DeReason: A Difficulty-Aware Curriculum Improves Decoupled SFT-then-RL Training for General Reasoning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Hu, Hanxu, Wang, Yuxuan, Huan, Maggie, Vamvas, Jannis, Huang, Yinya, Guo, Zhijiang, Sennrich, Rico
Format:	Preprint
Published:	2026
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2603.11193