SPPD: Self-training with Process Preference Learning Using Dynamic Value Margin.

Hao Yi, Qingyang Li, Yulan Hu, Fuzheng Zhang, Di Zhang, Yong Liu

Browse the full EMNLP paper archive.