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

    Neuronpedia

    INDEX
    Explanations

    9

    np_max-act · gemini-2.0-flash

    This neuron detects digit ranges in regex character classes (e.g. the “0-9” in patterns like `[a-zA-Z0-9._%+-]`).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Measurement
    -0.06
    .Sync
    -0.06
    survey
    -0.06
     sider
    -0.06
    GC
    -0.06
    _BIT
    -0.06
     wait
    -0.06
    .INFO
    -0.06
     disclosing
    -0.06
     polygon
    -0.06
    POSITIVE LOGITS
     dağ
    0.07
     техні
    0.07
     demos
    0.06
     seçim
    0.06
    _ctx
    0.06
    /use
    0.06
    htar
    0.06
    했던
    0.06
    toMatchSnapshot
    0.06
     dejting
    0.06
    Act Density 0.004%

    No Known Activations