{
  "data_revision": "e564e2ac567f774d1880aa1db6ce97afb8c519b7",
  "group_count": 3265,
  "group_overlap_count": 0,
  "grouping_note": "All observed groups are singleton rows; speaker/conversation/voice identity separation is not established.",
  "hyperparameters": {
    "epochs": 1000,
    "fpr_budget": 0.02,
    "l2": 0.001,
    "learning_rate": 0.05,
    "max_seconds": 4.0
  },
  "largest_group_rows": 1,
  "limitations": [
    "One of 83 upstream training shards was locally available.",
    "This is an interpretable sanity baseline, not the submitted neural model.",
    "The threshold was selected on this validation split and must not be tuned on test."
  ],
  "manifest": "data/processed/partial-iid-splits.jsonl",
  "manifest_sha256": "cae0765d18467186c0e890c35fd94c0fcc70ae3483b5a1ce31814dd8ffa37b50",
  "model": {
    "bytes": 1240,
    "path": "artifacts/partial-baseline/model.json",
    "sha256": "8a2c230027f07c99241392bc031d835586e3d5ce4e90ffbb22871ceeddcf6ff7"
  },
  "scope": "audited local training shard; official test remains sealed",
  "source_root": ".",
  "status": "development_only",
  "threshold_selection": {
    "accuracy": 0.5644171779141104,
    "average_precision": 0.7310146473284146,
    "balanced_accuracy": 0.5565888554216868,
    "brier_score": 0.20382910064244883,
    "count": 326,
    "expected_calibration_error": 0.08678660959363686,
    "f1": 0.22826086956521738,
    "false_negative_rate": 0.86875,
    "false_positive_rate": 0.018072289156626505,
    "fn": 139,
    "fp": 3,
    "log_loss": 0.591045865107666,
    "macro_f1": 0.462421033073207,
    "max_false_positive_rate": 0.02,
    "negative_count": 166,
    "positive_count": 160,
    "precision": 0.875,
    "recall": 0.13125,
    "roc_auc": 0.752183734939759,
    "specificity": 0.9819277108433735,
    "threshold": 0.7638751511278986,
    "tn": 163,
    "tp": 21
  },
  "train_examples": 2939,
  "train_metrics_at_validation_threshold": {
    "accuracy": 0.5440626063286832,
    "average_precision": 0.6528796660335541,
    "balanced_accuracy": 0.5349069558634532,
    "brier_score": 0.21565992251834837,
    "count": 2939,
    "expected_calibration_error": 0.05208070343574749,
    "f1": 0.18787878787878787,
    "false_negative_rate": 0.89221140472879,
    "false_positive_rate": 0.0379746835443038,
    "fn": 1283,
    "fp": 57,
    "log_loss": 0.6229624981492231,
    "macro_f1": 0.4354720334852785,
    "negative_count": 1501,
    "positive_count": 1438,
    "precision": 0.7311320754716981,
    "recall": 0.10778859527121001,
    "roc_auc": 0.7162675045565358,
    "specificity": 0.9620253164556962,
    "threshold": 0.7638751511278986,
    "tn": 1444,
    "tp": 155
  },
  "train_split": "train",
  "validation_examples": 326,
  "validation_metrics": {
    "accuracy": 0.5644171779141104,
    "average_precision": 0.7310146473284146,
    "balanced_accuracy": 0.5565888554216868,
    "brier_score": 0.20382910064244883,
    "count": 326,
    "expected_calibration_error": 0.08678660959363686,
    "f1": 0.22826086956521738,
    "false_negative_rate": 0.86875,
    "false_positive_rate": 0.018072289156626505,
    "fn": 139,
    "fp": 3,
    "log_loss": 0.591045865107666,
    "macro_f1": 0.462421033073207,
    "negative_count": 166,
    "positive_count": 160,
    "precision": 0.875,
    "recall": 0.13125,
    "roc_auc": 0.752183734939759,
    "specificity": 0.9819277108433735,
    "threshold": 0.7638751511278986,
    "tn": 163,
    "tp": 21
  },
  "validation_reliability": [
    {
      "count": 2,
      "lower": 0.0,
      "mean_probability": 0.07256123833644303,
      "observed_rate": 0,
      "upper": 0.1
    },
    {
      "count": 15,
      "lower": 0.1,
      "mean_probability": 0.1637974159539599,
      "observed_rate": 0,
      "upper": 0.2
    },
    {
      "count": 33,
      "lower": 0.2,
      "mean_probability": 0.24781361783797937,
      "observed_rate": 0.15151515151515152,
      "upper": 0.3
    },
    {
      "count": 60,
      "lower": 0.3,
      "mean_probability": 0.3501448253606763,
      "observed_rate": 0.3,
      "upper": 0.4
    },
    {
      "count": 72,
      "lower": 0.4,
      "mean_probability": 0.44860372731312476,
      "observed_rate": 0.5,
      "upper": 0.5
    },
    {
      "count": 62,
      "lower": 0.5,
      "mean_probability": 0.5475329234533823,
      "observed_rate": 0.7096774193548387,
      "upper": 0.6
    },
    {
      "count": 42,
      "lower": 0.6,
      "mean_probability": 0.6424729141775747,
      "observed_rate": 0.5714285714285714,
      "upper": 0.7
    },
    {
      "count": 25,
      "lower": 0.7,
      "mean_probability": 0.7485065215622557,
      "observed_rate": 0.72,
      "upper": 0.8
    },
    {
      "count": 10,
      "lower": 0.8,
      "mean_probability": 0.8286639570658632,
      "observed_rate": 1,
      "upper": 0.9
    },
    {
      "count": 5,
      "lower": 0.9,
      "mean_probability": 0.9319128283903566,
      "observed_rate": 1,
      "upper": 1.0
    }
  ],
  "validation_split": "validation"
}