Niclas Hart, Nicolas Weeger · 2026 · arXiv
DOI: 10.48550/arXiv.2610.27816 · open-access PDF
Retrieval-Augmented Generation (RAG) tornou-se o paradigma dominante para fundamentar as saídas dos grandes modelos de linguagem (LLM) no conhecimento externo. No entanto, os praticantes enfrentam um vasto espaço de configuração que inclui chunking, embedding, retrieval, reranking, prompting e modelos de geração, e há orientação limitada sobre como esses componentes interagem. Nós apresentamos RAG-NIMBF, uma estrutura de benchmarking modular que se liga aos tubos RAG existentes como uma camada de avaliação não invasiva. Ele avalia o espaço de configuração através de ablação controlada com rastreamento experimental de ponta a ponta.
Machine translation (Meta M2M100); the English original follows.
Retrieval-Augmented Generation (RAG) has become the dominant paradigm for grounding the outputs of large language models (LLM) in external knowledge. However, practitioners are faced with a vast configuration space comprising chunking, embedding, retrieval, reranking, prompting and generation models, and there is limited guidance on how these components interact. We present RAG-NIMBF, a modular benchmarking framework that attaches to existing RAG pipelines as a non-invasive evaluation layer. It evaluates the configuration space through controlled ablation with end-to-end experiment tracking. T
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.27816, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange