As far as I understand, that effect is just a result of the superposition characteristic of waves in general, so yeah it should also apply to gravitational waves. I think the difference is that there aren't a lot of sources of "carriers" in the gravitational wave universe.
The heterodyne effect comes from multiplication of signals. But you really just need to pass their superposition through a nonlinearity, because the Taylor series expansion will typically have an early term with the multiplication of the two signals' frequencies with a high coefficient. Of course, you'll have lot of artifacts from all the other terms of the expansion, but you can use a bandpass filter to drop all the unwanted terms.
