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

    Neuronpedia

    INDEX
    Explanations

    contractions where the subject is 'we' and the following word is a verb

    oai_token-act-pair · gpt-3.5-turbo

    negations, particularly the phrase "didn't."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    armed
    -0.75
     Dise
    -0.74
     DRAGON
    -0.72
    upp
    -0.68
     Methods
    -0.64
    OTOS
    -0.63
    mania
    -0.62
     VERS
    -0.61
     Supporting
    -0.60
     Duty
    -0.60
    POSITIVE LOGITS
    't
    1.26
    etsk
    0.86
    nt
    0.76
    okia
    0.72
    nels
    0.72
    atio
    0.70
     indeed
    0.70
    kered
    0.69
    geon
    0.68
    ":"/
    0.68
    Act Density 0.052%

    No Known Activations