Create an evaluation rubric for this task: [task] Return: - 5 criteria (0-2 each) - 3 test inputs - what a perfect output looks like
Opens on home with the prompt prefilled.
Open in PrompteaHow to apply it in prompts
Use this concept to reduce ambiguity and improve consistency. If you need strict formatting, combine it with a JSON guide and a per-model template.
What it means
- If you don't evaluate, you can't improve reliably.
- Use a small set of edge cases and score outputs against a rubric.
Copy-paste example