Exploring GPT-3 E-Book

Steve Tingiris

• Herausgeber: Packt Publishing
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 270

Veröffentlichungsjahr: 2021

Beschreibung

Generative Pre-trained Transformer 3 (GPT-3) is a highly advanced language model from OpenAI that can generate written text that is virtually indistinguishable from text written by humans. Whether you have a technical or non-technical background, this book will help you understand and start working with GPT-3 and the OpenAI API.

If you want to get hands-on with leveraging artificial intelligence for natural language processing (NLP) tasks, this easy-to-follow book will help you get started. Beginning with a high-level introduction to NLP and GPT-3, the book takes you through practical examples that show how to leverage the OpenAI API and GPT-3 for text generation, classification, and semantic search. You'll explore the capabilities of the OpenAI API and GPT-3 and find out which NLP use cases GPT-3 is best suited for. You’ll also learn how to use the API and optimize requests for the best possible results. With examples focusing on the OpenAI Playground and easy-to-follow JavaScript and Python code samples, the book illustrates the possible applications of GPT-3 in production.

By the end of this book, you'll understand the best use cases for GPT-3 and how to integrate the OpenAI API in your applications for a wide array of NLP tasks.

Leseprobe

Exploring GPT-3

An unofficial first look at the general-purpose language processing API from OpenAI

Steve Tingiris

BIRMINGHAM—MUMBAI

Exploring GPT-3

Copyright © 2021 Packt Publishing

All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews.

Every effort has been made in the preparation of this book to ensure the accuracy of the information presented. However, the information contained in this book is sold without warranty, either express or implied. Neither the author, nor Packt Publishing or its dealers and distributors, will be held liable for any damages caused or alleged to have been caused directly or indirectly by this book.

Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.

Group Product Manager: Kunal Parikh

Publishing Product Manager: Sunith Shetty

Senior Editor: David Sugarman

Content Development Editor: Nathanya Dias

Technical Editor: Devanshi Ayare

Copy Editor: Safis Editing

Project Coordinator: Aparna Ravikumar Nair

Proofreader: Safis Editing

Indexer: Rekha Nair

Production Designer: Alishon Mendonca

First published: July 2021

Production reference: 2310821

Published by Packt Publishing Ltd.

Livery Place

35 Livery Street

Birmingham

B3 2PB, UK.

ISBN 978-1-80056-319-3

www.packt.com

To my wife, Brigid, for putting up with my constant dabbling for all these years. To my son, Alex, for motivating me to push a little harder. There is no way I would have finished this book without you guys. Thank you!

– Steve

Foreword

What's all the fuss about? Maybe it's the demos. But that would be missing the point. GPT-3 burst into public consciousness in July 2020 not too long after OpenAI first introduced the technical breakthrough in a soberly presented blog post. Introduced in the middle of a global pandemic, rising social unrest, and a US presidential campaign, the message could have been lost entirely. The demos started trickling out in earnest a month later. They ensured that GPT-3 wouldn't be overlooked.  

Technology Review claimed that GPT-3 was shockingly good. Many long-time AI researchers expressed both enthusiasm and surprise at its capabilities. But it was the demos that really captured everyone's attention. With a few lines of sample content and a request, the technology was generating poetry, website programming code, analogies, and answers to math questions, to name just a few examples. No one had ever seen a computer create and respond creatively to such a wide range of queries. 

GPT-3 seemed to possess magical abilities. In many ways, GPT-3 is very simple. It predicts what words are the most likely to follow in a sequence. However, it is also the finest current example of the potential of Generative Adversarial Networks (GANs). And, it has shown that transformers can be applied to language models at an extremely large scale. GPT-3 provides an intriguing new technical capability while simultaneously resetting expectations about what is possible. As just one example, maybe chatbots don't have to choose between a set number of deterministic responses. 

Natural Language Processing (NLP) has improved quickly in recent years. That has led to the more accurate recognition and understanding of speech. At the same time, synthetic speech engines have improved immensely and sound more humanlike each year. What has changed very little is how the systems respond to requests. They are all picking from a predetermined set of responses. GPT-3 offers a capability that enables developers to rethink that approach. But that requires developers to understand the new technology and how to use it. 

Steve Tingiris is the first to take that task on with this book. With clear and precise presentation, Tingiris expertly walks new users through the journey from idea through production of a GPT-3 application. He lays out the principles and steps so developers can turn their ideas into (a new) reality. I look forward to seeing what you build. 

Bret Kinsella

Founder and CEO of Voicebot.ai

Contributors

About the author

Steve Tingiris is the founder and managing director of Dabble Lab, a technology research and services company that helps businesses accelerate learning and adoption of natural language processing, conversational AI, and other emerging technologies. He has been designing and building automation solutions for over 20 years and has consulted on conversational AI projects for companies including Amazon, Google, and Twilio. He also publishes technical tutorials on Dabble Lab's YouTube channel— one of the most popular educational resources for conversational AI developers—and manages several open-source projects, including the Autopilot CLI, Twilio's recommended tool for building Autopilot bots. To connect with Steve, you can find him on GitHub@tingiris, via email to [email protected], or on Twitter @tingiris.

Acknowledgments

This book is the result of contributions from friends, colleagues, and many members of the OpenAI community. There are far too many people to mention everyone by name but for those who contributed directly or reviewed early drafts, I want to extend a special thank you.

First, thank you to the entire Packt team. If Sunith had not reached out with the idea of creating a book on GPT-3, this project would never have started. If David, Gebin, Nathanya, Aishwarya, Roshan, and Devanshi weren't involved – the book might never have gotten finished. Thanks to each of you and everyone else on the Packt team who made this possible.

Thanks to Russell, Bakz, Ryan, and Bram for agreeing to be technical reviewers. You guys were so helpful to me in the OpenAI Slack channel when I first got started with GPT-3, I feel so fortunate to have had the opportunity to collaborate together on this book. Again, thank you!

Next, thank you to my work family at Dabble Lab. Kirk Owen for providing early feedback that helped refine the direction, and Mohamad Khalid, Manuel Fernandez, Mark Hovsepyan, Sohini Pattanayak, Shubham Prakash, and Daniela Ramirez, for picking up the slack while I spent much more time than I'd anticipated working on the book.

Finally, I'd like to thank the OpenAI team and the OpenAI community for all of the support and feedback throughout the project, including: Ashley Pilipiszyn, Mark Clintsman, Minal Chhatbar, Rene Diaz, Dan Shaw, Chris Fong, Dariusz Gross, Cristi Vlad, Jonathon Sauceda, Marc-Andre Schenk, Steven Kuo, Narendran Thillaisthanam, Matthew Benites, Manav Goel, Shubham Amraotkar, Mystici Mentis, Fred Zimmerman, Dmitry Kearo, CL Kim, Sudip Lingthep, Joakim Flink, Shubham Saboo, Pedro Ribeiro, Richard Klein, Steve Hoyt, Nicolas Garrel, Sebastian Derewicki, Vikram Pandya, Geoff Davis, Nelson Pereira, Heng Gu, Joey Bertschler, Surendra Reddy, James Morgan, Jon Oakes, Jeetendra K Sharma, Jim Taylor, Rebecca Johnson, Travis Barton, Herber Scrap, Pablo del Ser, Devin Bean, Nik K, Jason Boog, Mohak Agarwal, Sebastian Elliott, and Bjarne Carstensen.

About the reviewers

Russell Foltz-Smith has 20+ years of experience in tech as a developer, business development leader, executive, and researcher. He maintains a focus on search engines, scientific computing, and media platforms. Russ advises tech start-ups, mentors entrepreneurs, and is CTO of Maslo.ai, an empathetic computing platform. Russ is a visual artist and educator. He co-founded a k-12 school with his wife in 2012 in Venice, CA, where they live with their daughters and endless stacks of books.

Bakz Awan is an IT consultant and YouTube host of the channel Bakz T. Future. on YouTube. Bakz shares new concepts and ideas possible through GPT-3, while also providing tips and advice to beginners.

Ryan Sevey has been a leader for over 10 year in the machine learning space. Mr. Sevey founded Nexosis in 2015 and successfully exited to DataRobot where he became the General Manager responsible for self-service and developer experience. Prior to Nexosis, Mr. Sevey spent many years as an offensive cybersecurity professional working for HP and American Electric Power.

Bram Adams works with technology to tell stories. As technology evolves, the opportunities evolve as well. Bram will be there to invent dumb stuff with it.

Table of Contents

Preface

Section 1: Understanding GPT-3 and the OpenAI API

Chapter 1: Introducing GPT-3 and the OpenAI API

Technical requirements

Introduction to GPT-3

Simplifying NLP

What exactly is GPT-3?

Democratizing NLP

Understanding prompts, completions, and tokens

Prompts

Completions

Tokens

Introducing Davinci, Babbage, Curie, and Ada

Davinci

Curie

Babbage

Ada

Content filtering model

Instruct models

A snapshot in time

Understanding GPT-3 risks

Inappropriate or offensive results

Potential for malicious use

Summary

Chapter 2: GPT-3 Applications and Use Cases

Technical requirements

Understanding general GPT-3 use cases

Introducing the Playground

Getting started with the Playground

Handling text generation and classification tasks

Text generation

Text classification

Understanding semantic search

The Semantic Search tool

Summary

Section 2: Getting Started with GPT-3

Chapter 3: Working with the OpenAI Playground

Technical requirements

Exploring the OpenAI developer console

Developer documentation

Developer resources

Accounts and organizations

Pricing and billing

Usage reporting

Member management

Diving deeper into the Playground

Choosing the right engine

Response length

Temperature and Top P

Frequency and presence penalty

Best of

Stop sequence

Inject Start Text and Inject Restart Text

Show Probabilities

Working with presets

Grammatical Standard English

Text to command

Parse unstructured data

Summary

Chapter 4: Working with the OpenAI API

Technical requirements

Understanding APIs

Getting familiar with HTTP

Uniform resource identifiers

HTTP methods

The HTTP body

HTTP headers

HTTP response status codes

Reviewing the OpenAI API endpoints

List Engines

Retrieve Engine

Create Completions

Semantic Search

Introducing CURL and Postman

Understanding API authentication

Keeping API keys private

Making an authenticated request to the OpenAI API

Working with multiple organizations

Introducing JSON

Using the Completions endpoint

Using the Semantic Search endpoint

Summary

Chapter 5: Calling the OpenAI API in Code

Technical requirements

Choosing your programming language

Introducing replit

Creating a repl

Setting your OpenAI API key as an environment variable

Understanding and creating the .replit file

Using the OpenAI API with Node.js/JavaScript

Calling the engines endpoint

Calling the Completions endpoint

Calling the search endpoint

Using the OpenAI API in Python

Calling the completions endpoint

Calling the search endpoint

Using other programming languages

Summary

Section 3: Using the OpenAI API

Chapter 6: Content Filtering

Technical requirements

Preventing inappropriate and offensive results

Understanding content filtering

Testing the content filtering process

Filtering content with JavaScript

Flagging unsafe words with Node.js/JavaScript

Filtering content with Python

Flagging unsafe words with Python

Summary

Chapter 7: Generating and Transforming Text

Technical requirements

Using the examples

Generating content and lists

Dumb joke generator

Mars facts (in most cases)

Webinar description generator

Book suggestions

Children's book generator

Translating and transforming text

Acronym translator

English to Spanish

JavaScript to Python

Fifth-grade summary

Grammar correction

Extracting text

Extracting keywords

HTML parsing

Extracting a postal address

Extracting an email address

Creating chatbots

A simple chatbot

Summary

Chapter 8: Classifying and Categorizing Text

Technical requirements

Understanding text classification

Using the completions endpoint for text classification

Content filtering is a text classification task

Introducing the classifications endpoint

Uploading files

Implementing sentiment analysis

Assigning an ESRB rating to text

Classifying text by language

Classifying text from keywords

Summary

Chapter 9: Building a GPT-3-Powered Question-Answering App

Technical requirements

Introducing GPT Answers

GPT Answers technical overview

Hosting the app

Introducing the Answers endpoint

Setting up and testing Express

Creating the API endpoint for GPT Answers

Creating the API endpoint

Testing our API with Postman

Creating the GPT Answers user interface

Integrating the Answers endpoint

Generating relevant and factual answers

Using files with the Answers endpoint

Summary

Chapter 10: Going Live with OpenAI-Powered Apps

Technical requirements

Going live

Understanding use case guidelines

Addressing potential approval issues

Content filtering

Input and output lengths

Request rate limiting

Completing the pre-launch review request

High-level use case questions

Security and risk mitigation questions

Growth plan questions

Wrapping-up questions

Summary

Why subscribe?

Other Books You May Enjoy

Preface

What if this book was written by artificial intelligence? Would you read it? I hope so because parts of it were. Yes, GPT-3 was used to create parts of this book. It's a bit meta I know, a book about GPT-3 written by GPT-3. But creating content is one of the many great uses for GPT-3. So why not? Also, for me, content generation was the use case that most piqued my interest. I wondered if GPT-3 could be used in a product I was working on to automate the generation of technical learning material.

You probably also have a specific reason why you're interested in GPT-3. Perhaps it's intellectual curiosity. Or maybe you have an idea that you think GPT-3 can enable. You've likely seen online demos of GPT-3 generating content, writing code, penning poetry, or something else, and you're wondering if GPT-3 could be used for an idea you have. If so, this book was written specifically for you.

My goal for this book is to provide a practical resource to help you get started with GPT-3, as quickly as possible, without any required technical background. That said, as I write this, GPT-3 is still in private beta. So, everyone is learning as they go. But the one thing I've learned for sure is that the possible applications for GPT-3 are vast and there is no way to know all of what's possible, let alone get it into a book. So, I hope this book makes getting started easy, but I also hope it's just the beginning of your journey Exploring GPT-3

Who this book is for

This book was written for anyone with an interest in NLP or learning GPT-3 – with or without a technical background. Developers, product managers, entrepreneurs, and hobbyists who want to learn about NLP, AI, and GPT-3 will find this book useful. Basic computer skills are all you need to get the most out of the book. While experience with a modern programming language is helpful, it's not required. The code examples provided are beginner friendly and easy to follow, even if you're brand new to writing code.

What this book covers

Chapter 1, Introducing GPT-3 and the OpenAI API, is a high-level introduction to GPT-3 and the OpenAI API.

Chapter 2, GPT-3 Applications and Use Cases, is an overview of core GPT-3 use cases: text generation, classification, and semantic search.

Chapter 3, Working with the OpenAI Playground, is a semi-deep dive into the OpenAI Playground and the developer portal.

Chapter 4, Working with the OpenAI API, is an introduction to calling the OpenAI API using Postman.

Chapter 5, Calling the OpenAI API in Code, is an introduction to using the OpenAI API with both Node.js/JavaScript and Python.

Chapter 6, Content Filtering, explains how to implement content filtering.

Chapter 7, Generating and Transforming Text, contains code and prompt examples for generating and transforming text.

Chapter 8, Classifying and Categorizing Text, takes a closer look at text classification and the OpenAI API Classification endpoint.

Chapter 9, Building a GPT-3 Powered Question-Answering App, explains how to build a functional GPT-3 powered web knowledge base.

Chapter 10, Going Live with OpenAI-Powered Apps, explains the OpenAI application review and approval process and discusses getting ready for a review.

To get the most out of this book

All of the code examples in this book were written using a web-based Integrated Development Environment (IDE) from replit.com. A free replit.com account is sufficient to follow the examples. To use replit.com, all that is required is a modern web browser and a replit.com account. The code has also been tested on macOS using Visual Studio Code, although it should work with any code editor and properly configured operating system. Code examples are provided in both Node.js/JavaScript and Python. For Node.js, version 12.16.1 is used and for Python, version 3.8.2 is used.

All of the code examples will require an OpenAI API Key and access to the OpenAI API. You can request access to the OpenAI API by visiting https://openai.com/api.

If you are using the digital version of this book, we advise you to type the code yourself or access the code via the GitHub repository (link available in the next section). Doing so will help you avoid any potential errors related to the copying and pasting of code.

Download the example code files

You can download the example code files for this book from GitHub at https://github.com/PacktPublishing/Exploring-GPT-3. In case there's an update to the code, it will be updated on the existing GitHub repository.

We also have other code bundles from our rich catalog of books and videos available at https://github.com/PacktPublishing/. Check them out!

Download the color images

We also provide a PDF file that has color images of the screenshots/diagrams used in this book. You can download it here: https://static.packt-cdn.com/downloads/9781800563193_ColorImages.pdf.

Conventions used

There are a number of text conventions used throughout this book.

Code in text: Indicates code words in text, database table names, folder names, filenames, file extensions, pathnames, dummy URLs, user input, and Twitter handles. Here is an example: "However, suppose you don't want the completion to generate the human side of the conversation and you want to use the label AI: rather than Assistant:?"

A block of code is set as follows:

English: I do not speak Spanish Spanish:

Bold: Indicates a new term, an important word, or words that you see onscreen. For example, words in menus or dialog boxes appear in the text like this. Here is an example: "Each subsequent time the Submit button is clicked."

Tips or important notes

Appear like this.

Get in touch

Feedback from our readers is always welcome.

General feedback: If you have questions about any aspect of this book, mention the book title in the subject of your message and email us at [email protected].

Errata: Although we have taken every care to ensure the accuracy of our content, mistakes do happen. If you have found a mistake in this book, we would be grateful if you would report this to us. Please visit www.packtpub.com/support/errata, selecting your book, clicking on the Errata Submission Form link, and entering the details.

Piracy: If you come across any illegal copies of our works in any form on the Internet, we would be grateful if you would provide us with the location address or website name. Please contact us at [email protected] with a link to the material.

If you are interested in becoming an author: If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, please visit authors.packtpub.com.

Reviews

Please leave a review. Once you have read and used this book, why not leave a review on the site that you purchased it from? Potential readers can then see and use your unbiased opinion to make purchase decisions, we at Packt can understand what you think about our products, and our authors can see your feedback on their book. Thank you!

For more information about Packt, please visit packt.com.

Section 1: Understanding GPT-3 and the OpenAI API

The objective of this section is to provide you with a high-level introduction to GPT-3 and the OpenAI API and to show how easy it is to get started with. The goal is to engage you with fun examples that are quick and simple to implement.

This section comprises the following chapters:

Chapter 1, Introducing GPT-3 and the OpenAI APIChapter 2, GPT-3 Applications and Use Cases

Chapter 1: Introducing GPT-3 and the OpenAI API

The buzz about Generative Pre-trained Transformer Version 3 (GPT-3) started with a blog post from a leading Artificial Intelligence (AI) research lab, OpenAI, on June 11, 2020. The post began as follows:

We're releasing an API for accessing new AI models developed by OpenAI. Unlike most AI systems which are designed for one use-case, the API today provides a general-purpose "text in, text out" interface, allowing users to try it on virtually any English language task.

Online demos from early beta testers soon followed—some seemed too good to be true. GPT-3 was writing articles, penning poetry, answering questions, chatting with lifelike responses, translating text from one language to another, summarizing complex documents, and even writing code. The demos were incredibly impressive—things we hadn't seen a general-purpose AI system do before—but equally impressive was that many of the demos were created by people with a limited or no formal background in AI and Machine Learning (ML). GPT-3 had raised the bar, not just in terms of the technology, but also in terms of AI accessibility.

GPT-3 is a general-purpose language processing AI model that practically anybody can understand and start using in a matter of minutes. You don't need a Doctor of Philosophy (PhD) in computer science—you don't even need to know how to write code. In fact, everything you'll need to get started is right here in this book. We'll begin in this chapter with the following topics:

Introduction to GPT-3Democratizing NLPUnderstanding prompts, completions, and tokensIntroducing Davinci, Babbage, Curie, and AdaUnderstanding GPT-3 risks

Technical requirements

This chapter requires you to have access to the OpenAIApplication Programming Interface (API). You can register for API access by visiting https://openai.com/.

Introduction to GPT-3

In short, GPT-3 is a language model: a statistical model that calculates the probability distribution over a sequence of words. In other words, GPT-3 is a system for guessing which text comes next when text is given as an input.

Now, before we delve further into what GPT-3 is, let's cover a brief introduction (or refresher) on Natural Language Processing (NLP).

Simplifying NLP

NLP is a branch of AI that focuses on the use of natural human language for various computing applications. NLP is a broad category that encompasses many different types of language processing tasks, including sentiment analysis, speech recognition, machine translation, text generation, and text summarization, to name but a few.

In NLP, language models are used to calculate the probability distribution over a sequence of words. Language models are essential because of the extremely complex and nuanced nature of human languages. For example, pay in full and painful or tee time and teatime sound alike but have very different meanings. A phrase such as she's on fire could be literal or figurative, and words such as big and large can be used interchangeably in some cases but not in others—for example, using the word big to refer to an older sibling wouldn't have the same meaning as using the word large. Thus, language models are used to deal with this complexity, but that's easier said than done.

While understanding things such as word meanings and their appropriate usage seems trivial to humans, NLP tasks can be challenging for machines. This is especially true for more complex language processing tasks such as recognizing irony or sarcasm—tasks that even challenge humans at times.

Today, the best technical approach to a given NLP task depends on the task. So, most of the best-performing, state-of-the-art (SOTA) NLP systems are specialized systems that have been fine-tuned for a single purpose or a narrow range of tasks. Ideally, however, a single system could successfully handle any NLP task. That's the goal of GPT-3: to provide a general-purpose AI system for NLP. So, even though the best-performing NLP systems today tend to be specialized, purpose-built systems, GPT-3 achieves SOTA performance on a number of common NLP tasks, showing the potential for a future general-purpose NLP system that could provide SOTA performance for any NLP task.

What exactly is GPT-3?

Although GPT-3 is a general-purpose NLP system, it really just does one thing: it predicts what comes next based on the text that is provided as input. But it turns out that, with the right architecture and enough data, this one thing can handle a stunning array of language processing tasks.

GPT-3 is the third version of the GPT language model from OpenAI. So, although it started to become popular in the summer of 2020, the first version of GPT was announced 2 years earlier, and the following version, GPT-2, was announced in February 2019. But even though GPT-3 is the third version, the general system design and architecture hasn't changed much from GPT-2. There is one big difference, however, and that's the size of the dataset that was used for training.

GPT-3 was trained with a massive dataset comprised of text from the internet, books, and other sources, containing roughly 57 billion words and 175 billion parameters. That's 10 times larger than GPT-2 and the next-largest language model. To put the model size into perspective, the average human might read, write, speak, and hear upward of a billion words in an entire lifetime. So, GPT-3 has been trained on an estimated 57 times the number of words most humans will ever process.

The GPT-3 language model is massive, so it isn't something you'll be downloading and dabbling with on your laptop. But even if you could (which you can't because it's not available to download), it would cost millions of dollars in computing resources each time you wanted to build the model. This would put GPT-3 out of reach for most small companies and virtually all individuals if you had to rely on your own computer resource to use it. Thankfully, you don't. OpenAI makes GPT-3 available through an API that is both affordable and easy to use. So, anyone can use some of the most advanced AI ever created!

Democratizing NLP

Anyone can use GPT-3 with access to the OpenAI API. The API is a general-purpose text in, text out interface that could be used for virtually any language task. To use the API, you simply pass in text and get a text response back. The task might be to do sentiment analysis, write an article, answer a question, or summarize a document. It doesn't matter, as far as the API is concerned—it's all done the same way, which makes using the API easy enough for just about anyone to use, even non-programmers.

The text you pass in is referred to as a prompt, and the returned text is called a completion. A prompt is used by GPT-3 to determine how best to complete the task. In the simplest case, a prompt can provide a few words to get started with. For example, if the prompt was If today is Monday, tomorrow is, GPT-3 would likely respond with Tuesday, along with some additional text such as If today is Tuesday, tomorrow is Wednesday, and so on. This means that what you get out of GPT-3 depends on what you send to it.

As you might guess, the quality of a completion depends heavily on the prompt. GPT-3 uses all of the text in a prompt to help generate the most relevant completion. Each and every word, along with how the prompt is structured, helps improve the language model prediction results. So, understanding how to write and test prompts is the key to unlocking GPT-3's true potential.

Understanding prompts, completions, and tokens

Literally any text can be used as a prompt—send some text in and get some text back. However, as entertaining as it can be to see what GPT-3 does with random strings, the real power comes from understanding how to write effective prompts.

Prompts

Prompts are how you get GPT-3 to do what you want. It's like programming, but with plain English. So, you have to know what you're trying to accomplish, but rather than writing code, you use words and plain text.

When you're writing prompts, the main thing to keep in mind is that GPT-3 is trying to figure out which text should come next, so including things such as instructions and examples provides context that helps the model figure out the best possible completion. Also, quality matters— for example, spelling, unclear text, and the number of examples provided will have an effect on the quality of the completion.

Another key consideration is the prompt size. While a prompt can be any text, the prompt and the resulting completion must add up to fewer than 2,048 tokens. We'll discuss tokens a bit later in this chapter, but that's roughly 1,500 words.

So, a prompt can be any text, and there aren't hard and fast rules that must be followed like there are when you're writing code. However, there are some guidelines for structuring your prompt text that can be helpful in getting the best results.

Different kinds of prompts

We'll dive deep into prompt writing throughout this book, but let's start with the different prompt types. These are outlined as follows:

Zero-shot promptsOne-shot promptsFew-shot prompts

Zero-shot prompts

A zero-shot prompt is the simplest type of prompt. It only provides a description of a task, or some text for GPT-3 to get started with. Again, it could literally be anything: a question, the start of a story, instructions—anything, but the clearer your prompt text is, the easier it will be for GPT-3 to understand what should come next. Here is an example of a zero-shot prompt for generating an email message. The completion will pick up where the prompt ends—in this case, after Subject::

Write an email to my friend Jay from me Steve thanking him for covering my shift this past Friday. Tell him to let me know if I can ever return the favor.

Subject:

The following screenshot is taken from a web-based testing tool called the Playground. We'll discuss the Playground more in Chapter 2, GPT-3 Applications and Use Cases, and Chapter 3, Working with the OpenAI Playground, but for now we'll just use it to show the completion generated by GPT-3 as a result of the preceding prompt. Note that the original prompt text is bold, and the completion shows as regular text:

Figure 1.1 – Zero-shot prompt example

So, a zero-shot prompt is just a few words or a short description of a task without any examples. Sometimes this is all GPT-3 needs to complete the task. Other times, you may need to include one or more examples. A prompt that provides a single example is referred to as a one-shot prompt.

One-shot prompts

A one-shot prompt provides one example that GPT-3 can use to learn how to best complete a task. Here is an example of a one-shot prompt that provides a task description (the first line) and a single example (the second line):

A list of actors in the movie Star Wars

1. Mark Hamill: Luke Skywalker

From just the description and the one example, GPT-3 learns what the task is and that it should be completed. In this example, the task is to create a list of actors from the movie Star Wars. The following screenshot shows the completion generated from this prompt:

Figure 1.2 – One-shot prompt example

The one-shot prompt works great for lists and commonly understood patterns. But sometimes you'll need more than one example. When that's the case you'll use a few-shot prompt.

Few-shot prompts

A few-shot prompt provides multiple examples—typically, 10 to 100. Multiple examples can be useful for showing a pattern that GPT-3 should continue. Few-shot prompts and more examples will likely increase the quality of the completion because the prompt provides more for GPT-3 to learn from.

Here is an example of a few-shot prompt to generate a simulated conversation. Notice that the examples provide a back-and-forth dialog, with things that might be said in a conversation:

This is a conversation between Steve, the author of the book Exploring GPT-3 and someone who is reading the book.

Reader: Why did you decide to write the book?

Steve: Because I'm super fascinated by GPT-3 and emerging technology in general.

Reader: What will I learn from this book?

Steve: The book provides an introduction to GPT-3 from OpenAI. You'll learn what GPT-3 is and how to get started using it.

Reader: Do I need to be a coder to follow along?

Steve: No. Even if you've never written a line of code before, you'll be able to follow along just fine.

Reader:

In the following screenshot, you can see that GPT-3 continues the simulated conversation that was started in the examples provided in the prompt:

Figure 1.3 – Few-shot prompt example

Now that you understand the different prompt types, let's take a look at some prompt examples.

Prompt examples

The OpenAI API can handle a variety of tasks. The possibilities range from generating original stories to performing complex text analysis, and everything in between. To get familiar with the kinds of tasks GPT-3 can perform, OpenAI provides a number of prompt examples. You can find example prompts in the Playground and in the OpenAI documentation.

In the Playground, the examples are referred to as presets