LLM

A large language model is trained to predict text, and from that simple objective it learns to answer questions, summarize, translate, and write in a natural voice. It is the component doing the writing in an AI assistant.

On its own it answers from patterns baked in during training, which is why it can sound sure of facts it does not have. Pairing it with retrieval and evaluation is how you turn a fluent generalist into something dependable on your specific material.