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

    Neuronpedia

    INDEX
    Explanations

    unrelated phrases with dashes in-between

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
    eering
    -0.98
     metic
    -0.94
     oven
    -0.93
    iple
    -0.92
     indemn
    -0.92
     palm
    -0.91
     redress
    -0.90
     hemor
    -0.89
     Lumpur
    -0.87
     nuts
    -0.87
    POSITIVE LOGITS
    particularly
    1.50
    feat
    1.48
    especially
    1.44
    meaning
    1.42
    along
    1.42
    something
    1.40
    where
    1.39
    which
    1.39
    perhaps
    1.38
    these
    1.37
    Act Density 1.683%

    No Known Activations