Usage
Prompt
The input text given to a language model that frames the task, provides context, and guides the model's output. Prompt engineering is the practice of crafting effective prompts.
Explained at five levels
Level 1
The question or instruction you type to ask the AI to do something for you.
Level 2
The text you write to tell an AI what you want — like giving it a task or asking a question. Better prompts get better answers.
Level 3
The input text given to a language model that frames the task, provides context, and guides the model's output. Prompt engineering is the practice of crafting effective prompts.
Level 4
A structured input to an LLM that establishes context, constraints, and desired output format. Effective prompting leverages techniques like few-shot examples, chain-of-thought reasoning, and role specification.
Level 5
The conditioning sequence prepended to a generative model's context window that steers the posterior distribution over completions — manipulable through in-context learning, instruction tuning alignment, and constitutional constraints.
Definitions are educational summaries. Terminology can vary by source and context.
Sources
- NIST Trustworthy and Responsible AI Resource Center: Glossary — National Institute of Standards and Technology. Accessed 2026-07-20.