Mitigating Entity-Level Hallucination in Large Language Models

TL;DR

This paper introduces DRAD, a dynamic retrieval augmentation method that detects and mitigates hallucinations in large language models, significantly improving factual accuracy without external models.

Contribution

The paper presents DRAD, a novel approach combining real-time hallucination detection and self-correction using external knowledge to reduce hallucinations in LLMs.

Findings

DRAD outperforms existing methods in hallucination detection.

DRAD effectively reduces factual inaccuracies in LLM outputs.

Open-source code and data are provided for reproducibility.

Abstract

The emergence of Large Language Models (LLMs) has revolutionized how users access information, shifting from traditional search engines to direct question-and-answer interactions with LLMs. However, the widespread adoption of LLMs has revealed a significant challenge known as hallucination, wherein LLMs generate coherent yet factually inaccurate responses. This hallucination phenomenon has led to users' distrust in information retrieval systems based on LLMs. To tackle this challenge, this paper proposes Dynamic Retrieval Augmentation based on hallucination Detection (DRAD) as a novel method to detect and mitigate hallucinations in LLMs. DRAD improves upon traditional retrieval augmentation by dynamically adapting the retrieval process based on real-time hallucination detection. It features two main components: Real-time Hallucination Detection (RHD) for identifying potential hallucinations without external models, and Self-correction based on External Knowledge (SEK) for correcting these errors using external knowledge. Experiment results show that DRAD demonstrates superior performance in both detecting and mitigating hallucinations in LLMs. All of our code and data are open-sourced at https://github.com/oneal2000/EntityHallucination.