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

    Neuronpedia

    INDEX
    Explanations

    speech

    np_max-act · gemini-2.0-flash

    The neuron detects mentions of part-of-speech tagging—i.e. tokens like “speech” in “part-of-speech” and the abbreviation “POS.”

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ochastic
    -0.06
    .toolbar
    -0.06
     cakes
    -0.06
    .prop
    -0.06
     ViewPager
    -0.06
     Hyp
    -0.06
    appendTo
    -0.06
    印
    -0.06
     ubyt
    -0.06
    asString
    -0.06
    POSITIVE LOGITS
    ?></
    0.06
    ja
    0.06
     Dungeon
    0.06
    state
    0.06
    ジ
    0.06
    par
    0.06
    sub
    0.06
     coorden
    0.06
    unter
    0.06
    	se
    0.06
    Act Density 0.003%

    No Known Activations