12 papers
Auditing and Decomposing Feedback-Driven Evolution in LLM Test Generation under the Oracle Problem
Yunhao Liang, Chengguang Gan, Ruixuan Ying +3
Execution feedback is often treated as a self-verifying signal for improving LLM-generated tests. However, when generated inputs are executed on a single accepted program and its o…
Security Tests as Executable Specifications for LLM Code Generation: Benefits, Trade-offs, and Coverage Limits
Yunhao Liang, Chengguang Gan, Ruixuan Ying +3
Large language models (LLMs) can generate functionally useful code that remains vulnerable, while security-focused interventions may break intended behavior. We investigate securit…
Do Code Language Models Use Tests? A Behavioral and Representational Study of Test-Driven Code Generation
Yunhao Liang, Chengguang Gan, Ruixuan Ying +3
The paper investigates how code language models respond to test cases in prompts, analyzing whether tests act as executable specifications or merely extra context, and finds that t…
MAG: A Web-Agent Benchmark and Harness for Multimodal Action and Guide Generation
Chengguang Gan, Hanjun Wei, Yunhao Liang +3
The paper presents MAG, a benchmark and harness that combine web‑agent action execution and guide text generation into a single multimodal task using screenshot‑based grounding, an…
A Learning-Rate-Gated Failure of GRPO in a Small Language and Vision-Language Model Web Agent: A Controlled Null and Its Mechanism
Chengguang Gan, Zhixi Cai, Yunhao Liang +3
The paper evaluates whether Group Relative Policy Optimization (GRPO) improves the performance of small (4‑8 B parameter) language and vision‑language web agents and finds that it…
A Multilingual Dataset and Empirical Validation for the Mutual Reinforcement Effect in Information Extraction
Chengguang Gan, Sunbowen Lee, Qingyu Yin +9
The Mutual Reinforcement Effect (MRE) describes a phenomenon in information extraction where word-level and sentence-level tasks can mutually improve each other when jointly modele…