Elamur Research⌕
☰
home / papers / 2026 / AdaMARP: An Adaptive Multi-Agent Interaction Framework for General Immersive Role-Playing

AdaMARP: An Adaptive Multi-Agent Interaction Framework for General Immersive Role-Playing

Paper · 2026 · Zhenhua Xu, Dongsheng Chen, Shuo Wang et al.

Summary

LLM role-playing systems often feel flat. They mostly generate dialogue, treat the surroundings as static backdrop, and assume one fixed scene with a fixed cast. AdaMARP tackles both problems. Each turn interleaves four marked elements: [Thought], (Action), <Environment> and plain Speech, so world states and events become part of the story loop. A role-play episode is split among three agents:

  • an Actor Model that voices every non-user character,
  • a User Model, which is an LLM by default but can be replaced by a real person,
  • a Scene Manager that picks discrete control actions and justifies each one.

To train these agents, the authors build AdaRPSet for actors and AdaSMSet for orchestration decisions. They also release AdaptiveBench, which simulates whole dialogue trajectories and scores them with an LLM judge. Fine-tuned open models of modest size match or beat several commercial systems in this evaluation. This matters for building open, adaptive multi-character storytelling tools that do not rely on proprietary APIs.

Key points

  • Scene Manager actions: init_scene (always issued first), pick_speaker, switch_scene, add_role and end, each with a rationale. Character profiles cover seven dimensions, from identity and appearance to relationships.
  • AdaRPSet has two parts:
    • AdaRPSet-Extracted, drawn from narrative books, where an LLM helps pull out scenes and synthesize character profiles;
    • AdaRPSet-Synthesis, LLM-generated across 20 topics, with scene switches and role additions.
  • Actor Model results: an 8B actor trained on AdaRPSet beat GPT-4o-mini, Gemini-2.5-Pro and Doubao-1.5-Pro-Character on AdaptiveBench.
  • Scene Manager results: Qwen2.5-14B-Instruct trained on AdaSMSet scored 8.37 overall, up from 7.63 for the base model. This beat Claude Sonnet 4.5 (8.17) and came close to QwQ-32B (8.41).
  • Ablation (judge: GPT-5-Chat): Qwen2.5-7B-Instruct improved from 8.37 to 8.49 with Extracted data and to 8.72 with both subsets. Llama-3.1-8B rose from 7.70 to 8.89. Experiments were limited to the Llama 3.1 and Qwen 2.5 families.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🏢TencentShenzhen · China134 works · 19,743 citations🎓Zhejiang UniversityHangzhou · China172 works · 12,988 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksLiterature and Literary Theory · 519 works

Cite

Cite as Zhenhua Xu, Dongsheng Chen, Shuo Wang, Jian Li, Chengjie Wang, Meng Han and Yabiao Wang. 2026. AdaMARP: An Adaptive Multi-Agent Interaction Framework for General Immersive Role-Playing. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.1563

@inproceedings{xu2026adamarp,
  title = {{AdaMARP: An Adaptive Multi-Agent Interaction Framework for General Immersive Role-Playing}},
  author = {Zhenhua Xu, Dongsheng Chen, Shuo Wang, Jian Li, Chengjie Wang, Meng Han and Yabiao Wang},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1563},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1563}
}

References

43 references, 17 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Jaewoo Ahn, Taehyun Lee, Junyoung Lim, Jin-Hwa Kim, Sangdoo Yun, Hwaran Lee, and Gunhee Kim. 2024. TimeChara: Evaluating Point-in-Time Character Hallucination of Role-Playing Large Language Models. In Findings of the Association for Computational Linguistics: ACL 2024, pages 3291–3325. Association f… find
  2. Anthropic. 2024. Claude 4 model family. https://ww w.anthropic.com. Claude Sonnet 4.5. find
  3. Faeze Brahman, Meng Huang, Oyvind Tafjord, Chao Zhao, Mrinmaya Sachan, and Snigdha Chaturvedi. 2021. “Let Your Characters Tell Their Story”: A Dataset for Character-Centric Narrative Understanding. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 1734–1752. Association… find
  4. ByteDance. 2024. Doubao large language models. https://www.volcengine.com. Doubao-1.5-Pro-Character. ↗
  5. Jiangjie Chen, Xintao Wang, Rui Xu, Siyu Yuan, Yikai Zhang, Wei Shi, Jian Xie, Shuang Li, Ruihan Yang, Tinghui Zhu, Aili Chen, Nianqi Li, Lida Chen, Caiyu Hu, Siye Wu, Scott Ren, Ziquan Fu, and Yanghua Xiao. 2024. From Persona to Personalization: {A} Survey on Role-Playing Language Agents. 2024.
  6. Nuo Chen, Yan Wang, Haiyun Jiang, Deng Cai, Yuhan Li, Ziyang Chen, Longyue Wang, and Jia Li. 2023. Large Language Models Meet Harry Potter: A Dataset for Aligning Dialogue Agents with Characters. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 8506–8520. Association f…
  7. Gheorghe Comanici and 1 others. 2025. Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities. Preprint, arXiv:2507.06261.
  8. Zhenpeng Gao, Xiaofen Xing, and Xiangmin Xu. 2025. TailorRPA: A Retrieval-Based Framework for Eliciting Personalized and Coherent Role-Playing Agents in General Domain. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 5381–5412. Association for Computational Linguistic… find
  9. Aaron Grattafiori and 1 others. 2024. The llama 3 herd of models. Preprint, arXiv:2407.21783.
  10. Kai He, Yucheng Huang, Wenqing Wang, Delong Ran, Dongming Sheng, Junxuan Huang, Qika Lin, Jiaxing Xu, Wenqiang Liu, and Mengling Feng. 2025. Crab: A Novel Configurable Role-Playing LLM with Assessing Benchmark. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistic… find
  11. Ke Ji, Yixin Lian, Linxu Li, Jingsheng Gao, Weiyuan Li, and Bin Dai. 2025. Enhancing Persona Consistency for LLMs’ Role-Playing using Persona-Aware Contrastive Learning. In Findings of the Association for Computational Linguistics: ACL 2025, pages 26221– 26238. Association for Computational Linguist… find
  12. Nathan Lambert, Jacob Morrison, Valentina Pyatkin, Shengyi Huang, Hamish Ivison, Faeze Brahman, Lester James V. Miranda, Alisa Liu, Nouha Dziri, Shane Lyu, Yuling Gu, Saumya Malik, Victoria Graf, Jena D. Hwang, Jiangjiang Yang, Ronan Le Bras, Oyvind Tafjord, Chris Wilhelm, Luca Soldaini, and 4 other… — Tulu 3: Pushing Frontiers in Open Language Model Post-Training
  13. Cheng Li, Ziang Leng, Chenxi Yan, Junyi Shen, Hao Wang, Weishi Mi, Yaying Fei, Xiaoyang Feng, Song Yan, HaoSheng Wang, Linkang Zhan, Yaokai Jia, Pingyu Wu, and Haozhen Sun. 2023. ChatHaruhi: Reviving Anime Character in Reality via Large Language Model.
  14. Jian Li, Dongsheng Chen, Zhenhua Xu, Yizhang Jin, Jiafu Wu, Chengjie Wang, Xiaotong Yuan, and Yabiao Wang. 2026. Improving search agent with one line of code. Preprint, arXiv:2603.10069. — [arXiv:2603.10069](https://arxiv.org/abs/2603.10069) find
  15. Cheng Liu, Yifei Lu, Fanghua Ye, Jian Li, Xingyu Chen, Feiliang Ren, Zhaopeng Tu, and Xiaolong Li. 2025. CogDual: Enhancing Dual Cognition of LLMs via Reinforcement Learning with Implicit Rule-Based Rewards. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, p… find
  16. Junru Lu, Jiazheng Li, Guodong Shen, Lin Gui, Siyu An, Yulan He, Di Yin, and Xing Sun. 2025. RoleMRC: A Fine-Grained Composite Benchmark for Role-Playing and Instruction-Following. In Findings of the Association for Computational Linguistics: ACL 2025, pages 21008–21030. Association for Computationa… find
  17. Keming Lu, Bowen Yu, Chang Zhou, and Jingren Zhou. 2024. Large Language Models are Superpositions of All Characters: Attaining Arbitrary Role-play via Self-Alignment. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7828–7840.…
  18. OpenAI. 2025. Gpt-5 system card. https://cdn.open ai.com/gpt-5-system-card.pdf. Version dated August 13, 2025. ↗
  19. OpenAI, Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, Red Avila, Igor Babuschkin, Suchir Balaji, Valerie Balcom, Paul Baltescu, Haiming Bao, Mohammad Bavarian, Jeff Belgum, and 262 others.… — GPT-4 Technical Report
  20. Haiming Qin, Jiwei Zhang, Wei Zhang, KeZhong Lu, Mingyang Zhou, Hao Liao, and Rui Mao. 2025. R-CHAR: A Metacognition-Driven Framework for Role-Playing in Large Language Models. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 26984–27002. Association f… find
  21. Qwen Team and 1 others. 2025. Qwen2.5 technical report. Preprint, arXiv:2412.15115. find
  22. Yiting Ran, Xintao Wang, Rui Xu, Xinfeng Yuan, Jiaqing Liang, Yanghua Xiao, and Deqing Yang. 2024. Capturing Minds, Not Just Words: Enhancing Role-Playing Language Models with Personality-Indicative Data. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 14566–14576. As… find
  23. Vinay Samuel, Henry Peng Zou, Yue Zhou, Shreyas Chaudhari, Ashwin Kalyan, Tanmay Rajpurohit, Ameet Deshpande, Karthik R Narasimhan, and Vishvak Murahari. 2025. PersonaGym: Evaluating Persona Agents and LLMs. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 6999–7022. A… find
  24. Murray Shanahan, Kyle McDonell, and Laria Reynolds. 2023. Role-Play with Large Language Models. Preprint, arXiv:2305.16367.
  25. Yunfan Shao, Linyang Li, Junqi Dai, and Xipeng Qiu. 2023. Character-LLM: A Trainable Agent for Role-Playing. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 13153–13187. Association for Computational Linguistics.
  26. Zhen Tan, Jun Yan, I-Hung Hsu, Rujun Han, Zifeng Wang, Long Le, Yiwen Song, Yanfei Chen, Hamid Palangi, George Lee, Anand Rajan Iyer, Tianlong Chen, Huan Liu, Chen-Yu Lee, and Tomas Pfister. 2025. In prospect and retrospect: Reflective memory management for long-term personalized dialogue agents. In…
  27. Yu-Min Tseng, Yu-Chao Huang, Teng-Yun Hsiao, Wei-Lin Chen, Chao-Wei Huang, Yu Meng, and Yun-Nung Chen. 2024. Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 16612–16631. Association for Comput…
  28. Quan Tu, Shilong Fan, Zihang Tian, Tianhao Shen, Shuo Shang, Xin Gao, and Rui Yan. 2024. CharacterEval: A Chinese Benchmark for Role-Playing Conversational Agent Evaluation. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 31239 10 p…
  29. Noah Wang, Z.y. Peng, Haoran Que, Jiaheng Liu, Wangchunshu Zhou, Yuhan Wu, Hongcheng Guo, Ruitong Gan, Zehao Ni, Jian Yang, Man Zhang, Zhaoxiang Zhang, Wanli Ouyang, Ke Xu, Wenhao Huang, Jie Fu, and Junran Peng. 2024. RoleLLM: Benchmarking, Eliciting, and Enhancing Role-Playing Abilities of Large La…
  30. Xintao Wang, Heng Wang, Yifei Zhang, Xinfeng Yuan, Rui Xu, Jen-tse Huang, Siyu Yuan, Haoran Guo, Jiangjie Chen, Shuchang Zhou, Wei Wang, and Yanghua Xiao. 2025. CoSER: Coordinating LLM-Based Persona Simulation of Established Roles. Preprint, arXiv:2502.09082. — [arXiv:2502.09082](https://arxiv.org/abs/2502.09082) find
  31. An Yang and 1 others. 2025a. Qwen3 technical report. Preprint, arXiv:2505.09388.
  32. Bohao Yang, Dong Liu, Chenghao Xiao, Kun Zhao, Chen Tang, Chao Li, Lin Yuan, Yang Guang, and Chenghua Lin. 2025b. Crafting Customisable Characters with LLMs: A Persona-Driven Role-Playing Agent Framework. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 20216–20240. As… find
  33. Shihao Yang, Zhicong Lu, Yong Yang, Bo Lv, Yang Shen, and Nayu Liu. 2025c. HyCoRA: Hyper-Contrastive Role-Adaptive Learning for Role-Playing. Preprint, arXiv:2511.08017 [cs]. — [arXiv:2511.08017](https://arxiv.org/abs/2511.08017) find
  34. Jing Ye, Rui Wang, Yuchuan Wu, Victor Ma, Feiteng Fang, Fei Huang, and Yongbin Li. 2025. CPO: Addressing Reward Ambiguity in Role-playing Dialogue via Comparative Policy Optimization. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 297–323. Association for Computation… find
  35. Xiaoyan Yu, Tongxu Luo, Yifan Wei, Fangyu Lei, Yiming Huang, Hao Peng, and Liehuang Zhu. 2024. Neeko: Leveraging Dynamic LoRA for Efficient Multi-Character Role-Playing Agent. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 12540–12557. Association fo… find
  36. Yeyong Yu, Runsheng Yu, Haojie Wei, Zhanqiu Zhang, and Quan Qian. 2025. Beyond Dialogue: A Profile-Dialogue Alignment Framework Towards General Role-Playing Language Model. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 11992…
  37. Xinfeng Yuan, Siyu Yuan, Yuhan Cui, Tianhe Lin, Xintao Wang, Rui Xu, Jiangjie Chen, and Deqing Yang. 2024. Evaluating Character Understanding of Large Language Models via Character Profiling from Fictional Works. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processi… find
  38. Pinyi Zhang, Siyu An, Lingfeng Qiao, Yifei Yu, Jingyang Chen, Jie Wang, Di Yin, Xing Sun, and Kai Zhang. 2025a. RolePlot: A Systematic Framework for Evaluating and Enhancing the Plot-Progression Capabilities of Role-Playing Agents. In Proceedings of the 63rd Annual Meeting of the Association for Com… find
  39. Taolin Zhang, Dongyang Li, Qizhou Chen, Chengyu Wang, and Xiaofeng He. 2025b. BELLE: A bi-level multi-agent reasoning framework for multi-hop question answering. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 4184–4202, Vienn…
  40. Jinfeng Zhou, Zhuang Chen, Dazhen Wan, Bosi Wen, Yi Song, Jifan Yu, Yongkang Huang, Pei Ke, Guanqun Bi, Libiao Peng, JiaMing Yang, Xiyao Xiao, Sahand Sabour, Xiaohan Zhang, Wenjing Hou, Yijia Zhang, Yuxiao Dong, Hongning Wang, Jie Tang, and Minlie Huang. 2024a. CharacterGLM: Customizing social chara… find
  41. Jinfeng Zhou, Zhuang Chen, Dazhen Wan, Bosi Wen, Yi Song, Jifan Yu, Yongkang Huang, Pei Ke, Guanqun Bi, Libiao Peng, JiaMing Yang, Xiyao Xiao, Sahand Sabour, Xiaohan Zhang, Wenjing Hou, Yijia Zhang, Yuxiao Dong, Hongning Wang, Jie Tang, and Minlie Huang. 2024b. CharacterGLM: Customizing Social Chara… find
  42. Jinfeng Zhou, Yongkang Huang, Bosi Wen, Guanqun Bi, Yuxuan Chen, Pei Ke, Zhuang Chen, Xiyao Xiao, Libiao Peng, Kuntian Tang, Rongsheng Zhang, Le Zhang, Tangjie Lv, Zhipeng Hu, Hongning Wang, and Minlie Huang. 2024c. CharacterBench: Benchmarking Character Customization of Large Language Models. Prepr… — [arXiv:2412.11912](https://arxiv.org/abs/2412.11912) find
  43. Yufan Zhuang, Xiaodong Yu, Jialian Wu, Ximeng Sun, Ze Wang, Jiang Liu, Yusheng Su, Jingbo Shang, Zicheng Liu, and Emad Barsoum. 2025. Self-taught agentic long context understanding. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pa… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/xu-2026-adamarp-adaptive-multi-agent-interaction/