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

    Neuronpedia

    INDEX
    Explanations

    imperfections, limitations

    np_max-act · gemini-2.0-flash

    This neuron detects words signaling negation or limitation (e.g. “not,” “never,” “imperfect,” “inaccurate,” “always” in a limiting sense).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    	cc
    -0.07
    *cos
    -0.07
     Ctrl
    -0.07
    GameManager
    -0.06
    Universal
    -0.06
    .step
    -0.06
    GraphQL
    -0.06
    inputs
    -0.06
     кури
    -0.06
     oneself
    -0.06
    POSITIVE LOGITS
    AK
    0.08
    mie
    0.07
    ARK
    0.07
     casualty
    0.06
     HERO
    0.06
    _TITLE
    0.06
     छ
    0.06
     AK
    0.06
     gider
    0.06
    !,↵
    0.06
    Act Density 0.086%

    No Known Activations