Metraipolis › archive › dc_7b9ded796d6d

Predicting Multilingual Classification and Translation Performance of LLMs with Cross-Lingual Alignment -- Is English Enough? — arXiv:2608.03446

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.03446
License (as archived)
CC0-1.0
Archived
2026-10-31T12:31:03.077Z
SHA-256 of the archived text
7b4c4fba49887a5e65ba97b7dd3810ef05be987836ee406799a9bacf896d4e61
Stable URL of this page
https://metraipolis.com/documento/dc_7b9ded796d6d

How to cite

Predicting Multilingual Classification and Translation Performance of LLMs with Cross-Lingual Alignment -- Is English Enough? — arXiv:2608.03446 — arXiv (en), https://arxiv.org/abs/2608.03446 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7b9ded796d6d (Metraipolis archive dc_7b9ded796d6d, sha256 7b4c4fba49887a5e65ba97b7dd3810ef05be987836ee406799a9bacf896d4e61)
BibTeX
@misc{dc7b9ded796d6d,
  title = {Predicting Multilingual Classification and Translation Performance of LLMs with Cross-Lingual Alignment -- Is English Enough? — arXiv:2608.03446},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.03446},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7b9ded796d6d; sha256 7b4c4fba49887a5e65ba97b7dd3810ef05be987836ee406799a9bacf896d4e61},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.03446. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multilingual large language models (LLMs) have been shown to perform better on non-English classification tasks when the representations of the given language are more aligned to English within the model. Several cross-lingual alignment (CLA) scores have been proposed for use with LLMs, along with multiple approaches for extracting embeddings from the models. We provide a comparative analysis of 27 CLA score variants, examining how they differ and how well each predicts downstream performance across three tasks. Crucially, while LLMs are widely used for generative tasks such as machine translation, prior work has focused almost exclusively on classification. We therefore investigate whether CLA scores are similarly predictive of translation performance. To enable computing correlations across target languages, we propose a PMI-based translation metric, which is less dependent on the target language and correlates strongly with chrF. We find that CLA with English predicts translation quality comparably to or better than source-target CLA, providing new evidence that LLMs use English as an internal pivot language.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-05 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.03446 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T12:17:37.758Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).