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

    Neuronpedia

    INDEX
    Explanations

    instances where objects or people are physically close together or in contact

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "touch" and its variations, particularly in relation to people and their proximity

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     mathemat
    -0.76
    edition
    -0.75
    wcs
    -0.75
     enthusi
    -0.72
    price
    -0.71
     redes
    -0.70
     traged
    -0.70
     reconc
    -0.67
     disapp
    -0.66
     incorpor
    -0.66
    POSITIVE LOGITS
     objects
    1.13
     bushes
    0.97
    objects
    0.94
     rocks
    0.92
     surfaces
    0.91
     corpses
    0.90
     enemy
    0.89
     walls
    0.87
     anything
    0.87
     trees
    0.87
    Act Density 0.529%

    No Known Activations