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

    Neuronpedia

    INDEX
    Explanations

    digits

    np_max-act · gemini-2.0-flash

    This neuron detects date and time tokens (numeric date parts, timestamps, and time-zone/time-of-day markers).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     lez
    -0.07
     Huffman
    -0.06
     Decoration
    -0.06
     UTC
    -0.06
     thirteen
    -0.06
    istrict
    -0.06
     Ton
    -0.06
     #'
    -0.06
     Mug
    -0.06
     tur
    -0.06
    POSITIVE LOGITS
    	TRACE
    0.07
    то
    0.07
    intro
    0.06
     planned
    0.06
    copies
    0.06
    bage
    0.06
     varsa
    0.06
    exit
    0.06
    	field
    0.06
    ımızı
    0.06
    Act Density 0.017%

    No Known Activations