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

    Neuronpedia

    INDEX
    Explanations

    comparative or superlative structures in sentences

    oai_token-act-pair · gpt-3.5-turbo

    repetitive structures in phrases, especially those starting with "Most."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     Mellon
    -0.70
    abad
    -0.66
    OX
    -0.66
    hirt
    -0.65
    vette
    -0.64
    ettel
    -0.62
    multipl
    -0.62
    essage
    -0.61
    itud
    -0.61
    instead
    -0.60
    POSITIVE LOGITS
     importantly
    1.75
     notably
    1.32
     interestingly
    1.10
    afa
    1.06
     famously
    0.97
     likely
    0.97
     prominently
    0.95
     important
    0.92
     notable
    0.92
     observers
    0.90
    Act Density 0.039%

    No Known Activations