Lowdin population analysis

  According to the manual, constrained DFT uses Lowdin populations by default (Mulliken not recommended). I would like to set up a calculation by taking converged MOs as a guess, then running CDFT with a cluster of atomic charges constrained to the value in the original calculation (just as a check, to ensure the same or similar wavefunction is produced). To do this, I guess I'd need the Lowdin charges for the original WF. Is there a database key/value pair that can be set to produce Lowdin population analysis, and if so what is it?