Metraipolis › archive › dc_ca116a7c7158

AuxMark: Defending Against Unauthorized Agent Distillation via Auxiliary Behavioral Watermarking — arXiv:2609.34597

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.34597
License (as archived): CC0-1.0
Archived: 2026-10-07T22:34:49.037Z
SHA-256 of the archived text: eb3e4dea6ff870697a11e0ab65d13b05a2a6b19199805525e58d50e8c4e2023a
Stable URL of this page: https://metraipolis.com/documento/dc_ca116a7c7158

How to cite

AuxMark: Defending Against Unauthorized Agent Distillation via Auxiliary Behavioral Watermarking — arXiv:2609.34597 — arXiv (en), https://arxiv.org/abs/2609.34597 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ca116a7c7158 (Metraipolis archive dc_ca116a7c7158, sha256 eb3e4dea6ff870697a11e0ab65d13b05a2a6b19199805525e58d50e8c4e2023a)

BibTeX

@misc{dcca116a7c7158,
  title = {AuxMark: Defending Against Unauthorized Agent Distillation via Auxiliary Behavioral Watermarking — arXiv:2609.34597},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.34597},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ca116a7c7158; sha256 eb3e4dea6ff870697a11e0ab65d13b05a2a6b19199805525e58d50e8c4e2023a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34597. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language model agents can acquire complex capabilities through multi-step interaction and tool use, but their trajectories can also be illegally collected to dis- till student agents. However, existing watermarking methods either do not fit the structured and interactive nature of agent environments or lack reliable effective- ness across tasks and model architectures. We introduce AuxMark, a behavioral watermarking framework for tracing unauthorized agent distillation. AuxMark dynamically inserts safe, non-essential auxiliary action into teacher trajectories, and stores the associated contexts as private evidence cards. To audit a suspicious student model, AuxMark constructs paired real and fake probes from these cards and applies a card-level sign test. This black-box protocol supports both model- level detection and trace-level attribution. Across three agent benchmarks, two teacher agents, and four student architectures, AuxMark detects all 24 distilled models with zero false positives on 48 clean models. It also preserves task utility and remains effective against data flooding, paraphrasing, truncation, and adaptive cleaning attacks. Our code will be released at this URL.

Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CR).
Categorie arXiv: Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2609.34597 — 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 2026-10-07T22:13:03.723Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).