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

    Neuronpedia

    INDEX
    Explanations

    technology-related terms and mentions of research projects

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
     defe
    -0.63
     corrid
    -0.62
     pit
    -0.59
     footing
    -0.59
     tyr
    -0.58
    untarily
    -0.58
    ikuman
    -0.57
     skelet
    -0.56
     metab
    -0.55
    pection
    -0.54
    POSITIVE LOGITS
     Related
    0.88
     Their
    0.86
     Specifically
    0.84
     Additionally
    0.82
     Written
    0.81
     Unfortunately
    0.80
     Meanwhile
    0.78
     However
    0.77
     Below
    0.77
     Along
    0.77
    Act Density 4.973%

    No Known Activations