Welcome to this course on ChatGPT Prom Engineering for Developers. I am thrilled to have with me Isa, who is a member of the technical staff of OpenAI and has built the popular ChatGPT retrieval plug-in. She has also contributed to the OpenAI Cookbook that teaches people prompting. I am thrilled to be here and share some prompting best practices with you.
There has been a lot of material on the internet for prompting with articles like ‘30 Prompts Everyone Has to Know’. However, most of it has been focused on the ChatGPT web user interface, which is used for specific and often one-off tasks. But the power of large language models as a developer lies in using API calls to quickly build software applications.
In this course, we will share with you the possibilities of what you can do with language models and best practices for using them. We will cover prompting best practices for software development and common use cases such as summarizing, inferring, transforming, and expanding. You will also learn how to build a chatbot using a language model.
There are broadly two types of language models: base language models (Bas LM) and instruction-tuned language models (Instruction-Tuned LM). Bas LM is trained to predict the next word based on text training data, while Instruction-Tuned LM is trained to follow instructions. Instruction-Tuned LM is recommended for most practical applications as it is easier to use and safer.
Before we move on, I would like to acknowledge the team from OpenAI and Deep Learning who have contributed to the materials for this course. They have been involved in brainstorming and vetting the materials to put together the curriculum.
When using an instruction-tuned language model, think of giving instructions to another person who is smart but doesn’t know the specifics of your task. Clear instructions and specifying the tone of the text can help the language model generate the desired output. It is also important to give the language model time to think.
In the next video, you will see examples of how to be clear and specific in your instructions, which are important principles of prompting language models.
