I'd like to use the MEAM potential to calculate a subset of the interactions in the system, along with a tabulated potential for the remainder of the interactions. For example, in a two specie system (X and Y) I would like to use the MEAM potential for X-X and Y-Y but not for X-Y. I think it should be possible to do this using "pair_style hybrid/overlay" but this would require the MEAM potential to return zero for X-Y interactions. Is there a way I can set the parameters to get the MEAM to return zero for X-Y interactions? So far I haven't been able to work out if it is possible to do this.

I'd like to use the MEAM potential to calculate a subset of the interactions in the system, along with a tabulated potential for the remainder of the interactions. For example, in a two specie system (X and Y) I would like to use the MEAM potential for X-X and Y-Y but not for X-Y. I think it should be possible to do this using "pair_style hybrid/overlay" but this would require the MEAM potential to return zero for X-Y interactions. Is there a way I can set the parameters to get the MEAM to return zero for X-Y interactions? So far I haven't been able to work out if it is possible to do this.

first off, you would have to use pair_style hybrid not hybrid/overlay
and since MEAM is a manybody interaction, where you cannot separate
interactions by pairs, but only by atom types, you'd have to use MEAM
twice (once for type X and once for type Y).
however, since the pair style meam is a fortran library, it may not
support having two instances of the meam pair style.
have a closer look at the docs and make some test.

second, this whole enterprise would not make much sense, if the atoms
of type X are intermixed with atoms of type Y. combining multiple
instances of manybody interactions, via pairwise additive
interactions, is a reasonable approach primarily when combining well
separated regions, e.g. a piece of diamond scraping along a silicon
surface.

third, you can always modify the MEAM implementation itself (so that
you get M^2EAM ;-))

You might also be able to set some params in the MEAM

potential file (read in by pair_style meam) to set
the cross interactions to 0.0. However, I agree with
Axel that it seems you are going down a questionable
path.