This appears as normal behavior. The scope inputs are very sensitive, and you will easily pick up EMI. If you make a loop by clipping the ground clip to the input tip you'll make an antenna and stabilize the readings you see from ambient.
The Rigol does not automatically set probe scale, you must manually assign it or the readouts will be 1/10. The hardware and software counters do not necessarily match: the hardware counter counts input pulses while the software counter measures waveform peaks on the screen [ likewise other measurements are based and scaled to the screen, not actual measurements by a specific device like a voltmeter], ... pretty much what you see is unreliable until you ground the probe properly.
What appears as more noise in the what I think is the x10 setting is because the default probes, assuming you have RP2200 are rated to ~ 200 MHz in x10, and only 6 MHz in X1, so the filtering effect of the X1 is removed.
A lot of good info is in the manual [well written compared to many Chinese label DSOs] and it will pay to work through its examples on usage.
Enjoy.