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

    Neuronpedia

    INDEX
    Explanations

    code

    np_max-act · gemini-2.0-flash

    The neuron is looking for Unicode escape sequences (e.g. “\uXXXX”) in string literals.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     wit
    -0.07
     bust
    -0.06
    散
    -0.06
    離
    -0.06
    rated
    -0.06
    istance
    -0.06
     pension
    -0.06
    .join
    -0.06
    _args
    -0.06
     barely
    -0.06
    POSITIVE LOGITS
     σαν
    0.07
     Fakültesi
    0.07
     historically
    0.07
    ье
    0.06
    czas
    0.06
    ές
    0.06
    .AP
    0.06
     "__
    0.06
     DESIGN
    0.06
     millennia
    0.06
    Act Density 0.004%

    No Known Activations