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

    Neuronpedia

    INDEX
    Explanations

    datasets

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens containing digits—i.e. it picks out numbers (including years, floating‐point values, version numbers, coordinates, etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     completed
    -0.07
     indicted
    -0.07
     records
    -0.07
     power
    -0.06
     house
    -0.06
     image
    -0.06
     stations
    -0.06
    氏
    -0.06
     طرف
    -0.06
     Mosque
    -0.06
    POSITIVE LOGITS
     важ
    0.07
     uLocal
    0.06
    irtschaft
    0.06
    zsche
    0.06
    fclose
    0.06
    UIButton
    0.06
     useForm
    0.06
     ACM
    0.06
     mustard
    0.06
    .forName
    0.06
    Act Density 0.021%

    No Known Activations