Reference · Glossary
Overfitting
When a model **memorizes training examples** but fails on new data — great accuracy in the lab, poor results in the wild.
When to use
As a warning label when eval scores look too perfect on a tiny set.
When not to
As an excuse without measuring — always compare train vs holdout metrics.
Example
100% accuracy on 10 training rows and 40% on a fresh validation slice — classic overfit on a toy dataset.