How can cross$-$validation help in managing overfitting?

Question $3$ options:

It divides the dataset into multiple subsets to ensure the model is evaluated on different data portions

It provides a more reliable estimate of model performance on unseen data.

It forces the model to generalize better across different subsets of the data.

It eliminates the need for separate training and validation datasets.

It ensures the model is trained on less data to make it more robust.