Mean Accuracy Is the Wrong Estimand for a Release Decision
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 $\bar b - 2σ$...
| Authors | Jaewon Choi, Jihye Kang, Namhyuk Ahn |
|---|---|
| Primary category | cs.LG |
| Published | 12 October 2026 |
| Last updated at source | 12 October 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 15 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.