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

    Neuronpedia

    INDEX
    Explanations

    errors or inaccuracies marked with "sic" in the text

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "sic," often used to indicate a direct quote or to signal that the preceding text should be attributed as it was found

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    VIEW
    -0.79
    MQ
    -0.78
    eus
    -0.76
    Winged
    -0.71
    Arm
    -0.69
    ships
    -0.69
    EVA
    -0.66
    foot
    -0.64
    Joy
    -0.63
    ĪĴ
    -0.62
    POSITIVE LOGITS
    sic
    0.82
     offend
    0.73
    querade
    0.71
    rano
    0.69
    aska
    0.66
    otics
    0.66
    pecially
    0.66
    atin
    0.66
    odynamic
    0.65
    orship
    0.65
    Act Density 0.012%

    No Known Activations