Also known as: Prompt / Prompt Engineering · prompting
Ein Prompt ist die Eingabe an ein generatives KI-Modell, die dessen Antwort steuert. Er kann Anweisungen, Kontext, Beispiele, Beschränkungen oder multimodale Inhalte wie Text, Bilder, Audio, Video oder Dokumente enthalten. Wirksame Prompts definieren die Aufgabe klar und prägen Umfang, Format, Ton und Nützlichkeit der Ausgabe.
Any text entered as input to a large language model to condition the model to behave in a certain way. Prompts can be as short as a phrase or arbitrarily long (for example, the entire text of a novel). Prompts fall into multiple categories, including those shown in the following table: Example --- How fast can a pigeon fly? | Write a funny poem about arbitrage. Translate Markdown code to HTML. For example: Markdown: \\* list item HTML: \\ \\list item\\ \\ Explain why gradient descent is used in machine learning training to a PhD in Physics. The Prime Minister of the United Kingdom lives at A generative AI model can respond to a prompt with text, code, images, embeddings, videos...almost anything.
A prompt is a natural language request that's sent to a generative AI model to elicit a response. Depending on the model, a prompt can contain text, images, videos, audio, documents, and other modalities or even multiple modalities (multimodal). An effective prompt consists of content and structure. Content provides all relevant task information, such as instructions, examples, and context. Structure ensures efficient parsing through organization, including ordering, labeling, and delimiters.
The inputs, instructions and context provided to a generative model. Good prompt design improves reliability, safety and usefulness.