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

    Neuronpedia

    INDEX
    Explanations

    while, can, were, focus, always, are

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    el
    0.94
    endroit
    0.88
     поэтому
    0.88
    $.
    0.84
    %.
    0.83
    waż
    0.81
    at
    0.80
    amist
    0.80
    wenn
    0.80
    ఫ్‌
    0.79
    POSITIVE LOGITS
     **,
    1.16
     *,
    0.98
    ³,
    0.94
    ,
    0.93
     »,
    0.91
    +,
    0.90
    **,
    0.89
    »,
    0.89
    «,
    0.87
    ?",
    0.86
    Act Density 0.669%

    No Known Activations