2025
SAKI-RAG: Mitigating Context Fragmentation in Long-Document RAG via Sentence-level Attention Knowledge Integration
EMNLP 2025
Traditional Retrieval-Augmented Generation (RAG) frameworks often segment documents into larger chunks to preserve contextual coherence, inadvertently introducing redundant noise. Recent advanced RAG frameworks have shifted toward finer-grained chunking to improve precision. However, in long-documen