CulturalCrosswords benchmark tests multilingual answer constraints in grid puzzles
A new benchmark called CulturalCrosswords evaluates how answer constraints affect large language models on multilingual crossword-style grid completion, where exact-match scoring can penalise valid responses.
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-11-10 05:00 · arXiv cs.CL
CulturalCrosswords: A Multilingual Benchmark comprised of Answer Constraints and Grid Completion