The Shape of Learning: Anisotropy and Intrinsic Dimensions in Transformer-Based Models

Add code
Nov 10, 2023

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: