---
格式版本: 2
标题: "Natural Language Code Retrieval for 1C:Enterprise: An Open Benchmark and Efficient Bi-Encoder"
原文链接: "https://arxiv.org/abs/2608.19957"
发布日期: "2026-08-20"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Thu, 20 Aug 2026 12:24:53 UTC (113 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-21T16:18:10+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-21T16:16:59+08:00"
入库时间: "2026-08-21T08:18:10.365Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=Google&searchtype=all"
匹配关键词:
  []
相关厂家:
  - "Google"
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 5
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "论文主题为1C:Enterprise代码检索，与超节点/AI Rack/机柜级AI基础设施完全无关，仅提及google/gemma模型，无任何相关技术或商业信息。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-21T16:18:26+08:00"
AI主题相关性: 0
AI来源权威性: 5
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
AI摘要: "研究者为1C:Enterprise构建了含3,413对真实查询-代码的开放基准，并训练了专用bi-encoder模型；模型在基准上macro nDCG@10达0.5992，且用MRL压缩到256维仍保留99.9%的检索质量。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T02:12:13.406Z"
采集批次: "2026年8月21日13点55分53秒"
采集批次ID: "20260821-135553-743"
去重键: "https://arxiv.org/abs/2608.19957"
---

## Computer Science > Computation and Language

## Title:Natural Language Code Retrieval for 1C:Enterprise: An Open Benchmark and Efficient Bi-Encoder

Authors:[Konstantin Chesnokov](https://arxiv.org/search/cs?searchtype=author&query=Chesnokov,+K), [Chingiz Mingazov](https://arxiv.org/search/cs?searchtype=author&query=Mingazov,+C)

[View PDF](https://arxiv.org/pdf/2608.19957) [HTML (experimental)](https://arxiv.org/html/2608.19957v1)

> Abstract:Natural language code retrieval is a rapidly evolving task in computer science. However, the 1C:Enterprise ecosystem combines Russian syntax with highly domain-specific terminology, for which open datasets and specialized models have been virtually non-existent. We present a comprehensive pipeline for 1C code retrieval: an open benchmark of 3,413 real-world, PII-scrubbed query-code pairs, a reproducible evaluation harness, and a specialized bi-encoder. To overcome scarce labeled data, we fine-tune on 784,057 synthetic triplets generated by google/gemma-4-26B-A4B-it from public code repositories, using Matryoshka Representation Learning (MRL) and a privacy-aware tokenizer. Because the benchmark subsets differ in size, we report balanced-subset macro, query-weighted micro, and forum-only results. Our model reaches 0.5992 balanced macro nDCG@10, 0.5044 micro, and 0.4617 on forum, versus 0.4932 macro for the baseline architecture and 0.5404 for google/embeddinggemma-300m. Removing every benchmark example flagged by the conservative exact/13-gram overlap audit leaves 0.6011 balanced macro (0.5010 micro), indicating that detected train-benchmark overlap does not explain the headline result. MRL truncation to 256 dimensions preserves 99.9% of retrieval quality while reducing dense-index storage and exact similarity arithmetic by a factor of three.

| Subjects: | Computation and Language (cs.CL) |
| --- | --- |
| Cite as: | [arXiv:2608.19957](https://arxiv.org/abs/2608.19957) \[cs.CL\] |
|  | (or [arXiv:2608.19957v1](https://arxiv.org/abs/2608.19957v1) \[cs.CL\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.19957](https://doi.org/10.48550/arXiv.2608.19957) |

## Submission history

From: Konstantin Chesnokov Antonovich \[[view email](https://arxiv.org/show-email/ffd4649b/2608.19957)\]  
**\[v1\]** Thu, 20 Aug 2026 12:24:53 UTC (113 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.19957) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))