Judge Like Human Examiners: A Weighted Importance Multi-Point Evaluation Framework for Generative Tasks with Long-form Answers

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Yu, Guoxin, Zhou, Chulun, Liu, Lemao, Wang, Qi, Yu, Mo, Tang, Jialong, Yang, Baosong, Ao, Xiang, Lam, Wai, Yu, Yue
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/2604.11246