INDEX
    Explanations

    compression

    New Auto-Interp
    Negative Logits
    ok
    -0.07
     Fa
    -0.07
     Stories
    -0.06
    inge
    -0.06
     Tale
    -0.06
     Eck
    -0.06
     Xu
    -0.06
     Tony
    -0.06
     Ren
    -0.06
    oy
    -0.06
    POSITIVE LOGITS
     compress
    0.09
     compressor
    0.08
    compress
    0.08
     compressed
    0.08
    .compress
    0.08
    sending
    0.08
     compression
    0.07
     Compression
    0.07
    suppress
    0.07
     narcotics
    0.07
    Act Density 0.004%

    No Known Activations