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

    Neuronpedia

    INDEX
    Explanations

    is for defining purpose

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    明らか
    0.46
    вард
    0.41
     สาว
    0.39
     jūs
    0.38
     odlu
    0.38
    特点
    0.38
    цвет
    0.38
    可以选择
    0.37
    ionar
    0.37
     ricordare
    0.36
    POSITIVE LOGITS
     intended
    0.64
     offered
    0.61
     meant
    0.60
    intended
    0.54
     for
    0.52
     bedo
    0.50
     presented
    0.48
     solely
    0.45
    Just
    0.45
     Offered
    0.45
    Act Density 0.002%

    No Known Activations