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

    Neuronpedia

    INDEX
    Explanations

    economy and markets

    np_max-act · gemini-2.0-flash

    This neuron detects financial-news event phrases—short items like earnings surprises, analyst upgrades, insider buying, economic recovery, etc., especially when listed as events related to stock price movements.

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31

    positive sentiment Method used: 2

    np_max-act · gpt-5-miniTriggered by @yooniel31

    This neuron responds to list enumeration markers, in particular digits (and their trailing punctuation) that denote numbered list items.

    oai_token-act-pair · o4-miniTriggered by @yooniel31

    pipe-separated sentiment labels

    np_max-act · gpt-5Triggered by @yooniel31
    New Auto-Interp
    Negative Logits
     gre
    -0.07
    δί
    -0.06
    ancies
    -0.06
     humour
    -0.06
    arker
    -0.06
     زبان
    -0.06
     solicit
    -0.06
    ازی
    -0.06
    Calling
    -0.06
     veri
    -0.06
    POSITIVE LOGITS
    _CONV
    0.07
     Θεσσα
    0.07
     assemblies
    0.06
     přest
    0.06
    .locals
    0.06
    Claim
    0.06
    .StackTrace
    0.06
    '''
    0.06
    SOEVER
    0.06
    (tok
    0.06
    Act Density 0.444%

    No Known Activations