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

    Neuronpedia

    INDEX
    Explanations

    phrases where information or opinions are being highlighted or emphasized

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating acknowledgment or observation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    quer
    -0.93
     ILCS
    -0.78
    imeters
    -0.76
    enic
    -0.75
    ene
    -0.72
    agnetic
    -0.72
    ibur
    -0.71
    ãĥ¼ãĥ«
    -0.69
    taboola
    -0.69
    nect
    -0.69
    POSITIVE LOGITS
     how
    1.04
     similarities
    1.04
     inconsistencies
    1.03
     discrepancies
    0.99
     that
    0.97
     shortcomings
    0.90
     differences
    0.89
     inaccur
    0.87
     prominently
    0.85
     approving
    0.85
    Act Density 0.047%

    No Known Activations