Correct ModInv classical simulation

[fpapa250]

In order to test the ECAdd() bloq's classical simulation I implemented a hacky solution to make it invertible without implementing the full logic of the circuit described in https://arxiv.org/abs/2306.08585. Information is stored in garbage registers which allows us to invert the mod inversion in the actual bloq; for the meantime I just store the original value x in the garbage registers so if we want to uncompute the ModInv bloq we can get the original value without doing all the fancy logic.