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

    Neuronpedia

    INDEX
    Explanations

    references to the term "Sorcery."

    oai_token-act-pair · gpt-3.5-turbo

    references to sorcery or magical themes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     checkpoint
    -0.72
     bowel
    -0.69
     ledger
    -0.69
    mitt
    -0.65
     craw
    -0.65
     jog
    -0.65
    bill
    -0.64
     bust
    -0.63
    ct
    -0.63
    urg
    -0.62
    POSITIVE LOGITS
     Sor
    4.06
     sor
    2.43
     sorcery
    1.35
     Sorcerer
    1.27
     Sorce
    1.20
     Bard
    1.09
     Shir
    1.05
     Sar
    1.05
     Vor
    1.04
     Mou
    1.03
    Act Density 0.014%

    No Known Activations