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

    Neuronpedia

    INDEX
    Explanations

    phrases related to self-identification and personal characteristics

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
     Pwr
    -0.84
    displayText
    -0.82
    akedown
    -0.80
    inventoryQuantity
    -0.74
    effective
    -0.72
    cipled
    -0.71
    overy
    -0.70
    Policy
    -0.70
    ensing
    -0.69
    PRE
    -0.69
    POSITIVE LOGITS
     characters
    1.43
     objects
    1.22
     animals
    1.21
     humans
    1.19
     silhou
    1.19
     bodies
    1.15
     faces
    1.14
     creatures
    1.14
     dolls
    1.12
     portraits
    1.12
    Act Density 0.686%

    No Known Activations