SHARP: Unlocking Interactive Hallucination via Stance Transfer in Role-Playing LLMs

TL;DR

This paper introduces SHARP, a benchmark and paradigm for analyzing interactive hallucinations in role-playing LLMs through stance transfer, revealing limitations of current methods and emphasizing the importance of explicit social interaction modeling.

Contribution

The paper presents a novel stance transfer-based paradigm and benchmark, SHARP, for uncovering interactive hallucination patterns in role-playing LLMs, addressing a gap in social interaction research.

Findings

SHARP effectively uncovers interactive hallucination patterns.

Post-training methods tend to obscure knowledge, leading to monotonous behaviors.

The paradigm challenges existing hallucination mitigation approaches.

Abstract

The advanced role-playing capabilities of Large Language Models (LLMs) have enabled rich interactive scenarios, yet existing research in social interactions neglects hallucination while struggling with poor generalizability and implicit character fidelity judgments. To bridge this gap, motivated by human behaviour, we introduce a generalizable and explicit paradigm for uncovering interactive patterns of LLMs across diverse worldviews. Specifically, we first define interactive hallucination through stance transfer, then construct SHARP, a benchmark built by extracting relations from commonsense knowledge graphs and utilizing LLMs' inherent hallucination properties to simulate multi-role interactions. Extensive experiments confirm our paradigm's effectiveness and stability, examine the factors that influence these metrics, and challenge conventional hallucination mitigation solutions. More broadly, our work reveals a fundamental limitation in popular post-training methods for role-playing LLMs: the tendency to obscure knowledge beneath style, resulting in monotonous yet human-like behaviors - interactive hallucination.