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

    Neuronpedia

    INDEX
    Explanations

    text indicating a direct link or connection

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "directly."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    gerald
    -0.89
    glers
    -0.82
    ulton
    -0.72
     Garry
    -0.70
    ifully
    -0.67
     Kirin
    -0.67
    lis
    -0.65
    ĸļ
    -0.63
     Daily
    -0.62
     Turtle
    -0.62
    POSITIVE LOGITS
     contradicted
    0.89
     identifiable
    0.81
     contradicts
    0.80
    ebted
    0.74
    forward
    0.73
     contradict
    0.72
     addressed
    0.71
    obin
    0.70
     direct
    0.70
    cedented
    0.70
    Act Density 0.016%

    No Known Activations