Encoder-decoder PLM embeddings are better able to distinguish nearest neighbors than encoder-only PLMs

Summary

The encoder-decoder protein language model ProtT5 is better able to distinguish near neighbors than the encoder-only PLM ESM C (1).

Figures

Ref (1)

1.

Rissom PF, Sarmiento PY, Safer J, Coley CW, Renard BY, Heyne HO, et al. Decoding protein language models: insights from embedding space analysis. openRxiv; 2024. Available from: https://doi.org/10.1101/2024.06.21.600139