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

    Neuronpedia

    INDEX
    Explanations

    phrases or words related to admissions of guilt or ownership

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "admit" and its variations, indicating admissions of guilt or truthfulness

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     ILCS
    -0.82
    rior
    -0.71
    osi
    -0.70
    ighth
    -0.68
    acements
    -0.67
    atility
    -0.66
    ãĤ¡
    -0.65
    asus
    -0.64
    rouse
    -0.63
    adesh
    -0.63
    POSITIVE LOGITS
     wrongdoing
    1.07
     defeat
    1.01
     guilt
    0.99
     responsibility
    0.78
     ignorance
    0.77
     admitting
    0.77
     mistakes
    0.77
     fault
    0.76
     culp
    0.74
     admit
    0.74
    Act Density 0.057%

    No Known Activations