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

    Neuronpedia

    INDEX
    Explanations

    >

    np_max-act · gemini-2.0-flash

    This neuron selectively activates on HTML/XML markup tags (the angle-bracketed elements like `<head>`, `</body>`, `</html>`, etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     które
    -0.08
     Surely
    -0.07
    _SELECT
    -0.07
     dream
    -0.06
     JA
    -0.06
    467
    -0.06
    dream
    -0.06
     Role
    -0.06
     aggressive
    -0.06
     surely
    -0.06
    POSITIVE LOGITS
     temperature
    0.07
    .waitFor
    0.07
    .per
    0.06
    cosity
    0.06
     pores
    0.06
     infinitely
    0.06
    centration
    0.06
     Attribution
    0.06
    /tcp
    0.06
    Positions
    0.06
    Act Density 0.008%

    No Known Activations