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

    Neuronpedia

    INDEX
    Explanations

    references to closing or shutting down something, such as schools or businesses

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "close" related to closures or restrictions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ãĥīãĥ©ãĤ´ãĥ³
    -0.80
    orum
    -0.69
    anse
    -0.67
    icist
    -0.66
    alis
    -0.64
    ican
    -0.63
    ial
    -0.63
     endors
    -0.62
    MN
    -0.62
    ager
    -0.62
    POSITIVE LOGITS
     loopholes
    1.02
     doors
    0.98
     Doors
    0.92
    doors
    0.81
     curtains
    0.80
     shutter
    0.77
     door
    0.75
     down
    0.73
    door
    0.71
     throttle
    0.71
    Act Density 0.035%

    No Known Activations