Associate Professor, Beijing University of Aeronautics and Astronautics
1 paper at NeurIPS 2025
A new learning framework that improves LLM inference by learning from a Mistake Log collected during fine-tuning.
