SOURCE-LINKED INTELLIGENCE

The Cost of Compression: A Rate-Distortion Limit on Factual Hallucination

arXiv · AI, language, vision and robotics · article · Sep 10, 2026 · UTC

Factual hallucination in closed-book question answering is often treated as a coverage problem: a model fails because the relevant fact is absent from its internal memory. This view misses a second source of error. Even when a fact has been observed, finite memory may force it to be stored only approximately. We study this effect through a simple coverage--compression model of factual recall. We consider an unstructured question-answering task with $N$ possible queries and $K$ possible answers. A learner observes $M$ training facts, compresses them into at most $B$ bits, and answers uniformly

Read original source ↗ Open in workspace

recordType

paper

region

Global