MAgIC: Investigation of Large Language Model Powered Multi-Agent in Cognition, Adaptability, Rationality and Collaboration

TL;DR

This paper introduces a new benchmark framework for evaluating large language models in multi-agent social and cognitive scenarios, highlighting their reasoning, collaboration, and rationality capabilities.

Contribution

It presents a novel competition-based benchmark with probabilistic graphic modeling to assess LLMs' multi-agent social and cognitive abilities, including a comprehensive evaluation of seven models.

Findings

Significant capability gap between strongest and weakest models.

PGM enhancement improves model abilities by an average of 37%.

Evaluation across diverse social deduction and game-theory scenarios.

Abstract

Large Language Models (LLMs) have significantly advanced natural language processing, demonstrating exceptional reasoning, tool usage, and memory capabilities. As their applications expand into multi-agent environments, there arises a need for a comprehensive evaluation framework that captures LLMs' reasoning, planning, collaboration, and other social abilities. This work introduces a novel competition-based benchmark framework specifically designed to assess LLMs within multi-agent settings, providing quantitative metrics to evaluate their judgment, reasoning, deception, self-awareness, cooperation, coordination, and rationality. We utilize two social deduction games alongside three game-theory scenarios to create diverse environments. Our frame is fortified with the probabilistic graphic modeling (PGM) method, enhancing the LLMs' capabilities in navigating complex social and cognitive dimensions. We evaluate seven LLMs, quantitatively highlighting a significant capability gap of over threefold between the strongest, GPT o1, and the weakest, Llama-2-70B. It also confirms that our PGM enhancement boosts the abilities of all selected models by an average of 37%. Our data and code can be found here https://github.com/cathyxl/MAgIC.