CogDual: Enhancing Dual Cognition of LLMs via Reinforcement Learning with Implicit Rule-Based Rewards

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Liu, Cheng, Lu, Yifei, Ye, Fanghua, Li, Jian, Chen, Xingyu, Ren, Feiliang, Tu, Zhaopeng, Li, Xiaolong
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913955789668352
author Liu, Cheng
Lu, Yifei
Ye, Fanghua
Li, Jian
Chen, Xingyu
Ren, Feiliang
Tu, Zhaopeng
Li, Xiaolong
author_facet Liu, Cheng
Lu, Yifei
Ye, Fanghua
Li, Jian
Chen, Xingyu
Ren, Feiliang
Tu, Zhaopeng
Li, Xiaolong
contents Role-Playing Language Agents (RPLAs) have emerged as a significant application direction for Large Language Models (LLMs). Existing approaches typically rely on prompt engineering or supervised fine-tuning to enable models to imitate character behaviors in specific scenarios, but often neglect the underlying \emph{cognitive} mechanisms driving these behaviors. Inspired by cognitive psychology, we introduce \textbf{CogDual}, a novel RPLA adopting a \textit{cognize-then-respond } reasoning paradigm. By jointly modeling external situational awareness and internal self-awareness, CogDual generates responses with improved character consistency and contextual alignment. To further optimize the performance, we employ reinforcement learning with two general-purpose reward schemes designed for open-domain text generation. Extensive experiments on the CoSER benchmark, as well as Cross-MR and LifeChoice, demonstrate that CogDual consistently outperforms existing baselines and generalizes effectively across diverse role-playing tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2507_17147
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CogDual: Enhancing Dual Cognition of LLMs via Reinforcement Learning with Implicit Rule-Based Rewards
Liu, Cheng
Lu, Yifei
Ye, Fanghua
Li, Jian
Chen, Xingyu
Ren, Feiliang
Tu, Zhaopeng
Li, Xiaolong
Computation and Language
Role-Playing Language Agents (RPLAs) have emerged as a significant application direction for Large Language Models (LLMs). Existing approaches typically rely on prompt engineering or supervised fine-tuning to enable models to imitate character behaviors in specific scenarios, but often neglect the underlying \emph{cognitive} mechanisms driving these behaviors. Inspired by cognitive psychology, we introduce \textbf{CogDual}, a novel RPLA adopting a \textit{cognize-then-respond } reasoning paradigm. By jointly modeling external situational awareness and internal self-awareness, CogDual generates responses with improved character consistency and contextual alignment. To further optimize the performance, we employ reinforcement learning with two general-purpose reward schemes designed for open-domain text generation. Extensive experiments on the CoSER benchmark, as well as Cross-MR and LifeChoice, demonstrate that CogDual consistently outperforms existing baselines and generalizes effectively across diverse role-playing tasks.
title CogDual: Enhancing Dual Cognition of LLMs via Reinforcement Learning with Implicit Rule-Based Rewards
topic Computation and Language
url https://arxiv.org/abs/2507.17147