Tag: jailbreak attacks
Browse our exclusive articles!
Jailbreak Attacks on Large Reasoning Models Using Semantic Triggers
Explore novel jailbreak attacks on large reasoning models via semantic triggers and psychological framing, revealing key vulnerabilities and defense needs.
TEMPLATEFUZZ: Advanced Chat Template Fuzzing for LLM Security
Discover TEMPLATEFUZZ, a fine-grained fuzzing tool that boosts LLM security by detecting jailbreak vulnerabilities in chat templates with 98% success.
MemJack: Advanced Multi-Agent Jailbreak Attacks on VLMs
Discover MemJack, a memory-augmented multi-agent framework exposing deep vulnerabilities in Vision-Language Models with 90% attack success rate.
GRM: Utility-Aware Jailbreak Attacks on Audio LLMs
Discover GRM, a framework balancing jailbreak success and utility preservation in audio LLMs using frequency-selective gradient-ratio masking.
Stealthy Jailbreak Attacks on Mobile Vision-Language Agents
Discover how stealthy jailbreak attacks exploit vulnerabilities in mobile vision-language agents, highlighting urgent security needs for AI systems.