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

    Neuronpedia

    INDEX
    Explanations

    non-English characters

    np_max-act · gemini-2.0-flash

    This neuron activates on non-ASCII accented characters (i.e. letters with diacritical marks).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Pent
    -0.09
     fart
    -0.07
     prefab
    -0.07
    приєм
    -0.07
     cort
    -0.07
    	Get
    -0.07
    ueling
    -0.06
    DT
    -0.06
     filtr
    -0.06
    059
    -0.06
    POSITIVE LOGITS
     Musk
    0.07
     �
    0.07
     se
    0.06
     Monster
    0.06
    ятся
    0.06
    _logging
    0.06
     쪽
    0.06
    gré
    0.06
    chure
    0.06
     Draws
    0.06
    Act Density 0.004%

    No Known Activations