BriefGPT.xyz
Sep, 2023
大型语言模型的思路链测试
Stress Testing Chain-of-Thought Prompting for Large Language Models
HTML
PDF
Aayush Mishra, Karan Thakkar
TL;DR
通过对Chain-of-Thought(CoT)提示方式的研究,我们发现CoT的正确操作和数值是GPT-3在各种任务中获得准确答案的关键,而错误操作会对性能产生负面影响,这项研究加深了我们对CoT提示方式的理解,也提出了有关大语言模型在上下文中学习推理能力的新问题。
Abstract
This report examines the effectiveness of
chain-of-thought
(CoT)
prompting
in improving the
multi-step reasoning
abilities of
→