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

    Neuronpedia

    INDEX
    Explanations

    words related to closure or completion

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "close" and its variations, indicating actions related to closure or termination

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ãĥīãĥ©ãĤ´ãĥ³
    -0.75
    nis
    -0.71
    lain
    -0.71
    orum
    -0.68
    alis
    -0.67
    ican
    -0.67
    icist
    -0.67
    rouse
    -0.64
    ial
    -0.64
    inen
    -0.64
    POSITIVE LOGITS
     loopholes
    0.98
     doors
    0.96
     Doors
    0.91
     curtains
    0.83
     shutter
    0.81
    doors
    0.71
     abruptly
    0.70
     shop
    0.70
     down
    0.69
     door
    0.68
    Act Density 0.031%

    No Known Activations