R-Tuning received the Outstanding Paper Award at NAACL 2024.
Released the Eurus model and UltraInteract dataset in 2024, supporting SFT and preference learning for complex reasoning tasks.
Proposed the CodeAct agent (2024), which uses executable Python code actions to significantly improve LLM agent performance.
Papers MINT and CRAFT accepted to ICLR 2024.
Published multiple papers at top venues including ICLR, ICML, ACL, and TMLR, such as OpenHands, Executable Code Actions Elicit Better LLM Agents, and A Single Transformer for Scalable Vision-Language Modeling.