Too many thanks for your highly valuable and informative comments.
A)- My mean was using "no star" within $Del keyword of NBO.

Sincerely,
Saeed

(A) There are many different ways of using NBO deletion feature via $del, I don't know how you want to delete.

(B) Intrafragment E2 is completely different to E_pol. These E2 terms mostly correspond to electron delocalization from Lewis structure to actual electronic structure. Even if another fragment doesn't occur, there are still large intrafragment E2 terms.

Interfragment E2 mostly exhibits energy change corresponding to interfragment charge-transfer, however its value must be very different from the E_CT estimated by much more rigorous energy decomposition analysis. First, all interfragment E2 terms cannot be directly summed up to derive total energy variation due to interfragment  CT; second, E2 is only a very crude estimation by second-order perturbation theory on the top of idealized and thus artificial Lewis electronic structure.

Best regards,

Tian

Please suppose a non-covalent bonded adduct XY. It is well know that a SAPT analysis over the optimized structure of XY leads to different components of the non counterpoise-corrected interaction energy (non CP-IE). Among different components, induction energy (E_ind) is here going to particularly be paid attention.

As defined, E_ind includes charge transfer (CT) and polarization. CT indicates electron density delocalization between filled orbitals of X and empty orbitals of Y and vise versa. Meanwhile, polarization referees to as electron density delocalization between filled and empty orbitals of X due to presence of electric field of Y, and electron density delocalization between filled and empty orbitals of Y due to presence of electric field of X. Indeed:

E_ind=E_CT(X--->Y)+E_CT(X<---Y)+E_pol(of X by Y)+E_pol(of Y by X).

Thus,

Question (A)- using "$Del" keyword in NBO analysis, we can remove all empty orbitals of both X and Y fragments within the Fock matrix of the optimized adduct XY. Consequently, all CTs and polarizations would be prevented and the energy difference is expected to be comparable with E_ind of the SAPT analysis.

On the other hand,

Question (B)- within the E(2) perturbation analysis section of NBO output file, we can find all filled-empty delocalization through fragment 1 (X), through fragment 2 (Y), and between fragments 1 and 2 (1---> and 1<---2). I think the sum over energy of "delocalization through fragment 1 (X)" and he sum over energy of "delocalization through fragment 2 (Y)" gives us E_pol(of X by Y) and E_pol(of Y by X) separately. However, sum of all energies between fragments 1 and 2 (1---> and 1<---2) should give us E_CT(X--->Y)+E_CT(X<---Y). In this way, we can obtain two components of E_ind

Please, if possible, let me know your highly valuable opinion regarding the reasonableness of my point of view as well as interpretations in questions (A) and (B)

Sincerely,
Saeed