Towards Comprehensive Stage-wise Benchmarking of Large Language Models in Fact-Checking

2026-01-06

arXiv.org

AI Summary

Current fact-checking evaluations are largely confined to claim verification, neglecting critical upstream components such as claim extraction and evidence retrieval, thereby failing to comprehensively assess large language models’ systematic reasoning capabilities and factual robustness. To address this limitation, this work proposes FactArena—the first competitive evaluation framework encompassing the entire fact-checking pipeline. FactArena enables automated assessment across all stages—claim decomposition, evidence acquisition, and judgment reasoning—through LLM-driven process standardization, tool-augmented evidence retrieval, a multi-agent adjudication consensus mechanism, and semantically controlled adversarial claim generation. Experiments on 16 mainstream large language models reveal a significant gap between end-to-end fact-checking performance and static verification accuracy, underscoring the necessity and efficacy of holistic, full-pipeline evaluation.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Large Language Models (LLMs) are increasingly deployed in real-world fact-checking systems, yet existing evaluations focus predominantly on claim verification and overlook the broader fact-checking workflow, including claim extraction and evidence retrieval. This narrow focus prevents current benchmarks from revealing systematic reasoning failures, factual blind spots, and robustness limitations of modern LLMs. To bridge this gap, we present FactArena, a fully automated arena-style evaluation framework that conducts comprehensive, stage-wise benchmarking of LLMs across the complete fact-checking pipeline. FactArena integrates three key components: (i) an LLM-driven fact-checking process that standardizes claim decomposition, evidence retrieval via tool-augmented interactions, and justification-based verdict prediction; (ii) an arena-styled judgment mechanism guided by consolidated reference guidelines to ensure unbiased and consistent pairwise comparisons across heterogeneous judge agents; and (iii) an arena-driven claim-evolution module that adaptively generates more challenging and semantically controlled claims to probe LLMs'factual robustness beyond fixed seed data. Across 16 state-of-the-art LLMs spanning seven model families, FactArena produces stable and interpretable rankings. Our analyses further reveal significant discrepancies between static claim-verification accuracy and end-to-end fact-checking competence, highlighting the necessity of holistic evaluation. The proposed framework offers a scalable and trustworthy paradigm for diagnosing LLMs'factual reasoning, guiding future model development, and advancing the reliable deployment of LLMs in safety-critical fact-checking applications.

Problem

• fact-checking
• large language models
• benchmarking
• evaluation framework
• reasoning failures

Innovation

• stage-wise benchmarking
• arena-style evaluation
• fact-checking pipeline
• tool-augmented reasoning
• claim evolution