Regarding the first excited triplet states of organic molecules, it is basically expected that their spin density distributions can be evaluated from the coefficients of HOMO and LUMO, for example, like HOMO^2+LUMO^2. But spin densities of organic triplet states calculated by Orca is different from HOMO^2+LUMO^2. Actually, I don't understand well how spin densities are calculated in molecular orbital calculations. Could someone tell me hot to do that ?