When writing few-shot examples for classification, which practice improves robustness?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Few-shot works when examples look like real work: varied cases, clear labels, and the exact output shape you want.
Full explanation below image
Full Explanation
Diverse, correctly labeled few-shot examples teach both decision boundaries and output format. Single monolithic examples, only adversarial noise, or unlabeled inventiveness reduce reliability.