INDEX
    Explanations

    The neuron responds to descriptive style and quality terms in image‐generation prompts (e.g. words that specify artistic detail, composition, and rendering qualities).

    New Auto-Interp
    Negative Logits
     ******************************************************************************/↵↵
    -0.07
    ึก
    -0.06
     Sik
    -0.06
    oğunluk
    -0.06
     Bbw
    -0.06
    ющих
    -0.06
     Mayweather
    -0.06
     sürec
    -0.06
    mts
    -0.05
    ,/
    -0.05
    POSITIVE LOGITS
     volcan
    0.08
    .descripcion
    0.07
     dental
    0.06
    َج
    0.06
     implies
    0.06
     Vish
    0.06
     Thermal
    0.06
     Fuel
    0.06
    غان
    0.06
     bölge
    0.06
    Act Density 0.056%

    No Known Activations