Characterizing Inter-Layer Functional Mappings of Deep Learning Models