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

    Neuronpedia

    INDEX
    Explanations

    special character pairings

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    ו
    0.57
    larının
    0.54
    l
    0.53
     offsetting
    0.50
    ุ
    0.50
    ।]
    0.49
    Thickness
    0.48
     framing
    0.48
    <unused61>
    0.48
     ٹرسٹ
    0.47
    POSITIVE LOGITS
    semble
    0.50
    é
    0.50
     vâr
    0.47
    un
    0.47
    acije
    0.47
     castle
    0.46
    igne
    0.46
    hé
    0.46
    yce
    0.45
    ujourd
    0.45
    Act Density 0.001%

    No Known Activations