---
title: "The Stackelberg Speaker — paper by Zheng et al. (2026)"
description: "LLM agents for social deduction games (SDGs) usually concentrate on deducing hidden roles and choosing a strategy. 38 references."
authors: ["Zhang Zheng", "Deheng Ye", "Peilin Zhao", "Hao Wang"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.250/"
doi: "10.18653/v1/2026.acl-long.250"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/zheng-2026-stackelberg-speaker-optimizing-persuasive-communication/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zheng-2026-stackelberg-speaker-optimizing-persuasive-communication/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/53/) / The Stackelberg Speaker: Optimizing Persuasive Communication in Social Deduction Games

# The Stackelberg Speaker: Optimizing Persuasive Communication in Social Deduction Games

**Authors:** Zhang Zheng, [Deheng Ye](https://md.research.elamur.ai/free/en/authors/deheng-ye/), [Peilin Zhao](https://md.research.elamur.ai/free/en/authors/peilin-zhao/) and [Hao Wang](https://md.research.elamur.ai/free/en/authors/hao-wang/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.250/) · [PDF](https://aclanthology.org/2026.acl-long.250.pdf) · [DOI 10.18653/v1/2026.acl-long.250](https://doi.org/10.18653/v1/2026.acl-long.250) · [OpenAlex](https://openalex.org/W4417120004)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
LLM agents for social deduction games (SDGs) usually concentrate on deducing hidden roles and choosing a strategy. They largely ignore persuasion: wording a turn so that the next speaker reacts as the agent wants. The authors model each pair of consecutive turns as a Stackelberg competition. The current speaker is the leader and picks an utterance while anticipating the distribution of replies from the follower, who is the next speaker. Training uses a self-play dataset and runs in three stages:
- A backend LLM states the player's intent as desired and undesired replies from the follower.
- A Refiner model rewrites a base utterance.
- A Measurer scores each rewrite by how much it raises the likelihood of the desired replies and lowers that of the undesired ones.

The Refiner is optimised with GRPO, so no separate critic is needed. The method was tested on Werewolf, Avalon, ONUW and the social simulation Sotopia, and it beats the baselines for both cooperative and deceptive roles. The work matters because it optimises free-form language directly instead of choosing from a fixed set of candidate actions, and it reaches strong results against human players.

## Key points
- **Reward and data:** The reward is a sum of log-probabilities of K desired follower replies minus a sum over K undesired ones. K = 3 is the best setting; larger values add noise. The self-play data comes from GPT-4o, Gemini-2.5-Flash and Claude-3.5-Haiku.
- **Ablation (50 matches per setting):** The full reward beats both single-objective variants. Positive-Only still helps substantially. Negative-Only performs about the same as the ReAct baseline.
- **Generalisation:** The trained Refiner improves results on GPT-5 and Qwen3-14B as backend LLMs without any further fine-tuning.
- **Human study:** 16 volunteers each played 5 Werewolf games alongside 6 AI agents. "Ours + LSPO" had the highest win rate (44.1%), above humans (40.0%) and LSPO (41.3%). It also received the fewest votes from humans (0.39 per game).
- **Risks and release:** The authors note that the persuasion techniques could be misused for manipulation. Code and data are published on the project page (3dagentworld.github.io/leader_follower).

*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

- [Shanghai Jiao Tong University](https://md.research.elamur.ai/free/en/universities/shanghai-jiao-tong-university/) — University, Shanghai, China · 212 works · 19,544 citations
- [Tencent](https://md.research.elamur.ai/free/en/companies/tencent/) — Company, Shenzhen, China · 163 works · 21,032 citations
- [The Hong Kong University of Science and Technology (Guangzhou)](https://md.research.elamur.ai/free/en/universities/the-hong-kong-university-of-science-and-technology-guangzhou/) — University, Guangzhou, China · 67 works · 1,765 citations

📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🎯 Decision Sciences](https://md.research.elamur.ai/free/en/shelves/decision-sciences/) · 249 works · [🎯 Management Science and Operations Research](https://md.research.elamur.ai/free/en/shelves/decision-sciences/management-science-and-operations-research/) · 134 works

## Cite

Cite as: Zhang Zheng, Deheng Ye, Peilin Zhao and Hao Wang. 2026. The Stackelberg Speaker: Optimizing Persuasive Communication in Social Deduction Games. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{zheng2026stackelberg,
  title = {{The Stackelberg Speaker: Optimizing Persuasive Communication in Social Deduction Games}},
  author = {Zhang Zheng, Deheng Ye, Peilin Zhao and Hao Wang},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.250},
  url = {https://doi.org/10.18653/v1/2026.acl-long.250}
}
```

- [← previous The Spectrum of Sentiment: Optimistic, Pessimistic, and Neutral Voices in Online Depression Discourse](https://md.research.elamur.ai/free/en/papers/tabusca-2026-spectrum-sentiment-optimistic-pessimistic-neutral/)
- [next → The Structure-Content Trade-off in Knowledge Graph Retrieval: A Diagnostic Study of Question Decomposition](https://md.research.elamur.ai/free/en/papers/six-2026-structure-content-trade-off-knowledge/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/53/)

## References

38 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Anton Bakhtin, Noam Brown, Emily Dinan, Gabriele Farina, Colin Flaherty, Daniel Fried, Andrew Goff, Jonathan Gray, Hengyuan Hu, Athul Paul Jacob, Mo jtaba Komeili, Karthik Konath, Minae Kwon, Adam Lerer, Mike Lewis, Alexander H. Miller, Sandra Mitts, Adithya Renduchintala, Stephen Roller, and 7 othe… [find](https://scholar.google.com/scholar?q=Anton+Bakhtin%2C+Noam+Brown%2C+Emily+Dinan%2C+Gabriele+Farina%2C+Colin+Flaherty%2C+Daniel+Fried%2C+Andrew+Goff%2C+Jonathan+Gray%2C+Hengyuan+Hu%2C+Athul+Paul+Jacob%2C+Mo+jtaba+Komeili%2C+Karthik+Konath%2C+Minae+Kwon%2C+Adam+Ler)
2. Qi Chai, Zhang Zheng, Junlong Ren, Deheng Ye, Zichuan Lin, and Hao Wang. 2025. Causalmace: Causality empowered multi-agents in minecraft cooperative tasks. Preprint, arXiv:2508.18797. [find](https://scholar.google.com/scholar?q=Qi+Chai%2C+Zhang+Zheng%2C+Junlong+Ren%2C+Deheng+Ye%2C+Zichuan+Lin%2C+and+Hao+Wang%2E+2025%2E+Causalmace%3A+Causality+empowered+multi%2Dagents+in+minecraft+cooperative+tasks%2E+Preprint%2C+arXiv%3A2508%2E18797%2E)
3. Honghao Fu, Junlong Ren, Qi Chai, Deheng Ye, Yujun Cai, and Hao Wang. 2025. Vistawise: Building cost-effective agent with cross-modal knowledge graph for minecraft. Preprint, arXiv:2508.18722. [find](https://scholar.google.com/scholar?q=Honghao+Fu%2C+Junlong+Ren%2C+Qi+Chai%2C+Deheng+Ye%2C+Yujun+Cai%2C+and+Hao+Wang%2E+2025%2E+Vistawise%3A+Building+cost%2Deffective+agent+with+cross%2Dmodal+knowledge+graph+for+minecraft%2E+Preprint%2C+arXiv%3A2508%2E18722%2E)
4. Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. [LoRA: Low-rank adaptation of large language models](https://md.research.elamur.ai/free/en/papers/hu-2021-lora-low-rank-adaptation-large/). In International Conference on Learning Representations.
5. Xuanfa Jin, Ziyan Wang, Yali Du, Meng Fang, Haifeng Zhang, and Jun Wang. 2024. Learning to discuss strategically: A case study on one night ultimate werewolf. In The Thirty-eighth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Xuanfa+Jin%2C+Ziyan+Wang%2C+Yali+Du%2C+Meng+Fang%2C+Haifeng+Zhang%2C+and+Jun+Wang%2E+2024%2E+Learning+to+discuss+strategically%3A+A+case+study+on+one+night+ultimate+werewolf%2E+In+The+Thirty%2Deighth+Annual+Conference+on)
6. Aobo Kong, Wentao Ma, Shiwan Zhao, Yongbin Li, Yuchuan Wu, Ke Wang, Xiaoqian Liu, Qicheng Li, Yong Qin, and Fei Huang. 2025. SDPO: Segment-level direct preference optimization for social agents. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Lon… [find](https://scholar.google.com/scholar?q=Aobo+Kong%2C+Wentao+Ma%2C+Shiwan+Zhao%2C+Yongbin+Li%2C+Yuchuan+Wu%2C+Ke+Wang%2C+Xiaoqian+Liu%2C+Qicheng+Li%2C+Yong+Qin%2C+and+Fei+Huang%2E+2025%2E+SDPO%3A+Segment%2Dlevel+direct+preference+optimization+for+social+agents%2E+In+Pr)
7. Bolin Lai, Hongxin Zhang, Miao Liu, Aryan Pariani, Fiona Ryan, Wenqi Jia, Shirley Anugrah Hayati, James Rehg, and Diyi Yang. 2023. Werewolf among us: Multimodal resources for modeling persuasion behaviors in social deduction games. In Findings of the Association for Computational Linguistics: ACL 20… [find](https://scholar.google.com/scholar?q=Bolin+Lai%2C+Hongxin+Zhang%2C+Miao+Liu%2C+Aryan+Pariani%2C+Fiona+Ryan%2C+Wenqi+Jia%2C+Shirley+Anugrah+Hayati%2C+James+Rehg%2C+and+Diyi+Yang%2E+2023%2E+Werewolf+among+us%3A+Multimodal+resources+for+modeling+persuasion+behav)
8. Yihuai Lan, Zhiqiang Hu, Lei Wang, Yang Wang, Deheng Ye, Peilin Zhao, Ee-Peng Lim, Hui Xiong, and Hao Wang. 2024. LLM-based agent society investigation: Collaboration and confrontation in avalon gameplay. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, page… [find](https://scholar.google.com/scholar?q=Yihuai+Lan%2C+Zhiqiang+Hu%2C+Lei+Wang%2C+Yang+Wang%2C+Deheng+Ye%2C+Peilin+Zhao%2C+Ee%2DPeng+Lim%2C+Hui+Xiong%2C+and+Hao+Wang%2E+2024%2E+LLM%2Dbased+agent+society+investigation%3A+Collaboration+and+confrontation+in+avalon+gamep)
9. Juhao Liang, Zhenyang Cai, Jianqing Zhu, Huang Huang, Kewei Zong, Bang An, Mosen Alharthi, Juncai He, Lian Zhang, Haizhou Li, Benyou Wang, and Jinchao Xu. 2024. [Alignment at pre-training! towards native alignment for arabic LLMs](https://md.research.elamur.ai/free/en/papers/liang-2024-alignment-pre-training-towards-native/). In The Thirty-eighth Annual Conference on Neural Information Processin…
10. Jonathan Light, Min Cai, Weiqin Chen, Guanzhi Wang, Xiusi Chen, Wei Cheng, Yisong Yue, and Ziniu Hu. 2025. Strategist: Self-improvement of LLM decision making via bi-level tree search. In The Thirteenth International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Jonathan+Light%2C+Min+Cai%2C+Weiqin+Chen%2C+Guanzhi+Wang%2C+Xiusi+Chen%2C+Wei+Cheng%2C+Yisong+Yue%2C+and+Ziniu+Hu%2E+2025%2E+Strategist%3A+Self%2Dimprovement+of+LLM+decision+making+via+bi%2Dlevel+tree+search%2E+In+The+Thirteen)
11. Jonathan Light, Min Cai, Sheng Shen, and Ziniu Hu. 2023. From text to tactic: Evaluating LLMs playing the game of avalon. In NeurIPS 2023 Foundation Models for Decision Making Workshop. [find](https://scholar.google.com/scholar?q=Jonathan+Light%2C+Min+Cai%2C+Sheng+Shen%2C+and+Ziniu+Hu%2E+2023%2E+From+text+to+tactic%3A+Evaluating+LLMs+playing+the+game+of+avalon%2E+In+NeurIPS+2023+Foundation+Models+for+Decision+Making+Workshop%2E)
12. Weiyu Ma, Qirui Mi, Yongcheng Zeng, Xue Yan, Runji Lin, Yuqiao Wu, Jun Wang, and Haifeng Zhang. 2024. Large language models play starcraft II:benchmarks and a chain of summarization approach. In The Thirty-eighth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Weiyu+Ma%2C+Qirui+Mi%2C+Yongcheng+Zeng%2C+Xue+Yan%2C+Runji+Lin%2C+Yuqiao+Wu%2C+Jun+Wang%2C+and+Haifeng+Zhang%2E+2024%2E+Large+language+models+play+starcraft+II%3Abenchmarks+and+a+chain+of+summarization+approach%2E+In+The+T)
13. Shravan Nayak, Xiangru Jian, Kevin Qinghong Lin, Juan A. Rodriguez, Montek Kalsi, Nicolas Chapados, M. Tamer Özsu, Aishwarya Agrawal, David Vazquez, Christopher Pal, Perouz Taslakian, Spandana Gella, and Sai Rajeswar. 2025. UI-vision: A desktop-centric GUI benchmark for visual perception and interac… [find](https://scholar.google.com/scholar?q=Shravan+Nayak%2C+Xiangru+Jian%2C+Kevin+Qinghong+Lin%2C+Juan+A%2E+Rodriguez%2C+Montek+Kalsi%2C+Nicolas+Chapados%2C+M%2E+Tamer+%C3%96zsu%2C+Aishwarya+Agrawal%2C+David+Vazquez%2C+Christopher+Pal%2C+Perouz+Taslakian%2C+Spandana+Gella%2C+)
14. H Osawa, F Toriumi, D Katagami, K Shinoda, and M Inaba. 2014. Designing protocol of werewolf game: protocol for inference and persuasion. The 24th Fuzzy, Artificial Intelligence, Neural Networks and Computational Intelligence (FAN 2014). [find](https://scholar.google.com/scholar?q=H+Osawa%2C+F+Toriumi%2C+D+Katagami%2C+K+Shinoda%2C+and+M+Inaba%2E+2014%2E+Designing+protocol+of+werewolf+game%3A+protocol+for+inference+and+persuasion%2E+The+24th+Fuzzy%2C+Artificial+Intelligence%2C+Neural+Networks+and+C)
15. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Gray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul Christiano, Jan Leike, and Ryan Lowe. 2022. Trai… — [Training language models to follow instructions with human feedback](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/)
16. Qwen, :, An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, and 25 others. 2025. Qwen2.5 technical report. Preprint, arXiv:2412.15115. [find](https://scholar.google.com/scholar?q=Qwen%2C+%3A%2C+An+Yang%2C+Baosong+Yang%2C+Beichen+Zhang%2C+Binyuan+Hui%2C+Bo+Zheng%2C+Bowen+Yu%2C+Chengyuan+Li%2C+Dayiheng+Liu%2C+Fei+Huang%2C+Haoran+Wei%2C+Huan+Lin%2C+Jian+Yang%2C+Jianhong+Tu%2C+Jianwei+Zhang%2C+Jianxin+Yang%2C+Jiaxi+)
17. Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. 2023. [Direct preference optimization: Your language model is secretly a reward model](https://md.research.elamur.ai/free/en/papers/rafailov-2023-direct-preference-optimization-language-model/). In Thirty-seventh Conference on Neural Information Processing Systems.
18. Jack Serrino, Max Kleiman-Weiner, David C. Parkes, and Joshua B. Tenenbaum. 2019. Finding friend and foe in multi-agent games. Curran Associates Inc., Red Hook, NY, USA. [find](https://scholar.google.com/scholar?q=Jack+Serrino%2C+Max+Kleiman%2DWeiner%2C+David+C%2E+Parkes%2C+and+Joshua+B%2E+Tenenbaum%2E+2019%2E+Finding+friend+and+foe+in+multi%2Dagent+games%2E+Curran+Associates+Inc%2E%2C+Red+Hook%2C+NY%2C+USA%2E)
19. Xiao Shao, Weifu Jiang, Fei Zuo, and Mengqing Liu. 2024a. Swarmbrain: Embodied agent for real-time strategy game starcraft ii via large language models. Preprint, arXiv:2401.17749. — [arXiv:2401.17749](https://arxiv.org/abs/2401.17749) [find](https://scholar.google.com/scholar?q=Xiao+Shao%2C+Weifu+Jiang%2C+Fei+Zuo%2C+and+Mengqing+Liu%2E+2024a%2E+Swarmbrain%3A+Embodied+agent+for+real%2Dtime+strategy+game+starcraft+ii+via+large+language+models%2E+Preprint%2C+arXiv%3A2401%2E17749%2E+%E2%80%94+%5BarXiv%3A2401%2E17749)
20. Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, Y. K. Li, Y. Wu, and Daya Guo. 2024b. [Deepseekmath: Pushing the limits of mathematical reasoning in open language models](https://md.research.elamur.ai/free/en/papers/shao-2024-deepseekmath-pushing-limits-mathematical-reasoning/). Preprint, arXiv:2402.03300.
21. Zijing Shi, Meng Fang, Shunfeng Zheng, Shilong Deng, Ling Chen, and Yali Du. 2023. Cooperation on the fly: Exploring language agents for ad hoc teamwork in the avalon game. Preprint, arXiv:2312.17515. — [arXiv:2312.17515](https://arxiv.org/abs/2312.17515) [find](https://scholar.google.com/scholar?q=Zijing+Shi%2C+Meng+Fang%2C+Shunfeng+Zheng%2C+Shilong+Deng%2C+Ling+Chen%2C+and+Yali+Du%2E+2023%2E+Cooperation+on+the+fly%3A+Exploring+language+agents+for+ad+hoc+teamwork+in+the+avalon+game%2E+Preprint%2C+arXiv%3A2312%2E17515%2E)
22. Andrew Szot, Max Schwarzer, Harsh Agrawal, Bogdan Mazoure, Rin Metcalf, Walter Talbott, Natalie Mackraz, R Devon Hjelm, and Alexander T Toshev. 2024. Large language models as generalizable policies for embodied tasks. In The Twelfth International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Andrew+Szot%2C+Max+Schwarzer%2C+Harsh+Agrawal%2C+Bogdan+Mazoure%2C+Rin+Metcalf%2C+Walter+Talbott%2C+Natalie+Mackraz%2C+R+Devon+Hjelm%2C+and+Alexander+T+Toshev%2E+2024%2E+Large+language+models+as+generalizable+policies+fo)
23. Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anandkumar. 2024a. Voyager: An open-ended embodied agent with large language models. Transactions on Machine Learning Research. [find](https://scholar.google.com/scholar?q=Guanzhi+Wang%2C+Yuqi+Xie%2C+Yunfan+Jiang%2C+Ajay+Mandlekar%2C+Chaowei+Xiao%2C+Yuke+Zhu%2C+Linxi+Fan%2C+and+Anima+Anandkumar%2E+2024a%2E+Voyager%3A+An+open%2Dended+embodied+agent+with+large+language+models%2E+Transactions+on+)
24. Ruiyi Wang, Haofei Yu, Wenxin Zhang, Zhengyang Qi, Maarten Sap, Yonatan Bisk, Graham Neubig, and Hao Zhu. 2024b. SOTOPIA-π: Interactive learning of socially intelligent language agents. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)… [find](https://scholar.google.com/scholar?q=Ruiyi+Wang%2C+Haofei+Yu%2C+Wenxin+Zhang%2C+Zhengyang+Qi%2C+Maarten+Sap%2C+Yonatan+Bisk%2C+Graham+Neubig%2C+and+Hao+Zhu%2E+2024b%2E+SOTOPIA%2D%CF%80%3A+Interactive+learning+of+socially+intelligent+language+agents%2E+In+Proceedings)
25. Shenzhi Wang, Chang Liu, Zilong Zheng, Siyuan Qi, Shuo Chen, Qisen Yang, Andrew Zhao, Chaofei Wang, Shiji Song, and Gao Huang. 2023. Avalon’s game of thoughts: Battle against deception through recursive contemplation. Preprint, arXiv:2310.01320. — [arXiv:2310.01320](https://arxiv.org/abs/2310.01320) [find](https://scholar.google.com/scholar?q=Shenzhi+Wang%2C+Chang+Liu%2C+Zilong+Zheng%2C+Siyuan+Qi%2C+Shuo+Chen%2C+Qisen+Yang%2C+Andrew+Zhao%2C+Chaofei+Wang%2C+Shiji+Song%2C+and+Gao+Huang%2E+2023%2E+Avalon%E2%80%99s+game+of+thoughts%3A+Battle+against+deception+through+recursi)
26. Shenzhi Wang, Chang Liu, Zilong Zheng, Siyuan Qi, Shuo Chen, Qisen Yang, Andrew Zhao, Chaofei Wang, Shiji Song, and Gao Huang. 2024c. Boosting LLM agents with recursive contemplation for effective deception handling. In Findings of the Association for Computational Linguistics: ACL 2024, pages 9909–… [find](https://scholar.google.com/scholar?q=Shenzhi+Wang%2C+Chang+Liu%2C+Zilong+Zheng%2C+Siyuan+Qi%2C+Shuo+Chen%2C+Qisen+Yang%2C+Andrew+Zhao%2C+Chaofei+Wang%2C+Shiji+Song%2C+and+Gao+Huang%2E+2024c%2E+Boosting+LLM+agents+with+recursive+contemplation+for+effective+dec)
27. Tianhe Wang and Tomoyuki Kaneko. 2018. Application of deep reinforcement learning in werewolf game agents. In 2018 Conference on Technologies and Applications of Artificial Intelligence (TAAI), pages 28–33. IEEE. [find](https://scholar.google.com/scholar?q=Tianhe+Wang+and+Tomoyuki+Kaneko%2E+2018%2E+Application+of+deep+reinforcement+learning+in+werewolf+game+agents%2E+In+2018+Conference+on+Technologies+and+Applications+of+Artificial+Intelligence+%28TAAI%29%2C+pages+)
28. Xiaoqiang Wang and Bang Liu. 2025. OSCAR: Operating system control via state-aware reasoning and re-planning. In The Thirteenth International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Xiaoqiang+Wang+and+Bang+Liu%2E+2025%2E+OSCAR%3A+Operating+system+control+via+state%2Daware+reasoning+and+re%2Dplanning%2E+In+The+Thirteenth+International+Conference+on+Learning+Representations%2E)
29. Yiqin Wang, Haoji Zhang, Jingqi Tian, and Yansong Tang. 2025. Ponder & press: Advancing visual GUI agent towards general computer control. In Findings of the Association for Computational Linguistics: ACL 2025, pages 1461–1473, Vienna, Austria. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Yiqin+Wang%2C+Haoji+Zhang%2C+Jingqi+Tian%2C+and+Yansong+Tang%2E+2025%2E+Ponder+%26+press%3A+Advancing+visual+GUI+agent+towards+general+computer+control%2E+In+Findings+of+the+Association+for+Computational+Linguistics%3A)
30. Shuang Wu, Liwen Zhu, Tao Yang, Shiwei Xu, Qiang Fu, Yang Wei, and Haobo Fu. 2024. Enhance reasoning for large language models in the game werewolf. Preprint, arXiv:2402.02330. others. 2025. [Qwen3 technical report](https://md.research.elamur.ai/free/en/papers/yang-2025-qwen3-technical-report/). Preprint, arXiv:2505.09388.
31. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. 2023. [ReAct: Synergizing reasoning and acting in language models](https://md.research.elamur.ai/free/en/papers/yao-2022-react-synergizing-reasoning-acting-language/). In International Conference on Learning Representations (ICLR).
32. Wenyuan Zhang, Tianyun Liu, Mengxiao Song, Xiaodong Li, and Tingwen Liu. 2025a. SOTOPIA-Ω: Dynamic strategy injection learning and social instruction following evaluation for social agents. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Pap… [find](https://scholar.google.com/scholar?q=Wenyuan+Zhang%2C+Tianyun+Liu%2C+Mengxiao+Song%2C+Xiaodong+Li%2C+and+Tingwen+Liu%2E+2025a%2E+SOTOPIA%2D%E2%84%A6%3A+Dynamic+strategy+injection+learning+and+social+instruction+following+evaluation+for+social+agents%2E+In+Proceed)
33. Xuanming Zhang, Yuxuan Chen, Samuel Yeh, and Sharon Li. 2025b. Metamind: Modeling human social thoughts with metacognitive multi-agent systems. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Xuanming+Zhang%2C+Yuxuan+Chen%2C+Samuel+Yeh%2C+and+Sharon+Li%2E+2025b%2E+Metamind%3A+Modeling+human+social+thoughts+with+metacognitive+multi%2Dagent+systems%2E+In+The+Thirty%2Dninth+Annual+Conference+on+Neural+Informat)
34. Zheng Zhang, Yihuai Lan, Yangsen Chen, Lei Wang, Xiang Wang, and Hao Wang. 2025c. Dvm: Towards controllable llm agents in social deduction games. In ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5. [find](https://scholar.google.com/scholar?q=Zheng+Zhang%2C+Yihuai+Lan%2C+Yangsen+Chen%2C+Lei+Wang%2C+Xiang+Wang%2C+and+Hao+Wang%2E+2025c%2E+Dvm%3A+Towards+controllable+llm+agents+in+social+deduction+games%2E+In+ICASSP+2025+%2D+2025+IEEE+International+Conference+on)
35. Xuhui Zhou, Hao Zhu, Leena Mathur, Ruohong Zhang, Haofei Yu, Zhengyang Qi, Louis-Philippe Morency, Yonatan Bisk, Daniel Fried, Graham Neubig, and Maarten Sap. 2024. [SOTOPIA: Interactive evaluation for social intelligence in language agents](https://md.research.elamur.ai/free/en/papers/zhou-2023-sotopia-interactive-evaluation-social-intelligence/). In The Twelfth International Conference on Learning Represe…
36. Yuzhuang Xu, Shuo Wang, Peng Li, Fuwen Luo, Xiaolong Wang, Weidong Liu, and Yang Liu. 2023. [Exploring large language models for communication games: An empirical study on werewolf](https://md.research.elamur.ai/free/en/papers/xu-2023-exploring-large-language-models-communication/). arXiv preprint arXiv:2309.04658.
37. Zelai Xu, Wanjun Gu, Chao Yu, Yi Wu, and Yu Wang. 2025. Learning strategic language agents in the werewolf game with iterative latent space policy optimization. In Forty-second International Conference on Machine Learning. [find](https://scholar.google.com/scholar?q=Zelai+Xu%2C+Wanjun+Gu%2C+Chao+Yu%2C+Yi+Wu%2C+and+Yu+Wang%2E+2025%2E+Learning+strategic+language+agents+in+the+werewolf+game+with+iterative+latent+space+policy+optimization%2E+In+Forty%2Dsecond+International+Conferenc)
38. Zelai Xu, Chao Yu, Fei Fang, Yu Wang, and Yi Wu. 2024. Language agents with reinforcement learning for strategic play in the werewolf game. In Proceedings of the 41st International Conference on Machine Learning, ICML’24. JMLR.org. [find](https://scholar.google.com/scholar?q=Zelai+Xu%2C+Chao+Yu%2C+Fei+Fang%2C+Yu+Wang%2C+and+Yi+Wu%2E+2024%2E+Language+agents+with+reinforcement+learning+for+strategic+play+in+the+werewolf+game%2E+In+Proceedings+of+the+41st+International+Conference+on+Machi)

*Updated 2 November 2026*