Large Language Model (LLM)
Working knowledgeCore AI Concepts
Also called: LLM
A large language model (LLM) is a foundation model with billions of parameters, trained on vast textual datasets using transformer neural network architectures. It comprehends, translates, summarises, analyses and generates human language with contextual nuance. That is what powers general-purpose conversational agents, enterprise search platforms and automated content processing pipelines.
In practice
The LLM has become the default interface layer for enterprise software, so the buying decision has moved from whether to have one to which one and where it runs. Many vendors resell someone else's model, so ask whose model sits underneath a product and what happens to it if that supplier changes its price or its terms.