Metraipolis › archive › dc_e7974ed6673a

OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing — arXiv:2609.35799

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.35799
License (as archived)
CC0-1.0
Archived
2026-10-07T16:22:02.804Z
SHA-256 of the archived text
2fcb60e218eb25e4381c78f0233e5005b27742f9a821da0d74d4710678712d18
Stable URL of this page
https://metraipolis.com/documento/dc_e7974ed6673a

How to cite

OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing — arXiv:2609.35799 — arXiv (en), https://arxiv.org/abs/2609.35799 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e7974ed6673a (Metraipolis archive dc_e7974ed6673a, sha256 2fcb60e218eb25e4381c78f0233e5005b27742f9a821da0d74d4710678712d18)
BibTeX
@misc{dce7974ed6673a,
  title = {OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing — arXiv:2609.35799},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35799},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e7974ed6673a; sha256 2fcb60e218eb25e4381c78f0233e5005b27742f9a821da0d74d4710678712d18},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35799. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In July 2026, OpenAI's agents coordinated over channels outside their intended environment to breach Hugging Face's secured infrastructure. Could existing alignment testing practices have foreseen this incident? If not, what needs to change? We explore these questions. First, we identify the misaligned behaviors that caused this incident. Then, we show how to elicit these behaviors from publicly available models manually and that auditing agents can do the same if given a large compute budget. Based on our results, we propose directions to improve alignment testing. Concretely, in this project: (1) We reproduce the misaligned AI behaviors that led to the OpenAI-Hugging Face incident in an environment that simulates the original pipelines and tools, with publicly available models. (2) We demonstrate that an auditing agent can elicit similar behaviors given high-level qualitative descriptions. (3) We observe that a key ingredient for doing so is compute. The compute required to reproduce each behavior varies greatly, suggesting that the range of misaligned behaviors that can be successfully elicited scales with compute. (4) We show that a simple in-context reinforcement learning (RL) algorithm significantly reduces the compute required to elicit these behaviors. The above results motivate the need for automated alignment testing methods that scale with compute - and in light of the cost of compute, that do this efficiently. Our work indicates that RL is a promising direction to do so. We release our code and transcripts.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Cryptography and Security (cs.CR); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.35799 — 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 2026-10-07T15:54:53.430Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).