ryan_greenblatt comments on Alignment Faking in Large Language Models

[ryan_greenblatt]

I think this it probably mostly not what is going on, see my comment here.