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

    Neuronpedia

    INDEX
    Explanations

    phrases related to making up for something that is lacking or missing

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to compensating or making up for losses or deficiencies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    eki
    -0.75
    raq
    -0.75
    eks
    -0.71
    english
    -0.71
    ailand
    -0.69
    usa
    -0.67
    oly
    -0.67
    abad
    -0.65
    ek
    -0.65
    ura
    -0.65
    POSITIVE LOGITS
     deficiencies
    1.08
     losses
    1.08
     lost
    1.03
     shortfall
    1.00
     shortcomings
    0.95
     weaknesses
    0.92
     missing
    0.92
     discrepancies
    0.92
     deficiency
    0.91
     inconsistencies
    0.90
    Act Density 0.159%

    No Known Activations