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

    Neuronpedia

    INDEX
    Explanations

    conditional clauses or phrases that indicate exceptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     ANY
    -0.24
    ä»»ä½ķ
    -0.17
     EVER
    -0.17
    _ANY
    -0.17
     anytime
    -0.17
     SHOULD
    -0.16
    even
    -0.16
     nawet
    -0.16
    _any
    -0.16
    emplates
    -0.15
    POSITIVE LOGITS
     absolutely
    0.30
     specifically
    0.29
     explicitly
    0.27
     Absolutely
    0.25
     absolute
    0.24
     expressly
    0.24
    explicit
    0.23
     explicit
    0.23
     somehow
    0.22
     specific
    0.22
    Act Density 0.321%

    No Known Activations