Dataset Files

Dataset Format

Corpus rows contain a document ID, metadata, and sentence objects with stable sentence IDs.

{"doc_id":"doc_000001","sentences":[{"sent_id":1,"text":"..."}]}

Test rows contain only the claim input.

{"id":"test_000001","claim":"..."}

Submission Format

Submit exactly one JSONL prediction for every row in test.jsonl.

{"id":"test_000001","label":"NOT_ENOUGH_INFO","evidence":[]}

{"doc_id":"doc_000001","sent_id":1}

Allowed labels are SUPPORTED, REFUTED, and NOT_ENOUGH_INFO.