Glossary
The vocabulary of prompting, defined in plain English — no hand-waving, no hype. Every term links to the guide or tutorial that covers it in depth.
Agent
A model that works toward a goal in a loop — calling tools, reading results, and deciding the next step until the task is done.
Read the definition →Chain-of-thought
Prompting a model to write out intermediate reasoning before its answer, which can improve accuracy on multi-step problems.
Read the definition →Context window
The maximum amount of text, measured in tokens, a model can consider in a single request.
Read the definition →Embedding
A numeric representation of text that places similar meanings close together, used to search by meaning rather than keywords.
Read the definition →Few-shot prompting
Including one or more worked examples in the prompt so the model copies the pattern instead of guessing it.
Read the definition →Fine-tuning
Further training a base model on your own examples to specialise its behaviour — a heavier alternative to prompting.
Read the definition →Grounding
Constraining a model to answer from provided source material rather than its own training, to keep answers factual and checkable.
Read the definition →Hallucination
When a model produces confident, fluent output that is factually wrong or fabricated.
Read the definition →Prompt
The text you send an AI model to get it to do something — a question, an instruction, or a task with its data.
Read the definition →Prompt engineering
The practice of writing prompts so a model reliably returns what you actually want — mostly clarity, not tricks.
Read the definition →Prompt injection
An attack where text in the data a model processes is crafted to be read as instructions, hijacking its behaviour.
Read the definition →Retrieval-augmented generation (RAG)
Retrieving relevant text at query time and putting it in the prompt, so the model answers from your data instead of memory.
Read the definition →Structured output
Getting a model to return data in a strict, parseable shape — usually JSON matching a schema — instead of prose.
Read the definition →System prompt
The standing instruction sent with every request that sets the model's role, rules, and output conventions.
Read the definition →Temperature
A setting that controls randomness in a model's output — low for consistent, deterministic answers, higher for varied, creative ones.
Read the definition →Token
The unit a model reads and generates text in — roughly a word-piece; usage and cost are measured in tokens.
Read the definition →Tool calling
Giving a model functions it can ask to run — look something up, do an action — so it can go beyond what is in the prompt.
Read the definition →Zero-shot prompting
Asking a model to do a task with no worked examples in the prompt — just the instruction.
Read the definition →