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

    Neuronpedia

    INDEX
    Explanations

    academic texts

    np_max-act · gemini-2.0-flash

    The neuron is triggered by non-English tokens—especially pieces of words containing diacritic marks (e.g. “ã,” “ç,” “ó”) common in Portuguese/Spanish.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ilon
    -0.07
    Fault
    -0.06
    aro
    -0.06
     сид
    -0.06
    ojis
    -0.06
     particul
    -0.06
    FUNCTION
    -0.06
     toddlers
    -0.06
    tek
    -0.06
    Mike
    -0.06
    POSITIVE LOGITS
    _tra
    0.07
     싸
    0.07
     taxable
    0.07
    _shadow
    0.06
    ่อส
    0.06
     [["
    0.06
    .newaxis
    0.06
     الذه
    0.06
    	onChange
    0.06
    �
    0.06
    Act Density 0.048%

    No Known Activations