%0 Conference Paper %1 volker-etal-2025-salt %A Völker, Tom %A Pfister, Jan %A Hotho, Andreas %B Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025) %C Vienna, Austria %D 2025 %E Rosenthal, Sara %E Rosá, Aiala %E Ghosh, Debanjan %E Zampieri, Marcos %I Association for Computational Linguistics %K myown SALT SemEval LLM-free entity-aware-translation machine-translation SQL from:tomvoelker %P 852--864 %T SALT at SemEval-2025 Task 2: A SQL-based Approach for LLM-Free Entity-Aware-Translation %U https://aclanthology.org/2025.semeval-1.117/ %X Entity-aware machine translation faces significant challenges when translating culturally-adapted named entities that require knowledge beyond the source text. We present SALT (SQL-based Approach for LLM- Free Entity-Aware-Translation), a parameter-efficient system for the SemEval-2025 Task 2. Our approach combines SQL-based entity retrieval with constrained neural translation via logit biasing and explicit entity annotations. Despite its simplicity, it achieves state-of-the-art performance (First Place) among approaches not using gold-standard data, while requiring far less computation than LLM-based methods. Our ablation studies show simple SQL- based retrieval rivals complex neural models, and strategic model refinement outperforms increased model complexity. SALT offers an alternative to resource- intensive LLM-based approaches, achieving comparable results with only a fraction of the parameters. %@ 979-8-89176-273-2

@inproceedings{volker-etal-2025-salt, abstract = {Entity-aware machine translation faces significant challenges when translating culturally-adapted named entities that require knowledge beyond the source text. We present SALT (SQL-based Approach for LLM- Free Entity-Aware-Translation), a parameter-efficient system for the SemEval-2025 Task 2. Our approach combines SQL-based entity retrieval with constrained neural translation via logit biasing and explicit entity annotations. Despite its simplicity, it achieves state-of-the-art performance (First Place) among approaches not using gold-standard data, while requiring far less computation than LLM-based methods. Our ablation studies show simple SQL- based retrieval rivals complex neural models, and strategic model refinement outperforms increased model complexity. SALT offers an alternative to resource- intensive LLM-based approaches, achieving comparable results with only a fraction of the parameters.}, added-at = {2026-04-08T06:09:25.000+0200}, address = {Vienna, Austria}, author = {V{"o}lker, Tom and Pfister, Jan and Hotho, Andreas}, biburl = {https://www.bibsonomy.org/bibtex/2daff7d970b278774e0eb69ab3de6c4ba/dmir}, booktitle = {Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025)}, editor = {Rosenthal, Sara and Ros{\'a}, Aiala and Ghosh, Debanjan and Zampieri, Marcos}, interhash = {2a76d26d7cdf7c2e2284e3f4ce95021c}, intrahash = {daff7d970b278774e0eb69ab3de6c4ba}, isbn = {979-8-89176-273-2}, keywords = {myown SALT SemEval LLM-free entity-aware-translation machine-translation SQL from:tomvoelker}, month = {07}, pages = {852--864}, publisher = {Association for Computational Linguistics}, timestamp = {2026-04-08T06:09:25.000+0200}, title = {{SALT} at {S}em{E}val-2025 Task 2: A {SQL}-based Approach for {LLM}-Free Entity-Aware-Translation}, url = {https://aclanthology.org/2025.semeval-1.117/}, year = 2025 }