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

    Neuronpedia

    INDEX
    Explanations

    Internet/code fragments

    np_max-act · gemini-2.0-flash

    This neuron activates on hyphen‐delimited token fragments—i.e. pieces of words or URL slugs split by “-” (compound/adjective parts and URL path components).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     burn
    -0.07
    EXTERNAL
    -0.07
    ofile
    -0.07
     дир
    -0.06
     Serbia
    -0.06
     dips
    -0.06
    stick
    -0.06
     scrape
    -0.06
    	code
    -0.06
     Recent
    -0.06
    POSITIVE LOGITS
    	↵	↵↵
    0.07
    %"><
    0.07
     Iterator
    0.07
    elsea
    0.06
     pObj
    0.06
    UCT
    0.06
    rafted
    0.06
    _IDX
    0.06
    都市
    0.06
     سل
    0.06
    Act Density 0.042%

    No Known Activations