Summary
ESMFold has massive activations (3000x the median) that disproportionately contribute to structure prediction quality (1). Authors describe it as “pathologically disorganized”.
Figures

Ref (1)
1.
Lu AX, Yan W, Yang KK, Gligorijevic V, Cho K, Abbeel P, et al. Tokenized and continuous embedding compressions of protein sequence and structure. Patterns. 2025;6(6):101289. Available from: https://doi.org/10.1016/j.patter.2025.101289