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

    Neuronpedia

    INDEX
    Explanations

    URLs with date numbers

    np_max-act · gemini-2.0-flash

    The neuron selectively fires on short numeric tokens (e.g. single‐ or two‐digit numbers) that typically appear as date or version markers in URLs or text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    your
    -0.07
     himself
    -0.07
     aired
    -0.06
    _version
    -0.06
     Trade
    -0.06
     yourself
    -0.06
    upport
    -0.06
    bal
    -0.06
    ived
    -0.06
     lantern
    -0.06
    POSITIVE LOGITS
    .environment
    0.07
    گو
    0.07
     السي
    0.07
     분류
    0.07
    DispatchToProps
    0.07
     Tata
    0.06
     Pb
    0.06
     đúng
    0.06
     انتخاب
    0.06
     fraction
    0.06
    Act Density 0.002%

    No Known Activations