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

    Neuronpedia

    INDEX
    Explanations

    `raise` or `new Todo`

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ;
    0.83
    )
    0.72
    atud
    0.68
    )_
    0.68
    :
    0.66
    atakse
    0.66
    ating
    0.64
    ereal
    0.64
    ISI
    0.61
    )((
    0.58
    POSITIVE LOGITS
    id
    0.75
    an
    0.70
    un
    0.67
    im
    0.64
    να
    0.64
    マ
    0.61
     increasing
    0.59
    ের
    0.59
     콜
    0.59
    ка
    0.58
    Act Density 0.032%

    No Known Activations