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

    Neuronpedia

    INDEX
    Explanations

    sentences that convey advice or cautionary guidance

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     royalty
    -0.75
     birthplace
    -0.74
     descendants
    -0.72
     flagship
    -0.70
     namesake
    -0.68
     ancestry
    -0.67
     crashed
    -0.67
     crude
    -0.67
     boarded
    -0.66
     leaked
    -0.66
    POSITIVE LOGITS
     Ideally
    1.54
     Otherwise
    1.48
     Include
    1.30
     Especially
    1.29
     Doing
    1.28
     Lastly
    1.25
     If
    1.25
     Then
    1.25
     Avoid
    1.24
     Remember
    1.23
    Act Density 0.153%

    No Known Activations