Debating Truth: Debate-driven Claim Verification with Multiple Large Language Model Agents

TL;DR

DebateCV introduces a multi-agent debate framework for claim verification, leveraging synthetic data to improve adjudication accuracy and justification quality over existing single-agent methods.

Contribution

The paper presents DebateCV, a novel debate-driven verification framework with synthetic training for better multi-agent adjudication, surpassing state-of-the-art approaches.

Findings

DebateCV outperforms non-debate methods in accuracy.

Synthetic data improves debate adjudication.

Debate-based approach enhances justification quality.

Abstract

State-of-the-art single-agent claim verification methods struggle with complex claims that require nuanced analysis of multifaceted evidence. Inspired by real-world professional fact-checkers, we propose \textbf{DebateCV}, the first debate-driven claim verification framework powered by multiple LLM agents. In DebateCV, two \textit{Debaters} argue opposing stances to surface subtle errors in single-agent assessments. A decisive \textit{Moderator} is then required to weigh the evidential strength of conflicting arguments to deliver an accurate verdict. Yet, zero-shot Moderators are biased toward neutral judgments, and no datasets exist for training them. To bridge this gap, we propose \textbf{Debate-SFT}, a post-training framework that leverages synthetic data to enhance agents' ability to effectively adjudicate debates for claim verification. Results show that our methods surpass state-of-the-art non-debate approaches in both accuracy (across various evidence conditions) and justification quality.