# Acceptance Cards:A Four-Diagnostic Standard for Safe Fine-Tuning Defense Claims

DevFeed: [Acceptance Cards:A Four-Diagnostic Standard for Safe Fine-Tuning Defense Claims](<https://devfeed.tech/articles/acceptance-cards-a-four-diagnostic-standard-for-safe-fine-tuning-defense-claims-113988.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.10575>)

Author: Phongsakon Mark Konrad, Toygar Tanyel, Serkan Ayvaz

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [benchmark overfitting machine learning](<https://devfeed.tech/topics/benchmark-overfitting-machine-learning.md>), [gemma](<https://devfeed.tech/topics/gemma.md>)

Tags: [acceptance](<https://devfeed.tech/tags/acceptance.md>), [audit](<https://devfeed.tech/tags/audit.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [defense](<https://devfeed.tech/tags/defense.md>), [diagnostics](<https://devfeed.tech/tags/diagnostics.md>), [documentation](<https://devfeed.tech/tags/documentation.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [gap](<https://devfeed.tech/tags/gap.md>), [mechanism](<https://devfeed.tech/tags/mechanism.md>), [model](<https://devfeed.tech/tags/model.md>), [noise](<https://devfeed.tech/tags/noise.md>), [reliability](<https://devfeed.tech/tags/reliability.md>), [standard](<https://devfeed.tech/tags/standard.md>)

## AI overview

Acceptance Cards requires four checks before treating a safe fine-tuning defense's held-out gap reduction as evidence: statistical reliability, fresh semantic generalization, mechanism alignment, and cross-task transfer. In the reported installed-gap audit, SafeLoRA did not pass all four checks on Gemma-2-2B-it; it failed all four under strict mechanism-class coding and three under a permissive shrinkage relabel. None of 46 audited cells met the strict conjunction. The closest family passed reliability and mechanism checks where data were available, but missed the fresh-subject threshold, lacked a strict transfer pass, and incurred a measurable deployment-accuracy cost. The authors characterize this as a narrow audit on one model family, not a global judgment of SafeLoRA's effectiveness.

## Source excerpt

arXiv:2605.10575v2 Announce Type: replace Abstract: Safe fine-tuning defenses are often endorsed on the basis of a held-out gap reduction, but the same reduction can come from sampling noise, subject artifacts, capability loss, or a mechanism that does not transfer. We introduce Acceptance Cards: an evaluation protocol, a documentation object, an executable audit package, and a claim-specific evidential standard for safe fine-tuning defense claims. The protocol checks statistical reliability, fresh semantic generalization, mechanism alignment, and cross-task transfer before treating a gap reduction as a full-card pass. Re-scored under this installed-gap protocol, SafeLoRA fails the full-card pass on Gemma-2-2B-it: under strict mechanism-class coding it fails all four diagnostics, and under a permissive shrinkage relabel it still fails three of four. This is a narrow installed-gap audit on one model family, not a global judgment of SafeLoRA's effectiveness. In a 46-cell audit, no cell satisfies the strict conjunction. The closest family is a near miss that passes reliability and mechanism checks where the required data are available, but fails the fresh-subject threshold, lacks a strict transfer pass, and carries a measurable deployment-accuracy cost.