highly significantp < 0.001
The overall macro-averaged F1 score reaches 0.935 (95% CI: 0.928–0.942), representing a highly significant improvement compared to the pure large-model baseline (F1 = 0.780, 95% CI: 0.766–0.794, p < 0.001).
The overall macro-averaged F1 score reaches 0.935 (95% CI: 0.928–0.942), representing a highly significant improvement compared to the pure large-model baseline (F1 = 0.780, 95% CI: 0.766–0.794, p < 0.001).