Author(s): Boris Meinardus
Originally published on Towards AI.
Everyone knows ChatGPT, but do you know how it works?
You perhaps know it predicts one word at a time. But do you know how it chooses the right word? Because it’s probably not how you think! To change that, in this blog, I’ll walk you through some basics of ChatGPT all the way up to how the new custom GPTs work so that you will understand ChatGPT, and pretty much all other LLMs better than 99% of people.
Oh, and by the way, stick around, because you will get the chance to win a sick NVIDIA RTX 4080 Super GPU!!
Let’s get started!
Okay, what is the goal of ChatGPT?
Well, the goal is to generate text that makes sense. How do we achieve thatWell, just like us humans, we write one word at a time.
We write
‘The’ ‘cat’ ‘chased’ ‘the’ …
And, you have probably already filled in the blank with the word ‘mouse’.
But how does the model, ChatGPT, understand these words?
ChatGPT needs to understand what has been written so far to have the context of the current conversation but also have an understanding of grammar and the general world to know what to write next.
Let’s first focus on the main mechanism of how the model can understand the words we or it have… Read the full blog for free on Medium.
Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.
Published via Towards AI