also: large language modelalso: language model
What is LLM?
Last updated · 1 min read
The name is literal: large (billions of tunable parameters), language (text in, text out), model (a statistical system learned from data, not hand-programmed rules).
What makes it useful is scale. Predicting the next word on a few paragraphs gives you autocomplete; doing it across trillions of words gives you a system that drafts contracts, explains calculus and writes code. For the full picture, read What is a Large Language Model?.
Related: RAG gives an LLM fresh facts, hallucination is its failure mode, embeddings are how machines represent meaning.