data_4class/split_summary.json

{
  "label_order": [
    "reject",
    "low_value",
    "keep",
    "high_value"
  ],
  "mapping": {
    "X": "reject",
    "R": "reject",
    "KL": "low_value",
    "D": "low_value",
    "K": "keep",
    "KH": "high_value"
  },
  "rows": 198874,
  "train_rows": 178986,
  "test_rows": 19888,
  "source_d_counts": {
    "KL": 17014,
    "D": 137652,
    "X": 30408,
    "R": 1374,
    "K": 12082,
    "KH": 344
  },
  "label_counts": {
    "low_value": 154666,
    "reject": 31782,
    "keep": 12082,
    "high_value": 344
  },
  "train_label_counts": {
    "reject": 28604,
    "low_value": 139199,
    "keep": 10874,
    "high_value": 309
  },
  "test_label_counts": {
    "low_value": 15467,
    "reject": 3178,
    "keep": 1208,
    "high_value": 35
  },
  "test_size": 0.1,
  "seed": 20260604,
  "head_chars": 10000,
  "tail_chars": 2000,
  "weight_floor": 0.2
}