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

    Neuronpedia

    INDEX
    Explanations

    conjunctions that introduce contrast or exception

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     wireType
    -0.50
    ouage
    -0.49
    cadora
    -0.47
    allage
    -0.45
    trine
    -0.44
    antro
    -0.43
    ondata
    -0.43
     folle
    -0.42
    GOAT
    -0.42
    REGON
    -0.42
    POSITIVE LOGITS
     but
    3.23
     pero
    3.02
     nhưng
    2.89
     BUT
    2.77
     tetapi
    2.72
    但
    2.69
    but
    2.64
     But
    2.56
     mutta
    2.56
    But
    2.55
    Act Density 0.813%

    No Known Activations