Mitigating Shortcut Reasoning in Language Models: A Gradient-Aware Training Approach

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Cao, Hongyu, Liu, Kunpeng, Wang, Dongjie, Fu, Yanjie
Format:	Preprint
Published:	2026
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2603.20899