There are two (in)famous cases of pacesetters continuing the race and going on to win: Paul Pilkington at the 1994 Los Angeles marathon, and Tom Byers at the 1981 Bislett Games.
Byers built a ten-second lead when "the rest of the field refused to follow his pace" and Pilkington similarly was well away from the rest of the pack at the point he was supposed to step off.
Other answers have correctly noted that some events make their rabbits sign contracts to drop out, and IAAF rules require that any pacemakers start the race with everyone else (otherwise the pacemakers constitute unfair assistance for some runners).
For events where dropping out is not a part of the contract, though, consider this: what motivation do the other runners have to follow the pace set by a rabbit when they know the rabbit will drop out? Put another way, what's the point of having a pacemaker run 3/4 of a race out in front and then drop out while still in the lead?
Perfect pacemaking requires the rabbit to hit exactly the requested splits at the designated milestones, but also requires the actual competitors to be close in with the pacemakers. Allowing the pacemakers to finish the race motivates the rest of the runners to at least keep them in sight, to avoid piracy. I've had a few things to say on this subject elsewhere.
If a pacemaker reaches their designated drop-out point but is still leading and (important point) still feels capable of holding the lead, they can actually continue to be useful by staying in the race. Some races pay pacemakers more the longer they can continue leading at the agreed-upon pace. In these cases the rabbit would drop once they found themselves falling off the pace (or, on the track, if the rest of the field begins passing them).