{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.78125,
  "eval_steps": 500,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 914.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1183.0,
      "completions/max_terminated_length": 1183.0,
      "completions/mean_length": 914.375,
      "completions/mean_terminated_length": 914.375,
      "completions/min_length": 538.0,
      "completions/min_terminated_length": 538.0,
      "epoch": 0.00390625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.5925750732421875,
      "kl": 0.0005806067929370329,
      "learning_rate": 0.0,
      "loss": 0.0,
      "num_tokens": 48251.0,
      "reward": 0.17561426758766174,
      "reward_std": 0.5293996334075928,
      "rewards/reward_function/mean": 0.17561426758766174,
      "rewards/reward_function/std": 0.5293996930122375,
      "step": 1
    },
    {
      "completion_length": 939.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1141.0,
      "completions/max_terminated_length": 1141.0,
      "completions/mean_length": 939.875,
      "completions/mean_terminated_length": 939.875,
      "completions/min_length": 781.0,
      "completions/min_terminated_length": 781.0,
      "epoch": 0.0078125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4351307153701782,
      "kl": 0.00043600249045994133,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0,
      "num_tokens": 96706.0,
      "reward": 0.07213374972343445,
      "reward_std": 0.6617448329925537,
      "rewards/reward_function/mean": 0.07213374972343445,
      "rewards/reward_function/std": 0.6617448925971985,
      "step": 2
    },
    {
      "completion_length": 957.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1241.0,
      "completions/max_terminated_length": 1241.0,
      "completions/mean_length": 957.625,
      "completions/mean_terminated_length": 957.625,
      "completions/min_length": 684.0,
      "completions/min_terminated_length": 684.0,
      "epoch": 0.01171875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.274860680103302,
      "kl": 0.0005422348549473099,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0,
      "num_tokens": 145303.0,
      "reward": 0.1806439459323883,
      "reward_std": 0.5124807357788086,
      "rewards/reward_function/mean": 0.1806439459323883,
      "rewards/reward_function/std": 0.5124807357788086,
      "step": 3
    },
    {
      "completion_length": 918.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1079.0,
      "completions/max_terminated_length": 1079.0,
      "completions/mean_length": 918.0,
      "completions/mean_terminated_length": 918.0,
      "completions/min_length": 628.0,
      "completions/min_terminated_length": 628.0,
      "epoch": 0.015625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33157476782798767,
      "kl": 0.00042391348688397557,
      "learning_rate": 1.5e-06,
      "loss": 0.0,
      "num_tokens": 193583.0,
      "reward": 0.07469659298658371,
      "reward_std": 0.6634011268615723,
      "rewards/reward_function/mean": 0.07469659298658371,
      "rewards/reward_function/std": 0.6634011268615723,
      "step": 4
    },
    {
      "completion_length": 1034.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1189.0,
      "completions/max_terminated_length": 1189.0,
      "completions/mean_length": 1034.5,
      "completions/mean_terminated_length": 1034.5,
      "completions/min_length": 887.0,
      "completions/min_terminated_length": 887.0,
      "epoch": 0.01953125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2924421429634094,
      "kl": 0.00045540891733253375,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0,
      "num_tokens": 249859.0,
      "reward": 0.016167104244232178,
      "reward_std": 0.6689337491989136,
      "rewards/reward_function/mean": 0.016167104244232178,
      "rewards/reward_function/std": 0.6689338088035583,
      "step": 5
    },
    {
      "completion_length": 876.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1272.0,
      "completions/max_terminated_length": 1272.0,
      "completions/mean_length": 876.0,
      "completions/mean_terminated_length": 876.0,
      "completions/min_length": 645.0,
      "completions/min_terminated_length": 645.0,
      "epoch": 0.0234375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4938257038593292,
      "kl": 0.0005481051848619245,
      "learning_rate": 2.5e-06,
      "loss": 0.0,
      "num_tokens": 304867.0,
      "reward": 0.0005262494087219238,
      "reward_std": 0.6533803343772888,
      "rewards/reward_function/mean": 0.0005262494087219238,
      "rewards/reward_function/std": 0.6533803343772888,
      "step": 6
    },
    {
      "completion_length": 1123.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1490.0,
      "completions/max_terminated_length": 1490.0,
      "completions/mean_length": 1123.625,
      "completions/mean_terminated_length": 1123.625,
      "completions/min_length": 184.0,
      "completions/min_terminated_length": 184.0,
      "epoch": 0.02734375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.5299341082572937,
      "kl": 0.0005529591653612442,
      "learning_rate": 3e-06,
      "loss": 0.0,
      "num_tokens": 361856.0,
      "reward": -0.2599995732307434,
      "reward_std": 0.7911489009857178,
      "rewards/reward_function/mean": -0.2599995732307434,
      "rewards/reward_function/std": 0.7911489605903625,
      "step": 7
    },
    {
      "completion_length": 1054.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1215.0,
      "completions/max_terminated_length": 1215.0,
      "completions/mean_length": 1054.5,
      "completions/mean_terminated_length": 1054.5,
      "completions/min_length": 935.0,
      "completions/min_terminated_length": 935.0,
      "epoch": 0.03125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29235586524009705,
      "kl": 0.0004442112476681359,
      "learning_rate": 3.5e-06,
      "loss": 0.0,
      "num_tokens": 418292.0,
      "reward": -0.27060431241989136,
      "reward_std": 0.7798133492469788,
      "rewards/reward_function/mean": -0.27060431241989136,
      "rewards/reward_function/std": 0.7798133492469788,
      "step": 8
    },
    {
      "completion_length": 1041.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1649.0,
      "completions/max_terminated_length": 1649.0,
      "completions/mean_length": 1041.25,
      "completions/mean_terminated_length": 1041.25,
      "completions/min_length": 694.0,
      "completions/min_terminated_length": 694.0,
      "epoch": 0.03515625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29960209131240845,
      "kl": 0.0004553622711682692,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0,
      "num_tokens": 463294.0,
      "reward": -0.09699174016714096,
      "reward_std": 0.7479231357574463,
      "rewards/reward_function/mean": -0.09699174016714096,
      "rewards/reward_function/std": 0.7479231357574463,
      "step": 9
    },
    {
      "completion_length": 986.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1359.0,
      "completions/max_terminated_length": 1359.0,
      "completions/mean_length": 986.375,
      "completions/mean_terminated_length": 986.375,
      "completions/min_length": 445.0,
      "completions/min_terminated_length": 445.0,
      "epoch": 0.0390625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.44665637612342834,
      "kl": 0.0005360567811294459,
      "learning_rate": 4.5e-06,
      "loss": 0.0,
      "num_tokens": 507857.0,
      "reward": 0.07969788461923599,
      "reward_std": 0.6664658784866333,
      "rewards/reward_function/mean": 0.07969788461923599,
      "rewards/reward_function/std": 0.6664658784866333,
      "step": 10
    },
    {
      "completion_length": 925.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1388.0,
      "completions/max_terminated_length": 1388.0,
      "completions/mean_length": 925.875,
      "completions/mean_terminated_length": 925.875,
      "completions/min_length": 598.0,
      "completions/min_terminated_length": 598.0,
      "epoch": 0.04296875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4459737241268158,
      "kl": 0.0005316538081387989,
      "learning_rate": 5e-06,
      "loss": 0.0,
      "num_tokens": 551936.0,
      "reward": -0.1024932935833931,
      "reward_std": 0.7433716058731079,
      "rewards/reward_function/mean": -0.1024932935833931,
      "rewards/reward_function/std": 0.7433716058731079,
      "step": 11
    },
    {
      "completion_length": 1036.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1326.0,
      "completions/max_terminated_length": 1326.0,
      "completions/mean_length": 1036.875,
      "completions/mean_terminated_length": 1036.875,
      "completions/min_length": 856.0,
      "completions/min_terminated_length": 856.0,
      "epoch": 0.046875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2756291329860687,
      "kl": 0.00041996776417363435,
      "learning_rate": 4.99847706754774e-06,
      "loss": 0.0,
      "num_tokens": 596903.0,
      "reward": 0.36005687713623047,
      "reward_std": 0.2220204770565033,
      "rewards/reward_function/mean": 0.36005687713623047,
      "rewards/reward_function/std": 0.2220204472541809,
      "step": 12
    },
    {
      "completion_length": 1167.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1468.0,
      "completions/max_terminated_length": 1468.0,
      "completions/mean_length": 1167.25,
      "completions/mean_terminated_length": 1167.25,
      "completions/min_length": 614.0,
      "completions/min_terminated_length": 614.0,
      "epoch": 0.05078125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3536391854286194,
      "kl": 0.0005555427414947189,
      "learning_rate": 4.993910125649561e-06,
      "loss": 0.0,
      "num_tokens": 650185.0,
      "reward": -0.10552466660737991,
      "reward_std": 0.7408891320228577,
      "rewards/reward_function/mean": -0.10552466660737991,
      "rewards/reward_function/std": 0.7408891320228577,
      "step": 13
    },
    {
      "completion_length": 978.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1308.0,
      "completions/max_terminated_length": 1308.0,
      "completions/mean_length": 978.75,
      "completions/mean_terminated_length": 978.75,
      "completions/min_length": 612.0,
      "completions/min_terminated_length": 612.0,
      "epoch": 0.0546875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3985886871814728,
      "kl": 0.000563694593438413,
      "learning_rate": 4.986304738420684e-06,
      "loss": 0.0,
      "num_tokens": 701959.0,
      "reward": -0.11333049833774567,
      "reward_std": 0.7342555522918701,
      "rewards/reward_function/mean": -0.11333049833774567,
      "rewards/reward_function/std": 0.7342556118965149,
      "step": 14
    },
    {
      "completion_length": 1253.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1684.0,
      "completions/max_terminated_length": 1684.0,
      "completions/mean_length": 1253.875,
      "completions/mean_terminated_length": 1253.875,
      "completions/min_length": 1103.0,
      "completions/min_terminated_length": 1103.0,
      "epoch": 0.05859375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29394710063934326,
      "kl": 0.0005726947129005566,
      "learning_rate": 4.975670171853926e-06,
      "loss": 0.0,
      "num_tokens": 755934.0,
      "reward": 0.19782808423042297,
      "reward_std": 0.5165873169898987,
      "rewards/reward_function/mean": 0.19782808423042297,
      "rewards/reward_function/std": 0.5165873169898987,
      "step": 15
    },
    {
      "completion_length": 1230.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1535.0,
      "completions/max_terminated_length": 1535.0,
      "completions/mean_length": 1230.0,
      "completions/mean_terminated_length": 1230.0,
      "completions/min_length": 934.0,
      "completions/min_terminated_length": 934.0,
      "epoch": 0.0625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.282123863697052,
      "kl": 0.000479752998217009,
      "learning_rate": 4.962019382530521e-06,
      "loss": 0.0,
      "num_tokens": 809718.0,
      "reward": -0.1809864193201065,
      "reward_std": 0.7100936770439148,
      "rewards/reward_function/mean": -0.1809864193201065,
      "rewards/reward_function/std": 0.7100936770439148,
      "step": 16
    },
    {
      "completion_length": 1231.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2350.0,
      "completions/max_terminated_length": 2350.0,
      "completions/mean_length": 1231.0,
      "completions/mean_terminated_length": 1231.0,
      "completions/min_length": 907.0,
      "completions/min_terminated_length": 907.0,
      "epoch": 0.06640625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2648545503616333,
      "kl": 0.0004779834271175787,
      "learning_rate": 4.9453690018345144e-06,
      "loss": 0.0,
      "num_tokens": 868222.0,
      "reward": -0.18590302765369415,
      "reward_std": 0.7067751884460449,
      "rewards/reward_function/mean": -0.18590302765369415,
      "rewards/reward_function/std": 0.7067751884460449,
      "step": 17
    },
    {
      "completion_length": 1052.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1399.0,
      "completions/max_terminated_length": 1399.0,
      "completions/mean_length": 1052.5,
      "completions/mean_terminated_length": 1052.5,
      "completions/min_length": 626.0,
      "completions/min_terminated_length": 626.0,
      "epoch": 0.0703125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.30565333366394043,
      "kl": 0.0004614779900293797,
      "learning_rate": 4.925739315689991e-06,
      "loss": 0.0,
      "num_tokens": 925298.0,
      "reward": -0.2848556339740753,
      "reward_std": 0.764528214931488,
      "rewards/reward_function/mean": -0.2848556339740753,
      "rewards/reward_function/std": 0.7645282745361328,
      "step": 18
    },
    {
      "completion_length": 1201.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1488.0,
      "completions/max_terminated_length": 1488.0,
      "completions/mean_length": 1201.75,
      "completions/mean_terminated_length": 1201.75,
      "completions/min_length": 985.0,
      "completions/min_terminated_length": 985.0,
      "epoch": 0.07421875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2542005181312561,
      "kl": 0.00046427045890595764,
      "learning_rate": 4.903154239845798e-06,
      "loss": 0.0,
      "num_tokens": 983568.0,
      "reward": 0.26851925253868103,
      "reward_std": 0.292377233505249,
      "rewards/reward_function/mean": 0.26851925253868103,
      "rewards/reward_function/std": 0.29237720370292664,
      "step": 19
    },
    {
      "completion_length": 1270.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1647.0,
      "completions/max_terminated_length": 1647.0,
      "completions/mean_length": 1270.0,
      "completions/mean_terminated_length": 1270.0,
      "completions/min_length": 930.0,
      "completions/min_terminated_length": 930.0,
      "epoch": 0.078125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.26307252049446106,
      "kl": 0.0004189292449154891,
      "learning_rate": 4.8776412907378845e-06,
      "loss": 0.0,
      "num_tokens": 1042384.0,
      "reward": 0.07484199106693268,
      "reward_std": 0.6634507179260254,
      "rewards/reward_function/mean": 0.07484199106693268,
      "rewards/reward_function/std": 0.6634507179260254,
      "step": 20
    },
    {
      "completion_length": 997.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1256.0,
      "completions/max_terminated_length": 1256.0,
      "completions/mean_length": 997.375,
      "completions/mean_terminated_length": 997.375,
      "completions/min_length": 807.0,
      "completions/min_terminated_length": 807.0,
      "epoch": 0.08203125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.35027605295181274,
      "kl": 0.0005483183849719353,
      "learning_rate": 4.849231551964771e-06,
      "loss": 0.0,
      "num_tokens": 1099571.0,
      "reward": 0.2563348412513733,
      "reward_std": 0.5079569220542908,
      "rewards/reward_function/mean": 0.2563348412513733,
      "rewards/reward_function/std": 0.5079569220542908,
      "step": 21
    },
    {
      "completion_length": 947.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1186.0,
      "completions/max_terminated_length": 1186.0,
      "completions/mean_length": 947.25,
      "completions/mean_terminated_length": 947.25,
      "completions/min_length": 756.0,
      "completions/min_terminated_length": 756.0,
      "epoch": 0.0859375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3029741048812866,
      "kl": 0.0004957405471941456,
      "learning_rate": 4.817959636416969e-06,
      "loss": 0.0,
      "num_tokens": 1156357.0,
      "reward": -0.1654960811138153,
      "reward_std": 0.7127606272697449,
      "rewards/reward_function/mean": -0.1654960811138153,
      "rewards/reward_function/std": 0.7127606272697449,
      "step": 22
    },
    {
      "completion_length": 1005.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1239.0,
      "completions/max_terminated_length": 1239.0,
      "completions/mean_length": 1005.5,
      "completions/mean_terminated_length": 1005.5,
      "completions/min_length": 750.0,
      "completions/min_terminated_length": 750.0,
      "epoch": 0.08984375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33245348930358887,
      "kl": 0.0005613294633803889,
      "learning_rate": 4.783863644106502e-06,
      "loss": 0.0,
      "num_tokens": 1213609.0,
      "reward": 0.19226206839084625,
      "reward_std": 0.5118025541305542,
      "rewards/reward_function/mean": 0.19226206839084625,
      "rewards/reward_function/std": 0.5118025541305542,
      "step": 23
    },
    {
      "completion_length": 888.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1100.0,
      "completions/max_terminated_length": 1100.0,
      "completions/mean_length": 888.25,
      "completions/mean_terminated_length": 888.25,
      "completions/min_length": 629.0,
      "completions/min_terminated_length": 629.0,
      "epoch": 0.09375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34459948539733887,
      "kl": 0.0005078868562122807,
      "learning_rate": 4.746985115747918e-06,
      "loss": 0.0,
      "num_tokens": 1269923.0,
      "reward": 0.006138928234577179,
      "reward_std": 0.6435980200767517,
      "rewards/reward_function/mean": 0.006138928234577179,
      "rewards/reward_function/std": 0.6435979604721069,
      "step": 24
    },
    {
      "completion_length": 988.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1209.0,
      "completions/max_terminated_length": 1209.0,
      "completions/mean_length": 988.5,
      "completions/mean_terminated_length": 988.5,
      "completions/min_length": 650.0,
      "completions/min_terminated_length": 650.0,
      "epoch": 0.09765625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3348124027252197,
      "kl": 0.0005601778539130464,
      "learning_rate": 4.707368982147318e-06,
      "loss": 0.0,
      "num_tokens": 1316151.0,
      "reward": 0.01586177945137024,
      "reward_std": 0.6504648923873901,
      "rewards/reward_function/mean": 0.01586177945137024,
      "rewards/reward_function/std": 0.6504649519920349,
      "step": 25
    },
    {
      "completion_length": 1129.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1498.0,
      "completions/max_terminated_length": 1498.0,
      "completions/mean_length": 1129.25,
      "completions/mean_terminated_length": 1129.25,
      "completions/min_length": 867.0,
      "completions/min_terminated_length": 867.0,
      "epoch": 0.1015625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3254358470439911,
      "kl": 0.0005182483946555294,
      "learning_rate": 4.665063509461098e-06,
      "loss": 0.0,
      "num_tokens": 1363505.0,
      "reward": 0.10026979446411133,
      "reward_std": 0.679513156414032,
      "rewards/reward_function/mean": 0.10026979446411133,
      "rewards/reward_function/std": 0.6795132160186768,
      "step": 26
    },
    {
      "completion_length": 1122.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1480.0,
      "completions/max_terminated_length": 1480.0,
      "completions/mean_length": 1122.25,
      "completions/mean_terminated_length": 1122.25,
      "completions/min_length": 855.0,
      "completions/min_terminated_length": 855.0,
      "epoch": 0.10546875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.27222707867622375,
      "kl": 0.0005421710084192455,
      "learning_rate": 4.620120240391065e-06,
      "loss": 0.0,
      "num_tokens": 1410803.0,
      "reward": 0.018940787762403488,
      "reward_std": 0.6555877923965454,
      "rewards/reward_function/mean": 0.018940787762403488,
      "rewards/reward_function/std": 0.6555877923965454,
      "step": 27
    },
    {
      "completion_length": 956.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1243.0,
      "completions/max_terminated_length": 1243.0,
      "completions/mean_length": 956.5,
      "completions/mean_terminated_length": 956.5,
      "completions/min_length": 805.0,
      "completions/min_terminated_length": 805.0,
      "epoch": 0.109375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3711608648300171,
      "kl": 0.0004645055887522176,
      "learning_rate": 4.572593931387604e-06,
      "loss": 0.0,
      "num_tokens": 1456775.0,
      "reward": 0.3047725558280945,
      "reward_std": 0.2678278088569641,
      "rewards/reward_function/mean": 0.3047725558280945,
      "rewards/reward_function/std": 0.2678278088569641,
      "step": 28
    },
    {
      "completion_length": 998.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1578.0,
      "completions/max_terminated_length": 1578.0,
      "completions/mean_length": 998.0,
      "completions/mean_terminated_length": 998.0,
      "completions/min_length": 723.0,
      "completions/min_terminated_length": 723.0,
      "epoch": 0.11328125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3207095265388489,
      "kl": 0.0005635232664644718,
      "learning_rate": 4.522542485937369e-06,
      "loss": 0.0,
      "num_tokens": 1513687.0,
      "reward": -0.11509008705615997,
      "reward_std": 0.7328100800514221,
      "rewards/reward_function/mean": -0.11509008705615997,
      "rewards/reward_function/std": 0.7328100800514221,
      "step": 29
    },
    {
      "completion_length": 1042.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1424.0,
      "completions/max_terminated_length": 1424.0,
      "completions/mean_length": 1042.5,
      "completions/mean_terminated_length": 1042.5,
      "completions/min_length": 839.0,
      "completions/min_terminated_length": 839.0,
      "epoch": 0.1171875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3066699504852295,
      "kl": 0.0005217162106418982,
      "learning_rate": 4.470026884016805e-06,
      "loss": 0.0,
      "num_tokens": 1570955.0,
      "reward": -0.1107390969991684,
      "reward_std": 0.7364593148231506,
      "rewards/reward_function/mean": -0.1107390969991684,
      "rewards/reward_function/std": 0.7364593148231506,
      "step": 30
    },
    {
      "completion_length": 1029.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1752.0,
      "completions/max_terminated_length": 1752.0,
      "completions/mean_length": 1029.875,
      "completions/mean_terminated_length": 1029.875,
      "completions/min_length": 737.0,
      "completions/min_terminated_length": 737.0,
      "epoch": 0.12109375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2864665985107422,
      "kl": 0.0004613926066667773,
      "learning_rate": 4.415111107797445e-06,
      "loss": 0.0,
      "num_tokens": 1628122.0,
      "reward": 0.24289178848266602,
      "reward_std": 0.5025007128715515,
      "rewards/reward_function/mean": 0.24289178848266602,
      "rewards/reward_function/std": 0.5025007128715515,
      "step": 31
    },
    {
      "completion_length": 904.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1088.0,
      "completions/max_terminated_length": 1088.0,
      "completions/mean_length": 904.5,
      "completions/mean_terminated_length": 904.5,
      "completions/min_length": 605.0,
      "completions/min_terminated_length": 605.0,
      "epoch": 0.125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34721776843070984,
      "kl": 0.00046865284093655646,
      "learning_rate": 4.357862063693486e-06,
      "loss": 0.0,
      "num_tokens": 1684286.0,
      "reward": 0.4172564446926117,
      "reward_std": 0.015801435336470604,
      "rewards/reward_function/mean": 0.4172564446926117,
      "rewards/reward_function/std": 0.015801437199115753,
      "step": 32
    },
    {
      "completion_length": 748.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1067.0,
      "completions/max_terminated_length": 1067.0,
      "completions/mean_length": 748.125,
      "completions/mean_terminated_length": 748.125,
      "completions/min_length": 345.0,
      "completions/min_terminated_length": 345.0,
      "epoch": 0.12890625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.5278727412223816,
      "kl": 0.0005870972745469771,
      "learning_rate": 4.2983495008466285e-06,
      "loss": 0.0,
      "num_tokens": 1734087.0,
      "reward": -0.012617290019989014,
      "reward_std": 0.6455253958702087,
      "rewards/reward_function/mean": -0.012617290019989014,
      "rewards/reward_function/std": 0.6455253958702087,
      "step": 33
    },
    {
      "completion_length": 717.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 824.0,
      "completions/max_terminated_length": 824.0,
      "completions/mean_length": 717.375,
      "completions/mean_terminated_length": 717.375,
      "completions/min_length": 624.0,
      "completions/min_terminated_length": 624.0,
      "epoch": 0.1328125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34698522090911865,
      "kl": 0.0004982690297765657,
      "learning_rate": 4.236645926147493e-06,
      "loss": 0.0,
      "num_tokens": 1783642.0,
      "reward": 0.15846197307109833,
      "reward_std": 0.5145913362503052,
      "rewards/reward_function/mean": 0.15846197307109833,
      "rewards/reward_function/std": 0.5145912766456604,
      "step": 34
    },
    {
      "completion_length": 594.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 874.0,
      "completions/max_terminated_length": 874.0,
      "completions/mean_length": 594.625,
      "completions/mean_terminated_length": 594.625,
      "completions/min_length": 134.0,
      "completions/min_terminated_length": 134.0,
      "epoch": 0.13671875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.5477731823921204,
      "kl": 0.0005064220604253933,
      "learning_rate": 4.172826515897146e-06,
      "loss": 0.0,
      "num_tokens": 1832215.0,
      "reward": -0.2889060974121094,
      "reward_std": 0.7602130770683289,
      "rewards/reward_function/mean": -0.2889060974121094,
      "rewards/reward_function/std": 0.7602130770683289,
      "step": 35
    },
    {
      "completion_length": 835.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1236.0,
      "completions/max_terminated_length": 1236.0,
      "completions/mean_length": 835.625,
      "completions/mean_terminated_length": 835.625,
      "completions/min_length": 624.0,
      "completions/min_terminated_length": 624.0,
      "epoch": 0.140625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3209806978702545,
      "kl": 0.0005167589188204147,
      "learning_rate": 4.106969024216348e-06,
      "loss": 0.0,
      "num_tokens": 1882716.0,
      "reward": 0.23565331101417542,
      "reward_std": 0.4997692406177521,
      "rewards/reward_function/mean": 0.23565331101417542,
      "rewards/reward_function/std": 0.4997692406177521,
      "step": 36
    },
    {
      "completion_length": 1067.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1422.0,
      "completions/max_terminated_length": 1422.0,
      "completions/mean_length": 1067.125,
      "completions/mean_terminated_length": 1067.125,
      "completions/min_length": 886.0,
      "completions/min_terminated_length": 886.0,
      "epoch": 0.14453125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31051820516586304,
      "kl": 0.0005310940687195398,
      "learning_rate": 4.039153688314146e-06,
      "loss": 0.0,
      "num_tokens": 1933333.0,
      "reward": 0.3604900538921356,
      "reward_std": 0.22474592924118042,
      "rewards/reward_function/mean": 0.3604900538921356,
      "rewards/reward_function/std": 0.22474592924118042,
      "step": 37
    },
    {
      "completion_length": 983.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1199.0,
      "completions/max_terminated_length": 1199.0,
      "completions/mean_length": 983.0,
      "completions/mean_terminated_length": 983.0,
      "completions/min_length": 791.0,
      "completions/min_terminated_length": 791.0,
      "epoch": 0.1484375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.28662100434303284,
      "kl": 0.0004848391836276278,
      "learning_rate": 3.969463130731183e-06,
      "loss": 0.0,
      "num_tokens": 1983277.0,
      "reward": -0.09930499643087387,
      "reward_std": 0.7459418773651123,
      "rewards/reward_function/mean": -0.09930499643087387,
      "rewards/reward_function/std": 0.7459418773651123,
      "step": 38
    },
    {
      "completion_length": 1083.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1410.0,
      "completions/max_terminated_length": 1410.0,
      "completions/mean_length": 1083.375,
      "completions/mean_terminated_length": 1083.375,
      "completions/min_length": 937.0,
      "completions/min_terminated_length": 937.0,
      "epoch": 0.15234375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2731831967830658,
      "kl": 0.0004829144090763293,
      "learning_rate": 3.897982258676867e-06,
      "loss": 0.0,
      "num_tokens": 2034024.0,
      "reward": -0.18277299404144287,
      "reward_std": 0.7117413878440857,
      "rewards/reward_function/mean": -0.18277299404144287,
      "rewards/reward_function/std": 0.7117413878440857,
      "step": 39
    },
    {
      "completion_length": 1012.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1179.0,
      "completions/max_terminated_length": 1179.0,
      "completions/mean_length": 1012.875,
      "completions/mean_terminated_length": 1012.875,
      "completions/min_length": 803.0,
      "completions/min_terminated_length": 803.0,
      "epoch": 0.15625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.24166430532932281,
      "kl": 0.00040847257332643494,
      "learning_rate": 3.824798160583012e-06,
      "loss": 0.0,
      "num_tokens": 2084207.0,
      "reward": 0.37631669640541077,
      "reward_std": 0.19093753397464752,
      "rewards/reward_function/mean": 0.37631669640541077,
      "rewards/reward_function/std": 0.19093753397464752,
      "step": 40
    },
    {
      "completion_length": 928.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1163.0,
      "completions/max_terminated_length": 1163.0,
      "completions/mean_length": 928.875,
      "completions/mean_terminated_length": 928.875,
      "completions/min_length": 579.0,
      "completions/min_terminated_length": 579.0,
      "epoch": 0.16015625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.32170408964157104,
      "kl": 0.000433544373663608,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0,
      "num_tokens": 2131326.0,
      "reward": 0.17959481477737427,
      "reward_std": 0.5257959365844727,
      "rewards/reward_function/mean": 0.17959481477737427,
      "rewards/reward_function/std": 0.5257958769798279,
      "step": 41
    },
    {
      "completion_length": 941.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1077.0,
      "completions/max_terminated_length": 1077.0,
      "completions/mean_length": 941.625,
      "completions/mean_terminated_length": 941.625,
      "completions/min_length": 762.0,
      "completions/min_terminated_length": 762.0,
      "epoch": 0.1640625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.40420642495155334,
      "kl": 0.0004869201438850723,
      "learning_rate": 3.6736789069647273e-06,
      "loss": 0.0,
      "num_tokens": 2178547.0,
      "reward": 0.3620530366897583,
      "reward_std": 0.2077617347240448,
      "rewards/reward_function/mean": 0.3620530366897583,
      "rewards/reward_function/std": 0.207761749625206,
      "step": 42
    },
    {
      "completion_length": 1078.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1457.0,
      "completions/max_terminated_length": 1457.0,
      "completions/mean_length": 1078.625,
      "completions/mean_terminated_length": 1078.625,
      "completions/min_length": 908.0,
      "completions/min_terminated_length": 908.0,
      "epoch": 0.16796875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3174184262752533,
      "kl": 0.0005323004734236747,
      "learning_rate": 3.595927866972694e-06,
      "loss": 0.0,
      "num_tokens": 2226864.0,
      "reward": -0.10311660170555115,
      "reward_std": 0.7427304983139038,
      "rewards/reward_function/mean": -0.10311660170555115,
      "rewards/reward_function/std": 0.7427304983139038,
      "step": 43
    },
    {
      "completion_length": 920.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1060.0,
      "completions/max_terminated_length": 1060.0,
      "completions/mean_length": 920.125,
      "completions/mean_terminated_length": 920.125,
      "completions/min_length": 758.0,
      "completions/min_terminated_length": 758.0,
      "epoch": 0.171875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3262939751148224,
      "kl": 0.0004862127752858214,
      "learning_rate": 3.516841607689501e-06,
      "loss": 0.0,
      "num_tokens": 2273913.0,
      "reward": -0.10695745050907135,
      "reward_std": 0.7395932078361511,
      "rewards/reward_function/mean": -0.10695745050907135,
      "rewards/reward_function/std": 0.7395932078361511,
      "step": 44
    },
    {
      "completion_length": 806.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 943.0,
      "completions/max_terminated_length": 943.0,
      "completions/mean_length": 806.875,
      "completions/mean_terminated_length": 806.875,
      "completions/min_length": 462.0,
      "completions/min_terminated_length": 462.0,
      "epoch": 0.17578125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.37206342816352844,
      "kl": 0.00043304747669026256,
      "learning_rate": 3.436516483539781e-06,
      "loss": 0.0,
      "num_tokens": 2326376.0,
      "reward": 0.08102156221866608,
      "reward_std": 0.6678274869918823,
      "rewards/reward_function/mean": 0.08102156221866608,
      "rewards/reward_function/std": 0.6678274869918823,
      "step": 45
    },
    {
      "completion_length": 912.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1138.0,
      "completions/max_terminated_length": 1138.0,
      "completions/mean_length": 912.125,
      "completions/mean_terminated_length": 912.125,
      "completions/min_length": 768.0,
      "completions/min_terminated_length": 768.0,
      "epoch": 0.1796875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31978893280029297,
      "kl": 0.0003971458208980039,
      "learning_rate": 3.3550503583141726e-06,
      "loss": 0.0,
      "num_tokens": 2379681.0,
      "reward": -0.09052923321723938,
      "reward_std": 0.7533401846885681,
      "rewards/reward_function/mean": -0.09052923321723938,
      "rewards/reward_function/std": 0.7533401846885681,
      "step": 46
    },
    {
      "completion_length": 906.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1003.0,
      "completions/max_terminated_length": 1003.0,
      "completions/mean_length": 906.375,
      "completions/mean_terminated_length": 906.375,
      "completions/min_length": 725.0,
      "completions/min_terminated_length": 725.0,
      "epoch": 0.18359375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3475249707698822,
      "kl": 0.00039272347203223035,
      "learning_rate": 3.272542485937369e-06,
      "loss": 0.0,
      "num_tokens": 2432940.0,
      "reward": 0.455754816532135,
      "reward_std": 0.018690278753638268,
      "rewards/reward_function/mean": 0.455754816532135,
      "rewards/reward_function/std": 0.01869027502834797,
      "step": 47
    },
    {
      "completion_length": 849.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 953.0,
      "completions/max_terminated_length": 953.0,
      "completions/mean_length": 849.875,
      "completions/mean_terminated_length": 849.875,
      "completions/min_length": 757.0,
      "completions/min_terminated_length": 757.0,
      "epoch": 0.1875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3427131772041321,
      "kl": 0.0004229802143527195,
      "learning_rate": 3.189093389542498e-06,
      "loss": 0.0,
      "num_tokens": 2485747.0,
      "reward": 0.2003743201494217,
      "reward_std": 0.5182459950447083,
      "rewards/reward_function/mean": 0.2003743201494217,
      "rewards/reward_function/std": 0.5182459950447083,
      "step": 48
    },
    {
      "completion_length": 890.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1381.0,
      "completions/max_terminated_length": 1381.0,
      "completions/mean_length": 890.25,
      "completions/mean_terminated_length": 890.25,
      "completions/min_length": 201.0,
      "completions/min_terminated_length": 201.0,
      "epoch": 0.19140625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.5939275026321411,
      "kl": 0.0005514047734322958,
      "learning_rate": 3.1048047389991693e-06,
      "loss": 0.0,
      "num_tokens": 2539709.0,
      "reward": -0.11278533935546875,
      "reward_std": 0.7347214221954346,
      "rewards/reward_function/mean": -0.11278533935546875,
      "rewards/reward_function/std": 0.7347214221954346,
      "step": 49
    },
    {
      "completion_length": 1022.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1454.0,
      "completions/max_terminated_length": 1454.0,
      "completions/mean_length": 1022.125,
      "completions/mean_terminated_length": 1022.125,
      "completions/min_length": 797.0,
      "completions/min_terminated_length": 797.0,
      "epoch": 0.1953125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.339578777551651,
      "kl": 0.00044445088860811666,
      "learning_rate": 3.019779227044398e-06,
      "loss": 0.0,
      "num_tokens": 2594726.0,
      "reward": 0.05837273597717285,
      "reward_std": 0.653872549533844,
      "rewards/reward_function/mean": 0.05837273597717285,
      "rewards/reward_function/std": 0.653872549533844,
      "step": 50
    },
    {
      "completion_length": 872.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1108.0,
      "completions/max_terminated_length": 1108.0,
      "completions/mean_length": 872.25,
      "completions/mean_terminated_length": 872.25,
      "completions/min_length": 447.0,
      "completions/min_terminated_length": 447.0,
      "epoch": 0.19921875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.0404114723205566,
      "kl": 0.0005189829680602998,
      "learning_rate": 2.9341204441673267e-06,
      "loss": 0.0,
      "num_tokens": 2648544.0,
      "reward": 0.24048176407814026,
      "reward_std": 0.5014303922653198,
      "rewards/reward_function/mean": 0.24048176407814026,
      "rewards/reward_function/std": 0.5014303922653198,
      "step": 51
    },
    {
      "completion_length": 967.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1227.0,
      "completions/max_terminated_length": 1227.0,
      "completions/mean_length": 967.125,
      "completions/mean_terminated_length": 967.125,
      "completions/min_length": 698.0,
      "completions/min_terminated_length": 698.0,
      "epoch": 0.203125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29855239391326904,
      "kl": 0.0004989826556993648,
      "learning_rate": 2.847932752400164e-06,
      "loss": 0.0,
      "num_tokens": 2703121.0,
      "reward": -0.1780097782611847,
      "reward_std": 0.7012866139411926,
      "rewards/reward_function/mean": -0.1780097782611847,
      "rewards/reward_function/std": 0.7012866139411926,
      "step": 52
    },
    {
      "completion_length": 1005.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1206.0,
      "completions/max_terminated_length": 1206.0,
      "completions/mean_length": 1005.625,
      "completions/mean_terminated_length": 1005.625,
      "completions/min_length": 864.0,
      "completions/min_terminated_length": 864.0,
      "epoch": 0.20703125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2760545015335083,
      "kl": 0.0004165990831097588,
      "learning_rate": 2.761321158169134e-06,
      "loss": 0.0,
      "num_tokens": 2759446.0,
      "reward": 0.24998876452445984,
      "reward_std": 0.505235493183136,
      "rewards/reward_function/mean": 0.24998876452445984,
      "rewards/reward_function/std": 0.5052355527877808,
      "step": 53
    },
    {
      "completion_length": 1015.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1532.0,
      "completions/max_terminated_length": 1532.0,
      "completions/mean_length": 1015.75,
      "completions/mean_terminated_length": 1015.75,
      "completions/min_length": 660.0,
      "completions/min_terminated_length": 660.0,
      "epoch": 0.2109375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3195282220840454,
      "kl": 0.0004872979479841888,
      "learning_rate": 2.6743911843603134e-06,
      "loss": 0.0,
      "num_tokens": 2815852.0,
      "reward": -0.2854868173599243,
      "reward_std": 0.7638800740242004,
      "rewards/reward_function/mean": -0.2854868173599243,
      "rewards/reward_function/std": 0.7638801336288452,
      "step": 54
    },
    {
      "completion_length": 1141.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1444.0,
      "completions/max_terminated_length": 1444.0,
      "completions/mean_length": 1141.125,
      "completions/mean_terminated_length": 1141.125,
      "completions/min_length": 883.0,
      "completions/min_terminated_length": 883.0,
      "epoch": 0.21484375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2491772174835205,
      "kl": 0.0004129245280637406,
      "learning_rate": 2.587248741756253e-06,
      "loss": 0.0,
      "num_tokens": 2873261.0,
      "reward": -0.18247561156749725,
      "reward_std": 0.7091330289840698,
      "rewards/reward_function/mean": -0.18247561156749725,
      "rewards/reward_function/std": 0.7091330885887146,
      "step": 55
    },
    {
      "completion_length": 1048.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1244.0,
      "completions/max_terminated_length": 1244.0,
      "completions/mean_length": 1048.75,
      "completions/mean_terminated_length": 1048.75,
      "completions/min_length": 811.0,
      "completions/min_terminated_length": 811.0,
      "epoch": 0.21875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.24889753758907318,
      "kl": 0.00038504171243403107,
      "learning_rate": 2.5e-06,
      "loss": 0.0,
      "num_tokens": 2929931.0,
      "reward": 0.25053682923316956,
      "reward_std": 0.5054035186767578,
      "rewards/reward_function/mean": 0.25053682923316956,
      "rewards/reward_function/std": 0.5054035186767578,
      "step": 56
    },
    {
      "completion_length": 763.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1179.0,
      "completions/max_terminated_length": 1179.0,
      "completions/mean_length": 763.375,
      "completions/mean_terminated_length": 763.375,
      "completions/min_length": 93.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.22265625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.4766223430633545,
      "kl": 0.0004478975461097434,
      "learning_rate": 2.4127512582437486e-06,
      "loss": 0.0,
      "num_tokens": 2981454.0,
      "reward": 0.09302115440368652,
      "reward_std": 0.6747010946273804,
      "rewards/reward_function/mean": 0.09302115440368652,
      "rewards/reward_function/std": 0.6747010946273804,
      "step": 57
    },
    {
      "completion_length": 868.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1033.0,
      "completions/max_terminated_length": 1033.0,
      "completions/mean_length": 868.125,
      "completions/mean_terminated_length": 868.125,
      "completions/min_length": 728.0,
      "completions/min_terminated_length": 728.0,
      "epoch": 0.2265625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3186846673488617,
      "kl": 0.0003867191553581506,
      "learning_rate": 2.325608815639687e-06,
      "loss": -0.0,
      "num_tokens": 3033815.0,
      "reward": 0.4501683712005615,
      "reward_std": 0.012572327628731728,
      "rewards/reward_function/mean": 0.4501683712005615,
      "rewards/reward_function/std": 0.012572328560054302,
      "step": 58
    },
    {
      "completion_length": 850.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1053.0,
      "completions/max_terminated_length": 1053.0,
      "completions/mean_length": 850.625,
      "completions/mean_terminated_length": 850.625,
      "completions/min_length": 293.0,
      "completions/min_terminated_length": 293.0,
      "epoch": 0.23046875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4074990153312683,
      "kl": 0.0004984224360669032,
      "learning_rate": 2.238678841830867e-06,
      "loss": 0.0,
      "num_tokens": 3086036.0,
      "reward": -0.2775750756263733,
      "reward_std": 0.772331714630127,
      "rewards/reward_function/mean": -0.2775750756263733,
      "rewards/reward_function/std": 0.7723317742347717,
      "step": 59
    },
    {
      "completion_length": 824.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1076.0,
      "completions/max_terminated_length": 1076.0,
      "completions/mean_length": 824.0,
      "completions/mean_terminated_length": 824.0,
      "completions/min_length": 111.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.234375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.8724004030227661,
      "kl": 0.0008119349731714465,
      "learning_rate": 2.1520672475998374e-06,
      "loss": 0.0,
      "num_tokens": 3138044.0,
      "reward": -0.4572715163230896,
      "reward_std": 0.7490780353546143,
      "rewards/reward_function/mean": -0.4572715163230896,
      "rewards/reward_function/std": 0.7490780353546143,
      "step": 60
    },
    {
      "completion_length": 1087.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1528.0,
      "completions/max_terminated_length": 1528.0,
      "completions/mean_length": 1087.375,
      "completions/mean_terminated_length": 1087.375,
      "completions/min_length": 905.0,
      "completions/min_terminated_length": 905.0,
      "epoch": 0.23828125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2991656959056854,
      "kl": 0.00046085825306363404,
      "learning_rate": 2.0658795558326745e-06,
      "loss": 0.0,
      "num_tokens": 3192767.0,
      "reward": -0.007034048438072205,
      "reward_std": 0.6521883606910706,
      "rewards/reward_function/mean": -0.007034048438072205,
      "rewards/reward_function/std": 0.6521883606910706,
      "step": 61
    },
    {
      "completion_length": 1113.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1320.0,
      "completions/max_terminated_length": 1320.0,
      "completions/mean_length": 1113.5,
      "completions/mean_terminated_length": 1113.5,
      "completions/min_length": 815.0,
      "completions/min_terminated_length": 815.0,
      "epoch": 0.2421875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2844128906726837,
      "kl": 0.0004267314507160336,
      "learning_rate": 1.9802207729556023e-06,
      "loss": 0.0,
      "num_tokens": 3247699.0,
      "reward": 0.25297480821609497,
      "reward_std": 0.506567656993866,
      "rewards/reward_function/mean": 0.25297480821609497,
      "rewards/reward_function/std": 0.506567656993866,
      "step": 62
    },
    {
      "completion_length": 977.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1705.0,
      "completions/max_terminated_length": 1705.0,
      "completions/mean_length": 977.5,
      "completions/mean_terminated_length": 977.5,
      "completions/min_length": 177.0,
      "completions/min_terminated_length": 177.0,
      "epoch": 0.24609375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4520728886127472,
      "kl": 0.0006246675111469813,
      "learning_rate": 1.895195261000831e-06,
      "loss": 0.0,
      "num_tokens": 3301543.0,
      "reward": -0.12443238496780396,
      "reward_std": 0.7253407835960388,
      "rewards/reward_function/mean": -0.12443238496780396,
      "rewards/reward_function/std": 0.7253408432006836,
      "step": 63
    },
    {
      "completion_length": 1241.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2091.0,
      "completions/max_terminated_length": 2091.0,
      "completions/mean_length": 1241.875,
      "completions/mean_terminated_length": 1241.875,
      "completions/min_length": 840.0,
      "completions/min_terminated_length": 840.0,
      "epoch": 0.25,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.28819605708122253,
      "kl": 0.0004499929418670945,
      "learning_rate": 1.8109066104575023e-06,
      "loss": 0.0,
      "num_tokens": 3357502.0,
      "reward": -0.2883961498737335,
      "reward_std": 0.760738730430603,
      "rewards/reward_function/mean": -0.2883961498737335,
      "rewards/reward_function/std": 0.7607387900352478,
      "step": 64
    },
    {
      "completion_length": 969.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1196.0,
      "completions/max_terminated_length": 1196.0,
      "completions/mean_length": 969.875,
      "completions/mean_terminated_length": 969.875,
      "completions/min_length": 877.0,
      "completions/min_terminated_length": 877.0,
      "epoch": 0.25390625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3114400804042816,
      "kl": 0.0005156347615411505,
      "learning_rate": 1.7274575140626318e-06,
      "loss": 0.0,
      "num_tokens": 3410565.0,
      "reward": 0.06768127530813217,
      "reward_std": 0.6590126156806946,
      "rewards/reward_function/mean": 0.06768127530813217,
      "rewards/reward_function/std": 0.6590126156806946,
      "step": 65
    },
    {
      "completion_length": 1037.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1212.0,
      "completions/max_terminated_length": 1212.0,
      "completions/mean_length": 1037.5,
      "completions/mean_terminated_length": 1037.5,
      "completions/min_length": 834.0,
      "completions/min_terminated_length": 834.0,
      "epoch": 0.2578125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.28374701738357544,
      "kl": 0.0004491405125008896,
      "learning_rate": 1.6449496416858285e-06,
      "loss": 0.0,
      "num_tokens": 3464169.0,
      "reward": 0.28254395723342896,
      "reward_std": 0.27969422936439514,
      "rewards/reward_function/mean": 0.28254395723342896,
      "rewards/reward_function/std": 0.27969422936439514,
      "step": 66
    },
    {
      "completion_length": 945.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1096.0,
      "completions/max_terminated_length": 1096.0,
      "completions/mean_length": 945.125,
      "completions/mean_terminated_length": 945.125,
      "completions/min_length": 720.0,
      "completions/min_terminated_length": 720.0,
      "epoch": 0.26171875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3035137951374054,
      "kl": 0.0005614978290395811,
      "learning_rate": 1.56348351646022e-06,
      "loss": 0.0,
      "num_tokens": 3517034.0,
      "reward": 0.06603994220495224,
      "reward_std": 0.6580838561058044,
      "rewards/reward_function/mean": 0.06603994220495224,
      "rewards/reward_function/std": 0.6580838561058044,
      "step": 67
    },
    {
      "completion_length": 889.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1034.0,
      "completions/max_terminated_length": 1034.0,
      "completions/mean_length": 889.25,
      "completions/mean_terminated_length": 889.25,
      "completions/min_length": 761.0,
      "completions/min_terminated_length": 761.0,
      "epoch": 0.265625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31817343831062317,
      "kl": 0.0004019583066110499,
      "learning_rate": 1.4831583923105e-06,
      "loss": 0.0,
      "num_tokens": 3569452.0,
      "reward": -0.1899593323469162,
      "reward_std": 0.7019972801208496,
      "rewards/reward_function/mean": -0.1899593323469162,
      "rewards/reward_function/std": 0.7019972801208496,
      "step": 68
    },
    {
      "completion_length": 978.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1133.0,
      "completions/max_terminated_length": 1133.0,
      "completions/mean_length": 978.75,
      "completions/mean_terminated_length": 978.75,
      "completions/min_length": 788.0,
      "completions/min_terminated_length": 788.0,
      "epoch": 0.26953125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2677120566368103,
      "kl": 0.000421192504290957,
      "learning_rate": 1.4040721330273063e-06,
      "loss": 0.0,
      "num_tokens": 3622058.0,
      "reward": 0.26337215304374695,
      "reward_std": 0.5106496810913086,
      "rewards/reward_function/mean": 0.26337215304374695,
      "rewards/reward_function/std": 0.5106497406959534,
      "step": 69
    },
    {
      "completion_length": 1027.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1741.0,
      "completions/max_terminated_length": 1741.0,
      "completions/mean_length": 1027.5,
      "completions/mean_terminated_length": 1027.5,
      "completions/min_length": 785.0,
      "completions/min_terminated_length": 785.0,
      "epoch": 0.2734375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3545612096786499,
      "kl": 0.0004943152016494423,
      "learning_rate": 1.3263210930352737e-06,
      "loss": 0.0,
      "num_tokens": 3675054.0,
      "reward": 0.4391711354255676,
      "reward_std": 0.04140199348330498,
      "rewards/reward_function/mean": 0.4391711354255676,
      "rewards/reward_function/std": 0.041401997208595276,
      "step": 70
    },
    {
      "completion_length": 1002.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1330.0,
      "completions/max_terminated_length": 1330.0,
      "completions/mean_length": 1002.125,
      "completions/mean_terminated_length": 1002.125,
      "completions/min_length": 812.0,
      "completions/min_terminated_length": 812.0,
      "epoch": 0.27734375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31723299622535706,
      "kl": 0.0004918482736684382,
      "learning_rate": 1.2500000000000007e-06,
      "loss": 0.0,
      "num_tokens": 3727847.0,
      "reward": -0.09308577328920364,
      "reward_std": 0.751098096370697,
      "rewards/reward_function/mean": -0.09308577328920364,
      "rewards/reward_function/std": 0.7510981559753418,
      "step": 71
    },
    {
      "completion_length": 1050.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1413.0,
      "completions/max_terminated_length": 1413.0,
      "completions/mean_length": 1050.375,
      "completions/mean_terminated_length": 1050.375,
      "completions/min_length": 815.0,
      "completions/min_terminated_length": 815.0,
      "epoch": 0.28125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2855725884437561,
      "kl": 0.00046525982179446146,
      "learning_rate": 1.1752018394169882e-06,
      "loss": 0.0,
      "num_tokens": 3781026.0,
      "reward": 0.2761775255203247,
      "reward_std": 0.5158948302268982,
      "rewards/reward_function/mean": 0.2761775255203247,
      "rewards/reward_function/std": 0.5158948302268982,
      "step": 72
    },
    {
      "completion_length": 1050.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1197.0,
      "completions/max_terminated_length": 1197.0,
      "completions/mean_length": 1050.0,
      "completions/mean_terminated_length": 1050.0,
      "completions/min_length": 909.0,
      "completions/min_terminated_length": 909.0,
      "epoch": 0.28515625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4394182562828064,
      "kl": 0.0005783203814644367,
      "learning_rate": 1.1020177413231334e-06,
      "loss": 0.0,
      "num_tokens": 3836474.0,
      "reward": 0.24491167068481445,
      "reward_std": 0.5031947493553162,
      "rewards/reward_function/mean": 0.24491167068481445,
      "rewards/reward_function/std": 0.5031947493553162,
      "step": 73
    },
    {
      "completion_length": 1046.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1290.0,
      "completions/max_terminated_length": 1290.0,
      "completions/mean_length": 1046.375,
      "completions/mean_terminated_length": 1046.375,
      "completions/min_length": 837.0,
      "completions/min_terminated_length": 837.0,
      "epoch": 0.2890625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3345458507537842,
      "kl": 0.0005103340663481504,
      "learning_rate": 1.0305368692688175e-06,
      "loss": 0.0,
      "num_tokens": 3891893.0,
      "reward": 0.18502177298069,
      "reward_std": 0.5084241628646851,
      "rewards/reward_function/mean": 0.18502177298069,
      "rewards/reward_function/std": 0.5084241628646851,
      "step": 74
    },
    {
      "completion_length": 1180.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1387.0,
      "completions/max_terminated_length": 1387.0,
      "completions/mean_length": 1180.125,
      "completions/mean_terminated_length": 1180.125,
      "completions/min_length": 939.0,
      "completions/min_terminated_length": 939.0,
      "epoch": 0.29296875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2869652509689331,
      "kl": 0.0005238868543528952,
      "learning_rate": 9.608463116858544e-07,
      "loss": 0.0,
      "num_tokens": 3948382.0,
      "reward": 0.4311676621437073,
      "reward_std": 0.01504568662494421,
      "rewards/reward_function/mean": 0.4311676621437073,
      "rewards/reward_function/std": 0.01504568662494421,
      "step": 75
    },
    {
      "completion_length": 1176.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1854.0,
      "completions/max_terminated_length": 1854.0,
      "completions/mean_length": 1176.0,
      "completions/mean_terminated_length": 1176.0,
      "completions/min_length": 870.0,
      "completions/min_terminated_length": 870.0,
      "epoch": 0.296875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.28590667247772217,
      "kl": 0.000421482436649967,
      "learning_rate": 8.930309757836517e-07,
      "loss": 0.0,
      "num_tokens": 4004838.0,
      "reward": 0.37514999508857727,
      "reward_std": 0.18447871506214142,
      "rewards/reward_function/mean": 0.37514999508857727,
      "rewards/reward_function/std": 0.18447871506214142,
      "step": 76
    },
    {
      "completion_length": 836.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1274.0,
      "completions/max_terminated_length": 1274.0,
      "completions/mean_length": 836.25,
      "completions/mean_terminated_length": 836.25,
      "completions/min_length": 694.0,
      "completions/min_terminated_length": 694.0,
      "epoch": 0.30078125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3467836380004883,
      "kl": 0.0005228954396443442,
      "learning_rate": 8.271734841028553e-07,
      "loss": 0.0,
      "num_tokens": 4056272.0,
      "reward": 0.2929620146751404,
      "reward_std": 0.27145516872406006,
      "rewards/reward_function/mean": 0.2929620146751404,
      "rewards/reward_function/std": 0.27145513892173767,
      "step": 77
    },
    {
      "completion_length": 821.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1038.0,
      "completions/max_terminated_length": 1038.0,
      "completions/mean_length": 821.0,
      "completions/mean_terminated_length": 821.0,
      "completions/min_length": 692.0,
      "completions/min_terminated_length": 692.0,
      "epoch": 0.3046875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3295641839504242,
      "kl": 0.0004961359809385613,
      "learning_rate": 7.633540738525066e-07,
      "loss": 0.0,
      "num_tokens": 4107584.0,
      "reward": -0.0900302305817604,
      "reward_std": 0.6223264932632446,
      "rewards/reward_function/mean": -0.0900302305817604,
      "rewards/reward_function/std": 0.6223265528678894,
      "step": 78
    },
    {
      "completion_length": 975.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1555.0,
      "completions/max_terminated_length": 1555.0,
      "completions/mean_length": 975.25,
      "completions/mean_terminated_length": 975.25,
      "completions/min_length": 711.0,
      "completions/min_terminated_length": 711.0,
      "epoch": 0.30859375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3428508937358856,
      "kl": 0.0005168440329725854,
      "learning_rate": 7.016504991533727e-07,
      "loss": 0.0,
      "num_tokens": 4160130.0,
      "reward": 0.17976805567741394,
      "reward_std": 0.5253881216049194,
      "rewards/reward_function/mean": 0.17976805567741394,
      "rewards/reward_function/std": 0.5253881216049194,
      "step": 79
    },
    {
      "completion_length": 896.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1078.0,
      "completions/max_terminated_length": 1078.0,
      "completions/mean_length": 896.625,
      "completions/mean_terminated_length": 896.625,
      "completions/min_length": 721.0,
      "completions/min_terminated_length": 721.0,
      "epoch": 0.3125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33451908826828003,
      "kl": 0.0004913711964036338,
      "learning_rate": 6.421379363065142e-07,
      "loss": 0.0,
      "num_tokens": 4212047.0,
      "reward": 0.2575118839740753,
      "reward_std": 0.5082336068153381,
      "rewards/reward_function/mean": 0.2575118839740753,
      "rewards/reward_function/std": 0.5082336068153381,
      "step": 80
    },
    {
      "completion_length": 1186.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1439.0,
      "completions/max_terminated_length": 1439.0,
      "completions/mean_length": 1186.5,
      "completions/mean_terminated_length": 1186.5,
      "completions/min_length": 1046.0,
      "completions/min_terminated_length": 1046.0,
      "epoch": 0.31640625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2777305543422699,
      "kl": 0.0004566588904708624,
      "learning_rate": 5.848888922025553e-07,
      "loss": 0.0,
      "num_tokens": 4270267.0,
      "reward": 0.08119191229343414,
      "reward_std": 0.6674035787582397,
      "rewards/reward_function/mean": 0.08119191229343414,
      "rewards/reward_function/std": 0.6674035787582397,
      "step": 81
    },
    {
      "completion_length": 1302.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1983.0,
      "completions/max_terminated_length": 1983.0,
      "completions/mean_length": 1302.625,
      "completions/mean_terminated_length": 1302.625,
      "completions/min_length": 849.0,
      "completions/min_terminated_length": 849.0,
      "epoch": 0.3203125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2776745855808258,
      "kl": 0.0004434596048668027,
      "learning_rate": 5.299731159831953e-07,
      "loss": 0.0,
      "num_tokens": 4329416.0,
      "reward": -0.09772323071956635,
      "reward_std": 0.7471930980682373,
      "rewards/reward_function/mean": -0.09772323071956635,
      "rewards/reward_function/std": 0.7471930980682373,
      "step": 82
    },
    {
      "completion_length": 1093.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1306.0,
      "completions/max_terminated_length": 1306.0,
      "completions/mean_length": 1093.875,
      "completions/mean_terminated_length": 1093.875,
      "completions/min_length": 877.0,
      "completions/min_terminated_length": 877.0,
      "epoch": 0.32421875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3191281259059906,
      "kl": 0.0004105157349840738,
      "learning_rate": 4.774575140626317e-07,
      "loss": 0.0,
      "num_tokens": 4386895.0,
      "reward": 0.256418377161026,
      "reward_std": 0.5078137516975403,
      "rewards/reward_function/mean": 0.256418377161026,
      "rewards/reward_function/std": 0.5078137516975403,
      "step": 83
    },
    {
      "completion_length": 1108.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1572.0,
      "completions/max_terminated_length": 1572.0,
      "completions/mean_length": 1108.375,
      "completions/mean_terminated_length": 1108.375,
      "completions/min_length": 853.0,
      "completions/min_terminated_length": 853.0,
      "epoch": 0.328125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.26989954710006714,
      "kl": 0.0005018782903789543,
      "learning_rate": 4.27406068612396e-07,
      "loss": 0.0,
      "num_tokens": 4444490.0,
      "reward": 0.12714649736881256,
      "reward_std": 0.5166128277778625,
      "rewards/reward_function/mean": 0.12714649736881256,
      "rewards/reward_function/std": 0.5166128277778625,
      "step": 84
    },
    {
      "completion_length": 1082.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1199.0,
      "completions/max_terminated_length": 1199.0,
      "completions/mean_length": 1082.25,
      "completions/mean_terminated_length": 1082.25,
      "completions/min_length": 984.0,
      "completions/min_terminated_length": 984.0,
      "epoch": 0.33203125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2745755612850189,
      "kl": 0.0005251698967185803,
      "learning_rate": 3.798797596089351e-07,
      "loss": 0.0,
      "num_tokens": 4502412.0,
      "reward": -0.08731149882078171,
      "reward_std": 0.7558506727218628,
      "rewards/reward_function/mean": -0.08731149882078171,
      "rewards/reward_function/std": 0.7558506727218628,
      "step": 85
    },
    {
      "completion_length": 960.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1336.0,
      "completions/max_terminated_length": 1336.0,
      "completions/mean_length": 960.625,
      "completions/mean_terminated_length": 960.625,
      "completions/min_length": 536.0,
      "completions/min_terminated_length": 536.0,
      "epoch": 0.3359375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3008907735347748,
      "kl": 0.0004591922843246721,
      "learning_rate": 3.3493649053890325e-07,
      "loss": 0.0,
      "num_tokens": 4559361.0,
      "reward": 0.1179749146103859,
      "reward_std": 0.5200047492980957,
      "rewards/reward_function/mean": 0.1179749146103859,
      "rewards/reward_function/std": 0.5200048089027405,
      "step": 86
    },
    {
      "completion_length": 1191.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1353.0,
      "completions/max_terminated_length": 1353.0,
      "completions/mean_length": 1191.625,
      "completions/mean_terminated_length": 1191.625,
      "completions/min_length": 1014.0,
      "completions/min_terminated_length": 1014.0,
      "epoch": 0.33984375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29382893443107605,
      "kl": 0.0004918071499560028,
      "learning_rate": 2.9263101785268253e-07,
      "loss": 0.0,
      "num_tokens": 4618158.0,
      "reward": 0.10288041085004807,
      "reward_std": 0.680895984172821,
      "rewards/reward_function/mean": 0.10288041085004807,
      "rewards/reward_function/std": 0.680895984172821,
      "step": 87
    },
    {
      "completion_length": 1024.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1330.0,
      "completions/max_terminated_length": 1330.0,
      "completions/mean_length": 1024.625,
      "completions/mean_terminated_length": 1024.625,
      "completions/min_length": 830.0,
      "completions/min_terminated_length": 830.0,
      "epoch": 0.34375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.30418917536735535,
      "kl": 0.0005036265138187446,
      "learning_rate": 2.53014884252083e-07,
      "loss": 0.0,
      "num_tokens": 4675619.0,
      "reward": 0.20726439356803894,
      "reward_std": 0.5138247013092041,
      "rewards/reward_function/mean": 0.20726439356803894,
      "rewards/reward_function/std": 0.5138247609138489,
      "step": 88
    },
    {
      "completion_length": 1076.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1415.0,
      "completions/max_terminated_length": 1415.0,
      "completions/mean_length": 1076.25,
      "completions/mean_terminated_length": 1076.25,
      "completions/min_length": 825.0,
      "completions/min_terminated_length": 825.0,
      "epoch": 0.34765625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29263392090797424,
      "kl": 0.0005031488981330767,
      "learning_rate": 2.1613635589349756e-07,
      "loss": 0.0,
      "num_tokens": 4727813.0,
      "reward": 0.18039438128471375,
      "reward_std": 0.5106222033500671,
      "rewards/reward_function/mean": 0.18039438128471375,
      "rewards/reward_function/std": 0.5106222033500671,
      "step": 89
    },
    {
      "completion_length": 1055.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1210.0,
      "completions/max_terminated_length": 1210.0,
      "completions/mean_length": 1055.375,
      "completions/mean_terminated_length": 1055.375,
      "completions/min_length": 936.0,
      "completions/min_terminated_length": 936.0,
      "epoch": 0.3515625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29214635491371155,
      "kl": 0.0004412992129800841,
      "learning_rate": 1.8204036358303173e-07,
      "loss": 0.0,
      "num_tokens": 4779840.0,
      "reward": 0.4273361563682556,
      "reward_std": 0.012010330334305763,
      "rewards/reward_function/mean": 0.4273361563682556,
      "rewards/reward_function/std": 0.012010332196950912,
      "step": 90
    },
    {
      "completion_length": 1136.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1483.0,
      "completions/max_terminated_length": 1483.0,
      "completions/mean_length": 1136.25,
      "completions/mean_terminated_length": 1136.25,
      "completions/min_length": 810.0,
      "completions/min_terminated_length": 810.0,
      "epoch": 0.35546875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2611209750175476,
      "kl": 0.0004386235450510867,
      "learning_rate": 1.507684480352292e-07,
      "loss": 0.0,
      "num_tokens": 4832514.0,
      "reward": 0.24031642079353333,
      "reward_std": 0.5015574097633362,
      "rewards/reward_function/mean": 0.24031642079353333,
      "rewards/reward_function/std": 0.5015574097633362,
      "step": 91
    },
    {
      "completion_length": 1092.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1367.0,
      "completions/max_terminated_length": 1367.0,
      "completions/mean_length": 1092.25,
      "completions/mean_terminated_length": 1092.25,
      "completions/min_length": 843.0,
      "completions/min_terminated_length": 843.0,
      "epoch": 0.359375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31895720958709717,
      "kl": 0.0005823632091050968,
      "learning_rate": 1.223587092621162e-07,
      "loss": 0.0,
      "num_tokens": 4884836.0,
      "reward": -0.11290903389453888,
      "reward_std": 0.7347157001495361,
      "rewards/reward_function/mean": -0.11290903389453888,
      "rewards/reward_function/std": 0.7347157001495361,
      "step": 92
    },
    {
      "completion_length": 1156.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1366.0,
      "completions/max_terminated_length": 1366.0,
      "completions/mean_length": 1156.5,
      "completions/mean_terminated_length": 1156.5,
      "completions/min_length": 944.0,
      "completions/min_terminated_length": 944.0,
      "epoch": 0.36328125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3020682632923126,
      "kl": 0.0004940428771078587,
      "learning_rate": 9.684576015420277e-08,
      "loss": 0.0,
      "num_tokens": 4941160.0,
      "reward": 0.06459739059209824,
      "reward_std": 0.6571158766746521,
      "rewards/reward_function/mean": 0.06459739059209824,
      "rewards/reward_function/std": 0.6571159362792969,
      "step": 93
    },
    {
      "completion_length": 1061.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1237.0,
      "completions/max_terminated_length": 1237.0,
      "completions/mean_length": 1061.625,
      "completions/mean_terminated_length": 1061.625,
      "completions/min_length": 781.0,
      "completions/min_terminated_length": 781.0,
      "epoch": 0.3671875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3015153706073761,
      "kl": 0.0005082947827759199,
      "learning_rate": 7.426068431000883e-08,
      "loss": 0.0,
      "num_tokens": 4996725.0,
      "reward": 0.2391453981399536,
      "reward_std": 0.5007839202880859,
      "rewards/reward_function/mean": 0.2391453981399536,
      "rewards/reward_function/std": 0.5007839798927307,
      "step": 94
    },
    {
      "completion_length": 947.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1387.0,
      "completions/max_terminated_length": 1387.0,
      "completions/mean_length": 947.75,
      "completions/mean_terminated_length": 947.75,
      "completions/min_length": 641.0,
      "completions/min_terminated_length": 641.0,
      "epoch": 0.37109375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31657955050468445,
      "kl": 0.0004573179394355975,
      "learning_rate": 5.463099816548578e-08,
      "loss": 0.0,
      "num_tokens": 5051379.0,
      "reward": -0.019695594906806946,
      "reward_std": 0.638481855392456,
      "rewards/reward_function/mean": -0.019695594906806946,
      "rewards/reward_function/std": 0.6384819149971008,
      "step": 95
    },
    {
      "completion_length": 980.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1112.0,
      "completions/max_terminated_length": 1112.0,
      "completions/mean_length": 980.875,
      "completions/mean_terminated_length": 980.875,
      "completions/min_length": 808.0,
      "completions/min_terminated_length": 808.0,
      "epoch": 0.375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.38886314630508423,
      "kl": 0.0004613845740095712,
      "learning_rate": 3.798061746947995e-08,
      "loss": -0.0,
      "num_tokens": 5106298.0,
      "reward": 0.41173434257507324,
      "reward_std": 0.008369507268071175,
      "rewards/reward_function/mean": 0.41173434257507324,
      "rewards/reward_function/std": 0.008369502611458302,
      "step": 96
    },
    {
      "completion_length": 1003.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1291.0,
      "completions/max_terminated_length": 1291.0,
      "completions/mean_length": 1003.5,
      "completions/mean_terminated_length": 1003.5,
      "completions/min_length": 668.0,
      "completions/min_terminated_length": 668.0,
      "epoch": 0.37890625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3679444193840027,
      "kl": 0.00046288836892927065,
      "learning_rate": 2.4329828146074096e-08,
      "loss": 0.0,
      "num_tokens": 5146238.0,
      "reward": 0.18697933852672577,
      "reward_std": 0.5134840607643127,
      "rewards/reward_function/mean": 0.18697933852672577,
      "rewards/reward_function/std": 0.5134840607643127,
      "step": 97
    },
    {
      "completion_length": 1033.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1228.0,
      "completions/max_terminated_length": 1228.0,
      "completions/mean_length": 1033.0,
      "completions/mean_terminated_length": 1033.0,
      "completions/min_length": 853.0,
      "completions/min_terminated_length": 853.0,
      "epoch": 0.3828125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2726477086544037,
      "kl": 0.00043194100726395845,
      "learning_rate": 1.3695261579316776e-08,
      "loss": 0.0,
      "num_tokens": 5186414.0,
      "reward": 0.07853780686855316,
      "reward_std": 0.6657574772834778,
      "rewards/reward_function/mean": 0.07853780686855316,
      "rewards/reward_function/std": 0.6657574772834778,
      "step": 98
    },
    {
      "completion_length": 1069.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1659.0,
      "completions/max_terminated_length": 1659.0,
      "completions/mean_length": 1069.5,
      "completions/mean_terminated_length": 1069.5,
      "completions/min_length": 922.0,
      "completions/min_terminated_length": 922.0,
      "epoch": 0.38671875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.30219021439552307,
      "kl": 0.000492539482365828,
      "learning_rate": 6.089874350439507e-09,
      "loss": 0.0,
      "num_tokens": 5226882.0,
      "reward": 0.07734433561563492,
      "reward_std": 0.6650528311729431,
      "rewards/reward_function/mean": 0.07734433561563492,
      "rewards/reward_function/std": 0.6650528907775879,
      "step": 99
    },
    {
      "completion_length": 918.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1200.0,
      "completions/max_terminated_length": 1200.0,
      "completions/mean_length": 918.0,
      "completions/mean_terminated_length": 918.0,
      "completions/min_length": 573.0,
      "completions/min_terminated_length": 573.0,
      "epoch": 0.390625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.44854360818862915,
      "kl": 0.00044552871986525133,
      "learning_rate": 1.5229324522605949e-09,
      "loss": 0.0,
      "num_tokens": 5266138.0,
      "reward": 0.25066205859184265,
      "reward_std": 0.5054365992546082,
      "rewards/reward_function/mean": 0.25066205859184265,
      "rewards/reward_function/std": 0.5054366588592529,
      "step": 100
    },
    {
      "completion_length": 971.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1311.0,
      "completions/max_terminated_length": 1311.0,
      "completions/mean_length": 971.75,
      "completions/mean_terminated_length": 971.75,
      "completions/min_length": 562.0,
      "completions/min_terminated_length": 562.0,
      "epoch": 0.39453125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.35067111253738403,
      "kl": 0.0005845300620421767,
      "learning_rate": 0.0,
      "loss": 0.0,
      "num_tokens": 5305824.0,
      "reward": -0.11205706000328064,
      "reward_std": 0.7353475689888,
      "rewards/reward_function/mean": -0.11205706000328064,
      "rewards/reward_function/std": 0.7353475689888,
      "step": 101
    },
    {
      "completion_length": 1009.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1224.0,
      "completions/max_terminated_length": 1224.0,
      "completions/mean_length": 1009.75,
      "completions/mean_terminated_length": 1009.75,
      "completions/min_length": 697.0,
      "completions/min_terminated_length": 697.0,
      "epoch": 0.3984375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33843955397605896,
      "kl": 0.0005709103206754662,
      "learning_rate": 2.8910861626005774e-06,
      "loss": 0.0,
      "num_tokens": 5352350.0,
      "reward": -0.10698533058166504,
      "reward_std": 0.739599883556366,
      "rewards/reward_function/mean": -0.10698533058166504,
      "rewards/reward_function/std": 0.7395999431610107,
      "step": 102
    },
    {
      "completion_length": 1043.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1264.0,
      "completions/max_terminated_length": 1264.0,
      "completions/mean_length": 1043.0,
      "completions/mean_terminated_length": 1043.0,
      "completions/min_length": 945.0,
      "completions/min_terminated_length": 945.0,
      "epoch": 0.40234375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3096908628940582,
      "kl": 0.0005897778464714065,
      "learning_rate": 2.847932752400164e-06,
      "loss": 0.0,
      "num_tokens": 5392606.0,
      "reward": 0.248357892036438,
      "reward_std": 0.5045785307884216,
      "rewards/reward_function/mean": 0.248357892036438,
      "rewards/reward_function/std": 0.5045785307884216,
      "step": 103
    },
    {
      "completion_length": 1079.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1243.0,
      "completions/max_terminated_length": 1243.0,
      "completions/mean_length": 1079.75,
      "completions/mean_terminated_length": 1079.75,
      "completions/min_length": 848.0,
      "completions/min_terminated_length": 848.0,
      "epoch": 0.40625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3232628107070923,
      "kl": 0.0005662211769958958,
      "learning_rate": 2.804673358512869e-06,
      "loss": 0.0,
      "num_tokens": 5439692.0,
      "reward": 0.24970494210720062,
      "reward_std": 0.505219578742981,
      "rewards/reward_function/mean": 0.24970494210720062,
      "rewards/reward_function/std": 0.5052196383476257,
      "step": 104
    },
    {
      "completion_length": 1118.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1456.0,
      "completions/max_terminated_length": 1456.0,
      "completions/mean_length": 1118.75,
      "completions/mean_terminated_length": 1118.75,
      "completions/min_length": 465.0,
      "completions/min_terminated_length": 465.0,
      "epoch": 0.41015625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3483864665031433,
      "kl": 0.0005037105802330188,
      "learning_rate": 2.761321158169134e-06,
      "loss": 0.0,
      "num_tokens": 5492746.0,
      "reward": -0.00016714632511138916,
      "reward_std": 0.64275723695755,
      "rewards/reward_function/mean": -0.00016714632511138916,
      "rewards/reward_function/std": 0.64275723695755,
      "step": 105
    },
    {
      "completion_length": 941.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1185.0,
      "completions/max_terminated_length": 1185.0,
      "completions/mean_length": 941.875,
      "completions/mean_terminated_length": 941.875,
      "completions/min_length": 637.0,
      "completions/min_terminated_length": 637.0,
      "epoch": 0.4140625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3742240071296692,
      "kl": 0.0005701555346604437,
      "learning_rate": 2.717889356869146e-06,
      "loss": 0.0,
      "num_tokens": 5549881.0,
      "reward": 0.2512204349040985,
      "reward_std": 0.505646288394928,
      "rewards/reward_function/mean": 0.2512204349040985,
      "rewards/reward_function/std": 0.5056463479995728,
      "step": 106
    },
    {
      "completion_length": 990.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1131.0,
      "completions/max_terminated_length": 1131.0,
      "completions/mean_length": 990.75,
      "completions/mean_terminated_length": 990.75,
      "completions/min_length": 906.0,
      "completions/min_terminated_length": 906.0,
      "epoch": 0.41796875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2908116579055786,
      "kl": 0.0004926583642372862,
      "learning_rate": 2.6743911843603134e-06,
      "loss": 0.0,
      "num_tokens": 5601911.0,
      "reward": 0.0615215003490448,
      "reward_std": 0.6552067399024963,
      "rewards/reward_function/mean": 0.0615215003490448,
      "rewards/reward_function/std": 0.6552067399024963,
      "step": 107
    },
    {
      "completion_length": 1138.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1816.0,
      "completions/max_terminated_length": 1816.0,
      "completions/mean_length": 1138.625,
      "completions/mean_terminated_length": 1138.625,
      "completions/min_length": 861.0,
      "completions/min_terminated_length": 861.0,
      "epoch": 0.421875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.26039204001426697,
      "kl": 0.0006796395464334637,
      "learning_rate": 2.6308398906073603e-06,
      "loss": 0.0,
      "num_tokens": 5660620.0,
      "reward": -0.8185928463935852,
      "reward_std": 0.5130969882011414,
      "rewards/reward_function/mean": -0.8185928463935852,
      "rewards/reward_function/std": 0.5130969882011414,
      "step": 108
    },
    {
      "completion_length": 1128.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1260.0,
      "completions/max_terminated_length": 1260.0,
      "completions/mean_length": 1128.375,
      "completions/mean_terminated_length": 1128.375,
      "completions/min_length": 983.0,
      "completions/min_terminated_length": 983.0,
      "epoch": 0.42578125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.28983625769615173,
      "kl": 0.000518458073202055,
      "learning_rate": 2.587248741756253e-06,
      "loss": 0.0,
      "num_tokens": 5713751.0,
      "reward": 0.09786057472229004,
      "reward_std": 0.520240306854248,
      "rewards/reward_function/mean": 0.09786057472229004,
      "rewards/reward_function/std": 0.520240306854248,
      "step": 109
    },
    {
      "completion_length": 980.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1334.0,
      "completions/max_terminated_length": 1334.0,
      "completions/mean_length": 980.5,
      "completions/mean_terminated_length": 980.5,
      "completions/min_length": 697.0,
      "completions/min_terminated_length": 697.0,
      "epoch": 0.4296875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3524278700351715,
      "kl": 0.0005516579622053541,
      "learning_rate": 2.543631016093209e-06,
      "loss": 0.0,
      "num_tokens": 5771195.0,
      "reward": 0.06997495889663696,
      "reward_std": 0.660901665687561,
      "rewards/reward_function/mean": 0.06997495889663696,
      "rewards/reward_function/std": 0.6609017252922058,
      "step": 110
    },
    {
      "completion_length": 996.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1295.0,
      "completions/max_terminated_length": 1295.0,
      "completions/mean_length": 996.125,
      "completions/mean_terminated_length": 996.125,
      "completions/min_length": 423.0,
      "completions/min_terminated_length": 423.0,
      "epoch": 0.43359375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3624061644077301,
      "kl": 0.000616533579886891,
      "learning_rate": 2.5e-06,
      "loss": 0.0,
      "num_tokens": 5823268.0,
      "reward": -0.29218366742134094,
      "reward_std": 0.7567039728164673,
      "rewards/reward_function/mean": -0.29218366742134094,
      "rewards/reward_function/std": 0.7567040324211121,
      "step": 111
    },
    {
      "completion_length": 993.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1481.0,
      "completions/max_terminated_length": 1481.0,
      "completions/mean_length": 993.625,
      "completions/mean_terminated_length": 993.625,
      "completions/min_length": 747.0,
      "completions/min_terminated_length": 747.0,
      "epoch": 0.4375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3491348624229431,
      "kl": 0.0005375376276788302,
      "learning_rate": 2.4563689839067913e-06,
      "loss": 0.0,
      "num_tokens": 5880817.0,
      "reward": 0.12125939130783081,
      "reward_std": 0.5074965357780457,
      "rewards/reward_function/mean": 0.12125939130783081,
      "rewards/reward_function/std": 0.5074965953826904,
      "step": 112
    },
    {
      "completion_length": 1085.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1872.0,
      "completions/max_terminated_length": 1872.0,
      "completions/mean_length": 1085.5,
      "completions/mean_terminated_length": 1085.5,
      "completions/min_length": 877.0,
      "completions/min_terminated_length": 877.0,
      "epoch": 0.44140625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33136165142059326,
      "kl": 0.0005395807238528505,
      "learning_rate": 2.4127512582437486e-06,
      "loss": 0.0,
      "num_tokens": 5926349.0,
      "reward": -0.11549624800682068,
      "reward_std": 0.7324655055999756,
      "rewards/reward_function/mean": -0.11549624800682068,
      "rewards/reward_function/std": 0.7324654459953308,
      "step": 113
    },
    {
      "completion_length": 956.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1136.0,
      "completions/max_terminated_length": 1136.0,
      "completions/mean_length": 956.625,
      "completions/mean_terminated_length": 956.625,
      "completions/min_length": 763.0,
      "completions/min_terminated_length": 763.0,
      "epoch": 0.4453125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.325693815946579,
      "kl": 0.000503135692270007,
      "learning_rate": 2.3691601093926406e-06,
      "loss": 0.0,
      "num_tokens": 5966210.0,
      "reward": -0.006239533424377441,
      "reward_std": 0.6517221927642822,
      "rewards/reward_function/mean": -0.006239533424377441,
      "rewards/reward_function/std": 0.651722252368927,
      "step": 114
    },
    {
      "completion_length": 1024.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1456.0,
      "completions/max_terminated_length": 1456.0,
      "completions/mean_length": 1024.625,
      "completions/mean_terminated_length": 1024.625,
      "completions/min_length": 864.0,
      "completions/min_terminated_length": 864.0,
      "epoch": 0.44921875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3105911612510681,
      "kl": 0.0004542986207525246,
      "learning_rate": 2.325608815639687e-06,
      "loss": 0.0,
      "num_tokens": 6011255.0,
      "reward": 0.0608128160238266,
      "reward_std": 0.6547765731811523,
      "rewards/reward_function/mean": 0.0608128160238266,
      "rewards/reward_function/std": 0.6547765731811523,
      "step": 115
    },
    {
      "completion_length": 1240.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1532.0,
      "completions/max_terminated_length": 1532.0,
      "completions/mean_length": 1240.75,
      "completions/mean_terminated_length": 1240.75,
      "completions/min_length": 942.0,
      "completions/min_terminated_length": 942.0,
      "epoch": 0.453125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.26683756709098816,
      "kl": 0.0005024799829698168,
      "learning_rate": 2.2821106431308546e-06,
      "loss": 0.0,
      "num_tokens": 6053389.0,
      "reward": 0.2592804431915283,
      "reward_std": 0.50887131690979,
      "rewards/reward_function/mean": 0.2592804431915283,
      "rewards/reward_function/std": 0.5088713765144348,
      "step": 116
    },
    {
      "completion_length": 1007.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1154.0,
      "completions/max_terminated_length": 1154.0,
      "completions/mean_length": 1007.875,
      "completions/mean_terminated_length": 1007.875,
      "completions/min_length": 794.0,
      "completions/min_terminated_length": 794.0,
      "epoch": 0.45703125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3248363137245178,
      "kl": 0.0005079211114207283,
      "learning_rate": 2.238678841830867e-06,
      "loss": 0.0,
      "num_tokens": 6098300.0,
      "reward": -0.29335445165634155,
      "reward_std": 0.7554491758346558,
      "rewards/reward_function/mean": -0.29335445165634155,
      "rewards/reward_function/std": 0.7554491758346558,
      "step": 117
    },
    {
      "completion_length": 1096.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1226.0,
      "completions/max_terminated_length": 1226.0,
      "completions/mean_length": 1096.25,
      "completions/mean_terminated_length": 1096.25,
      "completions/min_length": 902.0,
      "completions/min_terminated_length": 902.0,
      "epoch": 0.4609375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.32839903235435486,
      "kl": 0.0005834192561451346,
      "learning_rate": 2.195326641487132e-06,
      "loss": 0.0,
      "num_tokens": 6139278.0,
      "reward": -0.10875681042671204,
      "reward_std": 0.7381133437156677,
      "rewards/reward_function/mean": -0.10875681042671204,
      "rewards/reward_function/std": 0.7381134033203125,
      "step": 118
    },
    {
      "completion_length": 1068.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1261.0,
      "completions/max_terminated_length": 1261.0,
      "completions/mean_length": 1068.5,
      "completions/mean_terminated_length": 1068.5,
      "completions/min_length": 838.0,
      "completions/min_terminated_length": 838.0,
      "epoch": 0.46484375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3054167628288269,
      "kl": 0.0005187875867704861,
      "learning_rate": 2.1520672475998374e-06,
      "loss": 0.0,
      "num_tokens": 6184674.0,
      "reward": -0.356624573469162,
      "reward_std": 0.7082850933074951,
      "rewards/reward_function/mean": -0.356624573469162,
      "rewards/reward_function/std": 0.7082850933074951,
      "step": 119
    },
    {
      "completion_length": 1191.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1684.0,
      "completions/max_terminated_length": 1684.0,
      "completions/mean_length": 1191.875,
      "completions/mean_terminated_length": 1191.875,
      "completions/min_length": 929.0,
      "completions/min_terminated_length": 929.0,
      "epoch": 0.46875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2923835217952728,
      "kl": 0.0005189872026676312,
      "learning_rate": 2.1089138373994226e-06,
      "loss": 0.0,
      "num_tokens": 6226417.0,
      "reward": -0.10591146349906921,
      "reward_std": 0.7404077053070068,
      "rewards/reward_function/mean": -0.10591146349906921,
      "rewards/reward_function/std": 0.7404076457023621,
      "step": 120
    },
    {
      "completion_length": 987.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1249.0,
      "completions/max_terminated_length": 1249.0,
      "completions/mean_length": 987.125,
      "completions/mean_terminated_length": 987.125,
      "completions/min_length": 757.0,
      "completions/min_terminated_length": 757.0,
      "epoch": 0.47265625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31898999214172363,
      "kl": 0.0005351051659090444,
      "learning_rate": 2.0658795558326745e-06,
      "loss": 0.0,
      "num_tokens": 6273762.0,
      "reward": -0.007490452378988266,
      "reward_std": 0.645849347114563,
      "rewards/reward_function/mean": -0.007490452378988266,
      "rewards/reward_function/std": 0.645849347114563,
      "step": 121
    },
    {
      "completion_length": 1239.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2190.0,
      "completions/max_terminated_length": 2190.0,
      "completions/mean_length": 1239.125,
      "completions/mean_terminated_length": 1239.125,
      "completions/min_length": 834.0,
      "completions/min_terminated_length": 834.0,
      "epoch": 0.4765625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3210497498512268,
      "kl": 0.0005187717179069296,
      "learning_rate": 2.022977511558638e-06,
      "loss": 0.0,
      "num_tokens": 6332283.0,
      "reward": 0.05918067693710327,
      "reward_std": 0.6538223028182983,
      "rewards/reward_function/mean": 0.05918067693710327,
      "rewards/reward_function/std": 0.6538223624229431,
      "step": 122
    },
    {
      "completion_length": 1006.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1189.0,
      "completions/max_terminated_length": 1189.0,
      "completions/mean_length": 1006.125,
      "completions/mean_terminated_length": 1006.125,
      "completions/min_length": 898.0,
      "completions/min_terminated_length": 898.0,
      "epoch": 0.48046875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3229573667049408,
      "kl": 0.0005403026298154145,
      "learning_rate": 1.9802207729556023e-06,
      "loss": 0.0,
      "num_tokens": 6379780.0,
      "reward": 0.17201174795627594,
      "reward_std": 0.5175948143005371,
      "rewards/reward_function/mean": 0.17201174795627594,
      "rewards/reward_function/std": 0.5175948739051819,
      "step": 123
    },
    {
      "completion_length": 971.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1154.0,
      "completions/max_terminated_length": 1154.0,
      "completions/mean_length": 971.0,
      "completions/mean_terminated_length": 971.0,
      "completions/min_length": 845.0,
      "completions/min_terminated_length": 845.0,
      "epoch": 0.484375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.309548944234848,
      "kl": 0.00047888144035823643,
      "learning_rate": 1.937622364140338e-06,
      "loss": 0.0,
      "num_tokens": 6436156.0,
      "reward": -0.12267381697893143,
      "reward_std": 0.7265216708183289,
      "rewards/reward_function/mean": -0.12267381697893143,
      "rewards/reward_function/std": 0.7265217304229736,
      "step": 124
    },
    {
      "completion_length": 917.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1127.0,
      "completions/max_terminated_length": 1127.0,
      "completions/mean_length": 917.125,
      "completions/mean_terminated_length": 917.125,
      "completions/min_length": 698.0,
      "completions/min_terminated_length": 698.0,
      "epoch": 0.48828125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3436194658279419,
      "kl": 0.0005273657734505832,
      "learning_rate": 1.895195261000831e-06,
      "loss": 0.0,
      "num_tokens": 6482941.0,
      "reward": 0.06335102021694183,
      "reward_std": 0.6563363075256348,
      "rewards/reward_function/mean": 0.06335102021694183,
      "rewards/reward_function/std": 0.6563363075256348,
      "step": 125
    },
    {
      "completion_length": 981.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1088.0,
      "completions/max_terminated_length": 1088.0,
      "completions/mean_length": 981.25,
      "completions/mean_terminated_length": 981.25,
      "completions/min_length": 882.0,
      "completions/min_terminated_length": 882.0,
      "epoch": 0.4921875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3238428235054016,
      "kl": 0.00048635365965310484,
      "learning_rate": 1.852952387243698e-06,
      "loss": 0.0,
      "num_tokens": 6539399.0,
      "reward": 0.4136182963848114,
      "reward_std": 0.0051859593950212,
      "rewards/reward_function/mean": 0.4136182963848114,
      "rewards/reward_function/std": 0.005185958929359913,
      "step": 126
    },
    {
      "completion_length": 980.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1164.0,
      "completions/max_terminated_length": 1164.0,
      "completions/mean_length": 980.25,
      "completions/mean_terminated_length": 980.25,
      "completions/min_length": 865.0,
      "completions/min_terminated_length": 865.0,
      "epoch": 0.49609375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.32173460721969604,
      "kl": 0.0005382613235269673,
      "learning_rate": 1.8109066104575023e-06,
      "loss": -0.0,
      "num_tokens": 6586689.0,
      "reward": 0.4277029037475586,
      "reward_std": 0.009604846127331257,
      "rewards/reward_function/mean": 0.4277029037475586,
      "rewards/reward_function/std": 0.00960485078394413,
      "step": 127
    },
    {
      "completion_length": 1130.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1473.0,
      "completions/max_terminated_length": 1473.0,
      "completions/mean_length": 1130.5,
      "completions/mean_terminated_length": 1130.5,
      "completions/min_length": 904.0,
      "completions/min_terminated_length": 904.0,
      "epoch": 0.5,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.27207547426223755,
      "kl": 0.0005242059778538533,
      "learning_rate": 1.7690707381931585e-06,
      "loss": 0.0,
      "num_tokens": 6644341.0,
      "reward": -0.4683274030685425,
      "reward_std": 0.7338022589683533,
      "rewards/reward_function/mean": -0.4683274030685425,
      "rewards/reward_function/std": 0.7338022589683533,
      "step": 128
    },
    {
      "completion_length": 994.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1057.0,
      "completions/max_terminated_length": 1057.0,
      "completions/mean_length": 994.5,
      "completions/mean_terminated_length": 994.5,
      "completions/min_length": 866.0,
      "completions/min_terminated_length": 866.0,
      "epoch": 0.50390625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.30755338072776794,
      "kl": 0.0005628820799756795,
      "learning_rate": 1.7274575140626318e-06,
      "loss": 0.0,
      "num_tokens": 6698833.0,
      "reward": -0.016938716173171997,
      "reward_std": 0.645233154296875,
      "rewards/reward_function/mean": -0.016938716173171997,
      "rewards/reward_function/std": 0.6452332139015198,
      "step": 129
    },
    {
      "completion_length": 1191.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1623.0,
      "completions/max_terminated_length": 1623.0,
      "completions/mean_length": 1191.5,
      "completions/mean_terminated_length": 1191.5,
      "completions/min_length": 890.0,
      "completions/min_terminated_length": 890.0,
      "epoch": 0.5078125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2911682426929474,
      "kl": 0.0005079254988231696,
      "learning_rate": 1.686079613857109e-06,
      "loss": 0.0,
      "num_tokens": 6755397.0,
      "reward": -0.10992002487182617,
      "reward_std": 0.7371261119842529,
      "rewards/reward_function/mean": -0.10992002487182617,
      "rewards/reward_function/std": 0.7371261119842529,
      "step": 130
    },
    {
      "completion_length": 929.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1217.0,
      "completions/max_terminated_length": 1217.0,
      "completions/mean_length": 929.375,
      "completions/mean_terminated_length": 929.375,
      "completions/min_length": 202.0,
      "completions/min_terminated_length": 202.0,
      "epoch": 0.51171875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.3324370384216309,
      "kl": 0.0007534601754741743,
      "learning_rate": 1.6449496416858285e-06,
      "loss": 0.0,
      "num_tokens": 6809368.0,
      "reward": 0.1685469150543213,
      "reward_std": 0.5241107940673828,
      "rewards/reward_function/mean": 0.1685469150543213,
      "rewards/reward_function/std": 0.5241108536720276,
      "step": 131
    },
    {
      "completion_length": 1025.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1330.0,
      "completions/max_terminated_length": 1330.0,
      "completions/mean_length": 1025.625,
      "completions/mean_terminated_length": 1025.625,
      "completions/min_length": 854.0,
      "completions/min_terminated_length": 854.0,
      "epoch": 0.515625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34628283977508545,
      "kl": 0.0005288709871820174,
      "learning_rate": 1.6040801261367494e-06,
      "loss": 0.0,
      "num_tokens": 6864605.0,
      "reward": -0.10574899613857269,
      "reward_std": 0.7405577301979065,
      "rewards/reward_function/mean": -0.10574899613857269,
      "rewards/reward_function/std": 0.7405577898025513,
      "step": 132
    },
    {
      "completion_length": 1157.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1519.0,
      "completions/max_terminated_length": 1519.0,
      "completions/mean_length": 1157.75,
      "completions/mean_terminated_length": 1157.75,
      "completions/min_length": 886.0,
      "completions/min_terminated_length": 886.0,
      "epoch": 0.51953125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29528382420539856,
      "kl": 0.0005665956414304674,
      "learning_rate": 1.56348351646022e-06,
      "loss": 0.0,
      "num_tokens": 6920403.0,
      "reward": -0.111003577709198,
      "reward_std": 0.7361726760864258,
      "rewards/reward_function/mean": -0.111003577709198,
      "rewards/reward_function/std": 0.7361726760864258,
      "step": 133
    },
    {
      "completion_length": 1294.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2684.0,
      "completions/max_terminated_length": 2684.0,
      "completions/mean_length": 1294.25,
      "completions/mean_terminated_length": 1294.25,
      "completions/min_length": 824.0,
      "completions/min_terminated_length": 824.0,
      "epoch": 0.5234375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.19954568147659302,
      "kl": 0.0004810404570889659,
      "learning_rate": 1.5231721787768162e-06,
      "loss": 0.0,
      "num_tokens": 6977789.0,
      "reward": 0.2454666793346405,
      "reward_std": 0.503404974937439,
      "rewards/reward_function/mean": 0.2454666793346405,
      "rewards/reward_function/std": 0.503404974937439,
      "step": 134
    },
    {
      "completion_length": 1026.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1272.0,
      "completions/max_terminated_length": 1272.0,
      "completions/mean_length": 1026.875,
      "completions/mean_terminated_length": 1026.875,
      "completions/min_length": 663.0,
      "completions/min_terminated_length": 663.0,
      "epoch": 0.52734375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2833687961101532,
      "kl": 0.00051848181465175,
      "learning_rate": 1.4831583923105e-06,
      "loss": 0.0,
      "num_tokens": 7032540.0,
      "reward": 0.23018378019332886,
      "reward_std": 0.4979630708694458,
      "rewards/reward_function/mean": 0.23018378019332886,
      "rewards/reward_function/std": 0.4979631304740906,
      "step": 135
    },
    {
      "completion_length": 1143.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2065.0,
      "completions/max_terminated_length": 2065.0,
      "completions/mean_length": 1143.5,
      "completions/mean_terminated_length": 1143.5,
      "completions/min_length": 831.0,
      "completions/min_terminated_length": 831.0,
      "epoch": 0.53125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2916565239429474,
      "kl": 0.0005820317019242793,
      "learning_rate": 1.443454345648252e-06,
      "loss": 0.0,
      "num_tokens": 7088720.0,
      "reward": 0.0656561329960823,
      "reward_std": 0.6580207943916321,
      "rewards/reward_function/mean": 0.0656561329960823,
      "rewards/reward_function/std": 0.6580208539962769,
      "step": 136
    },
    {
      "completion_length": 1176.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1339.0,
      "completions/max_terminated_length": 1339.0,
      "completions/mean_length": 1176.875,
      "completions/mean_terminated_length": 1176.875,
      "completions/min_length": 1011.0,
      "completions/min_terminated_length": 1011.0,
      "epoch": 0.53515625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3021836280822754,
      "kl": 0.00046610408753622323,
      "learning_rate": 1.4040721330273063e-06,
      "loss": 0.0,
      "num_tokens": 7143431.0,
      "reward": -0.28037768602371216,
      "reward_std": 0.7693378329277039,
      "rewards/reward_function/mean": -0.28037768602371216,
      "rewards/reward_function/std": 0.7693378329277039,
      "step": 137
    },
    {
      "completion_length": 1035.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1316.0,
      "completions/max_terminated_length": 1316.0,
      "completions/mean_length": 1035.125,
      "completions/mean_terminated_length": 1035.125,
      "completions/min_length": 822.0,
      "completions/min_terminated_length": 822.0,
      "epoch": 0.5390625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.32248759269714355,
      "kl": 0.00046888978249626234,
      "learning_rate": 1.3650237506511333e-06,
      "loss": 0.0,
      "num_tokens": 7194392.0,
      "reward": 0.250247597694397,
      "reward_std": 0.5053042769432068,
      "rewards/reward_function/mean": 0.250247597694397,
      "rewards/reward_function/std": 0.5053042769432068,
      "step": 138
    },
    {
      "completion_length": 1030.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1268.0,
      "completions/max_terminated_length": 1268.0,
      "completions/mean_length": 1030.25,
      "completions/mean_terminated_length": 1030.25,
      "completions/min_length": 778.0,
      "completions/min_terminated_length": 778.0,
      "epoch": 0.54296875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3456158936023712,
      "kl": 0.0005144016322446987,
      "learning_rate": 1.3263210930352737e-06,
      "loss": 0.0,
      "num_tokens": 7247930.0,
      "reward": 0.2453567385673523,
      "reward_std": 0.503298282623291,
      "rewards/reward_function/mean": 0.2453567385673523,
      "rewards/reward_function/std": 0.503298282623291,
      "step": 139
    },
    {
      "completion_length": 1074.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1220.0,
      "completions/max_terminated_length": 1220.0,
      "completions/mean_length": 1074.125,
      "completions/mean_terminated_length": 1074.125,
      "completions/min_length": 887.0,
      "completions/min_terminated_length": 887.0,
      "epoch": 0.546875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2869755029678345,
      "kl": 0.00043533078132895753,
      "learning_rate": 1.2879759493841577e-06,
      "loss": 0.0,
      "num_tokens": 7299203.0,
      "reward": -0.10356661677360535,
      "reward_std": 0.7423522472381592,
      "rewards/reward_function/mean": -0.10356661677360535,
      "rewards/reward_function/std": 0.7423522472381592,
      "step": 140
    },
    {
      "completion_length": 1284.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1708.0,
      "completions/max_terminated_length": 1708.0,
      "completions/mean_length": 1284.0,
      "completions/mean_terminated_length": 1284.0,
      "completions/min_length": 1076.0,
      "completions/min_terminated_length": 1076.0,
      "epoch": 0.55078125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2680242955684662,
      "kl": 0.000499433335789945,
      "learning_rate": 1.2500000000000007e-06,
      "loss": 0.0,
      "num_tokens": 7354771.0,
      "reward": 0.014156922698020935,
      "reward_std": 0.6485656499862671,
      "rewards/reward_function/mean": 0.014156922698020935,
      "rewards/reward_function/std": 0.6485656499862671,
      "step": 141
    },
    {
      "completion_length": 1086.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1583.0,
      "completions/max_terminated_length": 1583.0,
      "completions/mean_length": 1086.625,
      "completions/mean_terminated_length": 1086.625,
      "completions/min_length": 876.0,
      "completions/min_terminated_length": 876.0,
      "epoch": 0.5546875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.28385043144226074,
      "kl": 0.0004665546293836087,
      "learning_rate": 1.2124048127248644e-06,
      "loss": 0.0,
      "num_tokens": 7406144.0,
      "reward": 0.07084375619888306,
      "reward_std": 0.6610893607139587,
      "rewards/reward_function/mean": 0.07084375619888306,
      "rewards/reward_function/std": 0.661089301109314,
      "step": 142
    },
    {
      "completion_length": 1445.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1970.0,
      "completions/max_terminated_length": 1970.0,
      "completions/mean_length": 1445.0,
      "completions/mean_terminated_length": 1445.0,
      "completions/min_length": 1009.0,
      "completions/min_terminated_length": 1009.0,
      "epoch": 0.55859375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.27446597814559937,
      "kl": 0.0005598404095508158,
      "learning_rate": 1.1752018394169882e-06,
      "loss": 0.0,
      "num_tokens": 7463000.0,
      "reward": -0.09585624933242798,
      "reward_std": 0.7488118410110474,
      "rewards/reward_function/mean": -0.09585624933242798,
      "rewards/reward_function/std": 0.7488118410110474,
      "step": 143
    },
    {
      "completion_length": 1051.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1423.0,
      "completions/max_terminated_length": 1423.0,
      "completions/mean_length": 1051.5,
      "completions/mean_terminated_length": 1051.5,
      "completions/min_length": 884.0,
      "completions/min_terminated_length": 884.0,
      "epoch": 0.5625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.28686314821243286,
      "kl": 0.000490581318445038,
      "learning_rate": 1.1384024124624324e-06,
      "loss": 0.0,
      "num_tokens": 7514092.0,
      "reward": 0.25037020444869995,
      "reward_std": 0.5052564740180969,
      "rewards/reward_function/mean": 0.25037020444869995,
      "rewards/reward_function/std": 0.5052564740180969,
      "step": 144
    },
    {
      "completion_length": 910.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1172.0,
      "completions/max_terminated_length": 1172.0,
      "completions/mean_length": 910.25,
      "completions/mean_terminated_length": 910.25,
      "completions/min_length": 111.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.56640625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 2.037623882293701,
      "kl": 0.0005943369615124539,
      "learning_rate": 1.1020177413231334e-06,
      "loss": 0.0,
      "num_tokens": 7570974.0,
      "reward": 0.25001391768455505,
      "reward_std": 0.5055700540542603,
      "rewards/reward_function/mean": 0.25001391768455505,
      "rewards/reward_function/std": 0.505570113658905,
      "step": 145
    },
    {
      "completion_length": 1040.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1460.0,
      "completions/max_terminated_length": 1460.0,
      "completions/mean_length": 1040.5,
      "completions/mean_terminated_length": 1040.5,
      "completions/min_length": 847.0,
      "completions/min_terminated_length": 847.0,
      "epoch": 0.5703125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4468381106853485,
      "kl": 0.0005619311414193362,
      "learning_rate": 1.0660589091223854e-06,
      "loss": 0.0,
      "num_tokens": 7620194.0,
      "reward": 0.23401594161987305,
      "reward_std": 0.49918097257614136,
      "rewards/reward_function/mean": 0.23401594161987305,
      "rewards/reward_function/std": 0.49918097257614136,
      "step": 146
    },
    {
      "completion_length": 1247.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1814.0,
      "completions/max_terminated_length": 1814.0,
      "completions/mean_length": 1247.5,
      "completions/mean_terminated_length": 1247.5,
      "completions/min_length": 901.0,
      "completions/min_terminated_length": 901.0,
      "epoch": 0.57421875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3102348744869232,
      "kl": 0.0004937511766911484,
      "learning_rate": 1.0305368692688175e-06,
      "loss": 0.0,
      "num_tokens": 7679774.0,
      "reward": 0.08763788640499115,
      "reward_std": 0.671554684638977,
      "rewards/reward_function/mean": 0.08763788640499115,
      "rewards/reward_function/std": 0.671554684638977,
      "step": 147
    },
    {
      "completion_length": 1088.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1270.0,
      "completions/max_terminated_length": 1270.0,
      "completions/mean_length": 1088.375,
      "completions/mean_terminated_length": 1088.375,
      "completions/min_length": 955.0,
      "completions/min_terminated_length": 955.0,
      "epoch": 0.578125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3440247178077698,
      "kl": 0.000519426612299867,
      "learning_rate": 9.95462442119879e-07,
      "loss": 0.0,
      "num_tokens": 7729377.0,
      "reward": 0.004164457321166992,
      "reward_std": 0.6441723108291626,
      "rewards/reward_function/mean": 0.004164457321166992,
      "rewards/reward_function/std": 0.6441722512245178,
      "step": 148
    },
    {
      "completion_length": 1232.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1499.0,
      "completions/max_terminated_length": 1499.0,
      "completions/mean_length": 1232.0,
      "completions/mean_terminated_length": 1232.0,
      "completions/min_length": 906.0,
      "completions/min_terminated_length": 906.0,
      "epoch": 0.58203125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2699142098426819,
      "kl": 0.0005399520887294784,
      "learning_rate": 9.608463116858544e-07,
      "loss": 0.0,
      "num_tokens": 7788833.0,
      "reward": -0.1629893034696579,
      "reward_std": 0.7170795202255249,
      "rewards/reward_function/mean": -0.1629893034696579,
      "rewards/reward_function/std": 0.7170795202255249,
      "step": 149
    },
    {
      "completion_length": 1058.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1149.0,
      "completions/max_terminated_length": 1149.0,
      "completions/mean_length": 1058.25,
      "completions/mean_terminated_length": 1058.25,
      "completions/min_length": 867.0,
      "completions/min_terminated_length": 867.0,
      "epoch": 0.5859375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2994002103805542,
      "kl": 0.0004978677097824402,
      "learning_rate": 9.266990223754069e-07,
      "loss": 0.0,
      "num_tokens": 7838195.0,
      "reward": -0.6465986371040344,
      "reward_std": 0.6543768048286438,
      "rewards/reward_function/mean": -0.6465986371040344,
      "rewards/reward_function/std": 0.6543768048286438,
      "step": 150
    },
    {
      "completion_length": 1092.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1344.0,
      "completions/max_terminated_length": 1344.0,
      "completions/mean_length": 1092.375,
      "completions/mean_terminated_length": 1092.375,
      "completions/min_length": 787.0,
      "completions/min_terminated_length": 787.0,
      "epoch": 0.58984375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.30443111062049866,
      "kl": 0.000516350322868675,
      "learning_rate": 8.930309757836517e-07,
      "loss": 0.0,
      "num_tokens": 7896534.0,
      "reward": 0.008587591350078583,
      "reward_std": 0.6500032544136047,
      "rewards/reward_function/mean": 0.008587591350078583,
      "rewards/reward_function/std": 0.6500033140182495,
      "step": 151
    },
    {
      "completion_length": 969.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1327.0,
      "completions/max_terminated_length": 1327.0,
      "completions/mean_length": 969.875,
      "completions/mean_terminated_length": 969.875,
      "completions/min_length": 94.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.59375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.7537317276000977,
      "kl": 0.0005359335409593768,
      "learning_rate": 8.598524275237321e-07,
      "loss": 0.0,
      "num_tokens": 7945189.0,
      "reward": -0.24850904941558838,
      "reward_std": 0.6657283306121826,
      "rewards/reward_function/mean": -0.24850904941558838,
      "rewards/reward_function/std": 0.6657283306121826,
      "step": 152
    },
    {
      "completion_length": 1033.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1727.0,
      "completions/max_terminated_length": 1727.0,
      "completions/mean_length": 1033.625,
      "completions/mean_terminated_length": 1033.625,
      "completions/min_length": 667.0,
      "completions/min_terminated_length": 667.0,
      "epoch": 0.59765625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34412840008735657,
      "kl": 0.000547852658201009,
      "learning_rate": 8.271734841028553e-07,
      "loss": 0.0,
      "num_tokens": 7996642.0,
      "reward": 0.16972142457962036,
      "reward_std": 0.5201033353805542,
      "rewards/reward_function/mean": 0.16972142457962036,
      "rewards/reward_function/std": 0.5201033353805542,
      "step": 153
    },
    {
      "completion_length": 828.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 993.0,
      "completions/max_terminated_length": 993.0,
      "completions/mean_length": 828.875,
      "completions/mean_terminated_length": 828.875,
      "completions/min_length": 545.0,
      "completions/min_terminated_length": 545.0,
      "epoch": 0.6015625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3883880376815796,
      "kl": 0.0005664439522661269,
      "learning_rate": 7.950040998437541e-07,
      "loss": 0.0,
      "num_tokens": 8046737.0,
      "reward": -0.36253300309181213,
      "reward_std": 0.7088689208030701,
      "rewards/reward_function/mean": -0.36253300309181213,
      "rewards/reward_function/std": 0.7088689208030701,
      "step": 154
    },
    {
      "completion_length": 1078.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1602.0,
      "completions/max_terminated_length": 1602.0,
      "completions/mean_length": 1078.375,
      "completions/mean_terminated_length": 1078.375,
      "completions/min_length": 809.0,
      "completions/min_terminated_length": 809.0,
      "epoch": 0.60546875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3525019586086273,
      "kl": 0.000516050225996878,
      "learning_rate": 7.633540738525066e-07,
      "loss": 0.0,
      "num_tokens": 8098548.0,
      "reward": 0.010962013155221939,
      "reward_std": 0.6493006348609924,
      "rewards/reward_function/mean": 0.010962013155221939,
      "rewards/reward_function/std": 0.6493006944656372,
      "step": 155
    },
    {
      "completion_length": 900.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1039.0,
      "completions/max_terminated_length": 1039.0,
      "completions/mean_length": 900.625,
      "completions/mean_terminated_length": 900.625,
      "completions/min_length": 755.0,
      "completions/min_terminated_length": 755.0,
      "epoch": 0.609375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4146842658519745,
      "kl": 0.0006649910937994719,
      "learning_rate": 7.322330470336314e-07,
      "loss": 0.0,
      "num_tokens": 8149217.0,
      "reward": 0.07593274116516113,
      "reward_std": 0.6641193628311157,
      "rewards/reward_function/mean": 0.07593274116516113,
      "rewards/reward_function/std": 0.6641193628311157,
      "step": 156
    },
    {
      "completion_length": 794.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 999.0,
      "completions/max_terminated_length": 999.0,
      "completions/mean_length": 794.0,
      "completions/mean_terminated_length": 794.0,
      "completions/min_length": 129.0,
      "completions/min_terminated_length": 129.0,
      "epoch": 0.61328125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.9788252711296082,
      "kl": 0.000908300731680356,
      "learning_rate": 7.016504991533727e-07,
      "loss": 0.0,
      "num_tokens": 8198753.0,
      "reward": -0.28955256938934326,
      "reward_std": 0.7595077753067017,
      "rewards/reward_function/mean": -0.28955256938934326,
      "rewards/reward_function/std": 0.7595077753067017,
      "step": 157
    },
    {
      "completion_length": 859.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1018.0,
      "completions/max_terminated_length": 1018.0,
      "completions/mean_length": 859.25,
      "completions/mean_terminated_length": 859.25,
      "completions/min_length": 712.0,
      "completions/min_terminated_length": 712.0,
      "epoch": 0.6171875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3605237305164337,
      "kl": 0.0005935588560532779,
      "learning_rate": 6.716157459520739e-07,
      "loss": 0.0,
      "num_tokens": 8249091.0,
      "reward": 0.25031226873397827,
      "reward_std": 0.5052971839904785,
      "rewards/reward_function/mean": 0.25031226873397827,
      "rewards/reward_function/std": 0.5052971839904785,
      "step": 158
    },
    {
      "completion_length": 1081.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1559.0,
      "completions/max_terminated_length": 1559.0,
      "completions/mean_length": 1081.75,
      "completions/mean_terminated_length": 1081.75,
      "completions/min_length": 846.0,
      "completions/min_terminated_length": 846.0,
      "epoch": 0.62109375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.39315265417099,
      "kl": 0.0004596119179041125,
      "learning_rate": 6.421379363065142e-07,
      "loss": 0.0,
      "num_tokens": 8300929.0,
      "reward": 0.24795930087566376,
      "reward_std": 0.5042985677719116,
      "rewards/reward_function/mean": 0.24795930087566376,
      "rewards/reward_function/std": 0.5042985677719116,
      "step": 159
    },
    {
      "completion_length": 1107.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1957.0,
      "completions/max_terminated_length": 1957.0,
      "completions/mean_length": 1107.25,
      "completions/mean_terminated_length": 1107.25,
      "completions/min_length": 792.0,
      "completions/min_terminated_length": 792.0,
      "epoch": 0.625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.352717787027359,
      "kl": 0.0006316443686955608,
      "learning_rate": 6.1322604944307e-07,
      "loss": 0.0,
      "num_tokens": 8353251.0,
      "reward": -0.1729724407196045,
      "reward_std": 0.7113326787948608,
      "rewards/reward_function/mean": -0.1729724407196045,
      "rewards/reward_function/std": 0.7113326787948608,
      "step": 160
    },
    {
      "completion_length": 1091.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1355.0,
      "completions/max_terminated_length": 1355.0,
      "completions/mean_length": 1091.875,
      "completions/mean_terminated_length": 1091.875,
      "completions/min_length": 734.0,
      "completions/min_terminated_length": 734.0,
      "epoch": 0.62890625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.303595632314682,
      "kl": 0.0004378617668407969,
      "learning_rate": 5.848888922025553e-07,
      "loss": 0.0,
      "num_tokens": 8406186.0,
      "reward": -0.46176353096961975,
      "reward_std": 0.7428471446037292,
      "rewards/reward_function/mean": -0.46176353096961975,
      "rewards/reward_function/std": 0.742847204208374,
      "step": 161
    },
    {
      "completion_length": 1162.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1402.0,
      "completions/max_terminated_length": 1402.0,
      "completions/mean_length": 1162.75,
      "completions/mean_terminated_length": 1162.75,
      "completions/min_length": 173.0,
      "completions/min_terminated_length": 173.0,
      "epoch": 0.6328125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.2087342739105225,
      "kl": 0.0008970491835498251,
      "learning_rate": 5.571350963575728e-07,
      "loss": 0.0,
      "num_tokens": 8459736.0,
      "reward": -0.175228551030159,
      "reward_std": 0.7207473516464233,
      "rewards/reward_function/mean": -0.175228551030159,
      "rewards/reward_function/std": 0.7207473516464233,
      "step": 162
    },
    {
      "completion_length": 1177.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1390.0,
      "completions/max_terminated_length": 1390.0,
      "completions/mean_length": 1177.125,
      "completions/mean_terminated_length": 1177.125,
      "completions/min_length": 978.0,
      "completions/min_terminated_length": 978.0,
      "epoch": 0.63671875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29434847831726074,
      "kl": 0.0004823799536097795,
      "learning_rate": 5.299731159831953e-07,
      "loss": 0.0,
      "num_tokens": 8513353.0,
      "reward": 0.08043806254863739,
      "reward_std": 0.6669427156448364,
      "rewards/reward_function/mean": 0.08043806254863739,
      "rewards/reward_function/std": 0.6669427156448364,
      "step": 163
    },
    {
      "completion_length": 1363.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1791.0,
      "completions/max_terminated_length": 1791.0,
      "completions/mean_length": 1363.25,
      "completions/mean_terminated_length": 1363.25,
      "completions/min_length": 1010.0,
      "completions/min_terminated_length": 1010.0,
      "epoch": 0.640625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.27553999423980713,
      "kl": 0.0005040961696067825,
      "learning_rate": 5.034112248817685e-07,
      "loss": 0.0,
      "num_tokens": 8568507.0,
      "reward": 0.09454244375228882,
      "reward_std": 0.6757357120513916,
      "rewards/reward_function/mean": 0.09454244375228882,
      "rewards/reward_function/std": 0.6757358312606812,
      "step": 164
    },
    {
      "completion_length": 1147.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1400.0,
      "completions/max_terminated_length": 1400.0,
      "completions/mean_length": 1147.0,
      "completions/mean_terminated_length": 1147.0,
      "completions/min_length": 667.0,
      "completions/min_terminated_length": 667.0,
      "epoch": 0.64453125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.31530848145484924,
      "kl": 0.0004901376814814284,
      "learning_rate": 4.774575140626317e-07,
      "loss": 0.0,
      "num_tokens": 8621883.0,
      "reward": -0.2817423939704895,
      "reward_std": 0.7679179310798645,
      "rewards/reward_function/mean": -0.2817423939704895,
      "rewards/reward_function/std": 0.7679178714752197,
      "step": 165
    },
    {
      "completion_length": 1058.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1364.0,
      "completions/max_terminated_length": 1364.0,
      "completions/mean_length": 1058.0,
      "completions/mean_terminated_length": 1058.0,
      "completions/min_length": 522.0,
      "completions/min_terminated_length": 522.0,
      "epoch": 0.6484375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.35542306303977966,
      "kl": 0.000599144957959652,
      "learning_rate": 4.5211988927752026e-07,
      "loss": 0.0,
      "num_tokens": 8674595.0,
      "reward": 0.014520317316055298,
      "reward_std": 0.6543052196502686,
      "rewards/reward_function/mean": 0.014520317316055298,
      "rewards/reward_function/std": 0.6543052196502686,
      "step": 166
    },
    {
      "completion_length": 1177.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2078.0,
      "completions/max_terminated_length": 2078.0,
      "completions/mean_length": 1177.875,
      "completions/mean_terminated_length": 1177.875,
      "completions/min_length": 863.0,
      "completions/min_terminated_length": 863.0,
      "epoch": 0.65234375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.290801465511322,
      "kl": 0.000524831673828885,
      "learning_rate": 4.27406068612396e-07,
      "loss": 0.0,
      "num_tokens": 8728218.0,
      "reward": -0.28122541308403015,
      "reward_std": 0.7684498429298401,
      "rewards/reward_function/mean": -0.28122541308403015,
      "rewards/reward_function/std": 0.7684498429298401,
      "step": 167
    },
    {
      "completion_length": 1099.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1303.0,
      "completions/max_terminated_length": 1303.0,
      "completions/mean_length": 1099.375,
      "completions/mean_terminated_length": 1099.375,
      "completions/min_length": 820.0,
      "completions/min_terminated_length": 820.0,
      "epoch": 0.65625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.2773479223251343,
      "kl": 0.00045157810382079333,
      "learning_rate": 4.033235801364402e-07,
      "loss": 0.0,
      "num_tokens": 8781261.0,
      "reward": 0.0877399668097496,
      "reward_std": 0.671479344367981,
      "rewards/reward_function/mean": 0.0877399668097496,
      "rewards/reward_function/std": 0.671479344367981,
      "step": 168
    },
    {
      "completion_length": 1257.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2235.0,
      "completions/max_terminated_length": 2235.0,
      "completions/mean_length": 1257.25,
      "completions/mean_terminated_length": 1257.25,
      "completions/min_length": 939.0,
      "completions/min_terminated_length": 939.0,
      "epoch": 0.66015625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.29819291830062866,
      "kl": 0.00048441492253914475,
      "learning_rate": 3.798797596089351e-07,
      "loss": 0.0,
      "num_tokens": 8838007.0,
      "reward": -0.043121978640556335,
      "reward_std": 0.630191445350647,
      "rewards/reward_function/mean": -0.043121978640556335,
      "rewards/reward_function/std": 0.630191445350647,
      "step": 169
    },
    {
      "completion_length": 1023.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1202.0,
      "completions/max_terminated_length": 1202.0,
      "completions/mean_length": 1023.0,
      "completions/mean_terminated_length": 1023.0,
      "completions/min_length": 802.0,
      "completions/min_terminated_length": 802.0,
      "epoch": 0.6640625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.338838666677475,
      "kl": 0.000568984636629466,
      "learning_rate": 3.5708174824471947e-07,
      "loss": 0.0,
      "num_tokens": 8887223.0,
      "reward": 0.05965143442153931,
      "reward_std": 0.6547546982765198,
      "rewards/reward_function/mean": 0.05965143442153931,
      "rewards/reward_function/std": 0.654754638671875,
      "step": 170
    },
    {
      "completion_length": 1429.75,
      "completions/clipped_ratio": 0.125,
      "completions/max_length": 3500.0,
      "completions/max_terminated_length": 1337.0,
      "completions/mean_length": 1429.75,
      "completions/mean_terminated_length": 1134.0,
      "completions/min_length": 860.0,
      "completions/min_terminated_length": 860.0,
      "epoch": 0.66796875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.27894267439842224,
      "kl": 0.0004727757041109726,
      "learning_rate": 3.3493649053890325e-07,
      "loss": 0.0,
      "num_tokens": 8945349.0,
      "reward": -0.09473268687725067,
      "reward_std": 0.7497490644454956,
      "rewards/reward_function/mean": -0.09473268687725067,
      "rewards/reward_function/std": 0.7497490048408508,
      "step": 171
    },
    {
      "completion_length": 1153.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1421.0,
      "completions/max_terminated_length": 1421.0,
      "completions/mean_length": 1153.25,
      "completions/mean_terminated_length": 1153.25,
      "completions/min_length": 959.0,
      "completions/min_terminated_length": 959.0,
      "epoch": 0.671875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3112727403640747,
      "kl": 0.0005989774799672887,
      "learning_rate": 3.134507321515107e-07,
      "loss": 0.0,
      "num_tokens": 8995607.0,
      "reward": -0.17081062495708466,
      "reward_std": 0.7099243998527527,
      "rewards/reward_function/mean": -0.17081062495708466,
      "rewards/reward_function/std": 0.7099243998527527,
      "step": 172
    },
    {
      "completion_length": 1032.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1514.0,
      "completions/max_terminated_length": 1514.0,
      "completions/mean_length": 1032.125,
      "completions/mean_terminated_length": 1032.125,
      "completions/min_length": 237.0,
      "completions/min_terminated_length": 237.0,
      "epoch": 0.67578125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.9038693308830261,
      "kl": 0.0006451614681282081,
      "learning_rate": 2.9263101785268253e-07,
      "loss": 0.0,
      "num_tokens": 9050552.0,
      "reward": 0.08472549915313721,
      "reward_std": 0.669560432434082,
      "rewards/reward_function/mean": 0.08472549915313721,
      "rewards/reward_function/std": 0.669560432434082,
      "step": 173
    },
    {
      "completion_length": 934.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1022.0,
      "completions/max_terminated_length": 1022.0,
      "completions/mean_length": 934.125,
      "completions/mean_terminated_length": 934.125,
      "completions/min_length": 730.0,
      "completions/min_terminated_length": 730.0,
      "epoch": 0.6796875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3780133128166199,
      "kl": 0.0005501867926795967,
      "learning_rate": 2.7248368952908055e-07,
      "loss": 0.0,
      "num_tokens": 9099057.0,
      "reward": 0.24916215240955353,
      "reward_std": 0.5048820376396179,
      "rewards/reward_function/mean": 0.24916215240955353,
      "rewards/reward_function/std": 0.5048820972442627,
      "step": 174
    },
    {
      "completion_length": 1211.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1936.0,
      "completions/max_terminated_length": 1936.0,
      "completions/mean_length": 1211.375,
      "completions/mean_terminated_length": 1211.375,
      "completions/min_length": 953.0,
      "completions/min_terminated_length": 953.0,
      "epoch": 0.68359375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.36097532510757446,
      "kl": 0.0005826732158311643,
      "learning_rate": 2.53014884252083e-07,
      "loss": 0.0,
      "num_tokens": 9155436.0,
      "reward": 0.2617029845714569,
      "reward_std": 0.5099496245384216,
      "rewards/reward_function/mean": 0.2617029845714569,
      "rewards/reward_function/std": 0.5099496245384216,
      "step": 175
    },
    {
      "completion_length": 1103.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1672.0,
      "completions/max_terminated_length": 1672.0,
      "completions/mean_length": 1103.25,
      "completions/mean_terminated_length": 1103.25,
      "completions/min_length": 870.0,
      "completions/min_terminated_length": 870.0,
      "epoch": 0.6875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.30660855770111084,
      "kl": 0.000493127474328503,
      "learning_rate": 2.3423053240837518e-07,
      "loss": 0.0,
      "num_tokens": 9205294.0,
      "reward": -0.10572409629821777,
      "reward_std": 0.7405680418014526,
      "rewards/reward_function/mean": -0.10572409629821777,
      "rewards/reward_function/std": 0.7405680418014526,
      "step": 176
    },
    {
      "completion_length": 964.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1215.0,
      "completions/max_terminated_length": 1215.0,
      "completions/mean_length": 964.5,
      "completions/mean_terminated_length": 964.5,
      "completions/min_length": 851.0,
      "completions/min_terminated_length": 851.0,
      "epoch": 0.69140625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3269523084163666,
      "kl": 0.00049879898870131,
      "learning_rate": 2.1613635589349756e-07,
      "loss": 0.0,
      "num_tokens": 9261322.0,
      "reward": -0.4611111581325531,
      "reward_std": 0.743747889995575,
      "rewards/reward_function/mean": -0.4611111581325531,
      "rewards/reward_function/std": 0.743747889995575,
      "step": 177
    },
    {
      "completion_length": 1027.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1438.0,
      "completions/max_terminated_length": 1438.0,
      "completions/mean_length": 1027.625,
      "completions/mean_terminated_length": 1027.625,
      "completions/min_length": 438.0,
      "completions/min_terminated_length": 438.0,
      "epoch": 0.6953125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.5114771127700806,
      "kl": 0.0005179118452360854,
      "learning_rate": 1.9873786636889908e-07,
      "loss": 0.0,
      "num_tokens": 9318415.0,
      "reward": 0.2489490509033203,
      "reward_std": 0.5048139095306396,
      "rewards/reward_function/mean": 0.2489490509033203,
      "rewards/reward_function/std": 0.5048139691352844,
      "step": 178
    },
    {
      "completion_length": 1043.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2555.0,
      "completions/max_terminated_length": 2555.0,
      "completions/mean_length": 1043.5,
      "completions/mean_terminated_length": 1043.5,
      "completions/min_length": 565.0,
      "completions/min_terminated_length": 565.0,
      "epoch": 0.69921875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.339715838432312,
      "kl": 0.0005509828915819526,
      "learning_rate": 1.8204036358303173e-07,
      "loss": 0.0,
      "num_tokens": 9375075.0,
      "reward": -0.17033275961875916,
      "reward_std": 0.7077411413192749,
      "rewards/reward_function/mean": -0.17033275961875916,
      "rewards/reward_function/std": 0.7077411413192749,
      "step": 179
    },
    {
      "completion_length": 1009.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1202.0,
      "completions/max_terminated_length": 1202.0,
      "completions/mean_length": 1009.5,
      "completions/mean_terminated_length": 1009.5,
      "completions/min_length": 620.0,
      "completions/min_terminated_length": 620.0,
      "epoch": 0.703125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.38060900568962097,
      "kl": 0.0005021176039008424,
      "learning_rate": 1.6604893375699594e-07,
      "loss": 0.0,
      "num_tokens": 9432023.0,
      "reward": 0.24431279301643372,
      "reward_std": 0.5027908682823181,
      "rewards/reward_function/mean": 0.24431279301643372,
      "rewards/reward_function/std": 0.5027908682823181,
      "step": 180
    },
    {
      "completion_length": 921.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1216.0,
      "completions/max_terminated_length": 1216.0,
      "completions/mean_length": 921.375,
      "completions/mean_terminated_length": 921.375,
      "completions/min_length": 742.0,
      "completions/min_terminated_length": 742.0,
      "epoch": 0.70703125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.405112624168396,
      "kl": 0.0005638250586343929,
      "learning_rate": 1.507684480352292e-07,
      "loss": 0.0,
      "num_tokens": 9487706.0,
      "reward": 0.2474900782108307,
      "reward_std": 0.5042015910148621,
      "rewards/reward_function/mean": 0.2474900782108307,
      "rewards/reward_function/std": 0.5042015910148621,
      "step": 181
    },
    {
      "completion_length": 1200.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2126.0,
      "completions/max_terminated_length": 2126.0,
      "completions/mean_length": 1200.25,
      "completions/mean_terminated_length": 1200.25,
      "completions/min_length": 815.0,
      "completions/min_terminated_length": 815.0,
      "epoch": 0.7109375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3271883726119995,
      "kl": 0.0006362074636854231,
      "learning_rate": 1.362035610017079e-07,
      "loss": 0.0,
      "num_tokens": 9546180.0,
      "reward": 0.05625756084918976,
      "reward_std": 0.6525231599807739,
      "rewards/reward_function/mean": 0.05625756084918976,
      "rewards/reward_function/std": 0.6525231599807739,
      "step": 182
    },
    {
      "completion_length": 926.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1297.0,
      "completions/max_terminated_length": 1297.0,
      "completions/mean_length": 926.0,
      "completions/mean_terminated_length": 926.0,
      "completions/min_length": 737.0,
      "completions/min_terminated_length": 737.0,
      "epoch": 0.71484375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34687018394470215,
      "kl": 0.0005171666198293678,
      "learning_rate": 1.223587092621162e-07,
      "loss": 0.0,
      "num_tokens": 9601900.0,
      "reward": 0.0704411044716835,
      "reward_std": 0.6608877182006836,
      "rewards/reward_function/mean": 0.0704411044716835,
      "rewards/reward_function/std": 0.6608877182006836,
      "step": 183
    },
    {
      "completion_length": 1005.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1611.0,
      "completions/max_terminated_length": 1611.0,
      "completions/mean_length": 1005.125,
      "completions/mean_terminated_length": 1005.125,
      "completions/min_length": 768.0,
      "completions/min_terminated_length": 768.0,
      "epoch": 0.71875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.35485363006591797,
      "kl": 0.0004917978003504686,
      "learning_rate": 1.0923811009241142e-07,
      "loss": 0.0,
      "num_tokens": 9658813.0,
      "reward": 0.2492428421974182,
      "reward_std": 0.5048900842666626,
      "rewards/reward_function/mean": 0.2492428421974182,
      "rewards/reward_function/std": 0.5048900842666626,
      "step": 184
    },
    {
      "completion_length": 990.625,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1283.0,
      "completions/max_terminated_length": 1283.0,
      "completions/mean_length": 990.625,
      "completions/mean_terminated_length": 990.625,
      "completions/min_length": 751.0,
      "completions/min_terminated_length": 751.0,
      "epoch": 0.72265625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.36808741092681885,
      "kl": 0.0005794434182462282,
      "learning_rate": 9.684576015420277e-08,
      "loss": 0.0,
      "num_tokens": 9708434.0,
      "reward": -0.28416281938552856,
      "reward_std": 0.7652737498283386,
      "rewards/reward_function/mean": -0.28416281938552856,
      "rewards/reward_function/std": 0.7652737498283386,
      "step": 185
    },
    {
      "completion_length": 1321.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2448.0,
      "completions/max_terminated_length": 2448.0,
      "completions/mean_length": 1321.0,
      "completions/mean_terminated_length": 1321.0,
      "completions/min_length": 747.0,
      "completions/min_terminated_length": 747.0,
      "epoch": 0.7265625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34419044852256775,
      "kl": 0.000558189676667098,
      "learning_rate": 8.518543427732951e-08,
      "loss": 0.0,
      "num_tokens": 9762722.0,
      "reward": -0.2857913076877594,
      "reward_std": 0.7635413408279419,
      "rewards/reward_function/mean": -0.2857913076877594,
      "rewards/reward_function/std": 0.7635413408279419,
      "step": 186
    },
    {
      "completion_length": 1260.375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1672.0,
      "completions/max_terminated_length": 1672.0,
      "completions/mean_length": 1260.375,
      "completions/mean_terminated_length": 1260.375,
      "completions/min_length": 913.0,
      "completions/min_terminated_length": 913.0,
      "epoch": 0.73046875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3106219470500946,
      "kl": 0.000590090683544986,
      "learning_rate": 7.426068431000883e-08,
      "loss": 0.0,
      "num_tokens": 9814501.0,
      "reward": -0.10060018301010132,
      "reward_std": 0.7447962164878845,
      "rewards/reward_function/mean": -0.10060018301010132,
      "rewards/reward_function/std": 0.7447961568832397,
      "step": 187
    },
    {
      "completion_length": 883.25,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1120.0,
      "completions/max_terminated_length": 1120.0,
      "completions/mean_length": 883.25,
      "completions/mean_terminated_length": 883.25,
      "completions/min_length": 576.0,
      "completions/min_terminated_length": 576.0,
      "epoch": 0.734375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.4308620095252991,
      "kl": 0.0005764737070421688,
      "learning_rate": 6.407483803691216e-08,
      "loss": 0.0,
      "num_tokens": 9865287.0,
      "reward": -0.11281464993953705,
      "reward_std": 0.7346830368041992,
      "rewards/reward_function/mean": -0.11281464993953705,
      "rewards/reward_function/std": 0.7346829771995544,
      "step": 188
    },
    {
      "completion_length": 987.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1191.0,
      "completions/max_terminated_length": 1191.0,
      "completions/mean_length": 987.0,
      "completions/mean_terminated_length": 987.0,
      "completions/min_length": 773.0,
      "completions/min_terminated_length": 773.0,
      "epoch": 0.73828125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.34522199630737305,
      "kl": 0.0005745336093241349,
      "learning_rate": 5.463099816548578e-08,
      "loss": -0.0,
      "num_tokens": 9914879.0,
      "reward": 0.43888330459594727,
      "reward_std": 0.013809502124786377,
      "rewards/reward_function/mean": 0.43888330459594727,
      "rewards/reward_function/std": 0.013809490017592907,
      "step": 189
    },
    {
      "completion_length": 1093.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1639.0,
      "completions/max_terminated_length": 1639.0,
      "completions/mean_length": 1093.0,
      "completions/mean_terminated_length": 1093.0,
      "completions/min_length": 577.0,
      "completions/min_terminated_length": 577.0,
      "epoch": 0.7421875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33963632583618164,
      "kl": 0.0006164438163978048,
      "learning_rate": 4.593204138084006e-08,
      "loss": 0.0,
      "num_tokens": 9967343.0,
      "reward": -0.4747491180896759,
      "reward_std": 0.7252565622329712,
      "rewards/reward_function/mean": -0.4747491180896759,
      "rewards/reward_function/std": 0.725256621837616,
      "step": 190
    },
    {
      "completion_length": 974.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1168.0,
      "completions/max_terminated_length": 1168.0,
      "completions/mean_length": 974.0,
      "completions/mean_terminated_length": 974.0,
      "completions/min_length": 856.0,
      "completions/min_terminated_length": 856.0,
      "epoch": 0.74609375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33787915110588074,
      "kl": 0.0005020303433411755,
      "learning_rate": 3.798061746947995e-08,
      "loss": 0.0,
      "num_tokens": 10016831.0,
      "reward": 0.07625642418861389,
      "reward_std": 0.6643387675285339,
      "rewards/reward_function/mean": 0.07625642418861389,
      "rewards/reward_function/std": 0.6643387675285339,
      "step": 191
    },
    {
      "completion_length": 1223.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1915.0,
      "completions/max_terminated_length": 1915.0,
      "completions/mean_length": 1223.5,
      "completions/mean_terminated_length": 1223.5,
      "completions/min_length": 377.0,
      "completions/min_terminated_length": 377.0,
      "epoch": 0.75,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3466210961341858,
      "kl": 0.0006273999460972846,
      "learning_rate": 3.077914851215585e-08,
      "loss": 0.0,
      "num_tokens": 10070339.0,
      "reward": -0.4622008502483368,
      "reward_std": 0.7422569990158081,
      "rewards/reward_function/mean": -0.4622008502483368,
      "rewards/reward_function/std": 0.7422570586204529,
      "step": 192
    },
    {
      "completion_length": 950.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1292.0,
      "completions/max_terminated_length": 1292.0,
      "completions/mean_length": 950.75,
      "completions/mean_terminated_length": 950.75,
      "completions/min_length": 504.0,
      "completions/min_terminated_length": 504.0,
      "epoch": 0.75390625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3800007402896881,
      "kl": 0.0005864633276360109,
      "learning_rate": 2.4329828146074096e-08,
      "loss": 0.0,
      "num_tokens": 10124977.0,
      "reward": -0.28742432594299316,
      "reward_std": 0.76183021068573,
      "rewards/reward_function/mean": -0.28742432594299316,
      "rewards/reward_function/std": 0.7618302702903748,
      "step": 193
    },
    {
      "completion_length": 977.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1252.0,
      "completions/max_terminated_length": 1252.0,
      "completions/mean_length": 977.0,
      "completions/mean_terminated_length": 977.0,
      "completions/min_length": 396.0,
      "completions/min_terminated_length": 396.0,
      "epoch": 0.7578125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.550014317035675,
      "kl": 0.0005290872577461414,
      "learning_rate": 1.8634620896695044e-08,
      "loss": 0.0,
      "num_tokens": 10175113.0,
      "reward": 0.24153178930282593,
      "reward_std": 0.5017148852348328,
      "rewards/reward_function/mean": 0.24153178930282593,
      "rewards/reward_function/std": 0.5017149448394775,
      "step": 194
    },
    {
      "completion_length": 1148.125,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1667.0,
      "completions/max_terminated_length": 1667.0,
      "completions/mean_length": 1148.125,
      "completions/mean_terminated_length": 1148.125,
      "completions/min_length": 707.0,
      "completions/min_terminated_length": 707.0,
      "epoch": 0.76171875,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3343985378742218,
      "kl": 0.0005916821173741482,
      "learning_rate": 1.3695261579316776e-08,
      "loss": 0.0,
      "num_tokens": 10231330.0,
      "reward": 0.09651876986026764,
      "reward_std": 0.511390209197998,
      "rewards/reward_function/mean": 0.09651876986026764,
      "rewards/reward_function/std": 0.511390209197998,
      "step": 195
    },
    {
      "completion_length": 1165.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1567.0,
      "completions/max_terminated_length": 1567.0,
      "completions/mean_length": 1165.5,
      "completions/mean_terminated_length": 1165.5,
      "completions/min_length": 1004.0,
      "completions/min_terminated_length": 1004.0,
      "epoch": 0.765625,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.33010783791542053,
      "kl": 0.0005538033583434299,
      "learning_rate": 9.513254770636138e-09,
      "loss": 0.0,
      "num_tokens": 10282974.0,
      "reward": -0.1761539876461029,
      "reward_std": 0.7018446922302246,
      "rewards/reward_function/mean": -0.1761539876461029,
      "rewards/reward_function/std": 0.7018447518348694,
      "step": 196
    },
    {
      "completion_length": 1035.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1474.0,
      "completions/max_terminated_length": 1474.0,
      "completions/mean_length": 1035.0,
      "completions/mean_terminated_length": 1035.0,
      "completions/min_length": 764.0,
      "completions/min_terminated_length": 764.0,
      "epoch": 0.76953125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.27371305227279663,
      "kl": 0.000488734214741271,
      "learning_rate": 6.089874350439507e-09,
      "loss": -0.0,
      "num_tokens": 10338286.0,
      "reward": 0.4277012050151825,
      "reward_std": 0.013616181910037994,
      "rewards/reward_function/mean": 0.4277012050151825,
      "rewards/reward_function/std": 0.013616186566650867,
      "step": 197
    },
    {
      "completion_length": 888.75,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1162.0,
      "completions/max_terminated_length": 1162.0,
      "completions/mean_length": 888.75,
      "completions/mean_terminated_length": 888.75,
      "completions/min_length": 248.0,
      "completions/min_terminated_length": 248.0,
      "epoch": 0.7734375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.8447083830833435,
      "kl": 0.0005265605868771672,
      "learning_rate": 3.4261631135654174e-09,
      "loss": 0.0,
      "num_tokens": 10387716.0,
      "reward": 0.40771353244781494,
      "reward_std": 0.02116406336426735,
      "rewards/reward_function/mean": 0.40771353244781494,
      "rewards/reward_function/std": 0.02116405963897705,
      "step": 198
    },
    {
      "completion_length": 1005.875,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1336.0,
      "completions/max_terminated_length": 1336.0,
      "completions/mean_length": 1005.875,
      "completions/mean_terminated_length": 1005.875,
      "completions/min_length": 863.0,
      "completions/min_terminated_length": 863.0,
      "epoch": 0.77734375,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.3126639127731323,
      "kl": 0.0005563631129916757,
      "learning_rate": 1.5229324522605949e-09,
      "loss": 0.0,
      "num_tokens": 10442795.0,
      "reward": 0.24751342833042145,
      "reward_std": 0.5041323304176331,
      "rewards/reward_function/mean": 0.24751342833042145,
      "rewards/reward_function/std": 0.5041323304176331,
      "step": 199
    },
    {
      "completion_length": 950.5,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1109.0,
      "completions/max_terminated_length": 1109.0,
      "completions/mean_length": 950.5,
      "completions/mean_terminated_length": 950.5,
      "completions/min_length": 681.0,
      "completions/min_terminated_length": 681.0,
      "epoch": 0.78125,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.40675655007362366,
      "kl": 0.0006997519376454875,
      "learning_rate": 3.8076210902182607e-10,
      "loss": 0.0,
      "num_tokens": 10492719.0,
      "reward": -0.18069900572299957,
      "reward_std": 0.7014786601066589,
      "rewards/reward_function/mean": -0.18069900572299957,
      "rewards/reward_function/std": 0.7014786601066589,
      "step": 200
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 10492719,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}