fixaiprompt
All templates
⚡ Advanced Techniques·Works on: claude, chatgpt, gemini

Few-Shot Example Builder

Technique: Few-Shot Prompting

Generate the few-shot examples that make any prompt 2× more reliable.

Advanced#few-shot#examples#in-context-learning
0/3
Fill the template
3 placeholders left.
Live preview
Act as a prompt-engineering expert specializing in few-shot example design.

Task the examples will support:


How many examples: 

Edge cases the examples MUST cover:


Protocol:
1. **Restate the task** in 1 sentence so I can verify you understood it.
2. **List the dimensions of variation** the model needs to learn (length, tone, structure, edge-case handling).
3. **Generate the N examples** in INPUT → OUTPUT format. Each example covers a different dimension. Mark which example covers which edge case.
4. **Anti-examples**: 2 examples that look like they should match but should be classified differently. Explain why.
5. **Test**: invent a tricky new input you haven't shown — predict the model's output under your few-shot prompt. If you can't, the example set is too narrow.

Return the examples in a copy-paste-ready format.
Inputs0 of 3
›See the lazy version this template replaces
Before — the lazy prompt
Help me write few-shot examples for <task>.

Why it works

  • Forces the example set to span dimensions of variation, not just N near-duplicates.
  • Anti-examples teach the model what NOT to match — preventing over-eager false positives.
  • The self-test step exposes coverage gaps before you deploy.
  • Restating the task catches misinterpretation cheaply.

Make this one yours

Replace the bracketed placeholders, then paste into the Prompt Fixer to lint your customisation before hitting send.

Open Prompt FixerBrowse more templates