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

    Neuronpedia

    INDEX
    Explanations

    you

    np_max-act · gemini-2.0-flash

    This neuron primarily fires on words that begin new paragraphs or major text segments (i.e. tokens at the start of a block).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    discourse-structuring elements that signal organization, such as section headings, list markers, and transitional connectors.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
     Levin
    -0.07
    èle
    -0.06
    eling
    -0.06
    _SPEED
    -0.06
    แนะนำ
    -0.06
    	active
    -0.06
    %D
    -0.06
    ypse
    -0.06
    �
    -0.06
    _em
    -0.06
    POSITIVE LOGITS
    >Password
    0.07
    =\"$
    0.06
    recv
    0.06
    -food
    0.06
     vagy
    0.06
     experiencing
    0.06
    _INS
    0.06
    ленні
    0.06
    、どう
    0.06
    ようです
    0.06
    Act Density 0.160%

    No Known Activations