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

    Neuronpedia

    INDEX
    Explanations

    words related to ultra or extreme qualities

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to extreme qualities or levels

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     Versus
    -0.75
     slate
    -0.74
     Origins
    -0.74
     Barron
    -0.72
     accuser
    -0.71
     afterwards
    -0.69
     history
    -0.69
     remainder
    -0.69
     Panc
    -0.68
     Gle
    -0.68
    POSITIVE LOGITS
    expensive
    1.60
    sensitive
    1.56
    efficient
    1.47
    powerful
    1.45
    sized
    1.43
    competitive
    1.43
    exclusive
    1.40
    strength
    1.40
    violent
    1.38
    popular
    1.37
    Act Density 0.025%

    No Known Activations