MultiGhostBench: A Multilingual Benchmark for Long-Form LLM-Generated Text Attribution under Distribution Shifts
MultiGhostBench is introduced, a multilingual benchmark comprising 928 books generated by five recent LLMs across six languages and three scripts, with an average length of approximately 59K words per book, that supports evaluation under domain, author, and language shifts.
M. Greco, Anudeex Shetty, Andrea Tagarelli et al.
· 0 citations