通过风格特征识别代码由哪个大模型生成，助力软件审计与安全溯源。
Code Fingerprints: Disentangled Attribution of LLM-Generated Code
- 分离代码语义与模型风格，用对比学习提取生成指纹。
- 在四种语言、四个模型的代码上准确率超90%。
- 适合做代码合规审查、漏洞溯源和模型责任追踪的人看。
大语言模型（LLM）的广泛应用改变了软件开发模式，实现了大规模自动化代码生成。但这也带来了治理、问责与合规性挑战。现有研究多关注区分机器与人类代码，而实际场景如漏洞排查、事故调查和许可证审计，更需识别具体生成模型。本文研究模型级代码溯源问题，旨在确定代码片段的来源LLM。尽管溯源困难，但训练数据、架构、对齐策略和解码机制的差异会引入模型特有的风格与结构特征，形成生成指纹。我们提出解耦代码溯源网络（DCAN），将通用语义信息与特定模型风格分离。通过对比学习目标，DCAN捕捉可区分的模型依赖信号，同时保留任务语义，实现跨模型、跨语言的多分类溯源。为系统评估，我们构建首个大规模基准数据集，包含四个主流LLM（DeepSeek、Claude、Qwen、ChatGPT）在四种语言（Python、Java、C、Go）上的生成代码。实验表明，DCAN在多种设置下表现可靠，验证了软件工程中模型级来源分析的可行性。数据集与代码已公开于 https://github.com/mtt500/DCAN。
原文摘要 · Abstract (English)
The rapid adoption of Large Language Models (LLMs) has transformed modern software development by enabling automated code generation at scale. While these systems improve productivity, they introduce new challenges for software governance, accountability, and compliance. Existing research primarily focuses on distinguishing machine-generated code from human-written code; however, many practical scenarios--such as vulnerability triage, incident investigation, and licensing audits--require identifying which LLM produced a given code snippet. In this paper, we study the problem of model-level code attribution, which aims to determine the source LLM responsible for generated code. Although attribution is challenging, differences in training data, architectures, alignment strategies, and decoding mechanisms introduce model-dependent stylistic and structural variations that serve as generative fingerprints. Leveraging this observation, we propose the Disentangled Code Attribution Network (DCAN), which separates Source-Agnostic semantic information from Source-Specific stylistic representations. Through a contrastive learning objective, DCAN isolates discriminative model-dependent signals while preserving task semantics, enabling multi-class attribution across models and programming languages. To support systematic evaluation, we construct the first large-scale benchmark dataset comprising code generated by four widely used LLMs (DeepSeek, Claude, Qwen, and ChatGPT) across four programming languages (Python, Java, C, and Go). Experimental results demonstrate that DCAN achieves reliable attribution performance across diverse settings, highlighting the feasibility of model-level provenance analysis in software engineering contexts. The dataset and implementation are publicly available at https://github.com/mtt500/DCAN.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。