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

    Neuronpedia

    INDEX
    Explanations

    phrases related to failure or neglect

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating failure or shortcomings in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    moil
    -0.88
    loss
    -0.72
    interrupted
    -0.68
    anian
    -0.67
     Panic
    -0.65
     downfall
    -0.64
    adel
    -0.62
    urry
    -0.62
    gdala
    -0.60
     Mayhem
    -0.59
    POSITIVE LOGITS
     adequately
    1.17
     properly
    0.99
     adequate
    0.97
     enough
    0.81
     timely
    0.80
     stewards
    0.80
     proper
    0.80
     truthful
    0.80
     Proper
    0.79
     nutritious
    0.78
    Act Density 1.005%

    No Known Activations