Fig. 4
From: Comparative Assessment of Protein Large Language Models for Enzyme Commission Number Prediction
Comparison of DNN-LLM models’ and BLAST’s F1 scores. a,c,e Distribution of DNN-LLM models’ and BLASTp’s F1 score differences per number of sequences in the training set (\(\hbox {log}_{{10}}\)), where each dot represents an EC class. b,d,f Number of classes with an \(\Delta \text {F1}_i\) above a given threshold