A simple LH20 matrix or say G25 or say strong ion-exchange column would help getting rid of the electrolytes being bound to matrix and thus getting your compounds (unless they are highly charged as well; with some loss definitely!) separated from the electrolytes. If your compounds are larger than 1kDa (1000 MW) then you cana lso use 1 kDa or other cut-off dialysis membranes to get rid of very small electrolytes which would be washed away. best wishes.
I think that in your case the dialysis membrane should be used selecting the size you need. In the past I used also Dalton ultrafilters that were available from 5 KDa to 100 KDa from Sartorious (planar filters with pores conically cutted)
Assuming the electrolytes are less retained than the analytes on a reversed phase column, you could divert the outlet to waste and use a make-up flow to the ion source.
A better and relatively glitch-free solution is 2-D LC, perhaps in an automated sample prep implementation. The two columns could have identical stationary and mobile phases here. There are plenty of good reasons for doing it this way, one of them being that you can purge the first column, perhaps with back-flush, while the analytes are on the second one.
Waters has a bunch of solid phase extraction materials that may be useful. You may choose one of those depending on the analyte in question to clean up your sample.
The first step should involve suitable organic phase extractant to separate the the target compound which is most likely organic in nature.Second step would involve separation of electrolyte impurities (in very small amount) which might have been extracted along with the target compound while extraction operation.This can be done by passing it through ultrafilteration membranes.
Thank you for your the ideas. I must add that the target analytes are unknown, small and very polar compounds. I will try ultrafiltration and the 2D-LC.