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

    Neuronpedia

    INDEX
    Explanations

    Adding information

    np_max-act · gemini-2.0-flash

    This neuron detects discourse-level transition adverbs (e.g., “Moreover,” “Furthermore,” “In addition,” “Therefore”) that signal connections between sentences or clauses.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     explorer
    -0.07
     classifiers
    -0.06
    Assembly
    -0.06
     Typed
    -0.06
    104
    -0.06
     assembly
    -0.06
     signatures
    -0.06
     Foreign
    -0.06
    <Mesh
    -0.06
    cmc
    -0.06
    POSITIVE LOGITS
     является
    0.07
     lä
    0.06
    制
    0.06
    ังไม
    0.06
    ��
    0.06
    0.06
     صاح
    0.06
     البلد
    0.06
    леж
    0.06
    .require
    0.06
    Act Density 0.033%

    No Known Activations