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

    Neuronpedia

    INDEX
    Explanations

    phrases related to emphasizing the importance or rationale behind a particular action or decision

    oai_token-act-pair · gpt-3.5-turbo

    references to concepts of value and purpose

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     Cosponsors
    -1.19
    DragonMagazine
    -0.82
    falls
    -0.77
    til
    -0.73
    aults
    -0.69
    venants
    -0.67
     ILCS
    -0.67
    glass
    -0.66
    offer
    -0.66
    hook
    -0.66
    POSITIVE LOGITS
     preserving
    1.10
     fairness
    1.05
     integrity
    0.97
     protecting
    0.95
     decency
    0.92
     confidentiality
    0.91
     justice
    0.91
     efficiency
    0.90
     simplicity
    0.90
     purity
    0.89
    Act Density 0.134%

    No Known Activations