Fuzzy Overclustering (FOC) In real-world datasets, we need consistent annotations between annotators to give a certain ground-truth label. However, in