[fromageGB]

awm gives a convincing argument for the NF 4x6x. Playing Gazzilli it has to be this, but even if not, then it makes sense, as a GF can rebid 2♦ and I would have thought continuations would discover the best fit regardless of the lengths.

I suppose the argument (certainly at matchpoints) for NF 5x5x is that opener may have a doubleton in both suits, and if he can give a preference it scores better. One of the reasons for using Kaplan inversion of course.