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

    Neuronpedia

    INDEX
    Explanations

    the word "like" followed by different scenarios or preferences expressed in the text

    oai_token-act-pair · gpt-3.5-turbo

    expressions of dislike or aversion

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    iggins
    -0.84
    rontal
    -0.83
    utical
    -0.81
    yrinth
    -0.81
    itech
    -0.79
    owa
    -0.76
    nown
    -0.75
    verty
    -0.74
    oided
    -0.73
    olon
    -0.73
    POSITIVE LOGITS
     anymore
    0.96
     anybody
    0.88
     nor
    0.87
     anything
    0.80
     anyone
    0.79
     any
    0.76
    ably
    0.75
     surprises
    0.71
     censorship
    0.67
     bullies
    0.66
    Act Density 0.048%

    No Known Activations