The paper compares two output regimes for code-editing language models: direct generation, where the model outputs the entire modified file, and iterative diff-based generation, where the model emits a sequence of localized edits. Experiments on Flutter/Dart tasks show that direct generation consistently outperforms diff-based generation across metrics such as compilation success, token efficiency, and quality judgments. However, diff-based generation can be competitive for short, spatially localized edits, particularly in refactoring and error-handling tasks with few edit steps.
By Andrej Andrejev
arXiv:2608.21074v1 Announce Type: cross
Abstract: Large language models (LLMs) are increasingly used in research workflows and software development pipelines, yet their output remains sensitive to in...
By Erik Thureck, Robert K\"uhnen, Tim Jacobowitz
The paper introduces a multi-granularity framework for detecting AI involvement in code, distinguishing human-written from AI-generated or modified files and localizing the specific lines or blocks. It uses a frozen code language model with low-rank adapters trained on machine- and human-written data, producing per-token log-likelihood ratios that feed lightweight heads for line, block, and file-level predictions. Evaluated on thousands of mixed-authorship files, the method achieves high macro F1 scores and maintains low false-positive rates on real human code.
By Yue Xu, Xiyang Wang, Hui Liu, Bolei He
arXiv:2609.38257v1 Announce Type: cross
Abstract: Code editing requires a model to decide where to make changes, generate the new content, and preserve everything else. We study how masked diffusion...
By Xijia Tao, Ziru Liu, Shansan Gong, Jiacheng Ye, Kecheng Chen, Zirui Wu, Lin Zheng, Xinyu Fu, Rui Liu, Lingpeng Kong
arXiv:2607. 17532v1 Announce Type: cross Abstract: Developers frequently write uninformative git commit messages such as "fix" or "update stuff", degrading the value of version-control history for code review, debugging, and onboarding.
By Md Rafid Haque, Poojan Narendrabhai Patel, Meetkumar Vijaybhai Raychura
HERO is a method that improves large language models for program generation by combining offline and test‑time harness optimization. It learns reusable prompt updates from execution‑verified successes and failures, validates candidate revisions before adoption, and then refines each generated code instance using execution feedback. On four SQL, Cypher, and Python benchmarks, HERO achieves the best offline performance, raising pass@1 by 7.6 points over the initial prompt and adding another 2.0 points through test‑time refinement while improving execution efficiency.
By Mingyu Lu, Timothy Ossowski, Tristan Naumann, Hoifung Poon, Peniel Argaw
AgenticCADedit introduces a stateful, tool‑mediated approach to multimodal 3D CAD editing, transforming the process from generating a single complete program to executing a sequence of incremental, verifiable actions on a persistent CAD state. By committing each step, inspecting geometry, and selectively reverting faulty operations, the method preserves partial progress and builds upon earlier edits. Experiments across three large language models show substantial gains in validity and acceptance, with the weakest baseline model’s validity rising from 51.0% to 94.8% and a token‑cost reduction of 66.7% compared to neuralCAD‑Edit.
By Saptarshi Neil Sinha, Mika Silvan Goschke, Paul Julius K\"uhn, Arjan Kuijper, Michael Weinmann
arXiv:2606. 20512v1 Announce Type: cross Abstract: LLM-based coding agents need higher-level operational knowledge about a repository (which files house which subsystems, how to run the test suite, which workflows have historically led to wrong fixes) that does not exist in the code itself.
By Asa Shepard, Jeannie Albrecht
arXiv:2602. 18548v2 Announce Type: replace-cross Abstract: Design-to-code translates high-fidelity UI designs into executable front-end implementations, but progress remains hard to compare due to inconsistent datasets, toolchains, and evaluation protocols.
By Qiao Xu, Yipeng Yu, Chengxiao Feng, Xu Liu
The paper investigates the problem of over‑editing by large language models when repairing code, showing that even state‑of‑the‑art models like GPT‑5.5 frequently rewrite more code than necessary. Using a benchmark of 400 BigCodeBench problems with controlled AST corruptions, the authors quantify excess edits and demonstrate that a simple preservation instruction can reduce unnecessary changes and improve pass rates. They further explore training strategies, finding that reinforcement learning yields the best balance between edit fidelity and performance retention, highlighting edit fidelity as a distinct, measurable dimension of code‑repair quality.
By Tongyao Zhu, Wei Hern Lim, Min-Yen Kan
arXiv:2607. 18161v1 Announce Type: cross Abstract: Coding agents are increasingly used to accelerate code generation in many downstream tasks, such as fixing bugs, building applications, and prototyping.
By Alex Mathai, Shobini Iyer, Aleksandr Nogikh, Petros Maniatis, Franjo Ivancic, Junfeng Yang, Baishakhi Ray
arXiv:2607. 28887v1 Announce Type: cross Abstract: Large language models increasingly write and repair production code, yet evidence is mounting that their test-passing patches leave codebases harder to maintain.
By Amir M. Ebrahimi, Mohammed Mehedi Hasan, Aaditya Bhatia, Gopi Krishnan Rajbahadur, Ahmed E. Hassan