TY  - RPRT
TI  - Can RLHF be More Efficient with Imperfect Reward Models? A Policy Coverage Perspective
AU  - Jiawei Huang
AU  - Bingcong Li
AU  - Christoph Dann
AU  - Niao He
PY  - 2025
UR  - https://arxiv.org/abs/2502.19255
ID  - 2502.19255
ER  -