Chengqing Zong

Enough is as good as a feast: A Comprehensive Analysis of How Reinforcement Learning Mitigates Task Conflicts in LLMs

Parallel Scaling Law: Unveiling Reasoning Generalization through A Cross-Linguistic Perspective

SimulPL: Aligning Human Preferences in Simultaneous Machine Translation

Unifying Vocabulary of Large Language Model with Statistical Token-level Alignment

Language Imbalance Driven Rewarding for Multilingual Self-improving

Boosting LLM Translation Skills without General Ability Loss via Rationale Distillation