HER: Human-like Reasoning and Reinforcement Learning for LLM Role-playing

TL;DR

HER introduces a cognitive simulation framework for LLM role-playing, utilizing dual-layer thinking, curated reasoning data, and human-aligned rewards to enhance persona consistency and reasoning capabilities.

Contribution

The paper presents HER, a novel approach combining reasoning-augmented data and dual-layer thinking for improved LLM role-playing with human-like cognition.

Findings

HER models outperform baseline Qwen3-32B significantly.

30.26 point improvement on CoSER benchmark.

14.97% gain on Minimax Role-Play Bench.

Abstract

LLM role-playing, i.e., using LLMs to simulate specific personas, has emerged as a key capability in various applications, such as companionship, content creation and digital games. While current models effectively capture character tones and knowledge, simulating the inner thoughts behind their behaviors remains a challenge. Towards cognitive simulation in LLM role-play, previous efforts mainly suffer from two deficiencies: lacking data with high-quality reasoning traces, and lacking reliable reward signals aligned with human preferences. In this paper, we propose HER, a unified framework for cognitive-level persona simulation. HER introduces dual-layer thinking, which distinguishes characters' first-person thinking from LLMs' third-person thinking. To bridge these gaps, we curate reasoning-augmented role-playing data via reverse engineering, and construct human-aligned principles and reward models. Leveraging these resources, we train HER models based on Qwen3-32B via supervised and reinforcement learning. Extensive experiments validate the effectiveness of our approach. Notably, our models significantly outperform the Qwen3-32B baseline, achieving a 30.26 improvement on the CoSER benchmark and a 14.97% gain on the Minimax Role-Play Bench. Our datasets, principles, and models are released to facilitate future research.