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

    Neuronpedia

    INDEX
    Explanations

    exceptions related to roles

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    und
    0.46
    wy
    0.45
    dy
    0.44
    ulation
    0.44
    baw
    0.44
    ulit
    0.44
    a
    0.42
    on
    0.41
    after
    0.41
    b
    0.41
    POSITIVE LOGITS
     exceptions
    0.73
    Exceptions
    0.66
     Exceptions
    0.65
     excepción
    0.64
    例外
    0.60
     excepciones
    0.59
    Exception
    0.56
    exceptions
    0.56
     exception
    0.55
     Ausnahme
    0.54
    Act Density 0.000%

    No Known Activations