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

    Neuronpedia

    INDEX
    Explanations

    references to materials or objects that are resistant or protective, specifically mentioning properties like being bulletproof or fireproof

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
    atro
    -0.88
    rera
    -0.87
    udeb
    -0.86
    onde
    -0.84
    iewicz
    -0.80
     Unch
    -0.78
    atown
    -0.78
     DRAG
    -0.76
    izarre
    -0.74
    YN
    -0.74
    POSITIVE LOGITS
    ness
    1.16
    nesses
    0.86
     coat
    0.80
     coating
    0.76
     storage
    0.74
    ened
    0.74
    iating
    0.73
     coats
    0.73
     discharge
    0.72
     retri
    0.72
    Act Density 1.227%

    No Known Activations