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

    Neuronpedia

    INDEX
    Explanations

    instances of emphasis or confirmation in text

    oai_token-act-pair · gpt-3.5-turbo

    expressions of emphasis or intensifiers, particularly the word "really."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    tnc
    -0.83
    lain
    -0.80
    ription
    -0.73
    oise
    -0.72
    hz
    -0.70
    ousel
    -0.69
    tailed
    -0.67
    eer
    -0.65
    glers
    -0.65
    vae
    -0.65
    POSITIVE LOGITS
     speaking
    0.98
    etheless
    0.80
     intending
    0.78
     conclud
    0.76
    speaking
    0.75
     Speaking
    0.74
    entimes
    0.72
     tho
    0.72
    !,
    0.70
     excluding
    0.70
    Act Density 0.160%

    No Known Activations