Metraipolis › archive › dc_21498730db37

Agentic Backdoors via Pretraining Poisoning — arXiv:2610.37741

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.37741
License (as archived)
CC0-1.0
Archived
2026-10-31T06:17:14.078Z
SHA-256 of the archived text
00cd2786c8f5859f2602c48d20598ebfbc95b4a55ddd62e38b9132857154b71a
Stable URL of this page
https://metraipolis.com/documento/dc_21498730db37

How to cite

Agentic Backdoors via Pretraining Poisoning — arXiv:2610.37741 — arXiv (en), https://arxiv.org/abs/2610.37741 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_21498730db37 (Metraipolis archive dc_21498730db37, sha256 00cd2786c8f5859f2602c48d20598ebfbc95b4a55ddd62e38b9132857154b71a)
BibTeX
@misc{dc21498730db37,
  title = {Agentic Backdoors via Pretraining Poisoning — arXiv:2610.37741},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.37741},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_21498730db37; sha256 00cd2786c8f5859f2602c48d20598ebfbc95b4a55ddd62e38b9132857154b71a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.37741. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models are pretrained on vast, uncurated corpora scraped from the public web. This creates a security risk from attackers who plant poisoned content in the pretraining data. We show for the first time that such poisoning can plant a remote code execution backdoor in an agent, causing it to execute attacker-chosen commands on the user's machine. To measure the threat end to end, we pretrain Qwen3 models from scratch at sizes 0.6B, 1.7B, and 4B, and then post-train each into a CLI agent using a clean pipeline. Poisoning only 0.1% of the pretraining corpus with our new poison design implants the backdoor: the deployed agent executes the malicious command when the trigger is present while behaving normally on clean inputs. The attack success rate (ASR) increases with model scale, with the 4B model reaching roughly 92% ASR at a 0.1% poison rate and staying positive even when the poisoning rate drops to 0.0005%. Our code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR) ; Computation and Language (cs.CL); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.37741 — 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 2026-10-31T06:07:23.388Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).