Introduction and recap

0:09 Hi. Welcome to lesson four of generative AI for beginners. My name is Nitya Narasimhan. I'm a senior cloud advocate on the AI advocacy team, and I'm excited to talk to you about prompt engineering fundamentals. So, today we're going to first start off with a quick recap. If you've been following the series, you know most of the terms. We'll just look at them really briefly. Then we'll talk about the three providers that you can use if you want to get started trying this out on your own.

0:32 But most of this is going to focus on prompt engineering. You're going to learn what it is, you're going to learn why it matters, you're going to learn how to do prompt engineering, and then I'll be leaving you with a sense of intuition for how you can build your own mindset for prompt engineering in practice. So, let's get started. A quick recap. If you've been following this course, these terms are probably familiar to you. What is a prompt?

0:56 So, we're talking generative AI, large language models. A prompt is just a natural language input, the text input that the user provides. And you can almost think of it as the way it programs the model to do what it wants, or the user programs the model to do what they want. So, the response is then what the model does to kind of support the user request by generating content. Fabrication, you might have also heard of it by the term hallucination, is when sometimes the model generates a response that may not be rooted in fact.

1:18 We will talk about that. A base LLM, or a large language model, is really this foundational model that's trained on massive amounts of data. Think of it as a general purpose model that knows everything. But sometimes you want specialized models, things that know how to do a task very well. And that's what you look at for instruction-tuned LLMs, things that are good at summarization, or translation, or code generation. So, what is prompt engineering? Prompt engineering is really this process where you iterate on the prompt, that the text input that you give to the model, study the response, and keep going till the response looks like what you want it to be.