Long-context summarization dataset (Findings of ACL 2024): 2,200 movie screenplays (~29K words each) paired with Wikipedia plot summaries — twice the size of prior screenplay datasets. Clean Edinburgh work (Saxena, Keller, ILCC), in Lapata's ILCC long-context summarization orbit.

Dataset

Size 2,200 screenplays
Format text
datasetnlpresearch