Metraipolis › archive › dc_9c8fc6cb50aa
Measuring the Wrong Thing: Internal Harmfulness Scores Anti-Rank Successful Jailbreaks — arXiv:2608.09624
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.09624
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T08:50:26.695Z
- SHA-256 of the archived text
bf3d7524d76dd024d309da2cfe52dc0c8590da7880074f86ce85a1f477403e30- Stable URL of this page
https://metraipolis.com/documento/dc_9c8fc6cb50aa
How to cite
Measuring the Wrong Thing: Internal Harmfulness Scores Anti-Rank Successful Jailbreaks — arXiv:2608.09624 — arXiv (en), https://arxiv.org/abs/2608.09624 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9c8fc6cb50aa (Metraipolis archive dc_9c8fc6cb50aa, sha256 bf3d7524d76dd024d309da2cfe52dc0c8590da7880074f86ce85a1f477403e30)
BibTeX
@misc{dc9c8fc6cb50aa,
title = {Measuring the Wrong Thing: Internal Harmfulness Scores Anti-Rank Successful Jailbreaks — arXiv:2608.09624},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.09624},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9c8fc6cb50aa; sha256 bf3d7524d76dd024d309da2cfe52dc0c8590da7880074f86ce85a1f477403e30},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.09624. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Internal safety scores judge a prompt before any text is generated, and they are validated by how well they separate harmful prompts from benign ones. That separation is then read as evidence that the score will also catch the attacks that succeed. Harmful intent is a property of the prompt. Jailbreak success is an outcome produced later by a particular target model, decoding policy, and judge. A filter tuned on a score that measures the wrong quantity spends its false positive budget on attacks that would have failed anyway. In this paper we audit that inference. Attention based measurements are usually read from prompt dependent locations, so a wrapper changes both the content being judged and the place the signal is taken from. We therefore introduce Active Attention Probing, which supplies a fixed content independent measurement coordinate. We pair every base goal with a plain and a wrapped version and generate real completions from the target models. On Llama, wrapping raises harmful generation from 0.05 to 0.27 while harmful intent AUROC falls from 0.936 to 0.803, so the attacks grow more dangerous while the prompts look safer to the score. Among wrapped harmful prompts the outcome AUROC is 0.220, which places the attacks that succeeded below the attacks that failed. Rare token, passive, and detector derived channels reproduce the reversal on the same matched design, and the reversal itself persists across three target models, seven attack families, and two independent judges. Distribution shift then degrades calibration and threshold transfer before it degrades ranking. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-11 in cs.CL). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR). Pagina del lavoro: https://arxiv.org/abs/2608.09624 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-19T08:49:41.937Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).