一句话：提出预算受限的反馈压缩机制，将判题执行的噪声输出转为紧凑可操作提示，驱动LLM多轮代码修订。
论文关注LLM生成程序时推理缺乏外部验证、单次生成易漂移或重复失败的问题，认为关键在于「有预算的反馈压缩」：在严格上下文限制下，把噪声较大的判题器输出转化为紧凑、可操作的提示，从而可靠地驱动多轮代码修订。作者在在线判题（OJ）风格的算法编程场景中实例化该思想，构建模块化交互式智能体，将LLM核心与沙箱判题器、反馈到提示的构造器以及轨迹记忆（可选错误分类器）耦合。初步实验中，基于执行的迭代在570个真实的Codeforces失败提交上把调试成功率从83.9%（单次生成）提升到93.2%，并在50道题上呈现清晰的难度趋势。作者进一步设想将预算化的「oracle到提示」压缩推广到与CI流水线、测试和性能剖析器交互的软件工程助手，并把评估重心从最终准确率转向收敛性、补丁局部性等过程指标。
Large language models (LLMs) can generate non-trivial programs, yet their reasoning often remains ungrounded: without external verification signals, one-shot generation may drift, repeat failure modes, or overfit to examples. We argue that the missing piece is budgeted feedback compression: turning noisy oracle outputs into compact, actionable hints that reliably drive multi-round code revision under tight context limits. We instantiate this idea in online judge (OJ) style algorithmic programming as a modular interactive agent that couples an LLM core with a sandboxed judger, a feedback-to-hint prompt constructor, and trajectory memory (optional error classifier). The key mechanism is feedback compression: converting noisy execution artifacts into compact, actionable hints within a tight prompt budget. In preliminary experiments, execution-grounded iteration improves debugging success from 83.9% (one-shot) to 93.2% on 570 real failed Codeforces submissions, and yields a clear difficulty trend in solving across 50 problems. Beyond OJ tasks, we envision budgeted oracle-to-hint compression as a general foundation for software engineering assistants that interact with continuous integration (CI) pipelines, tests, and profilers, shifting evaluation from final accuracy toward process metrics such as convergence and patch locality.
@inproceedings{GuZLSZ26,
title = {Context-Aware Feedback Compression in Online Judge Programming with LLMs},
author = {Jialiang Gu and Keren Zhou and Daming Li and Hanyuan Shi and Jialu Zhang},
booktitle = {FSE},
year = {2026},
doi = {10.1145/3803437.3805565}
}