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

    Neuronpedia

    INDEX
    Explanations

    period

    np_max-act · gemini-2.0-flash

    This neuron fires on sentence‐ or clause‐boundary cues—namely end‐of‐sentence punctuation (periods, quotes, commas, colons) and the very next token (often a capitalized word) marking the start of a new sentence or segment.

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

    references to popular culture and media phenomena.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     duck
    -0.07
    Tpl
    -0.07
     Địa
    -0.07
     Parsing
    -0.06
    ISE
    -0.06
     öğ
    -0.06
    _MOVE
    -0.06
    -neck
    -0.06
     Psy
    -0.06
     住
    -0.06
    POSITIVE LOGITS
    RequestMethod
    0.06
    astic
    0.06
    como
    0.06
    =session
    0.06
     unfold
    0.06
    .Environment
    0.06
     rebuild
    0.06
    \Api
    0.06
    ABB
    0.06
    qs
    0.06
    Act Density 0.114%

    No Known Activations