TPO: Aligning Large Language Models with Multi-branch & Multi-step Preference Trees.

Weibin Liao, Xu Chu, Yasha Wang

Browse the full ICLR paper archive.