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

    Neuronpedia

    INDEX
    Explanations

    words related to exceeding limits or expectations

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to exceeding specified limits or thresholds

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    olan
    -0.81
    uay
    -0.76
    NetMessage
    -0.74
    apple
    -0.70
    Downloadha
    -0.69
    issue
    -0.69
    pring
    -0.69
    hair
    -0.68
    kaya
    -0.67
    rug
    -0.67
    POSITIVE LOGITS
     exceeded
    1.18
     exceed
    1.18
     exceeding
    1.02
    ceed
    0.96
     exceeds
    0.95
     surpassed
    0.84
     surpass
    0.84
     heights
    0.73
     expectations
    0.73
     limits
    0.70
    Act Density 0.012%

    No Known Activations