precision recall f1-score support entailment 0.57 0.67 0.62 750 contradiction 0.60 0.46 0.52 737 neutral 0.60 0.64 0.62 777 accuracy 0.59 2264 macro avg 0.59 0.59 0.59 2264 weighted avg 0.59 0.59 0.59 2264 Confusion matrix: [[502 112 136] [207 338 192] [166 115 496]]