BriefGPT.xyz
Feb, 2024
温度计: 面向大型语言模型的通用校准
Thermometer: Towards Universal Calibration for Large Language Models
HTML
PDF
Maohao Shen, Subhro Das, Kristjan Greenewald, Prasanna Sattigeri, Gregory Wornell...
TL;DR
我们提出了一个适用于大规模语言模型（LLM）的校准方法——温度计(THERMOMETER)，该方法能够通过学习辅助模型来校准LLM，具有高效计算、保持LLM准确性以及为新任务生成更好校准响应的特点，经过多个基准测试的广泛实证评估证明该方法的有效性。
Abstract
We consider the issue of
calibration
in
large language models
(LLM). Recent studies have found that common interventions such as instruction tuning often result in poorly calibrated
→