Evaluating Retrieval-Augmented Generation Systems on Unanswerable, Uncheatable, Realistic, Multi-hop Queries
This addresses the need for more challenging and realistic benchmarks to uncover limitations in RAG systems, though it is incremental as it builds on existing datasets and methods.
The authors tackled the problem of evaluating retrieval-augmented generation (RAG) systems on complex, realistic queries by introducing a pipeline to automatically create uncheatable, realistic, unanswerable, and multi-hop queries (CRUMQs), which reduced cheatability scores by up to 81.0% compared to prior benchmarks.
Real-world use cases often present RAG systems with complex queries for which relevant information is missing from the corpus or is incomplete. In these settings, RAG systems must be able to reject unanswerable, out-of-scope queries and identify failures of retrieval and multi-hop reasoning. Despite this, existing RAG benchmarks rarely reflect realistic task complexity for multi-hop or out-of-scope questions, which often can be cheated via disconnected reasoning (i.e., solved without genuine multi-hop inference) or require only simple factual recall. This limits the ability for such benchmarks to uncover limitations of existing RAG systems. To address this gap, we present the first pipeline for automatic, difficulty-controlled creation of un$\underline{c}$heatable, $\underline{r}$ealistic, $\underline{u}$nanswerable, and $\underline{m}$ulti-hop $\underline{q}$uerie$\underline{s}$ (CRUMQs), adaptable to any corpus and domain. We use our pipeline to create CRUMQs over two popular RAG datasets and demonstrate its effectiveness via benchmark experiments on leading retrieval-augmented LLMs. Results show that compared to prior RAG benchmarks, CRUMQs are highly challenging for RAG systems and achieve up to 81.0\% reduction in cheatability scores. More broadly, our pipeline offers a simple way to enhance benchmark difficulty and realism and drive development of more capable RAG systems.