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

    Neuronpedia

    INDEX
    Explanations

    sentences that include periods, signaling complete thoughts or statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    technical and scientific text

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron detects sentence endings, strongly activating on sentence-final punctuation (periods) and the ends of sentences.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     Administrativna
    -0.88
     zwiſchen
    -0.86
    ロウィン
    -0.86
     autorytatywna
    -0.85
     Taktlose
    -0.84
    -0.82
    <unused41>
    -0.81
     パンチラ
    -0.81
    uxxxx
    -0.81
    <unused52>
    -0.81
    POSITIVE LOGITS
     nôtre
    0.48
    .
    0.48
    The
    0.47
     It
    0.46
     duquel
    0.44
    No
    0.44
    because
    0.43
     derfor
    0.43
     However
    0.42
     If
    0.42
    Act Density 2.054%

    No Known Activations