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

    Neuronpedia

    INDEX
    Explanations

    a

    np_max-act · gemini-2.0-flash

    The neuron detects mentions of “expects”-style type-mismatch error messages (e.g. “expects a string, not a control”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    combination
    -0.07
    کش
    -0.06
    奪
    -0.06
     Uncomment
    -0.06
     waar
    -0.06
     Calcium
    -0.06
     Tüm
    -0.06
    _off
    -0.06
    .localized
    -0.06
     hukuk
    -0.06
    POSITIVE LOGITS
     saver
    0.07
    	puts
    0.06
     Choose
    0.06
    /screens
    0.06
     singers
    0.06
    _CASE
    0.06
    _pass
    0.06
    inden
    0.06
     conten
    0.06
    /");↵
    0.06
    Act Density 0.050%

    No Known Activations