Extended Data Fig. 1: The details of similarity scores distributed in the evaluated gene sets.
From: GeneAgent: self-verification language agent for gene-set analysis using domain databases
a. Statistic of gene sets at different similarity levels obtained by GeneAgent and GPT-4 (Hu et al.). The minimum cutoff value of similarity score is set to 0.7. b. Distribution of gene sets in the Gene Ontology dataset (n = 1000) that exhibit a higher similarity score to the ancestors of ground truth than the ground truth itself. The similarity scores of the gene sets included in the analysis range from 0.7 to 0.9 when compared with the ground truth.
