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

    Neuronpedia

    INDEX
    Explanations

    items related to crime or illegal activities

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    the hyphen in "heavily-" when it appears as part of compound modifiers or adjectives.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jiahuipeng
    New Auto-Interp
    Negative Logits
     Jackets
    -0.90
     circle
    -0.81
     Centauri
    -0.77
     Dickinson
    -0.76
     Slip
    -0.75
     Cotton
    -0.74
     Brus
    -0.74
     Shelley
    -0.71
     Torrent
    -0.70
     Swordsman
    -0.70
    POSITIVE LOGITS
    developed
    1.39
    produced
    1.38
    defined
    1.35
    trained
    1.34
    equipped
    1.33
    connected
    1.32
    protected
    1.31
    appointed
    1.30
    recogn
    1.30
    funded
    1.28
    Act Density 0.089%

    No Known Activations