How the neural network values are interpreted?

When we get the values after applying neural networks how are they interpreted?

Some workaround is still needed on this feature,for time being one can ignore this.