Lyapunov Probes for Hallucination Detection in Large Foundation Models
This addresses the problem of unreliable outputs in large language and multimodal models, offering a novel detection method that is incremental in its approach.
The paper tackles hallucination detection in large foundation models by framing them as dynamical systems and proposing Lyapunov Probes, which use stability constraints to distinguish factual from hallucination-prone regions, achieving consistent improvements over baselines in experiments.
We address hallucination detection in Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) by framing the problem through the lens of dynamical systems stability theory. Rather than treating hallucination as a straightforward classification task, we conceptualize (M)LLMs as dynamical systems, where factual knowledge is represented by stable equilibrium points within the representation space. Our main insight is that hallucinations tend to arise at the boundaries of knowledge-transition regions separating stable and unstable zones. To capture this phenomenon, we propose Lyapunov Probes: lightweight networks trained with derivative-based stability constraints that enforce a monotonic decay in confidence under input perturbations. By performing systematic perturbation analysis and applying a two-stage training process, these probes reliably distinguish between stable factual regions and unstable, hallucination-prone regions. Experiments on diverse datasets and models demonstrate consistent improvements over existing baselines.