大模型
arxiv.org ·
EffiPair：利用差分执行反馈提升 LLM 生成代码的效率
该论文提出差分执行反馈（DEF）：对比结构相近但性能差异明显的程序实现，把执行与实现差异转化为方向性优化证据。基于此构建 EffiPair——一种免训练、测试时框架，将结构相似但效率不同的程序配对，蒸馏其相对执行行为为紧凑反馈，并迭代精炼候选池。在 EvalPerf、Mercury、ENAMEL 三个基准上，使用 GPT-4o mini、DeepSeek-V4.1 Flash、GPT-5 mini，作者称在全部九个模型-基准设置中取得官方效率指标最高值；两轮对比精炼在保持或提升 Pass@1 的同时改善效率。
AI推荐理由 arXiv 论文（cs.PL），提出差分执行反馈（DEF）与 EffiPair 框架：通过配对结构相近但效率不同的程序，将其执行与实现差异蒸馏为方向性优化信号，在测试时迭代精炼候选池。摘要给出具体基准（EvalPerf、Mercury、ENAMEL）与模型（GPT-4o mini、DeepSeek-V4.1 Flash、GPT-5 mini），称在九个模型-基准组合中均取得官方效率指标最高值，且两轮对比精炼在保持或提升 Pass@1 的同时改善效率。属训练无关的测试时方法，对 LLM 代码生成效率优化有实际参考价值；局限是仅依据摘要，未提供代码、数据与复现细节，效果需独立验证。
78
正文
Computer Science > Programming Languages
arXiv:2604.05137 (cs)
[Submitted on 6 Apr 2026 (v1), last revised 26 Sep 2026 (this version, v2)]
Title:EffiPair: Improving the Efficiency of LLM-generated Code with Differential Execution Feedback
Authors:Samira Hajizadeh, Suman Jana
View a PDF of the paper titled EffiPair: Improving the Efficiency of LLM-generated Code with Differential Execution Feedback, by Samira Hajizadeh and 1 other authors
View PDF
HTML (experimental)
Abstract:Large language models (LLMs) can generate functionally correct programs that differ substantially in execution efficiency. Existing inference-time optimization methods typically refine each candidate using pointwise runtime or profiling feedback, which identifies how costly an implementation is or where the cost arises, but offers limited guidance on how the computation should change. We introduce DIFFERENTIAL EXECUTION FEEDBACK (DEF), which compares implementations that are nearby in structural program space but separated in performance, turning their execution and implementation differences into directional optimization evidence. We demonstrate DEF in EffiPair, a training-free, test-time framework that pairs structurally similar programs with different efficiencies, distills their relative execution behavior into compact feedback, and iteratively refines a candidate pool. Across EvalPerf, Mercury, and ENAMEL, using GPT-4o mini, DeepSeek-V4.1 Flash, and GPT-5 mini, EffiPair achieves the highest value on each benchmark's official efficiency metric in all nine model-benchmark settings under matched evaluation conditions. Moreover, two contrastive refinement rounds improve efficiency over the selected initial draft in every setting while preserving or improving Pass@1. These results demonstrate the effectiveness of relational execution feedback as a lightweight signal for test-time code optimization.
Subjects:
Programming Languages (cs.PL); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG); Software Engineering (cs.SE)
Cite as:
arXiv:2604.05137 [cs.PL]
(or
arXiv:2604.05137v2 [cs.PL] for this version)
https://doi.org/10.48550/arXiv.2604.05137
Focus to learn more
arXiv-issued DOI via DataCite
Submission history
From: Samira Hajizadeh [view email]
[v1]
Mon, 6 Apr 2026 19:56:26 UTC (649 KB)
[v2]
Sat, 26 Sep 2026 17:37:54 UTC (869 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled EffiPair: Improving the Efficiency of LLM-generated Code with Differential Execution Feedback, by Samira Hajizadeh and 1 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.PL
new
|
recent
| 2026-04
Change to browse by:
cs
cs.AI
cs.CL
cs.LG
cs.SE
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
export BibTeX citation
Loading...
BibTeX formatted citation
×
loading...
Data provided by:
Bookmark
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? |
Disable MathJax (What is MathJax?)