Behavior probes reveal information flow in AI language model groups
Behavioral Probes for Information Flow in LLM Swarms
Multiagent SystemsArtificial Intelligence
Summary
When groups of AI language models work together to solve problems, it's hard to see how they share and use information inside the network. The authors study these groups by measuring how much each AI relies on outside information versus its own learned knowledge. They adapt tools originally meant to detect errors to track this flow of information. Their measurements help improve the group’s efficiency by cutting down unnecessary communication and speeding up problem solving.
What this means in practice
- •For ai system engineers: Optimize coordination in AI model groups by pruning unnecessary communications using behavioral probes.
- •For software developers: Accelerate AI group decision processes by guiding reasoning paths with information flow metrics.
Authors
Junhui Chen, Yu Zhang, Lantian Li, Dong Wang
Abstract
Most existing approaches to LLM swarm optimization are utility-driven, focusing on final task utility while leaving the processing, integration, and reliance of information within individual interactions largely implicit. To address this, we study LLM swarm interactions from an information-flow-based view and introduce behavioral probes to quantify information flow patterns. Specifically, we repurpose External Context Score (ECS) and Parametric Knowledge Score (PKS)---originally designed for hallucination analysis---to measure a node's external-context dependence and parametric-knowledge dependence, respectively. Across controlled and realistic swarm settings, we observe consistent patterns: ECS increases with input relevance and distinguishes useful multi-source contributions, while PKS decreases as informative context accumulates. Building on this heuristic interpretation, we demonstrate how these probes can complement utility-driven structure optimization in two controlled applications. Probe-guided node pruning reduces redundant communication and token consumption, and probe-guided reasoning-path search accelerates convergence compared to utility-only baselines. Our empirical results suggest that behavioral probes can make the internal information flow of LLM swarms observable and serve as effective auxiliary signals for structure optimization.