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

    Neuronpedia

    INDEX
    Explanations

    punctuation separators

    np_max-act · gemini-2.0-flash

    The neuron primarily detects section headings, titles, and other document-structure metadata (e.g., headings, subheadings, and labeled sections).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     GridBagConstraints
    -0.07
    ักด
    -0.07
     Workbook
    -0.07
     CARE
    -0.06
    _ARRAY
    -0.06
    lectual
    -0.06
    stal
    -0.06
     Zot
    -0.06
    '
    -0.06
     valve
    -0.06
    POSITIVE LOGITS
     美国
    0.06
    ворю
    0.06
    仔
    0.06
     },{↵
    0.06
    ($(
    0.06
     نويسنده
    0.06
    "),
↵
    0.06
     قبل
    0.06
    ()){
↵
    0.06
     ساخته
    0.06
    Act Density 0.051%

    No Known Activations