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

    Neuronpedia

    INDEX
    Explanations

    shadow or shadowing

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ه
    0.91
    ’
    0.90
    le
    0.88
    ت
    0.88
     robuste
    0.84
    由于
    0.83
    ens
    0.82
    .
    0.81
    konstru
    0.81
    यों
    0.80
    POSITIVE LOGITS
    ן
    1.06
    0
    1.05
    ם
    0.98
    ა
    0.93
    ۰
    0.93
    ду
    0.84
    Hunt
    0.84
    지와
    0.84
    গঞ্জ
    0.83
    ၃
    0.82
    Act Density 0.004%

    No Known Activations