Datasets:
Tasks:
Token Classification
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
10K - 100K
Tags:
hallucination-detection
fact-verification
RAG
span-detection
token-classification
summarization
License:
Commit History
-
v2.1: Update README for normalized prompt format
Verified signature5682bcd -
v2.1: Normalized prompts to match RAGTruth format (Summarize the following text...)
Verified signaturef9a5c89 -
Fix dates: 2025 -> 2026
Verified signature8809e21 -
Add YAML metadata to dataset card
Verified signature2ebb15a -
Update dataset card for v2.0: LLM-detected spans
Verified signature29e4770 -
Update to v2.0: LLM-detected spans with RAGTruth-compatible labels
Verified signatureaadee02
-
Add full HaluEval spans dataset (38K samples, QA+Summarization+Dialogue)
Verified signature1f425eb -
initial commit
Verified signatureec340f7