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

    Neuronpedia

    INDEX
    Explanations

    technical issues or errors mentioned in a text

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating failures or issues in functionality

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    mere
    -0.82
    gone
    -0.79
    ãĤ®
    -0.79
    bush
    -0.73
    gemony
    -0.72
    iety
    -0.71
    é¾
    -0.71
    bare
    -0.69
    soDeliveryDate
    -0.65
    quarters
    -0.64
    POSITIVE LOGITS
     properly
    1.13
     load
    1.12
     correctly
    1.09
     recognise
    1.04
     recognize
    1.03
     sync
    1.02
     detect
    1.02
     display
    0.95
     function
    0.95
     compile
    0.94
    Act Density 0.170%

    No Known Activations