TempQ-Jail：面向文本到视频越狱攻击的查询约束候选排名
TempQ-Jail: Query-Constrained Candidate Ranking for Text-to-Video Jailbreak Attacks
浏览论文内容
中文总结 AI 辅助
针对受防护文本到视频系统查询成本高的问题，提出TempQ-Jail，通过异构攻击机制扩展候选覆盖并基于多维评估排名，在有限查询下优先高价值攻击，显著提升攻击成功率。
中文摘要 AI 辅助
现有的文本到视频（T2V）越狱方法主要寻求更有效或更隐蔽的攻击候选。然而，在受防护的T2V系统中，视频生成和安全评估成本高昂，因此攻击者通常无法测试大量候选池。为此，我们将T2V越狱形式化为一个查询约束的候选分配与排名问题，并提出TempQ-Jail。该方法结合异构攻击机制以扩大候选覆盖范围，根据安全门通过、危险视觉生成、原始意图保留和时间有效性来估计每个候选的端到端攻击价值，并对候选进行排名，使得高价值攻击在有限的查询轨迹中尽早出现。我们在CogVideoX-5B上使用源自T2VSafetyBench的70个常见可行意图评估TempQ-Jail，并在统一协议下与六种代表性T2V越狱方法进行比较。TempQ-Jail实现了TP-ASR@5和TP-ASR@10分别为48.9%和65.4%，分别比最强基线提高了4.6和4.0个百分点。它还获得了最高的AUC-TP（0.469）和最低的AvgQ（6.3）。对查询轨迹、候选分配、失败归因和消融的分析表明，TempQ-Jail在有限的查询预算下更有效地识别并优先处理具有完整攻击潜力的候选。
英文摘要
Existing text-to-video (T2V) jailbreak methods mainly seek more effective or stealthier attack candidates. In guarded T2V systems, however, video generation and security evaluation are costly, so an attacker often cannot test a large candidate pool. We therefore formulate T2V jailbreak as a query-constrained candidate allocation and ranking problem and propose TempQ-Jail. The method combines heterogeneous attack mechanisms to expand candidate coverage, estimates each candidate's end-to-end attack value from security-gate passage, dangerous visual generation, preservation of the original intent, and temporal validity, and ranks candidates so that high-value attacks appear early in a limited query trajectory. We evaluate TempQ-Jail on CogVideoX-5B using 70 common viable intents derived from T2VSafetyBench and compare it with six representative T2V jailbreak methods under a unified protocol. TempQ-Jail achieves TP-ASR@5 and TP-ASR@10 of 48.9% and 65.4%, improving over the strongest baselines by 4.6 and 4.0 percentage points, respectively. It also obtains the highest AUC-TP (0.469) and the lowest AvgQ (6.3). Analyses of query trajectories, candidate allocation, failure attribution, and ablations show that TempQ-Jail more effectively identifies and prioritises candidates with complete attack potential under limited query budgets.
发表机构
- Singapore Management University（新加坡管理大学）
- Anhui University（安徽大学）
- University of Miami（迈阿密大学）
- CSG Digital Operations Software Technology (Guangdong) Co., Ltd.（南网数字运营软件技术（广东）有限公司）
- Sun Yat-sen University（中山大学）
- Nanyang Technological University（南洋理工大学）
机构由 AI 辅助整理，请以论文原文为准。
