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

    Neuronpedia

    INDEX
    Explanations

    contractions of "is" or "has" with another word following it

    oai_token-act-pair · gpt-3.5-turbo

    repeated phrases indicating a sense of certainty or existence

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    eal
    -0.86
    roit
    -0.77
    ares
    -0.71
    eals
    -0.71
    enth
    -0.68
     approves
    -0.66
     |--
    -0.66
    ear
    -0.64
    soever
    -0.61
    icut
    -0.61
    POSITIVE LOGITS
     been
    1.18
     gotta
    1.16
     plenty
    1.15
     nothing
    1.08
     always
    1.03
     gonna
    0.97
     no
    0.96
     lots
    0.95
     definitely
    0.94
     something
    0.91
    Act Density 0.057%

    No Known Activations