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

    Neuronpedia

    INDEX
    Explanations

    phrases expressing a contrast or contradiction

    oai_token-act-pair · gpt-3.5-turbo

    the conjunction "but" indicating contrasting statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     pione
    -0.91
    dayName
    -0.79
    umat
    -0.75
    iggurat
    -0.74
    ā
    -0.74
    ivered
    -0.74
    Ě
    -0.74
    ē
    -0.73
    umbn
    -0.73
    ö
    -0.73
    POSITIVE LOGITS
     alas
    1.08
     nevertheless
    1.06
     nonetheless
    0.99
     beware
    0.98
     it
    0.94
     unfortunately
    0.93
     unless
    0.90
     nowhere
    0.88
     surely
    0.88
     why
    0.87
    Act Density 0.183%

    No Known Activations