Training
Overfitting
Overfitting occurs when a model learns training-specific detail that does not generalize.
How it works
Training error can keep falling after validation error begins to rise. Regularization, more representative data, or earlier stopping can reduce the gap.
A concrete example
Illustrative: training loss falls from 0.30 to 0.08 while validation loss rises from 0.34 to 0.51.
Do not confuse it with Underfitting
overfitting has a large train-validation gap; underfitting performs poorly on both.