TY  - RPRT
TI  - A Gradient Analysis Framework for Rewarding Good and Penalizing Bad Examples in Language Models
AU  - Yi-Lin Tuan
AU  - William Yang Wang
PY  - 2024
UR  - https://arxiv.org/abs/2408.16751
ID  - 2408.16751
ER  -