Metraipolis › archive › dc_d42952858bb7

Safety Alignment Illusion: The Cross-Lingual Safety Gap in LLMs — arXiv:2608.18131

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.18131
License (as archived)
CC0-1.0
Archived
2026-10-18T13:44:18.815Z
SHA-256 of the archived text
9f959a26f500562ad1bbda5a4b91db8a0615791a576ac379c91d24be201bbffe
Stable URL of this page
https://metraipolis.com/documento/dc_d42952858bb7

How to cite

Safety Alignment Illusion: The Cross-Lingual Safety Gap in LLMs — arXiv:2608.18131 — arXiv (en), https://arxiv.org/abs/2608.18131 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d42952858bb7 (Metraipolis archive dc_d42952858bb7, sha256 9f959a26f500562ad1bbda5a4b91db8a0615791a576ac379c91d24be201bbffe)
BibTeX
@misc{dcd42952858bb7,
  title = {Safety Alignment Illusion: The Cross-Lingual Safety Gap in LLMs — arXiv:2608.18131},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.18131},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d42952858bb7; sha256 9f959a26f500562ad1bbda5a4b91db8a0615791a576ac379c91d24be201bbffe},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.18131. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Current safety alignment training for Large Language Models (LLMs) are heavily English-centric. When such safety filters fail for non-English languages, the consequences are immediate and user-facing: voice assistants and spoken dialogue systems may produce stereotype-reinforcing outputs, bypassing the standard English-focused safety alignments and propagating harmful bias to non-English speaking communities. For spoken language technologies deployed across India's linguistically diverse population, this represents a critical failure mode. To address this cross-lingual gap, we introduce INCLUDE (Indian Cultural Lens for Understanding and Detecting Embedded Biases), a multilingual evaluation benchmark designed to quantify Indian-centric socio-cultural biases. INCLUDE consists of 2,604 prompts spanning six prompt languages: English, Hindi, Bengali, Marathi, Tamil, and Hinglish (Hindi-English code-mix). We evaluate ten open- and closed-source LLMs against this benchmark, analyzing 14,988 bias scores. Our statistical results reveal two key findings. First, Bengali yielded the highest average bias score in open-source models. Second, English demonstrated a notable reversal, producing the lowest bias in open-source models but the highest bias in closed-source models.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-20 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Computers and Society (cs.CY).
Pagina del lavoro: https://arxiv.org/abs/2608.18131 — 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 2026-10-18T13:41:17.842Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-20?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).