continuation_benchmark/eval_set_656/summary.json

{
  "version": "eval_set_score_v1",
  "generated_utc": "2026-05-02T22:54:25.554929Z",
  "n_plasmids": 656,
  "n_comp_windows": 12000,
  "n_surp_windows": 3784,
  "split_counts": {
    "clean_seva": 88,
    "full_set": 656,
    "hard_ood_chassis": 50,
    "dual_host": 201,
    "stress_test": 7
  },
  "per_split_completion": [
    {
      "split": "clean_seva",
      "n_plasmids": 88,
      "n_windows": 2772,
      "Base": -4.1473,
      "SFT": -3.4047,
      "RL": -6.5865,
      "RL-Base": -2.4392,
      "RL-SFT": -3.1818
    },
    {
      "split": "hard_ood_chassis",
      "n_plasmids": 50,
      "n_windows": 1914,
      "Base": -5.8742,
      "SFT": -4.9244,
      "RL": -7.7061,
      "RL-Base": -1.8319,
      "RL-SFT": -2.7817
    },
    {
      "split": "dual_host",
      "n_plasmids": 201,
      "n_windows": 10599,
      "Base": -7.0408,
      "SFT": -5.7407,
      "RL": -7.9791,
      "RL-Base": -0.9383,
      "RL-SFT": -2.2384
    },
    {
      "split": "stress_test",
      "n_plasmids": 7,
      "n_windows": 531,
      "Base": -6.5049,
      "SFT": -5.3609,
      "RL": -7.5605,
      "RL-Base": -1.0556,
      "RL-SFT": -2.1996
    },
    {
      "split": "full_set",
      "n_plasmids": 656,
      "n_windows": 30402,
      "Base": -6.6224,
      "SFT": -5.4644,
      "RL": -7.749,
      "RL-Base": -1.1266,
      "RL-SFT": -2.2846
    }
  ],
  "per_split_surprisal": [
    {
      "split": "clean_seva",
      "n_plasmids": 88,
      "n_windows": 858,
      "Base": -13.5061,
      "SFT": -7.0421,
      "RL": -10.7135,
      "RL-Base": 2.7926,
      "RL-SFT": -3.6714
    },
    {
      "split": "hard_ood_chassis",
      "n_plasmids": 50,
      "n_windows": 669,
      "Base": -13.4235,
      "SFT": -7.7557,
      "RL": -10.7485,
      "RL-Base": 2.675,
      "RL-SFT": -2.9928
    },
    {
      "split": "dual_host",
      "n_plasmids": 201,
      "n_windows": 4077,
      "Base": -12.973,
      "SFT": -7.9611,
      "RL": -10.6786,
      "RL-Base": 2.2944,
      "RL-SFT": -2.7175
    },
    {
      "split": "stress_test",
      "n_plasmids": 7,
      "n_windows": 177,
      "Base": -12.9169,
      "SFT": -7.4829,
      "RL": -10.3571,
      "RL-Base": 2.5598,
      "RL-SFT": -2.8742
    },
    {
      "split": "full_set",
      "n_plasmids": 656,
      "n_windows": 11352,
      "Base": -12.9963,
      "SFT": -7.7041,
      "RL": -10.6209,
      "RL-Base": 2.3754,
      "RL-SFT": -2.9168
    }
  ]
}