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

    Neuronpedia

    INDEX
    Explanations

    cannot provide illegal/harmful information

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    0.77
     some
    0.69
     the
    0.65
    ,
    0.64
    :
    0.63
     (
    0.61
    .
    0.61
     polych
    0.58
     three
    0.57
     four
    0.57
    POSITIVE LOGITS
    would
    0.98
     Would
    0.87
    Would
    0.87
     WOULD
    0.80
     serait
    0.76
     impedir
    0.75
     avrebbe
    0.75
    latego
    0.74
     would
    0.73
    任何人
    0.73
    Act Density 0.001%

    No Known Activations