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

    Neuronpedia

    INDEX
    Explanations

    hyphenated words or phrases

    oai_token-act-pair · gpt-3.5-turbo

    phrases that include the word "newly" followed by various forms of the past participle of verbs

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     Torrent
    -0.76
     Survivors
    -0.75
     Rica
    -0.74
     Dickinson
    -0.73
     Jackets
    -0.72
     Kali
    -0.72
     Slave
    -0.72
     Howe
    -0.71
     Samp
    -0.71
     circle
    -0.71
    POSITIVE LOGITS
    defined
    1.23
    connected
    1.10
    focused
    1.07
    mint
    1.06
    derived
    1.04
    produced
    1.04
    protected
    1.04
    recogn
    1.02
    exclusive
    1.02
    designed
    1.02
    Act Density 0.051%

    No Known Activations