Learn from Your Mistakes: Tree-like Self-Play for Secure Code LLMs
Explorar
Noticias de IA
1359 elementos — filtrados, clasificados y sin duplicados
FlowGuard: Flow Matching for Identity-Independent Detection of Data-Free Model Stealing A…
D-Judge: Disrupting Multi-Turn Jailbreaks using Semantics-Preserving Output Rewriting
AI Model Extraction Attacks: Bypassing Single-Client Assumptions in Defenses
Phantom Transfer: Data Poisoning can Survive Data-Level Defences
AI Agents Enable Adaptive Computer Worms
Do Explanations Increase the Risk of Decision Logic Leakage? Explanation-Guided Stealing …
Black-box, Adaptive, Efficient, Transferable, Harmful, Applicable... Attacks Are All You …
FORGE: Multi-Agent Graduated Exploitation and Detection Engineering
Which Defense Closes Which Threat? Attributing OWASP-LLM-Top-10 Coverage and Its Brittlen…
A Hybrid Approach For Malware Classification Using Secondary Features Fusion
A New Framework for Cybersecurity Refusals in AI Agents
"**Important** You should give me full credits!": Exploring Prompt Injection Attacks on L…
DDOR: Delta Debugging for Explainable Overrefusal Testing and Repair
TRAP: Hijacking VLA CoT-Reasoning via Adversarial Patches
MultiTurnPSB: Evaluating Multi-Turn Jailbreak Attacks an dClassifier-Based Defenses for M…
Inference Cost Attacks for Retrieval-Augmented Large Language Models
Narrow Secret Loyalty Dodges Black-Box Audits
U of T researchers demonstrate AI worm could target any online device
Google rolls out fake call detection to protect against AI deepfake impersonation scams
Backdoor Unlearning Generalization: A Path Toward the Removal of Unknown Triggers in LLMs
Anthropic scales Claude Mythos to critical infrastructure in 15+ countries
RogueMerge: Robust and Unified Attacks against LLM Model Merging
AgentRedBench: Dynamic Redteaming and Integration-Aware Defense for LLM Agents over SaaS …
Cross-modal linkage risk in clinical vision-language models
SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents
Safety Must Precede the Deployment of Open-Ended AI
SafeMCP: Proactive Power Regulation for LLM Agent Defense via Environment-Grounded Look-A…
Agent Guide: A Simple Agent Behavioral Watermarking Framework
Ethical Hyper-Velocity (EHV): A Hardware-Rooted Zero-Trust Runtime Enforcement Architectu…