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

    Neuronpedia

    INDEX
    Explanations

    references to sources or attributions within a text

    oai_token-act-pair · gpt-3.5-turbo

    instances of the phrase "according to" followed by a source or reference

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    VERT
    -0.66
    wid
    -0.66
    ield
    -0.65
    arse
    -0.65
    fax
    -0.64
    izont
    -0.63
    cffffcc
    -0.62
    rette
    -0.62
    rouse
    -0.62
    ertodd
    -0.61
    POSITIVE LOGITS
     "[
    1.21
     "âĢ¦
    1.00
     "'
    0.99
     however
    0.92
     "...
    0.91
     "(
    0.87
     there
    0.86
     '[
    0.86
     citing
    0.82
     which
    0.82
    Act Density 0.108%

    No Known Activations