LLM

A large language model is a neural network trained on massive text corpora to understand and generate human language. Built on the Transformer architecture with attention, modern LLMs (GPT, Claude, Llama) handle everything from conversation to programming to analysis. Key dimensions are model size (parameter count), context window (how many tokens it can hold), and emergent abilities that only appear at scale. LLMs are the foundation of modern AI agents, giving them reasoning and natural-language understanding.