Bo Wang (王博) - Homepage Enhancing GraphRAG with Syntactic Knowledge and Mixture-of-Experts for Knowledge-Intensive QA
Aligning LLMs by Predicting Preferences from User Writing Samples However, existing methods often produce generic preference descriptions that fail to capture the unique and individualized nature of human preferences This paper introduces PROSE, a method designed to enhance the precision of preference descriptions inferred from user writing samples
Aligning LLMs with Individual Preferences via Interaction . . . - GitHub Our approach involves establishing a diverse pool of 3,310 distinct user personas using iterative self-generation Then guided by these personas, we leverage multi-LLMs collaboration to develop a multi-turn preference dataset containing 3K+ multi-turn conversations in tree structures
api. crossref. org We propose QueryAligner, an adaptive rewriting system implementing dynamic model-aware input transformation through architecture-specific preference modeling