# Deceive, Detect, and Disclose: Large Language Models Play Mini-Mafia

arXiv: https://arxiv.org/abs/2509.23023 · HTML: https://mlsanity.com/paper/2509.23023/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 45 references that are in the map.
- First posted on arXiv 27 Sep 2025.

## Builds on (most cited first, 23 of 23)

- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [A Survey on Large Language Model based Autonomous Agents](https://mlsanity.com/paper/2308.11432/) (2023)
- [MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework](https://mlsanity.com/paper/2308.00352/) (2023)
- [Large Language Model based Multi-Agents: A Survey of Progress and Challenges](https://mlsanity.com/paper/2402.01680/) (2024)
- [Frontier Models are Capable of In-context Scheming](https://mlsanity.com/paper/2412.04984/) (2024)
- [Large Language Models Empowered Agent-based Modeling and Simulation: A Survey and Perspectives](https://mlsanity.com/paper/2312.11970/) (2023)
- [AI Deception: A Survey of Examples, Risks, and Potential Solutions](https://mlsanity.com/paper/2308.14752/) (2023)
- [Language Agents with Reinforcement Learning for Strategic Play in the Werewolf Game](https://mlsanity.com/paper/2310.18940/) (2023)
- [Truthful AI: Developing and governing AI that does not lie](https://mlsanity.com/paper/2110.06674/) (2021)
- [Measuring Implicit Bias in Explicitly Unbiased Large Language Models](https://mlsanity.com/paper/2402.04105/) (2024)
- [Enhance Reasoning for Large Language Models in the Game Werewolf](https://mlsanity.com/paper/2402.02330/) (2024)
- [Werewolf Arena: A Case Study in LLM Evaluation via Social Deduction](https://mlsanity.com/paper/2407.13943/) (2024)
- [Among Us: A Sandbox for Measuring and Detecting Agentic Deception](https://mlsanity.com/paper/2504.04072/) (2025)
- [Learning to Discuss Strategically: A Case Study on One Night Ultimate Werewolf](https://mlsanity.com/paper/2405.19946/) (2024)
- [The Vulnerability of Language Model Benchmarks: Do They Accurately Reflect True LLM Performance?](https://mlsanity.com/paper/2412.03597/) (2024)
- [Playing the Werewolf game with artificial intelligence for language understanding](https://mlsanity.com/paper/2302.10646/) (2023)
- [When Benchmarks Talk: Re-Evaluating Code LLMs with Interactive Feedback](https://mlsanity.com/paper/2502.18413/) (2025)
- [Learning Strategic Language Agents in the Werewolf Game with Iterative Latent Space Policy Optimization](https://mlsanity.com/paper/2502.04686/) (2025)
- [Helmsman of the Masses? Evaluate the Opinion Leadership of Large Language Models in the Werewolf Game](https://mlsanity.com/paper/2404.01602/) (2024)
- [MultiMind: Enhancing Werewolf Agents with Multimodal Reasoning and Theory of Mind](https://mlsanity.com/paper/2504.18039/) (2025)
- [Werewolf Among Us: A Multimodal Dataset for Modeling Persuasion Behaviors in Social Deduction Games](https://mlsanity.com/paper/2212.08279/) (2022)
- [EgoSocialArena: Benchmarking the Social Intelligence of Large Language Models from a First-person Perspective](https://mlsanity.com/paper/2410.06195/) (2024)
- [Strategy Adaptation in Large Language Model Werewolf Agents](https://mlsanity.com/paper/2507.12732/) (2025)

## Built on by (most cited first, 9 of 9)

- [Moral Susceptibility and Robustness under Persona Role-Play in Large Language Models](https://mlsanity.com/paper/2511.08565/) (2025)
- [SNEAK: Evaluating Strategic Communication and Information Leakage in Large Language Models](https://mlsanity.com/paper/2603.29846/) (2026)
- [MafiaScope: Non-Invasive, Time-Resolved Belief Probing for LLM Agents in Social Deduction Games](https://mlsanity.com/paper/2607.10645/) (2026)
- [Evaluating Large Language Models in a Complex Hidden Role Game](https://mlsanity.com/paper/2605.22826/) (2026)
- [RogueAI: A Reverse Turing Test for Detecting Licensed AI Deception in Dialogue](https://mlsanity.com/paper/2606.13310/) (2026)
- [When Agents Lie: Premeditation, Persistence, and Exploitation in Repeated Games](https://mlsanity.com/paper/2607.05132/) (2026)
- [Even More Deception: Objective Misalignment in Mixed-Motive LLM Multi-Agent Systems](https://mlsanity.com/paper/2607.26120/) (2026)
- [Can Agents Deceive? Evaluating Reasoning and Deception in ParliamentBench using a Social Deduction Game](https://mlsanity.com/paper/2607.28146/) (2026)
- [Communication Shapes Collective Inference in Self-Adapting LLM Societies: Evidence from Mafia](https://mlsanity.com/paper/2610.05041/) (2026)

## Abstract (arXiv)

Large language models are increasingly deployed in multi-agent settings whose outcomes hinge on social intelligence, motivating evaluations of their interactive capabilities; yet existing studies remain overwhelmingly empirical, leaving us without a theoretical understanding of how agent interactions determine collective outcomes. To address this, we introduce textitMini-Mafia, a four-player simplification of the social deduction game Mafia in which a fixed night phase reduces the game to a single critical exchange among a mafioso, a detective, and a villager. In this setting, we show that the mafia win-rate p is predicted by the analytical formula logit(p) = v × (m - d), where m, d, and v represent the mafioso's deception, the detective's disclosure, and the villager's detection capabilities. We turn this analytical framework into the textitMini-Mafia Benchmark, where Bayesian inference over gameplay data yields per-model estimates of the intrinsic parameters m, d, and v. For I models, only 3I parameters suffice to predict the outcomes of all I3 tournament combinations; and in 5-fold cross-validation the formula achieves a 76.6% Brier-score reduction over a random baseline. The benchmark also reveals counterintuitive results: Grok 3 Mini is the strongest detector and GPT-5 Mini the strongest discloser, both ahead of DeepSeek V3.1, Claude Opus 4, and Claude Sonnet 4; while Claude Sonnet 4 is the weakest detector, near random chance. Together, these results show that Mini-Mafia, a simple but nontrivial multi-agent system, admits an analytical description and serves as a principled benchmark for language model interactions.