English
Share
Sign In
The easiest way to give commands to artificial intelligence
A language model is an artificial neural network modeled after the neural network of the human brain, and it is well structured to create a 'model' specialized for a specific task. Then, hundreds of millions to trillions of text data are put into this model to complete it. Of course, we don't call it a language model just by putting a lot of data into the artificial neural network. We call it a language model, and depending on the size of the data, a large language model, if we make it work meaningfully as intended.
The red box is an area of concern for those researching and developing artificial intelligence, and users should pay attention to the black box on the far right.
We give a command to the language model created in this way to trigger the desired action (Prompt). We call this a prompt by reading the pronunciation as it is. This is actually a fancy explanation, and it comes from the fact that developers often open a 'command prompt' when giving instructions on the terminal. (The etymology is the same.)
It is often called prompt engineering because it is designed like an engineering design , and prompt design because it is thoroughly planned and entered . Personally, I think it is nothing more than a marketing term. In fact, it is currently in a transitional period, so the names are changing subtly. However, I will use the term prompting, which is the most recently used term. In the end, it is all the same. Since it is a method of giving orders to artificial intelligence, any term will do as long as the meaning is consistent.
We will learn how to give orders to AI by reading these 42 documents. To be more precise, we will talk about how to make AI do what we want . If we know where the nail is driven and how to do it, it doesn't matter whether the tool we are given is a hammer, a shovel, or a rock. In other words, if we clearly understand how to give orders, it doesn't really matter who receives the orders. We understand how language models work and learn how to give orders effectively.
Fortunately, AI doesn't get angry when you bother it yet, and it doesn't have a vengeful desire to exterminate humans. So the scope of what we can do is quite wide, and we can even ask experimental(?) questions, so let's do what we can without being restricted by the tools.
This is not knowledge that changes every time a new model and tool comes out. The reason why I gave this book for free is simple. It is strange, but you are happier if you spend that money on delicious meat, alcohol, dessert, salad, and coffee. Just spend that money on something else.
ⓒ 2023. Haebom, all rights reserved.
It may be used for commercial purposes with permission from the copyright holder, provided the source is cited.