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

    Neuronpedia

    INDEX
    Explanations

    Document snippets

    np_max-act · gemini-2.0-flash

    This neuron detects system or user instructions that define or assign an AI model’s role or capabilities (e.g. “You are about to immerse yourself into the role of another AI model known as…”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Polar
    -0.07
    eties
    -0.06
     Měst
    -0.06
     polar
    -0.06
    Side
    -0.06
    crets
    -0.06
    Agency
    -0.06
     Day
    -0.06
    frame
    -0.06
     해
    -0.06
    POSITIVE LOGITS
     التع
    0.07
    .ribbon
    0.07
    ImageUrl
    0.07
     dashed
    0.06
     Arithmetic
    0.06
    .enumer
    0.06
    ((&___
    0.06
     Serialized
    0.06
     subcontract
    0.06
     групи
    0.06
    Act Density 0.004%

    No Known Activations