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

    Neuronpedia

    INDEX
    Explanations

    Movie plot synopses

    np_max-act · gemini-2.0-flash

    This neuron detects mentions of central characters (proper names) and strong narrative actions—tokens that mark who is acting in the story.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     SME
    -0.07
    intval
    -0.07
    blink
    -0.07
     taxing
    -0.06
     Stanley
    -0.06
     Off
    -0.06
     '',
    -0.06
     Trem
    -0.06
    luví
    -0.06
    482
    -0.06
    POSITIVE LOGITS
     حاضر
    0.07
    /system
    0.06
    ุธ
    0.06
     elle
    0.06
    ulously
    0.06
    스크
    0.06
     flute
    0.06
    /music
    0.06
    (nb
    0.06
     vinden
    0.06
    Act Density 0.045%

    No Known Activations