Paul Colognese comments on High-level interpretability: detecting an AI’s objectives

[Paul Colognese]

Thanks for pointing this out. I’ll look into it and modify the post accordingly.