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

    Neuronpedia

    INDEX
    Explanations

    evaluating importance and risk

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     얼마나
    0.50
    细节
    0.40
    áját
    0.40
     কহ
    0.38
    ുകയായിരുന്നു
    0.38
    ეც
    0.38
     কতটা
    0.38
    你怎么
    0.38
    èves
    0.38
    魉
    0.38
    POSITIVE LOGITS
     compared
    0.67
     (>
    0.64
     (<
    0.60
    compared
    0.55
     Compared
    0.51
    বিশিষ্ট
    0.51
     (~
    0.50
     عالية
    0.50
    Compared
    0.50
     (\<
    0.48
    Act Density 0.102%

    No Known Activations