SigMap Benchmark Suite: 240-Repository Large-Scale AI Context Extraction Dataset
收藏Zenodo2026-04-29 更新2026-05-26 收录
下载链接：
https://zenodo.org/doi/10.5281/zenodo.19898842
下载链接
链接失效反馈官方服务：
资源简介：
The SigMap Benchmark Suite presents a comprehensive evaluation of AI context extraction across 240 diverse open-source repositories spanning 30+ programming languages. This dataset comprises 1,775 benchmark operations capturing token reduction metrics, execution performance, and code complexity analysis.
Key Features:• 240 repositories across 30+ languages• 1,775 benchmark operations (5 modes per repository)• 50+ metadata fields per repository• 96.2% average token reduction• Complete reproducibility package• 4 data export formats (CSV, JSON, JSONL, SQL)
The dataset enables analysis of language-specific context extraction patterns, monorepo complexity, domain-specific compression characteristics, and AI context optimization strategies.
Complete methodology, reproducibility materials, and scripts are included.
提供机构：
Zenodo创建时间：
2026-04-29
