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

    Neuronpedia

    INDEX
    Explanations

    details and specific visual descriptions in texts

    oai_token-act-pair · gpt-3.5-turbo

    references to a high level of detail or intricacy in descriptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    onz
    -0.78
    ammers
    -0.76
    nder
    -0.74
    orea
    -0.70
    anski
    -0.70
    mire
    -0.68
    aden
    -0.65
    reath
    -0.65
    emouth
    -0.65
    si
    -0.64
    POSITIVE LOGITS
     detail
    1.17
     details
    0.96
     Detail
    0.87
    iveness
    0.76
    ariat
    0.73
     marks
    0.73
     Details
    0.71
     particulars
    0.70
    displayText
    0.70
     settings
    0.68
    Act Density 0.013%

    No Known Activations