face recognition vs. keyword list

lately i tried to get used to the face recognition feature of LRC. 

I think its great , for i shoot a lot of public events and like to find local celebs later. 

but

now all the countless names clog my keyword list and make it much harder to edit all other keywords.

I made Groups for them but its still annoying

for instance: automativ keywords was nice before, type in 2 letters and choose beween 2-3 keywords. Now all combination of 2 or 3 letters lead to a lot of names. 

do i make a mistake?

is there a way to separate the faces from the keywords?