Training
Overfitting
When a model memorises its training examples instead of learning general patterns, so it performs worse on new data.
Overfitting is why models are evaluated on data they have not seen. A model that has seen benchmark questions during training can look stronger than it is.
Frequently asked questions
What does Overfitting mean?
When a model memorises its training examples instead of learning general patterns, so it performs worse on new data. Overfitting is why models are evaluated on data they have not seen. A model that has seen benchmark questions during training can look stronger than it is.