If one tries to sort the abstracts by score, then I observe the following behaviour:
Sort ascending: lowest score to highest score, then unreviewed abstracts
Sort descending: unreview, then highest to lowest score.
As the latter sorting is what one would use for accepting proposals, I would prefer to get the unreviewed abstracts last. For acsending, unreviewed first would be OK.
I others also agree, I could try to prepare a patch.