AI Dictionary › Prompting

Demonstration Selection

Selezione degli Esempi

Demonstration selection is the strategic choice of which examples to place in a few-shot prompt, since example quality and relevance affect output far more than their quantity. Examples similar to the user's input, diverse among themselves, and correctly formatted guide the model better. Example: to classify a refund support ticket, you dynamically select 3 past refund-ticket examples (by semantic similarity) rather than 3 random ones; seeing cases close to the current problem, the model classifies more accurately.

Definition

Use it when doing few-shot on variable inputs: retrieve the examples most similar to the current input (often via embeddings and similarity search) instead of a fixed set. Watch order and class balance: all-same-category examples can bias the answer.

Related terms

• Few-shot
• In-Context Learning
• One-shot
• Prompt Template
• Embedding

More in Prompting

• Prompt
• Prompting
• Role Assignment
• Zero-shot
• Few-shot
• Chain of Thought
• Prompt Engineering
• Output Format

Put it into practice

• Real scenario: apply this concept with an instant AI verdict.
• Daily gym: short drills to train dictionary concepts.
• Playoff game: challenge other players, elimination-style.

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.