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

    Neuronpedia

    INDEX
    Explanations

    used for actions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     incurring
    0.42
    uert
    0.39
    uerung
    0.39
    予測
    0.39
     resorting
    0.38
    using
    0.38
    度は
    0.37
     relying
    0.37
    essive
    0.37
     wasteland
    0.36
    POSITIVE LOGITS
     used
    1.08
    用于
    1.06
    用来
    1.05
     digunakan
    1.04
     dipakai
    1.00
     brukes
    0.96
    Used
    0.95
     Used
    0.94
     用于
    0.91
    用於
    0.91
    Act Density 0.010%

    No Known Activations