标题： 模型代码：文本是您需要的一切吗？ 显示英文标题

标题： Modeling Code: Is Text All You Need?

摘要： 代码大语言模型最近在建模源代码方面变得非常流行，适用于生成、翻译和总结等多种任务。 然而，基于变压器的模型在通过代码的结构化、分析属性（如控制流和数据流）进行推理方面存在局限性。 先前的研究探索了使用结构化数据和图神经网络来建模这些属性。 然而，这些方法缺乏现代大语言模型的生成能力和规模。 在本工作中，我们介绍了一种新的方法，结合建模代码作为文本和更结构化形式的优势。 显示英文摘要

摘要： Code LLMs have become extremely popular recently for modeling source code across a variety of tasks, such as generation, translation, and summarization. However, transformer-based models are limited in their capabilities to reason through structured, analytical properties of code, such as control and data flow. Previous work has explored the modeling of these properties with structured data and graph neural networks. However, these approaches lack the generative capabilities and scale of modern LLMs. In this work, we introduce a novel approach to combine the strengths of modeling both code as text and more structured forms.