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

    Neuronpedia

    INDEX
    Explanations

    phrases or sentences indicating something is useful or beneficial to people

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "helpful" and related vocabulary emphasizing assistance or support

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     Dare
    -0.73
     Bam
    -0.73
    Vision
    -0.70
    Metro
    -0.70
    é¾
    -0.68
    origin
    -0.68
    thur
    -0.67
    anyon
    -0.65
    jong
    -0.65
    onde
    -0.65
    POSITIVE LOGITS
     aid
    0.82
     helpful
    0.78
    tip
    0.77
     aide
    0.77
     aids
    0.75
    glers
    0.75
    Desk
    0.74
     introdu
    0.69
     advice
    0.69
     tips
    0.67
    Act Density 0.020%

    No Known Activations