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

    Neuronpedia

    INDEX
    Explanations

    Attribution/quotes

    np_max-act · gemini-2.0-flash

    The neuron is keyed to punctuation-style delimiters (e.g. em-dashes, section breaks, quotes) that mark structural boundaries in the text.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    waves
    -0.07
    효
    -0.06
    вала
    -0.06
    ़त
    -0.06
    ystone
    -0.06
    iseconds
    -0.06
    �
    -0.06
    ErrorCode
    -0.06
     cords
    -0.06
     piracy
    -0.06
    POSITIVE LOGITS
    ->{'
    0.07
     corrupted
    0.07
    ≦
    0.07
    receive
    0.06
     �
    0.06
    _DIRECTION
    0.06
     Origin
    0.06
     altered
    0.06
     originate
    0.06
    _NAMESPACE
    0.06
    Act Density 0.022%

    No Known Activations