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

    Neuronpedia

    INDEX
    Explanations

    phrases containing the words "stay away from"

    oai_token-act-pair · gpt-3.5-turbo

    phrases instructing to avoid or stay away from something

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     teasp
    -0.81
    vic
    -0.79
    ©¶æ¥µ
    -0.79
    ounce
    -0.74
    ccording
    -0.72
    pel
    -0.72
    laughter
    -0.71
    CHR
    -0.71
    KC
    -0.71
    phabet
    -0.71
    POSITIVE LOGITS
     afar
    0.89
     thence
    0.75
     infancy
    0.74
     recess
    0.74
     anywhere
    0.72
     Hogwarts
    0.71
     whence
    0.71
     wherever
    0.65
     combust
    0.65
     ledge
    0.65
    Act Density 0.058%

    No Known Activations