INDEX
    Explanations

    description

    The neuron activates on HTML meta‐tag attributes (especially the “description” and “keywords” tags and their content values).

    New Auto-Interp
    Negative Logits
    Stories
    -0.07
     Lenovo
    -0.06
     ngữ
    -0.06
    .wh
    -0.06
     ceremony
    -0.06
    -0.06
     Gram
    -0.06
     Arabia
    -0.06
    .Ag
    -0.06
    gew
    -0.06
    POSITIVE LOGITS
    jištění
    0.07
     struct
    0.07
    ingle
    0.06
    azes
    0.06
     ".");↵
    0.06
     rebel
    0.06
    ющее
    0.06
     Explain
    0.06
    Street
    0.06
    циклопед
    0.06
    Act Density 0.002%

    No Known Activations