How does an LLM "know" the sky is blue?
For most people, LLMs are a magic box that spews out language. Somehow, that language is very similar to " what a human would say ". "LLMs work on neural networks", "LLMs predict the next words in a sentence", "They're trained on a large set of data", are some of the common things people say about these mysterious models. But HOW ?? How can a machine, which was mostly working on logic, 1s and 0s,…
We haven't written up this one. Dev.to has the full story — the link below goes straight to it.
