INDEX
    Explanations
    New Auto-Interp
    Negative Logits
     carriage
    -0.08
     enf
    -0.08
     kish
    -0.08
     gated
    -0.08
     hoger
    -0.08
     stamped
    -0.08
     dresser
    -0.08
     compounded
    -0.07
     ali
    -0.07
     školy
    -0.07
    POSITIVE LOGITS
    acies
    0.09
     verifying
    0.08
     überprüfen
    0.08
     finale
    0.07
    acy
    0.07
     Ziel
    0.07
     hinzu
    0.07
     auch
    0.07
     задерж
    0.07
    atex
    0.07
    Act Density 0.081%

    No Known Activations