InjecGuard: Benchmarking and Mitigating Over-defense in Prompt Injection Guardrail Models
Paper
•
2410.22770
•
Published
Liu
Xiaogeng-SheltonLiu
AI & ML interests
None yet
Recent Activity
authored
a paper
13 days ago
InjecGuard: Benchmarking and Mitigating Over-defense in Prompt Injection
Guardrail Models
authored
a paper
13 days ago
Doxing via the Lens: Revealing Privacy Leakage in Image Geolocation for
Agentic Multi-Modal Large Reasoning Model
authored
a paper
13 days ago
Code Agent can be an End-to-end System Hacker: Benchmarking Real-world
Threats of Computer-use Agent
Organizations
authored
5
papers
13 days ago
Doxing via the Lens: Revealing Privacy Leakage in Image Geolocation for Agentic Multi-Modal Large Reasoning Model
Paper
•
2504.19373
•
Published
•
1
Code Agent can be an End-to-end System Hacker: Benchmarking Real-world Threats of Computer-use Agent
Paper
•
2510.06607
•
Published
•
4
AutoDAN-Reasoning: Enhancing Strategies Exploration based Jailbreak Attacks with Test-Time Scaling
Paper
•
2510.05379
•
Published
ReasoningBomb: A Stealthy Denial-of-Service Attack by Inducing Pathologically Long Reasoning in Large Reasoning Models
Paper
•
2602.00154
•
Published
authored
2
papers
almost 2 years ago