Some questions regarding the details in the paper

[SummerRice]

Great work on this! I have a few questions I'd like to ask for clarification:

Are the F1 and accuracy (Acc) indicators mentioned in the article evaluated at the triple level, or are they assessed at the paragraph level or sentence level?

In the Case Study section of Appendix F, how are the extracted triples mapped back to the original text? Is this achieved solely by prompting the LLM? If so, could you please elaborate on whether this method might introduce any errors?

Thank you in advance for your insights!

[fxynudt]

Thank you for reading and for your question!
F1 and Accuracy are evaluated at the paragraph level.
In the case study in Appendix F, the highlighted parts simply show that our method can detect those errors.
In real applications, we only need to identify whether there are incorrect triples in the long text; it is not necessary to map them back to the original text.

[SummerRice]

Thank you for reading and for your question! F1 and Accuracy are evaluated at the paragraph level. In the case study in Appendix F, the highlighted parts simply show that our method can detect those errors. In real applications, we only need to identify whether there are incorrect triples in the long text; it is not necessary to map them back to the original text.

OK，thanks for your reply!