Newest papers

Better Specifications, Better Code: Fine-Tuning LLMs for Behavioral Contract Generation

Brhane Team, Tiantian Yang, Nikta Akbarpour, Simone Bartucci, Francesco Calimeri, Song Wang · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.25616 · cited by 0 works (OpenAlex) · open-access PDF

Chinese

大型行为准确模型(LLMs)表现出强大的代码生成能力,但通常会因为暗示或模糊的行为要求而面临功能准确性。我们介绍了CG2,一个以合同为导向的代码生成框架,该框架将自然语言的任务转化为结构化行为准确合约,指定接口,前条件,后条件和无变性。我们评估了CG2在使用五个代码LLM:DeepSeek-Coder-V2-Lite-Instruct, Ministral-3-14B-Instructness-2512,Q22-Wen-Coder-14Instruct-Man,Q2.Coder-Instruct-Instruct,Q2.058的不变代码生成模型之前。我们评估了CG2在使用五个代码LLM的

Machine translation (Meta M2M100); the English original follows.

Abstract

Large language models (LLMs) demonstrate strong code-generation capabilities but often struggle with functional correctness due to implicit or ambiguous behavioral requirements. We present CG2, a contract-guided code generation framework that translates natural-language tasks into structured behavioral contracts specifying interfaces, preconditions, postconditions, and invariants. CG2 fine-tunes an LLM to generate these contracts and introduces an iterative semantic validation and repair process to identify and correct incomplete, inconsistent, or incorrect specifications before guiding an unchanged downstream code-generation model. We evaluate CG2 on 1,007 held-out tasks using five code LLMs: DeepSeek-Coder-V2-Lite-Instruct, Ministral-3-14B-Instruct-2512, Qwen2.5-Coder-14B-Instruct, Qwen2.5-Coder-32B-Instruct, and Qwen3.5-27B. Results show that CG2 consistently improves pass@1 over direct code generation by 1.49-3.27 percentage points (7.58%-22.45% relative improvement) and over contract guidance without fine-tuning by 1.19-2.38 percentage points (7.04%-14.65% relative improvement). Manual analysis further reveals that semantically correct contracts are associated with higher code-generation accuracy. These findings demonstrate that explicitly specifying and validating behavioral requirements can improve functional correctness without modifying downstream code-generation models.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Correctness, Computer science, Process (computing), Code (set theory), Perspective (graphical), Programming language

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange