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

    Neuronpedia

    INDEX
    Explanations

    referenced page numbers within the text

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
     basket
    -0.59
    emouth
    -0.57
     Kirin
    -0.56
     drained
    -0.56
     jug
    -0.54
    anka
    -0.54
    Frames
    -0.53
    stood
    -0.53
     hugging
    -0.53
    bies
    -0.52
    POSITIVE LOGITS
    APTER
    0.81
     Conclusion
    0.79
    ¶
    0.79
     TABLE
    0.76
    CHAPTER
    0.76
     =================================================================
    0.76
    PART
    0.75
    xxxxxxxx
    0.75
    Marginal
    0.74
     Allaah
    0.73
    Act Density 2.997%

    No Known Activations