{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8192,
  "eval_steps": 500,
  "global_step": 128,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.07638888888888888,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7457.0,
      "completions/mean_length": 2484.15283203125,
      "completions/mean_terminated_length": 2012.1199951171875,
      "completions/min_length": 559.6666666666666,
      "completions/min_terminated_length": 559.6666666666666,
      "epoch": 0.0192,
      "grad_norm": 0.11786151677370071,
      "kl": 0.000663916269938151,
      "learning_rate": 3.948717948717949e-06,
      "loss": 0.0792,
      "num_tokens": 736454.0,
      "reward": 1.920138915379842,
      "reward_std": 0.06854866879681747,
      "rewards/accuracy_reward/mean": 0.9965277711550394,
      "rewards/accuracy_reward/std": 0.03402068714300791,
      "rewards/format_reward/mean": 0.9236111044883728,
      "rewards/format_reward/std": 0.2288934737443924,
      "step": 3
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09027777777777779,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7337.333333333333,
      "completions/mean_length": 2525.18408203125,
      "completions/mean_terminated_length": 1964.8008219401042,
      "completions/min_length": 538.0,
      "completions/min_terminated_length": 538.0,
      "epoch": 0.0384,
      "grad_norm": 0.10443487018346786,
      "kl": 0.001732190450032552,
      "learning_rate": 3.8717948717948715e-06,
      "loss": 0.0688,
      "num_tokens": 1486231.0,
      "reward": 1.90625003973643,
      "reward_std": 0.11128650108973186,
      "rewards/accuracy_reward/mean": 0.993055542310079,
      "rewards/accuracy_reward/std": 0.0680413767695427,
      "rewards/format_reward/mean": 0.9131944378217062,
      "rewards/format_reward/std": 0.28209365407625836,
      "step": 6
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.07638888888888891,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7734.0,
      "completions/mean_length": 2304.361124674479,
      "completions/mean_terminated_length": 1825.1893717447917,
      "completions/min_length": 428.6666666666667,
      "completions/min_terminated_length": 428.6666666666667,
      "epoch": 0.0576,
      "grad_norm": 0.11160896718502045,
      "kl": 0.0075155893961588545,
      "learning_rate": 3.7948717948717945e-06,
      "loss": 0.0532,
      "num_tokens": 2170377.0,
      "reward": 1.9062500794728596,
      "reward_std": 0.09471678485472997,
      "rewards/accuracy_reward/mean": 0.9791666666666666,
      "rewards/accuracy_reward/std": 0.11481705804665883,
      "rewards/format_reward/mean": 0.9270833333333334,
      "rewards/format_reward/std": 0.25792460640271503,
      "step": 9
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.07986111111111112,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7192.333333333333,
      "completions/mean_length": 2486.7847900390625,
      "completions/mean_terminated_length": 2002.9917805989583,
      "completions/min_length": 588.6666666666666,
      "completions/min_terminated_length": 588.6666666666666,
      "epoch": 0.0768,
      "grad_norm": 0.07670364528894424,
      "kl": 0.007045745849609375,
      "learning_rate": 3.717948717948718e-06,
      "loss": 0.0655,
      "num_tokens": 2906641.0,
      "reward": 1.9027778307596843,
      "reward_std": 0.08286526799201965,
      "rewards/accuracy_reward/mean": 0.9826388955116272,
      "rewards/accuracy_reward/std": 0.12973794837792715,
      "rewards/format_reward/mean": 0.9201388955116272,
      "rewards/format_reward/std": 0.25063540538152057,
      "step": 12
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.048611111111111084,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 5679.0,
      "completions/mean_length": 2111.114705403646,
      "completions/mean_terminated_length": 1806.3657633463542,
      "completions/min_length": 638.3333333333334,
      "completions/min_terminated_length": 638.3333333333334,
      "epoch": 0.096,
      "grad_norm": 0.08451474457979202,
      "kl": 0.01056671142578125,
      "learning_rate": 3.6410256410256406e-06,
      "loss": 0.0438,
      "num_tokens": 3534238.0,
      "reward": 1.93750003973643,
      "reward_std": 0.08955795814593633,
      "rewards/accuracy_reward/mean": 0.9861111044883728,
      "rewards/accuracy_reward/std": 0.11590000987052917,
      "rewards/format_reward/mean": 0.9513888756434122,
      "rewards/format_reward/std": 0.20225290457407633,
      "step": 15
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.013888888888888876,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 6870.666666666667,
      "completions/mean_length": 1940.4757486979167,
      "completions/mean_terminated_length": 1853.5159098307292,
      "completions/min_length": 632.0,
      "completions/min_terminated_length": 632.0,
      "epoch": 0.1152,
      "grad_norm": 0.10638942569494247,
      "kl": 0.013872782389322916,
      "learning_rate": 3.564102564102564e-06,
      "loss": 0.0318,
      "num_tokens": 4114947.0,
      "reward": 1.9236111640930176,
      "reward_std": 0.086211613068978,
      "rewards/accuracy_reward/mean": 0.9340277711550394,
      "rewards/accuracy_reward/std": 0.21398874868949255,
      "rewards/format_reward/mean": 0.9895833134651184,
      "rewards/format_reward/std": 0.10206206391255061,
      "step": 18
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.041666666666666664,
      "completions/max_length": 7178.333333333333,
      "completions/max_terminated_length": 5503.666666666667,
      "completions/mean_length": 2042.7118733723958,
      "completions/mean_terminated_length": 1780.8347981770833,
      "completions/min_length": 534.6666666666666,
      "completions/min_terminated_length": 534.6666666666666,
      "epoch": 0.1344,
      "grad_norm": 0.06415776163339615,
      "kl": 0.016947428385416668,
      "learning_rate": 3.487179487179487e-06,
      "loss": 0.0258,
      "num_tokens": 4722190.0,
      "reward": 1.9236111640930176,
      "reward_std": 0.03402068962653478,
      "rewards/accuracy_reward/mean": 0.9548611243565878,
      "rewards/accuracy_reward/std": 0.15458878378073374,
      "rewards/format_reward/mean": 0.96875,
      "rewards/format_reward/std": 0.12663276741902033,
      "step": 21
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.045138888888888916,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7599.333333333333,
      "completions/mean_length": 2192.9861653645835,
      "completions/mean_terminated_length": 1915.0709635416667,
      "completions/min_length": 501.6666666666667,
      "completions/min_terminated_length": 501.6666666666667,
      "epoch": 0.1536,
      "grad_norm": 0.08798938244581223,
      "kl": 0.02048492431640625,
      "learning_rate": 3.41025641025641e-06,
      "loss": 0.0258,
      "num_tokens": 5379972.0,
      "reward": 1.8645833730697632,
      "reward_std": 0.09762249266107877,
      "rewards/accuracy_reward/mean": 0.9027777711550394,
      "rewards/accuracy_reward/std": 0.286453440785408,
      "rewards/format_reward/mean": 0.9618055621782938,
      "rewards/format_reward/std": 0.18727133671442667,
      "step": 24
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.020833333333333332,
      "completions/max_length": 7410.0,
      "completions/max_terminated_length": 6817.666666666667,
      "completions/mean_length": 2016.982666015625,
      "completions/mean_terminated_length": 1887.3319091796875,
      "completions/min_length": 514.0,
      "completions/min_terminated_length": 514.0,
      "epoch": 0.1728,
      "grad_norm": 0.10734036564826965,
      "kl": 0.0256195068359375,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0323,
      "num_tokens": 5984941.0,
      "reward": 1.8993055820465088,
      "reward_std": 0.10998130093018214,
      "rewards/accuracy_reward/mean": 0.9201388955116272,
      "rewards/accuracy_reward/std": 0.2663828631242116,
      "rewards/format_reward/mean": 0.9791666666666666,
      "rewards/format_reward/std": 0.10847451289494832,
      "step": 27
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7886.666666666667,
      "completions/mean_length": 2276.7084147135415,
      "completions/mean_terminated_length": 2084.757283528646,
      "completions/min_length": 573.6666666666666,
      "completions/min_terminated_length": 573.6666666666666,
      "epoch": 0.192,
      "grad_norm": 0.09922217577695847,
      "kl": 0.022427876790364582,
      "learning_rate": 3.2564102564102564e-06,
      "loss": 0.0182,
      "num_tokens": 6662305.0,
      "reward": 1.8333333730697632,
      "reward_std": 0.1692849819858869,
      "rewards/accuracy_reward/mean": 0.875,
      "rewards/accuracy_reward/std": 0.31955354909102124,
      "rewards/format_reward/mean": 0.9583333333333334,
      "rewards/format_reward/std": 0.197714701294899,
      "step": 30
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 6856.333333333333,
      "completions/max_terminated_length": 5939.666666666667,
      "completions/mean_length": 2417.1563313802085,
      "completions/mean_terminated_length": 2233.5343424479165,
      "completions/min_length": 661.3333333333334,
      "completions/min_terminated_length": 661.3333333333334,
      "epoch": 0.2112,
      "grad_norm": 0.13434383273124695,
      "kl": 0.019597371419270832,
      "learning_rate": 3.179487179487179e-06,
      "loss": 0.0666,
      "num_tokens": 7382296.0,
      "reward": 1.87500003973643,
      "reward_std": 0.14174887910485268,
      "rewards/accuracy_reward/mean": 0.90625,
      "rewards/accuracy_reward/std": 0.27019980053106946,
      "rewards/format_reward/mean": 0.96875,
      "rewards/format_reward/std": 0.1349801371494929,
      "step": 33
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.010416666666666666,
      "completions/max_length": 6942.666666666667,
      "completions/max_terminated_length": 6595.333333333333,
      "completions/mean_length": 1717.2361653645833,
      "completions/mean_terminated_length": 1649.8081461588542,
      "completions/min_length": 371.3333333333333,
      "completions/min_terminated_length": 371.3333333333333,
      "epoch": 0.2304,
      "grad_norm": 0.11780474334955215,
      "kl": 0.028951009114583332,
      "learning_rate": 3.1025641025641025e-06,
      "loss": 0.0318,
      "num_tokens": 7898124.0,
      "reward": 1.9409722884496052,
      "reward_std": 0.08892839774489403,
      "rewards/accuracy_reward/mean": 0.9652777910232544,
      "rewards/accuracy_reward/std": 0.17682797213395438,
      "rewards/format_reward/mean": 0.975694457689921,
      "rewards/format_reward/std": 0.15401925643285116,
      "step": 36
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.045138888888888874,
      "completions/max_length": 7901.0,
      "completions/max_terminated_length": 7665.0,
      "completions/mean_length": 1984.8681233723958,
      "completions/mean_terminated_length": 1685.4075520833333,
      "completions/min_length": 326.3333333333333,
      "completions/min_terminated_length": 326.3333333333333,
      "epoch": 0.2496,
      "grad_norm": 0.1302698850631714,
      "kl": 0.0289154052734375,
      "learning_rate": 3.0256410256410256e-06,
      "loss": 0.0366,
      "num_tokens": 8500984.0,
      "reward": 1.8715278307596843,
      "reward_std": 0.13606221228837967,
      "rewards/accuracy_reward/mean": 0.9270833333333334,
      "rewards/accuracy_reward/std": 0.25223057965437573,
      "rewards/format_reward/mean": 0.9444444378217062,
      "rewards/format_reward/std": 0.2128016253312429,
      "step": 39
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.02430555555555558,
      "completions/max_length": 6943.666666666667,
      "completions/max_terminated_length": 6541.0,
      "completions/mean_length": 1753.7569986979167,
      "completions/mean_terminated_length": 1596.4458414713542,
      "completions/min_length": 336.0,
      "completions/min_terminated_length": 336.0,
      "epoch": 0.2688,
      "grad_norm": 0.12249383330345154,
      "kl": 0.030186971028645832,
      "learning_rate": 2.9487179487179486e-06,
      "loss": 0.0353,
      "num_tokens": 9033138.0,
      "reward": 1.8680555820465088,
      "reward_std": 0.19277991354465485,
      "rewards/accuracy_reward/mean": 0.906250019868215,
      "rewards/accuracy_reward/std": 0.29195430874824524,
      "rewards/format_reward/mean": 0.961805542310079,
      "rewards/format_reward/std": 0.18208982795476913,
      "step": 42
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.027777777777777752,
      "completions/max_length": 7697.333333333333,
      "completions/max_terminated_length": 6401.666666666667,
      "completions/mean_length": 1648.7222493489583,
      "completions/mean_terminated_length": 1464.5533447265625,
      "completions/min_length": 229.66666666666666,
      "completions/min_terminated_length": 229.66666666666666,
      "epoch": 0.288,
      "grad_norm": 0.1356375366449356,
      "kl": 0.034571329752604164,
      "learning_rate": 2.8717948717948717e-06,
      "loss": 0.0566,
      "num_tokens": 9541138.0,
      "reward": 1.87500003973643,
      "reward_std": 0.14326218763987222,
      "rewards/accuracy_reward/mean": 0.9236111044883728,
      "rewards/accuracy_reward/std": 0.24128751705090204,
      "rewards/format_reward/mean": 0.9513888955116272,
      "rewards/format_reward/std": 0.17212405304114023,
      "step": 45
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.059027777777777755,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7894.0,
      "completions/mean_length": 2024.1111653645833,
      "completions/mean_terminated_length": 1630.966552734375,
      "completions/min_length": 298.3333333333333,
      "completions/min_terminated_length": 298.3333333333333,
      "epoch": 0.3072,
      "grad_norm": 0.14614339172840118,
      "kl": 0.033233642578125,
      "learning_rate": 2.7948717948717948e-06,
      "loss": 0.0668,
      "num_tokens": 10152390.0,
      "reward": 1.8819444974263508,
      "reward_std": 0.14595978458722433,
      "rewards/accuracy_reward/mean": 0.9479166666666666,
      "rewards/accuracy_reward/std": 0.2198773274819056,
      "rewards/format_reward/mean": 0.9340277711550394,
      "rewards/format_reward/std": 0.24177277584870657,
      "step": 48
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 8080.333333333333,
      "completions/max_terminated_length": 6668.333333333333,
      "completions/mean_length": 2075.5382486979165,
      "completions/mean_terminated_length": 1875.0848795572917,
      "completions/min_length": 313.0,
      "completions/min_terminated_length": 313.0,
      "epoch": 0.3264,
      "grad_norm": 0.13130497932434082,
      "kl": 0.029052734375,
      "learning_rate": 2.717948717948718e-06,
      "loss": 0.0391,
      "num_tokens": 10781333.0,
      "reward": 1.8715277910232544,
      "reward_std": 0.173075832426548,
      "rewards/accuracy_reward/mean": 0.9131944378217062,
      "rewards/accuracy_reward/std": 0.27422525485356647,
      "rewards/format_reward/mean": 0.9583333532015482,
      "rewards/format_reward/std": 0.18832933902740479,
      "step": 51
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03472222222222221,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 6894.666666666667,
      "completions/mean_length": 1978.5972900390625,
      "completions/mean_terminated_length": 1754.5013020833333,
      "completions/min_length": 291.0,
      "completions/min_terminated_length": 291.0,
      "epoch": 0.3456,
      "grad_norm": 0.11747946590185165,
      "kl": 0.025873819986979168,
      "learning_rate": 2.641025641025641e-06,
      "loss": 0.0452,
      "num_tokens": 11385513.0,
      "reward": 1.8368055820465088,
      "reward_std": 0.1378952513138453,
      "rewards/accuracy_reward/mean": 0.8819444378217062,
      "rewards/accuracy_reward/std": 0.3174315243959427,
      "rewards/format_reward/mean": 0.9548611243565878,
      "rewards/format_reward/std": 0.20342316726843515,
      "step": 54
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.013888888888888876,
      "completions/max_length": 7534.0,
      "completions/max_terminated_length": 7530.666666666667,
      "completions/mean_length": 1764.263916015625,
      "completions/mean_terminated_length": 1680.0013834635417,
      "completions/min_length": 283.0,
      "completions/min_terminated_length": 283.0,
      "epoch": 0.3648,
      "grad_norm": 0.12080682069063187,
      "kl": 0.032323201497395836,
      "learning_rate": 2.5641025641025644e-06,
      "loss": 0.0415,
      "num_tokens": 11918815.0,
      "reward": 1.864583412806193,
      "reward_std": 0.17177063475052515,
      "rewards/accuracy_reward/mean": 0.9027777711550394,
      "rewards/accuracy_reward/std": 0.28834333022435504,
      "rewards/format_reward/mean": 0.961805542310079,
      "rewards/format_reward/std": 0.1363807593782743,
      "step": 57
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.048611111111111126,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 6987.666666666667,
      "completions/mean_length": 2011.3229573567708,
      "completions/mean_terminated_length": 1690.6886393229167,
      "completions/min_length": 294.0,
      "completions/min_terminated_length": 294.0,
      "epoch": 0.384,
      "grad_norm": 0.12239208817481995,
      "kl": 0.030438741048177082,
      "learning_rate": 2.487179487179487e-06,
      "loss": 0.0768,
      "num_tokens": 12522634.0,
      "reward": 1.8854167064030964,
      "reward_std": 0.15016684432824454,
      "rewards/accuracy_reward/mean": 0.9409722089767456,
      "rewards/accuracy_reward/std": 0.23519064486026764,
      "rewards/format_reward/mean": 0.944444457689921,
      "rewards/format_reward/std": 0.22536789874235788,
      "step": 60
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.034722222222222245,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 6971.333333333333,
      "completions/mean_length": 2042.4896647135417,
      "completions/mean_terminated_length": 1825.8701578776042,
      "completions/min_length": 314.0,
      "completions/min_terminated_length": 314.0,
      "epoch": 0.4032,
      "grad_norm": 0.10179228335618973,
      "kl": 0.031634012858072914,
      "learning_rate": 2.41025641025641e-06,
      "loss": 0.0337,
      "num_tokens": 13138633.0,
      "reward": 1.8541667064030964,
      "reward_std": 0.20934962232907614,
      "rewards/accuracy_reward/mean": 0.8958333333333334,
      "rewards/accuracy_reward/std": 0.3057904740174611,
      "rewards/format_reward/mean": 0.9583333333333334,
      "rewards/format_reward/std": 0.197714701294899,
      "step": 63
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.038194444444444454,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7261.666666666667,
      "completions/mean_length": 1903.9722493489583,
      "completions/mean_terminated_length": 1650.5534261067708,
      "completions/min_length": 227.66666666666666,
      "completions/min_terminated_length": 227.66666666666666,
      "epoch": 0.4224,
      "grad_norm": 0.10856553167104721,
      "kl": 0.031336466471354164,
      "learning_rate": 2.3333333333333336e-06,
      "loss": 0.0397,
      "num_tokens": 13714469.0,
      "reward": 1.8819444576899211,
      "reward_std": 0.15875921150048575,
      "rewards/accuracy_reward/mean": 0.9270833333333334,
      "rewards/accuracy_reward/std": 0.2572407325108846,
      "rewards/format_reward/mean": 0.9548611044883728,
      "rewards/format_reward/std": 0.19624211142460504,
      "step": 66
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.038194444444444454,
      "completions/max_length": 7431.666666666667,
      "completions/max_terminated_length": 7204.666666666667,
      "completions/mean_length": 1631.7222900390625,
      "completions/mean_terminated_length": 1369.1958414713542,
      "completions/min_length": 245.0,
      "completions/min_terminated_length": 245.0,
      "epoch": 0.4416,
      "grad_norm": 0.11824989318847656,
      "kl": 0.03619384765625,
      "learning_rate": 2.2564102564102562e-06,
      "loss": 0.0374,
      "num_tokens": 14208711.0,
      "reward": 1.81250003973643,
      "reward_std": 0.16117709378401437,
      "rewards/accuracy_reward/mean": 0.8611111044883728,
      "rewards/accuracy_reward/std": 0.3456921974817912,
      "rewards/format_reward/mean": 0.9513888955116272,
      "rewards/format_reward/std": 0.20108659317096075,
      "step": 69
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.041666666666666664,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 6924.666666666667,
      "completions/mean_length": 1506.6423746744792,
      "completions/mean_terminated_length": 1216.2833251953125,
      "completions/min_length": 246.0,
      "completions/min_terminated_length": 246.0,
      "epoch": 0.4608,
      "grad_norm": 0.1255887746810913,
      "kl": 0.053243001302083336,
      "learning_rate": 2.1794871794871793e-06,
      "loss": 0.0383,
      "num_tokens": 14664410.0,
      "reward": 1.9236111640930176,
      "reward_std": 0.10031624138355255,
      "rewards/accuracy_reward/mean": 0.9548611044883728,
      "rewards/accuracy_reward/std": 0.2063711335261663,
      "rewards/format_reward/mean": 0.968749980131785,
      "rewards/format_reward/std": 0.16793753455082575,
      "step": 72
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.024305555555555542,
      "completions/max_length": 7341.0,
      "completions/max_terminated_length": 7337.666666666667,
      "completions/mean_length": 1445.312520345052,
      "completions/mean_terminated_length": 1281.9166463216145,
      "completions/min_length": 166.66666666666666,
      "completions/min_terminated_length": 166.66666666666666,
      "epoch": 0.48,
      "grad_norm": 0.14921000599861145,
      "kl": 0.042795817057291664,
      "learning_rate": 2.1025641025641023e-06,
      "loss": 0.036,
      "num_tokens": 15103004.0,
      "reward": 1.87500003973643,
      "reward_std": 0.11799862856666248,
      "rewards/accuracy_reward/mean": 0.9027777711550394,
      "rewards/accuracy_reward/std": 0.28232812881469727,
      "rewards/format_reward/mean": 0.9722222089767456,
      "rewards/format_reward/std": 0.1211421936750412,
      "step": 75
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.045138888888888874,
      "completions/max_length": 7986.666666666667,
      "completions/max_terminated_length": 7702.333333333333,
      "completions/mean_length": 1538.5555826822917,
      "completions/mean_terminated_length": 1232.4976399739583,
      "completions/min_length": 201.0,
      "completions/min_terminated_length": 201.0,
      "epoch": 0.4992,
      "grad_norm": 0.09636721014976501,
      "kl": 0.04656982421875,
      "learning_rate": 2.0256410256410254e-06,
      "loss": 0.0377,
      "num_tokens": 15574308.0,
      "reward": 1.87500003973643,
      "reward_std": 0.10612767189741135,
      "rewards/accuracy_reward/mean": 0.9166666666666666,
      "rewards/accuracy_reward/std": 0.26633305847644806,
      "rewards/format_reward/mean": 0.9583333333333334,
      "rewards/format_reward/std": 0.16222142179807028,
      "step": 78
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.038194444444444454,
      "completions/max_length": 8137.666666666667,
      "completions/max_terminated_length": 7975.0,
      "completions/mean_length": 1834.2222900390625,
      "completions/mean_terminated_length": 1581.649169921875,
      "completions/min_length": 203.0,
      "completions/min_terminated_length": 203.0,
      "epoch": 0.5184,
      "grad_norm": 0.9437447786331177,
      "kl": 0.051630655924479164,
      "learning_rate": 1.9487179487179485e-06,
      "loss": 0.0323,
      "num_tokens": 16131508.0,
      "reward": 1.8506944974263508,
      "reward_std": 0.13935044904549918,
      "rewards/accuracy_reward/mean": 0.8923611044883728,
      "rewards/accuracy_reward/std": 0.30875815947850543,
      "rewards/format_reward/mean": 0.9583333333333334,
      "rewards/format_reward/std": 0.18958522379398346,
      "step": 81
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.045138888888888916,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7797.666666666667,
      "completions/mean_length": 1368.4097900390625,
      "completions/mean_terminated_length": 1047.0757649739583,
      "completions/min_length": 174.33333333333334,
      "completions/min_terminated_length": 174.33333333333334,
      "epoch": 0.5376,
      "grad_norm": 0.12685449421405792,
      "kl": 0.043909708658854164,
      "learning_rate": 1.8717948717948718e-06,
      "loss": 0.0486,
      "num_tokens": 16552250.0,
      "reward": 1.9166667064030964,
      "reward_std": 0.08621161431074142,
      "rewards/accuracy_reward/mean": 0.9583333333333334,
      "rewards/accuracy_reward/std": 0.19592776894569397,
      "rewards/format_reward/mean": 0.9583333333333334,
      "rewards/format_reward/std": 0.18958522627751032,
      "step": 84
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00694444444444442,
      "completions/max_length": 8058.0,
      "completions/max_terminated_length": 7524.0,
      "completions/mean_length": 1276.295166015625,
      "completions/mean_terminated_length": 1228.2986246744792,
      "completions/min_length": 176.33333333333334,
      "completions/min_terminated_length": 176.33333333333334,
      "epoch": 0.5568,
      "grad_norm": 0.10477358847856522,
      "kl": 0.041203816731770836,
      "learning_rate": 1.7948717948717948e-06,
      "loss": 0.03,
      "num_tokens": 16944123.0,
      "reward": 1.93750003973643,
      "reward_std": 0.0895579606294632,
      "rewards/accuracy_reward/mean": 0.9513888955116272,
      "rewards/accuracy_reward/std": 0.21386653184890747,
      "rewards/format_reward/mean": 0.9861111044883728,
      "rewards/format_reward/std": 0.11590000490347545,
      "step": 87
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.048611111111111126,
      "completions/max_length": 7305.0,
      "completions/max_terminated_length": 6282.333333333333,
      "completions/mean_length": 1492.5034993489583,
      "completions/mean_terminated_length": 1149.2331136067708,
      "completions/min_length": 252.0,
      "completions/min_terminated_length": 252.0,
      "epoch": 0.576,
      "grad_norm": 0.13828961551189423,
      "kl": 0.037953694661458336,
      "learning_rate": 1.7179487179487177e-06,
      "loss": 0.0362,
      "num_tokens": 17398714.0,
      "reward": 1.8854167064030964,
      "reward_std": 0.10882141316930453,
      "rewards/accuracy_reward/mean": 0.9375,
      "rewards/accuracy_reward/std": 0.24268604318300882,
      "rewards/format_reward/mean": 0.9479166666666666,
      "rewards/format_reward/std": 0.17973358432451883,
      "step": 90
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7929.0,
      "completions/mean_length": 1596.2466227213542,
      "completions/mean_terminated_length": 1390.1203206380208,
      "completions/min_length": 207.0,
      "completions/min_terminated_length": 207.0,
      "epoch": 0.5952,
      "grad_norm": 0.10115975886583328,
      "kl": 0.037958780924479164,
      "learning_rate": 1.641025641025641e-06,
      "loss": 0.0531,
      "num_tokens": 17882805.0,
      "reward": 1.8784722487131755,
      "reward_std": 0.08846472824613254,
      "rewards/accuracy_reward/mean": 0.906250019868215,
      "rewards/accuracy_reward/std": 0.28532982369263965,
      "rewards/format_reward/mean": 0.9722222288449606,
      "rewards/format_reward/std": 0.1563331459959348,
      "step": 93
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03472222222222221,
      "completions/max_length": 6898.333333333333,
      "completions/max_terminated_length": 6688.666666666667,
      "completions/mean_length": 1595.3646647135417,
      "completions/mean_terminated_length": 1364.8641357421875,
      "completions/min_length": 208.66666666666666,
      "completions/min_terminated_length": 208.66666666666666,
      "epoch": 0.6144,
      "grad_norm": 0.08329301327466965,
      "kl": 0.038645426432291664,
      "learning_rate": 1.564102564102564e-06,
      "loss": 0.0404,
      "num_tokens": 18366630.0,
      "reward": 1.9097222884496052,
      "reward_std": 0.1264843593041102,
      "rewards/accuracy_reward/mean": 0.9513888756434122,
      "rewards/accuracy_reward/std": 0.21502756575743356,
      "rewards/format_reward/mean": 0.9583333333333334,
      "rewards/format_reward/std": 0.18493476013342539,
      "step": 96
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.038194444444444454,
      "completions/max_length": 7504.666666666667,
      "completions/max_terminated_length": 6996.0,
      "completions/mean_length": 1620.4861246744792,
      "completions/mean_terminated_length": 1355.0765380859375,
      "completions/min_length": 182.66666666666666,
      "completions/min_terminated_length": 182.66666666666666,
      "epoch": 0.6336,
      "grad_norm": 0.07651077955961227,
      "kl": 0.038045247395833336,
      "learning_rate": 1.487179487179487e-06,
      "loss": 0.0408,
      "num_tokens": 18857216.0,
      "reward": 1.9131944576899211,
      "reward_std": 0.03627380480368932,
      "rewards/accuracy_reward/mean": 0.9479166666666666,
      "rewards/accuracy_reward/std": 0.19905283798774084,
      "rewards/format_reward/mean": 0.9652777711550394,
      "rewards/format_reward/std": 0.13169091691573462,
      "step": 99
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.038194444444444454,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7569.0,
      "completions/mean_length": 1601.21533203125,
      "completions/mean_terminated_length": 1339.799072265625,
      "completions/min_length": 140.33333333333334,
      "completions/min_terminated_length": 140.33333333333334,
      "epoch": 0.6528,
      "grad_norm": 0.1341271549463272,
      "kl": 0.035313924153645836,
      "learning_rate": 1.4102564102564104e-06,
      "loss": 0.0451,
      "num_tokens": 19348528.0,
      "reward": 1.87500003973643,
      "reward_std": 0.14174888034661612,
      "rewards/accuracy_reward/mean": 0.9270833333333334,
      "rewards/accuracy_reward/std": 0.26084429522355396,
      "rewards/format_reward/mean": 0.9479166666666666,
      "rewards/format_reward/std": 0.22252296408017477,
      "step": 102
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.045138888888888874,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7510.666666666667,
      "completions/mean_length": 1486.263916015625,
      "completions/mean_terminated_length": 1165.8067626953125,
      "completions/min_length": 135.33333333333334,
      "completions/min_terminated_length": 135.33333333333334,
      "epoch": 0.672,
      "grad_norm": 0.11015435308218002,
      "kl": 0.034245808919270836,
      "learning_rate": 1.3333333333333332e-06,
      "loss": 0.0214,
      "num_tokens": 19798676.0,
      "reward": 1.9166667064030964,
      "reward_std": 0.0998756115635236,
      "rewards/accuracy_reward/mean": 0.961805542310079,
      "rewards/accuracy_reward/std": 0.19221885005633035,
      "rewards/format_reward/mean": 0.9548611044883728,
      "rewards/format_reward/std": 0.20372549692789713,
      "step": 105
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.03125,
      "completions/max_length": 7331.0,
      "completions/max_terminated_length": 7054.666666666667,
      "completions/mean_length": 1494.9896647135417,
      "completions/mean_terminated_length": 1280.5859375,
      "completions/min_length": 221.0,
      "completions/min_terminated_length": 221.0,
      "epoch": 0.6912,
      "grad_norm": 0.11232707649469376,
      "kl": 0.0319671630859375,
      "learning_rate": 1.2564102564102563e-06,
      "loss": 0.0022,
      "num_tokens": 20265089.0,
      "reward": 1.8819444974263508,
      "reward_std": 0.10322195539871852,
      "rewards/accuracy_reward/mean": 0.9166666666666666,
      "rewards/accuracy_reward/std": 0.27038883169492084,
      "rewards/format_reward/mean": 0.9652777711550394,
      "rewards/format_reward/std": 0.1454235017299652,
      "step": 108
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.045138888888888874,
      "completions/max_length": 7708.333333333333,
      "completions/max_terminated_length": 6119.333333333333,
      "completions/mean_length": 1444.1944580078125,
      "completions/mean_terminated_length": 1120.3043619791667,
      "completions/min_length": 238.33333333333334,
      "completions/min_terminated_length": 238.33333333333334,
      "epoch": 0.7104,
      "grad_norm": 0.09397315979003906,
      "kl": 0.029179890950520832,
      "learning_rate": 1.1794871794871795e-06,
      "loss": 0.0349,
      "num_tokens": 20706937.0,
      "reward": 1.9027778307596843,
      "reward_std": 0.0833058978120486,
      "rewards/accuracy_reward/mean": 0.937499980131785,
      "rewards/accuracy_reward/std": 0.24120144049326578,
      "rewards/format_reward/mean": 0.9652777711550394,
      "rewards/format_reward/std": 0.1454235017299652,
      "step": 111
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.09027777777777779,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 7374.666666666667,
      "completions/mean_length": 1775.2361653645833,
      "completions/mean_terminated_length": 1136.4374389648438,
      "completions/min_length": 140.33333333333334,
      "completions/min_terminated_length": 140.33333333333334,
      "epoch": 0.7296,
      "grad_norm": 0.17233380675315857,
      "kl": 0.0392608642578125,
      "learning_rate": 1.1025641025641026e-06,
      "loss": 0.0398,
      "num_tokens": 21245775.0,
      "reward": 1.8645833730697632,
      "reward_std": 0.10707558691501617,
      "rewards/accuracy_reward/mean": 0.9513888756434122,
      "rewards/accuracy_reward/std": 0.21238192915916443,
      "rewards/format_reward/mean": 0.913194457689921,
      "rewards/format_reward/std": 0.27788423001766205,
      "step": 114
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.05902777777777779,
      "completions/max_length": 7707.333333333333,
      "completions/max_terminated_length": 6991.333333333333,
      "completions/mean_length": 1498.5417073567708,
      "completions/mean_terminated_length": 1070.8243408203125,
      "completions/min_length": 130.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.7488,
      "grad_norm": 0.10053375363349915,
      "kl": 0.030309041341145832,
      "learning_rate": 1.0256410256410255e-06,
      "loss": 0.0379,
      "num_tokens": 21700377.0,
      "reward": 1.8506944974263508,
      "reward_std": 0.1034339318672816,
      "rewards/accuracy_reward/mean": 0.9097222288449606,
      "rewards/accuracy_reward/std": 0.2692435334126155,
      "rewards/format_reward/mean": 0.9409722288449606,
      "rewards/format_reward/std": 0.18527215222517648,
      "step": 117
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 8080.333333333333,
      "completions/max_terminated_length": 8074.0,
      "completions/mean_length": 1560.0416666666667,
      "completions/mean_terminated_length": 1113.6520182291667,
      "completions/min_length": 227.0,
      "completions/min_terminated_length": 227.0,
      "epoch": 0.768,
      "grad_norm": 0.09212761372327805,
      "kl": 0.033299763997395836,
      "learning_rate": 9.487179487179486e-07,
      "loss": 0.0356,
      "num_tokens": 22176393.0,
      "reward": 1.90625003973643,
      "reward_std": 0.09762249638636906,
      "rewards/accuracy_reward/mean": 0.9583333333333334,
      "rewards/accuracy_reward/std": 0.197714701294899,
      "rewards/format_reward/mean": 0.9479166666666666,
      "rewards/format_reward/std": 0.1926976442337036,
      "step": 120
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.052083333333333336,
      "completions/max_length": 7240.0,
      "completions/max_terminated_length": 6940.666666666667,
      "completions/mean_length": 1487.6284993489583,
      "completions/mean_terminated_length": 1119.332051595052,
      "completions/min_length": 180.66666666666666,
      "completions/min_terminated_length": 180.66666666666666,
      "epoch": 0.7872,
      "grad_norm": 0.1242799237370491,
      "kl": 0.03277587890625,
      "learning_rate": 8.717948717948718e-07,
      "loss": 0.0342,
      "num_tokens": 22636012.0,
      "reward": 1.857638915379842,
      "reward_std": 0.10882141441106796,
      "rewards/accuracy_reward/mean": 0.9166666666666666,
      "rewards/accuracy_reward/std": 0.23966746032238007,
      "rewards/format_reward/mean": 0.9409722288449606,
      "rewards/format_reward/std": 0.18527215222517648,
      "step": 123
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.06597222222222221,
      "completions/max_length": 8192.0,
      "completions/max_terminated_length": 6789.333333333333,
      "completions/mean_length": 1534.1597493489583,
      "completions/mean_terminated_length": 1059.709248860677,
      "completions/min_length": 198.0,
      "completions/min_terminated_length": 198.0,
      "epoch": 0.8064,
      "grad_norm": 0.09035886079072952,
      "kl": 0.034245808919270836,
      "learning_rate": 7.948717948717948e-07,
      "loss": 0.0458,
      "num_tokens": 23103896.0,
      "reward": 1.8854167064030964,
      "reward_std": 0.1067802719771862,
      "rewards/accuracy_reward/mean": 0.9583333333333334,
      "rewards/accuracy_reward/std": 0.18493476261695227,
      "rewards/format_reward/mean": 0.9270833134651184,
      "rewards/format_reward/std": 0.2387408266464869,
      "step": 126
    }
  ],
  "logging_steps": 3,
  "max_steps": 156,
  "num_input_tokens_seen": 23493835,
  "num_train_epochs": 1,
  "save_steps": 32,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}