INDEX
    Explanations

    specific pairings and contexts

    New Auto-Interp
    Negative Logits
     next
    0.34
     three
    0.33
     two
    0.33
     in
    0.32
     four
    0.32
     on
    0.31
    四个
    0.31
     Performance
    0.30
     তিন
    0.30
     frontend
    0.29
    POSITIVE LOGITS
     школова
    0.36
    0.33
     monaster
    0.33
     moléculas
    0.32
     physicians
    0.32
    habitants
    0.31
     sabbam
    0.31
     peasants
    0.31
     სახელმწიფ
    0.31
    oniazid
    0.31
    Act Density 0.001%

    No Known Activations