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

    Neuronpedia

    INDEX
    Explanations

    phrases indicating contradiction or contrast

    oai_token-act-pair · gpt-3.5-turbo

    adversative conjunctions indicating a contrast or exception

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    tnc
    -0.83
    usting
    -0.76
    Synopsis
    -0.75
    è¦ļéĨĴ
    -0.71
    iggurat
    -0.71
    ermanent
    -0.68
    IZ
    -0.68
    代
    -0.68
    dayName
    -0.67
    sense
    -0.67
    POSITIVE LOGITS
     declined
    1.12
     lacked
    1.00
     alas
    1.00
     hasn
    1.00
     denies
    0.97
     didn
    0.95
     retains
    0.94
     did
    0.94
     refuses
    0.94
     hadn
    0.94
    Act Density 0.168%

    No Known Activations