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

    Neuronpedia

    INDEX
    Explanations

    phrases related to capability and possibility

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating capability or permission

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    alm
    -0.68
    igl
    -0.68
    figure
    -0.67
    arms
    -0.67
    zbollah
    -0.66
    gdala
    -0.66
    irmation
    -0.64
    Yep
    -0.63
    azar
    -0.63
    ignment
    -0.62
    POSITIVE LOGITS
     anymore
    1.23
     nor
    0.90
     necessarily
    0.79
     anybody
    0.78
     any
    0.74
     anyone
    0.71
     bothered
    0.70
     anywhere
    0.70
     whatsoever
    0.69
     ever
    0.69
    Act Density 0.198%

    No Known Activations