Lex Fridman Podcast
#452 – Dario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity
And so neural networks, my guess would be, can do very well in terms of going and packing things in. To the point, that's probably not the limiting factor. How does the problem of polysemanticity enter the picture here? And polysemanticity is this phenomenon we observe where we look at many neurons and the neuron doesn't just sort of represent one concept. It's not a clean feature.
0
💬
0
Comments
Log in to comment.
There are no comments yet.