OpenAI trained o1 and o3 for safety policy #1 Post by nancy_le » Mon, Dec 23, 2024, 4:23 AM UTC OpenAI trained o1 and o3 for safety policytechcrunch.com