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

    Neuronpedia

    INDEX
    Explanations

    questions and punctuation

    np_max-act · gemini-2.0-flash

    The neuron fires on tokens that introduce or mark a user’s question—i.e. the “first question to begin:” phrasing and the question-mark punctuation.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     carnival
    -0.07
     Handles
    -0.06
    udents
    -0.06
    -rate
    -0.06
     gated
    -0.06
     pupil
    -0.06
     excitement
    -0.06
    level
    -0.06
    وده
    -0.06
     graduates
    -0.06
    POSITIVE LOGITS
     české
    0.07
     opak
    0.07
    .ReadToEnd
    0.07
     europ
    0.07
    (hex
    0.06
    _VIRTUAL
    0.06
    ",$
    0.06
    meldung
    0.06
     refund
    0.06
     conna
    0.06
    Act Density 0.019%

    No Known Activations