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

    Neuronpedia

    INDEX
    Explanations

    words related to traffic accidents and crashes

    oai_token-act-pair · gpt-3.5-turbo

    references to various types of crashes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    achu
    -0.73
    arov
    -0.70
    tains
    -0.69
    ophers
    -0.67
    egal
    -0.67
    Ĭ±
    -0.66
    icrobial
    -0.65
    andom
    -0.63
    gypt
    -0.63
    abbit
    -0.63
    POSITIVE LOGITS
     crash
    1.22
    Crash
    1.16
     crashes
    1.14
     Crash
    1.07
     crashed
    0.97
     Dive
    0.93
     crater
    0.86
     crashing
    0.82
    worthiness
    0.81
     wreck
    0.80
    Act Density 0.009%

    No Known Activations