AI RESEARCH
FAB-Bench: A Framework for Adaptive RAG Benchmarking in Semiconductor Manufacturing
arXiv CS.CL
•
ArXi:2605.26476v1 Announce Type: new Retrieval-Augmented Generation (RAG) has become critical for knowledge-intensive applications, yet evaluating its performance in vertical domains remains difficult due to domain complexity, diverse context scales, and heavy reliance on expert assessments that are costly, inconsistent, and non-scalable. We