Revisiting Prompt Engineering: A Comprehensive Evaluation for LLM-based Personalized Recommendation

Abstract

This work revisits the role of prompt engineering in personalized recommendation tasks. The paper comprehensively evaluates various LLM-based prompting strategies for user preference modeling and shows that their effectiveness depends heavily on data context and personalization granularity.

Publication
Proceedings of the 19th ACM Conference on Recommender Systems (RecSys 2025)