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

    Neuronpedia

    INDEX
    Explanations

    mentioning models and names

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     росій
    0.28
     цей
    0.28
     quốc
    0.27
     மனித
    0.27
     سیاه
    0.27
     Russische
    0.27
    shakespeare
    0.27
     österreich
    0.26
     რუს
    0.26
    此同时
    0.26
    POSITIVE LOGITS
    +
    0.35
     reduction
    0.32
    '
    0.30
    ’
    0.28
    和
    0.27
     being
    0.27
     triggering
    0.27
     giving
    0.27
     and
    0.26
    -
    0.26
    Act Density 0.218%

    No Known Activations