I suspect that your post #1 results may be caused by leakage, try adding a high value resistor (100k) from output to ground and see if that makes a difference.
For your last circuit I suspect that is a function of gate-drain capacity which is charged to 5v when the gate is at 0v and then hits the parasitic source-drain diode when the gate is driven to +5v (It would pulse the drain to +10v if the diode were not present). Again you have no DC load so very small currents are needed.
Slow typing I see others have suggested leakage as well.
Jim