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

    Neuronpedia

    INDEX
    Explanations

    numerical values associated with statistics or mathematical calculations

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
     ende
    -0.88
     eleph
    -0.86
     obser
    -0.83
     dilig
    -0.77
    swer
    -0.74
    ñ
    -0.72
     accomp
    -0.72
     Hoff
    -0.70
     Bere
    -0.70
     observ
    -0.69
    POSITIVE LOGITS
    %
    2.00
    %"
    1.56
    %,
    1.53
    %]
    1.52
    %)
    1.46
    %:
    1.44
    %-
    1.40
    %;
    1.40
    %.
    1.38
     percent
    1.28
    Act Density 0.055%

    No Known Activations