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

    Neuronpedia

    INDEX
    Explanations

    instances where characters are staring or looking at something

    oai_token-act-pair · gpt-3.5-turbo

    phrases that describe visual attention or focused observation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    eno
    -0.82
    Claim
    -0.76
    buster
    -0.67
    GV
    -0.66
    construct
    -0.65
    Fact
    -0.63
    wcs
    -0.63
    Topic
    -0.62
     lobbied
    -0.62
    ItemImage
    -0.62
    POSITIVE LOGITS
     passers
    0.96
     camera
    0.87
     screens
    0.85
     sunrise
    0.84
     scoreboard
    0.83
     horizon
    0.83
     blank
    0.82
     cameras
    0.82
     screen
    0.78
    screen
    0.78
    Act Density 0.167%

    No Known Activations