Low-rank fine-tuning lies between lazy training and feature learning.

Arif Kerem Dayi, Sitan Chen

Browse the full COLT paper archive.