GPT-3: Building Innovative NLP Products Using Large Language Models

Description

GPT-3: NLP with LLMs is a unique, pragmatic take on Generative Pre-trained Transformer 3, the famous AI language model launched by OpenAI in 2020. This model is capable of tackling a wide array of tasks, like conversation, text completion, and even coding with stunningly good performance. Since its launch, the API has powered a staggering number of applications that have now grown into full-fledged startups generating business value. This book will be a deep dive into what GPT-3 is, why it is important, what it can do, what has already been done with it, how to get access to it, and how one can build a GPT-3 powered product from scratch. This book is for anyone who wants to understand the scope and nature of GPT-3. The book will evaluate the GPT-3 API from multiple perspectives and discuss the various components of the new, burgeoning economy enabled by GPT-3. This book will look at the influence of GPT-3 on important AI trends like creator economy, no-code, and Artificial General Intelligence and will equip the readers to structure their imaginative ideas and convert them from mere concepts to reality. Preface Conventions Used in This Book Using Code Examples O’Reilly Online Learning How to Contact Us Acknowledgments From Sandra From Shubham 1. The Era of Large Language Models Natural Language Processing: Under the Hood Language Models: Bigger and Better The Generative Pre-Trained Transformer: GPT-3 Generative Models Pre-trained Models Transformer Models A Brief History of GPT-3 GPT-1 GPT-2 GPT-3 Accessing the OpenAI API 2. Using the OpenAI API Navigating the OpenAI Playground Prompt Engineering and Design How the OpenAI API Works Execution Engine Response Length Temperature and Top P Frequency and Presence Penalties Best Of Stop Sequence Inject Start Text and Inject Restart Text Show Probabilities Execution Engines Davinci Curie Babbage Ada Instruct Series Endpoints List Engines Retrieve Engine Completions Semantic Search Files Classification (Beta) Answers (Beta) Embeddings Customizing GPT-3 Apps Powered by Customized GPT-3 Models How to Customize GPT-3 for Your Application Tokens Pricing GPT-3’s Performance on Conventional NLP Tasks Text Classification Named Entity Recognition Text Summarization Text Generation Conclusion 3. Programming with GPT-3 Using the OpenAI API with Python Using the OpenAI API with Go Using the OpenAI API with Java GPT-3 Sandbox Powered by Streamlit Going Live with GPT-3-Powered Applications Conclusion 4. GPT-3 as a Launchpad for Next-Generation Start-ups Model-as-a-Service The New Start-up Ecosystem: Case Studies Creative Applications of GPT-3: Fable Studio Data Analysis Applications of GPT-3: Viable Chatbot Applications of GPT-3: Quickchat Marketing Applications of GPT-3: Copysmith Coding Applications of GPT-3: Stenography An Investor’s Perspective on the GPT-3 Start-up Ecosystem Conclusion 5. GPT-3 for Corporations Case Study: GitHub Copilot How It Works Developing Copilot No-Code/Low-Code: Simplifying Software Development? Scaling with the API What’s Next for GitHub Copilot? Case Study: Algolia Answers Evaluating NLP Options Data Privacy Cost Speed and Latency Lessons Learned Case Study: Microsoft Azure OpenAI Service A Partnership That Was Meant to Be An Azure-Native OpenAI API Resource Management Security and Data Privacy Model-as-a-Service at the Enterprise Level Other Microsoft AI and ML Services Advice for Enterprises OpenAI or Azure OpenAI Service: Which Should You Use? Conclusion 6. Challenges, Controversies, and Shortcomings The Challenge of AI Bias Anti-Bias Countermeasures Low-Quality Content and the Spread of Misinformation The Environmental Impact of LLMs Proceeding with Caution Conclusion 7. Democratizing Access to AI No Code? No Problem! Access and Model-as-a-Service Conclusion Index About the Authors