Profiles2022-2026
Rui Zheng
Fudan University
17Papers
0Last author
18Unique co-authors
3First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers17
- BAPO: Stabilizing Off-Policy Reinforcement Learning for LLMs via Balanced Policy Optimization with Adaptive ClippingAuthor 17 of 21ICLR 2026 · 54 citations
- Critique-RL: Training Language Models For Critiquing Through Two-Stage Reinforcement LearningAuthor 15 of 18ICLR 2026 · 4 citations
- AgentGym-RL: An Open-Source Framework to Train LLM Agents for Long-Horizon Decision Making via Multi-Turn RLAuthor 8 of 24ICLR 2026
- MetaAct-RL: Training Language Models for Reasoning Through Meta-Action-Based Reinforcement LearningAuthor 17 of 20AAAI 2026
- What Makes a Good Speech Tokenizer for LLM-Centric Speech Generation? A Systematic StudyAuthor 16 of 22AAAI 2026
- Training Large Language Models for Reasoning through Reverse Curriculum Reinforcement LearningAuthor 5 of 21ICML 2024 · 70 citations
- Improving Generalization of Alignment with Human Preferences through Group Invariant LearningAuthor 1 of 12ICLR 2024 · 25 citations
- Uncertainty Aware Learning for Language Model AlignmentAuthor 2 of 6ACL 2024
- Self-Polish: Enhance Reasoning in Large Language Models via Problem RefinementAuthor 4 of 8EMNLP Findings 2023 · 16 citations
- CASN: Class-Aware Score Network for Textual Adversarial DetectionAuthor 2 of 5ACL 2023 · 3 citations
- Detecting Adversarial Samples through Sharpness of Loss LandscapeAuthor 1 of 9ACL Findings 2023 · 3 citations
- Characterizing the Impacts of Instances on RobustnessAuthor 1 of 10ACL Findings 2023 · 1 citation
- Modeling the Q-Diversity in a Min-max Play Game for Robust OptimizationAuthor 2 of 5ACL Findings 2023 · 1 citation
- RealBehavior: A Framework for Faithfully Characterizing Foundation Models' Human-like Behavior MechanismsAuthor 2 of 10EMNLP Findings 2023 · 1 citation
- Connectivity Patterns are Task EmbeddingsAuthor 2 of 9ACL Findings 2023
- Flooding-X: Improving BERT's Resistance to Adversarial Attacks via Loss-Restricted Fine-TuningAuthor 2 of 10ACL 2022 · 35 citations
- Efficient Adversarial Training with Robust Early-Bird TicketsAuthor 2 of 5EMNLP 2022 · 7 citations
