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

    Neuronpedia

    INDEX
    Explanations

    parentheses

    np_max-act · gemini-2.0-flash

    The neuron activates on programming‐language syntax (e.g. code keywords, punctuation, braces, semicolons), effectively spotting code snippets.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Growth
    -0.07
     віднос
    -0.06
     bedding
    -0.06
    _main
    -0.06
     buzz
    -0.06
    ecom
    -0.06
     slash
    -0.06
     XElement
    -0.06
     curry
    -0.06
    ียน
    -0.06
    POSITIVE LOGITS
    ificados
    0.08
    нього
    0.07
    @register
    0.07
     проводить
    0.07
     Ihrer
    0.07
    ified
    0.06
    .sell
    0.06
    );
↵
↵
↵
    0.06
    ification
    0.06
    irez
    0.06
    Act Density 0.016%

    No Known Activations