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

    Neuronpedia

    INDEX
    Explanations

    instances of the word "roll" followed by numbers, connoting a variety of contexts such as physical actions or procedural steps

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "roll" and its variations in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ILE
    -0.76
    acters
    -0.69
    xual
    -0.68
    orial
    -0.63
    bub
    -0.63
    Hunt
    -0.63
    selage
    -0.62
    raints
    -0.62
    rent
    -0.62
    ulia
    -0.61
    POSITIVE LOGITS
    icking
    0.92
     boulder
    0.86
    back
    0.86
     dice
    0.84
    outs
    0.77
    anut
    0.77
     hills
    0.76
     downhill
    0.76
    ahon
    0.76
     oats
    0.75
    Act Density 0.039%

    No Known Activations