Fig. 3From: Benchmarking for biomedical natural language processing tasks with a domain specific ALBERTComparison of BioALBERT versus ALBERT. The evaluation scale is same as previously reported in Table 7Back to article page