Models fine-tuned on the SFinD dataset for Question-Answer alignment classification, producing outputs labeled as aligned or misaligned.