ryan_greenblatt comments on Alignment Faking in Large Language Models

[ryan_greenblatt]

Ah yes this is a mistake, thanks. We noticed it and replied with a correction on Twitter, but I didn’t move the fix here.