Tag: AI security
Browse our exclusive articles!
AgentHazard: Benchmark for Detecting Harmful Agent Behavior
AgentHazard benchmark evaluates harmful behavior in computer-use agents, highlighting safety risks and the need for improved safeguards in AI models.
Divide-and-Conquer Side-Channel Attacks on DNN Models
Discover a novel divide-and-conquer strategy for extracting deep neural networks via side-channel attacks with high fidelity and accuracy.
Secure LLM Instructions: Automated Framework Against Encoding Attacks
Discover an automated framework to evaluate and harden LLM system instructions, protecting sensitive data from encoding attacks without costly retraining.
WARP: Guaranteed Inner-Layer Repair for NLP Transformers
Discover WARP, a novel method ensuring guaranteed inner-layer repair and enhanced robustness for NLP Transformer models against adversarial attacks.
Backdoor Attacks on Continuous Latent Reasoning Models
Explore how backdoor attacks exploit continuous latent reasoning in AI models, revealing vulnerabilities and evading defenses with the ThoughtSteer method.