Test-Time Preference Optimization: On-the-Fly Alignment via Iterative Textual Feedback.

Yafu Li, Xuyang Hu, Xiaoye Qu, Linjie Li, Yu Cheng

Browse the full ICML paper archive.