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

    Neuronpedia

    INDEX
    Explanations

    database table and field names

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     is
    0.39
     使用
    0.36
    لي
    0.36
    دين
    0.34
     مرور
    0.34
     طريق
    0.33
    しながら
    0.33
     使っ
    0.33
     很
    0.33
     細胞
    0.32
    POSITIVE LOGITS
    and
    0.58
    im
    0.57
     for
    0.55
    ap
    0.53
    ik
    0.53
    ong
    0.53
    ang
    0.50
    for
    0.50
    ad
    0.49
    an
    0.49
    Act Density 0.759%

    No Known Activations