AI RESEARCH
CRPO: Character-centric Group Relative Policy Optimization for Role-aware Reasoning in Role-playing Agents
arXiv CS.CL
•
ArXi:2605.25511v1 Announce Type: new Recent advancements in Reinforcement Learning (RL), particularly Group Relative Policy Optimization (GRPO), have significantly enhanced the reasoning capabilities of Large Language Models. However, applying these problem-centric optimization methods to role-playing agents often leads to a loss of character fidelity and style collapse, as they prioritize context-specific utility over persona alignment.