Reproduce “Cross-Lingual Consensus: Aligning Multilingual Cultural Knowledge via Multilingual Self-Consistency” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Andrew Ivan Soegeng, Patrick Sutanto, Tan Sang Nguyen
The authors use cross-lingual self-consistency to train a multilingual model on cultural questions without human answer annotations or a stronger teacher. They compare agreement among answers within each language, translate a representative answer from the more consistent language, and apply critique-augmented training. On BLEnD, English SEM-B rises from 53.56 to 58.93, while the local-language average changes little, from 45.72 to 45.87.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.