By now, you have probably heard of OpenAI’s ChatGPT, or any of the alternatives GPT-3, GPT-4, Microsoft’s Bing Chat, Facebook’s LLaMa or even Google’s Bard. They are artificial intelligence programs that can participate in a conversation. Impressively smart, they can easily be mistaken for humans, and are skilled in a variety of tasks, from writing a dissertation to the creation of a website.
How can a computer hold such a conversation?
You know when your typing on your phone and you have that bar above the keyboard showing you what word it thinks you are writing? If you click the word before you finish typing it, it can even show you the word it thinks you are going to write next. Gpt works the same way, it just has waaaay more data that it can sample from.
It’s all just very advanced predictive text algorithms.
Ask it a question about basketball. It looks through all documents it can find about basketball and sees often they reference, hoops, Michael Jordan, sneakers, NBA ect. And just outputs things that are highly referenced in a structure that makes grammatical sense.
For instance, if you had the word ‘basketball’ it knows it’s very unlikely for the word before it to be ‘radish’ and it’s more likely to be a word like ‘the’ or ‘play’ so it just strings it together logically.
That’s the basics anyway.
I get that this is a simplified explanation but want to add that this part can be misleading. The model doesn’t contain the original documents and doesn’t have internet access to look up the documents (though that can be added as an extra feature, but even then it’s used more as a source to show humans than something for the model to learn from on the fly). The actual word associations are all learned during training, and during inference it just uses the stored weights. One implication of this is that the model doesn’t know about anything that happened after its training data was collected.
I wonder what an ELI5 version of ‘stored weights’ would be in this context.
Not quite ELI5 but I’ll try “basic understanding of calculus” level.
In very broad terms, the model learns complex relationships between words (or tokens to be specific, explained below) as probabilistic scores. At its simplest, this could mean the likelihood of one word appearing next to another in the massive amounts of text the model was trained with: the words “apple” and “pie” are often found together, so they might have a high-ish score of 0.7, while the words “apple” and “chair” might have a lower score of just 0.2. Recent GPT models consist of several billions of these scores, known as the weights. Once their values have been estabilished by feeding lots of text through the model’s training process, they are all that’s needed to generate more text.
Without getting into the math too much, this is how a GPT model then uses these numbers to come up with words:
In reality we’re not quite so sure what the weights represent to the model exactly, but this is the gist of it. All we know is that they signify the importances or non-importances that the model places on some pattern that was present in the training data. Some of these patterns could be just simple two-word pairs, but many are probably much more complicated. Lots of researchers are currently trying to get a better idea of how these numbers are actually affecting the model’s output.
How closely related words and their attributes are to other words.
Edit: i see now it’s an article and not just you asking a question lol. I’ll leave it up anyway.