aiqbench analysis
This arXiv paper proposes SaraCoder, a retrieval-augmented method that improves repository-level code completion by pruning redundant retrieved code and selectively adding cross-file dependency definitions within a fixed context budget. The core claim is that the bottleneck is often the low information density of the context, not model capacity, and the reported gains are relative to baselines such as GraphCoder and RepoCoder on CrossCodeEval and RepoEval-Updated. Readers should treat these as benchmark results from a preprint, not yet independently validated, but the budget-aware context construction idea applies to any RAG system working under tight token limits.
Enhancing Knowledge Density via Hierarchical Context Pruning for Retrieval-Augmented Repository-Level Code Completion
arXiv:2508.10068v3 Announce Type: replace-cross Abstract: Retrieval-Augmented Generation (RAG) has advanced repository-level code completion by retrieving external code snippets, API documentation, or type definitions to expand the model's contextual awareness. However, under the finite token budget of the context window, existing retrieval-augmented methods often saturate the budget with highly similar or even redundant code, which not only wastes tokens but also suppresses critical cross-file dependencies, preventing synergistic interaction between semantically similar reference snippets and cross-file dependency definitions. Therefore, further gains are limited by low knowledge density in the retrieved context rather than by model capacity. We propose SaraCoder, a retrieval-augmented framework that enhances knowledge density via hierarchical context pruning under a fixed context budget. A Hierarchical Feature Optimization module prunes an over-complete candidate pool in four stages. The freed tokens are then allocated to an External-Aware Identifier Disambiguator, which selectively injects only the most necessary cross-file dependency definitions. SaraCoder transforms context construction into a budget-aware optimization problem, balancing reference snippets and cross-file dependencies to build a context that is both knowledge-dense and semantically complete. Experiments on CrossCodeEval and RepoEval-Updated, across multiple backbone code LLMs, show that SaraCoder consistently outperforms strong RAG baselines such as GraphCoder and Repocoder in both code-level and identifier-level metrics, without increasing token usage, and yields synergistic gains when combined with other cross-file methods.
Abstract:Retrieval-Augmented Generation (RAG) has advanced repository-level code completion by retrieving external code snippets, API documentation, or type definitions to expand the model's contextual awareness. However, under the finite token budget of the context window, existing retrieval-augmented methods often saturate the budget with highly similar or even redundant code, which not only wastes tokens but also suppresses critical cross-file dependencies, preventing synergistic interaction between semantically similar reference snippets and cross-file dependency definitions. Therefore, further gains are limited by low knowledge density in the retrieved context rather than by model capacity. We propose SaraCoder, a retrieval-augmented framework that enhances knowledge density via hierarchical context pruning under a fixed context budget. A Hierarchical Feature Optimization module prunes an over-complete candidate pool in four stages. The freed tokens are then allocated to an External-Aware Identifier Disambiguator, which selectively injects only the most necessary cross-file dependency definitions. SaraCoder transforms context construction into a budget-aware optimization problem, balancing reference snippets and cross-file dependencies to build a context that is both knowledge-dense and semantically complete. Experiments on CrossCodeEval and RepoEval-Updated, across multiple backbone code LLMs, show that SaraCoder consistently outperforms strong RAG baselines such as GraphCoder and Repocoder in both code-level and identifier-level metrics, without increasing token usage, and yields synergistic gains when combined with other cross-file methods.
Software Engineering (cs.SE); Computation and Language (cs.CL); Information Retrieval (cs.IR); Programming Languages (cs.PL)
arXiv:2508.10068 [cs.SE]
(or arXiv:2508.10068v3 [cs.SE] for this version)
https://doi.org/10.48550/arXiv.2508.10068 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 313 words
Related coverage
Core-based Hierarchies for Efficient GraphRAG
arXiv · Jun 3, 2026
A Unified Framework for Context-Aware and Relation-Aware Graph Retrieval-Augmented Generation
arXiv · Jun 17, 2026
CacheRAG: A Semantic Caching System for Retrieval-Augmented Generation in Knowledge Graph Question Answering
arXiv · Jun 9, 2026
CacheRAG: A Semantic Caching System for Retrieval-Augmented Generation in Knowledge Graph Question Answering
arXiv · Jun 3, 2026