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

    Neuronpedia

    INDEX
    Explanations

    melanin, skin, hair color

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    f
    1.48
    h
    1.42
    ق
    1.29
    ز
    1.22
    m
    1.17
    ö
    1.16
    i
    1.13
    an
    1.11
    is
    1.09
    t
    1.09
    POSITIVE LOGITS
     为
    1.04
     bằng
    0.93
    ↵
    0.89
     将
    0.88
     还
    0.85
    '
    0.82
     كان
    0.82
     数
    0.81
    까지
    0.80
     หน้า
    0.80
    Act Density 0.007%

    No Known Activations