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

    Neuronpedia

    INDEX
    Explanations

    phrases relating to speaking or acting on someone's behalf

    oai_token-act-pair · gpt-3.5-turbo

    phrases that indicate representation or advocacy for someone else

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    DIT
    -0.73
    HUD
    -0.67
    ################################
    -0.64
     Fulton
    -0.64
    Article
    -0.62
    teen
    -0.62
     Extrem
    -0.61
    TT
    -0.61
     Traps
    -0.60
    ANG
    -0.59
    POSITIVE LOGITS
     behalf
    1.13
     guiActiveUn
    0.80
    auga
    0.76
    maid
    0.75
    wcsstore
    0.69
     coerc
    0.68
    ouched
    0.67
    indal
    0.67
    oux
    0.67
     ende
    0.66
    Act Density 0.005%

    No Known Activations