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

    Neuronpedia

    INDEX
    Explanations

    superlatives and words related to intensity or extremity

    oai_token-act-pair · gpt-3.5-turbo

    superlative adjectives and comparative terms

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    conservancy
    -0.67
     TAG
    -0.66
     WARN
    -0.64
    pires
    -0.60
     kidnapped
    -0.59
     accomp
    -0.58
     mathemat
    -0.57
     assassinated
    -0.56
     Ellison
    -0.55
     mosqu
    -0.55
    POSITIVE LOGITS
    est
    2.45
    er
    2.33
    nesses
    1.71
    ness
    1.62
    ening
    1.55
    ened
    1.54
    ener
    1.52
    erd
    1.42
    EST
    1.42
    itudinal
    1.40
    Act Density 0.184%

    No Known Activations