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

    Neuronpedia

    INDEX
    Explanations

    ion

    np_max-act · gemini-2.0-flash

    This neuron detects uncommon capitalized tokens that are part of named entities (e.g., company names, place names, drug or product names, author names).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    _refer
    -0.07
    chemas
    -0.07
     winter
    -0.07
     Scotch
    -0.07
     skating
    -0.07
     curtain
    -0.06
     Arabia
    -0.06
    -0.06
     density
    -0.06
    ष
    -0.06
    POSITIVE LOGITS
    	↵		↵
    0.07
     JMP
    0.06
     bam
    0.06
    شنبه
    0.06
     veel
    0.06
    tab
    0.06
    ##↵
    0.06
     ++$
    0.06
     بلند
    0.06
    。”↵↵
    0.06
    Act Density 0.323%

    No Known Activations