Indeed from the physorg article illustration (cited up top) and their description, the N doesn't seem to form a molecular bond with F, but to diffuse into the iron crystal; with F binding to F and N being trapped but not bound. Fe16N2 is then a description of the crystal ratios, not a description of a molecule that can create a crystal.

My chemistry is prob as weak as yours though, so don't trust this too much.