{
  "accepted_for_ptq_calibration": 512,
  "approximate_token_records": 0,
  "category_counts": {
    "A": 140,
    "B": 110,
    "C": 90,
    "D": 70,
    "E": 60,
    "F": 42
  },
  "file": "data/variant-2/calibration_samples.jsonl",
  "format": "leanstral_public_split_manifest_v1",
  "incomplete_provenance_records": 0,
  "length_bucket_counts": {
    "long": 16,
    "medium": 43,
    "short": 175,
    "tiny": 278
  },
  "mathlib_commits": [
    "5e932f97dd25535344f80f9dd8da3aab83df0fe6"
  ],
  "records": 512,
  "role": "context-enriched calibration candidate",
  "selected": false,
  "sha256": "7450f6b0c70c88c59bf0a0949cdacaaaad8b1f676775923470cce0d6fabfdbd3",
  "split": "variant-2",
  "token_counts": {
    "maximum": 20899,
    "minimum": 129,
    "total": 677384
  },
  "tokenizer_ids": [
    "mistralai/Leanstral-1.5-119B-A6B:tekken.json"
  ],
  "unique_declarations": 512,
  "unique_duplicate_fingerprints": 512,
  "unique_sample_ids": 512,
  "unique_source_files": 462
}