INDEX
    Explanations
    New Auto-Interp
    Negative Logits
    shapes
    -0.07
     SUN
    -0.07
     Thick
    -0.07
    _FIRE
    -0.06
    _CHAIN
    -0.06
     accounts
    -0.06
     руки
    -0.06
     clay
    -0.06
     silica
    -0.06
    .MIN
    -0.06
    POSITIVE LOGITS
     lob
    0.10
     Lab
    0.08
     Tribal
    0.07
     exacerbated
    0.07
     đột
    0.06
     librarian
    0.06
    28
    0.06
     bacter
    0.06
     Đức
    0.06
     Refer
    0.06
    Act Density 0.002%

    No Known Activations