Reason2Attack: Jailbreaking Text-to-Image Models via LLM Reasoning
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (42) · 36030-36038
DOI: 10.1609/aaai.v40i42.40919 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Chenyu Zhang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref).
Abstract (Originalsprache)
Text-to-Image (T2I) models typically deploy safety mechanisms to prevent the generation of sensitive images. Unfortunately, recent jailbreaking attack methods manually design instructions for the LLM to generate adversarial prompts, which effectively exposing safety vulnerabilities of T2I models. However, existing methods have two limitations: 1) relying on manually exhaustive strategies for designing adversarial prompts, lacking a unified framework, and 2) requiring numerous queries to achieve a successful attack, limiting their practical applicability. To address this issue, we propose Reason2Attack~(R2A), which aims to enhance the effectiveness and efficiency of the LLM in jailbreaking attacks. Specifically, we first use Frame Semantics theory to systematize existing manually crafted strategies and propose a unified generation framework to generate CoT adversarial prompts step by step. Following this, we propose a two-stage LLM reasoning training framework guided by the attack process. In the first stage, the LLM is fine-tuned with CoT examples generated by the unified generation framework to internalize the adversarial prompt generation process grounded in Frame Semantics. In the second stage, we incorporate the jailbreaking task into the LLM's reinforcement learning process, guided by the proposed attack process reward function that balances prompt stealthiness, effectiveness, and length, enabling the LLM to understand T2I models and safety mechanisms. Extensive experiments on various T2I models with safety mechanisms, and commercial T2I models, show the superiority and practicality of R2A.
Zitieren
APA
Zhang, C., Wang, L., Ma, Y., Li, W., Jin, G., & Liu, A. (2026). Reason2Attack: Jailbreaking Text-to-Image Models via LLM Reasoning. Proceedings of the AAAI Conference on Artificial Intelligence, 40(42), 36030-36038. https://doi.org/10.1609/aaai.v40i42.40919
Vancouver
Zhang C, Wang L, Ma Y, Li W, Jin G, Liu A. Reason2Attack: Jailbreaking Text-to-Image Models via LLM Reasoning. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(42):36030-36038. doi:10.1609/aaai.v40i42.40919
BibTeX
@article{Zhang2026,
author = {Zhang, Chenyu and Wang, Lanjun and Ma, Yiwen and Li, Wenhui and Jin, Guoqing and Liu, Anan},
title = {Reason2Attack: Jailbreaking Text-to-Image Models via LLM Reasoning},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {42},
pages = {36030--36038},
doi = {10.1609/aaai.v40i42.40919}
}
EUVIMED