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

    Neuronpedia

    INDEX
    Explanations

    superlatives expressing extreme negativity

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of "worst" or negative experiences

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ynthesis
    -0.87
    ools
    -0.86
    andise
    -0.86
    agine
    -0.79
    uador
    -0.79
    sure
    -0.79
    itialized
    -0.78
    thouse
    -0.78
    arya
    -0.76
    icles
    -0.76
    POSITIVE LOGITS
     offenders
    1.03
     offender
    1.02
     nightmare
    0.93
     nightmares
    0.90
     behaved
    0.80
     plag
    0.78
     catast
    0.77
     atrocities
    0.76
     imaginable
    0.75
     plague
    0.75
    Act Density 0.042%

    No Known Activations