Fake
wrong +3, it was nothing detected
wrong +5, it says GT wrong but GT is right
correct +2, it should be semantically included
correct+8, it was nothing detected
correct +2, they are wrong
correct +1, they are both right

This leads to:
Equivalences: 9/68 (correct)
Inclusions: 25/68 (correct, +2)
Conflicts: 22/68 (correct +2)
Inconclusive: 12/68 (correct +8 + 1)
Wrong ground truth refinements: 12/24 (-5 because they're are right!)


True
nothing detected --> both correct
conflict --> inconclusive

This leads to:
Equivalences: 16/24 (correct)
Inclusions: 3/24 (correct)
Conflicts: 3/24 (+1 correct)
Inconclusive: 2/24 (+1 inconclusive)
Wrong ground truth refinements: 2/24



