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

    Neuronpedia

    INDEX
    Explanations

    words related to being impassable, unable to be crossed

    oai_token-act-pair · gpt-3.5-turbo

    instances of the substring "etr" across various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    croft
    -0.79
    Ĥ¬
    -0.71
    ãģ¦
    -0.70
    ships
    -0.68
    ĺħ
    -0.68
    berries
    -0.67
    conn
    -0.67
    e
    -0.66
     Hastings
    -0.66
     missionary
    -0.65
    POSITIVE LOGITS
    aining
    1.21
    etr
    1.16
    acer
    1.07
    ading
    1.06
    ifying
    1.01
    ained
    0.99
    abulary
    0.99
    aded
    0.97
    agonal
    0.94
    asso
    0.93
    Act Density 0.020%

    No Known Activations