What is this?
This is DAIR.AI's Prompt Engineering Guide: a large, continuously updated collection of guides, papers, lecture notes, and notebooks covering prompt engineering techniques, context engineering, retrieval-augmented generation (RAG), and building AI agents with LLMs.
This is DAIR.AI's Prompt Engineering Guide: a large, continuously updated collection of guides, papers, lecture notes, and notebooks covering prompt engineering techniques, context engineering, retrieval-augmented generation (RAG), and building AI agents with LLMs.Is this a software tool, or a collection of guides and papers to read?It is a reference and documentation site rather than a software library - a curated set of write-ups on techniques, models, and papers. The repo also builds the guide into a Next.js/Nextra web app, published online as the Web Version.
It is a reference and documentation site rather than a software library - a curated set of write-ups on techniques, models, and papers. The repo also builds the guide into a Next.js/Nextra web app, published online as the Web Version.Where should a beginner start in the guide?The guide is organized into numbered sections; a newcomer should start with the Introduction and LLM Settings pages before moving on to techniques like zero-shot, few-shot, and Chain-of-Thought prompting.
The guide is organized into numbered sections; a newcomer should start with the Introduction and LLM Settings pages before moving on to techniques like zero-shot, few-shot, and Chain-of-Thought prompting.Do I need anything special to run the guide locally?Only if you want to run the site locally, for example to check a translation: you need Node >=18 and pnpm, then install the Next.js/Nextra dependencies and run pnpm dev. Just reading the guide online requires nothing.
Only if you want to run the site locally, for example to check a translation: you need Node >=18 and pnpm, then install the Next.js/Nextra dependencies and run pnpm dev. Just reading the guide online requires nothing.Is the content kept up to date and available in multiple languages?The project is actively maintained - it has crossed 3 million learners, added a video lecture, and supports 13 languages with more translations welcomed, so content keeps expanding rather than going stale.
The project is actively maintained - it has crossed 3 million learners, added a video lecture, and supports 13 languages with more translations welcomed, so content keeps expanding rather than going stale.What topics does the guide cover beyond the basics?It goes deep into techniques like Zero-Shot and Few-Shot prompting, Chain-of-Thought, Tree of Thoughts, ReAct, RAG, and function calling, plus model-specific notes for ChatGPT, GPT-4, LLaMA, Gemini, Mistral, and Mixtral.
It goes deep into techniques like Zero-Shot and Few-Shot prompting, Chain-of-Thought, Tree of Thoughts, ReAct, RAG, and function calling, plus model-specific notes for ChatGPT, GPT-4, LLaMA, Gemini, Mistral, and Mixtral.Is there a video lecture or paid course alongside the guide?Yes - there is a free 1-hour video lecture with slides and a notebook, plus paid, self-paced prompt engineering, RAG, and AI Agents courses under DAIR.AI Academy.
Yes - there is a free 1-hour video lecture with slides and a notebook, plus paid, self-paced prompt engineering, RAG, and AI Agents courses under DAIR.AI Academy.Has this guide gotten any outside recognition?It hit number 1 on Hacker News in February 2023 and has been covered by the Wall Street Journal and Forbes.
It hit number 1 on Hacker News in February 2023 and has been covered by the Wall Street Journal and Forbes.Why did DAIR.AI create this guide?DAIR.AI built it to consolidate papers, techniques, and tools around prompting because of high developer and researcher interest in getting more reliable and capable output from LLMs.
DAIR.AI built it to consolidate papers, techniques, and tools around prompting because of high developer and researcher interest in getting more reliable and capable output from LLMs.What are similar projects?ai-system-design, ai-guide, generative-ai-for-beginners, awesome-chatgpt-zh, harness-engineering, Awesome-LLM, ai-engineering-hub, awesome-generative-ai-guide
ai-system-design, ai-guide, generative-ai-for-beginners, awesome-chatgpt-zh, harness-engineering, Awesome-LLM, ai-engineering-hub, awesome-generative-ai-guideWhich topics is it in?mdx, language-model, retrieval-augmented-generation, llms, papers, gpt, chatgpt, knowledge-base, prompt-engineering, nextjs, rag, generative-ai
mdx, language-model, retrieval-augmented-generation, llms, papers, gpt, chatgpt, knowledge-base, prompt-engineering, nextjs, rag, generative-ai