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

    Neuronpedia

    INDEX
    Explanations

    ato

    np_max-act · gemini-2.0-flash

    The neuron lights up on tokens that look like identifiers—e.g. proper names, acronyms, version‐numbers or other alphanumeric codes containing uppercase letters and/or digits.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     tuyệt
    -0.07
    .RichTextBox
    -0.06
     wall
    -0.06
     nech
    -0.06
     Č
    -0.06
    [ii
    -0.06
    _n
    -0.06
     �
    -0.06
    _texts
    -0.06
     frat
    -0.06
    POSITIVE LOGITS
    Patterns
    0.07
    _configs
    0.06
     *↵
    0.06
    (message
    0.06
     metallic
    0.06
    central
    0.06
     Gibson
    0.06
    istrates
    0.06
    .microsoft
    0.06
    lichkeit
    0.06
    Act Density 0.635%

    No Known Activations