Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Claim Verification in the Age of Large Language Models: A Survey

Alphaeus Dmonte, Roland Oruche, Marcos Zampieri, Prasad Calyam, Isabelle Augenstein

TL;DR

The paper addresses the problem of verifying claims in the era of large language models and pervasive online misinformation. It provides a comprehensive survey of LLM-based claim verification frameworks, detailing pipeline components such as retrieval, prompting, transfer learning, and generation, with a focus on retrieval-augmented generation (RAG). It catalogs public English datasets, metrics, and shared tasks, and discusses open challenges including irrelevant context, knowledge conflicts, and multilinguality, offering guidance for future research. The work serves as a foundational guide for researchers and practitioners aiming to build robust, explainable, and scalable fact-checking systems using LLMs.

Abstract

The large and ever-increasing amount of data available on the Internet coupled with the laborious task of manual claim and fact verification has sparked the interest in the development of automated claim verification systems. Several deep learning and transformer-based models have been proposed for this task over the years. With the introduction of Large Language Models (LLMs) and their superior performance in several NLP tasks, we have seen a surge of LLM-based approaches to claim verification along with the use of novel methods such as Retrieval Augmented Generation (RAG). In this survey, we present a comprehensive account of recent claim verification frameworks using LLMs. We describe the different components of the claim verification pipeline used in these frameworks in detail including common approaches to retrieval, prompting, and fine-tuning. Finally, we describe publicly available English datasets created for this task.

Claim Verification in the Age of Large Language Models: A Survey

TL;DR

The paper addresses the problem of verifying claims in the era of large language models and pervasive online misinformation. It provides a comprehensive survey of LLM-based claim verification frameworks, detailing pipeline components such as retrieval, prompting, transfer learning, and generation, with a focus on retrieval-augmented generation (RAG). It catalogs public English datasets, metrics, and shared tasks, and discusses open challenges including irrelevant context, knowledge conflicts, and multilinguality, offering guidance for future research. The work serves as a foundational guide for researchers and practitioners aiming to build robust, explainable, and scalable fact-checking systems using LLMs.

Abstract

The large and ever-increasing amount of data available on the Internet coupled with the laborious task of manual claim and fact verification has sparked the interest in the development of automated claim verification systems. Several deep learning and transformer-based models have been proposed for this task over the years. With the introduction of Large Language Models (LLMs) and their superior performance in several NLP tasks, we have seen a surge of LLM-based approaches to claim verification along with the use of novel methods such as Retrieval Augmented Generation (RAG). In this survey, we present a comprehensive account of recent claim verification frameworks using LLMs. We describe the different components of the claim verification pipeline used in these frameworks in detail including common approaches to retrieval, prompting, and fine-tuning. Finally, we describe publicly available English datasets created for this task.
Paper Structure (28 sections, 3 figures)

This paper contains 28 sections, 3 figures.

Table of Contents

  1. Introduction
  2. Search Criteria
  3. Claim Verification Pipeline
  4. Claim Detection
  5. Check-Worthy Claim Identification
  6. Claim Matching
  7. Document/Evidence Retrieval
  8. Rationale/Sentence Selection
  9. Veracity Label Prediction
  10. Explanation/Justification Generation
  11. LLM Approaches
  12. Evidence Retrieval Strategies
  13. Prompt Creation Strategies
  14. Transfer Learning Strategies
  15. Fine-Tuning.
  16. ...and 13 more sections

Figures (3)

  • Figure 1: Comparison of claim verification systems between NLP-based (traditional) and LLM-based for claim veracity.
  • Figure 2: A typical claim verification pipeline
  • Figure 3: LLM-based claim verification pipeline. This involves creating a prompt from the retrieved evidence and the input claim as input to the LLM to generate a label, sentence evidence, and/or explanation of its response.