INDEX
    Explanations

    instances where the word "instead" is used

    New Auto-Interp
    Negative Logits
    Ehh
    -0.64
    Lma
    -0.59
    Rgds
    -0.57
    Tbh
    -0.56
    Sooo
    -0.56
    Hahah
    -0.56
    Whoo
    -0.56
    Noice
    -0.55
    Ahhhh
    -0.55
     Perfor
    -0.53
    POSITIVE LOGITS
    Instead
    0.72
     Instead
    0.71
     instead
    0.65
    instead
    0.61
    Espèce
    0.58
     izvē
    0.58
    <bos>
    0.56
    actéristique
    0.56
     spē
    0.54
     ļ
    0.54
    Act Density 0.048%

    No Known Activations