Large Language Model(LLM)

A large language model is the engine inside most of the AI tools you will use. The name gives away how it works. It is “large” because it has been trained on a huge amount of writing, and it is a “language model” because its core skill is modelling language: working out which words tend to follow which.

Here is a useful way to picture it. Imagine someone who has read a substantial slice of everything ever written, and whose one job is to guess the next word in any sentence. Give them “The capital of France is” and they will confidently say “Paris.” Stretch that same skill across whole paragraphs and you get something that can draft an email, explain a concept, or suggest ideas. This next-word prediction is doing more heavy lifting than it first appears.

Because it works by predicting likely text rather than looking facts up, an LLM can sometimes state something wrong with total confidence. We call that a hallucination, and it is why checking important answers matters.

The practical upshot for you: an LLM is brilliant at anything language-shaped, which turns out to be a great deal of everyday work. The clearer you are about what you want, the better it can predict the response you are actually after. That is what prompting is all about.