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

    Neuronpedia

    INDEX
    Explanations

    Article

    np_max-act · gemini-2.0-flash

    The neuron activates on the section header token “Article,” detecting when a document is marking off numbered article headings.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Huffman
    -0.08
    _SHADOW
    -0.07
    :new
    -0.07
     Pump
    -0.07
     wink
    -0.07
     SH
    -0.06
     heavier
    -0.06
     Timeout
    -0.06
     wann
    -0.06
    Lov
    -0.06
    POSITIVE LOGITS
     article
    0.12
    Article
    0.11
     articles
    0.11
     Article
    0.10
     ARTICLE
    0.09
    article
    0.09
    ICLE
    0.08
    Circle
    0.08
    /article
    0.08
     Aristotle
    0.08
    Act Density 0.027%

    No Known Activations