PsyPlay: Personality-Infused Role-Playing Conversational Agents

2025-02-06

AI Summary

Existing role-playing conversational agents (RPCAs) primarily emulate superficial stylistic and contextual features, neglecting deep personality trait modeling. Method: This paper introduces the first OCEAN-based (Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism) deep personality modeling framework for RPCAs, integrating personality-aware prompt engineering, a novel role-consistency constraint mechanism, and personality-guided generation strategies within a multi-LLM collaborative dialogue system. Contribution/Results: We propose a principled personality consistency modeling mechanism and release PsyPlay-Bench—the first dedicated benchmark for personality-consistent dialogue evaluation, comprising 4,745 high-quality utterance-level annotated dialogues. Experiments show an 80.31% personality accuracy rate on GPT-3.5-turbo; value-aligned models further enhance positive personality expression. Our work transcends surface-level imitation, establishing a new paradigm for trustworthy, personality-driven conversational agents.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Natural Language Processing: Conversational AI/Dialog Systems
• Multiagent Systems: Modeling other Agents

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

The current research on Role-Playing Conversational Agents (RPCAs) with Large Language Models (LLMs) primarily focuses on imitating specific speaking styles and utilizing character backgrounds, neglecting the depiction of deeper personality traits.~In this study, we introduce personality-infused role-playing for LLM agents, which encourages agents to accurately portray their designated personality traits during dialogues. We then propose PsyPlay, a dialogue generation framework that facilitates the expression of rich personalities among multiple LLM agents. Specifically, PsyPlay enables agents to assume roles with distinct personality traits and engage in discussions centered around specific topics, consistently exhibiting their designated personality traits throughout the interactions. Validation on generated dialogue data demonstrates that PsyPlay can accurately portray the intended personality traits, achieving an overall success rate of 80.31% on GPT-3.5. Notably, we observe that LLMs aligned with positive values are more successful in portraying positive personality roles compared to negative ones. Moreover, we construct a dialogue corpus for personality-infused role-playing, called PsyPlay-Bench. The corpus, which consists of 4745 instances of correctly portrayed dialogues using PsyPlay, aims to further facilitate research in personalized role-playing and dialogue personality detection.

Problem

• Enhance personality portrayal in LLM agents
• Develop PsyPlay for rich personality expression
• Create PsyPlay-Bench for dialogue personality research

Innovation

• Personality-infused role-playing agents
• Dialogue generation framework
• PsyPlay-Bench dialogue corpus