A Survey on Personalized and Pluralistic Preference Alignment in Large Language Models
By: Zhouhang Xie , Junda Wu , Yiran Shen and more
Potential Business Impact:
Makes AI understand what you like best.
Personalized preference alignment for large language models (LLMs), the process of tailoring LLMs to individual users' preferences, is an emerging research direction spanning the area of NLP and personalization. In this survey, we present an analysis of works on personalized alignment and modeling for LLMs. We introduce a taxonomy of preference alignment techniques, including training time, inference time, and additionally, user-modeling based methods. We provide analysis and discussion on the strengths and limitations of each group of techniques and then cover evaluation, benchmarks, as well as open problems in the field.
Similar Papers
A Survey on Personalized Alignment -- The Missing Piece for Large Language Models in Real-World Applications
Computation and Language
Teaches AI to be helpful and kind, your way.
From 1,000,000 Users to Every User: Scaling Up Personalized Preference for User-level Alignment
Computation and Language
Teaches AI to understand what *you* want.
Aligning Multimodal LLM with Human Preference: A Survey
CV and Pattern Recognition
Makes AI understand pictures and sounds better.