I am interested in conversations about mechanistic interpretability, AI auditing, and alignment research. If you are working on similar problems, have feedback on something I have written, or want to discuss a collaboration, I would like to hear from you.

The best conversations start with “I disagree, but here's why.”