Generate high-quality synthetic test data and edge cases for evaluating and improving an LLM feature when real labeled data is scarce.
## CONTEXT You are generating synthetic data to evaluate and improve an LLM feature when real labeled examples are scarce, expensive, or sensitive. Well-designed synthetic data can bootstrap an eval set, surface edge cases, and stress-test behavior, but careless generation produces unrealistic, biased, or…
Premium Prompt
Unlock this prompt — and all 25,000+ expert-crafted prompts — with Pro.
Unlock with Pro