Mean Accuracy Is the Wrong Estimand for a Release Decision

Abstract:A release decision about a fine-tuned model is an operational commitment about the \emph{next} rerun of its training recipe. Yet standard benchmark reporting gives only the mean accuracy of past seeds. These target fundamentally different questions: release decisions depend on catastrophic tail risk, whereas standard evaluation reports only central tendency. Because fine-tuning induces bimodal mode collapse, high average accuracy frequently conceals severe failure rates. Auditing $2\,124$ fine-tuned models across eleven language models, we show that shipping models based on mean accuracy violates a $10\%$ deployment failure budget $48.0\%$ of the time. Common heuristics like $\bar b - 2\sigma$ prove equally dangerous: by underestimating variance on small samples, they become most permissive when evidence is weakest. We introduce \textbf{recipe cards}, an evaluation framework that directly certifies rerun tail failure rates using exact finite-sample binomial bounds. Recipe cards reduce held-out budget violations to $0.6\%$, enforce a mathematical sample floor ($N \ge 22$) that prevents certification on thin evidence, and explicitly refuse unverified claims under domain and task shifts.

Machine Learning (cs.LG)

arXiv:2610.16825 [cs.LG]

(or arXiv:2610.16825v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2610.16825 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.