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

    Neuronpedia

    INDEX
    Explanations

    Dates and numbers

    np_max-act · gemini-2.0-flash

    This neuron activates on solitary numeric digit tokens (e.g. the “3” in the date), i.e. it flags individual digits.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Cod
    -0.06
     Manuals
    -0.06
    板
    -0.06
    _RELEASE
    -0.06
    ��
    -0.06
    สาห
    -0.06
     hr
    -0.06
    ιαν
    -0.06
     tanın
    -0.06
     dispatched
    -0.06
    POSITIVE LOGITS
     Pet
    0.07
    χει
    0.06
     Ange
    0.06
    tığ
    0.06
    кс
    0.06
     Venture
    0.06
     OMG
    0.06
    .gridy
    0.06
    toInt
    0.06
     Dolphins
    0.06
    Act Density 0.015%

    No Known Activations