In MLS course 1 week3, we talked about what overfitting means and how to reduce it.
How do we know whether a model has overfitted the training set? To my knowledge, when the error on training is very low and high on test set, that indicates overfitting.
Typically this requires measuring the performance on some data that wasn’t used in training (a test set or a validation set).
You may also want to watch the videos in MLS course 2 week 3 that talks about overfitting and underfitting.