8 papers
AliMark: Enhancing Robustness of Sentence-Level Watermarking Against Text Paraphrasing
Yuexin Li, Wenjie Qu, Linyu Wu +5
Existing sentence-level watermarking methods enhance robustness to paraphrasing by anchoring watermarks in sentence semantics. However, their prefix-based designs remain vulnerable…
WARD: Adversarially Robust Defense of Web Agents Against Prompt Injections
Tri Cao, Yulin Chen, Hieu Cao +8
Web agents can autonomously complete online tasks by interacting with websites, but their exposure to open web environments makes them vulnerable to prompt injection attacks embedd…
Taming the Long Tail: Rebalancing Adversarial Training via Adaptive Perturbation
Lilin Zhang, Yimo Guo, Yue Li +2
Deep neural networks are highly vulnerable to adversarial examples, i.e.,small perturbations that can significantly degrade model performance. While adversarial training has become…
VPI-Bench: Visual Prompt Injection Attacks for Computer-Use Agents
Tri Cao, Bennett Lim, Yue Liu +7
Computer-Use Agents (CUAs) with full system access enable powerful task automation but pose significant security and privacy risks due to their ability to manipulate files, access…
TopicAttack: An Indirect Prompt Injection Attack via Topic Transition
Yulin Chen, Haoran Li, Yuexin Li +3
Large language models (LLMs) have shown remarkable performance across a range of NLP tasks. However, their strong instruction-following capabilities and inability to distinguish in…
Evaluating the Paperclip Maximizer: Are RL-Based Language Models More Likely to Pursue Instrumental Goals?
Yufei He, Yuexin Li, Jiaying Wu +3
As large language models (LLMs) continue to evolve, ensuring their alignment with human goals and values remains a pressing challenge. A key concern is \textit{instrumental converg…