Crafting Customisable Characters with LLMs: Introducing SimsChat, a Persona-Driven Role-Playing Agent Framework

2024-06-25

AI Summary

Existing large language models (LLMs) exhibit limited role customization flexibility, weak persona consistency, poor factual accuracy, and insufficient boundary awareness in role-playing scenarios. To address these challenges, we propose SimsChat—a novel framework featuring a customizable persona modeling paradigm that supports multi-dimensional attribute injection (e.g., occupation, personality traits, skills) and social relationship expansion. We further introduce SimsConv, the first large-scale, multi-turn role-playing dialogue dataset comprising 68 distinct personas, 13,971 dialogues, and 1,360 diverse scenarios. SimsChat incorporates a fine-grained persona consistency preservation mechanism and a topic-sensitive refusal strategy to enhance safety and coherence. Evaluated on SimsConv and WikiRoleEval, SimsChat achieves improvements of 18.7% in persona consistency, 22.3% in knowledge accuracy, and 31.5% in refusal appropriateness—enabling zero-shot role customization and robust cross-scenario generalization.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Natural Language Processing: Conversational AI/Dialog Systems

Application Category

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

Abstract

Large Language Models (LLMs) demonstrate remarkable ability to comprehend instructions and generate human-like text, enabling sophisticated agent simulation beyond basic behavior replication. However, the potential for creating freely customisable characters remains underexplored. We introduce the Customisable Conversation Agent Framework, which employs LLMs to simulate real-world characters through personalised characteristic feature injection, enabling diverse character creation according to user preferences. We propose the SimsConv dataset, comprising 68 customised characters and 13,971 multi-turn role-playing dialogues across 1,360 real-world scenes. Characters are initially customised using pre-defined elements (career, aspiration, traits, skills), then expanded through personal and social profiles. Building on this, we present SimsChat, a freely customisable role-playing agent incorporating various realistic settings and topic-specified character interactions. Experimental results on both SimsConv and WikiRoleEval datasets demonstrate SimsChat's superior performance in maintaining character consistency, knowledge accuracy, and appropriate question rejection compared to existing models. Our framework provides valuable insights for developing more accurate and customisable human simulacra. Our data and code are publicly available at https://github.com/Bernard-Yang/SimsChat.

Problem

• Large Language Models
• Role-playing
• Personalized Character Creation

Innovation

• SimsChat Framework
• Customizable Dialogue Agents
• SimsConv Dataset