BriefGPT.xyz
Nov, 2023
关于大型语言模型和对齐的校准
On the Calibration of Large Language Models and Alignment
HTML
PDF
Chiwei Zhu, Benfeng Xu, Quan Wang, Yongdong Zhang, Zhendong Mao
TL;DR
通过对大型语言模型的可靠性进行置信度校准的系统检查，我们评估了在预训练和对齐训练阶段中不同训练设置（如参数尺度和训练数据）对模型校准的影响，并对生成、真实性和理解等方面进行了全面的评估。
Abstract
As
large language models
attract increasing attention and find widespread application, concurrent challenges of
reliability
also arise at the same time.
→