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

    Neuronpedia

    INDEX
    Explanations

    phrases with the conjunction 'and' along with an element of comparison

    oai_token-act-pair · gpt-3.5-turbo

    the conjunction "and" to identify connected phrases or ideas

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    cised
    -0.66
    UGE
    -0.66
     Powered
    -0.61
    sung
    -0.60
    atan
    -0.59
    Share
    -0.59
    lections
    -0.58
    owered
    -0.57
    utenberg
    -0.56
    Constructed
    -0.56
    POSITIVE LOGITS
     hence
    1.49
     thus
    1.42
     therefore
    1.41
     consequently
    1.38
     thereby
    1.28
     secondly
    1.21
     furthermore
    1.13
     moreover
    0.97
     THEN
    0.97
     vice
    0.93
    Act Density 0.423%

    No Known Activations