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

    Neuronpedia

    INDEX
    Explanations

    circle

    np_max-act · gemini-2.0-flash

    The neuron detects the special control and protocol tokens used to structure the conversation (e.g. the loop-of-Thought/Action/Observation markers and header delimiters).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    39
    -0.07
    ٢
    -0.07
    Slider
    -0.06
    mpar
    -0.06
     monopoly
    -0.06
    ۱۲
    -0.06
     idle
    -0.06
    сен
    -0.06
    LL
    -0.06
     DispatchQueue
    -0.06
    POSITIVE LOGITS
     Shi
    0.07
     فت
    0.06
     vý
    0.06
     Anniversary
    0.06
    ัวร
    0.06
    อนท
    0.06
    (PyObject
    0.06
    歲
    0.06
     виды
    0.06
     Elvis
    0.05
    Act Density 0.015%

    No Known Activations