I've no practical experience of how electricity meters work, but I seem to remember when I was taught about power factors being told that these meters typically measure current, and that the mean square adjustment and integration over time was then used to calculate the transferred charge and hence energy. This was to explain why industrial consumers were concerned about power factor - because they are charged for current but get value from energy so they want the power factor to be as high as possible so they don't pay for energy they didn't use.
Given that these devices are specifically designed to measure current accurately over a wide range of conditions, I doubt that it will be easy to cheat them. You could demonstrate reducing the power factor so that the meter shows power that you didn't get, but there's no way (that I know of) to get a power factor greater than 1.
What scope do you have to interfere with the power supply coming in to the meter? Is it entirely under your control, are you allowed to use inductive coupling etc, or is it completely inaccessible to you? If you only need to demonstrate power going through the meter without being measured and the whole circuit is under your control, then you could try either using DC (I have no idea whether meters would detect that) of a very high frequency AC well outside the range the meter is designed to pick up.