Table 8 Pearson’s correlation between word-similarity/Bio-SimVerb and Bio-SimLex scores and the NER tasks evaluated on biomedical representation models trained with different window sizes

UMN-rel

-0.78^a

-0.56

-0.78^a

-0.46

UMN-sim

-0.73

-0.57^a

-0.81

-0.42^a

MayoSRS

-0.78

-0.69

-0.54^a

-0.47^a

Bio-SimVerb

0.63

0.36

0.42

0.40

Bio-SimLex

0.83 ^a

0.66

0.92 ^a

0.59