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

    Neuronpedia

    INDEX
    Explanations

    *Contradictions* or *unexpected outcomes* in the text

    oai_token-act-pair · gpt-3.5-turbo

    instances of the phrase "even though."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    vantage
    -0.83
    IGN
    -0.71
    eed
    -0.70
     pione
    -0.68
    utical
    -0.67
    OE
    -0.67
    ilan
    -0.66
    idency
    -0.66
    isively
    -0.66
    oeuv
    -0.65
    POSITIVE LOGITS
     they
    0.92
     technically
    0.87
     it
    0.81
     admitting
    0.79
     he
    0.77
     there
    0.74
     we
    0.73
     acknowledging
    0.71
     she
    0.70
     millions
    0.70
    Act Density 0.076%

    No Known Activations