etnb, you're just a little confused.
Think of the spinning rotor disk as a wing. Because it is. In forward flight it has the same characteristics as an aeroplane wing. Pull back on the stick to increase the angle of attack of the disk and you will get a corresponding increase in lift/drag. This occurs whether you're in powered flight OR autorotation, makes no difference. Just so long as you have some positive airspeed. And if our "wing" is producing lift, it must be producing a corresponding downforce.
Now, let's say we enter an auto. As we plummet to earth at 60 knots IAS or whatever, we'll need to start a flare at some point. As the ship passes through thirty feet or so (depending on the rotor diameter of which one we're in), believe it or not we'll start to feel the effects of our ground cushion. Will it be noticeable? That is, would you be able to tell any difference between an autorotative flare done close to the ground versus one done up at altitude? Personally, I've never compared them and the results would be very hard to quantify in any case. But it would be there. It *must* be there.
Once we're slowed to zero and the ship starts to settle, the airflow through the rotor has changed from horizontal to vertical, and the rotor is no longer acting like an airplane wing. We begin to pull up on the collective. With the now-downward flow of air from the rotor, the ship again gets to take advantage of ground-effect one more time.