Fine-Refine: Iterative Fine-grained Refinement for Mitigating Dialogue Hallucination

TL;DR

Fine-Refine is an iterative, fine-grained framework that decomposes dialogue responses into atomic units, verifies facts with external knowledge, and corrects errors to reduce hallucinations in large language models.

Contribution

It introduces a novel fine-grained refinement approach that improves factual accuracy in dialogue systems by verifying and correcting individual response units.

Findings

Achieves up to 7.63-point increase in factual accuracy score.

Effectively reduces hallucinations in dialogue responses.

Maintains high dialogue quality with minimal trade-offs.

Abstract

The tendency for hallucination in current large language models (LLMs) negatively impacts dialogue systems. Such hallucinations produce factually incorrect responses that may mislead users and undermine system trust. Existing refinement methods for dialogue systems typically operate at the response level, overlooking the fact that a single response may contain multiple verifiable or unverifiable facts. To address this gap, we propose Fine-Refine, a fine-grained refinement framework that decomposes responses into atomic units, verifies each unit using external knowledge, assesses fluency via perplexity, and iteratively corrects granular errors. We evaluate factuality across the HybriDialogue and OpendialKG datasets in terms of factual accuracy (fact score) and coverage (Not Enough Information Proportion), and experiments show that Fine-Refine substantially improves factuality, achieving up to a 7.63-point gain in dialogue fact score, with a small trade-off in dialogue quality.