Login / Signup

In-Context Learning of a Linear Transformer Block: Benefits of the MLP Component and One-Step GD Initialization.

Ruiqi ZhangJingfeng WuPeter L. Bartlett
Published in: CoRR (2024)
Keyphrases