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

    Neuronpedia

    INDEX
    Explanations

    Maintaining or limiting something

    np_max-act · gemini-2.0-flash

    The neuron fires specifically on numeric tokens representing floating‐point decimals (e.g. “2.2421875”, “0.98193359375”, etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    **,
    -0.06
    /";↵↵
    -0.06
    Integral
    -0.06
    capacity
    -0.06
    raně
    -0.06
     Riyadh
    -0.06
    Iron
    -0.06
    .*(
    -0.06
     использовани
    -0.06
    TOTYPE
    -0.06
    POSITIVE LOGITS
     koup
    0.07
     Nu
    0.06
     内
    0.06
    Canvas
    0.06
    ุก
    0.06
     Hist
    0.06
     risult
    0.06
     губер
    0.06
     USART
    0.06
    писание
    0.06
    Act Density 0.016%

    No Known Activations