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

    Neuronpedia

    INDEX
    Explanations

    phrases related to action or implementation

    oai_token-act-pair · gpt-3.5-turbo

    instances of words related to actions being properly applied or enacted

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    whatever
    -0.62
    ajo
    -0.62
     Vaughan
    -0.61
     Vie
    -0.60
     Ethiopia
    -0.58
     Filip
    -0.58
    feat
    -0.56
    eah
    -0.56
     Baal
    -0.56
    Hop
    -0.55
    POSITIVE LOGITS
     properly
    1.11
     correctly
    1.00
     individually
    0.82
     perpend
    0.74
     appropriately
    0.73
     incorrectly
    0.72
     improperly
    0.70
     aloud
    0.70
    urally
    0.69
     together
    0.69
    Act Density 0.102%

    No Known Activations