INDEX
    Explanations

    Results may vary

    This neuron fires on hedging and qualifier language in technical explanations—words and phrases like “depends on,” “specific configuration,” “may vary,” and similar qualifiers.

    New Auto-Interp
    Negative Logits
    Jackson
    -0.07
    andscape
    -0.07
     Jackson
    -0.07
    cluding
    -0.06
     Fro
    -0.06
     Registrar
    -0.06
    idious
    -0.06
    esinden
    -0.06
     умовах
    -0.06
    ilage
    -0.06
    POSITIVE LOGITS
     симптом
    0.07
     Meet
    0.06
     سبب
    0.06
    .timezone
    0.06
    Đ
    0.06
    .XPath
    0.06
    特別
    0.06
    (st
    0.06
    =[↵
    0.06
    /***************************************************************************↵
    0.06
    Act Density 0.059%

    No Known Activations