{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.993762993762994,
  "eval_steps": 50,
  "global_step": 360,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04158004158004158,
      "grad_norm": 57.63412014296559,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7145180702209473,
      "logits/rejected": -2.7382097244262695,
      "logps/chosen": -265.24481201171875,
      "logps/rejected": -201.1150665283203,
      "loss": 0.6915,
      "rewards/accuracies": 0.375,
      "rewards/chosen": 0.01524554193019867,
      "rewards/margins": 0.008573563769459724,
      "rewards/rejected": 0.006671978626400232,
      "step": 5
    },
    {
      "epoch": 0.08316008316008316,
      "grad_norm": 52.91443405505758,
      "learning_rate": 1e-06,
      "logits/chosen": -2.690777063369751,
      "logits/rejected": -2.689495325088501,
      "logps/chosen": -259.32598876953125,
      "logps/rejected": -236.3219757080078,
      "loss": 0.6616,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": 0.4881957471370697,
      "rewards/margins": 0.10135326534509659,
      "rewards/rejected": 0.3868425786495209,
      "step": 10
    },
    {
      "epoch": 0.12474012474012475,
      "grad_norm": 53.28329333963272,
      "learning_rate": 9.994965332706572e-07,
      "logits/chosen": -2.5831141471862793,
      "logits/rejected": -2.580876350402832,
      "logps/chosen": -237.8153839111328,
      "logps/rejected": -192.7242889404297,
      "loss": 0.6359,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 1.1336857080459595,
      "rewards/margins": 0.47578057646751404,
      "rewards/rejected": 0.6579049825668335,
      "step": 15
    },
    {
      "epoch": 0.16632016632016633,
      "grad_norm": 59.51771261732688,
      "learning_rate": 9.979871469976195e-07,
      "logits/chosen": -2.5097970962524414,
      "logits/rejected": -2.491774797439575,
      "logps/chosen": -241.24349975585938,
      "logps/rejected": -206.7909698486328,
      "loss": 0.6385,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 1.5591249465942383,
      "rewards/margins": 0.980699896812439,
      "rewards/rejected": 0.5784249305725098,
      "step": 20
    },
    {
      "epoch": 0.2079002079002079,
      "grad_norm": 43.932979300169734,
      "learning_rate": 9.954748808839674e-07,
      "logits/chosen": -2.428896427154541,
      "logits/rejected": -2.426640510559082,
      "logps/chosen": -260.3008117675781,
      "logps/rejected": -223.4193878173828,
      "loss": 0.6287,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 1.4255732297897339,
      "rewards/margins": 0.9219636917114258,
      "rewards/rejected": 0.5036097764968872,
      "step": 25
    },
    {
      "epoch": 0.2494802494802495,
      "grad_norm": 51.9520224365334,
      "learning_rate": 9.919647942993147e-07,
      "logits/chosen": -2.270231008529663,
      "logits/rejected": -2.2373533248901367,
      "logps/chosen": -239.1925811767578,
      "logps/rejected": -199.03244018554688,
      "loss": 0.6013,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 1.0962295532226562,
      "rewards/margins": 0.8090855479240417,
      "rewards/rejected": 0.28714415431022644,
      "step": 30
    },
    {
      "epoch": 0.2910602910602911,
      "grad_norm": 42.07074967218912,
      "learning_rate": 9.874639560909118e-07,
      "logits/chosen": -2.2837703227996826,
      "logits/rejected": -2.2773120403289795,
      "logps/chosen": -245.8435516357422,
      "logps/rejected": -218.3318328857422,
      "loss": 0.5782,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": 0.842710018157959,
      "rewards/margins": 0.7606425285339355,
      "rewards/rejected": 0.08206750452518463,
      "step": 35
    },
    {
      "epoch": 0.33264033264033266,
      "grad_norm": 40.52902304303911,
      "learning_rate": 9.819814303479267e-07,
      "logits/chosen": -2.3677680492401123,
      "logits/rejected": -2.3257031440734863,
      "logps/chosen": -255.2162322998047,
      "logps/rejected": -206.1927490234375,
      "loss": 0.5732,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.774177610874176,
      "rewards/margins": 0.9143842458724976,
      "rewards/rejected": -0.14020657539367676,
      "step": 40
    },
    {
      "epoch": 0.37422037422037424,
      "grad_norm": 42.34370895122762,
      "learning_rate": 9.755282581475767e-07,
      "logits/chosen": -2.39888596534729,
      "logits/rejected": -2.3836829662323,
      "logps/chosen": -237.30508422851562,
      "logps/rejected": -216.06588745117188,
      "loss": 0.5958,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.7707290649414062,
      "rewards/margins": 1.1094779968261719,
      "rewards/rejected": -0.3387489914894104,
      "step": 45
    },
    {
      "epoch": 0.4158004158004158,
      "grad_norm": 46.42422960165206,
      "learning_rate": 9.681174353198686e-07,
      "logits/chosen": -2.3984861373901367,
      "logits/rejected": -2.3900020122528076,
      "logps/chosen": -233.5660400390625,
      "logps/rejected": -216.6641082763672,
      "loss": 0.635,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.9606113433837891,
      "rewards/margins": 1.124942660331726,
      "rewards/rejected": -0.16433146595954895,
      "step": 50
    },
    {
      "epoch": 0.4158004158004158,
      "eval_logits/chosen": -2.3899078369140625,
      "eval_logits/rejected": -2.3907740116119385,
      "eval_logps/chosen": -237.2779541015625,
      "eval_logps/rejected": -211.79153442382812,
      "eval_loss": 0.598716676235199,
      "eval_rewards/accuracies": 0.7013888955116272,
      "eval_rewards/chosen": 0.8451206684112549,
      "eval_rewards/margins": 0.8839507102966309,
      "eval_rewards/rejected": -0.038830023258924484,
      "eval_runtime": 229.0134,
      "eval_samples_per_second": 14.934,
      "eval_steps_per_second": 0.236,
      "step": 50
    },
    {
      "epoch": 0.4573804573804574,
      "grad_norm": 38.39211344385862,
      "learning_rate": 9.597638862757253e-07,
      "logits/chosen": -2.456921339035034,
      "logits/rejected": -2.4394099712371826,
      "logps/chosen": -237.6243133544922,
      "logps/rejected": -233.6071014404297,
      "loss": 0.6009,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.8705752491950989,
      "rewards/margins": 1.126349687576294,
      "rewards/rejected": -0.2557744085788727,
      "step": 55
    },
    {
      "epoch": 0.498960498960499,
      "grad_norm": 37.3024427630972,
      "learning_rate": 9.504844339512094e-07,
      "logits/chosen": -2.521387815475464,
      "logits/rejected": -2.4951186180114746,
      "logps/chosen": -266.1732177734375,
      "logps/rejected": -242.85897827148438,
      "loss": 0.554,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.9916666150093079,
      "rewards/margins": 1.3694941997528076,
      "rewards/rejected": -0.3778277039527893,
      "step": 60
    },
    {
      "epoch": 0.5405405405405406,
      "grad_norm": 34.5258394053866,
      "learning_rate": 9.402977659283689e-07,
      "logits/chosen": -2.4310925006866455,
      "logits/rejected": -2.4042844772338867,
      "logps/chosen": -223.9085693359375,
      "logps/rejected": -203.2917022705078,
      "loss": 0.5364,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.5567647814750671,
      "rewards/margins": 0.8208805918693542,
      "rewards/rejected": -0.26411566138267517,
      "step": 65
    },
    {
      "epoch": 0.5821205821205822,
      "grad_norm": 37.746071231755344,
      "learning_rate": 9.29224396800933e-07,
      "logits/chosen": -2.3321871757507324,
      "logits/rejected": -2.2954535484313965,
      "logps/chosen": -243.26156616210938,
      "logps/rejected": -226.8199005126953,
      "loss": 0.5423,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.38314875960350037,
      "rewards/margins": 1.1013362407684326,
      "rewards/rejected": -0.7181875109672546,
      "step": 70
    },
    {
      "epoch": 0.6237006237006237,
      "grad_norm": 57.07804647920337,
      "learning_rate": 9.172866268606513e-07,
      "logits/chosen": -2.26948881149292,
      "logits/rejected": -2.2403252124786377,
      "logps/chosen": -251.2090606689453,
      "logps/rejected": -240.20169067382812,
      "loss": 0.5844,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": 0.9941379427909851,
      "rewards/margins": 1.1448538303375244,
      "rewards/rejected": -0.1507158726453781,
      "step": 75
    },
    {
      "epoch": 0.6652806652806653,
      "grad_norm": 36.48822907311496,
      "learning_rate": 9.045084971874737e-07,
      "logits/chosen": -2.212411403656006,
      "logits/rejected": -2.1935043334960938,
      "logps/chosen": -272.8021545410156,
      "logps/rejected": -241.2564239501953,
      "loss": 0.5453,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 1.1616096496582031,
      "rewards/margins": 1.850441336631775,
      "rewards/rejected": -0.6888317465782166,
      "step": 80
    },
    {
      "epoch": 0.7068607068607069,
      "grad_norm": 38.5389032336389,
      "learning_rate": 8.909157412340149e-07,
      "logits/chosen": -2.180285930633545,
      "logits/rejected": -2.126051902770996,
      "logps/chosen": -253.0643310546875,
      "logps/rejected": -237.400390625,
      "loss": 0.5269,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.9046308398246765,
      "rewards/margins": 1.5041005611419678,
      "rewards/rejected": -0.599469780921936,
      "step": 85
    },
    {
      "epoch": 0.7484407484407485,
      "grad_norm": 31.08445554774815,
      "learning_rate": 8.765357330018055e-07,
      "logits/chosen": -2.146207332611084,
      "logits/rejected": -2.1198434829711914,
      "logps/chosen": -246.0504608154297,
      "logps/rejected": -242.79714965820312,
      "loss": 0.5741,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.4694552421569824,
      "rewards/margins": 1.537427544593811,
      "rewards/rejected": -1.0679724216461182,
      "step": 90
    },
    {
      "epoch": 0.7900207900207901,
      "grad_norm": 32.25150255167185,
      "learning_rate": 8.613974319136957e-07,
      "logits/chosen": -2.114973306655884,
      "logits/rejected": -2.0754525661468506,
      "logps/chosen": -249.64132690429688,
      "logps/rejected": -233.89352416992188,
      "loss": 0.4787,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.09324260801076889,
      "rewards/margins": 1.446690320968628,
      "rewards/rejected": -1.5399329662322998,
      "step": 95
    },
    {
      "epoch": 0.8316008316008316,
      "grad_norm": 34.22313791470385,
      "learning_rate": 8.455313244934324e-07,
      "logits/chosen": -2.027057409286499,
      "logits/rejected": -1.9819841384887695,
      "logps/chosen": -250.86947631835938,
      "logps/rejected": -241.2573699951172,
      "loss": 0.4933,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": 0.10764148086309433,
      "rewards/margins": 1.9850435256958008,
      "rewards/rejected": -1.8774023056030273,
      "step": 100
    },
    {
      "epoch": 0.8316008316008316,
      "eval_logits/chosen": -1.9530320167541504,
      "eval_logits/rejected": -1.9127507209777832,
      "eval_logps/chosen": -247.99232482910156,
      "eval_logps/rejected": -229.55447387695312,
      "eval_loss": 0.5284922122955322,
      "eval_rewards/accuracies": 0.7523148059844971,
      "eval_rewards/chosen": -0.22631923854351044,
      "eval_rewards/margins": 1.5888041257858276,
      "eval_rewards/rejected": -1.815123438835144,
      "eval_runtime": 227.9451,
      "eval_samples_per_second": 15.004,
      "eval_steps_per_second": 0.237,
      "step": 100
    },
    {
      "epoch": 0.8731808731808732,
      "grad_norm": 43.67585736596051,
      "learning_rate": 8.289693629698563e-07,
      "logits/chosen": -1.952876329421997,
      "logits/rejected": -1.9095561504364014,
      "logps/chosen": -250.580322265625,
      "logps/rejected": -255.4891357421875,
      "loss": 0.5324,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.0729295164346695,
      "rewards/margins": 1.8105415105819702,
      "rewards/rejected": -1.8834712505340576,
      "step": 105
    },
    {
      "epoch": 0.9147609147609148,
      "grad_norm": 39.1719273308311,
      "learning_rate": 8.117449009293668e-07,
      "logits/chosen": -1.8559868335723877,
      "logits/rejected": -1.799195647239685,
      "logps/chosen": -248.75961303710938,
      "logps/rejected": -232.480712890625,
      "loss": 0.4943,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.30641165375709534,
      "rewards/margins": 1.7232319116592407,
      "rewards/rejected": -1.4168202877044678,
      "step": 110
    },
    {
      "epoch": 0.9563409563409564,
      "grad_norm": 30.704674411877335,
      "learning_rate": 7.938926261462365e-07,
      "logits/chosen": -1.7355577945709229,
      "logits/rejected": -1.6728878021240234,
      "logps/chosen": -255.2344970703125,
      "logps/rejected": -221.1670684814453,
      "loss": 0.4842,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": 0.09199291467666626,
      "rewards/margins": 1.9209611415863037,
      "rewards/rejected": -1.8289684057235718,
      "step": 115
    },
    {
      "epoch": 0.997920997920998,
      "grad_norm": 41.39067500024576,
      "learning_rate": 7.754484907260512e-07,
      "logits/chosen": -1.7632324695587158,
      "logits/rejected": -1.6800066232681274,
      "logps/chosen": -236.336669921875,
      "logps/rejected": -248.4853515625,
      "loss": 0.4959,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.09794483333826065,
      "rewards/margins": 2.035454273223877,
      "rewards/rejected": -1.937509536743164,
      "step": 120
    },
    {
      "epoch": 1.0395010395010396,
      "grad_norm": 18.479902808810362,
      "learning_rate": 7.564496387029531e-07,
      "logits/chosen": -1.835211157798767,
      "logits/rejected": -1.744879126548767,
      "logps/chosen": -246.5315704345703,
      "logps/rejected": -238.4007568359375,
      "loss": 0.2765,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.2002815008163452,
      "rewards/margins": 3.028635025024414,
      "rewards/rejected": -1.8283536434173584,
      "step": 125
    },
    {
      "epoch": 1.0810810810810811,
      "grad_norm": 20.239636952023396,
      "learning_rate": 7.369343312364993e-07,
      "logits/chosen": -1.8902709484100342,
      "logits/rejected": -1.841406226158142,
      "logps/chosen": -249.97555541992188,
      "logps/rejected": -225.74893188476562,
      "loss": 0.2499,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.3934506177902222,
      "rewards/margins": 3.1044392585754395,
      "rewards/rejected": -1.7109886407852173,
      "step": 130
    },
    {
      "epoch": 1.1226611226611227,
      "grad_norm": 27.020533908406442,
      "learning_rate": 7.16941869558779e-07,
      "logits/chosen": -2.057455062866211,
      "logits/rejected": -2.003194570541382,
      "logps/chosen": -232.67001342773438,
      "logps/rejected": -234.6235809326172,
      "loss": 0.2301,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.6992406845092773,
      "rewards/margins": 3.422689914703369,
      "rewards/rejected": -1.7234487533569336,
      "step": 135
    },
    {
      "epoch": 1.1642411642411643,
      "grad_norm": 22.833931438432327,
      "learning_rate": 6.965125158269618e-07,
      "logits/chosen": -2.0912320613861084,
      "logits/rejected": -2.0458855628967285,
      "logps/chosen": -244.6804656982422,
      "logps/rejected": -225.2838897705078,
      "loss": 0.228,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 2.007575511932373,
      "rewards/margins": 3.5606026649475098,
      "rewards/rejected": -1.5530271530151367,
      "step": 140
    },
    {
      "epoch": 1.2058212058212059,
      "grad_norm": 19.05547442329841,
      "learning_rate": 6.756874120406714e-07,
      "logits/chosen": -2.0981993675231934,
      "logits/rejected": -2.0623602867126465,
      "logps/chosen": -231.7192840576172,
      "logps/rejected": -239.6023712158203,
      "loss": 0.2467,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.801042914390564,
      "rewards/margins": 3.514739513397217,
      "rewards/rejected": -1.713696837425232,
      "step": 145
    },
    {
      "epoch": 1.2474012474012475,
      "grad_norm": 17.479166649922572,
      "learning_rate": 6.545084971874736e-07,
      "logits/chosen": -2.1294615268707275,
      "logits/rejected": -2.0636110305786133,
      "logps/chosen": -238.5365447998047,
      "logps/rejected": -244.5021514892578,
      "loss": 0.2495,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.3247355222702026,
      "rewards/margins": 3.6321330070495605,
      "rewards/rejected": -2.3073973655700684,
      "step": 150
    },
    {
      "epoch": 1.2474012474012475,
      "eval_logits/chosen": -2.1232447624206543,
      "eval_logits/rejected": -2.0982720851898193,
      "eval_logps/chosen": -240.15695190429688,
      "eval_logps/rejected": -225.6040802001953,
      "eval_loss": 0.5427103042602539,
      "eval_rewards/accuracies": 0.7592592835426331,
      "eval_rewards/chosen": 0.5572175979614258,
      "eval_rewards/margins": 1.9773017168045044,
      "eval_rewards/rejected": -1.4200841188430786,
      "eval_runtime": 227.6445,
      "eval_samples_per_second": 15.023,
      "eval_steps_per_second": 0.237,
      "step": 150
    },
    {
      "epoch": 1.288981288981289,
      "grad_norm": 23.820329780522155,
      "learning_rate": 6.330184227833375e-07,
      "logits/chosen": -2.1582248210906982,
      "logits/rejected": -2.1403584480285645,
      "logps/chosen": -229.8828582763672,
      "logps/rejected": -232.87924194335938,
      "loss": 0.2962,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.5254548788070679,
      "rewards/margins": 3.3456084728240967,
      "rewards/rejected": -1.8201534748077393,
      "step": 155
    },
    {
      "epoch": 1.3305613305613306,
      "grad_norm": 17.670853805471324,
      "learning_rate": 6.112604669781572e-07,
      "logits/chosen": -2.1896843910217285,
      "logits/rejected": -2.151392698287964,
      "logps/chosen": -246.44552612304688,
      "logps/rejected": -224.73641967773438,
      "loss": 0.2666,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.2775523662567139,
      "rewards/margins": 3.355015516281128,
      "rewards/rejected": -2.077463150024414,
      "step": 160
    },
    {
      "epoch": 1.3721413721413722,
      "grad_norm": 19.884309338358804,
      "learning_rate": 5.892784473993183e-07,
      "logits/chosen": -2.1978659629821777,
      "logits/rejected": -2.139770984649658,
      "logps/chosen": -244.56283569335938,
      "logps/rejected": -235.612060546875,
      "loss": 0.2784,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.9508459568023682,
      "rewards/margins": 3.806994676589966,
      "rewards/rejected": -1.8561484813690186,
      "step": 165
    },
    {
      "epoch": 1.4137214137214138,
      "grad_norm": 21.536221614158325,
      "learning_rate": 5.671166329088277e-07,
      "logits/chosen": -2.159837245941162,
      "logits/rejected": -2.1259591579437256,
      "logps/chosen": -235.0163116455078,
      "logps/rejected": -240.001953125,
      "loss": 0.2998,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 1.634421944618225,
      "rewards/margins": 3.504880428314209,
      "rewards/rejected": -1.870458960533142,
      "step": 170
    },
    {
      "epoch": 1.4553014553014554,
      "grad_norm": 19.683343698848628,
      "learning_rate": 5.448196544517167e-07,
      "logits/chosen": -2.1057746410369873,
      "logits/rejected": -2.0693142414093018,
      "logps/chosen": -242.95492553710938,
      "logps/rejected": -222.35372924804688,
      "loss": 0.2881,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.408005952835083,
      "rewards/margins": 3.477801561355591,
      "rewards/rejected": -2.069795846939087,
      "step": 175
    },
    {
      "epoch": 1.496881496881497,
      "grad_norm": 23.26853368506061,
      "learning_rate": 5.224324151752575e-07,
      "logits/chosen": -2.0776679515838623,
      "logits/rejected": -2.0356287956237793,
      "logps/chosen": -252.8184356689453,
      "logps/rejected": -263.45013427734375,
      "loss": 0.2876,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.3347493410110474,
      "rewards/margins": 3.841991901397705,
      "rewards/rejected": -2.5072426795959473,
      "step": 180
    },
    {
      "epoch": 1.5384615384615383,
      "grad_norm": 19.309579616953823,
      "learning_rate": 5e-07,
      "logits/chosen": -2.0446431636810303,
      "logits/rejected": -1.985939621925354,
      "logps/chosen": -235.72598266601562,
      "logps/rejected": -258.97344970703125,
      "loss": 0.2758,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.3530393838882446,
      "rewards/margins": 3.8046765327453613,
      "rewards/rejected": -2.451636791229248,
      "step": 185
    },
    {
      "epoch": 1.5800415800415801,
      "grad_norm": 18.625040552548324,
      "learning_rate": 4.775675848247427e-07,
      "logits/chosen": -2.0644099712371826,
      "logits/rejected": -2.0216641426086426,
      "logps/chosen": -241.5558624267578,
      "logps/rejected": -263.6393737792969,
      "loss": 0.292,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.0546023845672607,
      "rewards/margins": 3.451650619506836,
      "rewards/rejected": -2.397047996520996,
      "step": 190
    },
    {
      "epoch": 1.6216216216216215,
      "grad_norm": 25.624591590300348,
      "learning_rate": 4.5518034554828327e-07,
      "logits/chosen": -2.0048115253448486,
      "logits/rejected": -1.9765331745147705,
      "logps/chosen": -227.64968872070312,
      "logps/rejected": -238.43795776367188,
      "loss": 0.2831,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.2037407159805298,
      "rewards/margins": 3.231950044631958,
      "rewards/rejected": -2.0282092094421387,
      "step": 195
    },
    {
      "epoch": 1.6632016632016633,
      "grad_norm": 18.88665153638521,
      "learning_rate": 4.328833670911724e-07,
      "logits/chosen": -2.0011658668518066,
      "logits/rejected": -1.9661411046981812,
      "logps/chosen": -240.28115844726562,
      "logps/rejected": -233.8397674560547,
      "loss": 0.2753,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.5664188861846924,
      "rewards/margins": 3.7791213989257812,
      "rewards/rejected": -2.212702512741089,
      "step": 200
    },
    {
      "epoch": 1.6632016632016633,
      "eval_logits/chosen": -2.0067989826202393,
      "eval_logits/rejected": -1.9752123355865479,
      "eval_logps/chosen": -239.95291137695312,
      "eval_logps/rejected": -228.13818359375,
      "eval_loss": 0.5260343551635742,
      "eval_rewards/accuracies": 0.7870370149612427,
      "eval_rewards/chosen": 0.5776218175888062,
      "eval_rewards/margins": 2.2511181831359863,
      "eval_rewards/rejected": -1.6734964847564697,
      "eval_runtime": 227.8787,
      "eval_samples_per_second": 15.008,
      "eval_steps_per_second": 0.237,
      "step": 200
    },
    {
      "epoch": 1.7047817047817047,
      "grad_norm": 20.606041672053742,
      "learning_rate": 4.107215526006817e-07,
      "logits/chosen": -2.0168378353118896,
      "logits/rejected": -2.0030291080474854,
      "logps/chosen": -241.3817901611328,
      "logps/rejected": -229.66983032226562,
      "loss": 0.2676,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.1407381296157837,
      "rewards/margins": 3.4952163696289062,
      "rewards/rejected": -2.354478359222412,
      "step": 205
    },
    {
      "epoch": 1.7463617463617465,
      "grad_norm": 22.789146320944226,
      "learning_rate": 3.8873953302184283e-07,
      "logits/chosen": -2.0401172637939453,
      "logits/rejected": -2.0455780029296875,
      "logps/chosen": -259.087646484375,
      "logps/rejected": -266.7929382324219,
      "loss": 0.2835,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.659684181213379,
      "rewards/margins": 4.7479376792907715,
      "rewards/rejected": -3.0882537364959717,
      "step": 210
    },
    {
      "epoch": 1.7879417879417878,
      "grad_norm": 21.146209859360724,
      "learning_rate": 3.669815772166625e-07,
      "logits/chosen": -2.096433162689209,
      "logits/rejected": -2.0388777256011963,
      "logps/chosen": -226.2377471923828,
      "logps/rejected": -233.63345336914062,
      "loss": 0.2949,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 1.1128900051116943,
      "rewards/margins": 3.8045170307159424,
      "rewards/rejected": -2.691626787185669,
      "step": 215
    },
    {
      "epoch": 1.8295218295218296,
      "grad_norm": 23.211306221342724,
      "learning_rate": 3.454915028125263e-07,
      "logits/chosen": -2.142961025238037,
      "logits/rejected": -2.1109135150909424,
      "logps/chosen": -254.2344970703125,
      "logps/rejected": -226.5556182861328,
      "loss": 0.3116,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.4686962366104126,
      "rewards/margins": 3.649186611175537,
      "rewards/rejected": -2.180490493774414,
      "step": 220
    },
    {
      "epoch": 1.871101871101871,
      "grad_norm": 21.234299796191088,
      "learning_rate": 3.243125879593286e-07,
      "logits/chosen": -2.1594347953796387,
      "logits/rejected": -2.073580503463745,
      "logps/chosen": -236.852783203125,
      "logps/rejected": -205.26364135742188,
      "loss": 0.2623,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.4601980447769165,
      "rewards/margins": 3.3408331871032715,
      "rewards/rejected": -1.8806352615356445,
      "step": 225
    },
    {
      "epoch": 1.9126819126819128,
      "grad_norm": 21.98980691042499,
      "learning_rate": 3.034874841730382e-07,
      "logits/chosen": -2.1095290184020996,
      "logits/rejected": -2.1008989810943604,
      "logps/chosen": -216.3356475830078,
      "logps/rejected": -237.36105346679688,
      "loss": 0.273,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 0.8807503581047058,
      "rewards/margins": 3.266737461090088,
      "rewards/rejected": -2.385986566543579,
      "step": 230
    },
    {
      "epoch": 1.9542619542619541,
      "grad_norm": 21.318480650474577,
      "learning_rate": 2.8305813044122093e-07,
      "logits/chosen": -2.1163198947906494,
      "logits/rejected": -2.0800211429595947,
      "logps/chosen": -256.31964111328125,
      "logps/rejected": -238.0819549560547,
      "loss": 0.2932,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.1696745157241821,
      "rewards/margins": 3.4981822967529297,
      "rewards/rejected": -2.328507900238037,
      "step": 235
    },
    {
      "epoch": 1.995841995841996,
      "grad_norm": 27.57270765971842,
      "learning_rate": 2.6306566876350067e-07,
      "logits/chosen": -2.1104226112365723,
      "logits/rejected": -2.059446096420288,
      "logps/chosen": -266.0362243652344,
      "logps/rejected": -221.4217987060547,
      "loss": 0.2674,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.0522549152374268,
      "rewards/margins": 3.6826469898223877,
      "rewards/rejected": -2.630392074584961,
      "step": 240
    },
    {
      "epoch": 2.0374220374220373,
      "grad_norm": 11.929615902205445,
      "learning_rate": 2.4355036129704696e-07,
      "logits/chosen": -2.1031384468078613,
      "logits/rejected": -2.043654203414917,
      "logps/chosen": -227.5277557373047,
      "logps/rejected": -237.7144012451172,
      "loss": 0.1545,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.1089200973510742,
      "rewards/margins": 4.071263790130615,
      "rewards/rejected": -2.962343692779541,
      "step": 245
    },
    {
      "epoch": 2.079002079002079,
      "grad_norm": 15.338343100063272,
      "learning_rate": 2.2455150927394878e-07,
      "logits/chosen": -2.081404685974121,
      "logits/rejected": -2.0469887256622314,
      "logps/chosen": -244.6934356689453,
      "logps/rejected": -248.8526153564453,
      "loss": 0.1584,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": 1.563730239868164,
      "rewards/margins": 4.418831825256348,
      "rewards/rejected": -2.8551013469696045,
      "step": 250
    },
    {
      "epoch": 2.079002079002079,
      "eval_logits/chosen": -2.068942070007324,
      "eval_logits/rejected": -2.0353572368621826,
      "eval_logps/chosen": -240.47457885742188,
      "eval_logps/rejected": -230.4605255126953,
      "eval_loss": 0.5117811560630798,
      "eval_rewards/accuracies": 0.7939814925193787,
      "eval_rewards/chosen": 0.5254563689231873,
      "eval_rewards/margins": 2.4311861991882324,
      "eval_rewards/rejected": -1.905730128288269,
      "eval_runtime": 227.834,
      "eval_samples_per_second": 15.011,
      "eval_steps_per_second": 0.237,
      "step": 250
    },
    {
      "epoch": 2.1205821205821205,
      "grad_norm": 16.808807842784802,
      "learning_rate": 2.0610737385376348e-07,
      "logits/chosen": -2.070066452026367,
      "logits/rejected": -2.032356023788452,
      "logps/chosen": -222.1339874267578,
      "logps/rejected": -214.38522338867188,
      "loss": 0.1768,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.210902452468872,
      "rewards/margins": 3.55863881111145,
      "rewards/rejected": -2.347736358642578,
      "step": 255
    },
    {
      "epoch": 2.1621621621621623,
      "grad_norm": 19.515515285926487,
      "learning_rate": 1.8825509907063326e-07,
      "logits/chosen": -2.078474760055542,
      "logits/rejected": -2.0414175987243652,
      "logps/chosen": -213.68118286132812,
      "logps/rejected": -248.75192260742188,
      "loss": 0.1721,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.2889325618743896,
      "rewards/margins": 3.8671317100524902,
      "rewards/rejected": -2.5781993865966797,
      "step": 260
    },
    {
      "epoch": 2.2037422037422036,
      "grad_norm": 15.747420635068087,
      "learning_rate": 1.710306370301437e-07,
      "logits/chosen": -2.096172332763672,
      "logits/rejected": -2.0482194423675537,
      "logps/chosen": -232.9115753173828,
      "logps/rejected": -231.60861206054688,
      "loss": 0.158,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 1.6092205047607422,
      "rewards/margins": 4.0434794425964355,
      "rewards/rejected": -2.4342594146728516,
      "step": 265
    },
    {
      "epoch": 2.2453222453222454,
      "grad_norm": 16.455326384704957,
      "learning_rate": 1.5446867550656767e-07,
      "logits/chosen": -2.112180233001709,
      "logits/rejected": -2.03873872756958,
      "logps/chosen": -244.0580291748047,
      "logps/rejected": -238.3592071533203,
      "loss": 0.1733,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.4473066329956055,
      "rewards/margins": 3.8618369102478027,
      "rewards/rejected": -2.414530038833618,
      "step": 270
    },
    {
      "epoch": 2.286902286902287,
      "grad_norm": 15.072874984273307,
      "learning_rate": 1.3860256808630427e-07,
      "logits/chosen": -2.134932041168213,
      "logits/rejected": -2.070984363555908,
      "logps/chosen": -250.3199005126953,
      "logps/rejected": -252.6285400390625,
      "loss": 0.1657,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.7358472347259521,
      "rewards/margins": 4.2915215492248535,
      "rewards/rejected": -2.5556740760803223,
      "step": 275
    },
    {
      "epoch": 2.3284823284823286,
      "grad_norm": 13.48600883883482,
      "learning_rate": 1.2346426699819456e-07,
      "logits/chosen": -2.1141560077667236,
      "logits/rejected": -2.0565381050109863,
      "logps/chosen": -224.2864532470703,
      "logps/rejected": -223.65731811523438,
      "loss": 0.1445,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.6165220737457275,
      "rewards/margins": 4.022408485412598,
      "rewards/rejected": -2.40588641166687,
      "step": 280
    },
    {
      "epoch": 2.37006237006237,
      "grad_norm": 13.775658471233747,
      "learning_rate": 1.090842587659851e-07,
      "logits/chosen": -2.0836844444274902,
      "logits/rejected": -2.033290386199951,
      "logps/chosen": -235.9185333251953,
      "logps/rejected": -233.7434539794922,
      "loss": 0.1616,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.657187819480896,
      "rewards/margins": 4.302671432495117,
      "rewards/rejected": -2.6454834938049316,
      "step": 285
    },
    {
      "epoch": 2.4116424116424118,
      "grad_norm": 19.09270347908536,
      "learning_rate": 9.549150281252632e-08,
      "logits/chosen": -2.122148036956787,
      "logits/rejected": -2.049131393432617,
      "logps/chosen": -262.602294921875,
      "logps/rejected": -225.87008666992188,
      "loss": 0.1501,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.8016000986099243,
      "rewards/margins": 4.24440860748291,
      "rewards/rejected": -2.4428083896636963,
      "step": 290
    },
    {
      "epoch": 2.453222453222453,
      "grad_norm": 14.877658534967546,
      "learning_rate": 8.271337313934867e-08,
      "logits/chosen": -2.0919740200042725,
      "logits/rejected": -2.0363011360168457,
      "logps/chosen": -246.9591827392578,
      "logps/rejected": -240.05551147460938,
      "loss": 0.1505,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.5468189716339111,
      "rewards/margins": 4.176497459411621,
      "rewards/rejected": -2.629678249359131,
      "step": 295
    },
    {
      "epoch": 2.494802494802495,
      "grad_norm": 19.060658903318366,
      "learning_rate": 7.077560319906694e-08,
      "logits/chosen": -2.0859265327453613,
      "logits/rejected": -2.04348087310791,
      "logps/chosen": -243.56875610351562,
      "logps/rejected": -244.1935577392578,
      "loss": 0.1572,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 1.8503570556640625,
      "rewards/margins": 4.711921691894531,
      "rewards/rejected": -2.8615641593933105,
      "step": 300
    },
    {
      "epoch": 2.494802494802495,
      "eval_logits/chosen": -2.0940816402435303,
      "eval_logits/rejected": -2.0615904331207275,
      "eval_logps/chosen": -238.14688110351562,
      "eval_logps/rejected": -228.6629180908203,
      "eval_loss": 0.5261171460151672,
      "eval_rewards/accuracies": 0.7986111044883728,
      "eval_rewards/chosen": 0.7582256197929382,
      "eval_rewards/margins": 2.484196424484253,
      "eval_rewards/rejected": -1.725970983505249,
      "eval_runtime": 227.7859,
      "eval_samples_per_second": 15.014,
      "eval_steps_per_second": 0.237,
      "step": 300
    },
    {
      "epoch": 2.5363825363825363,
      "grad_norm": 16.633910939695856,
      "learning_rate": 5.9702234071631e-08,
      "logits/chosen": -2.1052534580230713,
      "logits/rejected": -2.059093952178955,
      "logps/chosen": -238.5347442626953,
      "logps/rejected": -266.62548828125,
      "loss": 0.1602,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.5432864427566528,
      "rewards/margins": 4.184087753295898,
      "rewards/rejected": -2.6408019065856934,
      "step": 305
    },
    {
      "epoch": 2.577962577962578,
      "grad_norm": 14.797157748323585,
      "learning_rate": 4.951556604879048e-08,
      "logits/chosen": -2.0979275703430176,
      "logits/rejected": -2.0617194175720215,
      "logps/chosen": -222.8668975830078,
      "logps/rejected": -248.1284942626953,
      "loss": 0.1382,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.5113767385482788,
      "rewards/margins": 4.292911529541016,
      "rewards/rejected": -2.7815351486206055,
      "step": 310
    },
    {
      "epoch": 2.6195426195426195,
      "grad_norm": 14.963028269861777,
      "learning_rate": 4.023611372427471e-08,
      "logits/chosen": -2.0911357402801514,
      "logits/rejected": -2.0792324542999268,
      "logps/chosen": -242.3371124267578,
      "logps/rejected": -251.0138702392578,
      "loss": 0.1609,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.7089687585830688,
      "rewards/margins": 4.530019760131836,
      "rewards/rejected": -2.8210511207580566,
      "step": 315
    },
    {
      "epoch": 2.6611226611226613,
      "grad_norm": 13.588642778305575,
      "learning_rate": 3.188256468013139e-08,
      "logits/chosen": -2.1075356006622314,
      "logits/rejected": -2.0502302646636963,
      "logps/chosen": -231.9967041015625,
      "logps/rejected": -232.72366333007812,
      "loss": 0.157,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 1.5209684371948242,
      "rewards/margins": 4.352679252624512,
      "rewards/rejected": -2.8317112922668457,
      "step": 320
    },
    {
      "epoch": 2.7027027027027026,
      "grad_norm": 15.25792795245387,
      "learning_rate": 2.4471741852423233e-08,
      "logits/chosen": -2.125161647796631,
      "logits/rejected": -2.066580057144165,
      "logps/chosen": -240.263916015625,
      "logps/rejected": -235.22427368164062,
      "loss": 0.1656,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.425862431526184,
      "rewards/margins": 4.251040458679199,
      "rewards/rejected": -2.8251776695251465,
      "step": 325
    },
    {
      "epoch": 2.7442827442827444,
      "grad_norm": 13.021461954217658,
      "learning_rate": 1.8018569652073378e-08,
      "logits/chosen": -2.107436418533325,
      "logits/rejected": -2.0602996349334717,
      "logps/chosen": -213.11068725585938,
      "logps/rejected": -219.4556427001953,
      "loss": 0.1491,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.2270827293395996,
      "rewards/margins": 3.6631665229797363,
      "rewards/rejected": -2.4360833168029785,
      "step": 330
    },
    {
      "epoch": 2.785862785862786,
      "grad_norm": 16.103241816777906,
      "learning_rate": 1.253604390908819e-08,
      "logits/chosen": -2.1171445846557617,
      "logits/rejected": -2.0631439685821533,
      "logps/chosen": -223.59353637695312,
      "logps/rejected": -245.2403564453125,
      "loss": 0.1297,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.4035135507583618,
      "rewards/margins": 4.3274335861206055,
      "rewards/rejected": -2.923920154571533,
      "step": 335
    },
    {
      "epoch": 2.8274428274428276,
      "grad_norm": 22.633338136184896,
      "learning_rate": 8.035205700685165e-09,
      "logits/chosen": -2.111961841583252,
      "logits/rejected": -2.0561139583587646,
      "logps/chosen": -235.6881561279297,
      "logps/rejected": -224.3719482421875,
      "loss": 0.173,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.6374908685684204,
      "rewards/margins": 4.2424821853637695,
      "rewards/rejected": -2.6049914360046387,
      "step": 340
    },
    {
      "epoch": 2.869022869022869,
      "grad_norm": 15.753996313911308,
      "learning_rate": 4.5251191160326495e-09,
      "logits/chosen": -2.1138930320739746,
      "logits/rejected": -2.0703036785125732,
      "logps/chosen": -238.215087890625,
      "logps/rejected": -241.9582977294922,
      "loss": 0.1658,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.6625125408172607,
      "rewards/margins": 4.42767333984375,
      "rewards/rejected": -2.7651610374450684,
      "step": 345
    },
    {
      "epoch": 2.9106029106029108,
      "grad_norm": 13.276623107827307,
      "learning_rate": 2.0128530023804656e-09,
      "logits/chosen": -2.134644031524658,
      "logits/rejected": -2.074676036834717,
      "logps/chosen": -253.3544921875,
      "logps/rejected": -243.7964630126953,
      "loss": 0.1557,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.9221925735473633,
      "rewards/margins": 4.552640438079834,
      "rewards/rejected": -2.6304478645324707,
      "step": 350
    },
    {
      "epoch": 2.9106029106029108,
      "eval_logits/chosen": -2.102605104446411,
      "eval_logits/rejected": -2.070624589920044,
      "eval_logps/chosen": -239.3154296875,
      "eval_logps/rejected": -230.46446228027344,
      "eval_loss": 0.5264930129051208,
      "eval_rewards/accuracies": 0.7986111044883728,
      "eval_rewards/chosen": 0.6413695216178894,
      "eval_rewards/margins": 2.5474934577941895,
      "eval_rewards/rejected": -1.9061238765716553,
      "eval_runtime": 227.8625,
      "eval_samples_per_second": 15.009,
      "eval_steps_per_second": 0.237,
      "step": 350
    },
    {
      "epoch": 2.952182952182952,
      "grad_norm": 27.5358847385431,
      "learning_rate": 5.034667293427053e-10,
      "logits/chosen": -2.1214892864227295,
      "logits/rejected": -2.085740327835083,
      "logps/chosen": -227.8209991455078,
      "logps/rejected": -250.01199340820312,
      "loss": 0.1754,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.3233916759490967,
      "rewards/margins": 3.8978431224823,
      "rewards/rejected": -2.574451446533203,
      "step": 355
    },
    {
      "epoch": 2.993762993762994,
      "grad_norm": 15.431922041744329,
      "learning_rate": 0.0,
      "logits/chosen": -2.109636068344116,
      "logits/rejected": -2.076749563217163,
      "logps/chosen": -246.1726531982422,
      "logps/rejected": -251.31954956054688,
      "loss": 0.1308,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.7782920598983765,
      "rewards/margins": 4.2874860763549805,
      "rewards/rejected": -2.5091946125030518,
      "step": 360
    }
  ],
  "logging_steps": 5,
  "max_steps": 360,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4244978440077312.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}