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

    Neuronpedia

    INDEX
    Explanations

    references to negative outcomes or quantitative losses

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "losses" and related terms

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    Created
    -0.71
    pol
    -0.66
    dayName
    -0.66
    JB
    -0.64
    Fram
    -0.63
    cart
    -0.63
    ISTER
    -0.63
    Offic
    -0.62
    bara
    -0.61
    pter
    -0.61
    POSITIVE LOGITS
     losses
    3.83
     loss
    2.33
    loss
    2.18
     Loss
    2.13
     defeats
    1.90
     loses
    1.74
     setbacks
    1.67
     losers
    1.65
     victories
    1.58
     failures
    1.55
    Act Density 0.014%

    No Known Activations