Herd Behavior: Investigating Peer Influence in LLM-based Multi-Agent Systems

2025-05-27

AI Summary

This study investigates herding behavior—i.e., output convergence among agents—in LLM-driven multi-agent systems, induced by peer influence. Method: Using a controlled experimental paradigm, we systematically examine the modulation of group consensus via confidence bias and information presentation format across leading open-source LLMs, integrating prompt engineering, confidence modeling, and multi-round interaction analysis. Contribution/Results: We首次 demonstrate that “moderate conformity” can be deliberately induced: when an agent perceives peer confidence marginally higher than its own, moderate consensus enhances collaborative efficacy. We quantitatively identify key drivers of group-level conformity tendencies and achieve interpretable, controllable intervention. Experiments show significant improvements in answer consistency and task completion rate. This work establishes a novel paradigm for designing controllable, cooperative multi-agent systems grounded in socially informed behavioral regulation.

Technology Category

• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Recent advancements in Large Language Models (LLMs) have enabled the emergence of multi-agent systems where LLMs interact, collaborate, and make decisions in shared environments. While individual model behavior has been extensively studied, the dynamics of peer influence in such systems remain underexplored. In this paper, we investigate herd behavior, the tendency of agents to align their outputs with those of their peers, within LLM-based multi-agent interactions. We present a series of controlled experiments that reveal how herd behaviors are shaped by multiple factors. First, we show that the gap between self-confidence and perceived confidence in peers significantly impacts an agent's likelihood to conform. Second, we find that the format in which peer information is presented plays a critical role in modulating the strength of herd behavior. Finally, we demonstrate that the degree of herd behavior can be systematically controlled, and that appropriately calibrated herd tendencies can enhance collaborative outcomes. These findings offer new insights into the social dynamics of LLM-based systems and open pathways for designing more effective and adaptive multi-agent collaboration frameworks.

Problem

• Investigating herd behavior in LLM-based multi-agent systems
• Exploring factors influencing peer conformity in agent interactions
• Controlling herd tendencies to improve collaborative outcomes

Innovation

• Study herd behavior in LLM multi-agent systems
• Control herd behavior via confidence and presentation
• Enhance collaboration with calibrated herd tendencies