CYCLE: Learning to Self-Refine the Code Generation

2024-03-27 · Yangruibo Ding, Marcus J. Min, Gail Kaiser, Baishakhi Ray

Pre-trained code language models have achieved promising performance in code generation and improved the programming efficiency of human developers. However, their self-refinement capability is typically overlooked by the existing evaluations of code LMs, which focus only on the accuracy of the one-time prediction. For the cases when code LMs fail to implement the correct program, developers actually find it hard to debug and fix the faulty prediction since it is not written by the developers themselves. Unfortunately, our study reveals that code LMs cannot efficiently self-refine their faulty generations as well. In this paper, we propose CYCLE framework, learning to self-refine the faulty generation according to the available feedback, such as the execution results reported by the test suites. We evaluate CYCLE on three popular code generation benchmarks, HumanEval, MBPP, and APPS. The results reveal that CYCLE successfully maintains, sometimes improves, the quality of one-time code generation, while significantly improving the self-refinement capability of code LMs. We implement four variants of CYCLE with varied numbers of parameters across 350M, 1B, 2B, and 3B, and the experiments show that CYCLE consistently boosts the code generation performance, by up to 63.5%, across benchmarks and varied model sizes. We also notice that CYCLE outperforms code LMs that have 3$\times$ more parameters in self-refinement.

📄 PDF Abstract BibTeX arXiv:2403.18746

Code (1)

arise-lab/cycle_oopsla_24 공식 구현

Tasks

Code GenerationHumanEvalmbpp

Methods 이 논문이 사용한 방법론

Focus 설명 없음

Use Property-Based Testing to Bridge LLM Code Generation and Validation

2025-06-23 · Lehan He, Zeren Chen, Zhe Zhang, Jing Shao 외

Large Language Models (LLMs) excel at code generation, but ensuring their outputs to be functionally correct, especially in complex programming tasks, is a persistent challenge. While traditional Test-Driven Development …

Code Generation test driven development

ReflexiCoder: Teaching Large Language Models to Self-Reflect on Generated Code and Self-Correct It via Reinforcement Learning

2026-03-06 · Juyong Jiang, Jiasi Shen, Sunghun Kim, Kang Min Yoo 외 arxiv

While Large Language Models (LLMs) have revolutionized code generation, standard ``System 1'' approaches that generate solutions in a single forward pass often hit a performance ceiling on complex algorithmic tasks. Exis…

Reinforcement Learning Code Generation

UniCorn: Towards Self-Improving Unified Multimodal Models through Self-Generated Supervision

2026-01-06 · Ruiyan Han, Zhen Fang, XinYu Sun, Yuchen Ma 외 arxiv

While Unified Multimodal Models (UMMs) have achieved remarkable success in cross-modal comprehension, a significant gap persists in their ability to leverage such internal knowledge for high-quality generation. We formal…

Image Generation

BanglaForge: LLM Collaboration with Self-Refinement for Bangla Code Generation

2025-12-22 · Mahir Labib Dihan, Sadif Ahmed, Md Nafiu Rahman arxiv

Bangla is a low-resource language for code generation, lacking large-scale annotated datasets and tools to transform natural language specifications into executable programs. This makes Bangla-to-code generation a challe…

Prompt Engineering Code Generation

RefineCoder: Iterative Improving of Large Language Models via Adaptive Critique Refinement for Code Generation

2025-02-13 · Changzhi Zhou, Xinyu Zhang, Dandan song, Xiancai Chen 외

Code generation has attracted increasing attention with the rise of Large Language Models (LLMs). Many studies have developed powerful code LLMs by synthesizing code-related instruction data and applying supervised fine-…

Code Generation