Output vector y hat

Through each cell of our RNN we tend to yield a y hat but I haven’t seen where it could be utilized. Can you give an example of using y hat?

\hat{y} represents the output value(s) of the network. In production, it is the predicted values. During training, it is used along with the actual (or known or ground truth) values y to compute and minimize the loss or error between actual and predicted y - \hat{y}