trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.18,
  "eval_steps": 500,
  "global_step": 450,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 294.14584255218506,
      "epoch": 0.0008,
      "grad_norm": 3.741899013519287,
      "kl": 0.0,
      "learning_rate": 7.142857142857142e-08,
      "loss": 0.1235,
      "num_tokens": 77390.0,
      "reward": -0.44646846456453204,
      "reward_std": 0.05219822392973583,
      "rewards/SMILES_validity_reward": -0.5666666901670396,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.08333333698101342,
      "rewards/repetition_penalty_reward": -0.06572717940434813,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0,
      "step": 2
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 314.82292652130127,
      "epoch": 0.0016,
      "grad_norm": 3.488884925842285,
      "kl": 0.0005727410316467285,
      "learning_rate": 1.4285714285714285e-07,
      "loss": 0.1119,
      "num_tokens": 156765.0,
      "reward": -0.45767286978662014,
      "reward_std": 0.04655311269743834,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.05902777938172221,
      "rewards/repetition_penalty_reward": -0.04981988831423223,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 4
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 329.7812581062317,
      "epoch": 0.0024,
      "grad_norm": 5.260374069213867,
      "kl": 0.0007290840148925781,
      "learning_rate": 2.1428571428571426e-07,
      "loss": 0.1747,
      "num_tokens": 237576.0,
      "reward": -0.4684038292616606,
      "reward_std": 0.017661503254203126,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.07291666837409139,
      "rewards/repetition_penalty_reward": -0.05695587984519079,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 6
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 227.34375858306885,
      "epoch": 0.0032,
      "grad_norm": 4.617069244384766,
      "kl": 0.0006937384605407715,
      "learning_rate": 2.857142857142857e-07,
      "loss": 0.1561,
      "num_tokens": 308553.0,
      "reward": -0.4692091681063175,
      "reward_std": 0.013479128843755461,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.04513889038935304,
      "rewards/repetition_penalty_reward": -0.037231406051432714,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 8
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 290.1770935058594,
      "epoch": 0.004,
      "grad_norm": 3.2441182136535645,
      "kl": 0.0006629228591918945,
      "learning_rate": 3.5714285714285716e-07,
      "loss": 0.0965,
      "num_tokens": 385562.0,
      "reward": -0.46172465570271015,
      "reward_std": 0.031036741798743606,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.11111111380159855,
      "rewards/repetition_penalty_reward": -0.05700426420662552,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 316.9166741371155,
      "epoch": 0.0048,
      "grad_norm": 4.713688850402832,
      "kl": 0.0006747245788574219,
      "learning_rate": 4.285714285714285e-07,
      "loss": 0.1268,
      "num_tokens": 465138.0,
      "reward": -0.4592954497784376,
      "reward_std": 0.041717870262800716,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.027777778217568994,
      "rewards/repetition_penalty_reward": -0.053024730485049076,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.02083333395421505,
      "step": 12
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 291.4479274749756,
      "epoch": 0.0056,
      "grad_norm": 3.528653621673584,
      "kl": 0.0005544424057006836,
      "learning_rate": 5e-07,
      "loss": 0.0348,
      "num_tokens": 542269.0,
      "reward": -0.46717050671577454,
      "reward_std": 0.020807951921597123,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.07986111333593726,
      "rewards/repetition_penalty_reward": -0.05156701215310022,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 14
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 268.8750042915344,
      "epoch": 0.0064,
      "grad_norm": 3.8491642475128174,
      "kl": 0.0007927417755126953,
      "learning_rate": 4.999740409224932e-07,
      "loss": 0.139,
      "num_tokens": 617233.0,
      "reward": -0.4442032668739557,
      "reward_std": 0.06988084255863214,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.07986111240461469,
      "rewards/repetition_penalty_reward": -0.047415557550266385,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 16
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 267.59375953674316,
      "epoch": 0.0072,
      "grad_norm": 4.2977681159973145,
      "kl": 0.0007317066192626953,
      "learning_rate": 4.998961690809627e-07,
      "loss": 0.038,
      "num_tokens": 692074.0,
      "reward": -0.4525126740336418,
      "reward_std": 0.05021150383981876,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.11458333558402956,
      "rewards/repetition_penalty_reward": -0.053773476160131395,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 18
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 263.87500870227814,
      "epoch": 0.008,
      "grad_norm": 3.244373321533203,
      "kl": 0.0007187128067016602,
      "learning_rate": 4.997664006472578e-07,
      "loss": 0.0803,
      "num_tokens": 766558.0,
      "reward": -0.46448596753180027,
      "reward_std": 0.02012665472284425,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.09375000186264515,
      "rewards/repetition_penalty_reward": -0.04381883708992973,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.27083921432495,
      "epoch": 0.0088,
      "grad_norm": 4.975526332855225,
      "kl": 0.001035928726196289,
      "learning_rate": 4.995847625707292e-07,
      "loss": 0.0001,
      "num_tokens": 839640.0,
      "reward": -0.4649157803505659,
      "reward_std": 0.019527284224750474,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.1006944456603378,
      "rewards/repetition_penalty_reward": -0.03943644787068479,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 22
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 290.88542461395264,
      "epoch": 0.0096,
      "grad_norm": 4.3505401611328125,
      "kl": 0.0008978843688964844,
      "learning_rate": 4.993512925726318e-07,
      "loss": 0.1512,
      "num_tokens": 916717.0,
      "reward": -0.4671813119202852,
      "reward_std": 0.018046872515697032,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.07638889132067561,
      "rewards/repetition_penalty_reward": -0.04820289777126163,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 24
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 328.51042556762695,
      "epoch": 0.0104,
      "grad_norm": 4.017609596252441,
      "kl": 0.0010578632354736328,
      "learning_rate": 4.990660391382923e-07,
      "loss": -0.0418,
      "num_tokens": 997406.0,
      "reward": -0.46294988691806793,
      "reward_std": 0.03339050061185844,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.1041666695382446,
      "rewards/repetition_penalty_reward": -0.06491638510487974,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 26
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 266.8333377838135,
      "epoch": 0.0112,
      "grad_norm": 5.310181617736816,
      "kl": 0.0015530586242675781,
      "learning_rate": 4.987290615070384e-07,
      "loss": 0.2128,
      "num_tokens": 1072174.0,
      "reward": -0.4547987822443247,
      "reward_std": 0.04966713820613222,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.0659722238779068,
      "rewards/repetition_penalty_reward": -0.056669211451662704,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 28
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 268.0312581062317,
      "epoch": 0.012,
      "grad_norm": 5.723649978637695,
      "kl": 0.0017752647399902344,
      "learning_rate": 4.983404296598978e-07,
      "loss": -0.0013,
      "num_tokens": 1147057.0,
      "reward": -0.45876461267471313,
      "reward_std": 0.039007039871648885,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.1284722255077213,
      "rewards/repetition_penalty_reward": -0.055181623916723765,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.013020833721384406,
      "step": 30
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 208.0520886182785,
      "epoch": 0.0128,
      "grad_norm": 4.876113414764404,
      "kl": 0.0021886825561523438,
      "learning_rate": 4.979002243050646e-07,
      "loss": 0.0591,
      "num_tokens": 1216182.0,
      "reward": -0.4663495346903801,
      "reward_std": 0.01864687688066624,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.0833333358168602,
      "rewards/repetition_penalty_reward": -0.04422530399460811,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 32
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 332.57292652130127,
      "epoch": 0.0136,
      "grad_norm": 4.4281511306762695,
      "kl": 0.0018911361694335938,
      "learning_rate": 4.974085368611381e-07,
      "loss": 0.0615,
      "num_tokens": 1297261.0,
      "reward": -0.45041274279356003,
      "reward_std": 0.05302129179472104,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.07291666767559946,
      "rewards/repetition_penalty_reward": -0.04839907615678385,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.007812500232830644,
      "step": 34
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 227.93750929832458,
      "epoch": 0.0144,
      "grad_norm": 6.5882248878479,
      "kl": 0.0024008750915527344,
      "learning_rate": 4.968654694381379e-07,
      "loss": 0.0836,
      "num_tokens": 1368295.0,
      "reward": -0.4686909671872854,
      "reward_std": 0.013386835009441711,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.03819444542750716,
      "rewards/repetition_penalty_reward": -0.03552157600643113,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.015625000465661287,
      "step": 36
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 233.9479250907898,
      "epoch": 0.0152,
      "grad_norm": 5.5114569664001465,
      "kl": 0.0028159618377685547,
      "learning_rate": 4.962711348162987e-07,
      "loss": -0.024,
      "num_tokens": 1439906.0,
      "reward": -0.4651654362678528,
      "reward_std": 0.025145536681520753,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.06597222411073744,
      "rewards/repetition_penalty_reward": -0.048877436958719045,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.010416666744276881,
      "step": 38
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 246.72917079925537,
      "epoch": 0.016,
      "grad_norm": 4.999541282653809,
      "kl": 0.0029511451721191406,
      "learning_rate": 4.956256564226487e-07,
      "loss": -0.0201,
      "num_tokens": 1512744.0,
      "reward": -0.4640405811369419,
      "reward_std": 0.024847639388099196,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.07291666837409139,
      "rewards/repetition_penalty_reward": -0.04978172353003174,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.015625000465661287,
      "step": 40
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 230.2604217529297,
      "epoch": 0.0168,
      "grad_norm": 4.764357089996338,
      "kl": 0.0031957626342773438,
      "learning_rate": 4.949291683053768e-07,
      "loss": -0.1051,
      "num_tokens": 1584001.0,
      "reward": -0.4276710934937,
      "reward_std": 0.10604065289953724,
      "rewards/SMILES_validity_reward": -0.5500000212341547,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.12500000349245965,
      "rewards/repetition_penalty_reward": -0.038690910892910324,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 42
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 162.5000033378601,
      "epoch": 0.0176,
      "grad_norm": 5.813767433166504,
      "kl": 0.0059719085693359375,
      "learning_rate": 4.941818151059955e-07,
      "loss": 0.0334,
      "num_tokens": 1648753.0,
      "reward": -0.4659581948071718,
      "reward_std": 0.017741082614520565,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.07638889085501432,
      "rewards/repetition_penalty_reward": -0.033367487922078,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 44
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 244.57292413711548,
      "epoch": 0.0184,
      "grad_norm": 3.382213830947876,
      "kl": 0.005543231964111328,
      "learning_rate": 4.933837520293017e-07,
      "loss": 0.1422,
      "num_tokens": 1721384.0,
      "reward": -0.4495168598368764,
      "reward_std": 0.0532101350290759,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.07638889062218368,
      "rewards/repetition_penalty_reward": -0.04291246473439969,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 46
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 250.81250488758087,
      "epoch": 0.0192,
      "grad_norm": 3.784503936767578,
      "kl": 0.005627632141113281,
      "learning_rate": 4.925351448111454e-07,
      "loss": 0.0332,
      "num_tokens": 1794614.0,
      "reward": -0.4602485578507185,
      "reward_std": 0.03407564776716754,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.09722222457639873,
      "rewards/repetition_penalty_reward": -0.04918775928672403,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.007812500232830644,
      "step": 48
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 210.15625739097595,
      "epoch": 0.02,
      "grad_norm": 4.50042200088501,
      "kl": 0.009564399719238281,
      "learning_rate": 4.91636169684011e-07,
      "loss": 0.0973,
      "num_tokens": 1863941.0,
      "reward": -0.4667747635394335,
      "reward_std": 0.015594157837767852,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.07638889038935304,
      "rewards/repetition_penalty_reward": -0.04413735281559639,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 50
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 209.0416705608368,
      "epoch": 0.0208,
      "grad_norm": 3.6077358722686768,
      "kl": 0.010011672973632812,
      "learning_rate": 4.906870133404186e-07,
      "loss": 0.1673,
      "num_tokens": 1933161.0,
      "reward": -0.4403855809941888,
      "reward_std": 0.07992339776683366,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.04513888992369175,
      "rewards/repetition_penalty_reward": -0.03441216413921211,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0,
      "step": 52
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 207.75000667572021,
      "epoch": 0.0216,
      "grad_norm": 4.577792644500732,
      "kl": 0.011927604675292969,
      "learning_rate": 4.896878728941531e-07,
      "loss": 0.0505,
      "num_tokens": 2002257.0,
      "reward": -0.4528631530702114,
      "reward_std": 0.042765807069372386,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.08680555806495249,
      "rewards/repetition_penalty_reward": -0.029500389107852243,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 54
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 180.7500035762787,
      "epoch": 0.0224,
      "grad_norm": 6.483789443969727,
      "kl": 0.013889312744140625,
      "learning_rate": 4.886389558393284e-07,
      "loss": 0.1321,
      "num_tokens": 2068761.0,
      "reward": -0.4464299641549587,
      "reward_std": 0.07374286159756593,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.02083333395421505,
      "rewards/repetition_penalty_reward": -0.03148799104383215,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.007812500232830644,
      "step": 56
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 229.35417246818542,
      "epoch": 0.0232,
      "grad_norm": 3.902653932571411,
      "kl": 0.013338088989257812,
      "learning_rate": 4.875404800072976e-07,
      "loss": -0.0142,
      "num_tokens": 2139931.0,
      "reward": -0.45120589807629585,
      "reward_std": 0.04398537143060821,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.10763889225199819,
      "rewards/repetition_penalty_reward": -0.031157015706412494,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0,
      "step": 58
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 197.83333957195282,
      "epoch": 0.024,
      "grad_norm": 4.851426601409912,
      "kl": 0.02254486083984375,
      "learning_rate": 4.86392673521415e-07,
      "loss": 0.3276,
      "num_tokens": 2208075.0,
      "reward": -0.469699963927269,
      "reward_std": 0.015661527457268676,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.017361111473292112,
      "rewards/repetition_penalty_reward": -0.04040325006644707,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0,
      "step": 60
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 174.0625078678131,
      "epoch": 0.0248,
      "grad_norm": 4.745796203613281,
      "kl": 0.026813507080078125,
      "learning_rate": 4.851957747496606e-07,
      "loss": 0.0778,
      "num_tokens": 2273937.0,
      "reward": -0.46857298351824284,
      "reward_std": 0.016187770637770882,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.05208333418704569,
      "rewards/repetition_penalty_reward": -0.030001492603332736,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 62
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 135.04167068004608,
      "epoch": 0.0256,
      "grad_norm": 6.314318656921387,
      "kl": 0.023064613342285156,
      "learning_rate": 4.839500322551386e-07,
      "loss": 0.0664,
      "num_tokens": 2336053.0,
      "reward": -0.4527385290712118,
      "reward_std": 0.0479089121290599,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.05208333441987634,
      "rewards/repetition_penalty_reward": -0.022177818129421212,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0,
      "step": 64
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 172.0937558412552,
      "epoch": 0.0264,
      "grad_norm": 6.215081214904785,
      "kl": 0.0323638916015625,
      "learning_rate": 4.826557047444563e-07,
      "loss": 0.1401,
      "num_tokens": 2401726.0,
      "reward": -0.46973296627402306,
      "reward_std": 0.012096146092517301,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.04166666814126074,
      "rewards/repetition_penalty_reward": -0.03378878638613969,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0,
      "step": 66
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 115.77083659172058,
      "epoch": 0.0272,
      "grad_norm": 5.195652961730957,
      "kl": 0.04044342041015625,
      "learning_rate": 4.813130610139993e-07,
      "loss": 0.1761,
      "num_tokens": 2461992.0,
      "reward": -0.4498000517487526,
      "reward_std": 0.06054470940580359,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.0034722222480922937,
      "rewards/repetition_penalty_reward": -0.016577735252212733,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 68
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 214.18750643730164,
      "epoch": 0.028,
      "grad_norm": 5.4179182052612305,
      "kl": 0.028334617614746094,
      "learning_rate": 4.799223798941089e-07,
      "loss": 0.0926,
      "num_tokens": 2531706.0,
      "reward": -0.456182649359107,
      "reward_std": 0.04664806976506952,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.08333333558402956,
      "rewards/repetition_penalty_reward": -0.046202394034480676,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0,
      "step": 70
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 137.51041901111603,
      "epoch": 0.0288,
      "grad_norm": 4.121431827545166,
      "kl": 0.051410675048828125,
      "learning_rate": 4.78483950191177e-07,
      "loss": -0.0549,
      "num_tokens": 2594059.0,
      "reward": -0.4664184954017401,
      "reward_std": 0.013460547677823342,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.0625000016298145,
      "rewards/repetition_penalty_reward": -0.026685798191465437,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 72
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 184.61458683013916,
      "epoch": 0.0296,
      "grad_norm": 5.257875442504883,
      "kl": 0.050022125244140625,
      "learning_rate": 4.769980706276687e-07,
      "loss": 0.0014,
      "num_tokens": 2660934.0,
      "reward": -0.4514905624091625,
      "reward_std": 0.053132650078623556,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.07986111333593726,
      "rewards/repetition_penalty_reward": -0.04268423894245643,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.010416666977107525,
      "step": 74
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 112.86458694934845,
      "epoch": 0.0304,
      "grad_norm": 4.131363868713379,
      "kl": 0.058177947998046875,
      "learning_rate": 4.7546504978008595e-07,
      "loss": 0.1018,
      "num_tokens": 2720921.0,
      "reward": -0.46908595971763134,
      "reward_std": 0.0056425063230562955,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.020833333721384406,
      "rewards/repetition_penalty_reward": -0.011693818843923509,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 76
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 119.19792026281357,
      "epoch": 0.0312,
      "grad_norm": 5.673096656799316,
      "kl": 0.06604766845703125,
      "learning_rate": 4.738852060148848e-07,
      "loss": -0.087,
      "num_tokens": 2781516.0,
      "reward": -0.4646188300102949,
      "reward_std": 0.014277806236350443,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.06250000139698386,
      "rewards/repetition_penalty_reward": -0.013897485987399705,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 78
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 106.760418176651,
      "epoch": 0.032,
      "grad_norm": 5.955003261566162,
      "kl": 0.0661163330078125,
      "learning_rate": 4.722588674223593e-07,
      "loss": 0.0259,
      "num_tokens": 2840917.0,
      "reward": -0.46621129661798477,
      "reward_std": 0.01656446641209186,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.03819444519467652,
      "rewards/repetition_penalty_reward": -0.02374570633401163,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 80
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 143.8645884990692,
      "epoch": 0.0328,
      "grad_norm": 5.669166564941406,
      "kl": 0.05596160888671875,
      "learning_rate": 4.70586371748506e-07,
      "loss": 0.0061,
      "num_tokens": 2903880.0,
      "reward": -0.4342301404103637,
      "reward_std": 0.08722544111515163,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0416666679084301,
      "rewards/reasoning_steps_reward": 0.041666667675599456,
      "rewards/repetition_penalty_reward": -0.02146888236165978,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.0,
      "step": 82
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 152.95833730697632,
      "epoch": 0.0336,
      "grad_norm": 5.404734134674072,
      "kl": 0.07171249389648438,
      "learning_rate": 4.6886806632488363e-07,
      "loss": 0.1132,
      "num_tokens": 2967716.0,
      "reward": -0.47036008536815643,
      "reward_std": 0.013487992946465965,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.027777778450399637,
      "rewards/repetition_penalty_reward": -0.02617112870211713,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 84
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 132.33333587646484,
      "epoch": 0.0344,
      "grad_norm": 5.923622131347656,
      "kl": 0.0746612548828125,
      "learning_rate": 4.6710430799648143e-07,
      "loss": 0.1332,
      "num_tokens": 3029572.0,
      "reward": -0.4556627534329891,
      "reward_std": 0.04119024146348238,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.027777778450399637,
      "rewards/repetition_penalty_reward": -0.021906152804149315,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0,
      "step": 86
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 124.13541960716248,
      "epoch": 0.0352,
      "grad_norm": 5.512033939361572,
      "kl": 0.091949462890625,
      "learning_rate": 4.652954630476127e-07,
      "loss": 0.1804,
      "num_tokens": 3090641.0,
      "reward": -0.47078334353864193,
      "reward_std": 0.007824531650840072,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.017361111473292112,
      "rewards/repetition_penalty_reward": -0.025195291149429977,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 88
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 84.52083671092987,
      "epoch": 0.036,
      "grad_norm": 3.839015007019043,
      "kl": 0.09572601318359375,
      "learning_rate": 4.6344190712584713e-07,
      "loss": 0.0499,
      "num_tokens": 3147907.0,
      "reward": -0.4675292409956455,
      "reward_std": 0.008711494119779672,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.03472222317941487,
      "rewards/repetition_penalty_reward": -0.010015465391916223,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 90
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 93.9270852804184,
      "epoch": 0.0368,
      "grad_norm": 3.3888838291168213,
      "kl": 0.116241455078125,
      "learning_rate": 4.615440251639995e-07,
      "loss": -0.0578,
      "num_tokens": 3206076.0,
      "reward": -0.4662612807005644,
      "reward_std": 0.01008661445666803,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.05208333465270698,
      "rewards/repetition_penalty_reward": -0.00948869220155757,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0,
      "step": 92
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 106.22916853427887,
      "epoch": 0.0376,
      "grad_norm": 3.9039151668548584,
      "kl": 0.122039794921875,
      "learning_rate": 4.596022113001894e-07,
      "loss": -0.0583,
      "num_tokens": 3265426.0,
      "reward": -0.4652919042855501,
      "reward_std": 0.013997638350701891,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.05902777914889157,
      "rewards/repetition_penalty_reward": -0.019760221184697002,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0078125,
      "step": 94
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 91.39583814144135,
      "epoch": 0.0384,
      "grad_norm": 6.249267101287842,
      "kl": 0.128753662109375,
      "learning_rate": 4.576168687959895e-07,
      "loss": 0.0203,
      "num_tokens": 3323352.0,
      "reward": -0.4529070910066366,
      "reward_std": 0.0469353504377068,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.027777778217568994,
      "rewards/repetition_penalty_reward": -0.01257872223504819,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 96
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 110.9479192495346,
      "epoch": 0.0392,
      "grad_norm": 6.070929050445557,
      "kl": 0.122650146484375,
      "learning_rate": 4.555884099526793e-07,
      "loss": 0.0258,
      "num_tokens": 3383155.0,
      "reward": -0.439573897048831,
      "reward_std": 0.07561502030876,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.03819444542750716,
      "rewards/repetition_penalty_reward": -0.019350912392837927,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 98
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 107.42708563804626,
      "epoch": 0.04,
      "grad_norm": 6.182693004608154,
      "kl": 0.1309814453125,
      "learning_rate": 4.5351725602562174e-07,
      "loss": 0.0129,
      "num_tokens": 3442620.0,
      "reward": -0.465021351352334,
      "reward_std": 0.018675321242881182,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.0416666679084301,
      "rewards/repetition_penalty_reward": -0.020526854172203457,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 100
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 106.75000429153442,
      "epoch": 0.0408,
      "grad_norm": 4.344884395599365,
      "kl": 0.119537353515625,
      "learning_rate": 4.514038371367791e-07,
      "loss": 0.1713,
      "num_tokens": 3502020.0,
      "reward": -0.4704951122403145,
      "reward_std": 0.0033068859920604154,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.0069444444961845875,
      "rewards/repetition_penalty_reward": -0.011896428157342598,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 102
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 65.79166901111603,
      "epoch": 0.0416,
      "grad_norm": 4.156596660614014,
      "kl": 0.143524169921875,
      "learning_rate": 4.4924859218538936e-07,
      "loss": -0.0151,
      "num_tokens": 3557488.0,
      "reward": -0.4587331060320139,
      "reward_std": 0.02372456392913591,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.031250000931322575,
      "rewards/reasoning_steps_reward": 0.031250000931322575,
      "rewards/repetition_penalty_reward": -0.007123569957911968,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 104
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 113.0312557220459,
      "epoch": 0.0424,
      "grad_norm": 5.499018669128418,
      "kl": 0.11688232421875,
      "learning_rate": 4.470519687568185e-07,
      "loss": 0.0769,
      "num_tokens": 3617491.0,
      "reward": -0.46708683855831623,
      "reward_std": 0.011621264442510437,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.04513889015652239,
      "rewards/repetition_penalty_reward": -0.0160081133217318,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.0,
      "step": 106
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 103.69792056083679,
      "epoch": 0.0432,
      "grad_norm": 2.9395508766174316,
      "kl": 0.1288299560546875,
      "learning_rate": 4.4481442302960923e-07,
      "loss": -0.0733,
      "num_tokens": 3676598.0,
      "reward": -0.4545927792787552,
      "reward_std": 0.03454983566189185,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.045138889690861106,
      "rewards/repetition_penalty_reward": -0.01033839286537841,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0078125,
      "step": 108
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 90.92708659172058,
      "epoch": 0.044,
      "grad_norm": 4.17449426651001,
      "kl": 0.1243133544921875,
      "learning_rate": 4.4253641968074505e-07,
      "loss": 0.0349,
      "num_tokens": 3734479.0,
      "reward": -0.4687845651060343,
      "reward_std": 0.0076231868952163495,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.03125000069849193,
      "rewards/repetition_penalty_reward": -0.011283979780273512,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 110
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 129.86458790302277,
      "epoch": 0.0448,
      "grad_norm": 4.147829055786133,
      "kl": 0.10516357421875,
      "learning_rate": 4.402184317891501e-07,
      "loss": -0.0648,
      "num_tokens": 3796098.0,
      "reward": -0.45566146075725555,
      "reward_std": 0.04058586481369275,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.04861111217178404,
      "rewards/repetition_penalty_reward": -0.014080686727538705,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 112
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 137.92708706855774,
      "epoch": 0.0456,
      "grad_norm": 8.529430389404297,
      "kl": 0.115264892578125,
      "learning_rate": 4.37860940737443e-07,
      "loss": -0.1091,
      "num_tokens": 3858491.0,
      "reward": -0.4289623722434044,
      "reward_std": 0.09471672396102804,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.052083334885537624,
      "rewards/reasoning_steps_reward": 0.04861111263744533,
      "rewards/repetition_penalty_reward": -0.01740230650466401,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 114
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 63.87500262260437,
      "epoch": 0.0464,
      "grad_norm": 9.70301342010498,
      "kl": 0.1183013916015625,
      "learning_rate": 4.354644361119671e-07,
      "loss": 0.0586,
      "num_tokens": 3913775.0,
      "reward": -0.4454812277108431,
      "reward_std": 0.0836324201318348,
      "rewards/SMILES_validity_reward": -0.571875024586916,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.003771467447222676,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.0,
      "step": 116
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 90.41666865348816,
      "epoch": 0.0472,
      "grad_norm": 7.300216197967529,
      "kl": 0.114227294921875,
      "learning_rate": 4.3302941560111716e-07,
      "loss": -0.0711,
      "num_tokens": 3971607.0,
      "reward": -0.45174217596650124,
      "reward_std": 0.04705615748389391,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.05208333441987634,
      "rewards/repetition_penalty_reward": -0.017422605837055016,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.010416666977107525,
      "step": 118
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 120.2187534570694,
      "epoch": 0.048,
      "grad_norm": 7.793805122375488,
      "kl": 0.114410400390625,
      "learning_rate": 4.3055638489198236e-07,
      "loss": -0.0088,
      "num_tokens": 4032300.0,
      "reward": -0.46079872362315655,
      "reward_std": 0.022261769554461353,
      "rewards/SMILES_validity_reward": -0.6000000238418579,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.04861111240461469,
      "rewards/repetition_penalty_reward": -0.01649503846419975,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.007812500232830644,
      "step": 120
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 112.37500369548798,
      "epoch": 0.0488,
      "grad_norm": 5.610187530517578,
      "kl": 0.097442626953125,
      "learning_rate": 4.280458575653296e-07,
      "loss": 0.0941,
      "num_tokens": 4092240.0,
      "reward": -0.45120035484433174,
      "reward_std": 0.04598578553486732,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.03472222317941487,
      "rewards/repetition_penalty_reward": -0.015476670654607005,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.015624999767169356,
      "step": 122
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 92.35416865348816,
      "epoch": 0.0496,
      "grad_norm": 5.858239650726318,
      "kl": 0.114837646484375,
      "learning_rate": 4.2549835498894665e-07,
      "loss": -0.0426,
      "num_tokens": 4150258.0,
      "reward": -0.44725739397108555,
      "reward_std": 0.054796224998426624,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.031250000931322575,
      "rewards/reasoning_steps_reward": 0.041666667675599456,
      "rewards/repetition_penalty_reward": -0.011637324947514571,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.007812500232830644,
      "step": 124
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 145.7083374261856,
      "epoch": 0.0504,
      "grad_norm": 4.300314426422119,
      "kl": 0.092559814453125,
      "learning_rate": 4.229144062093679e-07,
      "loss": -0.067,
      "num_tokens": 4213398.0,
      "reward": -0.4413000885397196,
      "reward_std": 0.07192132115596905,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.04861111217178404,
      "rewards/repetition_penalty_reward": -0.018383800808805972,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0078125,
      "step": 126
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 124.31250333786011,
      "epoch": 0.0512,
      "grad_norm": 9.1367769241333,
      "kl": 0.102203369140625,
      "learning_rate": 4.2029454784200675e-07,
      "loss": -0.0685,
      "num_tokens": 4274484.0,
      "reward": -0.45298329181969166,
      "reward_std": 0.05172415471315617,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.052083334885537624,
      "rewards/repetition_penalty_reward": -0.016812930087326095,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.007812500232830644,
      "step": 128
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 116.47917032241821,
      "epoch": 0.052,
      "grad_norm": 6.13424825668335,
      "kl": 0.106658935546875,
      "learning_rate": 4.1763932395971433e-07,
      "loss": -0.0741,
      "num_tokens": 4334818.0,
      "reward": -0.43802906200289726,
      "reward_std": 0.08000406731298426,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.07986111333593726,
      "rewards/repetition_penalty_reward": -0.016923470440815436,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.0078125,
      "step": 130
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 150.53125488758087,
      "epoch": 0.0528,
      "grad_norm": 8.036478996276855,
      "kl": 0.09299468994140625,
      "learning_rate": 4.1494928597979117e-07,
      "loss": -0.1005,
      "num_tokens": 4398421.0,
      "reward": -0.43908108957111835,
      "reward_std": 0.0789313922796282,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.05902777914889157,
      "rewards/repetition_penalty_reward": -0.022235317796003073,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.0026041667442768812,
      "step": 132
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 150.88542008399963,
      "epoch": 0.0536,
      "grad_norm": 6.453376293182373,
      "kl": 0.0865478515625,
      "learning_rate": 4.122249925494726e-07,
      "loss": 0.0604,
      "num_tokens": 4462058.0,
      "reward": -0.4095195522531867,
      "reward_std": 0.10667452452980797,
      "rewards/SMILES_validity_reward": -0.5166666880249977,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.041666667675599456,
      "rewards/repetition_penalty_reward": -0.025404679967323318,
      "rewards/smiles_len_reward": -0.03645833395421505,
      "rewards/tag_count_reward": 0.010416666977107525,
      "step": 134
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 96.39583551883698,
      "epoch": 0.0544,
      "grad_norm": 8.152026176452637,
      "kl": 0.1044464111328125,
      "learning_rate": 4.094670094299131e-07,
      "loss": -0.0086,
      "num_tokens": 4520464.0,
      "reward": -0.45102883502840996,
      "reward_std": 0.0681979734054039,
      "rewards/SMILES_validity_reward": -0.588541692122817,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0416666679084301,
      "rewards/reasoning_steps_reward": 0.04513889015652239,
      "rewards/repetition_penalty_reward": -0.024178073908842634,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.0,
      "step": 136
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 112.1979204416275,
      "epoch": 0.0552,
      "grad_norm": 4.7342000007629395,
      "kl": 0.08402252197265625,
      "learning_rate": 4.066759093786931e-07,
      "loss": 0.096,
      "num_tokens": 4580387.0,
      "reward": -0.4373946124687791,
      "reward_std": 0.07959776032657828,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.031250000931322575,
      "rewards/reasoning_steps_reward": 0.03819444542750716,
      "rewards/repetition_penalty_reward": -0.0209956017206423,
      "rewards/smiles_len_reward": -0.03125000046566129,
      "rewards/tag_count_reward": 0.013020833721384406,
      "step": 138
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 115.69791889190674,
      "epoch": 0.056,
      "grad_norm": 6.3178863525390625,
      "kl": 0.0914154052734375,
      "learning_rate": 4.038522720308732e-07,
      "loss": -0.0181,
      "num_tokens": 4640646.0,
      "reward": -0.39627854742866475,
      "reward_std": 0.10544795080204494,
      "rewards/SMILES_validity_reward": -0.5166666894219816,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0416666679084301,
      "rewards/reasoning_steps_reward": 0.05555555666796863,
      "rewards/repetition_penalty_reward": -0.018862670767703094,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.013020833721384406,
      "step": 140
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 125.94791948795319,
      "epoch": 0.0568,
      "grad_norm": 7.152596950531006,
      "kl": 0.10680389404296875,
      "learning_rate": 4.009966837786194e-07,
      "loss": 0.0079,
      "num_tokens": 4701889.0,
      "reward": -0.45083487406373024,
      "reward_std": 0.05493604081857484,
      "rewards/SMILES_validity_reward": -0.5833333563059568,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.02083333395421505,
      "rewards/reasoning_steps_reward": 0.055555556900799274,
      "rewards/repetition_penalty_reward": -0.0222385291417595,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 142
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 170.20833539962769,
      "epoch": 0.0576,
      "grad_norm": 5.538232803344727,
      "kl": 0.09381866455078125,
      "learning_rate": 3.981097376494259e-07,
      "loss": -0.1594,
      "num_tokens": 4767381.0,
      "reward": -0.4077454451471567,
      "reward_std": 0.15711192146409303,
      "rewards/SMILES_validity_reward": -0.5333333536982536,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0416666679084301,
      "rewards/reasoning_steps_reward": 0.06944444682449102,
      "rewards/repetition_penalty_reward": -0.025545662763761356,
      "rewards/smiles_len_reward": -0.0416666679084301,
      "rewards/tag_count_reward": 0.02864583395421505,
      "step": 144
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 98.79166948795319,
      "epoch": 0.0584,
      "grad_norm": 5.379575729370117,
      "kl": 0.1139678955078125,
      "learning_rate": 3.951920331829592e-07,
      "loss": 0.0907,
      "num_tokens": 4826017.0,
      "reward": -0.44052676670253277,
      "reward_std": 0.07575159638145124,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.031250000931322575,
      "rewards/reasoning_steps_reward": 0.0069444444961845875,
      "rewards/repetition_penalty_reward": -0.013254568788397592,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.0052083334885537624,
      "step": 146
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 115.91666913032532,
      "epoch": 0.0592,
      "grad_norm": 5.238061904907227,
      "kl": 0.0884857177734375,
      "learning_rate": 3.922441763065506e-07,
      "loss": -0.0319,
      "num_tokens": 4886297.0,
      "reward": -0.4026124000083655,
      "reward_std": 0.12813380375882844,
      "rewards/SMILES_validity_reward": -0.5333333550952375,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.07291666883975267,
      "rewards/reasoning_steps_reward": 0.024305556202307343,
      "rewards/repetition_penalty_reward": -0.0176178389810957,
      "rewards/smiles_len_reward": -0.0416666679084301,
      "rewards/tag_count_reward": 0.02343750069849193,
      "step": 148
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 81.14583551883698,
      "epoch": 0.06,
      "grad_norm": 9.144115447998047,
      "kl": 0.10345458984375,
      "learning_rate": 3.8926677920936093e-07,
      "loss": 0.0174,
      "num_tokens": 4943239.0,
      "reward": -0.4181996285915375,
      "reward_std": 0.12576395435098675,
      "rewards/SMILES_validity_reward": -0.5500000212341547,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.06250000186264515,
      "rewards/reasoning_steps_reward": 0.02083333395421505,
      "rewards/repetition_penalty_reward": -0.014288760961790103,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.010416666977107525,
      "step": 150
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 126.05208683013916,
      "epoch": 0.0608,
      "grad_norm": 6.60600471496582,
      "kl": 0.105255126953125,
      "learning_rate": 3.862604602152464e-07,
      "loss": -0.1057,
      "num_tokens": 5004492.0,
      "reward": -0.4281045887619257,
      "reward_std": 0.09912917913607089,
      "rewards/SMILES_validity_reward": -0.5666666887700558,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0416666679084301,
      "rewards/reasoning_steps_reward": 0.10416667023673654,
      "rewards/repetition_penalty_reward": -0.02531764625746291,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.007812500232830644,
      "step": 152
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 138.1041705608368,
      "epoch": 0.0616,
      "grad_norm": 8.543445587158203,
      "kl": 0.08979034423828125,
      "learning_rate": 3.8322584365434934e-07,
      "loss": -0.0847,
      "num_tokens": 5066902.0,
      "reward": -0.4112464152276516,
      "reward_std": 0.1456986589182634,
      "rewards/SMILES_validity_reward": -0.5333333536982536,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.052083334885537624,
      "rewards/reasoning_steps_reward": 0.03819444542750716,
      "rewards/repetition_penalty_reward": -0.02149280160665512,
      "rewards/smiles_len_reward": -0.06250000186264515,
      "rewards/tag_count_reward": 0.010416666744276881,
      "step": 154
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 179.20833945274353,
      "epoch": 0.0624,
      "grad_norm": 7.902252197265625,
      "kl": 0.11370849609375,
      "learning_rate": 3.8016355973344173e-07,
      "loss": -0.1289,
      "num_tokens": 5133258.0,
      "reward": -0.40283326152712107,
      "reward_std": 0.13701566337113036,
      "rewards/SMILES_validity_reward": -0.5333333550952375,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.06250000186264515,
      "rewards/reasoning_steps_reward": 0.0902777798473835,
      "rewards/repetition_penalty_reward": -0.025902962646796368,
      "rewards/smiles_len_reward": -0.06250000186264515,
      "rewards/tag_count_reward": 0.015625000465661287,
      "step": 156
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 183.17708587646484,
      "epoch": 0.0632,
      "grad_norm": 5.164620399475098,
      "kl": 0.089508056640625,
      "learning_rate": 3.7707424440504863e-07,
      "loss": 0.0345,
      "num_tokens": 5199995.0,
      "reward": -0.36004658944148105,
      "reward_std": 0.23757518313504988,
      "rewards/SMILES_validity_reward": -0.4833333524875343,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.09375000279396772,
      "rewards/reasoning_steps_reward": 0.08333333511836827,
      "rewards/repetition_penalty_reward": -0.03744592401199043,
      "rewards/smiles_len_reward": -0.06250000139698386,
      "rewards/tag_count_reward": 0.018229166977107525,
      "step": 158
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 107.97916984558105,
      "epoch": 0.064,
      "grad_norm": 5.103909015655518,
      "kl": 0.102081298828125,
      "learning_rate": 3.739585392353787e-07,
      "loss": -0.1095,
      "num_tokens": 5259513.0,
      "reward": -0.37965655652806163,
      "reward_std": 0.18476388306226,
      "rewards/SMILES_validity_reward": -0.5000000200234354,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.0729166679084301,
      "rewards/reasoning_steps_reward": 0.03819444542750716,
      "rewards/repetition_penalty_reward": -0.011844190346891992,
      "rewards/smiles_len_reward": -0.05208333441987634,
      "rewards/tag_count_reward": 0.010416666977107525,
      "step": 160
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 133.76042115688324,
      "epoch": 0.0648,
      "grad_norm": 10.518568992614746,
      "kl": 0.1388397216796875,
      "learning_rate": 3.7081709127108767e-07,
      "loss": -0.0174,
      "num_tokens": 5321506.0,
      "reward": -0.3726199440425262,
      "reward_std": 0.1925114190817112,
      "rewards/SMILES_validity_reward": -0.5166666875593364,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.13541666977107525,
      "rewards/reasoning_steps_reward": 0.08333333465270698,
      "rewards/repetition_penalty_reward": -0.03140865158638917,
      "rewards/smiles_len_reward": -0.08854166837409139,
      "rewards/tag_count_reward": 0.020833333488553762,
      "step": 162
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 121.31250333786011,
      "epoch": 0.0656,
      "grad_norm": 13.173834800720215,
      "kl": 0.153289794921875,
      "learning_rate": 3.6765055290490513e-07,
      "loss": 0.0415,
      "num_tokens": 5382304.0,
      "reward": -0.3137630212586373,
      "reward_std": 0.2859425003753131,
      "rewards/SMILES_validity_reward": -0.4333333526737988,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.14583333674818277,
      "rewards/reasoning_steps_reward": 0.05208333465270698,
      "rewards/repetition_penalty_reward": -0.023568583332234994,
      "rewards/smiles_len_reward": -0.10416666930541396,
      "rewards/tag_count_reward": 0.03385416744276881,
      "step": 164
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 97.02083551883698,
      "epoch": 0.0664,
      "grad_norm": 8.68799877166748,
      "kl": 0.1730804443359375,
      "learning_rate": 3.644595817401501e-07,
      "loss": -0.0409,
      "num_tokens": 5440770.0,
      "reward": -0.27725822292268276,
      "reward_std": 0.28745744668412954,
      "rewards/SMILES_validity_reward": -0.38333335146307945,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.1458333358168602,
      "rewards/reasoning_steps_reward": 0.03472222341224551,
      "rewards/repetition_penalty_reward": -0.009388637714437209,
      "rewards/smiles_len_reward": -0.06770833535119891,
      "rewards/tag_count_reward": 0.015625000232830644,
      "step": 166
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 121.968754529953,
      "epoch": 0.0672,
      "grad_norm": 11.079466819763184,
      "kl": 0.1530609130859375,
      "learning_rate": 3.6124484045416483e-07,
      "loss": 0.0816,
      "num_tokens": 5501631.0,
      "reward": -0.25446695898426697,
      "reward_std": 0.32407376743503846,
      "rewards/SMILES_validity_reward": -0.383333352394402,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.21875000465661287,
      "rewards/reasoning_steps_reward": 0.03472222317941487,
      "rewards/repetition_penalty_reward": -0.013246808550320566,
      "rewards/smiles_len_reward": -0.07812500232830644,
      "rewards/tag_count_reward": 0.039062499767169356,
      "step": 168
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 153.052086353302,
      "epoch": 0.068,
      "grad_norm": 8.2036714553833,
      "kl": 0.171844482421875,
      "learning_rate": 3.580069966606949e-07,
      "loss": -0.1069,
      "num_tokens": 5565476.0,
      "reward": -0.17629894718993455,
      "reward_std": 0.47014068998396397,
      "rewards/SMILES_validity_reward": -0.2833333467133343,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.26041666977107525,
      "rewards/reasoning_steps_reward": 0.052083334885537624,
      "rewards/repetition_penalty_reward": -0.03746950729691889,
      "rewards/smiles_len_reward": -0.13541666883975267,
      "rewards/tag_count_reward": 0.05989583395421505,
      "step": 170
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 130.13542079925537,
      "epoch": 0.0688,
      "grad_norm": 12.656828880310059,
      "kl": 0.278564453125,
      "learning_rate": 3.547467227712444e-07,
      "loss": -0.0281,
      "num_tokens": 5627121.0,
      "reward": -0.14627268894400913,
      "reward_std": 0.501163106149761,
      "rewards/SMILES_validity_reward": -0.2333333482965827,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.23958333674818277,
      "rewards/reasoning_steps_reward": 0.031250000931322575,
      "rewards/repetition_penalty_reward": -0.022102794086094946,
      "rewards/smiles_len_reward": -0.09375000232830644,
      "rewards/tag_count_reward": 0.03645833348855376,
      "step": 172
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 136.18750500679016,
      "epoch": 0.0696,
      "grad_norm": 8.628015518188477,
      "kl": 0.2374114990234375,
      "learning_rate": 3.5146469585543386e-07,
      "loss": 0.0807,
      "num_tokens": 5689347.0,
      "reward": -0.11965863051591441,
      "reward_std": 0.4950985286559444,
      "rewards/SMILES_validity_reward": -0.21666668402031064,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.3020833367481828,
      "rewards/reasoning_steps_reward": 0.031250000931322575,
      "rewards/repetition_penalty_reward": -0.015858068407396786,
      "rewards/smiles_len_reward": -0.11979166977107525,
      "rewards/tag_count_reward": 0.01822916720993817,
      "step": 174
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 109.19791996479034,
      "epoch": 0.0704,
      "grad_norm": 9.866938591003418,
      "kl": 0.29345703125,
      "learning_rate": 3.481615975003922e-07,
      "loss": -0.0523,
      "num_tokens": 5748982.0,
      "reward": -0.16320179387548706,
      "reward_std": 0.46159220617846586,
      "rewards/SMILES_validity_reward": -0.25520835211500525,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.2500000074505806,
      "rewards/reasoning_steps_reward": 0.0243055559694767,
      "rewards/repetition_penalty_reward": -0.01569942681817338,
      "rewards/smiles_len_reward": -0.140625003259629,
      "rewards/tag_count_reward": 0.03645833395421505,
      "step": 176
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 141.4479193687439,
      "epoch": 0.0712,
      "grad_norm": 7.250740051269531,
      "kl": 0.243499755859375,
      "learning_rate": 3.448381136692089e-07,
      "loss": 0.0395,
      "num_tokens": 5811713.0,
      "reward": -0.009984066942706704,
      "reward_std": 0.5580916641047224,
      "rewards/SMILES_validity_reward": -0.10520835034549236,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.40625000838190317,
      "rewards/reasoning_steps_reward": 0.0416666679084301,
      "rewards/repetition_penalty_reward": -0.03265416223439388,
      "rewards/smiles_len_reward": -0.13541666883975267,
      "rewards/tag_count_reward": 0.04427083348855376,
      "step": 178
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 140.208336353302,
      "epoch": 0.072,
      "grad_norm": 13.146002769470215,
      "kl": 0.4273834228515625,
      "learning_rate": 3.4149493455847897e-07,
      "loss": 0.0279,
      "num_tokens": 5874325.0,
      "reward": -0.06133632222190499,
      "reward_std": 0.522064303047955,
      "rewards/SMILES_validity_reward": -0.1666666786186397,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.3854166707023978,
      "rewards/reasoning_steps_reward": 0.013888889225199819,
      "rewards/repetition_penalty_reward": -0.043919739313423634,
      "rewards/smiles_len_reward": -0.1302083362825215,
      "rewards/tag_count_reward": 0.05729166814126074,
      "step": 180
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 106.71875250339508,
      "epoch": 0.0728,
      "grad_norm": 10.344687461853027,
      "kl": 0.3924560546875,
      "learning_rate": 3.3813275445496766e-07,
      "loss": -0.0579,
      "num_tokens": 5933722.0,
      "reward": 0.10793787596048787,
      "reward_std": 0.6260992332245223,
      "rewards/SMILES_validity_reward": 0.011458318680524826,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.5312500074505806,
      "rewards/reasoning_steps_reward": 0.038194445660337806,
      "rewards/repetition_penalty_reward": -0.020795845077373087,
      "rewards/smiles_len_reward": -0.15104166930541396,
      "rewards/tag_count_reward": 0.039062500931322575,
      "step": 182
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 129.43750321865082,
      "epoch": 0.0736,
      "grad_norm": 11.922599792480469,
      "kl": 0.5929412841796875,
      "learning_rate": 3.347522715914262e-07,
      "loss": 0.0735,
      "num_tokens": 5995300.0,
      "reward": 0.14200845459708944,
      "reward_std": 0.6097154561430216,
      "rewards/SMILES_validity_reward": 0.06666666129603982,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.5208333395421505,
      "rewards/reasoning_steps_reward": 0.0034722222480922937,
      "rewards/repetition_penalty_reward": -0.021409569744719192,
      "rewards/smiles_len_reward": -0.12500000279396772,
      "rewards/tag_count_reward": 0.03385416744276881,
      "step": 184
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 72.20833587646484,
      "epoch": 0.0744,
      "grad_norm": 13.126518249511719,
      "kl": 0.7501220703125,
      "learning_rate": 3.313541880015877e-07,
      "loss": 0.0348,
      "num_tokens": 6051384.0,
      "reward": 0.2482103147485759,
      "reward_std": 0.6659532003104687,
      "rewards/SMILES_validity_reward": 0.14999999664723873,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.6562500074505806,
      "rewards/reasoning_steps_reward": 0.027777778450399637,
      "rewards/repetition_penalty_reward": -0.009738226246554404,
      "rewards/smiles_len_reward": -0.11979166930541396,
      "rewards/tag_count_reward": 0.06510416744276881,
      "step": 186
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 83.80208563804626,
      "epoch": 0.0752,
      "grad_norm": 11.092013359069824,
      "kl": 0.70703125,
      "learning_rate": 3.279392093743747e-07,
      "loss": 0.0463,
      "num_tokens": 6108581.0,
      "reward": 0.36337129410821944,
      "reward_std": 0.6607330553233624,
      "rewards/SMILES_validity_reward": 0.3333333395421505,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.6250000055879354,
      "rewards/reasoning_steps_reward": 0.010416666744276881,
      "rewards/repetition_penalty_reward": -0.012121445353841409,
      "rewards/smiles_len_reward": -0.10937500232830644,
      "rewards/tag_count_reward": 0.03645833418704569,
      "step": 188
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 85.01041889190674,
      "epoch": 0.076,
      "grad_norm": 14.43812370300293,
      "kl": 2.8946533203125,
      "learning_rate": 3.245080449073459e-07,
      "loss": 0.0825,
      "num_tokens": 6165894.0,
      "reward": 0.35271822242066264,
      "reward_std": 0.6430098190903664,
      "rewards/SMILES_validity_reward": 0.28333333088085055,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.7083333358168602,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.008235604094807059,
      "rewards/smiles_len_reward": -0.1145833358168602,
      "rewards/tag_count_reward": 0.03125000046566129,
      "step": 190
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 92.05208504199982,
      "epoch": 0.0768,
      "grad_norm": 12.526323318481445,
      "kl": 0.8994140625,
      "learning_rate": 3.210614071594162e-07,
      "loss": -0.0314,
      "num_tokens": 6223883.0,
      "reward": 0.4647933058440685,
      "reward_std": 0.5719528524205089,
      "rewards/SMILES_validity_reward": 0.43333333311602473,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.71875,
      "rewards/reasoning_steps_reward": 0.020833333488553762,
      "rewards/repetition_penalty_reward": -0.013005562825128436,
      "rewards/smiles_len_reward": -0.09375000279396772,
      "rewards/tag_count_reward": 0.04427083418704569,
      "step": 192
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 120.45833647251129,
      "epoch": 0.0776,
      "grad_norm": 10.05866813659668,
      "kl": 1.5377197265625,
      "learning_rate": 3.1760001190287695e-07,
      "loss": 0.056,
      "num_tokens": 6284599.0,
      "reward": 0.5278361081145704,
      "reward_std": 0.5975875779986382,
      "rewards/SMILES_validity_reward": 0.5000000088475645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.7708333283662796,
      "rewards/reasoning_steps_reward": 0.0069444444961845875,
      "rewards/repetition_penalty_reward": -0.03327204444212839,
      "rewards/smiles_len_reward": -0.06770833535119891,
      "rewards/tag_count_reward": 0.05989583441987634,
      "step": 194
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 91.43750321865082,
      "epoch": 0.0784,
      "grad_norm": 1621.8258056640625,
      "kl": 232.828125,
      "learning_rate": 3.141245779747502e-07,
      "loss": 0.2962,
      "num_tokens": 6342529.0,
      "reward": 0.5469494787976146,
      "reward_std": 0.5157439392060041,
      "rewards/SMILES_validity_reward": 0.5000000055879354,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8124999981373549,
      "rewards/reasoning_steps_reward": 0.031250000931322575,
      "rewards/repetition_penalty_reward": -0.012277308269403875,
      "rewards/smiles_len_reward": -0.06250000186264515,
      "rewards/tag_count_reward": 0.07552083465270698,
      "step": 196
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 94.01041972637177,
      "epoch": 0.0792,
      "grad_norm": 18.095582962036133,
      "kl": 9.997314453125,
      "learning_rate": 3.106358271275056e-07,
      "loss": 0.056,
      "num_tokens": 6400706.0,
      "reward": 0.5480379769578576,
      "reward_std": 0.5921958331018686,
      "rewards/SMILES_validity_reward": 0.5166666777804494,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.7812500037252903,
      "rewards/reasoning_steps_reward": 0.013888889225199819,
      "rewards/repetition_penalty_reward": -0.00955194083508104,
      "rewards/smiles_len_reward": -0.07291666837409139,
      "rewards/tag_count_reward": 0.08854166744276881,
      "step": 198
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 65.22916853427887,
      "epoch": 0.08,
      "grad_norm": 16.6898250579834,
      "kl": 2.7962646484375,
      "learning_rate": 3.0713448387917227e-07,
      "loss": 0.013,
      "num_tokens": 6456120.0,
      "reward": 0.6376010160893202,
      "reward_std": 0.48966592181750457,
      "rewards/SMILES_validity_reward": 0.6333333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8125,
      "rewards/reasoning_steps_reward": 0.017361111473292112,
      "rewards/repetition_penalty_reward": -0.01218536806118209,
      "rewards/smiles_len_reward": -0.04687500139698386,
      "rewards/tag_count_reward": 0.04687500116415322,
      "step": 200
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 82.77083480358124,
      "epoch": 0.0808,
      "grad_norm": 12.10589599609375,
      "kl": 1.421875,
      "learning_rate": 3.0362127536287636e-07,
      "loss": 0.0907,
      "num_tokens": 6513218.0,
      "reward": 0.7045793011784554,
      "reward_std": 0.3827786281472072,
      "rewards/SMILES_validity_reward": 0.7166666761040688,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8437499925494194,
      "rewards/reasoning_steps_reward": 0.02083333395421505,
      "rewards/repetition_penalty_reward": -0.009937431663274765,
      "rewards/smiles_len_reward": -0.06250000186264515,
      "rewards/tag_count_reward": 0.04947916720993817,
      "step": 202
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 47.45833492279053,
      "epoch": 0.0816,
      "grad_norm": 16.462682723999023,
      "kl": 1.3189697265625,
      "learning_rate": 3.0009693117583523e-07,
      "loss": 0.0142,
      "num_tokens": 6566926.0,
      "reward": 0.6629129499197006,
      "reward_std": 0.4532448905520141,
      "rewards/SMILES_validity_reward": 0.633333345875144,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0021217708999756724,
      "rewards/smiles_len_reward": -0.06250000186264515,
      "rewards/tag_count_reward": 0.0416666679084301,
      "step": 204
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 53.000001430511475,
      "epoch": 0.0824,
      "grad_norm": 11.786378860473633,
      "kl": 1.1549072265625,
      "learning_rate": 2.965621832278401e-07,
      "loss": 0.0448,
      "num_tokens": 6621166.0,
      "reward": 0.7589017078280449,
      "reward_std": 0.4288171596126631,
      "rewards/SMILES_validity_reward": 0.7666666805744171,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499925494194,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.006817507412051782,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.03645833418704569,
      "step": 206
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 69.76041889190674,
      "epoch": 0.0832,
      "grad_norm": 17.44184112548828,
      "kl": 4.21337890625,
      "learning_rate": 2.9301776558925875e-07,
      "loss": 0.0582,
      "num_tokens": 6677015.0,
      "reward": 0.7213433459401131,
      "reward_std": 0.4034923327853903,
      "rewards/SMILES_validity_reward": 0.7166666742414236,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499888241291,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.011567762594495434,
      "rewards/smiles_len_reward": -0.0416666679084301,
      "rewards/tag_count_reward": 0.031249999767169356,
      "step": 208
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 55.36458468437195,
      "epoch": 0.084,
      "grad_norm": 9.37515926361084,
      "kl": 2.1453857421875,
      "learning_rate": 2.894644143385885e-07,
      "loss": 0.0234,
      "num_tokens": 6731482.0,
      "reward": 0.7111485507339239,
      "reward_std": 0.4254695660783909,
      "rewards/SMILES_validity_reward": 0.7166666761040688,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8541666679084301,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.0015365626823040657,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.05468750116415322,
      "step": 210
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 65.73958432674408,
      "epoch": 0.0848,
      "grad_norm": 16.811309814453125,
      "kl": 3.41455078125,
      "learning_rate": 2.859028674095937e-07,
      "loss": 0.0406,
      "num_tokens": 6786945.0,
      "reward": 0.7542203683406115,
      "reward_std": 0.39707838895265013,
      "rewards/SMILES_validity_reward": 0.7666666749864817,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8958333320915699,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.019776776432991028,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.03906250069849193,
      "step": 212
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 47.010417461395264,
      "epoch": 0.0856,
      "grad_norm": 9.183852195739746,
      "kl": 1.72039794921875,
      "learning_rate": 2.823338644380566e-07,
      "loss": 0.0129,
      "num_tokens": 6840610.0,
      "reward": 0.8614385165274143,
      "reward_std": 0.19072945657535456,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9687499962747097,
      "rewards/reasoning_steps_reward": 0.0069444444961845875,
      "rewards/repetition_penalty_reward": -0.0029773336136713624,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.0416666679084301,
      "step": 214
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 40.33333420753479,
      "epoch": 0.0864,
      "grad_norm": 12.742477416992188,
      "kl": 2.152099609375,
      "learning_rate": 2.7875814660817504e-07,
      "loss": 0.0156,
      "num_tokens": 6893634.0,
      "reward": 0.8851364441215992,
      "reward_std": 0.17163086455548182,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9687499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0028036211151629686,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.05208333395421505,
      "step": 216
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 49.218750953674316,
      "epoch": 0.0872,
      "grad_norm": 13.92532730102539,
      "kl": 2.0343017578125,
      "learning_rate": 2.751764564986396e-07,
      "loss": 0.0608,
      "num_tokens": 6947511.0,
      "reward": 0.7820510920137167,
      "reward_std": 0.30305373720329953,
      "rewards/SMILES_validity_reward": 0.8000000081956387,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0008443895567324944,
      "rewards/smiles_len_reward": -0.0416666679084301,
      "rewards/tag_count_reward": 0.04427083395421505,
      "step": 218
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 47.88541829586029,
      "epoch": 0.088,
      "grad_norm": 15.1016206741333,
      "kl": 5.68994140625,
      "learning_rate": 2.715895379284194e-07,
      "loss": 0.0248,
      "num_tokens": 7001260.0,
      "reward": 0.7979313395917416,
      "reward_std": 0.287646692362614,
      "rewards/SMILES_validity_reward": 0.8166666738688946,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833246409893,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.01079208473674953,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.018229166977107525,
      "step": 220
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 53.968751072883606,
      "epoch": 0.0888,
      "grad_norm": 14.147798538208008,
      "kl": 1.9619140625,
      "learning_rate": 2.6799813580229174e-07,
      "loss": 0.0081,
      "num_tokens": 7055593.0,
      "reward": 0.8779301084578037,
      "reward_std": 0.15663258303538896,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0019502766517689452,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.03645833418704569,
      "step": 222
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 60.437501311302185,
      "epoch": 0.0896,
      "grad_norm": 10.820075988769531,
      "kl": 8.6739501953125,
      "learning_rate": 2.6440299595614606e-07,
      "loss": 0.0341,
      "num_tokens": 7110547.0,
      "reward": 0.7884009815752506,
      "reward_std": 0.3595552204642445,
      "rewards/SMILES_validity_reward": 0.8166666757315397,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8854166567325592,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.012345630770141724,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.04427083441987634,
      "step": 224
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 60.79166853427887,
      "epoch": 0.0904,
      "grad_norm": 8.08423900604248,
      "kl": 3.1444091796875,
      "learning_rate": 2.6080486500209347e-07,
      "loss": 0.04,
      "num_tokens": 7165535.0,
      "reward": 0.7426584232598543,
      "reward_std": 0.2605252732464578,
      "rewards/SMILES_validity_reward": 0.723958333954215,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.0031045216892380267,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.06510416767559946,
      "step": 226
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 37.729167222976685,
      "epoch": 0.0912,
      "grad_norm": 268.18206787109375,
      "kl": 43.56005859375,
      "learning_rate": 2.572044901734166e-07,
      "loss": 0.0062,
      "num_tokens": 7218309.0,
      "reward": 0.7288062907755375,
      "reward_std": 0.28436800511553884,
      "rewards/SMILES_validity_reward": 0.7020833417773247,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0015219146735034883,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.06250000046566129,
      "step": 228
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 38.33333432674408,
      "epoch": 0.092,
      "grad_norm": 12.932578086853027,
      "kl": 2.032470703125,
      "learning_rate": 2.536026191693893e-07,
      "loss": 0.0224,
      "num_tokens": 7271141.0,
      "reward": 0.8507591001689434,
      "reward_std": 0.25005635869456455,
      "rewards/SMILES_validity_reward": 0.8666666746139526,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9687499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0017853352474048734,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.0416666679084301,
      "step": 230
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 59.989585518836975,
      "epoch": 0.0928,
      "grad_norm": 77.25758361816406,
      "kl": 25.74658203125,
      "learning_rate": 2.5e-07,
      "loss": 0.098,
      "num_tokens": 7326052.0,
      "reward": 0.8249511867761612,
      "reward_std": 0.275203651515767,
      "rewards/SMILES_validity_reward": 0.8500000089406967,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.010416666744276881,
      "rewards/repetition_penalty_reward": -0.005176817707251757,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.07031250093132257,
      "step": 232
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 42.56250071525574,
      "epoch": 0.0936,
      "grad_norm": 12.249787330627441,
      "kl": 14.455810546875,
      "learning_rate": 2.4639738083061073e-07,
      "loss": -0.0119,
      "num_tokens": 7379290.0,
      "reward": 0.7708266153931618,
      "reward_std": 0.23000530700664967,
      "rewards/SMILES_validity_reward": 0.7520833369344473,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0016309260972775519,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.12239583325572312,
      "step": 234
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 48.91666758060455,
      "epoch": 0.0944,
      "grad_norm": 25.05546760559082,
      "kl": 10.76171875,
      "learning_rate": 2.4279550982658345e-07,
      "loss": 0.0693,
      "num_tokens": 7433138.0,
      "reward": 0.7949926014989614,
      "reward_std": 0.28037455228331964,
      "rewards/SMILES_validity_reward": 0.8166666748002172,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.011533659446286038,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.05208333395421505,
      "step": 236
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 59.07291853427887,
      "epoch": 0.0952,
      "grad_norm": 8.763030052185059,
      "kl": 1.9202880859375,
      "learning_rate": 2.3919513499790646e-07,
      "loss": 0.1179,
      "num_tokens": 7487961.0,
      "reward": 0.8150318637490273,
      "reward_std": 0.3204096044646576,
      "rewards/SMILES_validity_reward": 0.8333333432674408,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.011140864342451096,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.10937500186264515,
      "step": 238
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 71.48958492279053,
      "epoch": 0.096,
      "grad_norm": 8.304740905761719,
      "kl": 2.7021484375,
      "learning_rate": 2.3559700404385394e-07,
      "loss": 0.0967,
      "num_tokens": 7543976.0,
      "reward": 0.8159580808132887,
      "reward_std": 0.276808429392986,
      "rewards/SMILES_validity_reward": 0.8500000070780516,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8854166641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0065662864144542255,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.08593750256113708,
      "step": 240
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 69.59375298023224,
      "epoch": 0.0968,
      "grad_norm": 11.63044548034668,
      "kl": 1.60302734375,
      "learning_rate": 2.3200186419770823e-07,
      "loss": 0.064,
      "num_tokens": 7599809.0,
      "reward": 0.8393127992749214,
      "reward_std": 0.272423698563216,
      "rewards/SMILES_validity_reward": 0.8666666746139526,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.004269022305379622,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.049479167675599456,
      "step": 242
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 55.96875178813934,
      "epoch": 0.0976,
      "grad_norm": 9.17193603515625,
      "kl": 2.159423828125,
      "learning_rate": 2.284104620715807e-07,
      "loss": -0.0007,
      "num_tokens": 7654334.0,
      "reward": 0.7993007898330688,
      "reward_std": 0.13739498698851094,
      "rewards/SMILES_validity_reward": 0.8020833358168602,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.004389131907373667,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.08072916767559946,
      "step": 244
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 54.104168176651,
      "epoch": 0.0984,
      "grad_norm": 13.933065414428711,
      "kl": 4.1160888671875,
      "learning_rate": 2.2482354350136043e-07,
      "loss": -0.052,
      "num_tokens": 7708680.0,
      "reward": 0.8486762568354607,
      "reward_std": 0.256513943313621,
      "rewards/SMILES_validity_reward": 0.8666666746139526,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0034722222480922937,
      "rewards/repetition_penalty_reward": -0.0026484212721697986,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.13281250023283064,
      "step": 246
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 56.34375178813934,
      "epoch": 0.0992,
      "grad_norm": 11.744587898254395,
      "kl": 8.788330078125,
      "learning_rate": 2.2124185339182496e-07,
      "loss": 0.0587,
      "num_tokens": 7763241.0,
      "reward": 0.8349000215530396,
      "reward_std": 0.28379016066901386,
      "rewards/SMILES_validity_reward": 0.8666666746139526,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0015219679335132241,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.08593750023283064,
      "step": 248
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 56.43750286102295,
      "epoch": 0.1,
      "grad_norm": 15.010062217712402,
      "kl": 4.990234375,
      "learning_rate": 2.1766613556194344e-07,
      "loss": 0.0464,
      "num_tokens": 7817811.0,
      "reward": 0.8480795957148075,
      "reward_std": 0.25808324449462816,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.009830351526034065,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.07812500093132257,
      "step": 250
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 45.85416758060455,
      "epoch": 0.1008,
      "grad_norm": 41.247093200683594,
      "kl": 19.83984375,
      "learning_rate": 2.1409713259040628e-07,
      "loss": 0.0715,
      "num_tokens": 7871365.0,
      "reward": 0.8904657512903214,
      "reward_std": 0.1805916415178217,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.014093825098825619,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.14583333558402956,
      "step": 252
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 59.23958468437195,
      "epoch": 0.1016,
      "grad_norm": 13.48940372467041,
      "kl": 20.977294921875,
      "learning_rate": 2.105355856614115e-07,
      "loss": 0.0556,
      "num_tokens": 7926204.0,
      "reward": 0.7457279786467552,
      "reward_std": 0.2488641007221304,
      "rewards/SMILES_validity_reward": 0.7354166712611914,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499962747097,
      "rewards/reasoning_steps_reward": 0.013888889225199819,
      "rewards/repetition_penalty_reward": -0.004005965311080217,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.1171875,
      "step": 254
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 49.5208340883255,
      "epoch": 0.1024,
      "grad_norm": 11.00932502746582,
      "kl": 3.1722412109375,
      "learning_rate": 2.069822344107413e-07,
      "loss": 0.0286,
      "num_tokens": 7980110.0,
      "reward": 0.8951140381395817,
      "reward_std": 0.16878563386853784,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.009277544566430151,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.12500000069849193,
      "step": 256
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 42.94791805744171,
      "epoch": 0.1032,
      "grad_norm": 30.966983795166016,
      "kl": 6.7318115234375,
      "learning_rate": 2.034378167721599e-07,
      "loss": 0.0322,
      "num_tokens": 8033385.0,
      "reward": 0.8276818729937077,
      "reward_std": 0.30462851375341415,
      "rewards/SMILES_validity_reward": 0.8500000089406967,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8958333246409893,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0013075313763692975,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.15625,
      "step": 258
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 49.312501192092896,
      "epoch": 0.104,
      "grad_norm": 14.884501457214355,
      "kl": 382.56396484375,
      "learning_rate": 1.9990306882416485e-07,
      "loss": 0.3722,
      "num_tokens": 8087271.0,
      "reward": 0.8172137457877398,
      "reward_std": 0.259702468290925,
      "rewards/SMILES_validity_reward": 0.8333333358168602,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.002343074476812035,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.13802083441987634,
      "step": 260
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 52.562500953674316,
      "epoch": 0.1048,
      "grad_norm": 42.34991455078125,
      "kl": 13.2562255859375,
      "learning_rate": 1.9637872463712362e-07,
      "loss": -0.0042,
      "num_tokens": 8141469.0,
      "reward": 0.8395539987832308,
      "reward_std": 0.25605778163298965,
      "rewards/SMILES_validity_reward": 0.8614583387970924,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.0018571128166513517,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.12760416674427688,
      "step": 262
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 68.10416960716248,
      "epoch": 0.1056,
      "grad_norm": 15.944575309753418,
      "kl": 7.031494140625,
      "learning_rate": 1.9286551612082773e-07,
      "loss": 0.0701,
      "num_tokens": 8197159.0,
      "reward": 0.7490371651947498,
      "reward_std": 0.2526437883498147,
      "rewards/SMILES_validity_reward": 0.7354166731238365,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499962747097,
      "rewards/reasoning_steps_reward": 0.013888889225199819,
      "rewards/repetition_penalty_reward": -0.017789341538446024,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.15364583395421505,
      "step": 264
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 51.593751192092896,
      "epoch": 0.1064,
      "grad_norm": 15.867863655090332,
      "kl": 2011.879150390625,
      "learning_rate": 1.8936417287249446e-07,
      "loss": 2.0318,
      "num_tokens": 8251264.0,
      "reward": 0.8784895241260529,
      "reward_std": 0.19250585057307035,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833320915699,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.006772592620109208,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.11458333465270698,
      "step": 266
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 73.71875202655792,
      "epoch": 0.1072,
      "grad_norm": 17.923137664794922,
      "kl": 5.484130859375,
      "learning_rate": 1.8587542202524985e-07,
      "loss": 0.1186,
      "num_tokens": 8307493.0,
      "reward": 0.8079846333712339,
      "reward_std": 0.3059961917460896,
      "rewards/SMILES_validity_reward": 0.8333333414047956,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8749999925494194,
      "rewards/reasoning_steps_reward": 0.013888889225199819,
      "rewards/repetition_penalty_reward": -0.012168978486442938,
      "rewards/smiles_len_reward": -0.0416666679084301,
      "rewards/tag_count_reward": 0.16145833441987634,
      "step": 268
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 44.98958456516266,
      "epoch": 0.108,
      "grad_norm": 13.717728614807129,
      "kl": 193.90045166015625,
      "learning_rate": 1.82399988097123e-07,
      "loss": 0.172,
      "num_tokens": 8360964.0,
      "reward": 0.8983778692781925,
      "reward_std": 0.15485689049819484,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9375000037252903,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0026809167611645535,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.15625000023283064,
      "step": 270
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 33.04166758060455,
      "epoch": 0.1088,
      "grad_norm": 17.957460403442383,
      "kl": 26.424072265625,
      "learning_rate": 1.7893859284058378e-07,
      "loss": 0.0096,
      "num_tokens": 8413288.0,
      "reward": 0.9149191044270992,
      "reward_std": 0.12943543260917068,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0008101852145045996,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.12500000069849193,
      "step": 272
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 62.18750083446503,
      "epoch": 0.1096,
      "grad_norm": 25.32672119140625,
      "kl": 17.261962890625,
      "learning_rate": 1.7549195509265407e-07,
      "loss": 0.0172,
      "num_tokens": 8468410.0,
      "reward": 0.8183866981416941,
      "reward_std": 0.28250074468087405,
      "rewards/SMILES_validity_reward": 0.8500000070780516,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8749999925494194,
      "rewards/reasoning_steps_reward": 0.0069444444961845875,
      "rewards/repetition_penalty_reward": -0.010057752471766435,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.1484375016298145,
      "step": 274
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 51.82291805744171,
      "epoch": 0.1104,
      "grad_norm": 13.775248527526855,
      "kl": 12.86962890625,
      "learning_rate": 1.7206079062562536e-07,
      "loss": -0.031,
      "num_tokens": 8522537.0,
      "reward": 0.8880144283175468,
      "reward_std": 0.19194586825324222,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.00266948455828242,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.13802083488553762,
      "step": 276
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 50.718751668930054,
      "epoch": 0.1112,
      "grad_norm": 15.227043151855469,
      "kl": 6.1226806640625,
      "learning_rate": 1.6864581199841226e-07,
      "loss": 0.0581,
      "num_tokens": 8576558.0,
      "reward": 0.8818176276981831,
      "reward_std": 0.20186374394688755,
      "rewards/SMILES_validity_reward": 0.9000000059604645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.005262482751277275,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.19531250302679837,
      "step": 278
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 53.6250022649765,
      "epoch": 0.112,
      "grad_norm": 15.401267051696777,
      "kl": 45.770263671875,
      "learning_rate": 1.6524772840857388e-07,
      "loss": 0.0846,
      "num_tokens": 8630858.0,
      "reward": 0.8197834901511669,
      "reward_std": 0.32834628655109555,
      "rewards/SMILES_validity_reward": 0.8166666757315397,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.003207936490071006,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.16145833395421505,
      "step": 280
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 34.51041769981384,
      "epoch": 0.1128,
      "grad_norm": 26.33031463623047,
      "kl": 28.7257080078125,
      "learning_rate": 1.6186724554503237e-07,
      "loss": 0.0075,
      "num_tokens": 8683323.0,
      "reward": 0.9235701858997345,
      "reward_std": 0.1269298658007756,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9687499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0028410269296728075,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.19791666651144624,
      "step": 282
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 39.593750953674316,
      "epoch": 0.1136,
      "grad_norm": 20.33456802368164,
      "kl": 40.810302734375,
      "learning_rate": 1.5850506544152103e-07,
      "loss": 0.0284,
      "num_tokens": 8736276.0,
      "reward": 0.8979808054864407,
      "reward_std": 0.15943958633579314,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333320915699,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.001964016948477365,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.1953125004656613,
      "step": 284
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 41.322917342185974,
      "epoch": 0.1144,
      "grad_norm": 18.296335220336914,
      "kl": 10.76171875,
      "learning_rate": 1.5516188633079107e-07,
      "loss": 0.0069,
      "num_tokens": 8789395.0,
      "reward": 0.8857920374721289,
      "reward_std": 0.17018686461960897,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.009268370922654867,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.21093750023283064,
      "step": 286
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 59.20833480358124,
      "epoch": 0.1152,
      "grad_norm": 11.021564483642578,
      "kl": 40.314453125,
      "learning_rate": 1.5183840249960784e-07,
      "loss": 0.0934,
      "num_tokens": 8844231.0,
      "reward": 0.8584657311439514,
      "reward_std": 0.23017027985770255,
      "rewards/SMILES_validity_reward": 0.8666666727513075,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9375,
      "rewards/reasoning_steps_reward": 0.017361111473292112,
      "rewards/repetition_penalty_reward": -0.007184128480730578,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.21093749930150807,
      "step": 288
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 52.385417222976685,
      "epoch": 0.116,
      "grad_norm": 64.6243896484375,
      "kl": 17.584716796875,
      "learning_rate": 1.4853530414456612e-07,
      "loss": 0.0556,
      "num_tokens": 8898412.0,
      "reward": 0.8611728437244892,
      "reward_std": 0.2143052279134281,
      "rewards/SMILES_validity_reward": 0.8833333365619183,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.01952289731707424,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.21354166534729302,
      "step": 290
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 56.11458492279053,
      "epoch": 0.1168,
      "grad_norm": 34.26335906982422,
      "kl": 36.561767578125,
      "learning_rate": 1.4525327722875568e-07,
      "loss": 0.0401,
      "num_tokens": 8952951.0,
      "reward": 0.8726859986782074,
      "reward_std": 0.2261066745268181,
      "rewards/SMILES_validity_reward": 0.9000000059604645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.008037056235480122,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.2109374995343387,
      "step": 292
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 97.10416972637177,
      "epoch": 0.1176,
      "grad_norm": 45.39623260498047,
      "kl": 15.76025390625,
      "learning_rate": 1.4199300333930515e-07,
      "loss": 0.0902,
      "num_tokens": 9011425.0,
      "reward": 0.7532959198579192,
      "reward_std": 0.36259316198993474,
      "rewards/SMILES_validity_reward": 0.7666666712611914,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8229166604578495,
      "rewards/reasoning_steps_reward": 0.017361111473292112,
      "rewards/repetition_penalty_reward": -0.004715499031590298,
      "rewards/smiles_len_reward": -0.046875000931322575,
      "rewards/tag_count_reward": 0.23177083185873926,
      "step": 294
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 48.62500178813934,
      "epoch": 0.1184,
      "grad_norm": 12.377687454223633,
      "kl": 6.2955322265625,
      "learning_rate": 1.3875515954583523e-07,
      "loss": 0.0367,
      "num_tokens": 9065245.0,
      "reward": 0.8059816248714924,
      "reward_std": 0.1653146606986411,
      "rewards/SMILES_validity_reward": 0.7854166682809591,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0032058514771051705,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.2369791674427688,
      "step": 296
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 46.04166805744171,
      "epoch": 0.1192,
      "grad_norm": 12.048561096191406,
      "kl": 7.322998046875,
      "learning_rate": 1.3554041825985e-07,
      "loss": -0.0214,
      "num_tokens": 9118817.0,
      "reward": 0.8983332067728043,
      "reward_std": 0.13631578732747585,
      "rewards/SMILES_validity_reward": 0.9166666679084301,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333320915699,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0036484165393630974,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.21093750093132257,
      "step": 298
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 62.3750022649765,
      "epoch": 0.12,
      "grad_norm": 23.098913192749023,
      "kl": 142.0855712890625,
      "learning_rate": 1.323494470950949e-07,
      "loss": 0.2089,
      "num_tokens": 9173957.0,
      "reward": 0.8859845735132694,
      "reward_std": 0.20868572185281664,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.002134755515726283,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.2057291674427688,
      "step": 300
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 56.55208492279053,
      "epoch": 0.1208,
      "grad_norm": 19.67671012878418,
      "kl": 7.3123779296875,
      "learning_rate": 1.2918290872891236e-07,
      "loss": 0.0589,
      "num_tokens": 9228538.0,
      "reward": 0.8848470933735371,
      "reward_std": 0.213736486970447,
      "rewards/SMILES_validity_reward": 0.9000000059604645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.003613690787460655,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.21354166697710752,
      "step": 302
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 70.11458575725555,
      "epoch": 0.1216,
      "grad_norm": 6248.3740234375,
      "kl": 1183.59375,
      "learning_rate": 1.260414607646213e-07,
      "loss": 1.1913,
      "num_tokens": 9284421.0,
      "reward": 0.9230890087783337,
      "reward_std": 0.13145335903391242,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.005048613820690662,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.28385416651144624,
      "step": 304
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 78.30208432674408,
      "epoch": 0.1224,
      "grad_norm": 22.36941146850586,
      "kl": 9.85986328125,
      "learning_rate": 1.2292575559495143e-07,
      "loss": 0.0109,
      "num_tokens": 9341090.0,
      "reward": 0.9161470979452133,
      "reward_std": 0.16511259728576988,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166641831398,
      "rewards/reasoning_steps_reward": 0.0034722222480922937,
      "rewards/repetition_penalty_reward": -0.005544163461308926,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3020833330228925,
      "step": 306
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 59.104168176651,
      "epoch": 0.1232,
      "grad_norm": 33.61720275878906,
      "kl": 31.44189453125,
      "learning_rate": 1.1983644026655835e-07,
      "loss": 0.0418,
      "num_tokens": 9395916.0,
      "reward": 0.8362351339310408,
      "reward_std": 0.2932694805203937,
      "rewards/SMILES_validity_reward": 0.8333333432674408,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.002233220911875833,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.24479166604578495,
      "step": 308
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 44.29166758060455,
      "epoch": 0.124,
      "grad_norm": 15.68494987487793,
      "kl": 8.397705078125,
      "learning_rate": 1.1677415634565066e-07,
      "loss": 0.0084,
      "num_tokens": 9449320.0,
      "reward": 0.8074449226260185,
      "reward_std": 0.175186739070341,
      "rewards/SMILES_validity_reward": 0.7854166682809591,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.001593647408299148,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.2812500037252903,
      "step": 310
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 54.87500178813934,
      "epoch": 0.1248,
      "grad_norm": 50.4384765625,
      "kl": 22.9573974609375,
      "learning_rate": 1.1373953978475353e-07,
      "loss": 0.0812,
      "num_tokens": 9503740.0,
      "reward": 0.9299623742699623,
      "reward_std": 0.10308579099364579,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166679084301,
      "rewards/reasoning_steps_reward": 0.017361111473292112,
      "rewards/repetition_penalty_reward": -0.00315533101093024,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3020833311602473,
      "step": 312
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 77.62500095367432,
      "epoch": 0.1256,
      "grad_norm": 14.502692222595215,
      "kl": 13.63623046875,
      "learning_rate": 1.1073322079063913e-07,
      "loss": 0.0621,
      "num_tokens": 9560344.0,
      "reward": 0.8834315277636051,
      "reward_std": 0.24336056731408462,
      "rewards/SMILES_validity_reward": 0.9000000059604645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.009956820620573126,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.3098958358168602,
      "step": 314
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 64.33333539962769,
      "epoch": 0.1264,
      "grad_norm": 12.168563842773438,
      "kl": 8.08837890625,
      "learning_rate": 1.0775582369344946e-07,
      "loss": 0.0469,
      "num_tokens": 9615672.0,
      "reward": 0.8775706607848406,
      "reward_std": 0.17083232302684337,
      "rewards/SMILES_validity_reward": 0.900000000372529,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.011273908166913316,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.25260416651144624,
      "step": 316
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 48.479168176651,
      "epoch": 0.1272,
      "grad_norm": 51.8846321105957,
      "kl": 17.612548828125,
      "learning_rate": 1.0480796681704077e-07,
      "loss": 0.0651,
      "num_tokens": 9669478.0,
      "reward": 0.896970622241497,
      "reward_std": 0.15544916148064658,
      "rewards/SMILES_validity_reward": 0.9166666697710752,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.012065992181305774,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.2994791651144624,
      "step": 318
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 40.91666841506958,
      "epoch": 0.128,
      "grad_norm": 10.52868938446045,
      "kl": 17.12109375,
      "learning_rate": 1.018902623505741e-07,
      "loss": -0.0259,
      "num_tokens": 9722558.0,
      "reward": 0.8453612364828587,
      "reward_std": 0.27566418022615835,
      "rewards/SMILES_validity_reward": 0.8500000070780516,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.001597268710611388,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.2552083320915699,
      "step": 320
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 46.979167222976685,
      "epoch": 0.1288,
      "grad_norm": 13.913373947143555,
      "kl": 7.58154296875,
      "learning_rate": 9.900331622138063e-08,
      "loss": -0.0041,
      "num_tokens": 9776220.0,
      "reward": 0.9387280717492104,
      "reward_std": 0.11939947289647534,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9791666641831398,
      "rewards/reasoning_steps_reward": 0.0069444444961845875,
      "rewards/repetition_penalty_reward": -0.004039903054945171,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3124999958090484,
      "step": 322
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 46.88541793823242,
      "epoch": 0.1296,
      "grad_norm": 13.716207504272461,
      "kl": 13.66259765625,
      "learning_rate": 9.614772796912681e-08,
      "loss": 0.0425,
      "num_tokens": 9829873.0,
      "reward": 0.8513182587921619,
      "reward_std": 0.2821584241464734,
      "rewards/SMILES_validity_reward": 0.8666666727513075,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499888241291,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0040062003245111555,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.2578125037252903,
      "step": 324
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 43.86458432674408,
      "epoch": 0.1304,
      "grad_norm": 13.790403366088867,
      "kl": 10.61859130859375,
      "learning_rate": 9.332409062130686e-08,
      "loss": 0.0133,
      "num_tokens": 9883236.0,
      "reward": 0.8828205205500126,
      "reward_std": 0.21784398559248075,
      "rewards/SMILES_validity_reward": 0.9000000040978193,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666641831398,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.003046113546588458,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.2864583348855376,
      "step": 326
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 54.968750953674316,
      "epoch": 0.1312,
      "grad_norm": 20.811195373535156,
      "kl": 10.6666259765625,
      "learning_rate": 9.053299057008699e-08,
      "loss": 0.067,
      "num_tokens": 9937665.0,
      "reward": 0.8830222934484482,
      "reward_std": 0.2312118139816448,
      "rewards/SMILES_validity_reward": 0.8833333384245634,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.006757562659913674,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3203125020954758,
      "step": 328
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 59.46875238418579,
      "epoch": 0.132,
      "grad_norm": 11.719626426696777,
      "kl": 10.68682861328125,
      "learning_rate": 8.777500745052743e-08,
      "loss": 0.0699,
      "num_tokens": 9992526.0,
      "reward": 0.8845588695257902,
      "reward_std": 0.21500852517783642,
      "rewards/SMILES_validity_reward": 0.9000000059604645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.006495966023067012,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.296875,
      "step": 330
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 67.85416865348816,
      "epoch": 0.1328,
      "grad_norm": 18.956579208374023,
      "kl": 11.107421875,
      "learning_rate": 8.505071402020892e-08,
      "loss": 0.0666,
      "num_tokens": 10048192.0,
      "reward": 0.8577948585152626,
      "reward_std": 0.25394871982280165,
      "rewards/SMILES_validity_reward": 0.8666666746139526,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666567325592,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.017365171763231046,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.30468749813735485,
      "step": 332
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 55.46875178813934,
      "epoch": 0.1336,
      "grad_norm": 19.334678649902344,
      "kl": 17.1290283203125,
      "learning_rate": 8.236067604028562e-08,
      "loss": -0.0002,
      "num_tokens": 10102669.0,
      "reward": 0.9023284614086151,
      "reward_std": 0.16440001456066966,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9375,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.005362560274079442,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3098958330228925,
      "step": 334
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 40.61458444595337,
      "epoch": 0.1344,
      "grad_norm": 22.786741256713867,
      "kl": 20.921875,
      "learning_rate": 7.970545215799327e-08,
      "loss": -0.0046,
      "num_tokens": 10155720.0,
      "reward": 0.8433664571493864,
      "reward_std": 0.28101494431030005,
      "rewards/SMILES_validity_reward": 0.850000006146729,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666567325592,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.003316073620226234,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.27343749767169356,
      "step": 336
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 62.38541805744171,
      "epoch": 0.1352,
      "grad_norm": 19.285070419311523,
      "kl": 15.4423828125,
      "learning_rate": 7.708559379063204e-08,
      "loss": -0.0467,
      "num_tokens": 10210861.0,
      "reward": 0.8785608969628811,
      "reward_std": 0.20795345888473094,
      "rewards/SMILES_validity_reward": 0.8833333384245634,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166641831398,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.004496572495554574,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.2786458367481828,
      "step": 338
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 64.51041853427887,
      "epoch": 0.136,
      "grad_norm": 12.174135208129883,
      "kl": 4.55029296875,
      "learning_rate": 7.45016450110534e-08,
      "loss": 0.0704,
      "num_tokens": 10266206.0,
      "reward": 0.9585046321153641,
      "reward_std": 0.08200875978218392,
      "rewards/SMILES_validity_reward": 0.9833333343267441,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.009225877125572879,
      "rewards/smiles_len_reward": 0.0,
      "rewards/tag_count_reward": 0.3671874990686774,
      "step": 340
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 46.33333504199982,
      "epoch": 0.1368,
      "grad_norm": 14.097783088684082,
      "kl": 17.8291015625,
      "learning_rate": 7.195414243467029e-08,
      "loss": -0.0138,
      "num_tokens": 10319806.0,
      "reward": 0.8139520492404699,
      "reward_std": 0.19127862562891096,
      "rewards/SMILES_validity_reward": 0.7854166701436043,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.001626729812414851,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.2890624972060323,
      "step": 342
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 45.479167342185974,
      "epoch": 0.1376,
      "grad_norm": 20.031068801879883,
      "kl": 6.23046875,
      "learning_rate": 6.944361510801763e-08,
      "loss": 0.0154,
      "num_tokens": 10373324.0,
      "reward": 0.9225666001439095,
      "reward_std": 0.13941001996863633,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.005064387529273517,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3411458320915699,
      "step": 344
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 58.312501668930054,
      "epoch": 0.1384,
      "grad_norm": 11.070231437683105,
      "kl": 9.16748046875,
      "learning_rate": 6.697058439888283e-08,
      "loss": 0.0035,
      "num_tokens": 10428074.0,
      "reward": 0.9040811881422997,
      "reward_std": 0.20653777720872313,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.008668714319355786,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3411458311602473,
      "step": 346
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 38.60416758060455,
      "epoch": 0.1392,
      "grad_norm": 17.219825744628906,
      "kl": 11.9794921875,
      "learning_rate": 6.453556388803288e-08,
      "loss": -0.0453,
      "num_tokens": 10480932.0,
      "reward": 0.9238246716558933,
      "reward_std": 0.161299129773397,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0008169934735633433,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3463541669771075,
      "step": 348
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 39.60416758060455,
      "epoch": 0.14,
      "grad_norm": 41.384090423583984,
      "kl": 24.66162109375,
      "learning_rate": 6.213905926255697e-08,
      "loss": 0.0626,
      "num_tokens": 10533886.0,
      "reward": 0.8788461312651634,
      "reward_std": 0.22121598839294165,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833320915699,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.0016438955790363252,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.3463541669771075,
      "step": 350
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 54.70833432674408,
      "epoch": 0.1408,
      "grad_norm": 2251.369873046875,
      "kl": 260.109375,
      "learning_rate": 5.978156821084987e-08,
      "loss": 0.3363,
      "num_tokens": 10588290.0,
      "reward": 0.906851053237915,
      "reward_std": 0.20935551589354873,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0018032030784524977,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3619791641831398,
      "step": 352
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 41.07291758060455,
      "epoch": 0.1416,
      "grad_norm": 86.15303039550781,
      "kl": 25.509521484375,
      "learning_rate": 5.7463580319254853e-08,
      "loss": 0.0125,
      "num_tokens": 10641385.0,
      "reward": 0.9007619582116604,
      "reward_std": 0.2126168095273897,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833246409893,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.00540257606189698,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3359375009313226,
      "step": 354
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 53.343751668930054,
      "epoch": 0.1424,
      "grad_norm": 22.986886978149414,
      "kl": 9.517333984375,
      "learning_rate": 5.518557697039081e-08,
      "loss": -0.0054,
      "num_tokens": 10695658.0,
      "reward": 0.9346610344946384,
      "reward_std": 0.14231333124917,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.009120229282416403,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3671875037252903,
      "step": 356
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 36.20833420753479,
      "epoch": 0.1432,
      "grad_norm": 20.674697875976562,
      "kl": 6.1153564453125,
      "learning_rate": 5.294803124318145e-08,
      "loss": -0.0081,
      "num_tokens": 10748286.0,
      "reward": 0.9437504410743713,
      "reward_std": 0.11255369227728806,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9791666641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0015593872813042253,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.3567708330228925,
      "step": 358
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 51.54166805744171,
      "epoch": 0.144,
      "grad_norm": 98.24833679199219,
      "kl": 26.5751953125,
      "learning_rate": 5.07514078146106e-08,
      "loss": 0.072,
      "num_tokens": 10802386.0,
      "reward": 0.9014865681529045,
      "reward_std": 0.18204960267757997,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0007607677835039794,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3072916674427688,
      "step": 360
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 51.531251430511475,
      "epoch": 0.1448,
      "grad_norm": 16.738651275634766,
      "kl": 11.8388671875,
      "learning_rate": 4.859616286322094e-08,
      "loss": -0.0065,
      "num_tokens": 10856485.0,
      "reward": 0.8817995861172676,
      "reward_std": 0.22947271226439625,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0033597408328205347,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.32031249813735485,
      "step": 362
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 34.98958432674408,
      "epoch": 0.1456,
      "grad_norm": 24.977575302124023,
      "kl": 15.1845703125,
      "learning_rate": 4.648274397437829e-08,
      "loss": 0.0125,
      "num_tokens": 10908996.0,
      "reward": 0.8553803451359272,
      "reward_std": 0.26896312623284757,
      "rewards/SMILES_validity_reward": 0.8500000070780516,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0003645833348855376,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.3229166667442769,
      "step": 364
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 64.43750143051147,
      "epoch": 0.1464,
      "grad_norm": 23.089859008789062,
      "kl": 14.002685546875,
      "learning_rate": 4.4411590047320617e-08,
      "loss": 0.0537,
      "num_tokens": 10964334.0,
      "reward": 0.9408868439495564,
      "reward_std": 0.11353623877221253,
      "rewards/SMILES_validity_reward": 0.9666666686534882,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.008841203583870083,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.34895833395421505,
      "step": 366
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 87.88541948795319,
      "epoch": 0.1472,
      "grad_norm": 49.889408111572266,
      "kl": 17.040771484375,
      "learning_rate": 4.2383131204010494e-08,
      "loss": 0.089,
      "num_tokens": 11021923.0,
      "reward": 0.8763796538114548,
      "reward_std": 0.24184174370020628,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833283662796,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.01589235052233562,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.33593749767169356,
      "step": 368
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 56.270835399627686,
      "epoch": 0.148,
      "grad_norm": 12.946943283081055,
      "kl": 6.9974365234375,
      "learning_rate": 4.039778869981064e-08,
      "loss": 0.0186,
      "num_tokens": 11076477.0,
      "reward": 0.7805087044835091,
      "reward_std": 0.24972321733366698,
      "rewards/SMILES_validity_reward": 0.7520833369344473,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499925494194,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.006372742980602197,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.3437500016298145,
      "step": 370
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 55.385418176651,
      "epoch": 0.1488,
      "grad_norm": 14.288317680358887,
      "kl": 13.7529296875,
      "learning_rate": 3.845597483600049e-08,
      "loss": 0.0478,
      "num_tokens": 11130946.0,
      "reward": 0.9122310355305672,
      "reward_std": 0.16964219394139946,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.007378652589977719,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3046875009313226,
      "step": 372
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 68.54166889190674,
      "epoch": 0.1496,
      "grad_norm": 27.640518188476562,
      "kl": 46.751708984375,
      "learning_rate": 3.655809287415284e-08,
      "loss": 0.0663,
      "num_tokens": 11186678.0,
      "reward": 0.8932533636689186,
      "reward_std": 0.216169131686911,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.015384390106191859,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.30208333022892475,
      "step": 374
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 70.22916865348816,
      "epoch": 0.1504,
      "grad_norm": 20.095703125,
      "kl": 14.381103515625,
      "learning_rate": 3.4704536952387285e-08,
      "loss": 0.0798,
      "num_tokens": 11242572.0,
      "reward": 0.89638776704669,
      "reward_std": 0.19948702392866835,
      "rewards/SMILES_validity_reward": 0.9000000040978193,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.02083333395421505,
      "rewards/repetition_penalty_reward": -0.015811095014214516,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.37239583395421505,
      "step": 376
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 52.60416829586029,
      "epoch": 0.1512,
      "grad_norm": 100.34427642822266,
      "kl": 31.5345458984375,
      "learning_rate": 3.2895692003518575e-08,
      "loss": 0.0002,
      "num_tokens": 11296774.0,
      "reward": 0.8779818527400494,
      "reward_std": 0.2134934167843312,
      "rewards/SMILES_validity_reward": 0.8833333365619183,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666641831398,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.002474533444910776,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3593749990686774,
      "step": 378
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 41.437501192092896,
      "epoch": 0.152,
      "grad_norm": 22.127193450927734,
      "kl": 25.8388671875,
      "learning_rate": 3.113193367511635e-08,
      "loss": 0.053,
      "num_tokens": 11349904.0,
      "reward": 0.9402668289840221,
      "reward_std": 0.1094548690598458,
      "rewards/SMILES_validity_reward": 0.9666666686534882,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0020205874752718955,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3046875027939677,
      "step": 380
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 57.28125178813934,
      "epoch": 0.1528,
      "grad_norm": 33.59663391113281,
      "kl": 21.923828125,
      "learning_rate": 2.9413628251493934e-08,
      "loss": 0.054,
      "num_tokens": 11404555.0,
      "reward": 0.8744519427418709,
      "reward_std": 0.24664068571291864,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.013888889225199819,
      "rewards/repetition_penalty_reward": -0.004787582263816148,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.33854166604578495,
      "step": 382
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 48.187501668930054,
      "epoch": 0.1536,
      "grad_norm": 13.808212280273438,
      "kl": 10.502685546875,
      "learning_rate": 2.774113257764066e-08,
      "loss": -0.0248,
      "num_tokens": 11458333.0,
      "reward": 0.9046551585197449,
      "reward_std": 0.19689062098041177,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.010741465914179571,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.32291666651144624,
      "step": 384
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 45.23958504199982,
      "epoch": 0.1544,
      "grad_norm": 16.571672439575195,
      "kl": 9.0302734375,
      "learning_rate": 2.611479398511518e-08,
      "loss": 0.0151,
      "num_tokens": 11511828.0,
      "reward": 0.9085580185055733,
      "reward_std": 0.16660702659282833,
      "rewards/SMILES_validity_reward": 0.9166666697710752,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.02083333395421505,
      "rewards/repetition_penalty_reward": -0.002962835831567645,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3489583330228925,
      "step": 386
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 40.08333420753479,
      "epoch": 0.1552,
      "grad_norm": 17.277727127075195,
      "kl": 19.247802734375,
      "learning_rate": 2.4534950219914057e-08,
      "loss": 0.0401,
      "num_tokens": 11564828.0,
      "reward": 0.934512734413147,
      "reward_std": 0.12718994682654738,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9687499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.005394943415012676,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3046874990686774,
      "step": 388
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 53.80208480358124,
      "epoch": 0.156,
      "grad_norm": 16.590299606323242,
      "kl": 11.242431640625,
      "learning_rate": 2.300192937233128e-08,
      "loss": -0.0123,
      "num_tokens": 11619145.0,
      "reward": 0.8780521750450134,
      "reward_std": 0.1917848305311054,
      "rewards/SMILES_validity_reward": 0.8833333365619183,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.004375363059807569,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3098958320915699,
      "step": 390
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 69.8645852804184,
      "epoch": 0.1568,
      "grad_norm": 20.95870018005371,
      "kl": 25.18212890625,
      "learning_rate": 2.1516049808822935e-08,
      "loss": 0.0524,
      "num_tokens": 11675004.0,
      "reward": 0.9173811189830303,
      "reward_std": 0.17535327681980561,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.010565137352386955,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3593750009313226,
      "step": 392
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 45.48958468437195,
      "epoch": 0.1576,
      "grad_norm": 13.889383316040039,
      "kl": 14.10595703125,
      "learning_rate": 2.007762010589098e-08,
      "loss": 0.0016,
      "num_tokens": 11728523.0,
      "reward": 0.9576418250799179,
      "reward_std": 0.0714196574408561,
      "rewards/SMILES_validity_reward": 0.9833333343267441,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333320915699,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.004833227023482323,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.3177083320915699,
      "step": 394
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 46.35416781902313,
      "epoch": 0.1584,
      "grad_norm": 16.284929275512695,
      "kl": 12.9434814453125,
      "learning_rate": 1.8686938986000627e-08,
      "loss": -0.0031,
      "num_tokens": 11782125.0,
      "reward": 0.9024706967175007,
      "reward_std": 0.19274842808954418,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.00133611261844635,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3177083358168602,
      "step": 396
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 56.989585518836975,
      "epoch": 0.1592,
      "grad_norm": 16.189149856567383,
      "kl": 9.690673828125,
      "learning_rate": 1.734429525554365e-08,
      "loss": -0.0609,
      "num_tokens": 11836748.0,
      "reward": 0.9177911207079887,
      "reward_std": 0.17121762363240123,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.003861064527882263,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3255208311602473,
      "step": 398
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 66.229168176651,
      "epoch": 0.16,
      "grad_norm": 93.22771453857422,
      "kl": 26.119140625,
      "learning_rate": 1.604996774486145e-08,
      "loss": 0.0444,
      "num_tokens": 11892258.0,
      "reward": 0.8573517389595509,
      "reward_std": 0.29699486307799816,
      "rewards/SMILES_validity_reward": 0.8666666746139526,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666604578495,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.011379826988559216,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.2890625,
      "step": 400
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 39.26041758060455,
      "epoch": 0.1608,
      "grad_norm": 19.394777297973633,
      "kl": 15.634765625,
      "learning_rate": 1.4804225250339281e-08,
      "loss": 0.0331,
      "num_tokens": 11945179.0,
      "reward": 0.8919323273003101,
      "reward_std": 0.1985253794118762,
      "rewards/SMILES_validity_reward": 0.8833333365619183,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9687499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.000990475993603468,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3463541679084301,
      "step": 402
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 80.62500202655792,
      "epoch": 0.1616,
      "grad_norm": 52.34968185424805,
      "kl": 23.559814453125,
      "learning_rate": 1.360732647858498e-08,
      "loss": 0.1011,
      "num_tokens": 12002071.0,
      "reward": 0.8946188390254974,
      "reward_std": 0.20518372696824372,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499962747097,
      "rewards/reasoning_steps_reward": 0.0034722222480922937,
      "rewards/repetition_penalty_reward": -0.015618514036759734,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3437500009313226,
      "step": 404
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 61.145835757255554,
      "epoch": 0.1624,
      "grad_norm": 20.986385345458984,
      "kl": 24.005859375,
      "learning_rate": 1.2459519992702311e-08,
      "loss": 0.0417,
      "num_tokens": 12057093.0,
      "reward": 0.8893291372805834,
      "reward_std": 0.21797450783196837,
      "rewards/SMILES_validity_reward": 0.9000000040978193,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.013480799068929628,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3463541669771075,
      "step": 406
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 43.91666829586029,
      "epoch": 0.1632,
      "grad_norm": 278.88031005859375,
      "kl": 38.345947265625,
      "learning_rate": 1.1361044160671629e-08,
      "loss": -0.0215,
      "num_tokens": 12110461.0,
      "reward": 0.9176699332892895,
      "reward_std": 0.1729820077889599,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.002468873586622067,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3541666669771075,
      "step": 408
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 75.67708659172058,
      "epoch": 0.164,
      "grad_norm": 18.142961502075195,
      "kl": 17.00732421875,
      "learning_rate": 1.0312127105846947e-08,
      "loss": 0.097,
      "num_tokens": 12166878.0,
      "reward": 0.8851692602038383,
      "reward_std": 0.2130603661062196,
      "rewards/SMILES_validity_reward": 0.9000000040978193,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.01341284648515284,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.3151041669771075,
      "step": 410
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 53.270835399627686,
      "epoch": 0.1648,
      "grad_norm": 1134.2469482421875,
      "kl": 161.255859375,
      "learning_rate": 9.312986659581301e-09,
      "loss": 0.2592,
      "num_tokens": 12221144.0,
      "reward": 0.8985108807682991,
      "reward_std": 0.21337384218350053,
      "rewards/SMILES_validity_reward": 0.9000000059604645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166641831398,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.007600876037031412,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3541666651144624,
      "step": 412
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 44.82291769981384,
      "epoch": 0.1656,
      "grad_norm": 14.800688743591309,
      "kl": 10.4658203125,
      "learning_rate": 8.363830315988945e-09,
      "loss": 0.0247,
      "num_tokens": 12274599.0,
      "reward": 0.9208331219851971,
      "reward_std": 0.15400783298537135,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166679084301,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0020868037827312946,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3489583358168602,
      "step": 414
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 42.52083456516266,
      "epoch": 0.1664,
      "grad_norm": 40.16646194458008,
      "kl": 24.19287109375,
      "learning_rate": 7.46485518885462e-09,
      "loss": -0.024,
      "num_tokens": 12327833.0,
      "reward": 0.9045818336308002,
      "reward_std": 0.1850796421058476,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9375,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.0010579937370494008,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.3281250009313226,
      "step": 416
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 46.656251311302185,
      "epoch": 0.1672,
      "grad_norm": 16.928592681884766,
      "kl": 14.68310546875,
      "learning_rate": 6.616247970698319e-09,
      "loss": 0.0131,
      "num_tokens": 12381464.0,
      "reward": 0.933611448854208,
      "reward_std": 0.1313752842215763,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166604578495,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.001386914482282009,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3541666632518172,
      "step": 418
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 63.625001668930054,
      "epoch": 0.168,
      "grad_norm": 12.708135604858398,
      "kl": 8.35107421875,
      "learning_rate": 5.8181848940044855e-09,
      "loss": 0.0416,
      "num_tokens": 12436724.0,
      "reward": 0.8699160404503345,
      "reward_std": 0.2740453766891733,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9166666567325592,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.007611631946929265,
      "rewards/smiles_len_reward": -0.03645833441987634,
      "rewards/tag_count_reward": 0.29947916604578495,
      "step": 420
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 54.01041805744171,
      "epoch": 0.1688,
      "grad_norm": 13.965043067932129,
      "kl": 47.185302734375,
      "learning_rate": 5.070831694623135e-09,
      "loss": 0.0942,
      "num_tokens": 12491061.0,
      "reward": 0.859567554667592,
      "reward_std": 0.20586686272872612,
      "rewards/SMILES_validity_reward": 0.8666666708886623,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499990686774,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.010054960977868177,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.3463541669771075,
      "step": 422
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 36.406251072883606,
      "epoch": 0.1696,
      "grad_norm": 25.02899932861328,
      "kl": 15.2127685546875,
      "learning_rate": 4.374343577351336e-09,
      "loss": 0.0298,
      "num_tokens": 12543708.0,
      "reward": 0.9081405811011791,
      "reward_std": 0.18163511087186635,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.007137345732189715,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3072916651144624,
      "step": 424
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 85.25000143051147,
      "epoch": 0.1704,
      "grad_norm": 129.71121215820312,
      "kl": 40.9404296875,
      "learning_rate": 3.7288651837012745e-09,
      "loss": 0.1807,
      "num_tokens": 12601044.0,
      "reward": 0.8859259001910686,
      "reward_std": 0.22173287137411535,
      "rewards/SMILES_validity_reward": 0.9000000040978193,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.003242230974137783,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.3697916632518172,
      "step": 426
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 45.937501668930054,
      "epoch": 0.1712,
      "grad_norm": 17.502323150634766,
      "kl": 11.782958984375,
      "learning_rate": 3.134530561862081e-09,
      "loss": -0.0112,
      "num_tokens": 12654606.0,
      "reward": 0.9541554935276508,
      "reward_std": 0.09704665496246889,
      "rewards/SMILES_validity_reward": 0.9666666686534882,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9791666641831398,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.0037587993138004094,
      "rewards/smiles_len_reward": -0.0052083334885537624,
      "rewards/tag_count_reward": 0.3359375046566129,
      "step": 428
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 92.07291972637177,
      "epoch": 0.172,
      "grad_norm": 35.44276809692383,
      "kl": 17.183837890625,
      "learning_rate": 2.5914631388619103e-09,
      "loss": 0.1045,
      "num_tokens": 12712597.0,
      "reward": 0.847147922962904,
      "reward_std": 0.29424209939315915,
      "rewards/SMILES_validity_reward": 0.8666666746139526,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8749999962747097,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.01966797193745151,
      "rewards/smiles_len_reward": -0.031250000931322575,
      "rewards/tag_count_reward": 0.3203125,
      "step": 430
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 66.14583551883698,
      "epoch": 0.1728,
      "grad_norm": 27.857742309570312,
      "kl": 25.1279296875,
      "learning_rate": 2.0997756949353297e-09,
      "loss": 0.0567,
      "num_tokens": 12768099.0,
      "reward": 0.7844962626695633,
      "reward_std": 0.3947558330837637,
      "rewards/SMILES_validity_reward": 0.7833333443850279,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.8749999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.013371882087085396,
      "rewards/smiles_len_reward": -0.0416666679084301,
      "rewards/tag_count_reward": 0.29166666604578495,
      "step": 432
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 41.08333420753479,
      "epoch": 0.1736,
      "grad_norm": 18.345930099487305,
      "kl": 8.83349609375,
      "learning_rate": 1.6595703401020844e-09,
      "loss": -0.0442,
      "num_tokens": 12821195.0,
      "reward": 0.9463203363120556,
      "reward_std": 0.10566021921113133,
      "rewards/SMILES_validity_reward": 0.9666666686534882,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9583333283662796,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0013811642420478165,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3333333311602473,
      "step": 434
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 56.88541841506958,
      "epoch": 0.1744,
      "grad_norm": 13.988067626953125,
      "kl": 7.73828125,
      "learning_rate": 1.2709384929615596e-09,
      "loss": 0.0832,
      "num_tokens": 12875808.0,
      "reward": 0.9088719300925732,
      "reward_std": 0.1888645300641656,
      "rewards/SMILES_validity_reward": 0.9166666716337204,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.005031941807828844,
      "rewards/smiles_len_reward": -0.02083333395421505,
      "rewards/tag_count_reward": 0.38541666604578495,
      "step": 436
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 54.47916769981384,
      "epoch": 0.1752,
      "grad_norm": 16.43869972229004,
      "kl": 43.60986328125,
      "learning_rate": 9.339608617077165e-10,
      "loss": 0.0403,
      "num_tokens": 12930190.0,
      "reward": 0.8349853716790676,
      "reward_std": 0.13201338605722412,
      "rewards/SMILES_validity_reward": 0.8020833358168602,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9687499962747097,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.0017100500699598342,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.34635416558012366,
      "step": 438
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 36.86458420753479,
      "epoch": 0.176,
      "grad_norm": 17.14826011657715,
      "kl": 28.635498046875,
      "learning_rate": 6.487074273681114e-10,
      "loss": 0.0276,
      "num_tokens": 12982881.0,
      "reward": 0.9228818565607071,
      "reward_std": 0.1395076098269783,
      "rewards/SMILES_validity_reward": 0.9500000029802322,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833320915699,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.001911880768602714,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3151041688397527,
      "step": 440
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 57.86458492279053,
      "epoch": 0.1768,
      "grad_norm": 17.069561004638672,
      "kl": 9.9276123046875,
      "learning_rate": 4.152374292708538e-10,
      "loss": 0.034,
      "num_tokens": 13037588.0,
      "reward": 0.8790950290858746,
      "reward_std": 0.25983966258354485,
      "rewards/SMILES_validity_reward": 0.8833333402872086,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9374999925494194,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.01217609562445432,
      "rewards/smiles_len_reward": -0.026041667442768812,
      "rewards/tag_count_reward": 0.3333333320915699,
      "step": 442
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 52.05208492279053,
      "epoch": 0.1776,
      "grad_norm": 22.546499252319336,
      "kl": 13.678955078125,
      "learning_rate": 2.3359935274214204e-10,
      "loss": 0.0059,
      "num_tokens": 13091737.0,
      "reward": 0.9071716498583555,
      "reward_std": 0.14055689855013043,
      "rewards/SMILES_validity_reward": 0.9166666679084301,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.006409892288502306,
      "rewards/smiles_len_reward": -0.020833333488553762,
      "rewards/tag_count_reward": 0.3385416632518172,
      "step": 444
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 47.010418176651,
      "epoch": 0.1784,
      "grad_norm": 15.367889404296875,
      "kl": 19.07421875,
      "learning_rate": 1.0383091903720665e-10,
      "loss": 0.0431,
      "num_tokens": 13145402.0,
      "reward": 0.9057391881942749,
      "reward_std": 0.1908833612978924,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9062499925494194,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.004588633368257433,
      "rewards/smiles_len_reward": -0.010416666977107525,
      "rewards/tag_count_reward": 0.3098958358168602,
      "step": 446
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 61.66666793823242,
      "epoch": 0.1792,
      "grad_norm": 11.217643737792969,
      "kl": 11.876708984375,
      "learning_rate": 2.595907750671533e-11,
      "loss": 0.0867,
      "num_tokens": 13200474.0,
      "reward": 0.9117394872009754,
      "reward_std": 0.16433565050829202,
      "rewards/SMILES_validity_reward": 0.9333333373069763,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9270833358168602,
      "rewards/reasoning_steps_reward": 0.010416666977107525,
      "rewards/repetition_penalty_reward": -0.009689704194897786,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3177083330228925,
      "step": 448
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 49.55208456516266,
      "epoch": 0.18,
      "grad_norm": 20.237873077392578,
      "kl": 16.04541015625,
      "learning_rate": 0.0,
      "loss": -0.0015,
      "num_tokens": 13254383.0,
      "reward": 0.8980720601975918,
      "reward_std": 0.19482463505119085,
      "rewards/SMILES_validity_reward": 0.9000000059604645,
      "rewards/cosine_scaled_reward": -0.4999987781047821,
      "rewards/format_reward": 0.9479166641831398,
      "rewards/reasoning_steps_reward": 0.0,
      "rewards/repetition_penalty_reward": -0.006780669005820528,
      "rewards/smiles_len_reward": -0.015625000465661287,
      "rewards/tag_count_reward": 0.3593749962747097,
      "step": 450
    },
    {
      "epoch": 0.18,
      "step": 450,
      "total_flos": 0.0,
      "train_loss": 0.053622997744823805,
      "train_runtime": 6963.8787,
      "train_samples_per_second": 3.102,
      "train_steps_per_second": 0.065
    }
  ],
  "logging_steps": 2,
  "max_steps": 450,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 25,
  "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
}