RAG Comparison Hub

Reference Surfaces

Top Papers

• KohakuRAG: A simple RAG framework with hierarchical document indexing(8.0)

  KohakuRAG is an open-source hierarchical RAG framework that achieves state-of-the-art performance with precise citation and stable answers, ideal for technical documentation.

• FLANS at SemEval-2026 Task 7: RAG with Open-Sourced Smaller LLMs for Everyday Knowledge Across Diverse Languages and Cultures(8.0)

  Culturally aware AI-driven question-answering system for multilingual contexts using open-sourced LLMs.

• Hit-RAG: Learning to Reason with Long Contexts via Preference Alignment(7.0)

  Hit-RAG enhances retrieval-augmented generation by progressively optimizing context utilization, improving reasoning accuracy in long-context scenarios.

• A Systematic Investigation of Document Chunking Strategies and Embedding Sensitivity(7.0)

  Optimize RAG systems with our benchmarked document chunking strategies, improving retrieval accuracy and efficiency across diverse knowledge domains.

• SPD-RAG: Sub-Agent Per Document Retrieval-Augmented Generation(7.0)

  SPD-RAG is a hierarchical multi-agent framework for question answering that decomposes the problem along the document axis, improving scalability and answer quality.

• Towards Robust Retrieval-Augmented Generation Based on Knowledge Graph: A Comparative Analysis(7.0)

  GraphRAG enhances LLM responses by using knowledge graphs for more robust retrieval-augmented generation, outperforming standard RAG baselines in noisy environments.

• InfoFlow KV: Information-Flow-Aware KV Recomputation for Long Context(7.0)

  Optimize RAG for long-context question answering by intelligently recomputing key-value caches based on information flow, improving efficiency and accuracy.

• Revisiting RAG Retrievers: An Information Theoretic Benchmark(5.0)

  MIGRASCOPE enhances RAG system efficiency by providing metrics for retriever selection and ensemble configuration.

• HugRAG: Hierarchical Causal Knowledge Graph Design for RAG(5.0)

  HugRAG is a structured and scalable graph-based RAG framework explicitly modeling causal relationships for better retrieval and reasoning.

• AILS-NTUA at SemEval-2026 Task 8: Evaluating Multi-Turn RAG Conversations(3.0)

  AILS-NTUA presents a multi-turn retrieval-augmented generation system that excels in competitive evaluations.