One Size Does Not Fit All! Dynamic Retriever and Generator Selection for RAG
cs.IR, cs.AI
Submitted: 2026-09-15
Updated: 2026-09-15
License: http://creativecommons.org/licenses/by/4.0/
The gist: Retrieval-Augmented Generation (RAG) systems typically employ fixed retriever and generator configurations across queries, despite substantial differences in query complexity and information needs,
Terminology
Abstract
Retrieval-Augmented Generation (RAG) systems typically employ fixed retriever and generator configurations across queries, despite substantial differences in query complexity and information needs, leading to inefficient allocation of computational resources. While retrieval and generation adaptivity have been studied independently, their joint effect on end-to-end RAG performance remains underexplored. We systematically analyze how retriever and generator complexity interacts across factoid and multi-hop question answering (QA), including bridge and composition reasoning tasks. Our analysis shows that stronger retrieval generally yields larger gains than increased generation effort, but both exhibit diminishing and non-monotonic returns, indicating that higher-complexity configurations are not uniformly better across queries. Motivated by these findings, we introduce DRAG, a query-adaptive framework for selecting retriever-generator configurations. We first propose DRAG QPP, a training-free routing approach that uses Query Performance Prediction (QPP) signals to guide retriever selection and perplexity-based measures over retrieved context to guide generator selection. We further introduce DRAG SFT, a supervised routing approach that fine-tunes an LLM to jointly predict retriever-generator configurations. Across three LLM families and four QA benchmarks, achieves performance comparable to strong static RAG baselines while substantially reducing inference latency, whereas DRAG SFT consistently improves effectiveness over static and training-free adaptive baselines. Overall, DRAG demonstrates that jointly adapting retrieval and generation achieves a more favorable effectiveness-efficiency trade-off than static RAG pipelines.
Sources
- e1: Learning Adaptive Control of Reasoning Effort
- BGE Landmark Embedding: A Chunking-Free Embedding Method For Retrieval Augmented Long-Context Large Language Models
- Can QPP Choose the Right Query Variant? Evaluating Query Variant Selection for RAG Pipelines
- Retrieval-Augmented Generation for Large Language Models: A Survey
- Text Embeddings by Weakly-Supervised Contrastive Pre-training
- Retrieving and Reading: A Comprehensive Survey on Open-domain Question Answering
- Rank-R1: Enhancing Reasoning in LLM-based Document Rerankers via Reinforcement Learning
- gpt-oss-120b & gpt-oss-20b Model Card
- The Expando-Mono-Duo Design Pattern for Text Ranking with Pretrained Sequence-to-Sequence Models
- RankZephyr: Effective and Robust Zero-Shot Listwise Reranking is a Breeze!
- Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
- Scaling Down, LiTting Up: Efficient Zero-Shot Listwise Reranking with Seq2seq Encoder-Decoder Models
- Gemma 4 Technical Report
- Qwen3 Technical Report
- LLaMA: Open and Efficient Foundation Language Models
Related papers
- The Price of Isolation: Estimating the Ecosystem Cost of Symmetric Two-Sided A/B Testing
- SCAR: Semantic Continuity-Aware Retrieval for Efficient Context Expansion in RAG
- MixLoRA-DSI: Dynamically Expandable Mixture-of-LoRA Experts for Rehearsal-Free Generative Retrieval over Dynamic Corpora
- RRCM: Ranking-Driven Retrieval over Collaborative and Meta Memories for LLM Recommendation
- Right Family, Wrong Skill: Evaluating Risk Exposure in Agent Skill Retrieval
- UltRAG: a Universal Simple Scalable Recipe for Knowledge Graph RAG