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

    Neuronpedia

    INDEX
    Explanations

    safety tips related to securing objects to prevent accidents and damage

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to safety precautions and risks

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     streamlined
    -0.69
    iem
    -0.68
    ulously
    -0.67
    ellen
    -0.65
    mostly
    -0.64
    atorium
    -0.63
    onement
    -0.62
     simplified
    -0.62
    irmation
    -0.61
    orest
    -0.60
    POSITIVE LOGITS
     inappropriately
    1.18
     interfering
    1.08
     too
    1.07
    danger
    1.06
     improperly
    1.05
     accidentally
    1.05
     undue
    1.04
     interfere
    1.04
     contamin
    1.02
     unnecessarily
    1.02
    Act Density 0.636%

    No Known Activations