INDEX
    Explanations
    New Auto-Interp
    Negative Logits
     Gebühren
    -0.08
    ski
    -0.08
    furter
    -0.08
     neka
    -0.08
     west
    -0.08
     w
    -0.08
     Torino
    -0.08
     setzte
    -0.08
     western
    -0.07
    ziger
    -0.07
    POSITIVE LOGITS
    /global
    0.15
    -level
    0.14
    -scale
    0.14
    -global
    0.13
    -wide
    0.13
     overarching
    0.13
    _Global
    0.12
     масш
    0.12
    Global
    0.12
     масштаб
    0.12
    Act Density 0.064%

    No Known Activations