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

    Neuronpedia

    INDEX
    Explanations

    Text passages/excerpts

    np_max-act · gemini-2.0-flash

    This neuron activates on the literal curly‐brace characters (i.e. the “{” or “}” tokens) that delimit user annotations in braces.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    大
    -0.06
     tez
    -0.06
    GLOSS
    -0.06
    *T
    -0.06
     Christ
    -0.06
     avenues
    -0.06
    _loss
    -0.06
    ",↵↵
    -0.05
    $$$
    -0.05
    etCode
    -0.05
    POSITIVE LOGITS
    .j
    0.08
    .Sql
    0.07
     (--
    0.07
     chinese
    0.07
     racket
    0.07
    نة
    0.07
    ности
    0.07
     Vincent
    0.06
    .mybatis
    0.06
     áll
    0.06
    Act Density 0.148%

    No Known Activations