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

    Neuronpedia

    INDEX
    Explanations

    ]

    np_max-act · gemini-2.0-flash

    This neuron detects image‐placeholder tags (e.g. “[IMAGE]”) and other layout/formatting markers in the text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    국
    -0.07
    central
    -0.07
    工程
    -0.07
     describes
    -0.07
     Zum
    -0.06
    .LEFT
    -0.06
    .Blue
    -0.06
    asper
    -0.06
     magnet
    -0.06
     transistor
    -0.06
    POSITIVE LOGITS
    Change
    0.06
     삼
    0.06
    �
    0.06
     *)↵↵
    0.06
    х
    0.06
    imas
    0.06
     чай
    0.06
     tranh
    0.06
     strany
    0.06
    ию
    0.06
    Act Density 0.003%

    No Known Activations