The oxides correction and the nitrides correction on the Materials Project is not the same. For the oxide correction, the oxygen chemical potential is fixed to the solid-state O2 chemical potential, and all the Metal Oxide phases (MOx) are corrected to as if the O2 were a gas molecule at standard temperature and pressure. This is an implementation of the advanced scheme developed in the GGA/GGA+U mixing scheme discussed here: https://journals.aps.org/prb/abstract/10.1103/PhysRevB.84.045115
On the MaterialsProject, there is no +U benchmarking for metal nitrides (although, there should be, someone ought to do it). Instead, the N2 molecule is corrected to the chemical potential of N2 at STP, using the older scheme described here: https://journals.aps.org/prb/abstract/10.1103/PhysRevB.73.195107.
The corresponding chemical potential for N2 can be found at the bottom of the lowest-energy N2 entry, under “Corrected Energy”, https://materialsproject.org/materials/mp-1080711/#corrections-eqn. If you are generating phase diagrams on the MaterialsProject, or using the MPRester API in pymatgen, it should do this correction for nitrides automatically.