Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    This neuron detects document‐structuring elements such as section titles, headers, and metadata (e.g. headings, all‐caps labels, dates, bylines).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     enlarged
    -0.07
    ustos
    -0.06
     urlparse
    -0.06
     SUPPORT
    -0.06
    ovy
    -0.06
     Knife
    -0.06
    SpinBox
    -0.06
    締
    -0.06
     stronger
    -0.06
     waiver
    -0.06
    POSITIVE LOGITS
    —but
    0.07
    $total
    0.06
    ＿＿
    0.06
     cargar
    0.06
    $arity
    0.06
     escri
    0.06
     Cass
    0.06
    —we
    0.06
    Một
    0.06
    Xã
    0.06
    Act Density 0.066%

    No Known Activations