Which term describes the layer between the input layer and the output layer where learning occurs?

• A. Output layer
• B. Input layer
• C. Hidden layer
• D. Feature layer

Answer: (C)

The main idea is that learning happens through the internal processing done in the hidden layer. This layer sits between the input and the output, and it learns to transform the raw input into more meaningful representations by applying weighted sums and nonlinear activation functions. These internal representations—features—are what the network uses to make accurate predictions. During training, the learning signal (the error) flows back through these hidden layers, updating their weights (and the output layer’s weights as well) so the overall model improves. If there are multiple hidden layers, the network can learn increasingly abstract features, which is the hallmark of deeper architectures.

The other layers aren’t described as the place where learning happens in the same way. The input layer merely holds the data you feed in, and the output layer produces the final prediction. A “feature layer” isn’t a standard term for this exact position in the network, so the term that best captures the described location is the hidden layer.