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

    Neuronpedia

    INDEX
    Explanations

    instances where numbers have notably increased or multiplied

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
     Advice
    -1.00
    league
    -0.93
    rolled
    -0.91
    ãĤĬ
    -0.91
    cheat
    -0.88
    lev
    -0.82
    ndum
    -0.81
    ī
    -0.80
     Nationals
    -0.80
     sar
    -0.80
    POSITIVE LOGITS
    ĸļ
    1.13
    elsius
    0.97
     exponentially
    0.94
    xual
    0.93
    00007
    0.93
    uates
    0.89
     consecut
    0.88
     acceler
    0.87
    uate
    0.86
    frog
    0.84
    Act Density 0.611%

    No Known Activations