Communication Enhances LLMs' Stability in Strategic Thinking

TL;DR

This paper shows that simple, costless communication can significantly improve the stability and predictability of strategic behavior in large language models during multi-agent interactions, especially in volatile models.

Contribution

It demonstrates that cheap-talk style communication reduces variability in LLMs' strategic decisions, enhancing stability across different models and contexts.

Findings

Communication reduces trajectory noise in LLMs' strategic behavior

The stabilizing effect is consistent across multiple prompt variants

Higher baseline volatility models benefit most from communication

Abstract

Large Language Models (LLMs) often exhibit pronounced context-dependent variability that undermines predictable multi-agent behavior in tasks requiring strategic thinking. Focusing on models that range from 7 to 9 billion parameters in size engaged in a ten-round repeated Prisoner's Dilemma, we evaluate whether short, costless pre-play messages emulating the cheap-talk paradigm affect strategic stability. Our analysis uses simulation-level bootstrap resampling and nonparametric inference to compare cooperation trajectories fitted with LOWESS regression across both the messaging and the no-messaging condition. We demonstrate consistent reductions in trajectory noise across a majority of the model-context pairings being studied. The stabilizing effect persists across multiple prompt variants and decoding regimes, though its magnitude depends on model choice and contextual framing, with models displaying higher baseline volatility gaining the most. While communication rarely produces harmful instability, we document a few context-specific exceptions and identify the limited domains in which communication harms stability. These findings position cheap-talk style communication as a low-cost, practical tool for improving the predictability and reliability of strategic behavior in multi-agent LLM systems.