Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    easily susceptible

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    \}$,
    0.93
    ᱐
    0.90
    ле
    0.89
    ką
    0.88
    \}$.
    0.88
     நான்கு
    0.86
    sendKeyDown
    0.86
    qualiter
    0.85
     данной
    0.84
    ０
    0.84
    POSITIVE LOGITS
    is
    1.13
    (
    0.99
     easily
    0.97
    in
    0.89
    an
    0.89
    ل
    0.87
    T
    0.84
    EL
    0.82
     Easily
    0.78
    EM
    0.77
    Act Density 0.012%

    No Known Activations