Where Does Retrieval-Based Open-Ended Evaluation Fail? Automatic Taxonomy Induction from Long-Form Medical Answer Factuality Verification
A medical RAG study finds retrieve-then-verify factuality checks fail for structural reasons, not weak models.
1 stories, newest first · one card per story
A medical RAG study finds retrieve-then-verify factuality checks fail for structural reasons, not weak models.
Today 1.6K readers · 2.5K views · 779 reads · all time 189K views