15. Instructional Prompting (03:33)

Hey everyone, this is Shimron from BitesizedGenius, and Instructional Prompting is the most basic way of interacting with ChatGPT or Google Gemini. This involves providing the model with an instruction or a request before receiving a response, just like asking someone a question and getting an answer. By the end of this video, you will understand how you can use Instructional Prompting as an easy way to provide instructions and get responses. But let me give it to you Bitesized.

So, what is Instructional Prompting? This is where we provide an instruction to the large language model and get a response based on the question we asked. It’s the most basic way of interacting with the model. The benefit of Instructional Prompting is that it’s easy to use, and the information is delivered without much manual work. You ask it a question, and you get a response.

Now let’s move on to some practical examples.

Our first example is a creative one: I want five creative things I could do with my kids’ art, considering that I don’t want to throw them away but they’re also taking up space. ChatGPT generates, as requested, a list of five things we could do with our kids’ art. This is a great way to get creative ideas that you may not have otherwise thought of. Replicating this prompt in Google Gemini results in a similar list of ideas, so there’s not much difference between the two options.

The second example is to request an email template for informing customers about a delay in their order. This template should include an apology, an explanation of the cause of the delay, and the new expected delivery date. ChatGPT provides a template with space to fill in the details, saving us the time of having to formulate one ourselves. Google Gemini delivers a similar template, so there’s little difference here as well.

For the third example, I’ll ask for an explanation of the concept of machine learning in a simple way, along with three real-world examples. ChatGPT provides a simple explanation alongside the examples, which is useful for educational purposes. Google Gemini delivers a shorter and simpler response, including the requested real-world examples.

Our fourth example will be to summarize information—specifically, the plot of Romeo and Juliet in just three sentences. ChatGPT demonstrates how to take information and provide a quick overview, which can help save time or simplify a complex topic. Google Gemini offers a similar result for this simple request.

For the final example, I’ll ask ChatGPT to convince someone why they should start reading regularly, using this as a way to generate a persuasive argument. ChatGPT provides 10 ideas that could be used for various purposes, such as creating a leaflet for a local library. Google Gemini delivers five reasons in shorter responses, offering some concise ideas on why someone should read regularly.

In summary, use Instructional Prompting as a simple way to give an instruction and receive an output, much like a search engine. This can be useful for quickly mocking up written and numerical information, allowing you to focus your time on refinement. Don’t be afraid to experiment with different types of instructional prompts to see what tasks you can accomplish with this simple technique.

This is BitesizedGenius, and I hope you enjoyed!