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

    Neuronpedia

    INDEX
    Explanations

    expressions in text isolated by special characters (Ċ) with a higher activation for those separated by 4 instances of the special character

    oai_token-act-pair · gpt-3.5-turbo

    emphasis on emotional expressions and stylistic choices in writing

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    Ń·
    -0.65
    ĸļ
    -0.62
    ļé
    -0.61
    ensibly
    -0.57
    ¥µ
    -0.56
     neighb
    -0.55
    ãĥ¼ãĥĨ
    -0.55
     equival
    -0.54
    ylum
    -0.54
     dictated
    -0.54
    POSITIVE LOGITS
     largeDownload
    0.70
    Click
    0.69
    ccording
    0.69
    ³³³
    0.68
    PRESS
    0.68
    Lens
    0.68
    OVER
    0.68
    ³³³³
    0.68
    SELECT
    0.68
    Disclaimer
    0.68
    Act Density 0.260%

    No Known Activations