Can a Small Model Learn to Look Before It Leaps? Dynamic Learning and Proactive Correction for Hallucination Detection

TL;DR

This paper introduces LEAP, a dynamic learning framework that enables small language models to adaptively detect and correct hallucinations, significantly improving reliability and efficiency over fixed-strategy methods.

Contribution

LEAP is the first framework to enable small models to learn dynamic verification strategies for hallucination detection, guided by a powerful teacher model and a proactive correction mechanism.

Findings

LEAP outperforms state-of-the-art methods on three benchmarks.

Dynamic strategy learning improves hallucination detection accuracy.

Proactive correction enhances model reliability and adaptability.

Abstract

Hallucination in large language models (LLMs) remains a critical barrier to their safe deployment. For hallucination detection to be practical in real-world scenarios, the use of efficient small models is essential to ensure low latency and minimal resource consumption. However, existing methods rely on fixed verification strategies, where simply tuning small models to mimic fixed verification trajectories fails to capture the adaptability required for diverse hallucination patterns, thereby inducing planning instability. To address this limitation, we propose a ``Learning to Evaluate and Adaptively Plan'' (LEAP) framework, which shifts hallucination detection from fixed execution to dynamic strategy learning. Specifically, LEAP first employs a powerful teacher model to iteratively explore and refine verification strategies through a failure-driven loop. This dynamic planning capability is then distilled into an efficient student model, augmented by a novel proactive correction mechanism that enables the model to evaluate and optimize its verification strategy before execution. Experiments on three benchmarks demonstrate that LEAP outperforms state-of-the-art methods, offering an effective and scalable solution for reliable hallucination detection.