PhD student, Max Planck Institute for Software Systems
1 paper at NeurIPS 2025
We propose a novel inference-time personalized alignment method that elicits the user's preferences with a few preference queries.
