🤖 AI Summary
This paper systematically investigates critical challenges in large language model (LLM)-driven role-playing (RP), focusing on character authenticity and personalization. It identifies three core problems: weak personality consistency, difficulty in behavior alignment with role specifications, and insufficient user engagement. To address these, the paper proposes the first four-dimensional technical taxonomy—spanning data curation, model alignment, agent architecture, and evaluation—highlighting dynamic personality modeling and higher-order consistency as pivotal research directions. Methodologically, it integrates prompt engineering, supervised and reinforcement fine-tuning, multi-agent collaboration, and hybrid human-automated multidimensional evaluation. Key contributions include: (1) the first structured, comprehensive research landscape map for RP; (2) an open-sourced, authoritative RP literature repository on GitHub; and (3) a reproducible benchmark evaluation framework. Collectively, these advances provide both theoretical foundations and practical paradigms for immersive AI-character interaction.
📝 Abstract
This survey explores the burgeoning field of role-playing with language models, focusing on their development from early persona-based models to advanced character-driven simulations facilitated by Large Language Models (LLMs). Initially confined to simple persona consistency due to limited model capabilities, role-playing tasks have now expanded to embrace complex character portrayals involving character consistency, behavioral alignment, and overall attractiveness. We provide a comprehensive taxonomy of the critical components in designing these systems, including data, models and alignment, agent architecture and evaluation. This survey not only outlines the current methodologies and challenges, such as managing dynamic personal profiles and achieving high-level persona consistency but also suggests avenues for future research in improving the depth and realism of role-playing applications. The goal is to guide future research by offering a structured overview of current methodologies and identifying potential areas for improvement. Related resources and papers are available at https://github.com/nuochenpku/Awesome-Role-Play-Papers.