Metraipolis › archive › dc_30ea5695b8cf
RL-ATTACK: A Generalized Black-Box Adversarial Attack on Language Models via Reinforcement Learning — arXiv:2611.08711
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.08711
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-09T22:35:51.461Z
- SHA-256 of the archived text
7e3503a5c4117397f0fb7add5b2b16e30444cae71b877cb5ef4f875c7d081c8f- Stable URL of this page
https://metraipolis.com/documento/dc_30ea5695b8cf
How to cite
RL-ATTACK: A Generalized Black-Box Adversarial Attack on Language Models via Reinforcement Learning — arXiv:2611.08711 — arXiv (en), https://arxiv.org/abs/2611.08711 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_30ea5695b8cf (Metraipolis archive dc_30ea5695b8cf, sha256 7e3503a5c4117397f0fb7add5b2b16e30444cae71b877cb5ef4f875c7d081c8f)
BibTeX
@misc{dc30ea5695b8cf,
title = {RL-ATTACK: A Generalized Black-Box Adversarial Attack on Language Models via Reinforcement Learning — arXiv:2611.08711},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.08711},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_30ea5695b8cf; sha256 7e3503a5c4117397f0fb7add5b2b16e30444cae71b877cb5ef4f875c7d081c8f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.08711. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) exhibit vulnerabilities to adversarial perturbations that can induce unintended responses. These vulnerabilities can be identified via intentional attacks and subsequently used to improve model safety. In black-box settings where gradients are inaccessible, existing methods tend to guess adversarial transformations using external knowledge, ignoring the information provided by the target model. To mitigate this, we propose a trainable black-box framework based on reinforcement learning, which learns from the target model to generate tailored adversarial transformations. We demonstrate that our approach achieves lower post-attack accuracy, greater query efficiency and better semantic preservation on both natural language inference and generative Q&A tasks. On natural language inference, it drives post-attack accuracy to 2% within 20 queries. Against generative language models, our trained attacker consistently achieves lower post-attack accuracy than the untrained attacker and wins in 4 out of 6 settings when compared to a strong optimization-based baseline. We further experiment with multiple reward designs and training settings, and we find that no design or setting is consistently dominant. 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: Computation and Language (cs.CL) ; Cryptography and Security (cs.CR). Pagina del lavoro: https://arxiv.org/abs/2611.08711 — 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 2026-11-09T22:31:01.941Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).