INDEX
    Explanations

    The neuron detects descriptive adjectives of a woman’s physical appearance—especially hair color (e.g. brunette, blonde, redhead) and body descriptors like “busty.”

    New Auto-Interp
    Negative Logits
    :"↵
    -0.07
    samples
    -0.07
     Covers
    -0.06
    wor
    -0.06
       
    -0.06
     Old
    -0.06
    -0.06
     sided
    -0.06
    liest
    -0.06
     [{"
    -0.06
    POSITIVE LOGITS
     brunette
    0.08
     redhead
    0.07
     gazet
    0.07
     debuted
    0.07
     deve
    0.07
     approached
    0.07
    peg
    0.07
     streamline
    0.06
     retain
    0.06
     retire
    0.06
    Act Density 0.003%

    No Known Activations