{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 197,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.167724609375,
      "epoch": 0.005089058524173028,
      "grad_norm": 63.83088684082031,
      "learning_rate": 5e-06,
      "logits/chosen": -1.1478307765991018,
      "logits/rejected": -1.1531979437304007,
      "logps/chosen": -230.5,
      "logps/rejected": -250.5,
      "loss": 0.726806640625,
      "mean_token_accuracy": 0.9508575797080994,
      "num_tokens": 387774.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.0360107421875,
      "rewards/margins": 0.007598876953125,
      "rewards/rejected": -0.04305267333984375,
      "step": 1
    },
    {
      "entropy": 0.1708984375,
      "epoch": 0.010178117048346057,
      "grad_norm": 58.35269546508789,
      "learning_rate": 4.974619289340102e-06,
      "logits/chosen": -1.1465309061137527,
      "logits/rejected": -1.1435811743754352,
      "logps/chosen": -202.75,
      "logps/rejected": -220.75,
      "loss": 0.73284912109375,
      "mean_token_accuracy": 0.9518170654773712,
      "num_tokens": 796045.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.01251220703125,
      "rewards/margins": -0.01416015625,
      "rewards/rejected": 0.00146484375,
      "step": 2
    },
    {
      "entropy": 0.1708984375,
      "epoch": 0.015267175572519083,
      "grad_norm": 52.742557525634766,
      "learning_rate": 4.949238578680203e-06,
      "logits/chosen": -1.22388826039798,
      "logits/rejected": -1.2534275052969308,
      "logps/chosen": -217.0,
      "logps/rejected": -210.75,
      "loss": 0.695587158203125,
      "mean_token_accuracy": 0.9499112516641617,
      "num_tokens": 1158499.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.03204345703125,
      "rewards/margins": 0.07501220703125,
      "rewards/rejected": -0.1068115234375,
      "step": 3
    },
    {
      "entropy": 0.175537109375,
      "epoch": 0.020356234096692113,
      "grad_norm": 58.555728912353516,
      "learning_rate": 4.923857868020305e-06,
      "logits/chosen": -1.0887703188820197,
      "logits/rejected": -1.110711324029513,
      "logps/chosen": -247.5,
      "logps/rejected": -225.25,
      "loss": 0.7191162109375,
      "mean_token_accuracy": 0.9449715912342072,
      "num_tokens": 1526178.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.0124664306640625,
      "rewards/margins": 0.05133056640625,
      "rewards/rejected": -0.0640869140625,
      "step": 4
    },
    {
      "entropy": 0.181884765625,
      "epoch": 0.02544529262086514,
      "grad_norm": 57.25611114501953,
      "learning_rate": 4.898477157360406e-06,
      "logits/chosen": -1.1241738300809683,
      "logits/rejected": -1.1185368980659924,
      "logps/chosen": -197.25,
      "logps/rejected": -204.5,
      "loss": 0.746917724609375,
      "mean_token_accuracy": 0.9447457045316696,
      "num_tokens": 1898739.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.036865234375,
      "rewards/margins": 0.0240478515625,
      "rewards/rejected": -0.06103515625,
      "step": 5
    },
    {
      "entropy": 0.172119140625,
      "epoch": 0.030534351145038167,
      "grad_norm": 55.632781982421875,
      "learning_rate": 4.873096446700508e-06,
      "logits/chosen": -1.118042467736593,
      "logits/rejected": -1.1188858755620001,
      "logps/chosen": -195.5,
      "logps/rejected": -213.5,
      "loss": 0.706695556640625,
      "mean_token_accuracy": 0.9515382200479507,
      "num_tokens": 2280846.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.0267333984375,
      "rewards/margins": 0.051971435546875,
      "rewards/rejected": -0.07904052734375,
      "step": 6
    },
    {
      "entropy": 0.165283203125,
      "epoch": 0.035623409669211195,
      "grad_norm": 58.1217155456543,
      "learning_rate": 4.84771573604061e-06,
      "logits/chosen": -1.2056221529623334,
      "logits/rejected": -1.1698524733043854,
      "logps/chosen": -203.75,
      "logps/rejected": -237.5,
      "loss": 0.778106689453125,
      "mean_token_accuracy": 0.9536973237991333,
      "num_tokens": 2661565.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.047119140625,
      "rewards/margins": -0.0155029296875,
      "rewards/rejected": -0.031982421875,
      "step": 7
    },
    {
      "entropy": 0.16015625,
      "epoch": 0.04071246819338423,
      "grad_norm": 57.468605041503906,
      "learning_rate": 4.822335025380711e-06,
      "logits/chosen": -1.285398958255684,
      "logits/rejected": -1.2651885553215019,
      "logps/chosen": -185.25,
      "logps/rejected": -195.0,
      "loss": 0.76959228515625,
      "mean_token_accuracy": 0.9562991708517075,
      "num_tokens": 3011121.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.07479095458984375,
      "rewards/margins": -0.06903076171875,
      "rewards/rejected": -0.0052642822265625,
      "step": 8
    },
    {
      "entropy": 0.177978515625,
      "epoch": 0.04580152671755725,
      "grad_norm": 61.59230041503906,
      "learning_rate": 4.796954314720812e-06,
      "logits/chosen": -1.1463901211868202,
      "logits/rejected": -1.151384163231262,
      "logps/chosen": -262.0,
      "logps/rejected": -263.0,
      "loss": 0.77764892578125,
      "mean_token_accuracy": 0.9490418434143066,
      "num_tokens": 3364659.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.0624542236328125,
      "rewards/margins": -0.07977294921875,
      "rewards/rejected": 0.017181396484375,
      "step": 9
    },
    {
      "entropy": 0.17041015625,
      "epoch": 0.05089058524173028,
      "grad_norm": 66.9396743774414,
      "learning_rate": 4.771573604060914e-06,
      "logits/chosen": -1.1312320566718426,
      "logits/rejected": -1.114105189953177,
      "logps/chosen": -246.25,
      "logps/rejected": -253.5,
      "loss": 0.784912109375,
      "mean_token_accuracy": 0.9495462626218796,
      "num_tokens": 3756722.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.11962890625,
      "rewards/margins": -0.0777587890625,
      "rewards/rejected": -0.0416259765625,
      "step": 10
    },
    {
      "entropy": 0.17041015625,
      "epoch": 0.05597964376590331,
      "grad_norm": 60.69718933105469,
      "learning_rate": 4.746192893401016e-06,
      "logits/chosen": -1.2082123033021486,
      "logits/rejected": -1.240414757271401,
      "logps/chosen": -206.75,
      "logps/rejected": -253.0,
      "loss": 0.77130126953125,
      "mean_token_accuracy": 0.9514396339654922,
      "num_tokens": 4150502.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.122711181640625,
      "rewards/margins": -0.0660400390625,
      "rewards/rejected": -0.056396484375,
      "step": 11
    },
    {
      "entropy": 0.166015625,
      "epoch": 0.061068702290076333,
      "grad_norm": 68.75598907470703,
      "learning_rate": 4.7208121827411175e-06,
      "logits/chosen": -1.1497054731535286,
      "logits/rejected": -1.1323881415685357,
      "logps/chosen": -233.0,
      "logps/rejected": -214.0,
      "loss": 0.81524658203125,
      "mean_token_accuracy": 0.9499669224023819,
      "num_tokens": 4548017.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.14825439453125,
      "rewards/margins": -0.148681640625,
      "rewards/rejected": 0.0008544921875,
      "step": 12
    },
    {
      "entropy": 0.171142578125,
      "epoch": 0.06615776081424936,
      "grad_norm": 54.791534423828125,
      "learning_rate": 4.695431472081219e-06,
      "logits/chosen": -1.1963053139896285,
      "logits/rejected": -1.164336637055035,
      "logps/chosen": -204.0,
      "logps/rejected": -226.5,
      "loss": 0.71929931640625,
      "mean_token_accuracy": 0.952528104186058,
      "num_tokens": 4902525.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.0484619140625,
      "rewards/margins": 0.0706787109375,
      "rewards/rejected": -0.1187744140625,
      "step": 13
    },
    {
      "entropy": 0.165771484375,
      "epoch": 0.07124681933842239,
      "grad_norm": 55.923667907714844,
      "learning_rate": 4.67005076142132e-06,
      "logits/chosen": -1.227218462140038,
      "logits/rejected": -1.2252739974581663,
      "logps/chosen": -204.0,
      "logps/rejected": -245.75,
      "loss": 0.681884765625,
      "mean_token_accuracy": 0.9524703174829483,
      "num_tokens": 5247776.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": 0.017791748046875,
      "rewards/margins": 0.07401275634765625,
      "rewards/rejected": -0.056304931640625,
      "step": 14
    },
    {
      "entropy": 0.16796875,
      "epoch": 0.07633587786259542,
      "grad_norm": 54.7180290222168,
      "learning_rate": 4.644670050761422e-06,
      "logits/chosen": -1.1794452967378677,
      "logits/rejected": -1.213616516672937,
      "logps/chosen": -208.75,
      "logps/rejected": -199.5,
      "loss": 0.7034912109375,
      "mean_token_accuracy": 0.9460672587156296,
      "num_tokens": 5614879.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.097320556640625,
      "rewards/margins": 0.0595703125,
      "rewards/rejected": -0.15673828125,
      "step": 15
    },
    {
      "entropy": 0.176513671875,
      "epoch": 0.08142493638676845,
      "grad_norm": 62.994319915771484,
      "learning_rate": 4.6192893401015235e-06,
      "logits/chosen": -1.2212638279189771,
      "logits/rejected": -1.2088961146919388,
      "logps/chosen": -227.25,
      "logps/rejected": -213.0,
      "loss": 0.83746337890625,
      "mean_token_accuracy": 0.9471840560436249,
      "num_tokens": 5971613.0,
      "rewards/accuracies": 0.28125,
      "rewards/chosen": -0.173828125,
      "rewards/margins": -0.18463134765625,
      "rewards/rejected": 0.01068115234375,
      "step": 16
    },
    {
      "entropy": 0.169677734375,
      "epoch": 0.08651399491094147,
      "grad_norm": 60.76310348510742,
      "learning_rate": 4.593908629441624e-06,
      "logits/chosen": -1.2065562340385436,
      "logits/rejected": -1.2022840016395056,
      "logps/chosen": -218.5,
      "logps/rejected": -190.0,
      "loss": 0.77734375,
      "mean_token_accuracy": 0.9474756270647049,
      "num_tokens": 6368369.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.067138671875,
      "rewards/margins": -0.05255126953125,
      "rewards/rejected": -0.0137939453125,
      "step": 17
    },
    {
      "entropy": 0.1748046875,
      "epoch": 0.0916030534351145,
      "grad_norm": 61.58626174926758,
      "learning_rate": 4.568527918781726e-06,
      "logits/chosen": -1.1644959212587278,
      "logits/rejected": -1.1554610862237855,
      "logps/chosen": -213.5,
      "logps/rejected": -256.5,
      "loss": 0.744384765625,
      "mean_token_accuracy": 0.9518721997737885,
      "num_tokens": 6749942.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.1009521484375,
      "rewards/margins": 0.003753662109375,
      "rewards/rejected": -0.104736328125,
      "step": 18
    },
    {
      "entropy": 0.170654296875,
      "epoch": 0.09669211195928754,
      "grad_norm": 69.10458374023438,
      "learning_rate": 4.543147208121828e-06,
      "logits/chosen": -1.2018930915056671,
      "logits/rejected": -1.1882594101155102,
      "logps/chosen": -208.75,
      "logps/rejected": -223.0,
      "loss": 0.7578125,
      "mean_token_accuracy": 0.9503921419382095,
      "num_tokens": 7150669.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.16357421875,
      "rewards/margins": -0.0054931640625,
      "rewards/rejected": -0.157958984375,
      "step": 19
    },
    {
      "entropy": 0.1708984375,
      "epoch": 0.10178117048346055,
      "grad_norm": 59.20901870727539,
      "learning_rate": 4.5177664974619295e-06,
      "logits/chosen": -1.153684045887259,
      "logits/rejected": -1.1406723101806675,
      "logps/chosen": -214.25,
      "logps/rejected": -235.75,
      "loss": 0.742584228515625,
      "mean_token_accuracy": 0.9510545581579208,
      "num_tokens": 7577633.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.1868896484375,
      "rewards/margins": -0.008544921875,
      "rewards/rejected": -0.17822265625,
      "step": 20
    },
    {
      "entropy": 0.160888671875,
      "epoch": 0.10687022900763359,
      "grad_norm": 53.25274658203125,
      "learning_rate": 4.492385786802031e-06,
      "logits/chosen": -1.2115618534718147,
      "logits/rejected": -1.1697568144140837,
      "logps/chosen": -220.5,
      "logps/rejected": -224.5,
      "loss": 0.6365966796875,
      "mean_token_accuracy": 0.954043835401535,
      "num_tokens": 7935697.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.0498046875,
      "rewards/margins": 0.22021484375,
      "rewards/rejected": -0.2708282470703125,
      "step": 21
    },
    {
      "entropy": 0.1708984375,
      "epoch": 0.11195928753180662,
      "grad_norm": 64.42499542236328,
      "learning_rate": 4.467005076142132e-06,
      "logits/chosen": -1.163452744541083,
      "logits/rejected": -1.1574276551228084,
      "logps/chosen": -206.75,
      "logps/rejected": -249.0,
      "loss": 0.7249755859375,
      "mean_token_accuracy": 0.9557177871465683,
      "num_tokens": 8344719.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.064208984375,
      "rewards/margins": 0.043914794921875,
      "rewards/rejected": -0.108154296875,
      "step": 22
    },
    {
      "entropy": 0.1650390625,
      "epoch": 0.11704834605597965,
      "grad_norm": 57.46247863769531,
      "learning_rate": 4.441624365482234e-06,
      "logits/chosen": -1.184751779280805,
      "logits/rejected": -1.1601956141572203,
      "logps/chosen": -221.5,
      "logps/rejected": -223.5,
      "loss": 0.725341796875,
      "mean_token_accuracy": 0.9495507925748825,
      "num_tokens": 8715576.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.12018585205078125,
      "rewards/margins": 0.1531982421875,
      "rewards/rejected": -0.27362060546875,
      "step": 23
    },
    {
      "entropy": 0.166259765625,
      "epoch": 0.12213740458015267,
      "grad_norm": 88.59230041503906,
      "learning_rate": 4.4162436548223355e-06,
      "logits/chosen": -1.2269864585121328,
      "logits/rejected": -1.2167687340565112,
      "logps/chosen": -190.5,
      "logps/rejected": -223.25,
      "loss": 0.616790771484375,
      "mean_token_accuracy": 0.9541012793779373,
      "num_tokens": 9131729.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.0047607421875,
      "rewards/margins": 0.2396240234375,
      "rewards/rejected": -0.244873046875,
      "step": 24
    },
    {
      "entropy": 0.174560546875,
      "epoch": 0.1272264631043257,
      "grad_norm": 59.16816711425781,
      "learning_rate": 4.390862944162436e-06,
      "logits/chosen": -1.2085149622005857,
      "logits/rejected": -1.1962000528044974,
      "logps/chosen": -209.5,
      "logps/rejected": -248.25,
      "loss": 0.68780517578125,
      "mean_token_accuracy": 0.9498510509729385,
      "num_tokens": 9510884.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.14385986328125,
      "rewards/margins": 0.06866455078125,
      "rewards/rejected": -0.212890625,
      "step": 25
    },
    {
      "entropy": 0.166259765625,
      "epoch": 0.13231552162849872,
      "grad_norm": 58.58485412597656,
      "learning_rate": 4.365482233502538e-06,
      "logits/chosen": -1.2357831115334332,
      "logits/rejected": -1.2414339305565623,
      "logps/chosen": -205.25,
      "logps/rejected": -210.0,
      "loss": 0.735107421875,
      "mean_token_accuracy": 0.9495077282190323,
      "num_tokens": 9892399.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.144287109375,
      "rewards/margins": 0.02099609375,
      "rewards/rejected": -0.16522216796875,
      "step": 26
    },
    {
      "entropy": 0.17333984375,
      "epoch": 0.13740458015267176,
      "grad_norm": 51.31973648071289,
      "learning_rate": 4.34010152284264e-06,
      "logits/chosen": -1.1937940321742593,
      "logits/rejected": -1.20553321824887,
      "logps/chosen": -185.875,
      "logps/rejected": -204.75,
      "loss": 0.66217041015625,
      "mean_token_accuracy": 0.9475287944078445,
      "num_tokens": 10284784.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.072509765625,
      "rewards/margins": 0.17462158203125,
      "rewards/rejected": -0.247314453125,
      "step": 27
    },
    {
      "entropy": 0.16552734375,
      "epoch": 0.14249363867684478,
      "grad_norm": 61.93020248413086,
      "learning_rate": 4.3147208121827415e-06,
      "logits/chosen": -1.2497175985959397,
      "logits/rejected": -1.2349540219835924,
      "logps/chosen": -185.5,
      "logps/rejected": -200.5,
      "loss": 0.785888671875,
      "mean_token_accuracy": 0.9526190459728241,
      "num_tokens": 10640064.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.24853515625,
      "rewards/margins": -0.0711669921875,
      "rewards/rejected": -0.1773681640625,
      "step": 28
    },
    {
      "entropy": 0.168701171875,
      "epoch": 0.1475826972010178,
      "grad_norm": 47.82231903076172,
      "learning_rate": 4.289340101522843e-06,
      "logits/chosen": -1.2240958044550552,
      "logits/rejected": -1.2331657411234425,
      "logps/chosen": -164.75,
      "logps/rejected": -183.5,
      "loss": 0.64813232421875,
      "mean_token_accuracy": 0.9512129127979279,
      "num_tokens": 11043531.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.0230712890625,
      "rewards/margins": 0.1962890625,
      "rewards/rejected": -0.21942138671875,
      "step": 29
    },
    {
      "entropy": 0.166015625,
      "epoch": 0.15267175572519084,
      "grad_norm": 64.59899139404297,
      "learning_rate": 4.263959390862945e-06,
      "logits/chosen": -1.1938851832774966,
      "logits/rejected": -1.2063193481273322,
      "logps/chosen": -203.375,
      "logps/rejected": -209.25,
      "loss": 0.75030517578125,
      "mean_token_accuracy": 0.9482824355363846,
      "num_tokens": 11433645.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.160888671875,
      "rewards/margins": 0.0379638671875,
      "rewards/rejected": -0.1988525390625,
      "step": 30
    },
    {
      "entropy": 0.174560546875,
      "epoch": 0.15776081424936386,
      "grad_norm": 50.42904281616211,
      "learning_rate": 4.238578680203046e-06,
      "logits/chosen": -1.1847193683809207,
      "logits/rejected": -1.2039758508125116,
      "logps/chosen": -187.0,
      "logps/rejected": -196.0,
      "loss": 0.6805324554443359,
      "mean_token_accuracy": 0.9458457380533218,
      "num_tokens": 11802241.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.01849365234375,
      "rewards/margins": 0.2666015625,
      "rewards/rejected": -0.285400390625,
      "step": 31
    },
    {
      "entropy": 0.164794921875,
      "epoch": 0.1628498727735369,
      "grad_norm": 65.630126953125,
      "learning_rate": 4.2131979695431475e-06,
      "logits/chosen": -1.2242753640413924,
      "logits/rejected": -1.2582800165640973,
      "logps/chosen": -237.25,
      "logps/rejected": -239.75,
      "loss": 0.762481689453125,
      "mean_token_accuracy": 0.9494172036647797,
      "num_tokens": 12168742.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.256591796875,
      "rewards/margins": 0.0673828125,
      "rewards/rejected": -0.323486328125,
      "step": 32
    },
    {
      "entropy": 0.1669921875,
      "epoch": 0.16793893129770993,
      "grad_norm": 64.15123748779297,
      "learning_rate": 4.187817258883249e-06,
      "logits/chosen": -1.2054564969855512,
      "logits/rejected": -1.2242075139011555,
      "logps/chosen": -225.75,
      "logps/rejected": -242.25,
      "loss": 0.776153564453125,
      "mean_token_accuracy": 0.9507817625999451,
      "num_tokens": 12544019.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.259033203125,
      "rewards/margins": -0.0609130859375,
      "rewards/rejected": -0.1982421875,
      "step": 33
    },
    {
      "entropy": 0.16162109375,
      "epoch": 0.17302798982188294,
      "grad_norm": 59.748802185058594,
      "learning_rate": 4.162436548223351e-06,
      "logits/chosen": -1.1906290703191211,
      "logits/rejected": -1.2255973378124254,
      "logps/chosen": -185.5,
      "logps/rejected": -189.5,
      "loss": 0.81256103515625,
      "mean_token_accuracy": 0.951210081577301,
      "num_tokens": 12934809.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.261016845703125,
      "rewards/margins": -0.1204071044921875,
      "rewards/rejected": -0.140380859375,
      "step": 34
    },
    {
      "entropy": 0.160400390625,
      "epoch": 0.178117048346056,
      "grad_norm": 62.484031677246094,
      "learning_rate": 4.137055837563453e-06,
      "logits/chosen": -1.1568720249350068,
      "logits/rejected": -1.1932806025584015,
      "logps/chosen": -214.75,
      "logps/rejected": -264.25,
      "loss": 0.743255615234375,
      "mean_token_accuracy": 0.952626571059227,
      "num_tokens": 13333240.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.19916534423828125,
      "rewards/margins": 0.0906982421875,
      "rewards/rejected": -0.2904052734375,
      "step": 35
    },
    {
      "entropy": 0.168701171875,
      "epoch": 0.183206106870229,
      "grad_norm": 57.09201431274414,
      "learning_rate": 4.1116751269035535e-06,
      "logits/chosen": -1.244021900396991,
      "logits/rejected": -1.2053838438463438,
      "logps/chosen": -227.75,
      "logps/rejected": -235.75,
      "loss": 0.598724365234375,
      "mean_token_accuracy": 0.9517507553100586,
      "num_tokens": 13672135.0,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.099365234375,
      "rewards/margins": 0.307861328125,
      "rewards/rejected": -0.4072265625,
      "step": 36
    },
    {
      "entropy": 0.168212890625,
      "epoch": 0.18829516539440203,
      "grad_norm": 70.22441101074219,
      "learning_rate": 4.086294416243655e-06,
      "logits/chosen": -1.2293163289174895,
      "logits/rejected": -1.2229109217299612,
      "logps/chosen": -224.25,
      "logps/rejected": -227.75,
      "loss": 0.82135009765625,
      "mean_token_accuracy": 0.9494140148162842,
      "num_tokens": 14069079.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.0679931640625,
      "rewards/margins": -0.12286376953125,
      "rewards/rejected": 0.0546875,
      "step": 37
    },
    {
      "entropy": 0.16357421875,
      "epoch": 0.19338422391857507,
      "grad_norm": 76.26187896728516,
      "learning_rate": 4.060913705583757e-06,
      "logits/chosen": -1.210362411249237,
      "logits/rejected": -1.2507633848434636,
      "logps/chosen": -263.5,
      "logps/rejected": -245.75,
      "loss": 0.789825439453125,
      "mean_token_accuracy": 0.9464226067066193,
      "num_tokens": 14439718.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.00927734375,
      "rewards/margins": -0.00775146484375,
      "rewards/rejected": -0.001953125,
      "step": 38
    },
    {
      "entropy": 0.1767578125,
      "epoch": 0.1984732824427481,
      "grad_norm": 60.25480270385742,
      "learning_rate": 4.035532994923858e-06,
      "logits/chosen": -1.2676762543068074,
      "logits/rejected": -1.2517246495133931,
      "logps/chosen": -240.25,
      "logps/rejected": -250.0,
      "loss": 0.706390380859375,
      "mean_token_accuracy": 0.9461905360221863,
      "num_tokens": 14842983.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.234130859375,
      "rewards/margins": 0.1005859375,
      "rewards/rejected": -0.334716796875,
      "step": 39
    },
    {
      "entropy": 0.158447265625,
      "epoch": 0.2035623409669211,
      "grad_norm": 56.411930084228516,
      "learning_rate": 4.0101522842639595e-06,
      "logits/chosen": -1.253394155349274,
      "logits/rejected": -1.2763997522602422,
      "logps/chosen": -219.5,
      "logps/rejected": -248.0,
      "loss": 0.6590461730957031,
      "mean_token_accuracy": 0.9516128301620483,
      "num_tokens": 15207602.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.115966796875,
      "rewards/margins": 0.2529296875,
      "rewards/rejected": -0.3690185546875,
      "step": 40
    },
    {
      "entropy": 0.162109375,
      "epoch": 0.20865139949109415,
      "grad_norm": 50.47223663330078,
      "learning_rate": 3.984771573604061e-06,
      "logits/chosen": -1.221418931634342,
      "logits/rejected": -1.240179635098703,
      "logps/chosen": -198.75,
      "logps/rejected": -241.0,
      "loss": 0.61651611328125,
      "mean_token_accuracy": 0.9533202648162842,
      "num_tokens": 15615485.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": 0.01708984375,
      "rewards/margins": 0.32470703125,
      "rewards/rejected": -0.3079833984375,
      "step": 41
    },
    {
      "entropy": 0.164794921875,
      "epoch": 0.21374045801526717,
      "grad_norm": 63.18607711791992,
      "learning_rate": 3.959390862944163e-06,
      "logits/chosen": -1.2487682814246732,
      "logits/rejected": -1.2684586865119198,
      "logps/chosen": -230.5,
      "logps/rejected": -251.25,
      "loss": 0.73516845703125,
      "mean_token_accuracy": 0.9500314593315125,
      "num_tokens": 16006587.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.0408935546875,
      "rewards/margins": 0.07275390625,
      "rewards/rejected": -0.11383056640625,
      "step": 42
    },
    {
      "entropy": 0.16796875,
      "epoch": 0.21882951653944022,
      "grad_norm": 57.65352249145508,
      "learning_rate": 3.934010152284265e-06,
      "logits/chosen": -1.3110932592597404,
      "logits/rejected": -1.3089875033252198,
      "logps/chosen": -223.0,
      "logps/rejected": -271.25,
      "loss": 0.7037353515625,
      "mean_token_accuracy": 0.952717587351799,
      "num_tokens": 16376034.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.1783447265625,
      "rewards/margins": 0.10040283203125,
      "rewards/rejected": -0.2783203125,
      "step": 43
    },
    {
      "entropy": 0.162353515625,
      "epoch": 0.22391857506361323,
      "grad_norm": 51.37344741821289,
      "learning_rate": 3.9086294416243655e-06,
      "logits/chosen": -1.293287887477467,
      "logits/rejected": -1.2826172419774333,
      "logps/chosen": -169.0,
      "logps/rejected": -188.25,
      "loss": 0.68048095703125,
      "mean_token_accuracy": 0.9513245075941086,
      "num_tokens": 16729880.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.14501953125,
      "rewards/margins": 0.17425537109375,
      "rewards/rejected": -0.31884765625,
      "step": 44
    },
    {
      "entropy": 0.162109375,
      "epoch": 0.22900763358778625,
      "grad_norm": 60.11509704589844,
      "learning_rate": 3.883248730964467e-06,
      "logits/chosen": -1.2668976340142997,
      "logits/rejected": -1.2750247764479736,
      "logps/chosen": -227.0,
      "logps/rejected": -250.0,
      "loss": 0.6617431640625,
      "mean_token_accuracy": 0.9519101679325104,
      "num_tokens": 17112685.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.0929718017578125,
      "rewards/margins": 0.20703125,
      "rewards/rejected": -0.2998046875,
      "step": 45
    },
    {
      "entropy": 0.16552734375,
      "epoch": 0.2340966921119593,
      "grad_norm": 70.05858612060547,
      "learning_rate": 3.857868020304569e-06,
      "logits/chosen": -1.256115313351624,
      "logits/rejected": -1.292612336747009,
      "logps/chosen": -203.75,
      "logps/rejected": -204.25,
      "loss": 0.78472900390625,
      "mean_token_accuracy": 0.9469475299119949,
      "num_tokens": 17464688.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.1900482177734375,
      "rewards/margins": -0.072418212890625,
      "rewards/rejected": -0.118408203125,
      "step": 46
    },
    {
      "entropy": 0.16552734375,
      "epoch": 0.23918575063613232,
      "grad_norm": 63.72059631347656,
      "learning_rate": 3.832487309644671e-06,
      "logits/chosen": -1.2366188823977582,
      "logits/rejected": -1.2068790190894856,
      "logps/chosen": -215.75,
      "logps/rejected": -216.25,
      "loss": 0.70928955078125,
      "mean_token_accuracy": 0.947726234793663,
      "num_tokens": 17827406.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.126556396484375,
      "rewards/margins": 0.05908203125,
      "rewards/rejected": -0.18597412109375,
      "step": 47
    },
    {
      "entropy": 0.173095703125,
      "epoch": 0.24427480916030533,
      "grad_norm": 51.02957534790039,
      "learning_rate": 3.8071065989847715e-06,
      "logits/chosen": -1.2951891338286738,
      "logits/rejected": -1.2901734123706725,
      "logps/chosen": -186.375,
      "logps/rejected": -190.375,
      "loss": 0.686798095703125,
      "mean_token_accuracy": 0.9495810121297836,
      "num_tokens": 18168837.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.149169921875,
      "rewards/margins": 0.10107421875,
      "rewards/rejected": -0.25054931640625,
      "step": 48
    },
    {
      "entropy": 0.171142578125,
      "epoch": 0.24936386768447838,
      "grad_norm": 65.16993713378906,
      "learning_rate": 3.7817258883248736e-06,
      "logits/chosen": -1.2489111246586129,
      "logits/rejected": -1.2595230297930642,
      "logps/chosen": -226.25,
      "logps/rejected": -253.5,
      "loss": 0.6953582763671875,
      "mean_token_accuracy": 0.9487173557281494,
      "num_tokens": 18520757.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.05126953125,
      "rewards/margins": 0.1453857421875,
      "rewards/rejected": -0.197357177734375,
      "step": 49
    },
    {
      "entropy": 0.161865234375,
      "epoch": 0.2544529262086514,
      "grad_norm": 85.86858367919922,
      "learning_rate": 3.756345177664975e-06,
      "logits/chosen": -1.2788417924140663,
      "logits/rejected": -1.2818688140867107,
      "logps/chosen": -222.25,
      "logps/rejected": -228.0,
      "loss": 0.906097412109375,
      "mean_token_accuracy": 0.9481759369373322,
      "num_tokens": 18907465.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.32647705078125,
      "rewards/margins": -0.261962890625,
      "rewards/rejected": -0.06494140625,
      "step": 50
    },
    {
      "entropy": 0.1689453125,
      "epoch": 0.2595419847328244,
      "grad_norm": 63.7056999206543,
      "learning_rate": 3.7309644670050766e-06,
      "logits/chosen": -1.24593670349563,
      "logits/rejected": -1.2525011882667132,
      "logps/chosen": -204.5,
      "logps/rejected": -235.0,
      "loss": 0.809600830078125,
      "mean_token_accuracy": 0.9492675065994263,
      "num_tokens": 19271716.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.22967529296875,
      "rewards/margins": -0.108154296875,
      "rewards/rejected": -0.12200927734375,
      "step": 51
    },
    {
      "entropy": 0.16943359375,
      "epoch": 0.26463104325699743,
      "grad_norm": 66.75394439697266,
      "learning_rate": 3.705583756345178e-06,
      "logits/chosen": -1.2951451816324733,
      "logits/rejected": -1.2748576702987078,
      "logps/chosen": -220.0,
      "logps/rejected": -201.5,
      "loss": 0.7838897705078125,
      "mean_token_accuracy": 0.9473998248577118,
      "num_tokens": 19651670.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.283111572265625,
      "rewards/margins": -0.046630859375,
      "rewards/rejected": -0.235687255859375,
      "step": 52
    },
    {
      "entropy": 0.167724609375,
      "epoch": 0.2697201017811705,
      "grad_norm": 68.03031921386719,
      "learning_rate": 3.6802030456852796e-06,
      "logits/chosen": -1.2656714275142638,
      "logits/rejected": -1.2242606863621242,
      "logps/chosen": -184.375,
      "logps/rejected": -200.75,
      "loss": 0.81591796875,
      "mean_token_accuracy": 0.9486018866300583,
      "num_tokens": 20048489.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.284912109375,
      "rewards/margins": -0.12548828125,
      "rewards/rejected": -0.158782958984375,
      "step": 53
    },
    {
      "entropy": 0.164306640625,
      "epoch": 0.2748091603053435,
      "grad_norm": 52.74068832397461,
      "learning_rate": 3.654822335025381e-06,
      "logits/chosen": -1.3013878893228477,
      "logits/rejected": -1.3064163255128232,
      "logps/chosen": -192.5,
      "logps/rejected": -185.5,
      "loss": 0.67987060546875,
      "mean_token_accuracy": 0.9475405365228653,
      "num_tokens": 20384513.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.19287109375,
      "rewards/margins": 0.16943359375,
      "rewards/rejected": -0.3623046875,
      "step": 54
    },
    {
      "entropy": 0.162109375,
      "epoch": 0.27989821882951654,
      "grad_norm": 71.49076843261719,
      "learning_rate": 3.629441624365482e-06,
      "logits/chosen": -1.2508704209061283,
      "logits/rejected": -1.2312254570289394,
      "logps/chosen": -225.0,
      "logps/rejected": -227.5,
      "loss": 0.86474609375,
      "mean_token_accuracy": 0.9473535567522049,
      "num_tokens": 20754117.0,
      "rewards/accuracies": 0.25,
      "rewards/chosen": -0.34716796875,
      "rewards/margins": -0.2258148193359375,
      "rewards/rejected": -0.1211395263671875,
      "step": 55
    },
    {
      "entropy": 0.1640625,
      "epoch": 0.28498727735368956,
      "grad_norm": 70.81625366210938,
      "learning_rate": 3.6040609137055843e-06,
      "logits/chosen": -1.3251861075583002,
      "logits/rejected": -1.32016448634709,
      "logps/chosen": -237.75,
      "logps/rejected": -248.0,
      "loss": 0.78326416015625,
      "mean_token_accuracy": 0.948715329170227,
      "num_tokens": 21157270.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.26702880859375,
      "rewards/margins": -0.0032958984375,
      "rewards/rejected": -0.26409912109375,
      "step": 56
    },
    {
      "entropy": 0.16162109375,
      "epoch": 0.2900763358778626,
      "grad_norm": 57.185699462890625,
      "learning_rate": 3.5786802030456856e-06,
      "logits/chosen": -1.3391477581159092,
      "logits/rejected": -1.2969421235061,
      "logps/chosen": -167.5,
      "logps/rejected": -206.75,
      "loss": 0.7360153198242188,
      "mean_token_accuracy": 0.9557931572198868,
      "num_tokens": 21557901.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.13916015625,
      "rewards/margins": 0.08935546875,
      "rewards/rejected": -0.22784423828125,
      "step": 57
    },
    {
      "entropy": 0.167236328125,
      "epoch": 0.2951653944020356,
      "grad_norm": 67.23629760742188,
      "learning_rate": 3.5532994923857873e-06,
      "logits/chosen": -1.2344125058729924,
      "logits/rejected": -1.2176261940150734,
      "logps/chosen": -221.625,
      "logps/rejected": -227.5,
      "loss": 0.833160400390625,
      "mean_token_accuracy": 0.9486464709043503,
      "num_tokens": 21930023.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.220458984375,
      "rewards/margins": -0.0969696044921875,
      "rewards/rejected": -0.1231689453125,
      "step": 58
    },
    {
      "entropy": 0.158447265625,
      "epoch": 0.30025445292620867,
      "grad_norm": 61.011383056640625,
      "learning_rate": 3.5279187817258886e-06,
      "logits/chosen": -1.2608206675107227,
      "logits/rejected": -1.2702774965163959,
      "logps/chosen": -230.5,
      "logps/rejected": -221.0,
      "loss": 0.693756103515625,
      "mean_token_accuracy": 0.9499951750040054,
      "num_tokens": 22300544.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.32568359375,
      "rewards/margins": 0.140350341796875,
      "rewards/rejected": -0.466796875,
      "step": 59
    },
    {
      "entropy": 0.155517578125,
      "epoch": 0.3053435114503817,
      "grad_norm": 60.74739456176758,
      "learning_rate": 3.5025380710659903e-06,
      "logits/chosen": -1.2432721402471825,
      "logits/rejected": -1.1966404831484798,
      "logps/chosen": -188.25,
      "logps/rejected": -196.0,
      "loss": 0.8065185546875,
      "mean_token_accuracy": 0.9513755738735199,
      "num_tokens": 22699715.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.320556640625,
      "rewards/margins": -0.060028076171875,
      "rewards/rejected": -0.260986328125,
      "step": 60
    },
    {
      "entropy": 0.17333984375,
      "epoch": 0.3104325699745547,
      "grad_norm": 64.06476593017578,
      "learning_rate": 3.4771573604060916e-06,
      "logits/chosen": -1.23914246177565,
      "logits/rejected": -1.2589912793052047,
      "logps/chosen": -208.25,
      "logps/rejected": -225.25,
      "loss": 0.77069091796875,
      "mean_token_accuracy": 0.9473724365234375,
      "num_tokens": 23067391.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.082275390625,
      "rewards/margins": 0.08203125,
      "rewards/rejected": -0.166015625,
      "step": 61
    },
    {
      "entropy": 0.1689453125,
      "epoch": 0.3155216284987277,
      "grad_norm": 48.07145309448242,
      "learning_rate": 3.451776649746193e-06,
      "logits/chosen": -1.2898154412235743,
      "logits/rejected": -1.2884235645949569,
      "logps/chosen": -205.25,
      "logps/rejected": -217.125,
      "loss": 0.5599517822265625,
      "mean_token_accuracy": 0.947222501039505,
      "num_tokens": 23396708.0,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": 0.03094482421875,
      "rewards/margins": 0.41748046875,
      "rewards/rejected": -0.38671875,
      "step": 62
    },
    {
      "entropy": 0.1640625,
      "epoch": 0.32061068702290074,
      "grad_norm": 65.72078704833984,
      "learning_rate": 3.4263959390862946e-06,
      "logits/chosen": -1.2525791988685302,
      "logits/rejected": -1.2465728013183064,
      "logps/chosen": -226.5,
      "logps/rejected": -198.75,
      "loss": 0.852203369140625,
      "mean_token_accuracy": 0.9457215964794159,
      "num_tokens": 23771644.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.45166015625,
      "rewards/margins": -0.1153564453125,
      "rewards/rejected": -0.335205078125,
      "step": 63
    },
    {
      "entropy": 0.16943359375,
      "epoch": 0.3256997455470738,
      "grad_norm": 70.23268127441406,
      "learning_rate": 3.4010152284263963e-06,
      "logits/chosen": -1.2732464634523972,
      "logits/rejected": -1.2514687027641094,
      "logps/chosen": -235.75,
      "logps/rejected": -247.25,
      "loss": 0.781402587890625,
      "mean_token_accuracy": 0.9476925879716873,
      "num_tokens": 24154169.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.297607421875,
      "rewards/margins": -0.0504150390625,
      "rewards/rejected": -0.246826171875,
      "step": 64
    },
    {
      "entropy": 0.1630859375,
      "epoch": 0.33078880407124683,
      "grad_norm": 75.50202941894531,
      "learning_rate": 3.375634517766498e-06,
      "logits/chosen": -1.2872747032412404,
      "logits/rejected": -1.2523587882327665,
      "logps/chosen": -243.0,
      "logps/rejected": -231.0,
      "loss": 0.81689453125,
      "mean_token_accuracy": 0.9463559240102768,
      "num_tokens": 24486129.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.22308349609375,
      "rewards/margins": -0.103271484375,
      "rewards/rejected": -0.11993408203125,
      "step": 65
    },
    {
      "entropy": 0.162109375,
      "epoch": 0.33587786259541985,
      "grad_norm": 65.13658142089844,
      "learning_rate": 3.3502538071065993e-06,
      "logits/chosen": -1.2898683758825709,
      "logits/rejected": -1.2849647133801598,
      "logps/chosen": -220.25,
      "logps/rejected": -229.5,
      "loss": 0.7967147827148438,
      "mean_token_accuracy": 0.9507217556238174,
      "num_tokens": 24857818.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.2047119140625,
      "rewards/margins": 0.05828857421875,
      "rewards/rejected": -0.263427734375,
      "step": 66
    },
    {
      "entropy": 0.165283203125,
      "epoch": 0.34096692111959287,
      "grad_norm": 61.869327545166016,
      "learning_rate": 3.3248730964467006e-06,
      "logits/chosen": -1.2517319440883818,
      "logits/rejected": -1.2434228301098385,
      "logps/chosen": -205.625,
      "logps/rejected": -222.25,
      "loss": 0.697723388671875,
      "mean_token_accuracy": 0.9499448090791702,
      "num_tokens": 25270495.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.1441650390625,
      "rewards/margins": 0.16796875,
      "rewards/rejected": -0.31201171875,
      "step": 67
    },
    {
      "entropy": 0.156005859375,
      "epoch": 0.3460559796437659,
      "grad_norm": 69.01359558105469,
      "learning_rate": 3.2994923857868023e-06,
      "logits/chosen": -1.2596454913995498,
      "logits/rejected": -1.2822667709063957,
      "logps/chosen": -238.75,
      "logps/rejected": -219.25,
      "loss": 0.788330078125,
      "mean_token_accuracy": 0.9476147145032883,
      "num_tokens": 25639922.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.2823486328125,
      "rewards/margins": -0.060791015625,
      "rewards/rejected": -0.2213134765625,
      "step": 68
    },
    {
      "entropy": 0.170166015625,
      "epoch": 0.3511450381679389,
      "grad_norm": 65.33216094970703,
      "learning_rate": 3.2741116751269036e-06,
      "logits/chosen": -1.2824587990174645,
      "logits/rejected": -1.2777971502557461,
      "logps/chosen": -199.25,
      "logps/rejected": -190.5,
      "loss": 0.833953857421875,
      "mean_token_accuracy": 0.945390522480011,
      "num_tokens": 26010115.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.238037109375,
      "rewards/margins": -0.1005859375,
      "rewards/rejected": -0.1373443603515625,
      "step": 69
    },
    {
      "entropy": 0.162841796875,
      "epoch": 0.356234096692112,
      "grad_norm": 64.25736999511719,
      "learning_rate": 3.2487309644670053e-06,
      "logits/chosen": -1.332777326624499,
      "logits/rejected": -1.3164871807406162,
      "logps/chosen": -175.625,
      "logps/rejected": -220.75,
      "loss": 0.727630615234375,
      "mean_token_accuracy": 0.950916975736618,
      "num_tokens": 26385610.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.1827850341796875,
      "rewards/margins": 0.056396484375,
      "rewards/rejected": -0.2392578125,
      "step": 70
    },
    {
      "entropy": 0.1611328125,
      "epoch": 0.361323155216285,
      "grad_norm": 57.03504943847656,
      "learning_rate": 3.223350253807107e-06,
      "logits/chosen": -1.2973189023839775,
      "logits/rejected": -1.2664155194693132,
      "logps/chosen": -195.75,
      "logps/rejected": -225.25,
      "loss": 0.6987152099609375,
      "mean_token_accuracy": 0.9523173123598099,
      "num_tokens": 26772574.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.239013671875,
      "rewards/margins": 0.1385498046875,
      "rewards/rejected": -0.37744140625,
      "step": 71
    },
    {
      "entropy": 0.151123046875,
      "epoch": 0.366412213740458,
      "grad_norm": 106.73540496826172,
      "learning_rate": 3.1979695431472087e-06,
      "logits/chosen": -1.3368664755840716,
      "logits/rejected": -1.3149258298106234,
      "logps/chosen": -207.75,
      "logps/rejected": -193.75,
      "loss": 0.76409912109375,
      "mean_token_accuracy": 0.9506115764379501,
      "num_tokens": 27153798.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.32080078125,
      "rewards/margins": -0.001861572265625,
      "rewards/rejected": -0.318359375,
      "step": 72
    },
    {
      "entropy": 0.162841796875,
      "epoch": 0.37150127226463103,
      "grad_norm": 43.361202239990234,
      "learning_rate": 3.17258883248731e-06,
      "logits/chosen": -1.292466410490634,
      "logits/rejected": -1.2795658979837796,
      "logps/chosen": -174.25,
      "logps/rejected": -169.0,
      "loss": 0.6114501953125,
      "mean_token_accuracy": 0.9446200430393219,
      "num_tokens": 27531852.0,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.11895751953125,
      "rewards/margins": 0.2860107421875,
      "rewards/rejected": -0.40478515625,
      "step": 73
    },
    {
      "entropy": 0.165283203125,
      "epoch": 0.37659033078880405,
      "grad_norm": 258.8431701660156,
      "learning_rate": 3.1472081218274113e-06,
      "logits/chosen": -1.255416233032602,
      "logits/rejected": -1.2341775737485738,
      "logps/chosen": -251.25,
      "logps/rejected": -240.5,
      "loss": 0.8010711669921875,
      "mean_token_accuracy": 0.9430906176567078,
      "num_tokens": 27903266.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.41162109375,
      "rewards/margins": 0.0003662109375,
      "rewards/rejected": -0.4114990234375,
      "step": 74
    },
    {
      "entropy": 0.1513671875,
      "epoch": 0.3816793893129771,
      "grad_norm": 59.12688064575195,
      "learning_rate": 3.121827411167513e-06,
      "logits/chosen": -1.3708437654075616,
      "logits/rejected": -1.3749631542345564,
      "logps/chosen": -184.25,
      "logps/rejected": -207.75,
      "loss": 0.7006683349609375,
      "mean_token_accuracy": 0.9533696621656418,
      "num_tokens": 28291715.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.008056640625,
      "rewards/margins": 0.16162109375,
      "rewards/rejected": -0.169677734375,
      "step": 75
    },
    {
      "entropy": 0.168212890625,
      "epoch": 0.38676844783715014,
      "grad_norm": 70.00326538085938,
      "learning_rate": 3.0964467005076143e-06,
      "logits/chosen": -1.2278840581970798,
      "logits/rejected": -1.252243618402488,
      "logps/chosen": -241.25,
      "logps/rejected": -222.25,
      "loss": 0.7878265380859375,
      "mean_token_accuracy": 0.9443240016698837,
      "num_tokens": 28640895.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.22607421875,
      "rewards/margins": -0.0015869140625,
      "rewards/rejected": -0.2257080078125,
      "step": 76
    },
    {
      "entropy": 0.147216796875,
      "epoch": 0.39185750636132316,
      "grad_norm": 52.42808151245117,
      "learning_rate": 3.071065989847716e-06,
      "logits/chosen": -1.3088794461049085,
      "logits/rejected": -1.2848204770716634,
      "logps/chosen": -174.375,
      "logps/rejected": -186.75,
      "loss": 0.730010986328125,
      "mean_token_accuracy": 0.9527680724859238,
      "num_tokens": 29031104.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.24163818359375,
      "rewards/margins": 0.07861328125,
      "rewards/rejected": -0.31982421875,
      "step": 77
    },
    {
      "entropy": 0.162109375,
      "epoch": 0.3969465648854962,
      "grad_norm": 66.45071411132812,
      "learning_rate": 3.0456852791878177e-06,
      "logits/chosen": -1.2249955483759087,
      "logits/rejected": -1.2137169350336674,
      "logps/chosen": -216.0,
      "logps/rejected": -232.0,
      "loss": 0.7823486328125,
      "mean_token_accuracy": 0.9497682005167007,
      "num_tokens": 29411929.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.2440185546875,
      "rewards/margins": -0.053955078125,
      "rewards/rejected": -0.1904296875,
      "step": 78
    },
    {
      "entropy": 0.169189453125,
      "epoch": 0.4020356234096692,
      "grad_norm": 71.31132507324219,
      "learning_rate": 3.0203045685279194e-06,
      "logits/chosen": -1.311240796195369,
      "logits/rejected": -1.3000907150950454,
      "logps/chosen": -202.0,
      "logps/rejected": -218.75,
      "loss": 0.727569580078125,
      "mean_token_accuracy": 0.9468916058540344,
      "num_tokens": 29765589.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.289306640625,
      "rewards/margins": 0.17138671875,
      "rewards/rejected": -0.460205078125,
      "step": 79
    },
    {
      "entropy": 0.162353515625,
      "epoch": 0.4071246819338422,
      "grad_norm": 64.0061264038086,
      "learning_rate": 2.9949238578680207e-06,
      "logits/chosen": -1.3101197475281547,
      "logits/rejected": -1.2970568708741963,
      "logps/chosen": -220.25,
      "logps/rejected": -211.5,
      "loss": 0.729522705078125,
      "mean_token_accuracy": 0.9487328827381134,
      "num_tokens": 30163265.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.21240234375,
      "rewards/margins": 0.0606689453125,
      "rewards/rejected": -0.273193359375,
      "step": 80
    },
    {
      "entropy": 0.16357421875,
      "epoch": 0.4122137404580153,
      "grad_norm": 65.17308807373047,
      "learning_rate": 2.969543147208122e-06,
      "logits/chosen": -1.3028268301410426,
      "logits/rejected": -1.3046930569719115,
      "logps/chosen": -221.0,
      "logps/rejected": -223.25,
      "loss": 0.70806884765625,
      "mean_token_accuracy": 0.9474814534187317,
      "num_tokens": 30509121.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.2777099609375,
      "rewards/margins": 0.07781982421875,
      "rewards/rejected": -0.35546875,
      "step": 81
    },
    {
      "entropy": 0.165771484375,
      "epoch": 0.4173027989821883,
      "grad_norm": 50.35979080200195,
      "learning_rate": 2.9441624365482237e-06,
      "logits/chosen": -1.2836960568317857,
      "logits/rejected": -1.306755689640039,
      "logps/chosen": -194.25,
      "logps/rejected": -214.75,
      "loss": 0.6282501220703125,
      "mean_token_accuracy": 0.9473407566547394,
      "num_tokens": 30917508.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.16790771484375,
      "rewards/margins": 0.30712890625,
      "rewards/rejected": -0.47509765625,
      "step": 82
    },
    {
      "entropy": 0.156494140625,
      "epoch": 0.4223918575063613,
      "grad_norm": 55.4305419921875,
      "learning_rate": 2.918781725888325e-06,
      "logits/chosen": -1.3347602124239275,
      "logits/rejected": -1.297045195641747,
      "logps/chosen": -168.0,
      "logps/rejected": -182.5,
      "loss": 0.702239990234375,
      "mean_token_accuracy": 0.9523750394582748,
      "num_tokens": 31290108.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.17498779296875,
      "rewards/margins": 0.1236572265625,
      "rewards/rejected": -0.298828125,
      "step": 83
    },
    {
      "entropy": 0.1591796875,
      "epoch": 0.42748091603053434,
      "grad_norm": 72.86205291748047,
      "learning_rate": 2.8934010152284262e-06,
      "logits/chosen": -1.2560234762562408,
      "logits/rejected": -1.2293338403613354,
      "logps/chosen": -207.75,
      "logps/rejected": -219.0,
      "loss": 0.7640380859375,
      "mean_token_accuracy": 0.9502075463533401,
      "num_tokens": 31677755.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.6025390625,
      "rewards/margins": 0.078369140625,
      "rewards/rejected": -0.6806640625,
      "step": 84
    },
    {
      "entropy": 0.16015625,
      "epoch": 0.43256997455470736,
      "grad_norm": 60.04010009765625,
      "learning_rate": 2.8680203045685284e-06,
      "logits/chosen": -1.3011174292040961,
      "logits/rejected": -1.2539586523088446,
      "logps/chosen": -210.5,
      "logps/rejected": -211.25,
      "loss": 0.669281005859375,
      "mean_token_accuracy": 0.9515846818685532,
      "num_tokens": 32041342.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.1500244140625,
      "rewards/margins": 0.138671875,
      "rewards/rejected": -0.28857421875,
      "step": 85
    },
    {
      "entropy": 0.1572265625,
      "epoch": 0.43765903307888043,
      "grad_norm": 65.90755462646484,
      "learning_rate": 2.8426395939086297e-06,
      "logits/chosen": -1.2890849091113363,
      "logits/rejected": -1.2671961630083879,
      "logps/chosen": -212.0,
      "logps/rejected": -212.0,
      "loss": 0.8037109375,
      "mean_token_accuracy": 0.9498841613531113,
      "num_tokens": 32426894.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.328125,
      "rewards/margins": -0.07586669921875,
      "rewards/rejected": -0.2518310546875,
      "step": 86
    },
    {
      "entropy": 0.160888671875,
      "epoch": 0.44274809160305345,
      "grad_norm": 64.37542724609375,
      "learning_rate": 2.8172588832487314e-06,
      "logits/chosen": -1.2516031645716161,
      "logits/rejected": -1.2446139578943387,
      "logps/chosen": -238.5,
      "logps/rejected": -244.5,
      "loss": 0.734710693359375,
      "mean_token_accuracy": 0.9513941407203674,
      "num_tokens": 32747265.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.337158203125,
      "rewards/margins": 0.085693359375,
      "rewards/rejected": -0.42431640625,
      "step": 87
    },
    {
      "entropy": 0.164306640625,
      "epoch": 0.44783715012722647,
      "grad_norm": 64.61540222167969,
      "learning_rate": 2.7918781725888327e-06,
      "logits/chosen": -1.3129108468652448,
      "logits/rejected": -1.3204735199009792,
      "logps/chosen": -208.5,
      "logps/rejected": -216.5,
      "loss": 0.674041748046875,
      "mean_token_accuracy": 0.9465741366147995,
      "num_tokens": 33124296.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.34832763671875,
      "rewards/margins": 0.1859130859375,
      "rewards/rejected": -0.533203125,
      "step": 88
    },
    {
      "entropy": 0.166748046875,
      "epoch": 0.4529262086513995,
      "grad_norm": 62.742977142333984,
      "learning_rate": 2.7664974619289344e-06,
      "logits/chosen": -1.2541273019597161,
      "logits/rejected": -1.2661804942223065,
      "logps/chosen": -202.25,
      "logps/rejected": -204.0,
      "loss": 0.775146484375,
      "mean_token_accuracy": 0.9480140656232834,
      "num_tokens": 33498040.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.30126953125,
      "rewards/margins": -0.014892578125,
      "rewards/rejected": -0.28759765625,
      "step": 89
    },
    {
      "entropy": 0.1611328125,
      "epoch": 0.4580152671755725,
      "grad_norm": 70.147216796875,
      "learning_rate": 2.7411167512690357e-06,
      "logits/chosen": -1.2708255735449616,
      "logits/rejected": -1.2573322500480986,
      "logps/chosen": -226.0,
      "logps/rejected": -218.5,
      "loss": 0.7926025390625,
      "mean_token_accuracy": 0.9451356083154678,
      "num_tokens": 33845280.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.323974609375,
      "rewards/margins": -0.10205078125,
      "rewards/rejected": -0.2225341796875,
      "step": 90
    },
    {
      "entropy": 0.15625,
      "epoch": 0.4631043256997455,
      "grad_norm": 63.841976165771484,
      "learning_rate": 2.715736040609137e-06,
      "logits/chosen": -1.326140821141854,
      "logits/rejected": -1.3591119533293379,
      "logps/chosen": -200.25,
      "logps/rejected": -205.0,
      "loss": 0.790740966796875,
      "mean_token_accuracy": 0.9486561864614487,
      "num_tokens": 34242803.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.4453125,
      "rewards/margins": -0.0655517578125,
      "rewards/rejected": -0.380126953125,
      "step": 91
    },
    {
      "entropy": 0.16015625,
      "epoch": 0.4681933842239186,
      "grad_norm": 73.43412780761719,
      "learning_rate": 2.6903553299492387e-06,
      "logits/chosen": -1.252702561138765,
      "logits/rejected": -1.2816603315955593,
      "logps/chosen": -250.25,
      "logps/rejected": -238.75,
      "loss": 0.90521240234375,
      "mean_token_accuracy": 0.9443905353546143,
      "num_tokens": 34628251.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.48974609375,
      "rewards/margins": -0.2421875,
      "rewards/rejected": -0.2470703125,
      "step": 92
    },
    {
      "entropy": 0.171142578125,
      "epoch": 0.4732824427480916,
      "grad_norm": 63.70939254760742,
      "learning_rate": 2.6649746192893404e-06,
      "logits/chosen": -1.254758430715978,
      "logits/rejected": -1.2633837670267376,
      "logps/chosen": -201.5,
      "logps/rejected": -184.25,
      "loss": 0.7750244140625,
      "mean_token_accuracy": 0.9433987885713577,
      "num_tokens": 35007915.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.306884765625,
      "rewards/margins": -0.08544921875,
      "rewards/rejected": -0.220947265625,
      "step": 93
    },
    {
      "entropy": 0.162109375,
      "epoch": 0.47837150127226463,
      "grad_norm": 73.37699890136719,
      "learning_rate": 2.639593908629442e-06,
      "logits/chosen": -1.2285106824311274,
      "logits/rejected": -1.2533075679141386,
      "logps/chosen": -218.75,
      "logps/rejected": -242.25,
      "loss": 0.84307861328125,
      "mean_token_accuracy": 0.947221428155899,
      "num_tokens": 35387530.0,
      "rewards/accuracies": 0.28125,
      "rewards/chosen": -0.4287109375,
      "rewards/margins": -0.1712646484375,
      "rewards/rejected": -0.25738525390625,
      "step": 94
    },
    {
      "entropy": 0.16357421875,
      "epoch": 0.48346055979643765,
      "grad_norm": 51.54319381713867,
      "learning_rate": 2.6142131979695434e-06,
      "logits/chosen": -1.276040442817514,
      "logits/rejected": -1.2760191663203924,
      "logps/chosen": -208.5,
      "logps/rejected": -240.0,
      "loss": 0.6256561279296875,
      "mean_token_accuracy": 0.9505224972963333,
      "num_tokens": 35732544.0,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.03521728515625,
      "rewards/margins": 0.306884765625,
      "rewards/rejected": -0.3427734375,
      "step": 95
    },
    {
      "entropy": 0.165283203125,
      "epoch": 0.48854961832061067,
      "grad_norm": 64.21118927001953,
      "learning_rate": 2.588832487309645e-06,
      "logits/chosen": -1.2820478858157205,
      "logits/rejected": -1.2736327297402361,
      "logps/chosen": -211.5,
      "logps/rejected": -197.0,
      "loss": 0.8376922607421875,
      "mean_token_accuracy": 0.9446868598461151,
      "num_tokens": 36116587.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.236328125,
      "rewards/margins": -0.08966064453125,
      "rewards/rejected": -0.146087646484375,
      "step": 96
    },
    {
      "entropy": 0.15869140625,
      "epoch": 0.49363867684478374,
      "grad_norm": 60.89461135864258,
      "learning_rate": 2.5634517766497464e-06,
      "logits/chosen": -1.2462152995908475,
      "logits/rejected": -1.2528073331416407,
      "logps/chosen": -232.0,
      "logps/rejected": -247.25,
      "loss": 0.69921875,
      "mean_token_accuracy": 0.9487164616584778,
      "num_tokens": 36525364.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.1297607421875,
      "rewards/margins": 0.209716796875,
      "rewards/rejected": -0.3388671875,
      "step": 97
    },
    {
      "entropy": 0.162353515625,
      "epoch": 0.49872773536895676,
      "grad_norm": 58.411373138427734,
      "learning_rate": 2.5380710659898476e-06,
      "logits/chosen": -1.2857249020723578,
      "logits/rejected": -1.273766229120395,
      "logps/chosen": -194.0,
      "logps/rejected": -214.5,
      "loss": 0.718597412109375,
      "mean_token_accuracy": 0.949738010764122,
      "num_tokens": 36924408.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.32177734375,
      "rewards/margins": 0.137939453125,
      "rewards/rejected": -0.459716796875,
      "step": 98
    },
    {
      "entropy": 0.155029296875,
      "epoch": 0.5038167938931297,
      "grad_norm": 65.45690155029297,
      "learning_rate": 2.5126903553299493e-06,
      "logits/chosen": -1.3141924885583736,
      "logits/rejected": -1.3078939917330479,
      "logps/chosen": -222.25,
      "logps/rejected": -223.75,
      "loss": 0.7358856201171875,
      "mean_token_accuracy": 0.951283186674118,
      "num_tokens": 37320690.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.2308349609375,
      "rewards/margins": 0.076904296875,
      "rewards/rejected": -0.308990478515625,
      "step": 99
    },
    {
      "entropy": 0.175048828125,
      "epoch": 0.5089058524173028,
      "grad_norm": 64.2022705078125,
      "learning_rate": 2.487309644670051e-06,
      "logits/chosen": -1.2262440668616592,
      "logits/rejected": -1.23488919059254,
      "logps/chosen": -232.75,
      "logps/rejected": -212.375,
      "loss": 0.8157958984375,
      "mean_token_accuracy": 0.9397356808185577,
      "num_tokens": 37699769.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.3427734375,
      "rewards/margins": -0.0829620361328125,
      "rewards/rejected": -0.2600555419921875,
      "step": 100
    },
    {
      "entropy": 0.154541015625,
      "epoch": 0.5139949109414759,
      "grad_norm": 65.66177368164062,
      "learning_rate": 2.4619289340101523e-06,
      "logits/chosen": -1.318018712738355,
      "logits/rejected": -1.2828259768178074,
      "logps/chosen": -190.25,
      "logps/rejected": -202.5,
      "loss": 0.87933349609375,
      "mean_token_accuracy": 0.9535940736532211,
      "num_tokens": 38075154.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.4375,
      "rewards/margins": -0.1856842041015625,
      "rewards/rejected": -0.2518310546875,
      "step": 101
    },
    {
      "entropy": 0.16357421875,
      "epoch": 0.5190839694656488,
      "grad_norm": 63.887428283691406,
      "learning_rate": 2.436548223350254e-06,
      "logits/chosen": -1.2580108193399442,
      "logits/rejected": -1.2583606455369734,
      "logps/chosen": -215.25,
      "logps/rejected": -235.5,
      "loss": 0.760650634765625,
      "mean_token_accuracy": 0.9483280181884766,
      "num_tokens": 38449417.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.2845458984375,
      "rewards/margins": 0.07366943359375,
      "rewards/rejected": -0.3583984375,
      "step": 102
    },
    {
      "entropy": 0.161865234375,
      "epoch": 0.5241730279898219,
      "grad_norm": 73.05874633789062,
      "learning_rate": 2.4111675126903553e-06,
      "logits/chosen": -1.2512428178163009,
      "logits/rejected": -1.2793649019025604,
      "logps/chosen": -232.5,
      "logps/rejected": -254.5,
      "loss": 0.8672332763671875,
      "mean_token_accuracy": 0.9479686319828033,
      "num_tokens": 38849315.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.39599609375,
      "rewards/margins": -0.0771484375,
      "rewards/rejected": -0.317626953125,
      "step": 103
    },
    {
      "entropy": 0.150634765625,
      "epoch": 0.5292620865139949,
      "grad_norm": 59.51718521118164,
      "learning_rate": 2.385786802030457e-06,
      "logits/chosen": -1.3055115561137152,
      "logits/rejected": -1.2614773898545937,
      "logps/chosen": -214.0,
      "logps/rejected": -234.0,
      "loss": 0.663818359375,
      "mean_token_accuracy": 0.9556228965520859,
      "num_tokens": 39246450.0,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.277099609375,
      "rewards/margins": 0.1888427734375,
      "rewards/rejected": -0.465087890625,
      "step": 104
    },
    {
      "entropy": 0.159423828125,
      "epoch": 0.5343511450381679,
      "grad_norm": 47.1342658996582,
      "learning_rate": 2.3604060913705588e-06,
      "logits/chosen": -1.349871941358015,
      "logits/rejected": -1.3362314439717622,
      "logps/chosen": -193.5,
      "logps/rejected": -208.0,
      "loss": 0.577880859375,
      "mean_token_accuracy": 0.9497708082199097,
      "num_tokens": 39635764.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.0133056640625,
      "rewards/margins": 0.435546875,
      "rewards/rejected": -0.44970703125,
      "step": 105
    },
    {
      "entropy": 0.166748046875,
      "epoch": 0.539440203562341,
      "grad_norm": 58.260555267333984,
      "learning_rate": 2.33502538071066e-06,
      "logits/chosen": -1.3027896006955104,
      "logits/rejected": -1.2992721109006562,
      "logps/chosen": -216.75,
      "logps/rejected": -191.75,
      "loss": 0.73553466796875,
      "mean_token_accuracy": 0.9484568685293198,
      "num_tokens": 39969463.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.2978515625,
      "rewards/margins": 0.049072265625,
      "rewards/rejected": -0.347412109375,
      "step": 106
    },
    {
      "entropy": 0.16064453125,
      "epoch": 0.544529262086514,
      "grad_norm": 62.18130111694336,
      "learning_rate": 2.3096446700507618e-06,
      "logits/chosen": -1.2421911175554698,
      "logits/rejected": -1.3233783198097135,
      "logps/chosen": -200.5,
      "logps/rejected": -207.75,
      "loss": 0.804779052734375,
      "mean_token_accuracy": 0.946388453245163,
      "num_tokens": 40350675.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.3505859375,
      "rewards/margins": -0.074981689453125,
      "rewards/rejected": -0.275146484375,
      "step": 107
    },
    {
      "entropy": 0.159912109375,
      "epoch": 0.549618320610687,
      "grad_norm": 62.705894470214844,
      "learning_rate": 2.284263959390863e-06,
      "logits/chosen": -1.315684198825371,
      "logits/rejected": -1.3131661685802256,
      "logps/chosen": -188.5,
      "logps/rejected": -195.375,
      "loss": 0.77838134765625,
      "mean_token_accuracy": 0.9471890181303024,
      "num_tokens": 40724174.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.4033203125,
      "rewards/margins": -0.06640625,
      "rewards/rejected": -0.336669921875,
      "step": 108
    },
    {
      "entropy": 0.1669921875,
      "epoch": 0.55470737913486,
      "grad_norm": 66.83467102050781,
      "learning_rate": 2.2588832487309648e-06,
      "logits/chosen": -1.2883205988766404,
      "logits/rejected": -1.308848106252747,
      "logps/chosen": -258.0,
      "logps/rejected": -261.25,
      "loss": 0.7769012451171875,
      "mean_token_accuracy": 0.9474173188209534,
      "num_tokens": 41115013.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.186798095703125,
      "rewards/margins": 0.14091873168945312,
      "rewards/rejected": -0.327880859375,
      "step": 109
    },
    {
      "entropy": 0.163818359375,
      "epoch": 0.5597964376590331,
      "grad_norm": 67.90021514892578,
      "learning_rate": 2.233502538071066e-06,
      "logits/chosen": -1.2429459761599158,
      "logits/rejected": -1.237265815477489,
      "logps/chosen": -230.0,
      "logps/rejected": -218.5,
      "loss": 0.807220458984375,
      "mean_token_accuracy": 0.9451711028814316,
      "num_tokens": 41488642.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.345703125,
      "rewards/margins": -0.065887451171875,
      "rewards/rejected": -0.2799072265625,
      "step": 110
    },
    {
      "entropy": 0.15771484375,
      "epoch": 0.5648854961832062,
      "grad_norm": 69.03768920898438,
      "learning_rate": 2.2081218274111678e-06,
      "logits/chosen": -1.262367353042372,
      "logits/rejected": -1.2231653630458967,
      "logps/chosen": -223.25,
      "logps/rejected": -202.75,
      "loss": 0.890869140625,
      "mean_token_accuracy": 0.9496006816625595,
      "num_tokens": 41863681.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.42919921875,
      "rewards/margins": -0.21875,
      "rewards/rejected": -0.2099609375,
      "step": 111
    },
    {
      "entropy": 0.1640625,
      "epoch": 0.5699745547073791,
      "grad_norm": 65.44580841064453,
      "learning_rate": 2.182741116751269e-06,
      "logits/chosen": -1.230319660623082,
      "logits/rejected": -1.251510805200625,
      "logps/chosen": -253.25,
      "logps/rejected": -244.5,
      "loss": 0.7418975830078125,
      "mean_token_accuracy": 0.9484907984733582,
      "num_tokens": 42189294.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.26123046875,
      "rewards/margins": 0.162109375,
      "rewards/rejected": -0.42236328125,
      "step": 112
    },
    {
      "entropy": 0.166015625,
      "epoch": 0.5750636132315522,
      "grad_norm": 56.990074157714844,
      "learning_rate": 2.1573604060913707e-06,
      "logits/chosen": -1.228292510885018,
      "logits/rejected": -1.2380661145880048,
      "logps/chosen": -188.75,
      "logps/rejected": -210.25,
      "loss": 0.7041778564453125,
      "mean_token_accuracy": 0.9472708106040955,
      "num_tokens": 42566983.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.39013671875,
      "rewards/margins": 0.166259765625,
      "rewards/rejected": -0.5556640625,
      "step": 113
    },
    {
      "entropy": 0.1591796875,
      "epoch": 0.5801526717557252,
      "grad_norm": 73.63758850097656,
      "learning_rate": 2.1319796954314725e-06,
      "logits/chosen": -1.2304216080221377,
      "logits/rejected": -1.2626504818763358,
      "logps/chosen": -242.75,
      "logps/rejected": -261.25,
      "loss": 0.784637451171875,
      "mean_token_accuracy": 0.9489694982767105,
      "num_tokens": 42913983.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.3817138671875,
      "rewards/margins": 0.06597900390625,
      "rewards/rejected": -0.4482421875,
      "step": 114
    },
    {
      "entropy": 0.161865234375,
      "epoch": 0.5852417302798982,
      "grad_norm": 62.614219665527344,
      "learning_rate": 2.1065989847715737e-06,
      "logits/chosen": -1.3461430935037932,
      "logits/rejected": -1.3234544320772617,
      "logps/chosen": -185.5,
      "logps/rejected": -206.25,
      "loss": 0.7844467163085938,
      "mean_token_accuracy": 0.9493400007486343,
      "num_tokens": 43279452.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.310546875,
      "rewards/margins": 0.0008544921875,
      "rewards/rejected": -0.3109893798828125,
      "step": 115
    },
    {
      "entropy": 0.166748046875,
      "epoch": 0.5903307888040712,
      "grad_norm": 73.2928695678711,
      "learning_rate": 2.0812182741116755e-06,
      "logits/chosen": -1.2476711193063883,
      "logits/rejected": -1.229951835621045,
      "logps/chosen": -241.75,
      "logps/rejected": -224.25,
      "loss": 0.85125732421875,
      "mean_token_accuracy": 0.9468885958194733,
      "num_tokens": 43674027.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.483642578125,
      "rewards/margins": -0.1435546875,
      "rewards/rejected": -0.3389892578125,
      "step": 116
    },
    {
      "entropy": 0.166259765625,
      "epoch": 0.5954198473282443,
      "grad_norm": 66.14990234375,
      "learning_rate": 2.0558375634517767e-06,
      "logits/chosen": -1.1910507876796015,
      "logits/rejected": -1.1934789413274294,
      "logps/chosen": -231.25,
      "logps/rejected": -230.5,
      "loss": 0.78106689453125,
      "mean_token_accuracy": 0.950650081038475,
      "num_tokens": 44047350.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.2755126953125,
      "rewards/margins": 0.078369140625,
      "rewards/rejected": -0.354248046875,
      "step": 117
    },
    {
      "entropy": 0.15478515625,
      "epoch": 0.6005089058524173,
      "grad_norm": 52.3381233215332,
      "learning_rate": 2.0304568527918785e-06,
      "logits/chosen": -1.2695932504330432,
      "logits/rejected": -1.2736707511822316,
      "logps/chosen": -181.75,
      "logps/rejected": -203.25,
      "loss": 0.661346435546875,
      "mean_token_accuracy": 0.950680673122406,
      "num_tokens": 44459928.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.265625,
      "rewards/margins": 0.265625,
      "rewards/rejected": -0.53076171875,
      "step": 118
    },
    {
      "entropy": 0.159912109375,
      "epoch": 0.6055979643765903,
      "grad_norm": 49.49362564086914,
      "learning_rate": 2.0050761421319797e-06,
      "logits/chosen": -1.316048034713927,
      "logits/rejected": -1.3241544765179663,
      "logps/chosen": -168.0,
      "logps/rejected": -182.0,
      "loss": 0.6475830078125,
      "mean_token_accuracy": 0.9487863481044769,
      "num_tokens": 44836514.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.25537109375,
      "rewards/margins": 0.240570068359375,
      "rewards/rejected": -0.49462890625,
      "step": 119
    },
    {
      "entropy": 0.16357421875,
      "epoch": 0.6106870229007634,
      "grad_norm": 74.30496215820312,
      "learning_rate": 1.9796954314720814e-06,
      "logits/chosen": -1.2736612451556142,
      "logits/rejected": -1.2686709039286261,
      "logps/chosen": -208.25,
      "logps/rejected": -245.0,
      "loss": 0.751373291015625,
      "mean_token_accuracy": 0.9507658183574677,
      "num_tokens": 45174916.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.353515625,
      "rewards/margins": 0.05206298828125,
      "rewards/rejected": -0.4052734375,
      "step": 120
    },
    {
      "entropy": 0.1650390625,
      "epoch": 0.6157760814249363,
      "grad_norm": 64.3706283569336,
      "learning_rate": 1.9543147208121827e-06,
      "logits/chosen": -1.270622788271707,
      "logits/rejected": -1.2759695510336173,
      "logps/chosen": -203.0,
      "logps/rejected": -201.0,
      "loss": 0.756500244140625,
      "mean_token_accuracy": 0.9468925297260284,
      "num_tokens": 45574505.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.1610107421875,
      "rewards/margins": 0.05908203125,
      "rewards/rejected": -0.220703125,
      "step": 121
    },
    {
      "entropy": 0.169921875,
      "epoch": 0.6208651399491094,
      "grad_norm": 85.26793670654297,
      "learning_rate": 1.9289340101522844e-06,
      "logits/chosen": -1.2705119265476728,
      "logits/rejected": -1.2609238111160443,
      "logps/chosen": -246.25,
      "logps/rejected": -230.75,
      "loss": 0.944061279296875,
      "mean_token_accuracy": 0.9459521919488907,
      "num_tokens": 45911388.0,
      "rewards/accuracies": 0.25,
      "rewards/chosen": -0.48193359375,
      "rewards/margins": -0.3070068359375,
      "rewards/rejected": -0.1748046875,
      "step": 122
    },
    {
      "entropy": 0.16064453125,
      "epoch": 0.6259541984732825,
      "grad_norm": 63.35386657714844,
      "learning_rate": 1.9035532994923857e-06,
      "logits/chosen": -1.1881255673183586,
      "logits/rejected": -1.2287232149134593,
      "logps/chosen": -229.25,
      "logps/rejected": -229.25,
      "loss": 0.6989974975585938,
      "mean_token_accuracy": 0.949111670255661,
      "num_tokens": 46263528.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.358184814453125,
      "rewards/margins": 0.215087890625,
      "rewards/rejected": -0.57421875,
      "step": 123
    },
    {
      "entropy": 0.165771484375,
      "epoch": 0.6310432569974554,
      "grad_norm": 63.35451126098633,
      "learning_rate": 1.8781725888324874e-06,
      "logits/chosen": -1.2589330922736834,
      "logits/rejected": -1.2295728762675209,
      "logps/chosen": -230.0,
      "logps/rejected": -241.25,
      "loss": 0.821014404296875,
      "mean_token_accuracy": 0.9476511925458908,
      "num_tokens": 46617770.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.27587890625,
      "rewards/margins": 0.0260009765625,
      "rewards/rejected": -0.302093505859375,
      "step": 124
    },
    {
      "entropy": 0.15673828125,
      "epoch": 0.6361323155216285,
      "grad_norm": 64.3207778930664,
      "learning_rate": 1.852791878172589e-06,
      "logits/chosen": -1.2831963765459964,
      "logits/rejected": -1.2945043114039763,
      "logps/chosen": -208.0,
      "logps/rejected": -215.5,
      "loss": 0.807525634765625,
      "mean_token_accuracy": 0.9477943778038025,
      "num_tokens": 46994298.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.33453369140625,
      "rewards/margins": -0.073486328125,
      "rewards/rejected": -0.260833740234375,
      "step": 125
    },
    {
      "entropy": 0.1591796875,
      "epoch": 0.6412213740458015,
      "grad_norm": 55.1900749206543,
      "learning_rate": 1.8274111675126904e-06,
      "logits/chosen": -1.2910271166985685,
      "logits/rejected": -1.2747680953691005,
      "logps/chosen": -183.75,
      "logps/rejected": -211.0,
      "loss": 0.7557106018066406,
      "mean_token_accuracy": 0.9515034556388855,
      "num_tokens": 47389706.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.15460205078125,
      "rewards/margins": 0.17578125,
      "rewards/rejected": -0.3310546875,
      "step": 126
    },
    {
      "entropy": 0.1728515625,
      "epoch": 0.6463104325699746,
      "grad_norm": 56.20667266845703,
      "learning_rate": 1.8020304568527921e-06,
      "logits/chosen": -1.2981477149706437,
      "logits/rejected": -1.2675109223021184,
      "logps/chosen": -222.0,
      "logps/rejected": -232.5,
      "loss": 0.598602294921875,
      "mean_token_accuracy": 0.9497958421707153,
      "num_tokens": 47781197.0,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.27587890625,
      "rewards/margins": 0.317626953125,
      "rewards/rejected": -0.59326171875,
      "step": 127
    },
    {
      "entropy": 0.156005859375,
      "epoch": 0.6513994910941476,
      "grad_norm": 56.498069763183594,
      "learning_rate": 1.7766497461928936e-06,
      "logits/chosen": -1.2623762821907927,
      "logits/rejected": -1.274385917719584,
      "logps/chosen": -212.0,
      "logps/rejected": -233.75,
      "loss": 0.679718017578125,
      "mean_token_accuracy": 0.9534448385238647,
      "num_tokens": 48205592.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.3408203125,
      "rewards/margins": 0.1522216796875,
      "rewards/rejected": -0.493408203125,
      "step": 128
    },
    {
      "entropy": 0.1591796875,
      "epoch": 0.6564885496183206,
      "grad_norm": 64.94985961914062,
      "learning_rate": 1.7512690355329951e-06,
      "logits/chosen": -1.2347368781188934,
      "logits/rejected": -1.2598863054707647,
      "logps/chosen": -224.0,
      "logps/rejected": -214.25,
      "loss": 0.808380126953125,
      "mean_token_accuracy": 0.9472066313028336,
      "num_tokens": 48537876.0,
      "rewards/accuracies": 0.21875,
      "rewards/chosen": -0.46728515625,
      "rewards/margins": -0.0799560546875,
      "rewards/rejected": -0.3883056640625,
      "step": 129
    },
    {
      "entropy": 0.16357421875,
      "epoch": 0.6615776081424937,
      "grad_norm": 74.20825958251953,
      "learning_rate": 1.7258883248730964e-06,
      "logits/chosen": -1.2947300208859263,
      "logits/rejected": -1.2771773239258781,
      "logps/chosen": -243.75,
      "logps/rejected": -248.0,
      "loss": 0.887298583984375,
      "mean_token_accuracy": 0.9483547806739807,
      "num_tokens": 48920092.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.42333984375,
      "rewards/margins": -0.15228271484375,
      "rewards/rejected": -0.2706298828125,
      "step": 130
    },
    {
      "entropy": 0.169921875,
      "epoch": 0.6666666666666666,
      "grad_norm": 73.51078033447266,
      "learning_rate": 1.7005076142131981e-06,
      "logits/chosen": -1.2110933219118905,
      "logits/rejected": -1.2308418207144254,
      "logps/chosen": -235.75,
      "logps/rejected": -248.75,
      "loss": 0.8043212890625,
      "mean_token_accuracy": 0.9493052661418915,
      "num_tokens": 49305958.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.451171875,
      "rewards/margins": -0.077392578125,
      "rewards/rejected": -0.374267578125,
      "step": 131
    },
    {
      "entropy": 0.15869140625,
      "epoch": 0.6717557251908397,
      "grad_norm": 58.30695343017578,
      "learning_rate": 1.6751269035532996e-06,
      "logits/chosen": -1.3244597093243355,
      "logits/rejected": -1.3132443195709225,
      "logps/chosen": -206.0,
      "logps/rejected": -228.0,
      "loss": 0.668853759765625,
      "mean_token_accuracy": 0.9519393891096115,
      "num_tokens": 49693417.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.265869140625,
      "rewards/margins": 0.1748046875,
      "rewards/rejected": -0.43994140625,
      "step": 132
    },
    {
      "entropy": 0.160400390625,
      "epoch": 0.6768447837150128,
      "grad_norm": 53.58084487915039,
      "learning_rate": 1.6497461928934011e-06,
      "logits/chosen": -1.244864238209805,
      "logits/rejected": -1.2545958025840016,
      "logps/chosen": -183.625,
      "logps/rejected": -190.375,
      "loss": 0.700408935546875,
      "mean_token_accuracy": 0.9522906541824341,
      "num_tokens": 50059783.0,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.2659912109375,
      "rewards/margins": 0.1514892578125,
      "rewards/rejected": -0.4169921875,
      "step": 133
    },
    {
      "entropy": 0.164306640625,
      "epoch": 0.6819338422391857,
      "grad_norm": 68.5660171508789,
      "learning_rate": 1.6243654822335026e-06,
      "logits/chosen": -1.241397688917905,
      "logits/rejected": -1.278434530141986,
      "logps/chosen": -242.25,
      "logps/rejected": -234.75,
      "loss": 0.8605575561523438,
      "mean_token_accuracy": 0.9462583363056183,
      "num_tokens": 50413308.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.59716796875,
      "rewards/margins": -0.11083984375,
      "rewards/rejected": -0.48583984375,
      "step": 134
    },
    {
      "entropy": 0.164794921875,
      "epoch": 0.6870229007633588,
      "grad_norm": 64.64373016357422,
      "learning_rate": 1.5989847715736043e-06,
      "logits/chosen": -1.300048987756223,
      "logits/rejected": -1.3131778176200615,
      "logps/chosen": -192.75,
      "logps/rejected": -190.875,
      "loss": 0.850250244140625,
      "mean_token_accuracy": 0.9482599049806595,
      "num_tokens": 50756854.0,
      "rewards/accuracies": 0.28125,
      "rewards/chosen": -0.349609375,
      "rewards/margins": -0.200927734375,
      "rewards/rejected": -0.14849853515625,
      "step": 135
    },
    {
      "entropy": 0.1591796875,
      "epoch": 0.6921119592875318,
      "grad_norm": 60.40647506713867,
      "learning_rate": 1.5736040609137056e-06,
      "logits/chosen": -1.286467965109169,
      "logits/rejected": -1.2765258255496543,
      "logps/chosen": -209.25,
      "logps/rejected": -207.25,
      "loss": 0.673126220703125,
      "mean_token_accuracy": 0.951656237244606,
      "num_tokens": 51131502.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.367919921875,
      "rewards/margins": 0.16021728515625,
      "rewards/rejected": -0.52734375,
      "step": 136
    },
    {
      "entropy": 0.163330078125,
      "epoch": 0.6972010178117048,
      "grad_norm": 61.93556594848633,
      "learning_rate": 1.5482233502538071e-06,
      "logits/chosen": -1.3216607979661747,
      "logits/rejected": -1.331168170680353,
      "logps/chosen": -198.25,
      "logps/rejected": -213.5,
      "loss": 0.71380615234375,
      "mean_token_accuracy": 0.9489306807518005,
      "num_tokens": 51483225.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.2080078125,
      "rewards/margins": 0.1612548828125,
      "rewards/rejected": -0.36962890625,
      "step": 137
    },
    {
      "entropy": 0.159912109375,
      "epoch": 0.7022900763358778,
      "grad_norm": 63.112213134765625,
      "learning_rate": 1.5228426395939088e-06,
      "logits/chosen": -1.2680197798757142,
      "logits/rejected": -1.2732023051613852,
      "logps/chosen": -245.0,
      "logps/rejected": -265.25,
      "loss": 0.705902099609375,
      "mean_token_accuracy": 0.950348436832428,
      "num_tokens": 51849542.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.454833984375,
      "rewards/margins": 0.133056640625,
      "rewards/rejected": -0.58740234375,
      "step": 138
    },
    {
      "entropy": 0.162353515625,
      "epoch": 0.7073791348600509,
      "grad_norm": 65.93107604980469,
      "learning_rate": 1.4974619289340103e-06,
      "logits/chosen": -1.2531528664415568,
      "logits/rejected": -1.2751016177936147,
      "logps/chosen": -240.5,
      "logps/rejected": -234.75,
      "loss": 0.771026611328125,
      "mean_token_accuracy": 0.9475905448198318,
      "num_tokens": 52234352.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.31005859375,
      "rewards/margins": 0.0687255859375,
      "rewards/rejected": -0.378173828125,
      "step": 139
    },
    {
      "entropy": 0.16748046875,
      "epoch": 0.712468193384224,
      "grad_norm": 60.60008239746094,
      "learning_rate": 1.4720812182741118e-06,
      "logits/chosen": -1.284045350811356,
      "logits/rejected": -1.3076317590229236,
      "logps/chosen": -205.75,
      "logps/rejected": -207.0,
      "loss": 0.71966552734375,
      "mean_token_accuracy": 0.9487140476703644,
      "num_tokens": 52544764.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.2362060546875,
      "rewards/margins": 0.08050537109375,
      "rewards/rejected": -0.3160400390625,
      "step": 140
    },
    {
      "entropy": 0.159423828125,
      "epoch": 0.7175572519083969,
      "grad_norm": 61.68688201904297,
      "learning_rate": 1.4467005076142131e-06,
      "logits/chosen": -1.2626397077576366,
      "logits/rejected": -1.2523052221622069,
      "logps/chosen": -207.5,
      "logps/rejected": -230.5,
      "loss": 0.716705322265625,
      "mean_token_accuracy": 0.9503261744976044,
      "num_tokens": 52948022.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.290771484375,
      "rewards/margins": 0.0987548828125,
      "rewards/rejected": -0.3899383544921875,
      "step": 141
    },
    {
      "entropy": 0.161865234375,
      "epoch": 0.72264631043257,
      "grad_norm": 63.84259796142578,
      "learning_rate": 1.4213197969543148e-06,
      "logits/chosen": -1.3211154042459343,
      "logits/rejected": -1.324081790319996,
      "logps/chosen": -241.5,
      "logps/rejected": -234.25,
      "loss": 0.76654052734375,
      "mean_token_accuracy": 0.949401780962944,
      "num_tokens": 53288797.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.40576171875,
      "rewards/margins": 0.03076171875,
      "rewards/rejected": -0.4375,
      "step": 142
    },
    {
      "entropy": 0.16064453125,
      "epoch": 0.727735368956743,
      "grad_norm": 59.44301223754883,
      "learning_rate": 1.3959390862944163e-06,
      "logits/chosen": -1.2445779844461344,
      "logits/rejected": -1.2510412319436552,
      "logps/chosen": -179.125,
      "logps/rejected": -184.5,
      "loss": 0.8189239501953125,
      "mean_token_accuracy": 0.9503347277641296,
      "num_tokens": 53650401.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.379638671875,
      "rewards/margins": -0.0517578125,
      "rewards/rejected": -0.327880859375,
      "step": 143
    },
    {
      "entropy": 0.160400390625,
      "epoch": 0.732824427480916,
      "grad_norm": 57.853416442871094,
      "learning_rate": 1.3705583756345178e-06,
      "logits/chosen": -1.2394903485408175,
      "logits/rejected": -1.261283751051131,
      "logps/chosen": -214.0,
      "logps/rejected": -211.75,
      "loss": 0.7115325927734375,
      "mean_token_accuracy": 0.9477828443050385,
      "num_tokens": 54054341.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.356689453125,
      "rewards/margins": 0.125,
      "rewards/rejected": -0.48193359375,
      "step": 144
    },
    {
      "entropy": 0.16455078125,
      "epoch": 0.7379134860050891,
      "grad_norm": 65.62645721435547,
      "learning_rate": 1.3451776649746193e-06,
      "logits/chosen": -1.2286020505571353,
      "logits/rejected": -1.208450173191077,
      "logps/chosen": -223.0,
      "logps/rejected": -245.5,
      "loss": 0.75048828125,
      "mean_token_accuracy": 0.9515127539634705,
      "num_tokens": 54427493.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.38330078125,
      "rewards/margins": -0.000244140625,
      "rewards/rejected": -0.38189697265625,
      "step": 145
    },
    {
      "entropy": 0.156494140625,
      "epoch": 0.7430025445292621,
      "grad_norm": 68.8692398071289,
      "learning_rate": 1.319796954314721e-06,
      "logits/chosen": -1.2827174915867314,
      "logits/rejected": -1.290493482770122,
      "logps/chosen": -218.25,
      "logps/rejected": -206.75,
      "loss": 0.8368377685546875,
      "mean_token_accuracy": 0.9481632858514786,
      "num_tokens": 54791174.0,
      "rewards/accuracies": 0.34375,
      "rewards/chosen": -0.33935546875,
      "rewards/margins": -0.12493896484375,
      "rewards/rejected": -0.2144775390625,
      "step": 146
    },
    {
      "entropy": 0.164306640625,
      "epoch": 0.7480916030534351,
      "grad_norm": 55.04671859741211,
      "learning_rate": 1.2944162436548225e-06,
      "logits/chosen": -1.2610715380959399,
      "logits/rejected": -1.2777534346884782,
      "logps/chosen": -166.5,
      "logps/rejected": -155.25,
      "loss": 0.7591552734375,
      "mean_token_accuracy": 0.9474459439516068,
      "num_tokens": 55190542.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.2154541015625,
      "rewards/margins": -0.035186767578125,
      "rewards/rejected": -0.179931640625,
      "step": 147
    },
    {
      "entropy": 0.15673828125,
      "epoch": 0.7531806615776081,
      "grad_norm": 64.17411804199219,
      "learning_rate": 1.2690355329949238e-06,
      "logits/chosen": -1.2668030115539317,
      "logits/rejected": -1.2703246015537304,
      "logps/chosen": -179.25,
      "logps/rejected": -225.0,
      "loss": 0.754150390625,
      "mean_token_accuracy": 0.9535480886697769,
      "num_tokens": 55568891.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.3115234375,
      "rewards/margins": 0.02471923828125,
      "rewards/rejected": -0.336669921875,
      "step": 148
    },
    {
      "entropy": 0.167236328125,
      "epoch": 0.7582697201017812,
      "grad_norm": 77.12625885009766,
      "learning_rate": 1.2436548223350255e-06,
      "logits/chosen": -1.246196790872148,
      "logits/rejected": -1.2578459995620601,
      "logps/chosen": -221.25,
      "logps/rejected": -218.25,
      "loss": 0.7427825927734375,
      "mean_token_accuracy": 0.9458917081356049,
      "num_tokens": 55926373.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.1793212890625,
      "rewards/margins": 0.071319580078125,
      "rewards/rejected": -0.250244140625,
      "step": 149
    },
    {
      "entropy": 0.159912109375,
      "epoch": 0.7633587786259542,
      "grad_norm": 60.0277214050293,
      "learning_rate": 1.218274111675127e-06,
      "logits/chosen": -1.2921580835065267,
      "logits/rejected": -1.2859186866957593,
      "logps/chosen": -219.75,
      "logps/rejected": -219.75,
      "loss": 0.680755615234375,
      "mean_token_accuracy": 0.9510093927383423,
      "num_tokens": 56285700.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.2650146484375,
      "rewards/margins": 0.1168212890625,
      "rewards/rejected": -0.382568359375,
      "step": 150
    },
    {
      "entropy": 0.167724609375,
      "epoch": 0.7684478371501272,
      "grad_norm": 63.86808395385742,
      "learning_rate": 1.1928934010152285e-06,
      "logits/chosen": -1.3243436768126475,
      "logits/rejected": -1.289354653269572,
      "logps/chosen": -237.0,
      "logps/rejected": -260.25,
      "loss": 0.7232666015625,
      "mean_token_accuracy": 0.9495467245578766,
      "num_tokens": 56701491.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.33740234375,
      "rewards/margins": 0.0955810546875,
      "rewards/rejected": -0.4326171875,
      "step": 151
    },
    {
      "entropy": 0.163818359375,
      "epoch": 0.7735368956743003,
      "grad_norm": 74.54528045654297,
      "learning_rate": 1.16751269035533e-06,
      "logits/chosen": -1.2510128515313985,
      "logits/rejected": -1.2449032379965963,
      "logps/chosen": -252.75,
      "logps/rejected": -269.5,
      "loss": 0.78497314453125,
      "mean_token_accuracy": 0.9484715014696121,
      "num_tokens": 57047974.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.259765625,
      "rewards/margins": -0.009765625,
      "rewards/rejected": -0.249755859375,
      "step": 152
    },
    {
      "entropy": 0.17041015625,
      "epoch": 0.7786259541984732,
      "grad_norm": 66.40087127685547,
      "learning_rate": 1.1421319796954315e-06,
      "logits/chosen": -1.2434997907611427,
      "logits/rejected": -1.2419221284680912,
      "logps/chosen": -233.5,
      "logps/rejected": -222.25,
      "loss": 0.744964599609375,
      "mean_token_accuracy": 0.9437924772500992,
      "num_tokens": 57392596.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.3751220703125,
      "rewards/margins": 0.0340576171875,
      "rewards/rejected": -0.40966796875,
      "step": 153
    },
    {
      "entropy": 0.164306640625,
      "epoch": 0.7837150127226463,
      "grad_norm": 53.339717864990234,
      "learning_rate": 1.116751269035533e-06,
      "logits/chosen": -1.235711641356627,
      "logits/rejected": -1.2588020280263763,
      "logps/chosen": -199.5,
      "logps/rejected": -228.25,
      "loss": 0.61297607421875,
      "mean_token_accuracy": 0.9511121064424515,
      "num_tokens": 57805475.0,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.190673828125,
      "rewards/margins": 0.30908203125,
      "rewards/rejected": -0.4990234375,
      "step": 154
    },
    {
      "entropy": 0.155517578125,
      "epoch": 0.7888040712468194,
      "grad_norm": 58.77641677856445,
      "learning_rate": 1.0913705583756345e-06,
      "logits/chosen": -1.2875096398514654,
      "logits/rejected": -1.2338864169086345,
      "logps/chosen": -192.25,
      "logps/rejected": -199.75,
      "loss": 0.72454833984375,
      "mean_token_accuracy": 0.9548005163669586,
      "num_tokens": 58174003.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.3203125,
      "rewards/margins": 0.0784912109375,
      "rewards/rejected": -0.3978271484375,
      "step": 155
    },
    {
      "entropy": 0.159912109375,
      "epoch": 0.7938931297709924,
      "grad_norm": 57.71940231323242,
      "learning_rate": 1.0659898477157362e-06,
      "logits/chosen": -1.2822042524364599,
      "logits/rejected": -1.302401272896131,
      "logps/chosen": -230.25,
      "logps/rejected": -229.75,
      "loss": 0.7072296142578125,
      "mean_token_accuracy": 0.9481226801872253,
      "num_tokens": 58513712.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.398193359375,
      "rewards/margins": 0.172119140625,
      "rewards/rejected": -0.5712890625,
      "step": 156
    },
    {
      "entropy": 0.167724609375,
      "epoch": 0.7989821882951654,
      "grad_norm": 58.8786506652832,
      "learning_rate": 1.0406091370558377e-06,
      "logits/chosen": -1.292047990535198,
      "logits/rejected": -1.3054225527754584,
      "logps/chosen": -210.0,
      "logps/rejected": -242.25,
      "loss": 0.66375732421875,
      "mean_token_accuracy": 0.9478696882724762,
      "num_tokens": 58896680.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.280303955078125,
      "rewards/margins": 0.2080078125,
      "rewards/rejected": -0.48974609375,
      "step": 157
    },
    {
      "entropy": 0.162841796875,
      "epoch": 0.8040712468193384,
      "grad_norm": 59.46760940551758,
      "learning_rate": 1.0152284263959392e-06,
      "logits/chosen": -1.24424356589235,
      "logits/rejected": -1.238614479620221,
      "logps/chosen": -221.75,
      "logps/rejected": -230.0,
      "loss": 0.7036838531494141,
      "mean_token_accuracy": 0.9477695375680923,
      "num_tokens": 59263462.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.1445770263671875,
      "rewards/margins": 0.207275390625,
      "rewards/rejected": -0.3521728515625,
      "step": 158
    },
    {
      "entropy": 0.162841796875,
      "epoch": 0.8091603053435115,
      "grad_norm": 65.75154113769531,
      "learning_rate": 9.898477157360407e-07,
      "logits/chosen": -1.2624473891183055,
      "logits/rejected": -1.259654311209261,
      "logps/chosen": -228.25,
      "logps/rejected": -212.75,
      "loss": 0.778594970703125,
      "mean_token_accuracy": 0.944281741976738,
      "num_tokens": 59633558.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.264617919921875,
      "rewards/margins": -0.0078125,
      "rewards/rejected": -0.2572021484375,
      "step": 159
    },
    {
      "entropy": 0.162841796875,
      "epoch": 0.8142493638676844,
      "grad_norm": 66.01615142822266,
      "learning_rate": 9.644670050761422e-07,
      "logits/chosen": -1.2357071070169634,
      "logits/rejected": -1.2384324784738467,
      "logps/chosen": -256.25,
      "logps/rejected": -248.5,
      "loss": 0.7286834716796875,
      "mean_token_accuracy": 0.9499144405126572,
      "num_tokens": 60016025.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.320068359375,
      "rewards/margins": 0.130126953125,
      "rewards/rejected": -0.449462890625,
      "step": 160
    },
    {
      "entropy": 0.152587890625,
      "epoch": 0.8193384223918575,
      "grad_norm": 78.33038330078125,
      "learning_rate": 9.390862944162437e-07,
      "logits/chosen": -1.3029380474205265,
      "logits/rejected": -1.3006044617540562,
      "logps/chosen": -200.75,
      "logps/rejected": -223.75,
      "loss": 0.7779769897460938,
      "mean_token_accuracy": 0.9547528177499771,
      "num_tokens": 60444729.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.326904296875,
      "rewards/margins": 0.07421875,
      "rewards/rejected": -0.4010009765625,
      "step": 161
    },
    {
      "entropy": 0.161865234375,
      "epoch": 0.8244274809160306,
      "grad_norm": 63.31332778930664,
      "learning_rate": 9.137055837563452e-07,
      "logits/chosen": -1.2962333428603996,
      "logits/rejected": -1.2906292243153006,
      "logps/chosen": -203.75,
      "logps/rejected": -227.5,
      "loss": 0.84423828125,
      "mean_token_accuracy": 0.9507206827402115,
      "num_tokens": 60806152.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.3492431640625,
      "rewards/margins": -0.097412109375,
      "rewards/rejected": -0.2523193359375,
      "step": 162
    },
    {
      "entropy": 0.157470703125,
      "epoch": 0.8295165394402035,
      "grad_norm": 53.934791564941406,
      "learning_rate": 8.883248730964468e-07,
      "logits/chosen": -1.2821091434094367,
      "logits/rejected": -1.2567944198943195,
      "logps/chosen": -182.5,
      "logps/rejected": -213.0,
      "loss": 0.723388671875,
      "mean_token_accuracy": 0.9540557265281677,
      "num_tokens": 61201391.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.25140380859375,
      "rewards/margins": 0.138916015625,
      "rewards/rejected": -0.39111328125,
      "step": 163
    },
    {
      "entropy": 0.16064453125,
      "epoch": 0.8346055979643766,
      "grad_norm": 66.25235748291016,
      "learning_rate": 8.629441624365482e-07,
      "logits/chosen": -1.2868481912577527,
      "logits/rejected": -1.3049520468495386,
      "logps/chosen": -199.25,
      "logps/rejected": -213.5,
      "loss": 0.8221435546875,
      "mean_token_accuracy": 0.9519907087087631,
      "num_tokens": 61572941.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.24542236328125,
      "rewards/margins": -0.0406494140625,
      "rewards/rejected": -0.2047119140625,
      "step": 164
    },
    {
      "entropy": 0.1591796875,
      "epoch": 0.8396946564885496,
      "grad_norm": 58.21211624145508,
      "learning_rate": 8.375634517766498e-07,
      "logits/chosen": -1.3208748146171105,
      "logits/rejected": -1.3522801461996157,
      "logps/chosen": -193.75,
      "logps/rejected": -242.5,
      "loss": 0.65692138671875,
      "mean_token_accuracy": 0.9525688141584396,
      "num_tokens": 61950200.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.421875,
      "rewards/margins": 0.1826171875,
      "rewards/rejected": -0.60498046875,
      "step": 165
    },
    {
      "entropy": 0.156982421875,
      "epoch": 0.8447837150127226,
      "grad_norm": 66.12063598632812,
      "learning_rate": 8.121827411167513e-07,
      "logits/chosen": -1.301899044199523,
      "logits/rejected": -1.2575513722113114,
      "logps/chosen": -221.25,
      "logps/rejected": -218.25,
      "loss": 0.794952392578125,
      "mean_token_accuracy": 0.9528547078371048,
      "num_tokens": 62317017.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.30322265625,
      "rewards/margins": -0.009185791015625,
      "rewards/rejected": -0.2940673828125,
      "step": 166
    },
    {
      "entropy": 0.1611328125,
      "epoch": 0.8498727735368957,
      "grad_norm": 66.28915405273438,
      "learning_rate": 7.868020304568528e-07,
      "logits/chosen": -1.2902138038428934,
      "logits/rejected": -1.280293306008564,
      "logps/chosen": -230.5,
      "logps/rejected": -249.25,
      "loss": 0.7397384643554688,
      "mean_token_accuracy": 0.9489684700965881,
      "num_tokens": 62671349.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.3994140625,
      "rewards/margins": 0.1123046875,
      "rewards/rejected": -0.5123291015625,
      "step": 167
    },
    {
      "entropy": 0.150634765625,
      "epoch": 0.8549618320610687,
      "grad_norm": 75.95428466796875,
      "learning_rate": 7.614213197969544e-07,
      "logits/chosen": -1.2779956306537894,
      "logits/rejected": -1.2907999663209873,
      "logps/chosen": -184.5,
      "logps/rejected": -218.0,
      "loss": 0.876434326171875,
      "mean_token_accuracy": 0.9518905580043793,
      "num_tokens": 63052689.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.338134765625,
      "rewards/margins": -0.1540069580078125,
      "rewards/rejected": -0.18463134765625,
      "step": 168
    },
    {
      "entropy": 0.1669921875,
      "epoch": 0.8600508905852418,
      "grad_norm": 63.13179016113281,
      "learning_rate": 7.360406091370559e-07,
      "logits/chosen": -1.2887472754751863,
      "logits/rejected": -1.2670022003966717,
      "logps/chosen": -214.75,
      "logps/rejected": -228.25,
      "loss": 0.78338623046875,
      "mean_token_accuracy": 0.9487744569778442,
      "num_tokens": 63405602.0,
      "rewards/accuracies": 0.28125,
      "rewards/chosen": -0.312744140625,
      "rewards/margins": -0.020751953125,
      "rewards/rejected": -0.2926025390625,
      "step": 169
    },
    {
      "entropy": 0.166748046875,
      "epoch": 0.8651399491094147,
      "grad_norm": 69.41549682617188,
      "learning_rate": 7.106598984771574e-07,
      "logits/chosen": -1.2627873517587107,
      "logits/rejected": -1.2781351035738362,
      "logps/chosen": -219.5,
      "logps/rejected": -236.25,
      "loss": 0.81103515625,
      "mean_token_accuracy": 0.9494354575872421,
      "num_tokens": 63792124.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.3935546875,
      "rewards/margins": -0.071044921875,
      "rewards/rejected": -0.322509765625,
      "step": 170
    },
    {
      "entropy": 0.165771484375,
      "epoch": 0.8702290076335878,
      "grad_norm": 59.073123931884766,
      "learning_rate": 6.852791878172589e-07,
      "logits/chosen": -1.291131682486795,
      "logits/rejected": -1.3237879333881308,
      "logps/chosen": -203.5,
      "logps/rejected": -223.5,
      "loss": 0.6531982421875,
      "mean_token_accuracy": 0.9505943953990936,
      "num_tokens": 64155911.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.1978759765625,
      "rewards/margins": 0.2900390625,
      "rewards/rejected": -0.48779296875,
      "step": 171
    },
    {
      "entropy": 0.162353515625,
      "epoch": 0.8753180661577609,
      "grad_norm": 59.49985885620117,
      "learning_rate": 6.598984771573605e-07,
      "logits/chosen": -1.2140906585743236,
      "logits/rejected": -1.1976295247955515,
      "logps/chosen": -200.0,
      "logps/rejected": -205.25,
      "loss": 0.706451416015625,
      "mean_token_accuracy": 0.9494357705116272,
      "num_tokens": 64532812.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.234375,
      "rewards/margins": 0.100830078125,
      "rewards/rejected": -0.33447265625,
      "step": 172
    },
    {
      "entropy": 0.152587890625,
      "epoch": 0.8804071246819338,
      "grad_norm": 52.89236831665039,
      "learning_rate": 6.345177664974619e-07,
      "logits/chosen": -1.3571618690737455,
      "logits/rejected": -1.3450652315884195,
      "logps/chosen": -174.625,
      "logps/rejected": -204.25,
      "loss": 0.59918212890625,
      "mean_token_accuracy": 0.9549228101968765,
      "num_tokens": 64943462.0,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.1142578125,
      "rewards/margins": 0.2861328125,
      "rewards/rejected": -0.400390625,
      "step": 173
    },
    {
      "entropy": 0.162353515625,
      "epoch": 0.8854961832061069,
      "grad_norm": 63.51266098022461,
      "learning_rate": 6.091370558375635e-07,
      "logits/chosen": -1.2747322101681697,
      "logits/rejected": -1.278634245927865,
      "logps/chosen": -250.25,
      "logps/rejected": -267.0,
      "loss": 0.6451339721679688,
      "mean_token_accuracy": 0.9491883218288422,
      "num_tokens": 65286854.0,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.319091796875,
      "rewards/margins": 0.359130859375,
      "rewards/rejected": -0.67626953125,
      "step": 174
    },
    {
      "entropy": 0.156494140625,
      "epoch": 0.8905852417302799,
      "grad_norm": 54.96915817260742,
      "learning_rate": 5.83756345177665e-07,
      "logits/chosen": -1.265227401251292,
      "logits/rejected": -1.2394450329297166,
      "logps/chosen": -169.0,
      "logps/rejected": -184.375,
      "loss": 0.73223876953125,
      "mean_token_accuracy": 0.9530220776796341,
      "num_tokens": 65705185.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": -0.24835205078125,
      "rewards/margins": 0.1209716796875,
      "rewards/rejected": -0.368408203125,
      "step": 175
    },
    {
      "entropy": 0.1611328125,
      "epoch": 0.8956743002544529,
      "grad_norm": 72.51837158203125,
      "learning_rate": 5.583756345177665e-07,
      "logits/chosen": -1.291017567652124,
      "logits/rejected": -1.3191793802079457,
      "logps/chosen": -260.75,
      "logps/rejected": -237.25,
      "loss": 0.888946533203125,
      "mean_token_accuracy": 0.9461717158555984,
      "num_tokens": 66041900.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.342071533203125,
      "rewards/margins": -0.199951171875,
      "rewards/rejected": -0.14239501953125,
      "step": 176
    },
    {
      "entropy": 0.165771484375,
      "epoch": 0.9007633587786259,
      "grad_norm": 63.78479766845703,
      "learning_rate": 5.329949238578681e-07,
      "logits/chosen": -1.2883858909839578,
      "logits/rejected": -1.2964681724098532,
      "logps/chosen": -227.5,
      "logps/rejected": -206.25,
      "loss": 0.831298828125,
      "mean_token_accuracy": 0.9456988126039505,
      "num_tokens": 66397264.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.43017578125,
      "rewards/margins": -0.098876953125,
      "rewards/rejected": -0.33203125,
      "step": 177
    },
    {
      "entropy": 0.16552734375,
      "epoch": 0.905852417302799,
      "grad_norm": 64.87358093261719,
      "learning_rate": 5.076142131979696e-07,
      "logits/chosen": -1.2762111021043254,
      "logits/rejected": -1.277499721225321,
      "logps/chosen": -215.25,
      "logps/rejected": -232.0,
      "loss": 0.82025146484375,
      "mean_token_accuracy": 0.9479079395532608,
      "num_tokens": 66746118.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.335205078125,
      "rewards/margins": -0.015380859375,
      "rewards/rejected": -0.320556640625,
      "step": 178
    },
    {
      "entropy": 0.165771484375,
      "epoch": 0.910941475826972,
      "grad_norm": 63.49721145629883,
      "learning_rate": 4.822335025380711e-07,
      "logits/chosen": -1.2594111699857853,
      "logits/rejected": -1.2441778402503918,
      "logps/chosen": -203.75,
      "logps/rejected": -218.0,
      "loss": 0.7030181884765625,
      "mean_token_accuracy": 0.949649915099144,
      "num_tokens": 67123472.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.20211029052734375,
      "rewards/margins": 0.1322021484375,
      "rewards/rejected": -0.333984375,
      "step": 179
    },
    {
      "entropy": 0.16650390625,
      "epoch": 0.916030534351145,
      "grad_norm": 76.09004974365234,
      "learning_rate": 4.568527918781726e-07,
      "logits/chosen": -1.286043293513197,
      "logits/rejected": -1.2636617381085613,
      "logps/chosen": -236.25,
      "logps/rejected": -225.0,
      "loss": 0.9158935546875,
      "mean_token_accuracy": 0.9456801265478134,
      "num_tokens": 67465986.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.5185546875,
      "rewards/margins": -0.237213134765625,
      "rewards/rejected": -0.281982421875,
      "step": 180
    },
    {
      "entropy": 0.1552734375,
      "epoch": 0.9211195928753181,
      "grad_norm": 63.31582260131836,
      "learning_rate": 4.314720812182741e-07,
      "logits/chosen": -1.251146523022114,
      "logits/rejected": -1.2975083912016157,
      "logps/chosen": -198.25,
      "logps/rejected": -203.75,
      "loss": 0.7587127685546875,
      "mean_token_accuracy": 0.9466626644134521,
      "num_tokens": 67864018.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.345458984375,
      "rewards/margins": 0.0047607421875,
      "rewards/rejected": -0.349609375,
      "step": 181
    },
    {
      "entropy": 0.153564453125,
      "epoch": 0.926208651399491,
      "grad_norm": 54.91245651245117,
      "learning_rate": 4.0609137055837566e-07,
      "logits/chosen": -1.2824599495241058,
      "logits/rejected": -1.2501124767523026,
      "logps/chosen": -194.0,
      "logps/rejected": -206.5,
      "loss": 0.65789794921875,
      "mean_token_accuracy": 0.9545126259326935,
      "num_tokens": 68229219.0,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.26611328125,
      "rewards/margins": 0.213134765625,
      "rewards/rejected": -0.47900390625,
      "step": 182
    },
    {
      "entropy": 0.165771484375,
      "epoch": 0.9312977099236641,
      "grad_norm": 60.69124221801758,
      "learning_rate": 3.807106598984772e-07,
      "logits/chosen": -1.2306977794006666,
      "logits/rejected": -1.2199483238128237,
      "logps/chosen": -212.5,
      "logps/rejected": -250.25,
      "loss": 0.7246932983398438,
      "mean_token_accuracy": 0.9500255733728409,
      "num_tokens": 68606870.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.113128662109375,
      "rewards/margins": 0.128448486328125,
      "rewards/rejected": -0.24212646484375,
      "step": 183
    },
    {
      "entropy": 0.162353515625,
      "epoch": 0.9363867684478372,
      "grad_norm": 64.28182983398438,
      "learning_rate": 3.553299492385787e-07,
      "logits/chosen": -1.2587565773941798,
      "logits/rejected": -1.2312963361066895,
      "logps/chosen": -212.5,
      "logps/rejected": -213.5,
      "loss": 0.64691162109375,
      "mean_token_accuracy": 0.9500741362571716,
      "num_tokens": 68978668.0,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.16424560546875,
      "rewards/margins": 0.3056640625,
      "rewards/rejected": -0.47021484375,
      "step": 184
    },
    {
      "entropy": 0.166015625,
      "epoch": 0.9414758269720102,
      "grad_norm": 57.6767692565918,
      "learning_rate": 3.2994923857868026e-07,
      "logits/chosen": -1.2751947302167712,
      "logits/rejected": -1.2666428366190272,
      "logps/chosen": -235.0,
      "logps/rejected": -263.5,
      "loss": 0.6388931274414062,
      "mean_token_accuracy": 0.9481093883514404,
      "num_tokens": 69372058.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.0687255859375,
      "rewards/margins": 0.2685546875,
      "rewards/rejected": -0.3369140625,
      "step": 185
    },
    {
      "entropy": 0.172607421875,
      "epoch": 0.9465648854961832,
      "grad_norm": 60.594364166259766,
      "learning_rate": 3.0456852791878176e-07,
      "logits/chosen": -1.2357604930632382,
      "logits/rejected": -1.2510668876440354,
      "logps/chosen": -225.25,
      "logps/rejected": -216.875,
      "loss": 0.702392578125,
      "mean_token_accuracy": 0.9432621896266937,
      "num_tokens": 69753477.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.2430419921875,
      "rewards/margins": 0.139495849609375,
      "rewards/rejected": -0.3818359375,
      "step": 186
    },
    {
      "entropy": 0.1611328125,
      "epoch": 0.9516539440203562,
      "grad_norm": 63.02968215942383,
      "learning_rate": 2.7918781725888326e-07,
      "logits/chosen": -1.2820828380810267,
      "logits/rejected": -1.3018126825693557,
      "logps/chosen": -237.0,
      "logps/rejected": -264.5,
      "loss": 0.7537841796875,
      "mean_token_accuracy": 0.9488907307386398,
      "num_tokens": 70144262.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.1806640625,
      "rewards/margins": 0.0987548828125,
      "rewards/rejected": -0.280029296875,
      "step": 187
    },
    {
      "entropy": 0.157470703125,
      "epoch": 0.9567430025445293,
      "grad_norm": 62.73226547241211,
      "learning_rate": 2.538071065989848e-07,
      "logits/chosen": -1.3243303259905526,
      "logits/rejected": -1.3251137517496847,
      "logps/chosen": -177.125,
      "logps/rejected": -189.5,
      "loss": 0.8604354858398438,
      "mean_token_accuracy": 0.9508524388074875,
      "num_tokens": 70522847.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.4222412109375,
      "rewards/margins": -0.1492919921875,
      "rewards/rejected": -0.272705078125,
      "step": 188
    },
    {
      "entropy": 0.158935546875,
      "epoch": 0.9618320610687023,
      "grad_norm": 58.20997619628906,
      "learning_rate": 2.284263959390863e-07,
      "logits/chosen": -1.3108909963701287,
      "logits/rejected": -1.307712402955017,
      "logps/chosen": -180.125,
      "logps/rejected": -224.0,
      "loss": 0.732025146484375,
      "mean_token_accuracy": 0.9533163607120514,
      "num_tokens": 70940314.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.1817626953125,
      "rewards/margins": 0.1083984375,
      "rewards/rejected": -0.2891845703125,
      "step": 189
    },
    {
      "entropy": 0.156982421875,
      "epoch": 0.9669211195928753,
      "grad_norm": 64.7269287109375,
      "learning_rate": 2.0304568527918783e-07,
      "logits/chosen": -1.3086706742163743,
      "logits/rejected": -1.3384005897848779,
      "logps/chosen": -246.75,
      "logps/rejected": -256.75,
      "loss": 0.761688232421875,
      "mean_token_accuracy": 0.9508141130208969,
      "num_tokens": 71285208.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.1785888671875,
      "rewards/margins": 0.101318359375,
      "rewards/rejected": -0.279296875,
      "step": 190
    },
    {
      "entropy": 0.158935546875,
      "epoch": 0.9720101781170484,
      "grad_norm": 76.1530990600586,
      "learning_rate": 1.7766497461928935e-07,
      "logits/chosen": -1.2052610815729892,
      "logits/rejected": -1.2051475881810254,
      "logps/chosen": -245.0,
      "logps/rejected": -249.25,
      "loss": 0.84283447265625,
      "mean_token_accuracy": 0.9478296190500259,
      "num_tokens": 71692156.0,
      "rewards/accuracies": 0.40625,
      "rewards/chosen": -0.44970703125,
      "rewards/margins": -0.0787353515625,
      "rewards/rejected": -0.37158203125,
      "step": 191
    },
    {
      "entropy": 0.153564453125,
      "epoch": 0.9770992366412213,
      "grad_norm": 49.5556640625,
      "learning_rate": 1.5228426395939088e-07,
      "logits/chosen": -1.2858759972786784,
      "logits/rejected": -1.3171012184101922,
      "logps/chosen": -206.0,
      "logps/rejected": -227.5,
      "loss": 0.5926513671875,
      "mean_token_accuracy": 0.9544648975133896,
      "num_tokens": 72062381.0,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.1572265625,
      "rewards/margins": 0.33837890625,
      "rewards/rejected": -0.49658203125,
      "step": 192
    },
    {
      "entropy": 0.163818359375,
      "epoch": 0.9821882951653944,
      "grad_norm": 55.3388671875,
      "learning_rate": 1.269035532994924e-07,
      "logits/chosen": -1.2850746114469256,
      "logits/rejected": -1.2788264737681292,
      "logps/chosen": -187.75,
      "logps/rejected": -178.75,
      "loss": 0.7740325927734375,
      "mean_token_accuracy": 0.9483982175588608,
      "num_tokens": 72459408.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.284423828125,
      "rewards/margins": -0.025634765625,
      "rewards/rejected": -0.25927734375,
      "step": 193
    },
    {
      "entropy": 0.16015625,
      "epoch": 0.9872773536895675,
      "grad_norm": 62.76466751098633,
      "learning_rate": 1.0152284263959391e-07,
      "logits/chosen": -1.2936464840174513,
      "logits/rejected": -1.3121440655724232,
      "logps/chosen": -232.25,
      "logps/rejected": -268.25,
      "loss": 0.710235595703125,
      "mean_token_accuracy": 0.952901765704155,
      "num_tokens": 72853047.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.21728515625,
      "rewards/margins": 0.15203857421875,
      "rewards/rejected": -0.36962890625,
      "step": 194
    },
    {
      "entropy": 0.1572265625,
      "epoch": 0.9923664122137404,
      "grad_norm": 53.645687103271484,
      "learning_rate": 7.614213197969544e-08,
      "logits/chosen": -1.2712459537508989,
      "logits/rejected": -1.234670637217532,
      "logps/chosen": -163.875,
      "logps/rejected": -177.5,
      "loss": 0.69317626953125,
      "mean_token_accuracy": 0.9523515999317169,
      "num_tokens": 73277603.0,
      "rewards/accuracies": 0.3125,
      "rewards/chosen": -0.1658935546875,
      "rewards/margins": 0.06201171875,
      "rewards/rejected": -0.22802734375,
      "step": 195
    },
    {
      "entropy": 0.166015625,
      "epoch": 0.9974554707379135,
      "grad_norm": 65.67327880859375,
      "learning_rate": 5.076142131979696e-08,
      "logits/chosen": -1.2582488581214466,
      "logits/rejected": -1.2718061289438034,
      "logps/chosen": -210.0,
      "logps/rejected": -215.5,
      "loss": 0.739013671875,
      "mean_token_accuracy": 0.9474593698978424,
      "num_tokens": 73625261.0,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.23590087890625,
      "rewards/margins": 0.0318603515625,
      "rewards/rejected": -0.267578125,
      "step": 196
    },
    {
      "entropy": 0.15283203125,
      "epoch": 1.0,
      "grad_norm": 91.96894836425781,
      "learning_rate": 2.538071065989848e-08,
      "logits/chosen": -1.2407954088557835,
      "logits/rejected": -1.2307852017793066,
      "logps/chosen": -217.5,
      "logps/rejected": -237.0,
      "loss": 0.7696762084960938,
      "mean_token_accuracy": 0.9491404592990875,
      "num_tokens": 73770526.0,
      "rewards/accuracies": 0.375,
      "rewards/chosen": -0.31298828125,
      "rewards/margins": 0.15869140625,
      "rewards/rejected": -0.470703125,
      "step": 197
    }
  ],
  "logging_steps": 1,
  "max_steps": 197,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.5608290532420223e+19,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}