Lex Fridman Podcast
#452 – Dario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity
But it turns out that that works very, very well. To me, that seems like some non-trivial validation of linear representations and supersession. So with dictionary learning, you're not looking for particular kind of categories. You don't know what they are. Exactly. They just emerge. And this gets back to our earlier point, right?
0
💬
0
Comments
Log in to comment.
There are no comments yet.