Network Topology Impacts Consensus Speed and Adversarial Vulnerability in LLM Populations
October 23, 2026
Simulations of up to 10^5 LLM agents show that network topology dictates how quickly conventions form and how resilient they are to adversarial attacks. Sparse networks can slow consensus by twelvefold but may also prevent total convergence failure seen in well-mixed populations.
HOW THIS AFFECTS YOU
●
researcherYou should consider the interaction architecture when studying emergent behavior in multi-agent systems.
●
policyThis highlights how the structure of model deployments affects the systemic stability and susceptibility to manipulation.