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

    Neuronpedia

    INDEX
    Explanations

    front and forward direction

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     apunt
    0.83
    ъм
    0.82
    碎
    0.82
     mengh
    0.82
    ⿶
    0.81
     notas
    0.81
     pale
    0.79
     mesin
    0.78
    )}{(
    0.77
     apunta
    0.77
    POSITIVE LOGITS
    garde
    1.10
    wards
    1.06
    endment
    1.04
    🚀
    1.03
    瞻
    1.03
     والخ
    1.03
    aliers
    1.03
    matter
    1.02
    rances
    1.01
    髪
    1.00
    Act Density 0.145%

    No Known Activations