Table 7.

Entity-level performance of PubMedBERT on the gold validation set. Original scores refer to the performance of the model before self-training; self-training scores refer to the performance of the model at the best iteration (6th iteration) of self-training. ‘R’ and ‘P’ refer to recall and precision, respectively

	Original scores			Self-training scores
	F1	R	P	F1	R	P
Comparator	33.3	66.7	22.2	80.0	66.7	100.0
Induction	46.2	50.9	42.3	45.7	40.7	52.2
Intervention	67.3	69.6	65.2	69.6	75.0	64.9
Outcome	61.5	68.0	56.2	69.9	73.5	66.7
Species	96.4	99.1	93.9	96.4	99.1	93.9
Strain	80.0	80.0	80.0	90.9	100.0	83.3
Overall	68.1	73.0	63.8	73.6	76.4	71.0