INDEX
    Explanations

    native speakers, without, intending to

    New Auto-Interp
    Negative Logits
    uetooth
    0.65
     입니다
    0.62
    /
    0.62
     使用
    0.59
     環境
    0.59
     daß
    0.59
     মোঃ
    0.59
     /
    0.58
     mediante
    0.58
    bedaan
    0.58
    POSITIVE LOGITS
    がたくさん
    0.77
     unaffected
    0.73
     perks
    0.73
     grassroots
    0.71
     बिना
    0.70
     малень
    0.70
     всіх
    0.70
     fervent
    0.70
     handmade
    0.69
    ไม่มี
    0.69
    Act Density 0.158%

    No Known Activations