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

    Neuronpedia

    INDEX
    Explanations

    publishing and writing

    np_max-act · gemini-2.0-flash

    This neuron responds to multi-part tokens—especially hyphenated or compound terms and proper names (e.g. “copy-proof,” “SPi Global,” “BellMetrix”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Environmental
    -0.07
    osaic
    -0.07
    Mur
    -0.07
    (child
    -0.07
    _miss
    -0.07
     اد
    -0.07
     UIAlertController
    -0.06
     tiến
    -0.06
    ных
    -0.06
     GW
    -0.06
    POSITIVE LOGITS
    dcc
    0.06
    โอ
    0.06
     مساحت
    0.06
     Added
    0.06
     GetHashCode
    0.06
    ิทธ
    0.06
     tack
    0.06
    ングル
    0.06
     Hệ
    0.06
    _external
    0.06
    Act Density 0.460%

    No Known Activations