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

    Neuronpedia

    INDEX
    Explanations

    code/data

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on numeric tokens—especially sequences of digits and decimal points—i.e. it’s a detector for numbers (in particular floating‐point or multi‐digit numeric strings).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    є
    -0.08
     على
    -0.07
     throwError
    -0.07
    keterangan
    -0.07
    184
    -0.06
    iona
    -0.06
    :&
    -0.06
     imdb
    -0.06
    -br
    -0.06
    烧
    -0.06
    POSITIVE LOGITS
    _node
    0.07
    rud
    0.07
     simplified
    0.06
     hrs
    0.06
    -storage
    0.06
     soph
    0.06
    DR
    0.06
    Queen
    0.06
     DOWNLOAD
    0.06
     модель
    0.06
    Act Density 0.016%

    No Known Activations