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

    Neuronpedia

    INDEX
    Explanations

    former state of being lost

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     있도록
    0.75
    hidden
    0.73
    ěr
    0.73
     நடந்து
    0.73
    chodu
    0.71
    boten
    0.71
     czas
    0.71
    प्ति
    0.70
     ಅನ
    0.70
    ้าว
    0.70
    POSITIVE LOGITS
     once
    1.63
     formerly
    1.41
    Once
    1.31
    once
    1.28
     Once
    1.25
    曾经
    1.25
     previously
    1.23
    原本
    1.23
     hope
    1.20
    曾經
    1.19
    Act Density 0.172%

    No Known Activations