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

    Neuronpedia

    INDEX
    Explanations

    exclamation points combined with phrases that convey a sense of conviction or urgency

    oai_token-act-pair · gpt-3.5-turbo

    instances of special characters and unusual punctuation patterns

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     snacks
    -0.78
     cob
    -0.74
     reception
    -0.67
     tram
    -0.66
     telesc
    -0.66
     tranquil
    -0.65
     Particip
    -0.65
     mosaic
    -0.64
     snack
    -0.64
     seiz
    -0.64
    POSITIVE LOGITS
    Ļ
    1.37
    Ĵ
    1.27
    ¤
    1.23
    ¬
    1.19
    ĵ
    1.15
    ħ
    1.14
    Ķ
    1.13
    ¡
    1.12
    £
    1.11
    ĺ
    1.08
    Act Density 0.213%

    No Known Activations