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

    Neuronpedia

    INDEX
    Explanations

    superlatives or extremes such as "oldest", "shortest", "hottest"

    oai_token-act-pair · gpt-3.5-turbo

    superlative adjectives

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    yang
    -0.76
    yi
    -0.74
    tis
    -0.73
    skirts
    -0.71
    ending
    -0.70
    olor
    -0.66
    çīĪ
    -0.66
    alus
    -0.66
    ortium
    -0.66
     Cosponsors
    -0.66
    POSITIVE LOGITS
     imaginable
    1.44
     conceivable
    1.40
     possible
    1.37
     ever
    1.16
     practicable
    1.13
     EVER
    0.93
     Possible
    0.89
     extant
    0.88
     thing
    0.85
     feasible
    0.84
    Act Density 0.066%

    No Known Activations