INDEX
    Explanations

    Technical text snippets

    This neuron fires on the demonstrative “These” when it begins explanatory sentences (e.g. “These can only reduce…”), marking definition/description clauses.

    New Auto-Interp
    Negative Logits
     ByteString
    -0.07
     XIII
    -0.06
    -0.06
    valu
    -0.06
     plu
    -0.06
     labyrinth
    -0.06
                                                                                               
    -0.06
     вересня
    -0.06
     ADMIN
    -0.06
    voje
    -0.06
    POSITIVE LOGITS
     đúng
    0.07
    0.06
     settles
    0.06
     ironically
    0.06
     decorating
    0.06
    (fe
    0.06
    アル
    0.06
    Helmet
    0.06
     Designs
    0.06
    .algorithm
    0.06
    Act Density 0.081%

    No Known Activations