frozen specific layer in mnist model

Question

I have a dumb question about freezing specific layers in the MNIST model: the model I have is:

mnist_model = keras.Sequential(
    [
     keras.Input(shape=(2