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

    Neuronpedia

    INDEX
    Explanations

    comparative adjectives indicating levels of costliness

    oai_token-act-pair · gpt-3.5-turbo

    comparative and superlative forms of the word "dead" and related concepts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     combined
    -0.60
     Security
    -0.60
     Def
    -0.59
     captured
    -0.58
    Shield
    -0.58
     Dawn
    -0.55
     tit
    -0.55
     dun
    -0.54
     SEC
    -0.54
     chars
    -0.54
    POSITIVE LOGITS
    lier
    5.03
    liest
    3.64
    liness
    2.26
    lihood
    1.91
    lies
    1.50
    ly
    1.28
    LY
    1.15
    lie
    1.15
    lied
    1.05
    lying
    1.02
    Act Density 0.010%

    No Known Activations