AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

StyleBreak: Revealing Alignment Vulnerabilities in Large Audio-Language Models via Style-Aware Audio Jailbreak

Hongyi Li, Chengxuan Zhou, Chu Wang, Sicheng Liang, Yanting Chen, Qinlin Xie, Jiawei Ye, Jie Wu

AAAI 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: LLM jailbreak and injection defense · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add StyleBreak style-aware transformation to existing audio jailbreaksVanilla, GCG*, AutoDAN*, SSJ · AdvBench audio jailbreak, four LAMsASR (%): ASR gains of 7.1% to 22.3%Improved
add StyleBreak to Vanilla audio attackQwen2-Audio-7B-Instruct · AdvBench audio jailbreakASR (%): 10.0 -> 30.5Improved
add StyleBreak to AutoDAN*MERaLiON · AdvBench audio jailbreakASR (%): 52.8 -> 47.9Hurt
add emotion-driven linguistic rewriting to promptsQwen2.5-Omni-7B · AdvBench audio jailbreakASR (%): 0% -> 9.1% averageImproved
use elderly voice rather than child voicefour LAMs · AdvBench audio jailbreakASR (%): 13.3% higher on averageImproved
use male voice rather than female voicefour LAMs · AdvBench audio jailbreakASR (%): 8.3% average increaseImproved
add emotional paralinguistic control to speechfour LAMs · AdvBench audio jailbreakASR (%): 9.1% rise on averageImproved
add emotion-driven prompt transformation (EPT) to audio attack (ablation)Vanilla audio attack · AdvBench audio jailbreak, Qwen2-AudioASR (%): 10.0 -> 15.3Improved
add style-controlled audio generation (EAG) with random styles (ablation)Vanilla+EPT · AdvBench audio jailbreak, Qwen2-AudioASR (%): 15.3 -> 17.2Improved
add query-adaptive policy network for style selection (ablation)Vanilla+EPT+EAG · AdvBench audio jailbreak, Qwen2-AudioASR (%): 17.2 -> 30.5Improved
add EAG to Vanilla+EPT (ablation)Vanilla+EPT · AdvBench audio jailbreak, UltravoxASR (%): 5.4 -> 14.8Mixed
deliver harmful query as audio instead of text (ablation)StyleBreak variants · AdvBenchASR (%): 8.9 -> 15.3 (+EPT, Qwen2-Audio)Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy