Overfitting
Definition
When a model memorizes its training data instead of learning the general pattern, so it scores well in training but poorly on new data.
In plain English
Imagine a student who memorizes last year’s exam answers word for word. They ace that exact exam but fail a new one with different questions. An overfit model does the same thing: it learns the noise and quirks of its training examples instead of the real rule behind them.
The usual warning sign is a big gap between results on the training data and results on a validation set the model has never seen.
Example
A spam filter trained on 500 emails learns that “Dear friend” means spam, because that phrase appeared in many spam examples. On real inboxes, it starts blocking normal emails too.
Draft, not yet reviewed by a person · Updated Nov 2026