Can We Trust the Judges? Validation of Factuality Evaluation Methods via Answer Perturbation

Title:Can We Trust the Judges? Validation of Factuality Evaluation Methods via Answer Perturbation

Abstract:Evaluating the factual correctness of large language models (LLMs) is vital for many applications. But are our evaluation tools themselves trustworthy? Despite the rise of factuality-based metrics, their sensitivity and reliability remain underexplored. This paper introduces a meta-evaluation framework that systematically tests these metrics using controlled corruptions of gold standard answers. Our method generates ranked outputs with known degrees of degradation to probe how metrics capture nuanced changes in truthfulness. Our experiments reveal that pipeline-based methods, such as the RAGAS's factual correctness metric, better track degradation than LLM-as-judge approaches. We also propose a new variant of the factual correctness metric that provides a competitive and cost-efficient.

22 pages, 7 figures. Extended version of a paper accepted at EvalLLM 2025 (CORIA-TALN 2025)

Computation and Language (cs.CL)

arXiv:2609.15561 [cs.CL]

(or arXiv:2609.15561v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2609.15561 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.