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

    Neuronpedia

    INDEX
    Explanations

    superlatives indicating extreme negativity or crisis situations

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "worst" used to describe negative circumstances or events

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    arij
    -0.79
    arya
    -0.74
    ependence
    -0.73
    alde
    -0.73
    arist
    -0.72
    bara
    -0.70
    ulton
    -0.70
    yi
    -0.70
    uncture
    -0.70
    pher
    -0.70
    POSITIVE LOGITS
     worst
    1.03
    worst
    1.01
     Worst
    0.96
     nightmare
    0.93
     imaginable
    0.87
     offenders
    0.87
     offender
    0.83
     nightmares
    0.81
     losers
    0.80
     loser
    0.78
    Act Density 0.008%

    No Known Activations