Safety tuning on one region's preference data does not transfer cleanly; this measures how global versus local preference sources change harm rates when a model serves users in many languages.

• articleBam Just Like That Simple And Efficient Parameter Upcycling For Mixture Of Experts 2024 08 15
• articleMultilingual Arbitrage Optimizing Data Pools To Accelerate Multilingual Progress 2024 08 28
• articleOne Tokenizer To Rule Them All Emergent Language Plasticity Via Multilingual Tokenizers 2025 05 30
• articleInclude Evaluating Multilingual Language Understanding With Regional Knowledge 2024 11 29

• articleRlhf Can Speak Many Languages Unlocking Multilingual Preference Optimization For Llms 2024 07 05Cohere editorial sitemap
• articleROPD: Routing-based On-Policy Distillation for LLM Safety RealignmentYongjian Guo, Wanlun Ma, Lingyu Shen, Xi Xiao, Sheng Wen
• articleOptimismBench: Forecasting Bias and the Alignment Effect in Language Model JudgmentSeonglae Cho, Adriano Koshiyama

Checking sign-in…

Loading comments…