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

    Neuronpedia

    INDEX
    Explanations

    Programming/data analysis code

    np_max-act · gemini-2.0-flash

    The neuron primarily activates on numeric identifier tokens (e.g. the digit suffixes in “NAME_1” or “NAME_2”), effectively detecting references to numbered placeholders.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    website
    -0.07
    риг
    -0.07
    .website
    -0.07
     root
    -0.06
    兹
    -0.06
     yanında
    -0.06
     Nug
    -0.06
     requestOptions
    -0.06
     PCI
    -0.06
     Visa
    -0.06
    POSITIVE LOGITS
     eer
    0.07
    683
    0.07
    なく
    0.06
    kir
    0.06
    สอง
    0.06
     civic
    0.06
    credited
    0.06
    184
    0.06
     جور
    0.06
    Guardar
    0.06
    Act Density 0.038%

    No Known Activations