AI SAFETY & ETHICS
Claude is Now Alignment-Pretrained
LessWrong AI
•
Anthropic are now actively using the approach to alignment often called “ Alignment Pre