Fetching the paper…

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