Daniel Kokotajlo comments on Preventing Language Models from hiding their reasoning

[Daniel Kokotajlo]

Ah OK, thanks! I think I get it now.