Pixols aren’t in 3d space. They’re squares, like pixels, except in addition to color and value info, they also hold depth, lighting, and material information. The ZBrush canvas, which can be used for 2d, 2.5d, and 3d all at the same time, is just a grid of pixols. You can zoom in and see them. ZBrush models are polygon based, so sculpting isn’t “displaced” like with voxels, you’re actually manipulating polygons. Because of the pixol, ZBrush is an absolute beast with handling dense meshes, so it doesn’t need to use map displacement to achieve detail.
The pixol comes into play in two ways, both related to the canvas. The first is 2d/2.5d sculpting and painting, which is strictly pixol based with no polygons. The second is a way to display and manipulate 3d meshes without taxing your computer as much as an actual 3d scene does. The 2.5d canvas allows for you to work with a crazy poly count on each tool. Plus each subtool can push the maximum poly count that your computer can handle. If you’re just rendering out of ZBrush, you can push that even further and sculpt in hd, which is just ridiculous, and also possible because of the pixol. It’s actually all pretty genius.
To answer your question, pixols are just what makes up the canvas, not the model. The models are strictly polygons with no on screen displacement such as voxel sculpting, which is based on an rgb map.