16 papers
Practical Secrets Extraction against Black-box LLMs
Shiqian Zhao, Siwei Jiang, Xinfeng Li +5
Large language models (LLMs) increasingly power autonomous coding agents such as Codex and Claude Code, yet their training corpora may contain confidential credentials exposed in p…
ECLIPSE: Self-Evolving Stealthy Prompt Injection Attack against Long-Horizon Agentic Systems
Shiqian Zhao, Yangfan Zhou, Xinfeng Li +5
Recently, large language model (LLM) agents, such as Codex, Claude Code, and OpenClaw, have become capable of planning and executing long-horizon tasks through repeated tool calls.…
When Claws Remember but Do Not Tell: Stealthy Memory Injection in Persistent Personal Agents
Yechao Zhang, Shiqian Zhao, Jiawen Zhang +5
Persistent personal agents combine long-term memory with access to users' external environments, enabling personalized foreground assistance and proactive background execution. Thi…
Mind Your HEARTBEAT! Claw Background Execution Inherently Enables Silent Memory Pollution
Yechao Zhang, Shiqian Zhao, Jie Zhang +5
We identify a critical security vulnerability in mainstream Claw personal AI agents: untrusted content encountered during heartbeat-driven background execution can silently pollute…
Knowing without Acting: The Disentangled Geometry of Safety Mechanisms in Large Language Models
Jinman Wu, Yi Xie, Shen Lin +2
Safety alignment is often conceptualized as a monolithic process wherein harmfulness detection automatically triggers refusal. However, the persistence of jailbreak attacks suggest…
Depth Charge: Jailbreak Large Language Models from Deep Safety Attention Heads
Jinman Wu, Yi Xie, Shiqian Zhao +1
Currently, open-sourced large language models (OSLLMs) have demonstrated remarkable generative performance. However, as their structure and weights are made public, they are expose…
Inference-time Alignment via Sparse Junction Steering
Runyi Hu, Jie Zhang, Shiqian Zhao +7
Token-level steering has emerged as a pivotal approach for inference-time alignment, enabling fine grained control over large language models by modulating their output distributio…
WADBERT: Dual-Channel Web Attack Detection via Permutation-Invariant Parameter Modeling
Kangqiang Luo, Yi Xie, Shiqian Zhao +1
Web attack detection is the first line of defense for securing web applications, designed to preemptively identify malicious activities. Deep learning-based approaches are increasi…
P2P: A Poison-to-Poison Remedy for Reliable Backdoor Defense in LLMs
Shuai Zhao, Xinyi Wu, Shiqian Zhao +4
During fine-tuning, large language models (LLMs) are increasingly vulnerable to data-poisoning backdoor attacks, which compromise their reliability and trustworthiness. However, ex…
Towards Effective Prompt Stealing Attack against Text-to-Image Diffusion Models
Shiqian Zhao, Chong Wang, Yiming Li +7
Text-to-Image (T2I) models, represented by DALLE and Midjourney, have gained huge popularity for creating realistic images. The quality of these images relies on the careful…
T2V-OptJail: Discrete Prompt Optimization for Text-to-Video Jailbreak Attacks
Jiayang Liu, Siyuan Liang, Shiqian Zhao +5
In recent years, fueled by the rapid advancement of diffusion models, text-to-video (T2V) generation models have achieved remarkable progress, with notable examples including Pika,…
Three Minds, One Legend: Jailbreak Large Reasoning Model with Adaptive Stacked Ciphers
Viet-Anh Nguyen, Shiqian Zhao, Gia Dao +3
Recently, Large Reasoning Models (LRMs) have demonstrated superior logical capabilities compared to traditional Large Language Models (LLMs), gaining significant attention. Despite…
When Memory Becomes a Vulnerability: Towards Multi-turn Jailbreak Attacks against Text-to-Image Generation Systems
Shiqian Zhao, Jiayang Liu, Yiming Li +9
Modern text-to-image (T2I) generation systems (e.g., DALLE 3) exploit the memory mechanism, which captures key information in multi-turn interactions for faithful generation…
A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment
Kun Wang, Guibin Zhang, Zhenhong Zhou +100
The remarkable success of Large Language Models (LLMs) has illuminated a promising pathway toward achieving Artificial General Intelligence for both academic and industrial communi…
Mask Image Watermarking
Runyi Hu, Jie Zhang, Shiqian Zhao +5
We present MaskWM, a simple, efficient, and flexible framework for image watermarking. MaskWM has two variants: (1) MaskWM-D, which supports global watermark embedding, watermark l…
SAME: Sample Reconstruction against Model Extraction Attacks
Yi Xie, Jie Zhang, Shiqian Zhao +2
While deep learning models have shown significant performance across various domains, their deployment needs extensive resources and advanced computing infrastructure. As a solutio…