Beyond Dialogue: A Profile-Dialogue Alignment Framework Towards General Role-Playing Language Model
Yeyong Yu, Runsheng Yu, Haojie Wei, Zhanqiu Zhang, Quan Qian
Abstract
The rapid advancement of large language models (LLMs) has revolutionized role-playing, enabling the development of general role-playing models. However, current role-playing training has two significant issues: (I) Using a predefined role profile to prompt dialogue training for specific scenarios usually leads to biases and even conflicts between the dialogue and the profile, resulting in training biases. (II) Models learn to imitate the role based solely on the profile, neglecting profile-dialogue alignment at the sentence level. To overcome the aforementioned hurdles, we propose a novel framework BEYOND DIALOGUE, which introduces "beyond dialogue" tasks to align dialogue with profile traits for each scenario, eliminating biases during training. Furthermore, the framework achieves a sentence-level fine-grained alignment between profile and dialogue through an innovative prompting mechanism that generates reasoning data for training. Moreover, the aforementioned methods are fully automated and low-cost. Experimental results demonstrate our model excels in adhering to role profiles, outperforming most proprietary general and specialized role-playing baselines. The code and data are provided in https:// github.com/yuyouyu32/BeyondDialogue . 1. Split the novel by tokens & Filter chunks by roles frequency "Look at this," said Ron … "Light?" said Harry…Snape had already taken Harry's invisibility … "It's a trap," Ron said suddenly… "No," said Harry, trying to sound confident. "I think we'll be all right." 2. Extract scenes & Evaluate chunks using role expressiveness LLMs chunk-1: … chunk-2: … chunk-3: … chunk-1 chunk-2 scene-1: Quirrell … scene-2: Tension … scene: Harry struggles …. ① Discard non-single scene chunk chunk-3 chunk-2 ② Keep chunks with role profile reflection Score: 9.0. Score: 3.0. Harry's speech is plain… Harry is courageous in exploration … 5. Align profile and dialogue & Generate derivative beyond dialogue data Personality Analyze the MBTI personality reflected by the characters in the dialogue.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext 73717714-98b5-4d77-a1f1-750438cb78bbCited by top-tier papers5
- Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit ProfilesKuang Wang, Xianfei Li, Shenghao Yang, Li Zhou et al.ACL 2025 · 24 citations
- OmniCharacter: Towards Immersive Role-Playing Agents with Seamless Speech-Language Personality InteractionHaonan Zhang, Run Luo, Xiong Liu, Yuchuan Wu et al.ACL 2025 · 10 citations
- AgentMental: An Interactive Multi-Agent Framework for Explainable and Adaptive Mental Health AssessmentJinpeng Hu, Ao Wang, Qianqian Xie, Zhuo Li et al.AAAI 2026 · 5 citations
- Enhancing Persona Following at Decoding Time via Dynamic Importance Estimation for Role-Playing AgentsYuxin Liu, Mingye Zhu, Siyuan Liu, Bo Hu et al.ICLR 2026 · 2 citations
- Know Thyself, Know Thy User: Intrinsic Dual-Perspective Reasoning for Role-Playing LLMsHaotong Sun, Jianye Xie, Bocheng Xu, Yinghui JiangICML 2026
Builds on7
- Prometheus: Inducing Fine-Grained Evaluation Capability in Language ModelsSeungone Kim, Jamin Shin, Yejin Choi, Joel Jang et al.ICLR 2024 · 468 citations
- Adapting Large Language Models via Reading ComprehensionDaixuan Cheng, Shaohan Huang, Furu WeiICLR 2024 · 146 citations
- Character-LLM: A Trainable Agent for Role-PlayingYunfan Shao, Linyang Li, Junqi Dai, Xipeng QiuEMNLP 2023 · 97 citations
- NaturalConv: A Chinese Dialogue Dataset Towards Multi-turn Topic-driven ConversationXiaoyang Wang, Chen Li, Jianqiao Zhao, Dong YuAAAI 2021 · 54 citations
- Large Language Models are Superpositions of All Characters: Attaining Arbitrary Role-play via Self-AlignmentKeming Lu, Bowen Yu, Chang Zhou, Jingren ZhouACL 2024 · 16 citations
Related papers
- Neeko: Leveraging Dynamic LoRA for Efficient Multi-Character Role-Playing AgentXiaoyan Yu, Tongxu Luo, Yifan Wei, Fangyu Lei et al.EMNLP 2024 · 4 citations
- R4: Nested Reasoning-Retrieval for Reward Modeling in Role-Playing AgentsRenzhi Wang, Chongqiang Wei, Zhisheng Wang, Piji LiICLR 2026
- R-CHAR: A Metacognition-Driven Framework for Role-Playing in Large Language ModelsHaiming Qin, Jiwei Zhang, Wei Zhang, Kezhong Lu et al.EMNLP 2025
- BIG5-CHAT: Shaping LLM Personalities Through Training on Human-Grounded DataWenkai Li, Jiarui Liu, Andy Liu, Xuhui Zhou et al.ACL 2025
- Anchoring-Guidance Fine-Tuning (AnGFT): Elevating Professional Response Quality in Role-Playing Conversational AgentsQibin Li, Zhen Xu, Shengyuan Bai, Nianmin Yao et al.EMNLP 2025
