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

    Neuronpedia

    INDEX
    Explanations

    phrases denoting actions or processes that can be done or achieved easily and efficiently

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating the absence of certain conditions or obstacles

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    raq
    -0.94
    ounce
    -0.72
     Facts
    -0.72
    aez
    -0.72
    onen
    -0.72
    hai
    -0.70
    soon
    -0.69
    late
    -0.69
    lator
    -0.69
    elf
    -0.68
    POSITIVE LOGITS
     risking
    1.44
     sacrificing
    1.43
     harming
    1.39
     needing
    1.38
     compromising
    1.36
     worrying
    1.24
     relying
    1.22
     disrupting
    1.18
     bothering
    1.17
     requiring
    1.16
    Act Density 0.050%

    No Known Activations