Why Can't Transformers Learn Multiplication? Reverse-Engineering Reveals Long-Range Dependency Pitfalls

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bai, Xiaoyan, Pres, Itamar, Deng, Yuntian, Tan, Chenhao, Shieber, Stuart, Viégas, Fernanda, Wattenberg, Martin, Lee, Andrew
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!