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

    Neuronpedia

    INDEX
    Explanations

    conjunctions, particularly "and" and "or"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.16
    1:0.02
    2:0.06
    3:0.07
    4:0.07
    5:0.04
    6:0.20
    7:0.02
    8:0.08
    9:0.07
    10:0.05
    11:0.11
    Negative Logits
    ocrin
    -1.40
     Republic
    -1.37
     Vik
    -1.33
     bunk
    -1.30
    Republic
    -1.27
    guyen
    -1.25
     Tuls
    -1.25
     Tray
    -1.24
    itals
    -1.23
    archs
    -1.22
    POSITIVE LOGITS
    worthiness
    1.52
    CONCLUS
    1.44
     ACTIONS
    1.42
    iqueness
    1.40
    imov
    1.39
    Whereas
    1.31
    fortune
    1.27
     whereas
    1.27
    igure
    1.27
     WATCHED
    1.27
    Act Density 0.003%

    No Known Activations