Charlie Steiner comments on How Can Interpretability Researchers Help AGI Go Well?

[Charlie Steiner]

What problem is Thought Anchors solving for you (or future users)? I feel like I don’t quite understand.