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

    Neuronpedia

    INDEX
    Explanations

    absence, rejection, or failure

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     breaches
    0.92
    贋
    0.83
    罠
    0.83
    Cmd
    0.83
    যেমন
    0.82
    okość
    0.82
     লুট
    0.79
     потеря
    0.78
     거짓
    0.77
     inaccuracy
    0.77
    POSITIVE LOGITS
     altogether
    1.20
     entirely
    1.02
     unless
    0.96
    Unless
    0.92
     outright
    0.92
     completely
    0.91
    luster
    0.89
     except
    0.86
    Except
    0.86
    unless
    0.85
    Act Density 0.396%

    No Known Activations