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

    Neuronpedia

    INDEX
    Explanations

    differences or comparisons between items

    oai_token-act-pair · gpt-3.5-turbo

    comparative phrases that indicate differences or improvements between items

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    erity
    -0.88
    ilic
    -0.75
    nton
    -0.74
    elfare
    -0.72
    illions
    -0.71
    igious
    -0.70
    services
    -0.65
    arcity
    -0.65
    efe
    -0.65
    igent
    -0.65
    POSITIVE LOGITS
     normal
    1.31
     usual
    1.28
     previous
    1.26
     regular
    1.13
     vanilla
    1.11
     originals
    1.08
     original
    1.06
     standard
    1.05
     ordinary
    1.02
    normal
    0.99
    Act Density 0.305%

    No Known Activations