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

    Neuronpedia

    INDEX
    Explanations

    emissions

    np_max-act · gemini-2.0-flash

    The neuron chiefly activates on numeric tokens (especially floating‐point numbers), i.e. it finds numerical data.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     wakeup
    -0.07
    orphic
    -0.07
    .hist
    -0.06
    ropoda
    -0.06
     bak
    -0.06
     Созд
    -0.06
     rencontres
    -0.06
    pag
    -0.06
     guarded
    -0.06
     wurde
    -0.06
    POSITIVE LOGITS
     emissions
    0.11
    飞
    0.06
    ประม
    0.06
    .listen
    0.06
     electronics
    0.06
     accr
    0.06
    contin
    0.06
    mem
    0.06
     SR
    0.06
    \(
    0.06
    Act Density 0.004%

    No Known Activations