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

    Neuronpedia

    INDEX
    Explanations

    words and punctuation that often occur together in English writing.

    oai_token-act-pair · gemini-2.0-flash

    Complex sentences

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Negative Logits
     Namely
    -1.60
     namely
    -1.48
    namely
    -1.23
     viz
    -1.18
     cioè
    -1.14
     nemlig
    -1.11
     ossia
    -1.01
     یعنی
    -1.00
    viz
    -0.99
    即
    -0.99
    POSITIVE LOGITS
     which
    0.74
    gameserver
    0.74
    Autoritní
    0.59
    TagHelper
    0.56
    which
    0.56
     Which
    0.53
     للاسماء
    0.52
    Which
    0.51
     чему
    0.48
     vilket
    0.48
    Act Density 0.652%

    No Known Activations