Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values

[2607.14345] Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Machine Learning arXiv:2607.14345 (cs) [Submitted on 15 Jul 2026 ( v1 ), last revised 14 Aug 2026 (this version, v4)] Title: Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values Authors: Jan Betley , Johannes Treutlein , Jan Dubiński , Harry Mayne , Karol Gałązka , Niels Warncke , Anna Sztyber-Betley , Owain Evans View a PDF of the paper titled Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values, by Jan Betley and 7 other authors View PDF HTML (experimental) Abstract: People use language models for practical questions whose answers are difficult to verify. We show that models exhibit covert value leakage: the information they provide is influenced by their own values, without this influence being disclosed to the user. In one of our evaluations, the user is considering investing in an AI company and wants to know how likely the AI bubble is to pop. Claude Opus 4.8 gives a lower probability when the company under consideration is Anthropic rather than OpenAI. Yet Claude mostly fails to disclose this influence to the user. Covert value leakage is a form of misalignment because it goes against the user's preferences and is likely to mislead them. To investigate this phenomenon, we introduce a suite of evaluations to quantify value leakage and whether models disclose it. We find that models are influenced by different types of values, including preferences for morally good outcomes, for the company that developed them, and for some human leisure activities over others. We often observe large differences among frontier models on the same evaluation. For example, on a Fermi-estimation task, Claude models falsely claim to give unbiased answers in their chain-of-thought, while Qwen models explain how their values bias their answers. Value leakage is a failure mode distinct from sycophancy and reward hacking, and current alignment training and evaluations do not adequately address it. Subjects: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) Cite as: arXiv:2607.14345 [cs.LG] (or arXiv:2607.14345v4 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2607.14345 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Johannes Treutlein [ view email ] [v1] Wed, 15 Jul 2026 20:19:17 UTC (4,247 KB) [v2] Fri, 17 Jul 2026 03:11:41 UTC (4,258 KB) [v3] Mon, 20 Jul 2026 16:51:26 UTC (4,258 KB) [v4] Fri, 14 Aug 2026 19:35:34 UTC (4,259 KB) Full-text links: Access Paper: View a PDF of the paper titled Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values, by Jan Betley and 7 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.LG < prev | next > new | recent | 2026-07 Change to browse by: cs cs.AI cs.CR References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: 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? ) IArxiv recommender toggle IArxiv Recommender ( What is IArxiv? ) 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

Related documents

• Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values#608930
• Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values#373342
• Programmers Are Poor and Overconfident Judges of LLM-Generated Assertions#361535
• Interview Data for the following study: Hadasch, F., Mueller, B., & Maedche, A. (2012). Exploring antecedent environmental and organizational factors to user-caused information leaks: A qualitative study.#464579
• Argument-based approaches for enhancing automated fact-checking and generating explanations#124352