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

    Neuronpedia

    INDEX
    Explanations

    sentence endings with punctuation marks

    oai_token-act-pair · gpt-3.5-turbo

    sentences that convey a sense of noncommittal agreement or neutrality

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     purse
    -0.77
     allowance
    -0.72
     ambush
    -0.70
     robber
    -0.67
    hoe
    -0.62
     culture
    -0.61
     sponsor
    -0.61
     inund
    -0.60
    roud
    -0.60
     rapidly
    -0.57
    POSITIVE LOGITS
     Regardless
    1.07
     Nevertheless
    1.06
     Whether
    1.05
     Nonetheless
    1.03
     However
    1.03
     Unless
    1.01
     Instead
    1.01
     Especially
    1.00
     Whereas
    1.00
     Thankfully
    0.99
    Act Density 0.764%

    No Known Activations