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

    Neuronpedia

    INDEX
    Explanations

    phrases indicating multiplication or increase

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "fold" used in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    vae
    -0.71
     indo
    -0.69
     Vital
    -0.67
     Esper
    -0.66
    anwhile
    -0.66
    vil
    -0.65
    pez
    -0.63
     Lots
    -0.62
     Altern
    -0.61
     Julio
    -0.60
    POSITIVE LOGITS
    fold
    1.62
     fold
    1.25
    ername
    1.03
     Fold
    0.90
     folded
    0.89
     folding
    0.88
    ers
    0.79
    theless
    0.78
     folds
    0.72
    shr
    0.71
    Act Density 0.004%

    No Known Activations