INDEX
    Explanations
    New Auto-Interp
    Negative Logits
    ,is
    -0.08
    имент
    -0.07
     spont
    -0.07
    achelor
    -0.07
     approached
    -0.07
     maternity
    -0.07
    -0.07
    -0.07
     оны
    -0.07
     disent
    -0.07
    POSITIVE LOGITS
     loyal
    0.10
    之一
    0.09
    0.08
     duo
    0.08
    0.08
     accompany
    0.07
    0.07
    0.07
     questioning
    0.07
     prosecut
    0.07
    Act Density 0.014%

    No Known Activations