Chunk Twice, Embed Once: A Systematic Study of Segmentation and Representation Trade-offs in Chemistry-Aware Retrieval-Augmented Generation

By Mahmoud Amiri, Thomas Bocklitz. 2025

Subfields: Retrieval-Augmented Generation, Information Retrieval & Search, Computational Chemistry & Drug Discovery

Citations in the corpus

• ChEmbed: Enhancing Chemical Literature Search Through Domain-Specific Text Embeddings (2025, unknown): “However, recent evaluations on chemical RAG systems [27] indicate that these older domain-specific models are frequently outperformed by modern generalist retrievers.”
• Chunking Is Not the Bottleneck: Why RAG Evaluation Must Look beyond Segmentation Strategy (2026, unknown): “Amiri and Bocklitz (2025) systematically evaluate 25 chunking configurations across 48 embedding models on chemistry-specific RAG benchmarks and find that chunking strategy has a substantial impact on retrieval quality in that domain.”