In machine learning, model sensitivity (or model variance) measures how much a model’s predictions change when trained on different subsets of data.
High sensitivity means small changes in training data lead to large changes in the learned model - the model is “sensitive” to the specific training examples. This is a key component of overfitting: the model fits too closely to the particular training set rather than learning generalizable patterns.
Models with many parameters or high flexibility tend to have high sensitivity/variance without proper regularization.
Not to be confused with sensitivity (statistics), which measures true positive rate in classification tests.