INDEX
    Explanations

    wrestling matches

    This neuron fires on tokens related to professional wrestling broadcast commentary—particularly wrestler and show names (e.g. SmackDown, Champion, Joe, Sheamus) and match/action vocabulary.

    New Auto-Interp
    Negative Logits
    Index
    -0.07
    .ReadAllText
    -0.07
    PLEX
    -0.06
    orpion
    -0.06
    View
    -0.06
     Mahmoud
    -0.06
     مهندسی
    -0.06
    sql
    -0.06
    _mes
    -0.06
     cultures
    -0.06
    POSITIVE LOGITS
     대구
    0.07
    ázi
    0.07
    _ACTIVITY
    0.07
    тра
    0.06
    (create
    0.06
    بو
    0.06
    غراف
    0.06
     CGPoint
    0.06
    clazz
    0.06
     يح
    0.06
    Act Density 0.011%

    No Known Activations