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

    Neuronpedia

    INDEX
    Explanations

    Research papers

    np_max-act · gemini-2.0-flash

    The neuron detects Spanish words or stems that name research‐method actions (e.g. investigar, recopilar, analizar, experimentar).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    words and fragments related to research, data, and scientific-method vocabulary (often in Romance-language tokens).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     nowrap
    -0.06
    _signup
    -0.06
     پدر
    -0.06
     پزش
    -0.06
    	un
    -0.06
    anners
    -0.06
    iglia
    -0.06
     polynomial
    -0.06
    chemist
    -0.06
    096
    -0.06
    POSITIVE LOGITS
    ontvangst
    0.08
    fadeOut
    0.06
    :n
    0.06
     δικ
    0.06
    _skin
    0.06
     удар
    0.06
     Nicaragua
    0.06
     идет
    0.06
     downloaded
    0.06
    (predicate
    0.06
    Act Density 0.228%

    No Known Activations