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

    Neuronpedia

    INDEX
    Explanations

    friend

    np_max-act · gemini-2.0-flash

    This neuron responds to speaker-label tokens or role tags (e.g. NAME_, user/assistant headers) that mark who is speaking.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ТО
    -0.07
     orthodox
    -0.07
     stk
    -0.06
    |array
    -0.06
     Bent
    -0.06
    ське
    -0.06
     enviar
    -0.06
     Лу
    -0.06
     seeker
    -0.06
     PROC
    -0.06
    POSITIVE LOGITS
     amend
    0.07
    픽
    0.06
    fuck
    0.06
     Launch
    0.06
    Labels
    0.06
     elev
    0.06
    _until
    0.06
    �
    0.06
    ıda
    0.06
    Bird
    0.06
    Act Density 0.048%

    No Known Activations