HalluGuard: Demystifying Data-Driven and Reasoning-Driven Hallucinations in LLMs
This addresses reliability issues in high-stakes domains like healthcare and law by providing a generalized detection approach for hallucinations, though it builds incrementally on existing NTK-based methods.
The paper tackles the problem of hallucinations in Large Language Models (LLMs) by introducing a unified theoretical framework and a detection method called HalluGuard, which achieves state-of-the-art performance across 10 benchmarks, 11 baselines, and 9 LLM backbones.
The reliability of Large Language Models (LLMs) in high-stakes domains such as healthcare, law, and scientific discovery is often compromised by hallucinations. These failures typically stem from two sources: data-driven hallucinations and reasoning-driven hallucinations. However, existing detection methods usually address only one source and rely on task-specific heuristics, limiting their generalization to complex scenarios. To overcome these limitations, we introduce the Hallucination Risk Bound, a unified theoretical framework that formally decomposes hallucination risk into data-driven and reasoning-driven components, linked respectively to training-time mismatches and inference-time instabilities. This provides a principled foundation for analyzing how hallucinations emerge and evolve. Building on this foundation, we introduce HalluGuard, an NTK-based score that leverages the induced geometry and captured representations of the NTK to jointly identify data-driven and reasoning-driven hallucinations. We evaluate HalluGuard on 10 diverse benchmarks, 11 competitive baselines, and 9 popular LLM backbones, consistently achieving state-of-the-art performance in detecting diverse forms of LLM hallucinations.