split,attribute_reasoning,coarse_perception,finegrained_perception (cross-instance),finegrained_perception (instance-level),logic_reasoning,relation_reasoning
full,0.7164179104477612,0.7651006711409396,0.5586206896551724,0.67003367003367,0.39166666666666666,0.6869565217391305
dev,0.7164179104477612,0.7651006711409396,0.5586206896551724,0.67003367003367,0.39166666666666666,0.6869565217391305
test,,,,,,