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

    Neuronpedia

    INDEX
    Explanations

    phrases related to extreme or negative situations

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "worst" to highlight negative situations or conditions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    agine
    -0.81
    raltar
    -0.80
    sure
    -0.78
    trl
    -0.74
    dinand
    -0.73
    andise
    -0.73
    uador
    -0.72
    itialized
    -0.70
    bold
    -0.70
    leans
    -0.69
    POSITIVE LOGITS
     offender
    1.00
     imaginable
    0.96
     offenders
    0.94
     nightmare
    0.93
     behaved
    0.83
     nightmares
    0.77
     losers
    0.75
     possible
    0.75
    liest
    0.74
     loser
    0.73
    Act Density 0.022%

    No Known Activations