Qianchi Zhang

Qianchi Zhang, Hainan Zhang, Liang Pang et al.

Retrieval-Augmented Generation (RAG) has become a key paradigm for reducing factual hallucinations in Large Language Models (LLMs), yet little is known about how the order of retrieved documents affects model behavior. We empirically show that under a Top-5 retrieval setting with the gold document included, LLM answers vary substantially across permutations of the retrieved set, even when the gold document is fixed in the first position. This reveals a previously underexplored sensitivity to retrieval permutations. Although existing robust RAG methods focus primarily on enhancing LLM robustness to low-quality retrieval and mitigating positional bias to distribute attention fairly over long contexts, neither approach directly addresses permutation sensitivity. In this paper, we propose Stable-RAG, which exploits permutation sensitivity estimation to mitigate permutation-induced hallucinations. Stable-RAG runs the generator under multiple retrieval orders, clusters hidden states, and decodes from a cluster-center representation that captures the dominant reasoning pattern. It then uses these reasoning results to align hallucinated outputs toward the correct answer, encouraging the model to produce consistent and accurate predictions across document permutations. Experiments on three QA datasets show that Stable-RAG improves answer accuracy, reasoning consistency, and generalization across datasets, retrievers, and input lengths compared with strong baselines.

Qianchi Zhang, Hainan Zhang, Liang Pang et al.

Retrieved documents containing noise will hinder RAG from detecting answer clues and make the inference process slow and expensive. Therefore, context compression is necessary to enhance its accuracy and efficiency. Existing context compression methods use extractive or generative models to retain the most query-relevant sentences or apply the information bottleneck theory to preserve sufficient information. However, these methods may face issues such as over-compression or high computational costs. We observe that the retriever often ranks relevant documents at the top, but the exact number of documents needed to answer the query is uncertain due to the impact of query complexity and retrieval quality: complex queries like multi-hop questions may require retaining more documents than simpler queries, and a low-quality retrieval may need to rely on more documents to generate accurate outputs. Therefore, determining the minimum number of required documents (compression rate) is still a challenge for RAG. In this paper, we introduce AdaComp, a low-cost extractive context compression method that adaptively determines the compression rate based on both query complexity and retrieval quality. Specifically, we first annotate the minimum top-k documents necessary for the RAG system to answer the current query as the compression rate and then construct triplets of the query, retrieved documents, and its compression rate. Then, we use this triplet dataset to train a compression-rate predictor. Experiments on three QA datasets and one conversational Muiti-doc QA dataset show that AdaComp significantly reduces inference costs while maintaining performance nearly identical to uncompressed models, achieving a balance between efficiency and performance.

Qianchi Zhang, Hainan Zhang, Liang Pang et al.

Current RAG retrievers are designed primarily for human readers, emphasizing complete, readable, and coherent paragraphs. However, LLMs benefit more from precise, compact, and well-structured input, which enhances reasoning quality and efficiency. Existing methods often rely on reranking or summarization to identify key sentences, but may suffer from semantic breaks and unfaithfulness. Thus, efficiently extracting and organizing answer-relevant clues from large-scale documents while reducing LLM reasoning costs remains a challenge for RAG. Inspired by Occam's razor, we frame LLM-centric retrieval as a MinMax optimization: maximizing the extraction of potential clues and reranking them for well-organization, while minimizing reasoning costs by truncating to the smallest sufficient clues set. In this paper, we propose CompSelect, a Compact clue Selection mechanism for LLM-centric RAG, consisting of a clue extractor, a reranker, and a truncator. (1) The clue extractor first uses answer-containing sentences as fine-tuning targets, aiming to extract sufficient potential clues; (2) The reranker is trained to prioritize effective clues based on real LLM feedback; (3) The truncator uses the truncated text containing the minimum sufficient clues for answering the question as fine-tuning targets, thereby enabling efficient RAG reasoning. Experiments on three QA datasets show that CompSelect improves QA performance by approximately 11\% and reduces Total Latency and Online Latency by approximately 17\% and 67\% compared to various baseline methods on both LLaMA3 and Qwen3. Further analysis confirms its robustness to unreliable retrieval and generalization across different scenarios, offering a scalable and cost-efficient solution for web-scale RAG applications.