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 programming‐language identifier tokens—especially type or macro names made up of letters (often mixed‐case) and underscores.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    forc
    -0.07
    扶
    -0.06
    -war
    -0.06
    *u
    -0.06
    bery
    -0.06
     MMC
    -0.06
    -connect
    -0.06
    SetFont
    -0.06
    .enabled
    -0.06
     �
    -0.06
    POSITIVE LOGITS
    оби
    0.06
    変更
    0.06
     натураль
    0.06
    ^(
    0.06
    -loop
    0.06
     fondo
    0.06
    Fourth
    0.06
     пищ
    0.06
    로는
    0.06
     έν
    0.06
    Act Density 0.012%

    No Known Activations