Correct study data and cards; separate audit results

Upload dataset

Upload dataset

Upload dataset

Correct option-count and failure-tag claims; align to card template

Remove pre-push_to_hub data files superseded by config directories

Upload dataset

Upload dataset

Upload dataset

Add the judge's first-pass verdict and per-stage tokens

Describe the data only; drop baseline results

Remove audit/ — not part of the dataset

Remove raw/ — not part of the dataset

Remove results/ — not part of the dataset

Remove JSONL splits superseded by Parquet

Remove JSONL splits superseded by Parquet

Remove JSONL splits superseded by Parquet

Remove JSONL splits superseded by Parquet

Remove JSONL splits superseded by Parquet

Remove JSONL splits superseded by Parquet

Ship split files as typed Parquet; keep raw JSONL as artifacts of record

Initial upload: corpus, splits, results, provenance

initial commit