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

    Neuronpedia

    INDEX
    Explanations

    ING

    np_max-act · gemini-2.0-flash

    This neuron detects document-level headings, titles and other structural/metadata elements (table-of-contents and section headings).

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31
    New Auto-Interp
    Negative Logits
    .↵
    -0.07
     vile
    -0.06
     anv
    -0.06
    χρι
    -0.06
     genu
    -0.06
    ".↵
    -0.06
    ."↵
    -0.06
     solidity
    -0.06
    .↵↵
    -0.06
    	yield
    -0.06
    POSITIVE LOGITS
     TCHAR
    0.07
     трь
    0.07
    /apis
    0.06
    _rooms
    0.06
    γκο
    0.06
    شور
    0.06
    eya
    0.06
     Bazı
    0.06
     ileri
    0.06
     služeb
    0.06
    Act Density 41.574%

    No Known Activations