Golden Sets

Coming Soon

Detailed documentation for golden sets is coming soon. In the meantime, see Testing Prompts for an overview of golden set capabilities.

Golden sets are curated collections of input/output pairs used for systematic prompt testing. They allow you to:

• Define expected inputs and outputs for a prompt
• Run regression tests when prompt content changes
• Track test results over time
• Source test items from manual entry, query history, or previous test runs