Preprint
Aug 2026
Training-Free LLM-Based Recommendation with Post-LLM Item Refinement Using Collaborative Signals
CoRRe is proposed, a training-free recommendation framework with a post-LLM paradigm that injects CF signals into LLM-generated item representations, which are later matched with LLM-generated user interests for ranking.
Kyungho Kim, Sunwoo Kim, Ge-On Lee et al.
· 0 citations