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

    Neuronpedia

    INDEX
    Explanations

    terms indicating something surpassing a limit or threshold

    oai_token-act-pair · gpt-3.5-turbo

    references to limits or thresholds being surpassed

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    rug
    -0.78
    choice
    -0.68
    pring
    -0.66
    onna
    -0.66
    Ale
    -0.66
    seed
    -0.65
    Downloadha
    -0.65
    rol
    -0.65
    roller
    -0.64
    Sham
    -0.64
    POSITIVE LOGITS
    ingly
    0.86
     9000
    0.82
     expectations
    0.81
    ceed
    0.77
    =>
    0.76
    atos
    0.75
     400
    0.72
    capacity
    0.71
     exceed
    0.70
    imates
    0.66
    Act Density 0.017%

    No Known Activations