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

    Neuronpedia

    INDEX
    Explanations

    Code

    np_max-act · gemini-2.0-flash

    The neuron fires on code‐style tokens—i.e. programming‐language syntax (like braces, keywords, and backticks) rather than ordinary prose.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     Zero
    -0.07
    uela
    -0.07
     DeV
    -0.07
     Saudi
    -0.06
     đoạn
    -0.06
     Matches
    -0.06
    oplan
    -0.06
     sexy
    -0.06
     numero
    -0.06
    152
    -0.06
    POSITIVE LOGITS
    게시
    0.07
    ReceiveMemoryWarning
    0.06
     rum
    0.06
     остан
    0.06
     -:
    0.05
     silicone
    0.05
    ='{$
    0.05
    Nota
    0.05
    은
    0.05
    _stub
    0.05
    Act Density 0.039%

    No Known Activations