In revision r2, the maximum conductance density and "offset" values used for the NaF channel in the reference ChannelML- and mod-based versions of the cell model seem to differ from the default value in the ChannelML definition. Specifically, the values are:
* In GranuleCell_VSCS and GranuleCell_mod_all: gmax=4.0E-7 mS um^-2, shift=-15.0
* In the source:neuroConstruct/cellMechanisms/NaF/NaF_Chan.xml, default_gmax=400 mS cm^-2 (=4.0E-6 mS um^-2) and shift=-10.

If the default values are not in use anywhere else, I think they should be changed to coincide with the ones in the reference model.