source-evaluation

Tag

Cards List
#source-evaluation

Trust, but Don't Verify: Epistemic Blind Spots in LLM Source Evaluation

arXiv cs.LG · 2026-06-05 Cached

This paper identifies a failure mode in LLMs where they do not verify the validity of numerical statistics when synthesizing multiple sources, instead relying on the stylistic markers of analytical rigor. The authors term this 'epistemic alignment' and show that it persists across models and domains, resisting prompting-based mitigations.

0 favorites 0 likes
← Back to home

Submit Feedback