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

    Neuronpedia

    INDEX
    Explanations

    mentions of physical damage or breakage

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "cracked" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    enture
    -0.73
    endment
    -0.69
    gard
    -0.68
    amins
    -0.66
    anamo
    -0.64
    alla
    -0.64
    ==
    -0.62
     mater
    -0.60
    ript
    -0.60
    urai
    -0.60
    POSITIVE LOGITS
    pots
    1.10
    kered
    0.98
     cracking
    0.89
     cracked
    0.84
    Berry
    0.77
    ibly
    0.73
     cracks
    0.72
     jaws
    0.72
    ulously
    0.70
    etooth
    0.70
    Act Density 0.007%

    No Known Activations