INDEX
    Explanations

    cross validation training data

    New Auto-Interp
    Negative Logits
     novelas
    -0.96
     yall
    -0.88
    Hebrews
    -0.85
     bluetooth
    -0.84
     Ľ
    -0.82
     hice
    -0.80
     Viena
    -0.80
    AutoSize
    -0.79
    一样的
    -0.79
     mantan
    -0.79
    POSITIVE LOGITS
     discut
    0.91
    0.85
    Приготовление
    0.84
     надо
    0.81
     мона
    0.80
    ाप्त
    0.80
     discuter
    0.79
     superb
    0.78
    }$)
    0.77
     income
    0.77
    Act Density 0.001%

    No Known Activations