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

    Neuronpedia

    INDEX
    Explanations

    the word "step"

    oai_token-act-pair · gpt-4-turbo

    phrases related to progress or advancement

    oai_token-act-pair · gpt-3.5-turbo

    phrases that indicate progress or advancement

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    anguages
    -0.72
     Corpus
    -0.70
    ores
    -0.70
    ortunately
    -0.69
    eatures
    -0.65
     Pengu
    -0.64
    iqueness
    -0.64
    inately
    -0.62
     canned
    -0.62
    ciating
    -0.62
    POSITIVE LOGITS
    daughter
    1.00
    dad
    0.95
    DragonMagazine
    0.93
    Step
    0.93
    step
    0.90
     steps
    0.89
     step
    0.89
     Steps
    0.88
    steps
    0.87
    brother
    0.84
    Act Density 0.019%

    No Known Activations