Profiles2025
Songjun Tu
Chinese Academy of Sciences
3Papers
0Last author
7Unique co-authors
2First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers3
- Learning When to Think: Shaping Adaptive Reasoning in R1-Style Models via Multi-Stage RLAuthor 1 of 7NeurIPS 2025 · 69 citations
- In-Dataset Trajectory Return Regularization for Offline Preference-based Reinforcement LearningAuthor 1 of 7AAAI 2025 · 4 citations
- Unsupervised Zero-Shot Reinforcement Learning via Dual-Value Forward-Backward RepresentationAuthor 2 of 8ICLR 2025
