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

    Neuronpedia

    INDEX
    Explanations

    technical documents

    np_max-act · gemini-2.0-flash

    This neuron activates on capitalized words, especially those at the start of sentences or headings.

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

    domain-specific technical terminology, especially in formal or scientific/technical contexts (often capitalized or used as section headings, proper nouns, or key metrics).

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    エ
    -0.07
     Laguna
    -0.07
    光
    -0.07
     ог
    -0.07
     eliminar
    -0.06
    米
    -0.06
    자의
    -0.06
    朱
    -0.06
    _connected
    -0.06
    热
    -0.06
    POSITIVE LOGITS
    Crear
    0.06
     слож
    0.06
    .Menu
    0.06
     CRS
    0.06
    ]*(
    0.06
    олн
    0.06
     getSession
    0.06
    ้า
    0.06
    /login
    0.06
     prat
    0.06
    Act Density 0.777%

    No Known Activations