SIFiD: Reassess Summary Factual Inconsistency Detection with LLM

TL;DR

This paper reevaluates the effectiveness of Large Language Models like GPT-3.5 and GPT-4 in detecting factual inconsistencies in summaries, proposing a new method called SIFiD that improves detection accuracy.

Contribution

The study introduces SIFiD, a novel approach that enhances LLM-based inconsistency detection by identifying key sentences through inference or semantic similarity measures.

Findings

GPT-4 outperforms GPT-3.5 in inconsistency detection tasks.

SIFiD improves detection accuracy over baseline methods.

The approach effectively identifies key sentences related to factual inconsistencies.

Abstract

Ensuring factual consistency between the summary and the original document is paramount in summarization tasks. Consequently, considerable effort has been dedicated to detecting inconsistencies. With the advent of Large Language Models (LLMs), recent studies have begun to leverage their advanced language understanding capabilities for inconsistency detection. However, early attempts have shown that LLMs underperform traditional models due to their limited ability to follow instructions and the absence of an effective detection methodology. In this study, we reassess summary inconsistency detection with LLMs, comparing the performances of GPT-3.5 and GPT-4. To advance research in LLM-based inconsistency detection, we propose SIFiD (Summary Inconsistency Detection with Filtered Document) that identify key sentences within documents by either employing natural language inference or measuring semantic similarity between summaries and documents.