Earticle

초록

영어

We investigate whether Large Language Model (LLM)s can learn strategic reasoning and social deception abilities through Reinforcement Learning (RL) finetuning via a multi-agent “Mafia Game” simulation environment. We finetune a baseline 7B model using Proximal Policy Optimization (PPO) with sparse binary rewards based on game outcomes. Training samples are collected through an opponent pool consisting of different versions of the finetuned model. Our experiment results show that the finetuned model outperforms the baseline model by a significant margin and suggest that strategic capabilities unseen in baseline models emerge.

목차

Abstract
I. INTRODUCTION
II. METHODOLOGY
A. Game Environment Setup
B. Training Setup
C. Opponent Pool Design
III. EXPERIMENTS AND RESULTS
IV. CONCLUSION
ACKNOWLEDGMENT
REFERENCES

저자

• Jiho Jun [ School of Electrical Engineering Korea University Seoul, Korea ]
• Junhee Seok [ School of Electrical Engineering Korea University Seoul, Korea ] Corresponding Author

참고문헌