They're looking for all of the above, but I'd say they're looking for "average quality good" if you'll pardon that ungrammatical expression. What I mean is that if they look at all of the names one by one, the overall quality is "good" or better. A lot of junk names padding out a longer list would put them off, probably enough to kill any deal.
For example, dictionary word domains that have zero commercial potential (one word "generics" that are "special" only because they happen to have made it into the pages of a dictionary), typos, obvious TM violations etc. have no place in that kind of ready-for-primetime bulk portfolio. They would IMO actually reduce the overall value i.e. they have an incremental NEGATIVE value.
I doubt they'd be interested in "brandable" names either i.e. names that have no intrinsic value until somebody comes along and thinks "hmm, I could build a business around that name".
100% COMMERCIAL generics all the way is the way to go.
There would probably also have to be a strong core of stand-out "anchor" domains, domains that would make any serious buyer sit up and take notice. The opportunity of scooping them all up along with a swathe of good secondary names should be that much more appealing for their presence.
Of course, this is all speculation... until a big deal is done, it's hard to know what would work and what wouldn't!
