AI SAFETY & ETHICS

Claude is Now Alignment-Pretrained

LessWrong AI

Anthropic are now actively using the approach to alignment often called “ Alignment Pre