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

    Neuronpedia

    INDEX
    Explanations

    numeric values in the format "x.x", likely related to finance or technical data

    oai_token-act-pair · gpt-3.5-turbo

    instances of the number seven

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     lett
    -0.69
    ificate
    -0.64
     medicine
    -0.62
     rule
    -0.59
     tarn
    -0.58
     safety
    -0.58
     mutually
    -0.58
     utter
    -0.57
     decor
    -0.57
     tell
    -0.57
    POSITIVE LOGITS
    7
    3.17
    8
    2.43
    6
    2.39
    9
    2.37
    5
    2.30
    4
    1.97
    3
    1.94
    747
    1.74
    2
    1.69
    733
    1.67
    Act Density 0.027%

    No Known Activations