1.7141,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -27.426029205322266,
      "rewards/margins": 6.7889723777771,
      "rewards/rejected": -34.21500015258789,
      "semantic_entropy": 0.32845446467399597,
      "step": 395
    },
    {
      "epoch": 0.8743169398907104,
      "grad_norm": 132.6235448510247,
      "learning_rate": 4.204043857859128e-08,
      "logits/chosen": -0.9146641492843628,
      "logits/rejected": -0.8824892044067383,
      "logps/chosen": -2.5529472827911377,
      "logps/rejected": -3.324016571044922,
      "loss": 1.7533,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -25.52947235107422,
      "rewards/margins": 7.710694789886475,
      "rewards/rejected": -33.24016571044922,
      "semantic_entropy": 0.3849184811115265,
      "step": 400
    },
    {
      "epoch": 0.8743169398907104,
      "eval_logits/chosen": -1.217973232269287,
      "eval_logits/rejected": -1.1664073467254639,
      "eval_logps/chosen": -2.541196346282959,
      "eval_logps/rejected": -3.292675256729126,
      "eval_loss": 1.7650105953216553,
      "eval_rewards/accuracies": 0.8433734774589539,
      "eval_rewards/chosen": -25.411962509155273,
      "eval_rewards/margins": 7.514789581298828,
      "eval_rewards/rejected": -32.926753997802734,
      "eval_runtime": 37.4693,
      "eval_samples_per_second": 35.175,
      "eval_semantic_entropy": 0.371148943901062,
      "eval_steps_per_second": 2.215,
      "step": 400
    },
    {
      "epoch": 0.8852459016393442,
      "grad_norm": 144.66318541442615,
      "learning_rate": 3.508164917477564e-08,
      "logits/chosen": -0.9309503436088562,
      "logits/rejected": -0.8550406694412231,
      "logps/chosen": -2.522739887237549,
      "logps/rejected": -3.200605869293213,
      "loss": 1.8143,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -25.227399826049805,
      "rewards/margins": 6.778659820556641,
      "rewards/rejected": -32.00605773925781,
      "semantic_entropy": 0.38521987199783325,
      "step": 405
    },
    {
      "epoch": 0.8961748633879781,
      "grad_norm": 109.19303106749234,
      "learning_rate": 2.872885037273573e-08,
      "logits/chosen": -0.9444577097892761,
      "logits/rejected": -0.9762741923332214,
      "logps/chosen": -2.642904758453369,
      "logps/rejected": -3.4438624382019043,
      "loss": 1.7524,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -26.42904281616211,
      "rewards/margins": 8.0095796585083,
      "rewards/rejected": -34.438629150390625,
      "semantic_entropy": 0.35392093658447266,
      "step": 410
    },
    {
      "epoch": 0.907103825136612,
      "grad_norm": 140.70600085298142,
      "learning_rate": 2.2991320471108737e-08,
      "logits/chosen": -0.9723037481307983,
      "logits/rejected": -0.9694955945014954,
      "logps/chosen": -2.5781261920928955,
      "logps/rejected": -3.383686065673828,
      "loss": 2.0281,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -25.781265258789062,
      "rewards/margins": 8.055593490600586,
      "rewards/rejected": -33.836856842041016,
      "semantic_entropy": 0.37105756998062134,
      "step": 415
    },
    {
      "epoch": 0.9180327868852459,
      "grad_norm": 114.03524937369127,
      "learning_rate": 1.787743916481737e-08,
      "logits/chosen": -0.9833318591117859,
      "logits/rejected": -0.9596291780471802,
      "logps/chosen": -2.6041696071624756,
      "logps/rejected": -3.4229087829589844,
      "loss": 1.6196,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -26.041698455810547,
      "rewards/margins": 8.187392234802246,
      "rewards/rejected": -34.229087829589844,
      "semantic_entropy": 0.34770655632019043,
      "step": 420
    },
    {
      "epoch": 0.9289617486338798,
      "grad_norm": 105.90948195293248,
      "learning_rate": 1.3394675306478253e-08,
      "logits/chosen": -0.9838679432868958,
      "logits/rejected": -0.9677348136901855,
      "logps/chosen": -2.711045265197754,
      "logps/rejected": -3.550402879714966,
      "loss": 1.8917,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -27.110454559326172,
      "rewards/margins": 8.393570899963379,
      "rewards/rejected": -35.5040283203125,
      "semantic_entropy": 0.33215636014938354,
      "step": 425
    },
    {
      "epoch": 0.9398907103825137,
      "grad_norm": 116.33746873521659,
      "learning_rate": 9.549575998100024e-09,
      "logits/chosen": -0.9154612421989441,
      "logits/rejected": -0.9118291139602661,
      "logps/chosen": -2.4197230339050293,
      "logps/rejected": -3.2603530883789062,
      "loss": 1.5031,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": -24.197229385375977,
      "rewards/margins": 8.406302452087402,
      "rewards/rejected": -32.60353088378906,
      "semantic_entropy": 0.39185765385627747,
      "step": 430
    },
    {
      "epoch": 0.9508196721311475,
      "grad_norm": 129.942185849395,
      "learning_rate": 6.3477570290039974e-09,
      "logits/chosen": -0.967807412147522,
      "logits/rejected": -0.9948493242263794,
      "logps/chosen": -2.624267101287842,
      "logps/rejected": -3.4724323749542236,
      "loss": 1.7083,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -26.2426700592041,
      "rewards/margins": 8.481654167175293,
      "rewards/rejected": -34.72432327270508,
      "semantic_entropy": 0.3495063781738281,
      "step": 435
    },
    {
      "epoch": 0.9617486338797814,
      "grad_norm": 130.00917349864633,
      "learning_rate": 3.7938946739319575e-09,
      "logits/chosen": -0.9625166058540344,
      "logits/rejected": -0.9211773872375488,
      "logps/chosen": -2.5590291023254395,
      "logps/rejected": -3.1604175567626953,
      "loss": 1.7398,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -25.590290069580078,
      "rewards/margins": 6.013881683349609,
      "rewards/rejected": -31.604175567626953,
      "semantic_entropy": 0.3838638365268707,
      "step": 440
    },
    {
      "epoch": 0.9726775956284153,
      "grad_norm": 114.30411553161933,
      "learning_rate": 1.8917188633203684e-09,
      "logits/chosen": -0.9689086675643921,
      "logits/rejected": -0.9481277465820312,
      "logps/chosen": -2.5780527591705322,
      "logps/rejected": -3.3827826976776123,
      "loss": 1.7293,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -25.780529022216797,
      "rewards/margins": 8.0472993850708,
      "rewards/rejected": -33.82782745361328,
      "semantic_entropy": 0.3513355553150177,
      "step": 445
    },
    {
      "epoch": 0.9836065573770492,
      "grad_norm": 141.94643359249721,
      "learning_rate": 6.440077357159901e-10,
      "logits/chosen": -0.8894301652908325,
      "logits/rejected": -0.8833738565444946,
      "logps/chosen": -2.633645534515381,
      "logps/rejected": -3.3513991832733154,
      "loss": 1.7421,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -26.336456298828125,
      "rewards/margins": 7.177537441253662,
      "rewards/rejected": -33.51399230957031,
      "semantic_entropy": 0.34072810411453247,
      "step": 450
    },
    {
      "epoch": 0.994535519125683,
      "grad_norm": 119.0733411035781,
      "learning_rate": 5.258358028848442e-11,
      "logits/chosen": -0.9054166674613953,
      "logits/rejected": -0.8974553346633911,
      "logps/chosen": -2.7232465744018555,
      "logps/rejected": -3.295553684234619,
      "loss": 2.1021,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -27.23246192932129,
      "rewards/margins": 5.723072052001953,
      "rewards/rejected": -32.955535888671875,
      "semantic_entropy": 0.34370288252830505,
      "step": 455
    },
    {
      "epoch": 0.9989071038251366,
      "step": 457,
      "total_flos": 0.0,
      "train_loss": 2.585477126766347,
      "train_runtime": 5989.402,
      "train_samples_per_second": 9.777,
      "train_steps_per_second": 0.076
    }
  ],
  "logging_steps": 5,
  "max_steps": 457,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000000,
  "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": 2,
  "trial_name": null,
  "trial_params": null
}