The paper proposes an item-aware attention mechanism with intra-item and inter-item layers to let LLMs capture item-level collaborative relations instead of only token-level ones.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
UniDisc learns cross-modal representations from data lake context using a heterogeneous graph to unify natural language and table-based data discovery across diverse intents.
citing papers explorer
-
Beyong Tokens: Item-aware Attention for LLM-based Recommendation
The paper proposes an item-aware attention mechanism with intra-item and inter-item layers to let LLMs capture item-level collaborative relations instead of only token-level ones.
-
Unified Data Discovery across Query Modalities and User Intents
UniDisc learns cross-modal representations from data lake context using a heterogeneous graph to unify natural language and table-based data discovery across diverse intents.