INDEX
    Explanations

    Public speaking/events

    This neuron activates on words that describe public figures speaking or delivering remarks—typically verbs like "speaks," "delivers," or "gestures" in photo captions or reporting of speeches.

    New Auto-Interp
    Negative Logits
    ůj
    -0.06
    Translated
    -0.06
     Dubai
    -0.06
    _quit
    -0.06
    ัค
    -0.06
     shemale
    -0.06
     mất
    -0.06
    -0.06
    Signature
    -0.06
     crusher
    -0.06
    POSITIVE LOGITS
     scholars
    0.07
    0.07
    =__
    0.06
    -prepend
    0.06
     выс
    0.06
    ับความ
    0.06
    <Book
    0.06
    0.06
    bet
    0.06
     μην
    0.06
    Act Density 0.011%

    No Known Activations