[2503.03750] The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems

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. arXiv Operational Status Get status notifications via email or slack

[2503.03750] The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems Skip to main content arXiv is now an independent nonprofit! Learn more × Search arXiv Press Enter to search · Advanced search --> Computer Science > Machine Learning arXiv:2503.03750 (cs) [Submitted on 5 Mar 2025 ( v1 ), last revised 5 Jan 2026 (this version, v3)] Title: The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems Authors: Richard Ren , Arunim Agarwal , Mantas Mazeika , Cristina Menghini , Robert Vacareanu , Brad Kenstler , Mick Yang , Isabelle Barrass , Alice Gatti , Xuwan

related reading

• [2503.03750] The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systemsarxiv.org
• [2401.05566] Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Trainingarxiv.org
• [2503.10965] Auditing language models for hidden objectivesarxiv.org
• [2503.07003] Large Language Models Often Say One Thing and Do Anotherarxiv.org
• [2506.09501] Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inferencearxiv.org
• [2506.00751] Alignment Revisited: Are Large Language Models Consistent in Stated and Revealed Preferences?arxiv.org
• 2310.01405arxiv.org
• 2310.01405.pdfarxiv.org
• [2008.02275] Aligning AI With Shared Human Valuesarxiv.org
• [2602.22755] AuditBench: Evaluating Alignment Auditing Techniques on Models with Hidden Behaviorsarxiv.org
• [2406.02061] Alice in Wonderland: Simple Tasks Showing Complete Reasoning Breakdown in State-Of-the-Art Large Language Modelsarxiv.org
• [2505.05410] Reasoning Models Don't Always Say What They Thinkarxiv.org