Metraipolis › archive › dc_79e99a1bd3ee
Improving Cross-Lingual Token Representations by Adding a Pinch of SALT — arXiv:2609.09953
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.09953
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T07:26:39.376Z
- SHA-256 of the archived text
aa8b880b1a772cfa8e5a30aec05e8f151c2a4a7c9a481ee9e914a65c056dc759- Stable URL of this page
https://metraipolis.com/documento/dc_79e99a1bd3ee
How to cite
Improving Cross-Lingual Token Representations by Adding a Pinch of SALT — arXiv:2609.09953 — arXiv (en), https://arxiv.org/abs/2609.09953 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_79e99a1bd3ee (Metraipolis archive dc_79e99a1bd3ee, sha256 aa8b880b1a772cfa8e5a30aec05e8f151c2a4a7c9a481ee9e914a65c056dc759)
BibTeX
@misc{dc79e99a1bd3ee,
title = {Improving Cross-Lingual Token Representations by Adding a Pinch of SALT — arXiv:2609.09953},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.09953},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_79e99a1bd3ee; sha256 aa8b880b1a772cfa8e5a30aec05e8f151c2a4a7c9a481ee9e914a65c056dc759},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.09953. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Cross-lingual sentence encoders enable scalable transfer across hundreds of languages, powering applications such as translation mining and zero-shot learning in low-resource settings. Although trained for sentence-level alignment, they are increasingly also applied to token-level tasks such as hallucination detection and sequence tagging, exposing a mismatch between training and usage. We propose SALT, a lightweight post-training method that improves token representations by injecting span-level supervision into existing sentence encoders. Across five multilingual token-level benchmarks, SALT achieves the best overall results on four of them, outperforming alternative fine-tuning strategies and competitive encoders. It also improves sentence-level performance on cross-lingual retrieval and classification tasks. These results demonstrate that span-level supervision is an effective signal for improving both token and sentence representations. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-10 in cs.AI). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.09953 — 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 2026-10-16T07:00:16.367Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).