add notes to gene embedding benchmark

_data/transformer-evaluation.yml

@@ -9,7 +9,7 @@
   omic_modalities: Natural Language, scRNAseq, protein sequence, DNA sequence
   evaluated_transformers: Mistral, MPnet, CellPLM, GeneFormer, scGPT, DNABert2, ESM (plus bag-of-words, gene2vec)
   tasks: genomic properties, gene regulatory functions, gene localization, gene biological processes, protein properties
-  notes: '-'
+  notes: 'Evaluating model performance using gene embeddings allows comparing models utilizing different data modalities. The performance profile of a diverse set of models across ~300 gene related tasks was shown to be most influenced by the training modality.'
 - paper:
     type: preprint
     text: '[Ahlmann-Eltze et al. 2024](https://www.biorxiv.org/content/10.1101/2024.09.16.613342)'