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

    Neuronpedia

    INDEX
    Explanations

    phrases indicating a level of difficulty or timeliness

    oai_token-act-pair · gpt-3.5-turbo

    instances of the phrase "not too" indicating a qualitative assessment

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    igi
    -0.77
    swick
    -0.76
    hyde
    -0.72
    ERT
    -0.69
    ngth
    -0.68
    icipated
    -0.67
    intosh
    -0.65
    çļ
    -0.65
    Rus
    -0.64
    ridor
    -0.63
    POSITIVE LOGITS
     much
    0.91
     fancy
    0.87
     flashy
    0.83
     bothered
    0.80
     noticeable
    0.77
     pleasant
    0.76
     surprising
    0.76
     forgiving
    0.75
     distracting
    0.75
     bother
    0.75
    Act Density 0.030%

    No Known Activations