How did they decide OpenAI frontier model was safe?

1 points | by rstagi 9 hours ago