One doc tagged with "llm-safety"

Why safety filters trained on English fail against Russian, Chinese, or low-resource languages. Understanding Cross-Lingual Attacks.