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

    Neuronpedia

    INDEX
    Explanations

    adjectives describing emotions or characteristics like comical, wondrous, whimsical, heartwarming, macabre, and awe-inspiring

    oai_token-act-pair · gpt-3.5-turbo

    adjectives describing whimsical or comical qualities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ensor
    -0.78
    utor
    -0.76
    eded
    -0.75
    holding
    -0.70
    grow
    -0.67
    mediate
    -0.67
    vor
    -0.67
    mediated
    -0.66
    catentry
    -0.66
    ribut
    -0.66
    POSITIVE LOGITS
     conco
    0.89
     spectacle
    0.88
     rendition
    0.84
     creations
    0.83
     tale
    0.81
     twist
    0.81
     stunts
    0.80
     twists
    0.79
     glam
    0.77
     tales
    0.76
    Act Density 0.140%

    No Known Activations