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

    Neuronpedia

    INDEX
    Explanations

    colon

    np_max-act · gemini-2.0-flash

    This neuron detects the numeric vote‐score token (e.g. “5.921875”) that appears right after the “Q:” header in each question.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     glyphs
    -0.07
     strong
    -0.06
     bacter
    -0.06
    egin
    -0.06
     oppressed
    -0.06
    .uk
    -0.06
     Al
    -0.06
    .helpers
    -0.06
     trận
    -0.06
     cytok
    -0.06
    POSITIVE LOGITS
     Počet
    0.07
    订单
    0.06
    IFIED
    0.06
    _WEAPON
    0.06
    ウス
    0.06
     IPL
    0.06
     pomp
    0.06
    _PLAY
    0.06
     EXT
    0.06
    +l
    0.06
    Act Density 0.025%

    No Known Activations