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

    Neuronpedia

    INDEX
    Explanations

    phrases related to physical crashes or sudden declines

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of crashing or collisions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    arov
    -0.71
    icrobial
    -0.71
    achu
    -0.69
    inguishable
    -0.68
    iries
    -0.68
    esthetic
    -0.65
    åĮ
    -0.65
    inen
    -0.64
     suscept
    -0.64
    rov
    -0.63
    POSITIVE LOGITS
     Dive
    0.90
     crash
    0.89
     crashes
    0.88
    worthiness
    0.85
     Crash
    0.78
    ulent
    0.77
     wreck
    0.76
    audio
    0.75
    ulence
    0.75
     crashed
    0.74
    Act Density 0.021%

    No Known Activations