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

    Neuronpedia

    INDEX
    Explanations

    contrasts or comparisons using the word "Whereas"

    oai_token-act-pair · gpt-3.5-turbo

    phrases that contrast different situations or perspectives

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    entry
    -0.79
    anut
    -0.76
    Roll
    -0.75
    ve
    -0.72
    eat
    -0.72
    roll
    -0.71
    packed
    -0.71
    icious
    -0.70
    erion
    -0.68
    aqu
    -0.68
    POSITIVE LOGITS
    xual
    0.88
    soever
    0.82
    lihood
    0.78
     ours
    0.74
    theless
    0.70
    entimes
    0.67
    Ļ½
    0.67
     Whereas
    0.67
     ACTIONS
    0.64
     hitherto
    0.64
    Act Density 0.022%

    No Known Activations