INDEX
    Explanations
    New Auto-Interp
    Negative Logits
     repel
    -0.08
    verb
    -0.08
    vorming
    -0.08
     persone
    -0.07
    object
    -0.07
    ackage
    -0.07
    res
    -0.07
     res
    -0.07
     الهدف
    -0.07
     معين
    -0.07
    POSITIVE LOGITS
     қар
    0.09
     Ina
    0.08
     disgrace
    0.08
     velmi
    0.08
     Calc
    0.08
     Cozy
    0.08
    "So
    0.08
     Ako
    0.08
    テレビ
    0.08
     whom
    0.08
    Act Density 0.023%

    No Known Activations