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

    Neuronpedia

    INDEX
    Explanations

    contrasts and comparisons using the word "Whereas"

    oai_token-act-pair · gpt-3.5-turbo

    the word "whereas" indicating comparisons or contrasts between two subjects

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    entry
    -0.84
    hal
    -0.73
    ve
    -0.71
    icious
    -0.71
    omo
    -0.71
    eyed
    -0.71
    roll
    -0.70
    aqu
    -0.68
    anut
    -0.68
    gren
    -0.67
    POSITIVE LOGITS
    xual
    1.07
    soever
    0.93
    lihood
    0.86
    wcsstore
    0.77
     Whereas
    0.77
     hitherto
    0.72
    terness
    0.71
    entimes
    0.70
     we
    0.69
    awei
    0.66
    Act Density 0.010%

    No Known Activations