← Glossary / 87 of 100
n. / Quality and evaluation
Golden dataset
Definition
A golden dataset is a small, carefully reviewed set of inputs with trusted expected answers, used as the reference for judging changes. Quality matters more than size.
In more detail
Build it from real traces, including the failures you have already seen, and have a person confirm each expected answer. Keep it stable so scores are comparable between versions, and add new cases when new failures appear. A golden dataset that never changes slowly stops representing real traffic, so review it regularly.
Example
You pick 80 real support questions, including 15 that once caused wrong answers, and write the correct answer for each after review by a support lead.
