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

    Neuronpedia

    INDEX
    Explanations

    phrases related to stating facts or observations

    oai_token-act-pair · gpt-3.5-turbo

    repetitive phrases that introduce statements or facts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    gur
    -0.64
    mes
    -0.62
    borg
    -0.61
    Gy
    -0.61
    Guard
    -0.60
    throp
    -0.60
    iatric
    -0.58
    pling
    -0.58
     Canad
    -0.58
    roth
    -0.58
    POSITIVE LOGITS
     hindsight
    0.84
     contradicts
    0.78
     fateful
    0.78
     happened
    0.74
     arose
    0.73
    cher
    0.73
     they
    0.73
    soever
    0.72
     accompanies
    0.71
     we
    0.70
    Act Density 0.374%

    No Known Activations