LLM (large language model)

LLM stands for large language model. A model of this kind is trained on very large collections of text and learns by repeatedly predicting how a passage is likely to continue. It then writes its answers piece by piece, each time estimating what should come next.

A model predicts what comes next rather than checking whether it is true, so it writes just as fluently when it is wrong as when it is right. Check important claims against a source.

In my articles

In my second brain series, I describe a large language model as a program that can take over the job of keeping notes in order. People abandon collections of notes because that upkeep is tedious, and a model can do it for them.

Sources

• Large language model, Wikipedia (opens in a new tab)