Skill-RAG: Failure-State-Aware Retrieval Augmentation via Hidden-State Probing and Skill Routing

TL;DR

Skill-RAG introduces a failure-aware retrieval augmentation framework that diagnoses and corrects query-evidence misalignments using a skill router and hidden-state probing, significantly improving accuracy on challenging benchmarks.

Contribution

It proposes a novel failure-state-aware retrieval method with a skill router and hidden-state prober to address query-evidence misalignment in RAG models.

Findings

Substantially improves accuracy on hard cases in open-domain QA.

Achieves strong gains on out-of-distribution datasets.

Shows that failure states occupy structured, separable regions in representation space.

Abstract

Retrieval-Augmented Generation (RAG) has emerged as a foundational paradigm for grounding large language models in external knowledge. While adaptive retrieval mechanisms have improved retrieval efficiency, existing approaches treat post-retrieval failure as a signal to retry rather than to diagnose -- leaving the structural causes of query-evidence misalignment unaddressed. We observe that a significant portion of persistent retrieval failures stem not from the absence of relevant evidence but from an alignment gap between the query and the evidence space. We propose Skill-RAG, a failure-aware RAG framework that couples a lightweight hidden-state prober with a prompt-based skill router. The prober gates retrieval at two pipeline stages; upon detecting a failure state, the skill router diagnoses the underlying cause and selects among four retrieval skills -- query rewriting, question decomposition, evidence focusing, and an exit skill for truly irreducible cases -- to correct misalignment before the next generation attempt. Experiments across multiple open-domain QA and complex reasoning benchmarks show that Skill-RAG substantially improves accuracy on hard cases persisting after multi-turn retrieval, with particularly strong gains on out-of-distribution datasets. Representation-space analyses further reveal that the proposed skills occupy structured, separable regions of the failure state space, supporting the view that query-evidence misalignment is a typed rather than monolithic phenomenon.