Data

Dataset Files

Download All

File name: dev.jsonl
question: The question to answer
id: The question ID
reasoning_type: The reasoning category (contextual_single_hop, causal_chain, bridge_multi_doc, unanswerable, temporal_resolution, comparison_with_filtering, conflict_resolution, aggregation, bridge_multi_sentence)
evidence: The document and sentence references needed to answer the question, e.g. [{"doc_id": "", "sent_id": 1}]
answers: The gold answer label(s)

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

RAG-QA test rows contain question inputs and stable IDs. The server validates submissions against the organizer test IDs.

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

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

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

Each prediction must contain exactly id, answer, and evidence fields.

{"id":"test_000001","answer":"Mara Quinn","evidence":[{"doc_id":"doc_001542","sent_id":2}]}

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