LSPO: Length-aware Dynamic Sampling for Policy Optimization in LLM Reasoning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Chen, Weizhe, Koenig, Sven, Dilkina, Bistra
Format:	Preprint
Published:	2025
Subjects:	Machine Learning 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/2510.01459