NLP Framework for Organizing Classical Chinese Exegetical Knowledge

A new computational framework uses two-step prompt chaining and cross-source mention clustering to organize historical commentaries in Classical Chinese texts. The method achieved a CoNLL F1 score above 97% in classifying the relationship between inline notes and main text segments.