KDCM: Reducing Hallucination in LLMs through Explicit Reasoning Structures

TL;DR

This paper introduces KDCM, a framework that reduces hallucinations in large language models by integrating explicit reasoning structures and external knowledge through programmable modules within prompts, enhancing reliability and interpretability.

Contribution

The paper presents a novel method that extends knowledge distillation with programmable modules embedded in reasoning prompts to explicitly regulate intermediate reasoning steps in LLMs.

Findings

Significant improvements in HIT@1, HIT@3, and HIT@5 metrics.

Scores exceeding 95% on multiple benchmarks.

Effective reduction of prompt-induced hallucinations.

Abstract

To mitigate hallucinations in large language models (LLMs), we propose a framework that focuses on errors induced by prompts. Our method extends a chain-style knowledge distillation approach by incorporating a programmable module that guides knowledge graph exploration. This module is embedded as executable code within the reasoning prompt, allowing the model to leverage external structured knowledge during inference. Based on this design, we develop an enhanced distillation-based reasoning framework that explicitly regulates intermediate reasoning steps, resulting in more reliable predictions. We evaluate the proposed approach on multiple public benchmarks using GPT-4 and LLaMA-3.3. Experimental results show that code-guided reasoning significantly improves contextual modeling and reduces prompt-induced hallucinations. Specifically, HIT@1, HIT@3, and HIT@5 increase by 15.64%, 13.38%, and 13.28%, respectively, with scores exceeding 95% across several evaluation settings. These findings indicate that the proposed method effectively constrains erroneous reasoning while improving both accuracy and interpretability.