Yes, the A/P does flare noticeably later. It seems that the average autoland drifts a shade above the G/S below 70 ft. This is followed by a slight reduction in pitch immdeiately prior to the beginning of the flare. This is what made me wonder about the impact of ground effect. I presume the A/P would be trying to not be affected by ground effect until it was ready to flare, hence the effort to remain on G/S until the inertial reference takes over and the flare begins. Interesting.