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

    Neuronpedia

    INDEX
    Explanations

    phrases indicating a contrast to commonly held beliefs or assertions

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
    ilitating
    -0.83
    lov
    -0.79
    oler
    -0.78
    killer
    -0.76
    isol
    -0.75
    urated
    -0.75
    beans
    -0.75
    aqu
    -0.74
    ubes
    -0.74
    aquin
    -0.73
    POSITIVE LOGITS
     viewpoints
    1.05
     sides
    0.89
     opinions
    0.88
     counsel
    0.81
     sexes
    0.79
     shore
    0.79
     views
    0.78
     side
    0.77
     halves
    0.77
     minded
    0.77
    Act Density 0.054%

    No Known Activations