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

    Neuronpedia

    INDEX
    Explanations

    phrases related to actions or concepts that can be done without needing some form of external assistance or intervention

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating actions or processes that can be achieved without external interference

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    raq
    -0.89
    mun
    -0.78
    geist
    -0.74
    late
    -0.74
    oled
    -0.72
    =-=-=-=-=-=-=-=-
    -0.69
    ahime
    -0.69
     Facts
    -0.68
    lator
    -0.67
    ulum
    -0.65
    POSITIVE LOGITS
     sacrificing
    1.35
     risking
    1.33
     compromising
    1.28
     harming
    1.22
     needing
    1.20
     relying
    1.16
     mentioning
    1.13
     knowing
    1.13
     requiring
    1.10
     necessarily
    1.08
    Act Density 0.048%

    No Known Activations