The mutual solubility/insolubility of SAN/SEBS blends much depends on the composition of both polymers. The more styrenic (rich in styrene monomer) the polymers are, the more difficult it is to separate them by using sufficient solvent systems.
Acetone may be useful in removing SAN from the SAN/SEBS blend since it readily dissolves styrene-rich SAN random copolymers (e.g. SAN 24) and only swells SEBS triblock copolymers (e.g. Kraton G) without dissolving it.
Certainly, you can predict that behavior based on the solubility parameter values of both polymers, being 21.6 and 17.0 (J/cm3)^0.5 for SAN and SEBS respectively. The SAN copolymer is also much more polar than SEBS with larger contributions to the solubility parameter from polar components (10.0 and 3.5 (J/cm3)^0.5, for SAN and SEBS respectively).
For a better result, I would try extracting the blend with the solvent at a low temperature (-20 to 0 deg C). Ethyl acetate might be sufficient to do so as well.
You mentioned preferential dissolution as a tool for revealing morphology. This could be successful but it has its challenges as in removing the dissolved material so it does not redeposit also the possibility of structural collapse . Selective Etching in which the polymer is degraded in to more soluble sections using an oxidising agent in which the soluble sections can easily be transported from the surface without redeposition give a much higher resolution image. Equally staining, for example SEBS will stain with Ruthenium Tetroxide (make a thorough review of the H & S procedures before starting) and looking at a sample in the TEM or at a push in a SEM using back scattered electrons is probably your best approach.
Any new morphology revealing procedure needs to be extensively validated otherwise you think you have succeeded when you have images that match your expectations.