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

    Neuronpedia

    INDEX
    Explanations

    000

    np_max-act · gemini-2.0-flash

    This neuron activates on numeric identifier tokens (like “000” in section or reference tags), i.e. it detects document metadata numbers.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    /prom
    -0.07
     grill
    -0.06
    odb
    -0.06
    iator
    -0.06
    avi
    -0.06
    CTX
    -0.06
    丸
    -0.06
     रहत
    -0.06
    .INVISIBLE
    -0.06
     cổ
    -0.06
    POSITIVE LOGITS
    .childNodes
    0.07
    Contained
    0.07
    ，我们
    0.07
    pack
    0.06
     ending
    0.06
     paths
    0.06
     thorough
    0.06
    ились
    0.06
     Türkiye
    0.06
     ¥
    0.06
    Act Density 0.004%

    No Known Activations