ContextRAG constructs extraction-free hierarchical graphs via residual-quantization k-means and Formal Concept Analysis with Lukasiewicz residuated logic on embeddings, using 30 LLM calls and 22k tokens to reach 33.6% F1 on a 130-task UltraDomain subset.
arXiv preprint , volume =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
ContextRAG: Extraction-Free Hierarchical Graph Construction for Retrieval-Augmented Generation
ContextRAG constructs extraction-free hierarchical graphs via residual-quantization k-means and Formal Concept Analysis with Lukasiewicz residuated logic on embeddings, using 30 LLM calls and 22k tokens to reach 33.6% F1 on a 130-task UltraDomain subset.