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

    Neuronpedia

    INDEX
    Explanations

    contractions of "will not" with a focus on the negation aspect

    oai_token-act-pair · gpt-3.5-turbo

    instances of the contraction "wouldn't."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ULT
    -0.76
    dimensional
    -0.66
    PI
    -0.65
     Proced
    -0.59
    ierrez
    -0.57
    gran
    -0.57
     Butt
    -0.57
    æ³
    -0.57
     fib
    -0.56
    PU
    -0.55
    POSITIVE LOGITS
    't
    1.16
    geon
    0.90
    ģĸ
    0.83
    geons
    0.79
    etsk
    0.78
    atically
    0.77
    terness
    0.76
    ¹
    0.76
    anke
    0.75
    §
    0.74
    Act Density 0.014%

    No Known Activations