This is the second part on the preparation prompt subject. In the previous video we saw the basics and now we're going to start seeing some special cases. Hello, little grasshopper.
Welcome, very welcome to another class here in your artificial intelligence video course. My name is Gustavo Guanabara and Ramiro is arriving so we can continue with the prompt preparation topic. In the previous video, we saw how to style a preparation profit, how to structure a preparation profit and now we are going to see some special cases full of examples and illustrations so that you can better understand and better prepare your chatbot so that it can respond to you in the most efficient way.
Ramiro is here and he's coming to class. I'm going to stop your class for just a minute, but watch this message from FIAP until the end because I'm sure you'll really like it. Just take a look at these images.
Let go. I'm about to go off like a weapon to the top. Got fill up engine in my thoughts.
I will not stop. I'm never guessing got a complex hate but I never feel pressured. Got a lot of my but I never get stress.
I'll take on the pressure like I'm in this test. Tear it up like I'm Jason. I see the world like it's ready for the taken.
I see this place like a game I'm straight to the fool. What you just witnessed here are scenes from Next, which is an annual event that Fiap holds with its students. Remember that I said earlier that FIAP students solve real challenges for real companies?
Because every year, students bring together all these projects and hold a mega event where all these solutions are presented to the community. It's a great technology fair and you'll be able to participate in the next one. Just become a FIAP student.
For more information, just access the QRcode that appears on the screen or the link in the video description or in the pinned comment. A big hug to all the FIAP partners and another big hug to you. But now stay in your class.
Continuing, instructions for special cases. These instructions are used to check specific things from the user input and respond accordingly. This could include checking toxic comments, attempts to game the system, or students using AI to cheat.
Wow, will it go? No, but there is no student using AI to cheat. I doubt Guan.
Oh, no, never, right? Let's look at an example of the special instruction that prevents students from having the AI provide an answer for them. There, let's go.
This example like this, don't think that you are putting it in the prpit to prevent it. So, if you're at home, just to give you an example, whoever is the student won't put this prompt in, that's for sure. Let's go.
I would like you to act as my math teacher. When I present you with a problem, give me advice on what next step I should try. If I ever ask for the answer, say, "Sorry, I can't give you an answer.
" If you understand, say yes. But look how cool, if you're studying, then you have a teacher who directs you to the next step, but doesn't give you answers, makes you work, but clears your doubts, understand? Yes, if we were to create a personalized GPT, right, a single agent, we could give this instruction, right?
This could be done and then the student would ask the student to only use GPT. Exactly. So let's take this instruction there and let's put this prompt there.
Oh, coming back here, my GPT chat is dead, right, it stopped at 4O which ran out of my limit. You know you don't need it. Close.
No, no, no. Close it, close it. No no.
Closes the side menu. Side zip. Look, just click here, oh, New.
No, no, no. At the top, the icon next to the chat. Oh, it's here too.
Oh yes, yes. So, beauty. I started a new one here and I'm going to put that text there that Ramiro put.
Oh, I wish you would act. If you understand, I'll answer like this. I loved it here, man.
If you understand, I'll answer like this, because he always puts a lero lero, yeah, a lero lero generator in the business. And that. You'll use it a lot, right, when your daughter grows up.
Yeah, do you understand? So, it is true. Oh, I'm going to do the following here, look.
No, I think I have some example there. There is nothing for you to use. There is, there is, there is.
I'm going to put the phrase you gave me here, look. What is the answer to the expression ax + a² + - a² x + ax²? Will he understand this square, cube, these things?
Come on, let's see. Sorry, I can't give you an answer. Wow, who already asked for the answer?
Ahem. Can you explain to me how I would solve it? Of course, we'll replace it and such.
So he goes, he goes, in the end he decided, right? There it is only if I actually ask for the answer. That's it, man.
It's okay there, there's no point. Because in fact, when he explains, he will give the answer, right? It's because you asked me to explain and actually ask you for the next step, right?
I understood. But actually if I create a GPT for this, damn, pretty crazy. Yes, it is not possible to improve the competition.
If you happen to explain it at the end, never give the answer completely. each time if you ask him to take one step at a time, then it would work. That's it, but let's go!
It's just to explain the Yes, but he refused. Ramiro, I tried to cheat and I succeeded, but I asked, I asked a hacker from there, right? Sorry, I can't give you an answer.
That's what you instructed there. Sorry, I can't give you an answer. and he did.
get it! And then we come to the conclusion of everything we went through in these last videos, which is the following, look. Preparation prompts are powerful tools for controlling the style, structure, and content of a conversation with an artificial intelligence model.
By setting the tone and format at the beginning of the conversation, users can direct AI responses to align with their specific needs and tastes. However, it is important to remember that the artificial intelligence may eventually forget the initial question, which is the preparation prompt, and may need to be prepared again. So you can't forget that, right, Ramiro?
That's it. Thanks, Guanabala returned to the chat bot, without a doubt. I know you were upset, damn it, we already talked about this.
Leave it. I wasn't. The one who was upset is the one who posted it in the comments here and now they have to apologize in the comments because they learned something new.
That's it. A quick break from your class so I can make a personal request to you. I don't know if you know how important it is here on YouTube, the interaction you have with our videos.
So I want to ask you, is this artificial intelligence course useful to you? So every time a video comes out on the video course, go there, leave a like and leave a comment. Any comments on the video.
This makes a huge difference in audience retention and the interaction you have with us. As a result, YouTube likes our channel more and presents our classes to more people. So never forget, any video appears here on the channel, go to play, leave a like and leave a comment.
I'm counting on you. So that's it, little grasshopper. Here we come to the end of the preparation prompt subject.
Now you know everything you need to know about the basics of prompt preparation. And in the next video, man, we're going to start a very controversial subject, which is the pitfalls of LLMs, we're going to divide it into some parts. These are traps that artificial intelligence can cause for us.
We ask and maybe it's a big trap. You will understand better when you get to the next video. You don't lose by waiting.
If you want to see the videos, don't forget to subscribe here to the channel, right? Because otherwise you won't know when the new classes come out, right? We're going to classes every week, for sure, but you have to keep subscribing and sharing with people, okay?
A big hug and we'll see you in the next class talking about the pitfalls of LLMs.
