LLM (Large Language Model)
A large language model is a neural network trained on massive text corpora—often trillions of tokens—to predict the next word in a sequence. That simple objective, scaled up, produces systems that can write code, summarize legal briefs, translate languages, and hold surprisingly coherent conversations. GPT-4, Claude, Gemini, and Llama are all LLMs. The "large" refers to parameter count, typically in the billions, which correlates loosely with capability but tightly with compute cost. LLMs are general-purpose by default and useless for specific tasks until you shape their behavior through prompting, fine-tuning, or connecting them to your data. The model is the engine. Everything else—retrieval, guardrails, UI, evaluation—is what makes it drive straight.
Related terms:
System Prompt
A system prompt is an invisible set of instructions given to a language model—defining its persona, constraints, output format, and behavioral rules—and occupies the “system” role in APIs like OpenAI and Anthropic. It shapes every response by encoding business logic and is the most efficient way to control model behavior.
Structured Output
Structured output occurs when a language model returns data in predictable, machine-readable formats—such as JSON, XML, or typed objects—rather than free-form prose, enabling software systems to reliably parse fields like names, dates, and dollar amounts. By using constrained generation to enforce a JSON schema, structured output transforms AI from a conversational interface into a dependable system component.
Foundation Model
A foundation model is a large AI model trained on broad data at massive scale, designed to be adapted to a wide range of downstream tasks rather than built for any single one. Coined in 2021 by Stanford’s Center for Research on Foundation Models, this approach boosts efficiency but concentrates power among providers like OpenAI, Google, Meta, and Anthropic.