重新思考代码语言模型中的枢轴编程语言
Rethinking Pivot Programming Languages in Code Language Models
- Ulm University（乌尔姆大学）
- University of Southern Denmark（南丹麦大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本研究在控制混淆因素下重新审视代码语言模型中的枢轴编程语言，发现Python的特殊角色源于其面向英语的亲和性，而非普遍的几何中心性。
AI中文摘要：
多语言代码语言模型在编程语言（PL）之间迁移技能，但是否存在某种编程语言占据特权枢轴位置仍存争议：几何分析指向C族语言和Go，而行为证据则强调Python。我们在控制表征各向异性和跨PL长度变化（这两个混淆因素损害了先前基于余弦的分析）的条件下重新审视这一问题。在三个代码模型上，针对多语言竞争性编程数据，我们研究了跨PL组织的三个视角：成对PL几何、PL-英语对齐，以及通过候选PL表示空间的枢轴检索。结果是关系依赖的。代码-代码几何揭示了结构化的语言区域，但无普遍中心；代码-英语对齐偏向高级脚本语言；而枢轴检索在代码到代码和英语到代码迁移中偏好不同的中间空间。这些发现表明，Python的特殊角色更应被理解为面向英语的亲和性，而非普遍的几何中心性。
英文摘要：
Multilingual code language models transfer skills across programming languages (PLs), but whether any PL occupies a privileged pivot position remains contested: geometric analyses point to C-family languages and Go, while behavioral evidence highlights Python. We revisit this question under controls for representational anisotropy and length variation across PLs, two confounds that compromise prior cosine-based analyses. Across three code models on multilingual competitive-programming data, we study three views of cross-PL organization: pairwise PL geometry, PL-English alignment, and pivoted retrieval through candidate PL representation spaces. The results are relation-dependent. Code-code geometry reveals structured language regions but no universal center; code-English alignment favors high-level scripting languages; and pivoted retrieval favors different intermediate spaces for code-to-code and English-to-code transfer. These findings suggest that Python's special role is better understood as English-facing affinity than as universal geometric centrality.
