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

    Neuronpedia

    INDEX
    Explanations

    instances where the document mentions initiating dialogue or discussions

    oai_token-act-pair · gpt-3.5-turbo

    words and phrases indicating intention or purpose

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    _.
    -0.58
     disadvant
    -0.57
     horizont
    -0.57
    anooga
    -0.56
    listed
    -0.56
     nomine
    -0.55
     costing
    -0.53
     netted
    -0.52
    Required
    -0.52
     lines
    -0.52
    POSITIVE LOGITS
    ggles
    0.97
    asty
    0.92
     avoid
    0.91
     celebrate
    0.89
     satisfy
    0.87
     protect
    0.87
     ensure
    0.86
     relieve
    0.86
     accompany
    0.85
     maximize
    0.85
    Act Density 0.409%

    No Known Activations