Benchmarking concerns

#2
by Tech-Meld - opened

Dear Authors,
Have you ever considered benchmarking this model ? Do you consider doing that ? Are there anything that the community can work on to improve benchmarking long sequences of AI generated text ?

Knowledge Engineering Group (KEG) & Data Mining at Tsinghua University org

Good question! We release two benchmarks along our work: LongBench-Write and LongWrite-Ruler. Please see our github Repo for more details: https://github.com/THUDM/LongWriter?tab=readme-ov-file#evaluation

That's really helpful, thanks!

zRzRzRzRzRzRzR changed discussion status to closed

Sign up or log in to comment