I Discovered The Perfect ChatGPT Prompt Formula
If you use the perfect prompt when you prompt ChatGPT, or if you prompt Bard or Bing, you’re going to be really blown away by the result that you’re going to get. It’s going to save you a ton of time.
So, I’ve studied this for three to four hours a day since GPT 3 came out. I’ve taken a bunch of courses – anything I could get my hands on – and tested a whole lot of different ways to prompt these platforms. I think I’ve crafted a really useful formula that you could implement for pretty much any situation. I want to share that with you in this video, and hopefully, after this, you could just copy and paste this prompt.
Because most people that use ChatGPT, Bard, or Bing are stuck kind of in beginners mode. They just ask a question and then they just try to follow up with AI to get the answer they’re looking for. But if you give the right prompt up front, most of the time you’re gonna get to the answer without having to revise it a ton of different times.
Let me show you exactly how it works. Now, the way I crafted this perfect prompt came from seven best practices that I’ll share with you. This is basically the most effective way of prompt engineering. These seven different steps are going to go inside of our formula, so these steps will help you understand it a little bit better, and I’ll show you some real-life examples that are going to be handy.
Number one is you need to be specific. You want to give as much detail as you can. So, a lot of times, people speed this up because they want to see an output, and they’ll think they’ll revise it after that. But don’t do that; be very specific. Here’s a super simple example: you could ask, “Tell me about all the different dog breeds that exist.” Or the more specific example is, “What are the different breeds of small dogs suitable for apartment living?” Right? You get a lot more specific, so your answer is going to be a lot more relevant.
Number two is you need to state your intent. You need to state that in the prompt. A good example: A lot of times, people say, “Explain quantum physics,” and then you’re going to get some kind of a long response. But if you gave it your intent, so in this case, “I’m helping my son with his science homework. I want to explain quantum physics. Can you explain it in a very simple way?” Right? Then I get a very specific answer that is tailored for that. That is all about stating your intention so that the output is designed for that.
Number three is you gotta use correct spelling and grammar the best you can, even though the model is going to figure that out. A lot of times, a small spelling error is going to send it in the wrong direction. So, even if you do spell check by right-clicking on something that you misspelled, correcting it, there is going to, a lot of times, ensure that your prompt actually gets you the best response.
Number four is direct the output format. Give it some direction on what kind of output you’re looking for. So, “Could you list the steps to break down chocolate cake?” Right? Then it’s going to give you a step-by-step guide. Or, “Could you explain a process in a paragraph?” Then it’s going to know it’s going to give you a paragraph, not a whole bunch of text.
Number five is following up with questions because sometimes, even if you use the perfect prompt, you still need to give it some follow-up to clarify it or change one of the inputs in the formula that I’m going to give you now.
Number six, if you don’t get the result you’re looking for, experiment with different phrasing. Okay, so sometimes the model doesn’t understand how you phrased it, but this perfect formula, if you use it in the order I’m going to show you, is going to get you the result most of the time. But if it doesn’t, just experiment with phrasing and moving things around.
Number seven is sometimes I do a fact-checking prompt. So, I usually get my results, then I feed it back, and I say, “Can you fact-check this statement?” or whatever output I got. This just ensures that it runs through it one more time.
Now, let me show you the formula. This formula, if you use it based on all the different specific information I gave you already, in this order, and I’ll give you some examples, you’re going to get results that are going to be far better than what you’re getting right now. So, here’s the formula: context plus specific information plus your intent plus the response format that you want. That equals the perfect prompt.
So, I’ll give you a couple of different examples here. So, context, in this case, “I’m a beginner cook,” that’s the context, right? I set that up. Then, specific information, “I’m trying to make Italian cuisine.” Okay, then my intent is, “Can you provide me” – this is what I’m asking at my goal – “Can you provide me a simple and easy to follow recipe and put them in numbered order from one through ten?” Okay, and you could give it something else. “Make me lasagna, for example. Can you give me the perfect recipe?” So, you could even add more context in the formula, but this is the perfect prompt. So now, you copy and paste this prompt, and then you’re gonna get a result that’s far better than going back and forth a whole bunch of time, and it’ll save you sometimes five, ten times the amount of the back and forth it usually will take.
Here’s another example. Context is, “I’m a software developer.” The specific information is, “I’m working on a Python project.” Now, the intent is, “Can you explain how to implement exception handling in Python?” And the response format is, “Tell me in a simple paragraph.” So, this will be the perfect prompt: “I’m a software developer working on a Python project. Can you explain how to implement exception handling in Python? Give it to me in a simple paragraph.”
This perfect prompt formula works for ChatGPT, it works for Bard, it works for Bing. I will post it in the description below this video so you can copy and paste the formula, and then every time, just put in those four different pieces every time you give a prompt to any of these softwares. And what I recommend next is you actually learn something called prompt priming. That is things you do before you even give it a prompt. You never ask a question before you prime it. So, I’ll make a video about that that I’ll link here. And I’m also putting together an entire learning platform, Netflix-style learning platform, all about AI, so you don’t fall behind. You get the latest tools, you get the latest tutorials, and you get complete courses, much more detail than what I could do online. So, make sure you sign up for that. The link is in the description.
I hope you found this useful, and I’ll see you next time.