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

    Neuronpedia

    INDEX
    Explanations

    instances where contrasting elements are being compared

    oai_token-act-pair · gpt-3.5-turbo

    the phrase "while", indicating a contrast or a conditional statement

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    aer
    -0.81
    agin
    -0.76
    atron
    -0.74
    omet
    -0.70
    ahime
    -0.70
    red
    -0.70
    aja
    -0.69
    ison
    -0.69
    icted
    -0.68
    fleet
    -0.68
    POSITIVE LOGITS
     acknowledging
    1.06
     researching
    0.97
     browsing
    0.92
     respecting
    0.91
     conced
    0.90
     touring
    0.85
     agreeing
    0.85
     maintaining
    0.85
     admitting
    0.85
     compiling
    0.78
    Act Density 0.043%

    No Known Activations