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

    Neuronpedia

    INDEX
    Explanations

    phrases indicating a specific goal or purpose

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating intentions or objectives related to assistance or improvement

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    wine
    -0.87
    quartered
    -0.69
     descended
    -0.68
     Held
    -0.65
    ps
    -0.65
    ABLE
    -0.63
     flowed
    -0.62
    brace
    -0.62
    cup
    -0.61
    stars
    -0.61
    POSITIVE LOGITS
     solving
    0.83
    è¦ļéĨĴ
    0.79
     realism
    0.76
    onement
    0.76
     beginners
    0.75
     preserving
    0.74
     inflicting
    0.73
     combating
    0.73
     improving
    0.73
     dissu
    0.72
    Act Density 0.054%

    No Known Activations