TY  - RPRT
TI  - Design Conditions for Intra-Group Learning of Sequence-Level Rewards: Token Gradient Cancellation
AU  - Fei Ding
AU  - Yongkang Zhang
AU  - youwei wang
AU  - Zijian Zeng
PY  - 2026
UR  - https://arxiv.org/abs/2604.13088
ID  - 2604.13088
ER  -