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

    Neuronpedia

    INDEX
    Explanations

    phrases related to being out of place or out of proportion

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating something is out of context or out of place

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     ALSO
    -0.64
    erity
    -0.64
    dump
    -0.60
    tsky
    -0.60
    til
    -0.58
    uli
    -0.57
     Vengeance
    -0.56
     renamed
    -0.56
     silently
    -0.56
     Morning
    -0.56
    POSITIVE LOGITS
    vable
    0.86
    sighted
    0.79
    whelming
    0.76
     unbeliev
    0.74
    icative
    0.74
    adequ
    0.74
    athetic
    0.72
    awed
    0.71
    ibaba
    0.70
    pport
    0.70
    Act Density 0.181%

    No Known Activations