Discovering Differences in Strategic Behavior Between Humans and LLMs

TL;DR

This paper uses a novel program discovery approach to analyze and compare the strategic behaviors of humans and large language models in game scenarios, revealing that advanced LLMs can exhibit deeper strategic reasoning than humans.

Contribution

It introduces a method employing AlphaEvolve to discover interpretable models of human and LLM behavior, enabling structural analysis of their strategic differences.

Findings

Frontier LLMs demonstrate deeper strategic behavior than humans.

The approach uncovers structural factors influencing behavior in strategic interactions.

Analysis is based on iterated rock-paper-scissors game data.

Abstract

As Large Language Models (LLMs) are increasingly deployed in social and strategic scenarios, it becomes critical to understand where and why their behavior diverges from that of humans. While behavioral game theory (BGT) provides a framework for analyzing behavior, existing models do not fully capture the idiosyncratic behavior of humans or black-box, non-human agents like LLMs. We employ AlphaEvolve, a cutting-edge program discovery tool, to directly discover interpretable models of human and LLM behavior from data, thereby enabling open-ended discovery of structural factors driving human and LLM behavior. Our analysis on iterated rock-paper-scissors reveals that frontier LLMs can be capable of deeper strategic behavior than humans. These results provide a foundation for understanding structural differences driving differences in human and LLM behavior in strategic interactions.