Hi Yoonhoo,
As Wei Li suggest, the last parameter in the state definition is the energy correction. Lets say, you have a HOMO-LUMO gap from PDF to be 1.0 eV and it is underestimated with respect to say 2.0 eV what you know is right (e.g. experiment). Then you could define you ground and excited states as:
gs = ["GS", [2,-2, 3,-3], 0.0]
and
ex1 = ["S1", [2,-2,3,-4], 1.0]
Here, index 3 corresponds to the HOMO and 4 to the LUMO, so the gs -> ex1 is essentially HOMO->LUMO. In the definition of the ex1 state, you shift its energy by 1.0 eV, because you know that the DFT gap is underestimated.
Hope this helps.
Alexey