Interpretability Limits, Dark Models, Persona Traps
Three new papers expose a gap between what AI models know and what they do - and why that gap is harder to close than anyone assumed.
Prompt engineering
AI Hallucinations Explained: How to Catch Them
AI chatbots confidently state false information all the time - here's why it happens, which outputs to distrust most, and five strategies to catch mistakes before they cause problems.
