Beyond Demographics: Aligning Role-playing LLM-based Agents Using Human Belief Networks

TL;DR

This paper proposes using empirically-derived human belief networks to improve the alignment of LLM-based agents with human beliefs, demonstrating that seeding agents with a single belief enhances their social simulation fidelity.

Contribution

Introducing a method to incorporate human belief networks into LLM role-playing, improving belief alignment beyond demographic-based approaches.

Findings

Seeding LLM agents with a single belief improves alignment on related topics.

Demographic information alone does not produce realistic belief alignment.

Belief network structure guides effective belief propagation in LLM agents.

Abstract

Creating human-like large language model (LLM) agents is crucial for faithful social simulation. Having LLMs role-play based on demographic information sometimes improves human likeness but often does not. This study assessed whether LLM alignment with human behavior can be improved by integrating information from empirically-derived human belief networks. Using data from a human survey, we estimated a belief network encompassing 64 topics loading on nine non-overlapping latent factors. We then seeded LLM-based agents with an opinion on one topic, and assessed the alignment of its expressed opinions on remaining test topics with corresponding human data. Role-playing based on demographic information alone did not align LLM and human opinions, but seeding the agent with a single belief greatly improved alignment for topics related in the belief network, and not for topics outside the network. These results suggest a novel path for human-LLM belief alignment in work seeking to simulate and understand patterns of belief distributions in society.