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

    Neuronpedia

    INDEX
    Explanations

    How-to guides

    np_max-act · gemini-2.0-flash

    the neuron's sensitive to content-bearing topic words/nouns (important keywords like "sword", "survey", "speech", "AI") in user queries.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     rsp
    -0.09
     пять
    -0.07
    /window
    -0.07
    	top
    -0.06
    /list
    -0.06
     зг
    -0.06
    yecto
    -0.06
    (My
    -0.06
    دي
    -0.06
     alist
    -0.06
    POSITIVE LOGITS
    comm
    0.06
    $wp
    0.06
     volunteering
    0.06
    auer
    0.06
    .moves
    0.06
     antic
    0.06
    :"",
    0.06
    -handler
    0.06
     propel
    0.06
    ]"↵
    0.06
    Act Density 0.144%

    No Known Activations