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

    Neuronpedia

    INDEX
    Explanations

    food items that are typically consumed as part of a meal

    oai_token-act-pair · gpt-3.5-turbo

    mentions of soup and related food items

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    inel
    -0.74
    sburg
    -0.70
    rium
    -0.69
    uate
    -0.67
    ussen
    -0.64
    adden
    -0.61
     setbacks
    -0.59
    rics
    -0.59
     unforeseen
    -0.59
     midterm
    -0.58
    POSITIVE LOGITS
     soup
    1.20
     Soup
    1.13
    bowl
    1.04
     kitchens
    0.91
    ç
    0.91
     nood
    0.86
    ée
    0.84
     broth
    0.83
    velt
    0.82
     bowls
    0.79
    Act Density 0.013%

    No Known Activations