Metraipolis › archive › dc_072a86d14fdb
Are LLMs Safe Beyond Text: Do Emojis Expose Gaps in Safety Evaluation — arXiv:2608.18164
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.18164
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T13:44:54.930Z
- SHA-256 of the archived text
9d4785483df042c3417ae155cfee235a1e032a8a09f570022bb1312c6c60229c- Stable URL of this page
https://metraipolis.com/documento/dc_072a86d14fdb
How to cite
Are LLMs Safe Beyond Text: Do Emojis Expose Gaps in Safety Evaluation — arXiv:2608.18164 — arXiv (en), https://arxiv.org/abs/2608.18164 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_072a86d14fdb (Metraipolis archive dc_072a86d14fdb, sha256 9d4785483df042c3417ae155cfee235a1e032a8a09f570022bb1312c6c60229c)
BibTeX
@misc{dc072a86d14fdb,
title = {Are LLMs Safe Beyond Text: Do Emojis Expose Gaps in Safety Evaluation — arXiv:2608.18164},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.18164},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_072a86d14fdb; sha256 9d4785483df042c3417ae155cfee235a1e032a8a09f570022bb1312c6c60229c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.18164. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Safety evaluations of large language models (LLMs) predominantly rely on text-based adversarial prompts, potentially overlooking vulnerabilities arising from alternative input representations. This work examines emoji-augmented prompts as a test case for this gap, evaluating 50 prompts across four open-source LLMs (Mistral 7B, Qwen 2 7B, Gemma 2 9B, Llama 3 8B). Results show substantial variation in robustness: Gemma 2 9B and Mistral 7B exhibit non-zero success rates (10%), Llama 3 8B 6%, while Qwen 2 7B shows complete resistance (0% success rate). A chi-square test ($\chi^2 = 32.94, p < 0.001$) confirms significant differences in outcome distributions. These findings indicate that robustness is sensitive to input representation, and that evaluations restricted to standard text prompts may underrepresent model vulnerabilities. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-20 in cs.AI). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR). Pagina del lavoro: https://arxiv.org/abs/2608.18164 — 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 2026-10-18T13:41:17.916Z 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).