Why LLMs Hallucinate, and How to Get (Evidential) Closure: Perceptual, Intensional, and Extensional Learning for Faithful Natural Language Generation
Submission Type: Regular Long Paper
Submission Track: Theme Track: Large Language Models and the Future of NLP
Keywords: Truth, Semantics, Meaning, Hallucination, Identification, Natural Language Generation
TL;DR: LLMs hallucinate because their output is not constrained to be synonymous with claims for which they have evidence. We show conceptually how to constrain LLM output to be synonymous with claims for which the LLM has evidence.
Abstract: We show that LLMs hallucinate because their output is not constrained to be synonymous with claims for which they have evidence: a condition that we call evidential closure. Information about the truth or falsity of sentences is not statistically identified in the standard neural language generation setup, and so cannot be conditioned on to generate new strings. We then show how to constrain LLMs to produce output that satisfies evidential closure. A multimodal LLM must learn about the external world (perceptual learning); it must learn a mapping from strings to states of the world (extensional learning); and, to achieve fluency when generalizing beyond a body of evidence, it must learn mappings from strings to their synonyms (intensional learning). The output of a unimodal LLM must be synonymous with strings in a validated evidence set. Finally, we present a heuristic procedure, Learn-Babble-Prune, that yields faithful output from an LLM by rejecting output that is not synonymous with claims for which the LLM has evidence.
Submission Number: 2298
Loading