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

    Neuronpedia

    INDEX
    Explanations

    irony and unexpected twists

    oai_token-act-pair · gpt-3.5-turbo

    concepts related to irony and contradictions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    onding
    -0.78
     Doing
    -0.74
    Moving
    -0.68
     noticing
    -0.67
    ittal
    -0.63
    cking
    -0.63
     realizing
    -0.63
     deciding
    -0.62
     moving
    -0.62
    ivating
    -0.61
    POSITIVE LOGITS
     resembles
    1.05
     resembled
    0.98
     constitutes
    0.86
    pires
    0.86
     dominates
    0.83
     translates
    0.82
    itiz
    0.81
     accompanies
    0.80
     coincides
    0.79
     includes
    0.79
    Act Density 0.215%

    No Known Activations