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

    Neuronpedia

    INDEX
    Explanations

    Persian caviar functions cancelled

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     livre
    0.44
     bonus
    0.44
     "$
    0.44
     oorlog
    0.44
     reports
    0.43
     lista
    0.42
     bi
    0.42
     arre
    0.42
     ($
    0.42
     house
    0.41
    POSITIVE LOGITS
    Vue
    0.45
    uie
    0.45
    非常に
    0.44
    hidden
    0.43
    veu
    0.42
    Ｍ
    0.42
    τέ
    0.41
    gradient
    0.41
    oulders
    0.41
    ਵੇਂ
    0.40
    Act Density 0.001%

    No Known Activations