INDEX
    Explanations

    statistical punctuation

    comparisons and performance differences

    New Auto-Interp
    Negative Logits
    раб
    0.66
     arbete
    0.65
     interessante
    0.62
     woorden
    0.61
    Museum
    0.61
     পুস্তকের
    0.61
     архитек
    0.60
    änk
    0.60
     berühm
    0.59
     écrire
    0.59
    POSITIVE LOGITS
     gastro
    0.63
    s
    0.61
     wrongful
    0.60
    d
    0.57
    b
    0.57
    )\
    0.56
     dodgy
    0.55
     pesky
    0.55
    (
    0.55
     toxic
    0.55
    Act Density 0.023%

    No Known Activations