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

    Neuronpedia

    INDEX
    Explanations

    phrases containing contrasting or opposing information

    oai_token-act-pair · gpt-3.5-turbo

    the word "but" indicating contrasting situations or exceptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    mast
    -0.71
    Times
    -0.69
    cloth
    -0.69
    obyl
    -0.67
    ige
    -0.67
    analysis
    -0.67
    ampa
    -0.66
    tnc
    -0.65
    oin
    -0.64
    built
    -0.64
    POSITIVE LOGITS
     luckily
    0.94
     fortunately
    0.92
     nevertheless
    0.90
     alas
    0.87
    ersen
    0.81
     nonetheless
    0.79
    tons
    0.79
    chery
    0.77
     thankfully
    0.77
     ultimately
    0.77
    Act Density 0.203%

    No Known Activations