Overfitting
Also known as: over-fit
Overfitting is when a model learns the specific training examples so thoroughly that it can’t handle anything new. It’s the academic version of a student who memorizes the answer key instead of learning the subject.
- Good model: generalizes — handles unseen data well.
- Overfit model: nails the training set, fails the real world.
It’s one of the most common failure modes, and a key reason we hold out test data.
The goal is the middle: learn the underlying pattern (not the memorized examples) so you can apply it to new inputs. That's [generalization](/glossary/generalization/).