Poisoning Attacks on LLMs: A Direct Attack on LLMs with Less than 250 Samples

OpenAI API moonlights as malware HQ in Microsoft’s latest discovery

Annotating the Chain-of-Thought: A Behavior-Labeled Dataset for AI Safety

Anthropic AI Releases Petri: An Open-Source Framework for Automated Auditing by Using AI Agents to Test the Behaviors of Target Models on Diverse Scenarios

Beyond Classification: Evaluating LLMs for Fine-Grained Automatic Malware Behavior Auditing

Apple slips up on ChillyHell macOS malware, lets it past security . . . for 4 years

Anthropic details its AI safety strategy