IDEAS home Printed from https://ideas.repec.org/p/arx/papers/2605.23916.html

Advertising to Agents: Measuring Description-Framing Bias in LLM Tool Selection

Author

Listed:
  • Haochuan Kevin Wang

Abstract

AI agents often pick tools from registries, where each provider writes its tool's description; we ask whether sales language there changes which tool an agent picks. We built pairs of listings that differ in one way (praise, a verifiable specification, or list order) and asked two OpenAI models to call one tool. In a preregistered study, stacked praise (four kinds) raised a tool's pick rate by about 43 percentage points, matching or beating a verifiable specification. Praise also pulled some picks toward tools whose numeric limit was too small. With identical listings, the first-listed tool was picked about 72 points more often. Structured fields helped agents pick the capable tool; sales text beside them reduced or erased that gain. An exploratory study of five models shows the same direction; a "Sponsored" label dampened it, and a system-prompt warning moved one model. Registries could list limits as fields, hide sales text from agents, and randomize order. Stacked praise replicated on held-out domains; no single kind did after correcting for multiple tests. Preregistered results are provisional until blind phrase ratings are complete, and cover two small models.

Suggested Citation

  • Haochuan Kevin Wang, 2026. "Advertising to Agents: Measuring Description-Framing Bias in LLM Tool Selection," Papers 2605.23916, arXiv.org, revised Oct 2026.
    • Handle: RePEc:arx:papers:2605.23916
    as

    Download full text from publisher

    File URL: https://arxiv.org/pdf/2605.23916
    File Function: Latest version
    Download Restriction: no
    ---><---

    References listed on IDEAS

    as
    1. Milgrom, Paul R & Weber, Robert J, 1982. "A Theory of Auctions and Competitive Bidding," Econometrica, Econometric Society, vol. 50(5), pages 1089-1122, September.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Hongpeng Guo & Zhihao Lv & Junyi Hua & Hongxu Yuan & Qingyu Yu, 2021. "Design of Combined Auction Model for Emission Rights of International Forestry Carbon Sequestration and Other Pollutants Based on SMRA," Sustainability, MDPI, vol. 13(20), pages 1-18, October.
    2. Yusuke Matsuki, 2016. "A Distribution-Free Test of Monotonicity with an Application to Auctions," Working Papers e110, Tokyo Center for Economic Research.
    3. Vinci Chow, 2019. "Predicting Auction Price of Vehicle License Plate with Deep Residual Learning," Papers 1910.04879, arXiv.org.
    4. Yeon-Koo Che & Ian Gale, 1994. "Auctions with budget-constrained buyers: a nonequivalence result," Working Papers (Old Series) 9402, Federal Reserve Bank of Cleveland.
    5. Kathryn M. Dominguez, 1991. "Do Exchange Auctions Work? An Examination of the Bolivian Experience," NBER Working Papers 3683, National Bureau of Economic Research, Inc.
    6. Giuseppe Topa & Alain Karsenty & Carole Megevand & Laurent Debroux, 2009. "The Rainforests of Cameroon : Experience and Evidence from a Decade of Reform," World Bank Publications - Books, The World Bank Group, number 2683.
    7. Rasmusen Eric Bennett, 2006. "Strategic Implications of Uncertainty over One's Own Private Value in Auctions," The B.E. Journal of Theoretical Economics, De Gruyter, vol. 6(1), pages 1-24, November.
    8. Thomas D. Jeitschko, 1998. "Learning in Sequential Auctions," Southern Economic Journal, John Wiley & Sons, vol. 65(1), pages 98-112, July.
    9. Christian Gollier & James Hammitt & Nicolas Treich, 2013. "Risk and choice: A research saga," Journal of Risk and Uncertainty, Springer, vol. 47(2), pages 129-145, October.
    10. Renato Gomes & Alessandro Pavan, 2013. "Cross-Subsidization and Matching Design," Discussion Papers 1559, Northwestern University, Center for Mathematical Studies in Economics and Management Science.
    11. Dirk Bergemann & Marek Bojko & Paul DŸtting & Renato Paes Leme & Haifeng Xu & Song Zuo, 2024. "Data-Driven Mechanism Design: Jointly Eliciting Preferences and Information," Cowles Foundation Discussion Papers 2418, Cowles Foundation for Research in Economics, Yale University.
    12. Jinpeng Ma, 1997. "English Auctions and Walrasian Equilibria with Multiple Objects: a dynamic approach," Departmental Working Papers 199702, Rutgers University, Department of Economics.
    13. Vladimirov, Vladimir, 2015. "Financing bidders in takeover contests," Journal of Financial Economics, Elsevier, vol. 117(3), pages 534-557.
    14. Khurshid Alam & Sakiba Tasneem & Molla Huq, 2014. "Reservation wage of female volunteer community health workers in Dhaka urban slums: a bidding game approach," Health Economics Review, Springer, vol. 4(1), pages 1-10, December.
    15. Chi, Chang Koo & Murto, Pauli & Valimaki, Juuso, 2017. "All-Pay Auctions with Affiliated Values," MPRA Paper 80799, University Library of Munich, Germany.
    16. Ming Chen & Sareh Nabi & Marciano Siniscalchi, 2023. "Advancing Ad Auction Realism: Practical Insights & Modeling Implications," Papers 2307.11732, arXiv.org, revised Apr 2024.
    17. de Frutos, Maria-Angeles & Pechlivanos, Lambros, 2006. "Second-price common-value auctions under multidimensional uncertainty," Games and Economic Behavior, Elsevier, vol. 55(1), pages 43-71, April.
    18. John H. Kagel & Colin M. Campbell & Dan Levin, 1999. "The Winner's Curse and Public Information in Common Value Auctions: Reply," American Economic Review, American Economic Association, vol. 89(1), pages 325-334, March.
    19. Malueg, David A. & Orzach, Ram, 2009. "Revenue comparison in common-value auctions: Two examples," Economics Letters, Elsevier, vol. 105(2), pages 177-180, November.
    20. Nyborg, Kjell G. & Sundaresan, Suresh, 1996. "Discriminatory versus uniform Treasury auctions: Evidence from when-issued transactions," Journal of Financial Economics, Elsevier, vol. 42(1), pages 63-104, September.

    More about this item

    NEP fields

    This paper has been announced in the following NEP Reports:

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:arx:papers:2605.23916. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: arXiv administrators (email available below). General contact details of provider: https://arxiv.org/ .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.