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

    Neuronpedia

    INDEX
    Explanations

    aides

    np_max-act · gemini-2.0-flash

    The neuron fires on JSON document structure tokens—especially the “body” field label (and related key syntax) in JSON-formatted email outputs.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    pw
    -0.06
     cav
    -0.06
    ぁ
    -0.06
     lập
    -0.06
     tuyển
    -0.06
     cosmetics
    -0.06
    _GET
    -0.06
     nord
    -0.06
     bulunur
    -0.06
     Mehr
    -0.06
    POSITIVE LOGITS
     deception
    0.07
     eq
    0.06
    .cleanup
    0.06
     �
    0.06
     bunu
    0.06
    .dictionary
    0.06
    _entropy
    0.06
    "),
    0.06
     повыш
    0.06
     onFinish
    0.06
    Act Density 0.088%

    No Known Activations