Auditing LLM Benchmarks with Item Response Theory

[2605.30504] Auditing LLM Benchmarks with Item Response Theory Skip to main content Search Submit Donate Log in Computer Science > Computation and Language arXiv:2605.30504 (cs) [Submitted on 28 May 2026 ( v1 ), last revised 28 Aug 2026 (this version, v2)] Title: Auditing LLM Benchmarks with Item Response Theory Authors: Sander Land , Daniel M. Bikel View PDF HTML (experimental) Abstract: LLM benchmark labels are frozen at release and silently propagated into downstream benchmarks, errors and all. We introduce an Item Response Theory-based indicator that surfaces likely mislabels at 95% precision in the top 200 examples across seven preference and multiple-choice benchmarks using responses from 114 models, outperforming a supervised classifier. We trace these errors to mechanical labeling heuristics, upstream annotation mistakes inherited unchanged from source datasets, and fundamentally ambiguous items without a defensible single label. The same model fit reveals that reward models specialize in stylistic preference rather than factual knowledge, and identifies one frontier reward model that agrees with detected mislabels at 78% accuracy versus 38% for its peers, consistent with benchmark contamination or benchmark-specific over-optimization. Comments: Accepted at EMNLP 2026. Associated data at this https URL Subjects: Computation and Language (cs.CL) Cite as: arXiv:2605.30504 [cs.CL] (or arXiv:2605.30504v2 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2605.30504 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Sander Land [ view email ] [v1] Thu, 28 May 2026 19:38:44 UTC (421 KB) [v2] Fri, 28 Aug 2026 07:46:20 UTC (429 KB) Full-text links: Access Paper: View PDF HTML (experimental) TeX Source view license Additional Features Audio Summary Current browse context: cs.CL < prev | next > new | recent | 2026-05 Change to browse by: cs References & Citations SciX Google Scholar Semantic Scholar export BibTeX citation Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators 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 . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from (opens in new tab) (opens in new tab) (opens in new tab)

Related documents

• Auditing LLM Benchmarks with Item Response Theory#682701
• Quantifying and Auditing LLM Evaluation via Positive--Unlabeled Learning#287155
• ReproRepo: Scaling Reproducibility Audits with GitHub Repository Issues#820515
• From Evaluative Structure to Reviewability: A Minimal Analytical Schema#196677
• More Programs or More Rolls? Separating Coverage from Specialization in LLM Harnesses#1130715