← 研踪 PaperTrailarXiv:2606.27025 · v2
Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
首次提交 2026-06-25 21:34
版本更新 2026-11-12 15:08
雷达观测 2026-11-13 14:15
摘要
Building general-purpose role-playing agents that faithfully portray any character from a natural-language profile remains challenging. Explicit reasoning can help models deliberate before responding, but generic reasoning does not necessarily preserve character fidelity. We therefore propose Psy-CoT, a psychology-grounded chain-of-thought framework that decomposes pre-response reasoning into three role-specific steps: Interaction Perception, Psychological Empathy, and Logical Construction, so that the model thinks dynamically from the profile rather than merely mimicking surface patterns. While structured reasoning provides a foundation, it alone is insufficient; reinforcement learning is essential to further align the model with character fidelity. However, we observe that under LLM-based reward models, both generic phrases that hack the reward model and genuinely role-specific phrases receive identical gradient signals. This hacking accumulates over training, misleading the model into treating both as equally optimal choices. To address this, we propose Role-Aware Policy Optimization (RAPO), which uses profile-induced entropy differences to weight gradients asymmetrically, amplifying role-specific tokens under positive advantage while attenuating them under negative advantage. Experiments on CoSER, CharacterBench, and CharacterEval demonstrate that Psy-CoT outperforms existing role-playing CoT methods, and RAPO consistently surpasses GRPO across multiple model scales.
作者备注：NeurIPS2026
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-12 15:08 · 最新观测版本
Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
Building general-purpose role-playing agents that faithfully portray any character from a natural-language profile remains challenging. Explicit reasoning can help models deliberate before responding, but generic reasoning does not necessarily preserve character fidelity. We therefore propose Psy-CoT, a psychology-grounded chain-of-thought framework that decomposes pre-response reasoning into three role-specific steps: Interaction Perception, Psychological Empathy, and Logical Construction, so that the model thinks dynamically from the profile rather than merely mimicking surface patterns. While structured reasoning provides a foundation, it alone is insufficient; reinforcement learning is essential to further align the model with character fidelity. However, we observe that under LLM-based reward models, both generic phrases that hack the reward model and genuinely role-specific phrases receive identical gradient signals. This hacking accumulates over training, misleading the model into treating both as equally optimal choices. To address this, we propose Role-Aware Policy Optimization (RAPO), which uses profile-induced entropy differences to weight gradients asymmetrically, amplifying role-specific tokens under positive advantage while attenuating them under negative advantage. Experiments on CoSER, CharacterBench, and CharacterEval demonstrate that Psy-CoT outperforms existing role-playing CoT methods, and RAPO consistently surpasses GRPO across multiple model scales.
查看此版本解读 →查看此版本原文 ↗