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

    Neuronpedia

    INDEX
    Explanations

    words related to breaking or failing

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
    esthetic
    -0.73
    isable
    -0.72
    ndra
    -0.72
    ised
    -0.71
    pora
    -0.66
    isation
    -0.66
    raviolet
    -0.64
     Pradesh
    -0.64
    velength
    -0.62
     vomit
    -0.59
    POSITIVE LOGITS
    le
    0.93
    enegger
    0.86
    buster
    0.85
     bust
    0.83
    neck
    0.82
    aneers
    0.80
    les
    0.80
    enance
    0.80
    inelli
    0.79
    aign
    0.78
    Act Density 0.059%

    No Known Activations