Multiple LLM Agents Debate for Equitable Cultural Alignment

TL;DR

This paper introduces a multi-agent debate framework for LLMs to enhance cultural adaptability, demonstrating improved accuracy and fairness across diverse cultural contexts, even with smaller models.

Contribution

The paper proposes a novel multi-agent debate approach for LLMs to better adapt to cultural norms, outperforming single-LLM methods in accuracy and fairness.

Findings

Debate improves accuracy and cultural fairness.

Small LLMs achieve large-model performance through debate.

Approach validated on 7 open-weight LLMs and 75 countries.

Abstract

Large Language Models (LLMs) need to adapt their predictions to diverse cultural contexts to benefit diverse communities across the world. While previous efforts have focused on single-LLM, single-turn approaches, we propose to exploit the complementary strengths of multiple LLMs to promote cultural adaptability. We introduce a Multi-Agent Debate framework, where two LLM-based agents debate over a cultural scenario and collaboratively reach a final decision. We propose two variants: one where either LLM agents exclusively debate and another where they dynamically choose between self-reflection and debate during their turns. We evaluate these approaches on 7 open-weight LLMs (and 21 LLM combinations) using the NormAd-ETI benchmark for social etiquette norms in 75 countries. Experiments show that debate improves both overall accuracy and cultural group parity over single-LLM baselines. Notably, multi-agent debate enables relatively small LLMs (7-9B) to achieve accuracies comparable to that of a much larger model (27B parameters).