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

    Neuronpedia

    INDEX
    Explanations

    sentences that end with a period and appear to have a negative connotation

    oai_token-act-pair · gpt-3.5-turbo

    sentences that express strong opinions or evaluations of various subjects

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    usalem
    -0.76
     mosqu
    -0.76
    onde
    -0.74
     sponsor
    -0.72
    idal
    -0.70
    idium
    -0.68
    uese
    -0.68
     face
    -0.66
     escort
    -0.66
     authorised
    -0.66
    POSITIVE LOGITS
     Especially
    1.29
     Regardless
    1.18
     Nonetheless
    1.16
     Whereas
    1.14
     Nevertheless
    1.13
     Particularly
    1.12
     However
    1.09
     Otherwise
    1.08
     Therefore
    1.07
     Unless
    1.05
    Act Density 0.694%

    No Known Activations