Minionese: Comprehensive Benchmark and Mechanistic Study of Multilingual LLM Safety

Abstract

Safety alignment in large language models remains brittle across languages: prompts reliably refused in English can elicit harmful compliance in non-English and low-resource settings. We introduce \textsc{Minionese}, a multilingual jailbreak benchmark spanning 18 languages, 4 resource tiers, and 4 perturbation types (standard translation, code-switching, transliteration, and translationese), paired with a geometric mechanistic analysis of refusal failure across language tiers. We show that each

Categories

Framework mappings

Suggested from the entry's categories.

@misc{ifebi2026minionese,
  title = {{Minionese: Comprehensive Benchmark and Mechanistic Study of Multilingual LLM Safety}},
  author = {Chigozirim Ifebi and Brent Kong and Ayushi Mehrotra},
  year = {2026},
  month = jul,
  eprint = {2607.10112},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2607.10112}
}