A jargon-free explanation of how AI large language models work
Want to really understand large language models? Here’s a gentle primer.
That was really cool. It's absolutely fascinating, and scary, that were able to develop sophisticated software and we have so little idea of how it actually works. And it seems like we can't catch up! Researchers are still working on understanding the layers involved in GPT3, and we're already on to GPT4.
I don’t think there’s anything scary about LLMs. They’re just fancy computer algorithms that print out the most likely response to any given query based on the data it was trained on. The actual scary part are the idiots who somehow confuse it with real intelligence, who then choose to rely on it for important stuff like policy making.
LLMs would have told you the sun revolved around the earth back when that was standard thinking.
![[Other] A jargon-free explanation of how AI large language models work](https://lemmy.world/pictrs/image/9d54eb36-d1b3-4f99-b378-fbf49c97c553.jpeg?format=webp&thumbnail=256)