33 citations · 79 across the 31 of their papers we have counts for
36 papers
Teaching Code Agents to Ask Without Leaking Private Code
Wenjia Jiang, Jieke Shi, Zhensu Sun +9
Local coding agents can extend their reach by consulting a stronger cloud model. Such consultations, however, often expose the repository's private identifiers, file paths, and sou…
HSPR: Recovering Software-Agent Benchmark Scores via Hierarchical Posterior Inference
Wanyi Chen, Wanqin Shen, Yuanchen Shi +2
Evaluating software agents on a full benchmark is expensive. A small task subset can preserve agent rankings while yielding inaccurate scores. Historical responses provide useful s…
Diagonalization of skew-Hermitian dual quaternion matrices and -Hermitian dual quaternion matrices
Cailing Yao, Bingzhe Hou
In this article, we present the diagonalization of skew-Hermitian dual quaternion matrices in the sense of unitary equivalence. For a unit quaternion , we also obtain the diagon…
Trustworthy Runtime Error Healing in Real-World Repositories: A Benchmark and Guardrail
Gou Tan, Pengfei Chen, Zhensu Sun +9
Runtime error healing lets a crashed program continue by generating code that repairs its live runtime state. Recent work shows that LLMs can generate such healing code, but it is…
Path2Spec: Path-Aware Specification Generation via Large Language Models
Dan Huang, Zhensu Sun, Huihui Huang +4
Formal specifications are critical for program verification, comprehension, and maintenance. However, manually writing them is costly and difficult to scale. Recent studies have sh…
Talking to Itself While Coding: What Makes Comments Help Code Generation?
Dangfeng Pan, Zhensu Sun, Cenyuan Zhang +2
Large Language Models (LLMs) often generate natural-language comments while writing code, and these comments become part of the context used to generate the code that follows. Howe…
EarlyEval: Cheaper Agent Evaluation via Early Outcome Prediction
Yuling Shi, Zhensu Sun, Junsen Dong +3
Evaluating LLM agents is essential for guiding their development, yet it has grown prohibitively expensive: a single pass of a frontier model over an agentic benchmark can cost hun…
Second Thought: Reasoning in Parallel as LLM Agents Act and Observe
Zhensu Sun, Chengran Yang, Yunbo Lyu +2
LLM agents in the ReAct paradigm alternate between reasoning, acting, and observing, but deliberate reasoning is confined to the Thought phase: while the agent serializes an action…
AgentChaos: Chaos Engineering for Agent Systems via Programmatic Fault Injection
Gou Tan, Zhensu Sun, Jieke Shi +10
Agent systems rely on LLM APIs for every response, but these APIs can return server errors, truncated responses, or corrupted content that propagates through downstream agents and…
Lossless Tensor Compression as Program Synthesis
Jieke Shi, Junda He, Wenjia Jiang +11
Model checkpoints are growing in both number and size, which makes archival, transfer, and deployment increasingly costly. General-purpose compressors can reduce storage requiremen…
How Do Practitioners Build SE Agents? Insights from a Mixed-Methods Study
Yunbo Lyu, David Williams, Jieke Shi +5
The rise of Software Engineering (SE) agents, i.e., LLM-based agents that can understand large codebases and carry out engineering tasks with limited human intervention, has been m…
Are Performance-Optimization Benchmarks Reliably Measuring Coding Agents?
Zhi Chen, Zhensu Sun, Yuling Shi +2
Repository-level performance-optimization benchmarks such as GSO, SWE-Perf and SWE-fficiency evaluate coding agents by applying patches to real repositories and comparing runtime a…
RepoRescue: An Empirical Study of LLM Agents on Whole-Repository Compatibility Rescue
Zhihao Lin, Mingyi Zhou, Zhensu Sun +4
Open-source libraries and tools are widely reused, but compatibility maintenance is expensive. Once maintainers leave, useful repositories can stop working as runtimes and dependen…
To Run or Not to Run: Analyzing the Cost-Effectiveness of Code Execution in LLM-Based Program Repair
Zhihao Lin, Junhua Zhu, Mingyi Zhou +5
LLM-based agents for program repair are increasingly built on a "generate-run-revise" paradigm, iteratively executing tests to evaluate and refine patches. This execution-based app…
Executing as You Generate: Hiding Execution Latency in LLM Code Interpreters
Zhensu Sun, Zhihao Lin, Zhi Chen +4
Current LLM systems are increasingly equipped with a code interpreter that executes generated code to obtain results. This works serially: the model first generates the complete co…
Rethinking the Value of Agent-Generated Tests for LLM-Based Software Engineering Agents
Zhi Chen, Zhensu Sun, Yuling Shi +4
Large Language Model (LLM) code agents increasingly resolve repository-level issues by iteratively editing code, invoking tools, and validating candidate patches. In these workflow…
Seeing is Coding: On the Effectiveness of Vision Language Models in Code Understanding
Yuling Shi, Chaoxiang Xie, Zhensu Sun +7
Large Language Models (LLMs) have achieved remarkable success in source code understanding, yet as software systems grow in scale, computational efficiency has become a critical bo…
Autoregressive, Yet Revisable: In Decoding Revision for Secure Code Generation
Chengran Yang, Zichao Wei, Heminghao Deng +6
Large Language Model (LLM) based code generation is predominantly formulated as a strictly monotonic process, appending tokens linearly to an immutable prefix. This formulation con…
LIDL: LLM Integration Defect Localization via Knowledge Graph-Enhanced Multi-Agent Analysis
Gou Tan, Zilong He, Min Li +8
LLM-integrated software, which embeds or interacts with large language models (LLMs) as functional components, exhibits probabilistic and context-dependent behaviors that fundament…
Token Sugar: Making Source Code Sweeter for LLMs through Token-Efficient Shorthand
Zhensu Sun, Chengran Yang, Xiaoning Du +3
Large language models (LLMs) have shown exceptional performance in code generation and understanding tasks, yet their high computational costs hinder broader adoption. One importan…
Extending Fill-In-the-Middle with Instructions for Steerable Code Completion
Zhensu Sun, Chengran Yang, Chao Peng +4
Code completion models often fail when the developer's intent is under-specified in the code context. To mitigate this, developers frequently use natural language comments to clari…
The Hidden Cost of Readability: How Code Formatting Silently Consumes Your LLM Budget
Dangfeng Pan, Zhensu Sun, Cenyuan Zhang +2
Source code is usually formatted with elements like indentation and newlines to improve readability for human developers. However, these visual aids do not seem to be beneficial fo…
Think Like Human Developers: Harnessing Community Knowledge for Structured Code Reasoning
Chengran Yang, Zhensu Sun, Hong Jin Kang +2
Large Language Models (LLMs) have significantly advanced automated code generation, yet they struggle with complex coding tasks requiring multi-step logical reasoning. High-quality…
FDI: Attack Neural Code Generation Systems through User Feedback Channel
Zhensu Sun, Xiaoning Du, Xiapu Luo +3
Neural code generation systems have recently attracted increasing attention to improve developer productivity and speed up software development. Typically, these systems maintain a…