Hi, I've never tried it, but in principle I'd expect the result to be very much like making a b&w film mask after the fact, and sandwich it with the original neg. In essence it would seem to mainly increase contrast, almost like push processing. But this is just a guess, and in my experience, carrying out the actual experiment sometimes reveals an unexpected angle.
Another idea you might try is multiple developments of the dye image. In principle, if you were to do the bleach bypass, but do fix, it seems like you should then be able to rebleach the silver image, converting back to silver bromide. So presumably it can then be exposed and developed (in color developer) again, forming another dye image. And on and on. But I fear it will just be driving up the contrast.
Best of luck.
Ps, it occurs to me that, assuming successive color developments actually work, that it may be useful to use highly diluted developer so as to act as a compensating developer, holding the contrast in check. Again a complete guess, but it seems as though it might work.
Pps, given the extent of your tech knowledge I'm guessing that you are already familiar with the effect of "flashing" film to improve shadow detail. If not, the point is that we get what one might call a sub-latent image that is there, but not quite strong enough to be developed. But by giving a very weak additional exposure, an overall "fog," some of the sub-latent images can be upgraded to "latent," thus becoming capable of being developed. A useful exposure might be 3 to 4 stops under a metered exposure, so an out-of-focus double exposure of a blank wall, or whatever, could do it. If you overdo it you can get a bothersome "flat" look in the shadow areas, so the effect is limited. (The effect should be identical to flare light, which is sometimes demonstrated on sensitometric curves; one can see how the limitations play out.)