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

    Neuronpedia

    INDEX
    Explanations

    matrices

    np_max-act · gemini-2.0-flash

    The neuron activates on fragments of LaTeX‐style math expressions (e.g. delimiters like “\$” or “\\begin”, Greek letters, matrix syntax).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    661
    -0.07
     orn
    -0.07
    僕
    -0.06
     treff
    -0.06
    spi
    -0.06
     uomini
    -0.06
     Neuroscience
    -0.06
     бла
    -0.06
     wich
    -0.06
    §ط
    -0.06
    POSITIVE LOGITS
     protože
    0.07
    ektör
    0.06
    astery
    0.06
    렵
    0.06
    Surface
    0.06
     airst
    0.06
     Speed
    0.06
    =R
    0.06
    董事
    0.06
     fats
    0.06
    Act Density 0.001%

    No Known Activations