INDEX
    Explanations

    mathematical derivatives and expressions

    New Auto-Interp
    Negative Logits
     will
    -0.90
     during
    -0.85
     bendera
    -0.81
     what
    -0.81
     Jovi
    -0.80
    わけで
    -0.78
     {}",
    -0.77
    just
    -0.77
     giving
    -0.77
    なお
    -0.77
    POSITIVE LOGITS
     quizás
    1.02
     pow
    0.94
     ánimo
    0.94
    ^{*}\
    0.94
     debería
    0.93
     manchmal
    0.93
     jejich
    0.90
     parfois
    0.90
     monstrous
    0.90
     retraso
    0.89
    Act Density 0.485%

    No Known Activations