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

    Neuronpedia

    INDEX
    Explanations

    instances where the phrase "As you can see" is used

    oai_token-act-pair · gpt-3.5-turbo

    phrases that express potential or capability

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     Lug
    -0.79
    rang
    -0.61
    erville
    -0.61
     Abandon
    -0.61
    amba
    -0.58
    orno
    -0.58
    andon
    -0.58
    OUP
    -0.58
    rament
    -0.57
     SG
    -0.57
    POSITIVE LOGITS
     imagine
    1.17
     guessed
    1.16
     see
    1.05
     attest
    1.05
     guess
    1.04
     infer
    1.04
     plainly
    0.97
     glean
    0.94
     observe
    0.93
     ded
    0.90
    Act Density 0.041%

    No Known Activations