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

    Neuronpedia

    INDEX
    Explanations

    negative input handling

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     tycoon
    1.45
    ность
    1.43
    ن
    1.43
    ك
    1.40
    ную
    1.38
    وب
    1.38
    还得
    1.38
     dignitaries
    1.34
     mercantil
    1.33
    筵
    1.33
    POSITIVE LOGITS
    u
    2.09
    ir
    1.77
    IN
    1.65
    o
    1.40
    travis
    1.30
    Rah
    1.27
    conoc
    1.27
    lı
    1.25
     випадку
    1.24
    ுள்ளார்
    1.23
    Act Density 0.000%

    No Known Activations