skip to content
菜鸟文档

202310月记

/ 4 min read

TODO

事项明细

访问GPT-api的测试方法

curl --location 'https://tabmanager.cn/v1/chat/completions' \
--header 'Authorization: Bearer sk-todo' \
--header 'Content-Type: application/json' \
--data '{
   "model": "gpt-4",
  "messages": [{"role": "user", "content": "树上有9只鸟,打掉一只,还有几只?"}]
 }'

日记

  • 今天还是国庆节,来书店把一些琐事处理一下

学习记录

Astro

地址:https://docs.astro.build/zh-cn/tutorial/5-astro-api/ 学习进度:第六章 复习进度:第五章

学习明细

总结归纳

大模型论文

大模型相关综述论文 Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing

conclution

In this article, we have summarized and analyzed several paradigms in the development of sta-
tistical natural language processing techniques and have argued that prompt-based learning is a
promising new paradigm that may represent another major change in the way we look at NLP.
First, we hope this survey will help researchers more effectivel yan dcomprehensivel yunderstan
the paradigm of prompt-based learning and grasp its core challenges so that more scientificall
meaningful advances can be made in this field .I naddition ,lookin gal lth ewa ybac kt oth esum
mary of the four paradigms of NLP research presented in Section 1, we hope to highlight the
commonalities and difference sbetwee nthem ,makin gresearc ho nan yo fthes eparadigm smor
full-fledge dan dpotentiall yprovidin g acatalys tt oinspir ewor ktowar dth enex tparadig mshif ta
well.
"在这篇文章中,我们总结并分析了统计自然语言处理技术发展中的几种范例,并且我们认为基于提示的学习是一个有前景的新范例,它可能代表了我们看待NLP的另一种重大改变。首先,我们希望这个调查能帮助研究者更有效地、更全面地理解基于提示的学习的范例,并把握其核心挑战,以便在这个领域做出更有科学意义的进步。此外,回顾第一部分中介绍的NLP研究的四种范例的总结,我们希望突出它们之间的共性和差异,使对这些范例的研究更加全面,也可能提供一个催化剂来激发朝下一个范例转变的工作。"

阅读论文的网站

https://chatdoc.com/ https://arxiv.org/ https://lw.tabmanager.cn/

nlp相关网站,以及 Trend of Prompt-based Research http://pretrain.nlpedia.ai/

大模型相关内容综述: https://arxiv.org/pdf/2107.13586.pdf

html

盒子模型学习进度: https://developer.mozilla.org/zh-CN/docs/Learn/CSS/Building_blocks/The_box_model#%E5%9D%97%E7%BA%A7%E7%9B%92%E5%AD%90%EF%BC%88block_box%EF%BC%89%E5%92%8C_%E5%86%85%E8%81%94%E7%9B%92%E5%AD%90%EF%BC%88inline_box%EF%BC%89