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

    Neuronpedia

    INDEX
    Explanations

    capital letters with unusual symbols and low numbers

    oai_token-act-pair · gpt-3.5-turbo

    instances of discourse or communication topics

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     exting
    -0.83
     citiz
    -0.80
     newcom
    -0.78
     tremend
    -0.78
     undermin
    -0.77
     eleph
    -0.76
    aditional
    -0.75
     exha
    -0.74
    senal
    -0.74
     proport
    -0.73
    POSITIVE LOGITS
    ↵
    0.86
    SPONSORED
    0.82
    PHOTOS
    0.80
    Scroll
    0.76
    NPR
    0.74
    Specifically
    0.70
    Eh
    0.68
    DragonMagazine
    0.68
    iak
    0.68
    sonian
    0.67
    Act Density 0.541%

    No Known Activations