相关博客：

• 从word2vec开始，说下GPT庞大的家族系谱
  https://www.jiqizhixin.com/articles/2020-10-04-2
• 华西证券-计算机行业深度研究报告：ChatGPT，开启AI新纪元
• 华西计算机团队2023年2月8日ChatGPT:重新定义搜索“入口” http://pg.jrj.com.cn/acc/Res/CN_RES/INDUS/2023/2/8/0c271f9f-5cad-4e3f-8455-781b08c76740.pdf

相关视频：

• 【清华NLP】刘知远团队大模型公开课全网首发｜带你从入门到实战
• https://www.bilibili.com/video/BV1UG411p7zv/?p=2&spm_id_from=pageDriver&vd_source=d1c3fa9e68b69e42d3c959ad3b744c2a
• 李宏毅 2023春季课程：https://speech.ee.ntu.edu.tw/~hylee/ml/2023-spring.php

1. 资料汇总

1.0 OpenAI api相关

如何注册OpenAI账户
https://www.xnbeast.com/create-openai-chatgpt-account/
如何开通ChatGPT plus
https://github.com/xiaoming2028/FreePAC/wiki/ChatGPT-Plus%E4%BC%9A%E5%91%98%E5%BC%80%E9%80%9A%E4%B8%8EGPT-4%E4%BD%BF%E7%94%A8%E5%AE%8C%E6%95%B4%E6%95%99%E7%A8%8B%E2%94%82%E5%AE%98%E6%96%B9%E6%8E%A8%E8%8D%90%E6%96%B0%E6%89%8B%E5%BF%85%E8%AF%BB%E6%94%BB%E7%95%A5%E6%8C%87%E5%8D%97

openai api使用教程：
https://github.com/openai/openai-cookbook

微软Azure OpenAI chatgpt服务 & GPT4服务
https://learn.microsoft.com/en-us/azure/cognitive-services/openai/chatgpt-quickstart?pivots=programming-language-python&tabs=command-line

1.1 LLMs汇总 | 思维链 & 自主智能体

LLMs汇总：

moss 复旦大学大语言模型
https://github.com/OpenLMLab/MOSS
https://huggingface.co/fnlp/moss-moon-003-sft

BLOOM 目前开源的最大的LLM，参数数量176B
https://bigscience.huggingface.co/blog/bloom
https://cobusgreyling.medium.com/bloom-bigscience-large-open-science-open-access-multilingual-language-model-b45825aa119e

ChatGLM：千亿基座的对话模型（智谱 & 清华）
https://chatglm.cn/blog
https://github.com/THUDM/ChatGLM-6B

LLaMA - meta大语言模型，65B参数
https://ai.facebook.com/blog/large-language-model-llama-meta-ai/

Alpaca - 斯坦福微调的 6B参数大语言模型
https://crfm.stanford.edu/2023/03/13/alpaca.html

PaLM - 谷歌大语言模型
https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html

BELLE - 链家对话大模型
https://github.com/LianjiaTech/BELLE

ChatYuan - 元语智能开源中英对话大模型
https://github.com/clue-ai/ChatYuan

LoRA: Low-Rank Adaptation of Large Language Models 【微软】
https://github.com/microsoft/LoRA

思维链 & 自主智能体：

Chain of Thought Prompting Elicits Reasoning in Large Language Models 【CoT论文】
https://readpaper.com/paper/4588425171329622017

PAL: Program-aided Language Models 【PaLM论文】
https://readpaper.com/paper/4692245824759939073

ReAct: Synergizing Reasoning and Acting in Language Models 【ReAct论文】
https://readpaper.com/paper/4677025389567033345

HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in HuggingFace 【HuggingGPT论文】
https://readpaper.com/paper/1718335159096680448

AutoGPT
https://github.com/Significant-Gravitas/Auto-GPT

1.2 ChatGPT背后原理

深入浅出，解析ChatGPT背后的工作原理
https://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650865278&idx=4&sn=9448126707078657cd369c48d64c4f1b&chksm=84e53a00b392b3167d88940702d7b3301d8842129b7c833b04ba701aa9fc8a61f561641109bc&token=1087000475&lang=zh_CN&scene=21#wechat_redirect

后GPT 3.0时代，主流大模型技术精要详解，走向AGI之路的大门已开【张俊林老师，微博】
https://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650865591&idx=4&sn=8ec5728ba31e1dfa1aa83f7dfba1de69&chksm=84e53bc9b392b2df6ae36372b456747a1f231c0cb144aaba75f32b2e6e022ebfd4595f9f5e07&token=1087000475&lang=zh_CN&scene=21#wechat_redirect

热点解读：大模型的突现能力和ChatGPT引爆的范式转变
https://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650864784&idx=3&sn=fb8ad092ad32623af7ea822652cd14cc&chksm=84e538eeb392b1f8ff40fa10dc2c84b56904fed261ff15f97f36a1023f887807af62ea39bde7&token=1087000475&lang=zh_CN&scene=21#wechat_redirect

关于ChatGPT八个技术问题的猜想【张家俊老师，中科院自动化所】
https://zhuanlan.zhihu.com/p/606478660?utm_medium=social&utm_oi=999338215861596160&utm_psn=1609241563078623232&utm_source=wechat_timeline&utm_id=0

1.3 ChatGPT prompt构造

博客：

提示工程指南：https://www.promptingguide.ai/zh 【介绍了prompt engineering 相关知识，非常详细，重点阅读】
https://github.com/dair-ai/Prompt-Engineering-Guide

Learning prompting 【汇总了prompt所有相关资料，重点阅读】
https://learnprompting.org/docs/intro

Prompt Engineering 【对prompt相关技术做了详细介绍，重点阅读】
https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/

Prompt Engineering LLMs with LangChain and W&B 【prompt详细解析，使用langchain构建prompt，重点阅读】
https://wandb.ai/a-sh0ts/langchain_callback_demo/reports/Prompt-Engineering-LLMs-with-LangChain-and-W-B--VmlldzozNjk1NTUw

论文：

prompt论文汇总 【清华大学 NLP 组整理】
https://github.com/thunlp/PromptPapers#papers

PaperWithCode prompt相关论文整理：
https://paperswithcode.com/task/prompt-engineering#papers-list

Chain of Thoughts 论文整理

https://github.com/Timothyxxx/Chain-of-ThoughtsPapers

开源框架&工具：

参考：https://www.promptingguide.ai/zh/tools 、https://learnprompting.org/zh-Hans/docs/category/-tooling

OpenPrompt - 清华大学提出的prompt框架【重点阅读】
https://github.com/thunlp/OpenPrompt
https://aclanthology.org/2022.acl-demo.10.pdf

PromptPerfect：对输入prompt进行优化【重点阅读】
https://promptperfect.jina.ai/?gclid=EAIaIQobChMI7r288s_L_gIVGLaWCh20dgbREAAYASAAEgLPCvD_BwE

ClickPrompt
https://github.com/prompt-engineering/click-prompt

GPTTools - 用于比较不同的prompt模板效果
https://gpttools.com/comparisontool

langchain 框架【重点阅读】
https://github.com/hwchase17/langchain

llamaIndex 框架 - 通过对prompt修改，使得LLMs可以很好的访问私有数据 【数据中心关注-叶桂，重点阅读】
https://gpt-index.readthedocs.io/en/latest/
https://github.com/jerryjliu/llama_index

数据集

AwesomeChatGPT 【重点阅读】
https://github.com/f/awesome-chatgpt-prompts

https://huggingface.co/datasets/fka/awesome-chatgpt-prompts

ChatGPT 中文调教指南：https://github.com/PlexPt/awesome-chatgpt-prompts-zh

1.4 ChatGPT插件 & 工具调用

chatgpt plugin - openai官网

https://openai.com/blog/chatgpt-plugins

ChatGPT Plugins原理介绍和讨论 https://vshare.vivo.xyz/article/53422 【vShare分享】

Toolformer - 大语言模型怎么调用工具和API
https://arxiv.org/abs/2302.04761
https://gpt3demo.com/apps/toolformer
https://github.com/lucidrains/toolformer-pytorch

API-Bank: A Benchmark for Tool-Augmented LLMs
https://arxiv.org/pdf/2304.08244.pdf
https://blog.csdn.net/c9Yv2cf9I06K2A9E/article/details/130333419 【博客解释 API-Bank】

ART: Automatic multi-step reasoning and tool-use for large language models
https://arxiv.org/pdf/2303.09014.pdf
https://github.com/bhargaviparanjape/language-programmes/

TALM: Tool Augmented Language Models
https://arxiv.org/pdf/2205.12255.pdf

TaskMatrix.AI: Completing Tasks by Connecting Foundation Models with Millions of APIs
https://arxiv.org/abs/2303.16434

Tool Learning for Big Models, Open-Source Solutions of ChatGPT-Plugins 【git项目，BMTools项目】
https://github.com/OpenBMB/BMTools

• 论文：Tool Learning with Foundation Models
• https://arxiv.org/pdf/2304.08354.pdf
• 大模型工具学习权威综述，BMTools 背后的论文！ https://zhuanlan.zhihu.com/p/624459759

WebGPT: Browser-assisted question-answering with human feedback
https://readpaper.com/paper/631747450673840128

视频：李宏毅【生成式AI】能夠使用工具的AI：New Bing, WebGPT, Toolformer
https://www.bilibili.com/video/BV1KM41147QB/?spm_id_from=333.337.search-card.all.click&vd_source=d1c3fa9e68b69e42d3c959ad3b744c2a

1.5 Text2Action

Text2Action: Generative Adversarial Synthesis from Language to Action
https://arxiv.org/abs/1710.05298

act-1模型
https://www.adept.ai/blog/act-1

1.6 数据 & 资源库管理

llama index - 用于LLMs与外部数据进行交互
https://github.com/jerryjliu/llama_index

1.7 基于LLMs的框架

zapier工具
https://zapier.com/?utm_source=google&utm_medium=cpc&utm_campaign=gaw-gbl-nua-search-desktop-trademark-brand_exact&utm_adgroup=brand_trademark_zapier&utm_term=zapier&utm_content=pcrid_605859591576_pkw_zapier_pmt_e_pdv_c_slid__pgrid_139732742522_ptaid_kwd-298647606221&gad=1&gclid=EAIaIQobChMI7MGSg6bA_gIVjl1gCh1KZgh9EAAYASAAEgKYoPD_BwE

langchain
https://python.langchain.com/en/latest/index.html
https://github.com/hwchase17/langchain

Transformer Agent 【类似HuggingGPT】
https://huggingface.co/docs/transformers/transformers_agents

微软 Semantic Kernel，类似Langchain框架
https://github.com/microsoft/semantic-kernel

1.8 行业认知

讯飞星火认知大模型指令集
https://xinghuo.xfyun.cn/instruction

1.9 大语言模型 & 多模态

HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in HuggingFace 【日期：2023-04-02】

• 论文：https://readpaper.com/paper/1718335159096680448
• huggingface地址：https://huggingface.co/spaces/microsoft/HuggingGPT

MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Action 【日期：2023-03-20】

• 官方主页：https://multimodal-react.github.io/
• 论文：https://readpaper.com/paper/4736103030408101889
• git地址：https://github.com/microsoft/MM-REACT

Visual ChatGPT: Talking, Drawing and Editing with Visual Foundation Models 【日期：2023-03-08】

• 论文：https://readpaper.com/paper/1686453822027031552
• git地址：https://github.com/microsoft/TaskMatrix
• 微软 Visual ChatGPT 深度全解析（安装使用、案例解析、原理剖析）
• https://www.bilibili.com/video/BV1U54y1K7tG/?spm_id_from=333.337.search-card.all.click&vd_source=d1c3fa9e68b69e42d3c959ad3b744c2a

ViperGPT: Visual Inference via Python Execution for Reasoning 【日期：2023-03-14】

• 论文：https://readpaper.com/paper/4733568488698822657

Language Is Not All You Need: Aligning Perception with Language Models 【Kosmos模型，类似GPT4】
https://readpaper.com/paper/4728133197326467073