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

    Neuronpedia

    INDEX
    Explanations

    phrases related to toxic substances

    oai_token-act-pair · gpt-3.5-turbo

    references to toxins or harmful substances

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    noon
    -0.88
    dar
    -0.77
    pora
    -0.76
    aan
    -0.69
    quarter
    -0.66
    doms
    -0.64
    RM
    -0.64
    Merit
    -0.63
    romy
    -0.63
    ann
    -0.62
    POSITIVE LOGITS
     poisoning
    1.07
     poison
    1.00
     dart
    0.95
     poisoned
    0.91
     darts
    0.90
     iv
    0.90
    ously
    0.85
     Ivy
    0.84
     poisonous
    0.84
     gas
    0.84
    Act Density 0.012%

    No Known Activations