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

    Neuronpedia

    INDEX
    Explanations

    quantitative expressions indicating a portion or fraction, such as "half of"

    oai_token-act-pair · gpt-3.5-turbo

    the phrase "more than half" or variations of it

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ocr
    -0.65
    ffen
    -0.62
    berus
    -0.55
    igslist
    -0.55
    andr
    -0.54
     arrang
    -0.52
    spir
    -0.52
    orge
    -0.51
    hran
    -0.51
    kson
    -0.51
    POSITIVE LOGITS
    azo
    0.65
     of
    0.64
    century
    0.64
     dozen
    0.64
    wheel
    0.62
    rene
    0.61
     century
    0.61
    terness
    0.61
    atos
    0.60
     million
    0.59
    Act Density 0.052%

    No Known Activations