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

    Neuronpedia

    INDEX
    Explanations

    dates or time-related words, especially involving the day before the current one

    oai_token-act-pair · gpt-3.5-turbo

    references to the word "yesterday."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    eers
    -0.94
    eer
    -0.84
     Conquer
    -0.82
    OHN
    -0.79
    iframe
    -0.73
    egal
    -0.69
    Control
    -0.68
    lasses
    -0.67
    hips
    -0.67
    ordes
    -0.65
    POSITIVE LOGITS
     afternoon
    1.61
     evening
    1.45
     morning
    1.45
     night
    1.26
     mornings
    1.09
    morning
    1.03
     Evening
    0.98
     Night
    0.94
    night
    0.92
     Morning
    0.86
    Act Density 0.034%

    No Known Activations