How Do Artificial Intelligence Applications Like ChatGPT Work?

In today’s world, artificial intelligence The term has become a concept that almost everyone is familiar with. But pulling back the curtain behind this complex technology and truly understanding how it works can be a little more complicated. Looks like something out of a science fiction movie artificial intelligence; In fact, it integrates into our lives by swimming in the seas of algorithms and crossing the oceans of data.

So how exactly do these smart systems work? Especially ChatGPTArtificial intelligence tools such as , Midjourney and DALL-E amaze us with the answers they give us, naturally it comes to our mind “How do artificial intelligence applications work?” question comes up frequently. We will unravel the mysterious world of artificial intelligence for you and discuss in detail how sophisticated technology can answer our every question and create visuals.

How exactly do AI applications like ChatGPT work?

ChatGPTIt works by trying to understand your question and then coming up with strings of words that it predicts will best answer your question based on the data it was trained on. Although it sounds relatively simple, let’s face it, what’s going on is a bit complicated, so it’s best to go step by step.

• Supervised and unsupervised learning
• Transformer architecture (T in GPT)
• Tokens
• Reinforcement learning from human feedback (RLHF)
• Natural language processing (NLP)

Supervised and unsupervised learning

P in GPT “pre-trained” and it’s a very important part of why GPT can do what it does. The best-performing AI models before GPT used supervised learning to improve their core algorithms.

GPTwhere a few basic rules are given and then fed large amounts of unlabeled data (almost the entire open internet). pre-training uses. He is then asked to navigate through all this data and develop his own understanding of the rules and relationships that govern the text. unsupervised is left.

Naturally, it’s not possible to really know what you’re going to get when you use unsupervised learning. ChatGPT In order to make its behavior more predictable and appropriate, fine adjustment doing.

Transformer architecture (T in GPT)

The purpose of all this training is; teaching patterns and relationships in text data, predicting what text will be next, and generating human-like response. Of course, this process is incredibly complex and multi-layered. So in summary deep learning neural network We can also say that it is aimed to create .

Although it sounds complicated when you explain it, the transformer model is How artificial intelligence algorithms are designed simplified it quite a bit. It enabled calculations to be parallelized or performed simultaneously.

In this way, training times were significantly shortened. Only artificial intelligence He not only made his models better, but also made them faster and cheaper to produce.

At the core of transformers is a process called “self-attention.” Old recurrent neural networks (RNNs) read text from left to right. This method can be good when related words and concepts are next to each other, but things can get a little complicated when the words are opposite ends. In our opinion, the biggest example of this is the occasional deviation in the Turkish language.

transformers reads each word in the sentence at once and compares each word with the others. This allows them to direct their attention to the most relevant words, no matter where they are in the sentence.

Of course, everything we have said simplifies things greatly. transformers does not work with words, which are pieces of text encoded as a vector (a number containing position and direction) “with tokens” they work. Attention is also encoded as a vector, allowing transformer-based neural networks to remember important information at the beginning of a paragraph.

Tokens

GPT-3 trained on approximately 500 billion tokens, allowing language models to more easily assign meaning and by matching in vector space allowing him to predict plausible text. Many words matched a single token, but were longer or longer complex words it was often split into more than one token.

OpenAIWhile , remains silent about the inner workings of GPT-4, we can assume that it was trained on almost the same dataset.

All tokens are written by humans from a huge corpus of data is coming. Among these; books, articles and other documents on all different topics, styles and genres, as well as the incredible amount of content available on the open internet.

As a result of all this training GPT-3’s neural network It had 175 billion parameters or variables. Thanks to his training, he was able to take input and give the most appropriate output based on the values ​​and weights he gave to different parameters.

OpenAIdidn’t say how many parameters GPT-4 has, but it’s probably more than 175 billion. Part of GPT-4’s increased power probably comes from having more parameters than GPT-3 and from improvement in education stems from.

Reinforcement learning from human feedback (RLHF)

GPT’s first neural network It was not available for public use, meaning it was trained on the open internet with almost no guidance. Because ChatGPTTo further develop ‘s ability to respond safely, logically and consistently to a variety of different prompts reinforcement learning from human feedback Optimized for dialogue using a technique called

Substantially OpenAIcreated some demonstration and comparison data that shows the neural network how it should react in typical situations. Like this artificial intelligence was able to learn which was the best response in any given situation. RLHF Although it is not pure supervised learning GPT It allowed effective fine-tuning of networks such as

Natural language processing (NLP)

All these efforts, of course, destroy GPT. natural language processing aims to make it as effective as possible. NLP; including speech recognition, machine translation and chatbots artificial intelligence We can also say that it is a large category that covers many aspects. So, the NLP category teaches artificial intelligence to understand language rules and syntax.

But remember, he still hasn’t learned it completely. Especially many people ChatGPT and such artificial intelligence In this period where the models write thesis/homework/content, you can guess that the robot script is revealing itself. We created content so that you can understand whether the articles you read were written with artificial intelligence. You can take a detailed look below.

We have explained at length how it works, now let’s compare it to understand how ChatGPT has developed itself with data sets.

At first ChatGPT 3.5to “How can I protect myself from getting sick on cold winter days?” We asked a themed question. Their answers were adequate, but GPT-4 Naturally, he did not give such advanced answers and also explained some information incompletely. Here are the answers to help you understand the difference:

For example, in these answers just like ChatGPT-4 We can see that it is written like, but when it comes to detailing, naturally advanced artificial intelligence lags behind the model. The subtitle opens, but the follow-up sentences constantly repeat themselves. We are asked to have a balanced diet, but there is no information about what kind of food we should eat.

GPT-4 It catches our eye with its detailed explanations at first glance. While taking the step of balanced nutrition, it gives us guiding answers about what to eat and what to do, of course. to users’ feedback and depends on the advanced data sets.

If you don’t like an answer or find it inappropriate, leave feedback. “feedback” It is very important that you throw it away. You ask why? Because as we mentioned before, much of the GPT foundation depends on it. Human-based learning model, allowing it to provide you with more advanced answers. So you actually train GPT yourself.

You can find our other artificial intelligence content that may interest you below: