arXiv subjects

Qingyu Zhang

Publications and source records attributed to Qingyu Zhang.

22 records · Page 2Linked to original sources

arXiv2024

CONGRA: Benchmarking Automatic Conflict Resolution

Resolving conflicts from merging different software versions is a challenging task. To reduce the overhead of manual merging, researchers develop various program analysis-based tools which only solve specific types of conflicts and have a limited scope of application. With the development of language models, researchers treat conflict code as text, which theoretically allows for addressing almost all types of conflicts. However, the absence of effective conflict difficulty grading methods hinders a comprehensive evaluation of large language models (LLMs), making it difficult to gain a deeper understanding of their limitations. Furthermore, there is a notable lack of large-scale open benchmarks for evaluating the performance of LLMs in automatic conflict resolution. To address these issues, we introduce ConGra, a CONflict-GRAded benchmarking scheme designed to evaluate the performance of software merging tools under varying complexity conflict scenarios. We propose a novel approach to classify conflicts based on code operations and use it to build a large-scale evaluation dataset based on 44,948 conflicts from 34 real-world projects. We evaluate state-of-the-art LLMs on conflict resolution tasks using this dataset. By employing the dataset, we assess the performance of multiple state-of-the-art LLMs and code LLMs, ultimately uncovering two counterintuitive yet insightful phenomena. ConGra will be released at https://github.com/HKU-System-Security-Lab/ConGra.

cs.SE↗

arXiv2024

WizardMerge -- Save Us From Merging Without Any Clues

Modern software development necessitates efficient version-oriented collaboration among developers. While Git is the most popular version control system, it generates unsatisfactory version merging results due to textual-based workflow, leading to potentially unexpected results in the merged version of the project. Although numerous merging tools have been proposed for improving merge results, developers remain struggling to resolve the conflicts and fix incorrectly modified code without clues. We present WizardMerge, an auxiliary tool that leverages merging results from Git to retrieve code block dependency on text and LLVM-IR level and provide suggestions for developers to resolve errors introduced by textual merging. Through the evaluation, we subjected WizardMerge to testing on 227 conflicts within five large-scale projects. The outcomes demonstrate that WizardMerge diminishes conflict merging time costs, achieving a 23.85% reduction. Beyond addressing conflicts, WizardMerge provides merging suggestions for over 70% of the code blocks potentially affected by the conflicts. Notably, WizardMerge exhibits the capability to identify conflict-unrelated code blocks that require manual intervention yet are harmfully applied by Git during the merging.

cs.SE↗

arXiv2024

Base of RoPE Bounds Context Length

Position embedding is a core component of current Large Language Models (LLMs). Rotary position embedding (RoPE), a technique that encodes the position information with a rotation matrix, has been the de facto choice for position embedding in many LLMs, such as the Llama series. RoPE has been further utilized to extend long context capability, which is roughly based on adjusting the \textit{base} parameter of RoPE to mitigate out-of-distribution (OOD) problems in position embedding. However, in this paper, we find that LLMs may obtain a superficial long-context ability based on the OOD theory. We revisit the role of RoPE in LLMs and propose a novel property of long-term decay, we derive that the \textit{base of RoPE bounds context length}: there is an absolute lower bound for the base value to obtain certain context length capability. Our work reveals the relationship between context length and RoPE base both theoretically and empirically, which may shed light on future long context training.

cs.CL↗

arXiv2022

MDIA: A Benchmark for Multilingual Dialogue Generation in 46 Languages

Owing to the lack of corpora for low-resource languages, current works on dialogue generation have mainly focused on English. In this paper, we present mDIA, the first large-scale multilingual benchmark for dialogue generation across low- to high-resource languages. It covers real-life conversations in 46 languages across 19 language families. We present baseline results obtained by fine-tuning the multilingual, non-dialogue-focused pre-trained model mT5 as well as English-centric, dialogue-focused pre-trained chatbot DialoGPT. The results show that mT5-based models perform better on sacreBLEU and BertScore but worse on diversity. Even though promising results are found in few-shot and zero-shot scenarios, there is a large gap between the generation quality in English and other languages. We hope that the release of mDIA could encourage more works on multilingual dialogue generation to promote language diversity.

cs.CL↗