mlsanityData to 7 Oct 2026

MARBO: Relational Belief Grounding for LLM Agents in Social Deduction Games

arXiv 2609.06563 ↗First posted 6 Sep 2026cs.AI cs.LGcited by 0

On the map: LLM agents and tool use → Reinforcement learning for LLM agents

Abstract (arXiv)

Social deduction games (SDGs) require agents to reason under partial observability by maintaining relational beliefs about hidden roles and team alignments. While recent LLM-agent approaches improve gameplay through prompting and preference optimization, they often optimize actions and in-game speech without explicitly grounding them in such beliefs. This frequently leads to strategically inconsistent behavior, especially for compact LLM agents. We introduce Multi-Agent Relational Belief Optimization (MARBO), a belief-grounded preference optimization framework that leverages relational beliefs to guide strategic decisions and in-game speech. MARBO provides preference feedback only when behaviors are supported by reliable relational beliefs and lead to strategically favorable social outcomes, encouraging more consistent learning under uncertainty. Experiments on representative SDGs show that MARBO enables compact LLM agents to consistently outperform existing baselines. The Code is available on https://github.com/PleaseTakemeAway/MARBO.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use. Builds on 28 of its 52 references that are in the map. First posted on arXiv 6 Sep 2026.

Builds on

28 of 52 references.

Closest ancestors · 7

Same direction, newest first

  1. 2025MultiMind: Enhancing Werewolf Agents with Multimodal Reasoning and Theory of Mind25 Apr
  2. 2025Learning Strategic Language Agents in the Werewolf Game with Iterative Latent Space Policy Optimization7 Feb
  3. 2025Multi-agent KTO: Reinforcing Strategic Interactions of Large Language Model in Language Game24 Jan
  4. 2024Enhance Reasoning for Large Language Models in the Game Werewolf4 Feb
  5. 2023Language Agents with Reinforcement Learning for Strategic Play in the Werewolf Game29 Oct
  6. 2023AvalonBench: Evaluating LLMs Playing the Game of Avalon8 Oct
  7. 2023Reflexion: Language Agents with Verbal Reinforcement Learning20 Mar

Other references · 21

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  4. 2017Deep reinforcement learning from human preferencescited by 3,846
  5. 2023Let's Verify Step by Stepcited by 3,598
  6. 2021The Surprising Effectiveness of PPO in Cooperative, Multi-Agent Gamescited by 903
  7. 2023ExpeL: LLM Agents Are Experiential Learnerscited by 814
  8. 2018QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learningcited by 697
  9. 2023RRHF: Rank Responses to Align Language Models with Human Feedback without tearscited by 454
  10. 2023SLiC-HF: Sequence Likelihood Calibration with Human Feedbackcited by 395
  11. 2023Statistical Rejection Sampling Improves Preference Optimizationcited by 321
  12. 2024KTO: Model Alignment as Prospect Theoretic Optimizationcited by 256
9 more
  1. 2023Exploring Large Language Models for Communication Games: An Empirical Study on Werewolfcited by 238
  2. 2020"Other-Play" for Zero-Shot Coordinationcited by 194
  3. 2024Werewolf Arena: A Case Study in LLM Evaluation via Social Deductioncited by 42
  4. 2023Robust Multi-Agent Reinforcement Learning via Adversarial Regularization: Theoretical Foundation and Stable Algorithmscited by 25
  5. 2025Among Us: A Sandbox for Measuring and Detecting Agentic Deceptioncited by 24
  6. 2025Bayesian Social Deduction with Graph-Informed Language Modelscited by 8
  7. 2026One Model, All Roles: Multi-Turn, Multi-Agent Self-Play Reinforcement Learning for Conversational Social Intelligencecited by 7
  8. 2026Generalized Per-Agent Advantage Estimation for Multi-Agent Policy Optimizationcited by 2
  9. 2026Shaping Zero-Shot Coordination via State Blockingcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025DVM: Towards Controllable LLM Agents in Social Deduction Games
  2. 2026Relational Priors as Convergence Pressure in LLM-Based Multi-Agent Systems
  3. 2026MARO: Learning Stronger Reasoning from Social Interaction
  4. 2026Auditing Belief-Conditioned LLM Agents in Hidden-Information Social Deduction Games
  5. 2025From Debate to Equilibrium: Belief-Driven Multi-Agent LLM Reasoning via Bayesian Nash Equilibrium
  6. 2025Approximating Human Strategic Reasoning with LLM-Enhanced Recursive Reasoners Leveraging Multi-agent Hypergames
  7. 2025Reasoning and Behavioral Equilibria in LLM-Nash Games: From Mindsets to Actions
  8. 2024K-Level Reasoning: Establishing Higher Order Beliefs in Large Language Models for Strategic Reasoning