Large Language Model (LLM)

A large language model (LLM) is a deep learning model trained on vast amounts of text data to predict and generate natural language. LLMs are built on transformer architectures and contain billions to trillions of parameters that encode statistical patterns about language, knowledge, and reasoning. Prominent examples include GPT-4, Claude, Gemini, and Llama. They can perform a wide range of tasks — summarization, translation, question answering, code generation, and more — without task-specific training.

In commerce and enterprise contexts, LLMs serve as the foundation for capabilities including AI-powered search, product content generation, customer service automation, and intelligent document processing. They can be accessed via API (hosted models) or deployed on-premises. Their primary limitations — hallucination, context window constraints, and sensitivity to prompt phrasing — require careful system design. LLMs are often augmented with retrieval systems (RAG), tool use, and fine-tuning to adapt them to specific business domains and workflows.