Bridging Generation and Training: A Systematic Review of Quality Issues in LLMs for Code

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Kaifeng, Zhang, Xiaojun, Cai, Peiliang, Liu, Mingwei, Wang, Yanlin, Wang, Chong, Huang, Kaifeng, Chen, Bihuan, Peng, Xin, Zheng, Zibin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914536250933248
author He, Kaifeng
Zhang, Xiaojun
Cai, Peiliang
Liu, Mingwei
Wang, Yanlin
Wang, Chong
Huang, Kaifeng
Chen, Bihuan
Peng, Xin
Zheng, Zibin
author_facet He, Kaifeng
Zhang, Xiaojun
Cai, Peiliang
Liu, Mingwei
Wang, Yanlin
Wang, Chong
Huang, Kaifeng
Chen, Bihuan
Peng, Xin
Zheng, Zibin
contents Large language models (LLMs) frequently generate defective outputs in code generation tasks, ranging from logical bugs to security vulnerabilities. While these generation failures are often treated as model-level limitations, empirical evidence increasingly traces their root causes to imperfections within the training corpora. Yet, the specific mechanisms linking training data quality issues to generated code quality issues remain largely unmapped. This paper presents a systematic literature review of 114 primary studies to investigate how training data quality issues propagate into code generation. We establish a unified taxonomy that categorizes generated code quality issues across nine dimensions and training data quality issues into code and non-code attributes. Based on this taxonomy, we formalize a causal framework detailing 18 typical propagation mapping mechanisms. Furthermore, we synthesize state-of-the-art detection and mitigation techniques across the data, model, and generation lifecycles. The reviewed literature reveals a clear methodological shift: quality assurance is transitioning from reactive, heuristic-based post-generation filtering toward proactive, data-centric governance and closed-loop repair. Finally, we identify open challenges and outline research directions for developing reliable LLMs for code through integrated data curation and continuous evaluation. Our repository is available at https://github.com/SYSUSELab/From-Data-to-Code.
format Preprint
id arxiv_https___arxiv_org_abs_2605_05267
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Bridging Generation and Training: A Systematic Review of Quality Issues in LLMs for Code
He, Kaifeng
Zhang, Xiaojun
Cai, Peiliang
Liu, Mingwei
Wang, Yanlin
Wang, Chong
Huang, Kaifeng
Chen, Bihuan
Peng, Xin
Zheng, Zibin
Software Engineering
Artificial Intelligence
Large language models (LLMs) frequently generate defective outputs in code generation tasks, ranging from logical bugs to security vulnerabilities. While these generation failures are often treated as model-level limitations, empirical evidence increasingly traces their root causes to imperfections within the training corpora. Yet, the specific mechanisms linking training data quality issues to generated code quality issues remain largely unmapped. This paper presents a systematic literature review of 114 primary studies to investigate how training data quality issues propagate into code generation. We establish a unified taxonomy that categorizes generated code quality issues across nine dimensions and training data quality issues into code and non-code attributes. Based on this taxonomy, we formalize a causal framework detailing 18 typical propagation mapping mechanisms. Furthermore, we synthesize state-of-the-art detection and mitigation techniques across the data, model, and generation lifecycles. The reviewed literature reveals a clear methodological shift: quality assurance is transitioning from reactive, heuristic-based post-generation filtering toward proactive, data-centric governance and closed-loop repair. Finally, we identify open challenges and outline research directions for developing reliable LLMs for code through integrated data curation and continuous evaluation. Our repository is available at https://github.com/SYSUSELab/From-Data-to-Code.
title Bridging Generation and Training: A Systematic Review of Quality Issues in LLMs for Code
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2605.05267