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

    Neuronpedia

    INDEX
    Explanations

    Assistant/Bot

    np_max-act · gemini-2.0-flash

    chat role markers and references to the AI assistant’s identity within conversation transcripts.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    (Auth
    -0.07
     land
    -0.07
    /us
    -0.07
     da
    -0.07
    ighbour
    -0.06
    IDE
    -0.06
    国务院
    -0.06
     Noise
    -0.06
    ↓
    -0.06
    Forget
    -0.06
    POSITIVE LOGITS
    _TUN
    0.08
    etak
    0.07
    .'));↵
    0.07
    aceut
    0.07
    _PI
    0.07
    修剪
    0.07
    �
    0.07
     preco
    0.07
    냈
    0.07
    里面的
    0.07
    Act Density 0.014%

    No Known Activations