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

    Neuronpedia

    INDEX
    Explanations

    Code-related attributes/configurations

    np_max-act · gemini-2.0-flash

    This neuron detects identifiers used as configuration or attribute keys (e.g. HTML “class”, “style”, API or config option names) in code.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    .imwrite
    -0.07
    .breakpoints
    -0.07
    oise
    -0.07
    っている
    -0.07
    .articles
    -0.07
    (net
    -0.07
    StringRef
    -0.07
     ++$
    -0.07
    ERICA
    -0.06
    :^(
    -0.06
    POSITIVE LOGITS
    ORG
    0.07
     sog
    0.07
     Correction
    0.06
     soit
    0.06
     RA
    0.06
     descendants
    0.06
     magna
    0.06
    .events
    0.06
     parcel
    0.06
    .Te
    0.06
    Act Density 0.026%

    No Known Activations