We all enjoy using ChatGPT. It's like a magical helper that can answer any question we ask. It can guide us, make us laugh, or help us understand complex things. But as we tap away at our keyboards, pouring out inquiries into the ether and receiving instant, insightful responses, have you ever paused to wonder about the inner workings of this technology?
The secret behind ChatGPT, and many similar technologies, lies in the innovative application of a larger class of models called Large Language Models, or LLMs for short. What are these LLMs? How do they work? And how do they make ChatGPT so smart and helpful?
I couldn’t find an easily understandable infographic while researching about it. So I took the matter into my own hands and crafted an infographic that distills the complexities of LLMs into digestible nuggets of information. With the aid of this visual, I'll try to break down the technical jargon and illuminate the concept of Large Language Models in a manner that is easy to understand and engaging. Let's embark on this adventure together ✨:
Data input
You ask ChatGPT a question or tell it something.
Text embedding
ChatGPT converts these tokens into vectors, a process known as "embedding". Each token corresponds to a specific point in a high-dimensional space.
Neural Network processing
ChatGPT uses its 'brain' (a thing called a neural network) to think about what you said. It's like it's solving a big puzzle!
Context Understanding
Then it figures out how all the words you used to link together, using a trick called the 'transformer' model.
Response Generation
It thinks up the best tokens to answer your question.
Spitting out the answer
Finally, it changes the number code back into words, and that's the answer you get!
I hope this helps you understand LLMs better in some way :) Hot girls know how LLMs werkkk
Subscribe to my Newsletter: