KDCM: Reducing Hallucination in LLMs through Explicit Reasoning Structures

2026-01-07

arXiv.org

AI Summary

Large language models are prone to hallucinations induced by prompts. To address this issue, this work proposes a structured reasoning approach that integrates executable code embeddings with knowledge graph guidance. By embedding programmable modules directly into prompts, the method explicitly invokes external structured knowledge and constrains intermediate reasoning steps through a chain-of-knowledge distillation framework. This approach significantly reduces hallucination rates and achieves consistent performance gains across multiple public benchmarks, improving HIT@1, HIT@3, and HIT@5 by 15.64%, 13.38%, and 13.28%, respectively, with several metrics surpassing the 95% threshold.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Computer Vision: Large Vision Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Large language models for search

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.

Problem

• hallucination
• large language models
• prompt-induced errors
• reasoning reliability
• structured knowledge

Innovation

• hallucination reduction
• knowledge distillation
• reasoning structure
• code-guided reasoning
• structured knowledge