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

    Neuronpedia

    INDEX
    Explanations

    contrasts or contradictions within text

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "However" that indicate contrasting information or a change in perspective

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    gio
    -0.59
    "},"
    -0.58
    kamp
    -0.56
     âĢİ
    -0.55
    rounder
    -0.54
    prus
    -0.54
    chairs
    -0.53
     actionGroup
    -0.52
    letter
    -0.51
    iard
    -0.50
    POSITIVE LOGITS
    ,
    0.87
    ,.
    0.86
    .,
    0.83
    tons
    0.71
    oldown
    0.67
     importantly
    0.66
    ,,
    0.66
    tif
    0.62
    terday
    0.61
    ,-
    0.61
    Act Density 0.057%

    No Known Activations