Representational Similarity and Model Behavior in Multi-Agent Interaction

2026-06-05

AI Summary

This study investigates how representational similarity among large language models (LLMs) influences cooperative and innovative behaviors in multi-agent settings. By analyzing interactions between 276 pairs of LLMs across eight tasks spanning cooperation and novelty, the research integrates representational similarity analysis, controlled regression, and cross-task evaluation to systematically uncover a trade-off between inter-model representational similarity and performance outcomes. The findings reveal that model pairs with higher representational similarity achieve significantly better cooperation but exhibit lower innovation. Notably, similarity in early network layers demonstrates the strongest predictive power, suggesting that shared semantic foundations underlie this trade-off. This work provides the first systematic evidence linking internal representational structure to emergent collaborative and creative dynamics in LLM-based multi-agent systems.

Abstract

Researchers have shown that neural similarity among humans predicts social closeness and cooperative success, whereas innovation often emerges from interactions among dissimilar individuals. We investigate whether these principles extend to artificial intelligence by examining interactions between large language models. In our experiments, 276 model pairs interact across eight games spanning both cooperation and novelty. We find that pairs with more similar representation spaces achieve significantly higher cooperation but exhibit reduced novelty and creativity. The effects of representational similarity on cooperation and novelty remain robust even after controlling for other factors such as performance disparity and model size. We also find that similarity in the early layers consistently shows the strongest association with cooperation and novelty, compared to the middle and later layers. This suggests that a central factor underlying these patterns could be the extent to which the two models share lexical and semantic grounding. Overall, representational similarity can be an important consideration in multi-agent system design.

Problem

• representational similarity
• multi-agent interaction
• cooperation
• novelty
• large language models

Innovation

• representational similarity
• multi-agent interaction
• large language models
• cooperation vs. novelty
• neural representation