These parameter values are calculated from a fixed voltage step, right?
If, despite the applied voltage, the current has gone the opposite direction, the software may calculate negative resistances. It suggests that there is another source that contributes to the currents that you record, e.g. hum, opening of ion channels, changes in your seal with the membrane.
Make sure that you get a gigaohm seal before you break in. Compensate the pipette capacitance. Break in to your cell and wait a minute or two for the cell to stabilize.
You can do a continuous voltage-clamp recording to see what changes are occurring when you do not do the voltage step. Then, do your voltage step protocol, and check how the response looks.
Do you see instabilities also with the model cell? That would be a first step to realize if there's any issues with the rig itself. If the model cell's working properly, then try to tweak your prep. Btw, what is your amplifier?