Starvard University sealStarvard
๐ŸŒ English
EnglishFranรงaisEspaรฑolเคนเคฟเคจเฅเคฆเฅ€PortuguรชsBahasa IndonesiaDeutschItalianoNederlandsๆ—ฅๆœฌ่ชžํ•œ๊ตญ์–ดไธญๆ–‡ุงู„ุนุฑุจูŠุฉะ ัƒััะบะธะนTรผrkรงeTiแบฟng Viแป‡tTagalog

Translations open on starvard.org. Courses are translated in the first five; the library speaks all sixteen.

Join the Discord

Starvard University ยท AI and Money

Prompt Engineering Guide

The budget AI skills course: techniques that make any free chatbot give you better answers.

By Elvis Saravia and DAIR.AIFree to read in fullMIT licence
Start reading
Course cover art for the Prompt Engineering Guide

About this course

You do not need a paid AI tool to get good results. You need to know how to ask. This guide teaches the techniques that make any chatbot more useful, starting simple and building up.

It begins with zero-shot and few-shot prompting, moves through chain of thought, and goes on to retrieval-augmented generation (RAG) and agents. It finishes with model-by-model notes, so you can see how the techniques apply to the specific models you use.

New to prompting? Start with our AI Prompting 101, then come back here for the full technique library.

Contents

Every section, in the guide's own order.

Introduction

ยทIntroductionยทLLM SettingsยทBasics of PromptingยทElements of a PromptยทGeneral Tips for Designing PromptsยทExamples of Prompts

Prompting Techniques

ยทPrompting TechniquesยทZero-Shot PromptingยทFew-Shot PromptingยทChain-of-Thought PromptingยทMeta PromptingยทSelf-ConsistencyยทGenerated Knowledge PromptingยทPrompt ChainingยทTree of Thoughts (ToT)ยทRetrieval Augmented Generation (RAG)ยทAutomatic Reasoning and Tool-use (ART)ยทAutomatic Prompt Engineer (APE)ยทActive-PromptยทDirectional Stimulus PromptingยทPAL (Program-Aided Language Models)ยทReAct PromptingยทReflexionยทMultimodal CoT PromptingยทGraphPrompts

AI Agents

ยทAgentsยทIntroduction to AI AgentsยทAgent ComponentsยทAI Workflows vs. AI AgentsยทWhy Context Engineering?ยทContext Engineering Deep Dive: Building a Deep Research AgentยทFunction Calling in AI AgentsยทDeep Agents

Guides

ยทOptimizing PromptsยทDeep ResearchยทReasoning Llmsยท4O Image Generationยท**Context Engineering Guide**

Applications

ยทLLM Applications & GuidesยทFine-Tuning with GPT-4o ModelsยทFunction Calling with LLMsยทContext Caching with Gemini 1.5 FlashยทGenerating DataยทGenerating Synthetic Dataset for RAGยทTackling Generated Datasets DiversityยทGenerating CodeยทGraduate Job Classification Case StudyยทPrompt Function

Prompt Library

ยทPrompt HubยทLLMs for ClassificationยทLLMs for Code GenerationยทLLMs for CreativityยทLLM EvaluationยทInformation Extraction with LLMsยทImage GenerationยทMathematical Understanding with LLMsยทQuestion Answering with LLMsยทReasoning with LLMsยทText Summarization with LLMsยทTruthfulness in LLMsยทAdversarial Prompting in LLMs

Model Guides

ยทModel Prompting GuidesยทChatGPT Prompt EngineeringยทClaude 3ยทPrompting Guide for Code LlamaยทScaling Instruction-Finetuned Language ModelsยทGetting Started with GeminiยทGemini AdvancedยทGemini 1.5 ProยทGemmaยทGPT-4ยทGrok-1ยทKimi K2.5ยทLlamaยทLlama 3ยทMistral 7B LLMยทMistral LargeยทMixtralยทMixtral 8x22BยทOLMoยทPhi-2ยทSoraยทLLM Collection

Risks and Misuses

ยทRisks & MisusesยทAdversarial Prompting in LLMsยทFactualityยทBiases

Research

ยทLLM Research FindingsยทLLM AgentsยทRetrieval Augmented Generation (RAG) for LLMsยทLLM ReasoningยทHow Faithful are RAG Models?ยทLLM In-Context Recall is Prompt DependentยทReducing Hallucination in Structured Outputs via RAGยทBest Practices and Lessons Learned on Synthetic Data for Language ModelsยทReasoning with Intermediate Revision and Search for LLMsยทEfficient Infinite Context TransformersยทLM-Guided Chain-of-ThoughtยทTrustworthiness in LLMsยทLLM TokenizationยทWhat is Groq?

Papers

ยทPapers

Tools

ยทTools & Libraries

Notebooks

ยทPrompt Engineering Notebooks

Datasets

ยทDatasets

Readings

ยทAdditional Readings

Credits and licence

This course reproduces every section of the Prompt Engineering Guide by Elvis Saravia and DAIR.AI, taken from the guide's repository and released under the MIT licence, as stated in its README. The full licence text and copyright notice appear below.

Diagrams: several figures are reproduced from the research papers they illustrate, including the Tree of Thoughts, Automatic Prompt Engineer and Reflexion papers. Each figure's paper is cited in the text where the figure appears.

Prompt Engineering Guide by Elvis Saravia and DAIR.AI, MIT licence

Visit the original guide at promptingguide.ai

MIT licence

MIT License

Copyright (c) 2022 DAIR.AI

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Related courses

AI Prompting 101
Our friendly first course on asking AI well.Generative AI for Beginners
Go from prompting to building your own AI apps.
Coming soonMastering GitHub Copilot for AI Paired Programming
Put prompting to work writing code.
Coming soon