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

    Neuronpedia

    INDEX
    Explanations

    quotation marks

    np_max-act · gemini-2.0-flash

    tokens that occur inside or adjacent to direct speech/quotation (dialogue).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     coloured
    -0.07
     Tricks
    -0.06
     managed
    -0.06
     Watching
    -0.06
     Unternehmen
    -0.06
    �
    -0.06
    Ông
    -0.06
    ávka
    -0.06
    �
    -0.06
     newspaper
    -0.06
    POSITIVE LOGITS
    ATT
    0.06
    bons
    0.06
     prostřed
    0.06
    dT
    0.06
     зн
    0.06
    ึ้
    0.06
     опыт
    0.06
     surg
    0.06
    _THREADS
    0.06
    objc
    0.05
    Act Density 0.107%

    No Known Activations