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

    Neuronpedia

    INDEX
    Explanations

    disjointed sentences

    np_max-act · gemini-2.0-flash

    It detects user/identity and metadata tokens — handles, NAME_ placeholders, timestamps/numeric IDs and other forum/chat markup.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
     звіль
    -0.07
    Votre
    -0.07
    Vari
    -0.06
    �
    -0.06
     Profession
    -0.06
    ораль
    -0.06
    $/
    -0.06
    shapes
    -0.06
    ısıyla
    -0.06
    观看
    -0.06
    POSITIVE LOGITS
     getCurrent
    0.07
     friendly
    0.07
     burn
    0.07
     vd
    0.06
     suất
    0.06
    Iraq
    0.06
     Impro
    0.06
     automatic
    0.06
    ,-
    0.06
     menstrual
    0.06
    Act Density 0.111%

    No Known Activations