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

    Neuronpedia

    INDEX
    Explanations

    with

    np_max-act · gemini-2.0-flash

    The neuron reliably lights up on words in third-person past-tense narrative passages—that is, tokens used to describe events or actions (“he,” “was,” “attacked,” “said,” etc.) in story-telling prose.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Keith
    -0.06
    Tek
    -0.06
    ğan
    -0.06
    aka
    -0.06
     IX
    -0.06
     Semantic
    -0.06
     pan
    -0.06
     serr
    -0.06
     tỷ
    -0.06
     Sac
    -0.06
    POSITIVE LOGITS
    eselect
    0.07
     CGI
    0.06
     arises
    0.06
    _MINOR
    0.06
    ?>>↵
    0.06
     जय
    0.06
     jsou
    0.06
    /http
    0.06
     CX
    0.06
     çiz
    0.06
    Act Density 0.048%

    No Known Activations