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

    Neuronpedia

    INDEX
    Explanations

    contractions ending in 's', indicating possession or abbreviation.

    oai_token-act-pair · gpt-3.5-turbo

    possessive forms and contractions similar to "he's," "she's," and "who's."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    stal
    -0.70
    ren
    -0.66
    TAIN
    -0.65
    ={
    -0.64
    ESE
    -0.64
    sters
    -0.64
    rent
    -0.63
    DS
    -0.62
     Reduce
    -0.62
    ollow
    -0.62
    POSITIVE LOGITS
     gonna
    1.36
     gotta
    1.31
     been
    1.18
     got
    1.15
     gotten
    1.10
    been
    1.09
     gone
    0.89
     Been
    0.87
     going
    0.87
     supposed
    0.85
    Act Density 0.148%

    No Known Activations