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

    Neuronpedia

    INDEX
    Explanations

    phrases indicating contrast or opposition

    oai_token-act-pair · gpt-3.5-turbo

    conditional phrases or clauses that set up contrasts or exceptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    ahoo
    -0.74
    stadt
    -0.72
    eus
    -0.72
    asus
    -0.71
    enture
    -0.71
    iesta
    -0.70
    iot
    -0.70
    zag
    -0.70
    eva
    -0.69
    izons
    -0.69
    POSITIVE LOGITS
     imperfect
    1.01
     lacking
    1.00
     technically
    0.96
     composed
    0.92
     useful
    0.92
     understandable
    0.91
     ostensibly
    0.91
     capable
    0.91
     harmless
    0.90
     considered
    0.89
    Act Density 0.121%

    No Known Activations