INDEX
    Explanations
    New Auto-Interp
    Negative Logits
     attempts
    -0.09
     halal
    -0.08
    -0.08
     crosses
    -0.08
     convenient
    -0.08
    יכ
    -0.07
    Attempt
    -0.07
     новом
    -0.07
    려고
    -0.07
    тән
    -0.07
    POSITIVE LOGITS
     leaks
    0.09
     Leak
    0.08
     leak
    0.08
    .optimize
    0.08
     pérdidas
    0.08
     πυ
    0.08
     desperd
    0.08
    0.08
     anbef
    0.08
    ólicos
    0.07
    Act Density 0.019%

    No Known Activations