We introduce Trillion-7B, the most token-efficient Korean-centric multilingual LLM available. Our novel Cross-lingual Document Attention (XLDA) mechanism enables highly efficient and effective knowledge transfer from English to target languages like Korean and Japanese. Combined with optimized data mixtures, language-specific filtering, and tailored tokenizer construction, Trillion-7B achieves competitive performance while dedicating only 10\% of its 2T training tokens to multilingual data and requiring just 59.4K H100 GPU hours ($148K) for full training. Comprehensive evaluations across 27 benchmarks in four languages demonstrate Trillion-7B's robust multilingual performance and exceptional cross-lingual consistency.
Cite
@article{arxiv.2504.15431,
title = {Trillion 7B Technical Report},
author = {Sungjun Han and Juyoung Suk and Suyeong An and Hyungguk Kim and Kyuseok Kim and Wonsuk Yang and Seungtaek Choi and Jamin Shin},
journal= {arXiv preprint arXiv:2504.15431},
year = {2025}
}