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

    Neuronpedia

    INDEX
    Explanations

    personal pronouns in conjunction with the verbs "were" and "are"

    oai_token-act-pair · gpt-3.5-turbo

    pronouns related to personal relationships and identity

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    catentry
    -0.95
    Reviewer
    -0.78
     childbirth
    -0.67
    Synopsis
    -0.66
     concluding
    -0.63
     Lifetime
    -0.62
     Description
    -0.62
     Variety
    -0.60
     persuasion
    -0.60
     Infinite
    -0.58
    POSITIVE LOGITS
    've
    1.15
    're
    1.15
     belonged
    1.10
     belong
    1.09
     existed
    1.07
    'd
    1.03
    'm
    1.03
     cheated
    1.02
    'll
    0.97
     lied
    0.96
    Act Density 0.196%

    No Known Activations