I am not familiar with ORCA, but in general you cannot locate a particular electron on a particular atom with DFT. You tell the program how many up and how many down spins and it determines the lowest energy configuration. In some cases, you can bias a DFT method toward a local solution with an appropriate starting guess.
The only way I know to confine an electron to a region of a molecule is using Constrained DFT. This has not been implemented in all DFT codes; don't know whether it's in ORCA.
i don't know much about orca in this respect, but if you simply want to know where is a particular electron, there are schemes such as NBO that would transform the canonical (delocalized) orbitals into something that chemists are more familiar with.
ORCA uses the BS(M,N) approach, with M and N being the number of unpaired electrons in fragments A and B. The program first calculates a higher multiplet, then a localization scheme is applyed and finally a lower spin state is enforced in a calculation with the previous as a starting point. You can find more information in the manual and in a few papers by Frank Neese.
Its same as Daniel mentioned above where you can provide BS(N,M) approach, where M and N are the number of unpaired electron on the two fragments A and B. It uses the spin flip approach to estimate the Broken symmetry solutions. For more details you just look in to the mannual.