Skip to content
trending
AI Glossary / Prompt engineering
⌨️ Prompts

Prompt engineering

The art of writing effective prompts: context, goal, example, iteration.

Prompt engineering is the skill of formulating instructions so the model gives the best possible responses. It comprises specific techniques: providing context and role, breaking complex tasks into steps, showing examples, enforcing response format, and iterative refinement. Despite the name, it's less "engineering" and more a craft of communication: you need to know what the model understands well and what needs to be spelled out explicitly. As models evolve, some tricks lose importance, but clearly describing the task remains the cheapest way to better results.

Related concepts
PromptChain-of-thoughtFew-shotPrompt chaining

← All terms in the AI Glossary

// site map
Quantslant in numbers 1,427posts in archive26news from the last 7 days18video models in the observatory
🎬 AI Video Benchmarks: Higgsfield DoP vs OpenAI Sora • Higgsfield vs Runway Gen-3 • AI Video Credit Calculator
× ‹ enlarge ›