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

    Neuronpedia

    INDEX
    Explanations

    age descriptions, particularly the phrase "old"

    oai_token-act-pair · gpt-3.5-turbo

    references to ages in a specific format

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    anwhile
    -0.84
    akespe
    -0.77
    hod
    -0.76
    oldown
    -0.75
     destro
    -0.74
    ullivan
    -0.74
    ainer
    -0.73
    acly
    -0.72
    antha
    -0.72
    ipedia
    -0.71
    POSITIVE LOGITS
     boy
    0.89
     girl
    0.72
    ish
    0.72
     Tribune
    0.72
     Jah
    0.71
     Boy
    0.71
     Frenchman
    0.71
     Tav
    0.66
     York
    0.66
     Indonesian
    0.65
    Act Density 0.029%

    No Known Activations