PulseAugur
中
实时 17:03:57
English(EN) ExecRetrieval: Measuring the Functional-Correctness Gap in Code-Embedding Retrieval

新的ExecRetrieval基准揭示了代码检索中的功能正确性差距

作者 PulseAugur 编辑部 · [3 个来源] ·

引入了一个名为ExecRetrieval的新基准来衡量代码检索系统的功能正确性。该基准包含939个Python任务，每个任务都有一个正确的实现和多达四个执行验证过的、与正确代码几乎相同的错误变体。对23种密集嵌入配置和BM25的评估表明，虽然顶级系统在10个项目（exec@10 = 1.00）时实现了高召回率，但它们在1个项目（exec@1 = 0.331）时的召回率显著降低，大多数排名第一的遗漏是错误的干扰项。 AI

影响 突出了AI代理代码检索中的一个关键差距，表明当前的方法难以区分功能正确的代码和几乎相同的错误版本。

排序理由 该集群描述了一篇介绍代码检索基准的新学术论文。

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →

新的ExecRetrieval基准揭示了代码检索中的功能正确性差距

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
该集群描述了一篇介绍代码检索基准的新学术论文。
Source corroboration
3 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
paper, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
29 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [3]

  1. arXiv cs.CL TIER_1 English(EN) · Aaryan Kapoor, Md Abdullah Al Hafiz Khan ·

    ExecRetrieval：衡量代码嵌入检索中的功能正确性差距

    arXiv:2609.01865v1 Announce Type: cross Abstract: Embedding-based code retrieval is a core component of coding agents and retrieval-augmented code generation, where retrieving correct code matters more than retrieving lexically similar code. Existing code-retrieval benchmarks do …

  2. arXiv cs.IR (Information Retrieval) TIER_1 English(EN) · Md Abdullah Al Hafiz Khan ·

    ExecRetrieval：衡量代码嵌入检索中的功能正确性差距

    Embedding-based code retrieval is a core component of coding agents and retrieval-augmented code generation, where retrieving correct code matters more than retrieving lexically similar code. Existing code-retrieval benchmarks do not plant controlled, execution-verified single-ed…

  3. Hugging Face Daily Papers TIER_1 English(EN) ·

    ExecRetrieval：衡量代码嵌入检索中的功能正确性差距

    ExecRetrieval evaluates whether code embeddings can distinguish correct implementations from near-identical buggy variants, revealing that leading retrievers frequently rank incorrect near-clones above canonical solutions.

相关实体

相关话题