A prompt is your instruction to a Large Language Model (LLM). Effective prompting is the art and science of crafting these instructions to get precise, relevant, and creative results. It’s an iterative process that blends clarity, context, and clever design.
Prompt engineering is the key that unlocks the full potential of AI. It’s how we turn a general-purpose model into a specialized expert for any task.
Construct a detailed prompt using the fields below. The result will be copied to your clipboard.
Get feedback on your prompt to improve its effectiveness. (This is a simulation).
The simplest form: just ask the model to perform a task with no prior examples.
Classify the following movie review as POSITIVE, NEUTRAL, or NEGATIVE.
Review: "The movie was an interesting attempt, but the plot felt underdeveloped."
Sentiment:Provide a few examples (shots) to teach the model the desired pattern and format.
Translate English to French.
sea otter => loutre de mer
cheese => fromage
- - -
car =>Encourage the model to "think step-by-step" to solve complex reasoning problems.
Q: Roger has 5 tennis balls. He buys 2 more cans of tennis balls. Each can has 3 tennis balls. How many tennis balls does he have now?
A: Roger started with 5 balls. 2 cans of 3 tennis balls is 6 balls. 5 + 6 = 11. The answer is 11.Copy-and-paste templates for common tasks. Just fill in the blanks!
Summarize the following text in [number] key bullet points. Identify the main argument, the evidence used, and the conclusion.
[Paste text here]Explain the concept of [complex topic, e.g., "black holes"] to me as if I were 5 years old. Use a simple analogy.See how prompting can be applied in different professional fields.
Act as a high school history teacher. Create a lesson plan for a 1-hour class on the main causes of World War I. The plan should include learning objectives, key terms, a 15-minute lecture outline, a 20-minute group activity, and a simple assessment question.Techniques in this guide apply broadly, but different models have unique strengths. Explore platforms like Hugging Face (for thousands of open models) and specialized models like Llama 3.
Hugging Face is a hub for finding and testing models. Llama, by Meta, is a powerful open-source model you can run locally for greater control.
Act as Llama 3, the latest open-source model from Meta. Briefly introduce yourself and highlight two of your key improvements over previous versions.Be mindful of common pitfalls:
This guide builds on the work of many researchers. For a deeper dive, explore these sources.