TY  - RPRT
TI  - Eliminating Biased Length Reliance of Direct Preference Optimization via Down-Sampled KL Divergence
AU  - Junru Lu
AU  - Jiazheng Li
AU  - Siyu An
AU  - Meng Zhao
AU  - Yulan He
AU  - Di Yin
AU  - Xing Sun
PY  - 2024
UR  - https://arxiv.org/abs/2406.10957
ID  - 2406.10957
ER  -