Imagine there is reverb on a note - RipX will need more than one set of waveforms to describe each harmonic, since there are effectively multiple versions of it - and layers allows those extra sets to be stored.

There are also other reasons you might get more layers, for example if the sound is difficult to describe using pure sine waves.

Merging the layers often makes little audible difference.