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

    Neuronpedia

    INDEX
    Explanations

    ---

    np_max-act · gemini-2.0-flash

    This neuron detects top‐level document titles—that is, lines beginning with “# ” that mark the main heading.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    egg
    -0.06
     qreal
    -0.06
    bpp
    -0.06
     yog
    -0.06
    tube
    -0.06
    leet
    -0.06
     حديث
    -0.06
     bekannt
    -0.06
    ः
    -0.06
    Calculate
    -0.06
    POSITIVE LOGITS
     Edison
    0.08
     lumber
    0.06
    .Cursors
    0.06
    娱乐
    0.06
     może
    0.06
    (KEY
    0.06
     ク
    0.06
    waves
    0.06
    -binary
    0.06
    cly
    0.06
    Act Density 0.000%

    No Known Activations