Role-Play Paradox in Large Language Models: Reasoning Performance Gains and Ethical Dilemmas

Jinman Zhao, Zifan Qian, Linbo Cao +5

Role-play in large language models (LLMs) enhances their ability to generate contextually relevant and high-quality responses by simulating diverse cognitive perspectives. However,…