Reproduce “Discovering Lexical Gaps Using Embeddings from Multilingual LLMs” 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
Yoonwon Jung, Aaron S. Cohen, Benjamin K. Bergen
The authors develop a method to rank possible lexical gaps, concepts without a single-word equivalent, using multilingual word representations. It compares each word with its nearest neighbor in another language, treating weaker matches as candidates for expert review. In Korean–English emotion adjectives, 94% of Korean-source and 97% of English-source embedding spaces separated known gaps above chance; generalization remains untested.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.