Mur d'images - Cyberveille

curated by Decio

Recent Jailbreaks Demonstrate Emerging Threat to DeepSeek

Many-shot jailbreaking \ Anthropic

Bad Likert Judge: A Novel Multi-Turn Technique to Jailbreak LLMs by Misusing Their Evaluation Capability

EPFL: des failles de sécurité dans les modèles d'IA

Here is Apple's official 'jailbroken' iPhone for security researchers | TechCrunch

Using AI to Automatically Jailbreak GPT-4 and Other LLMs in Under a Minute