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

    Neuronpedia

    INDEX
    Explanations

    capitalized keywords or acronyms representing commands or options

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "OR."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    kered
    -0.78
     Mug
    -0.65
    asia
    -0.60
    ench
    -0.60
     Desmond
    -0.60
    cephal
    -0.59
     feudal
    -0.59
    joy
    -0.57
    zac
    -0.56
    aments
    -0.56
    POSITIVE LOGITS
     OR
    3.44
    OR
    1.65
     AND
    1.43
    ORS
    1.37
     Or
    1.33
    Or
    1.29
     ON
    1.23
     AS
    1.20
     NOR
    1.19
     IN
    1.19
    Act Density 0.011%

    No Known Activations