Pith. sign in

REVIEW 4 cited by

Enhance Reasoning for Large Language Models in the Game Werewolf

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.02330 v2 pith:YQPCVDSS submitted 2024-02-04 cs.AI cs.CL

Shuang Wu , Liwen Zhu , Tao Yang , Shiwei Xu , Qiang Fu , Yang Wei , Haobo Fu This is my paper · ORCID
classification cs.AIcs.CL
keywords thinkerreasoningframeworkllmsgamelanguageenhanceknowledge
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

This paper presents an innovative framework that integrates Large Language Models (LLMs) with an external Thinker module to enhance the reasoning capabilities of LLM-based agents. Unlike augmenting LLMs with prompt engineering, Thinker directly harnesses knowledge from databases and employs various optimization techniques. The framework forms a reasoning hierarchy where LLMs handle intuitive System-1 tasks such as natural language processing, while the Thinker focuses on cognitive System-2 tasks that require complex logical analysis and domain-specific knowledge. Our framework is presented using a 9-player Werewolf game that demands dual-system reasoning. We introduce a communication protocol between LLMs and the Thinker, and train the Thinker using data from 18800 human sessions and reinforcement learning. Experiments demonstrate the framework's effectiveness in deductive reasoning, speech generation, and online game evaluation. Additionally, we fine-tune a 6B LLM to surpass GPT4 when integrated with the Thinker. This paper also contributes the largest dataset for social deduction games to date.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. CaM-Wolf: Causal-Aware Multimodal Agents for Social Deduction Games

    cs.AI 2026-07 conditional novelty 6.0 of 10

    CaM-Wolf is a multimodal Werewolf agent that perceives player video, reasons about hidden roles with a counterfactual-intervention-trained RL reasoner, and responds through an animated avatar.

  2. Strategy Adaptation in Large Language Model Werewolf Agents

    cs.CL 2025-07 conditional novelty 6.0 of 10

    Dynamically switching between Support and Attack strategies based on role estimates raises win rates for Werewolf LLM agents, with mixed effects for Villagers.

  3. Empowering Economic Simulation for Massively Multiplayer Online Games through Generative Agent-Based Modeling

    cs.AI 2025-06 conditional novelty 6.0 of 10

    LLM-driven agents in a simulated MMO economy reproduce role specialization and price responses to supply and demand, though the price result is partly shaped by what the AI is told.

  4. Ethical Considerations of Large Language Models in Game Playing

    cs.CL 2025-08 conditional novelty 5.0 of 10

    In Werewolf games, LLM agents change their kills, votes, and trust scores based on explicit gender labels and even based on gender-implied first names, behaving differently for male and female players.

Pith tools