INDEX
    Explanations

    precedes "other" or "audit"

    New Auto-Interp
    Negative Logits
    一边
    0.48
     redact
    0.46
     zupeł
    0.44
     procéder
    0.42
     Roy
    0.41
     Jeżeli
    0.40
     实现
    0.40
     ONLY
    0.40
    一张
    0.40
     Nutrition
    0.39
    POSITIVE LOGITS
    disturbance
    0.48
    และการ
    0.45
    круг
    0.44
     precedes
    0.44
     لاز
    0.43
    зву
    0.43
     gangguan
    0.41
    attacks
    0.40
    fing
    0.40
     disturbs
    0.40
    Act Density 0.036%

    No Known Activations