paper-with-me

Papers

DebugBench: Evaluating Debugging Capability of Large Language Models

2024-01-09 · Runchu Tian, Yining Ye, Yujia Qin, Xin Cong, Yankai Lin, Yinxu Pan, Yesai Wu, Haotian Hui, Weichuan Liu, Zhiyuan Liu, Maosong Sun

Large Language Models (LLMs) have demonstrated exceptional coding capability. However, as another critical component of programming proficiency, the debugging capability of LLMs remains relatively unexplored. Previous evaluations of LLMs' debugging ability are significantly limited by the risk of data leakage, the scale of the dataset, and the variety of tested bugs. To overcome these deficiencies, we introduce `DebugBench', an LLM debugging benchmark consisting of 4,253 instances. It covers four major bug categories and 18 minor types in C++, Java, and Python. To construct DebugBench, we collect code snippets from the LeetCode community, implant bugs into source data with GPT-4, and assure rigorous quality checks. We evaluate two commercial and four open-source models in a zero-shot scenario. We find that (1) while closed-source models exhibit inferior debugging performance compared to humans, open-source models relatively lower pass rate scores; (2) the complexity of debugging notably fluctuates depending on the bug category; (3) incorporating runtime feedback has a clear impact on debugging performance which is not always helpful. As an extension, we also compare LLM debugging and code generation, revealing a strong correlation between them for closed-source models. These findings will benefit the development of LLMs in debugging.

📄 PDF Abstract BibTeX arXiv:2401.04621

Code (1)

thunlp/debugbench 공식 구현

Tasks

Code Generation

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Adam 설명 없음
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Residual Connection 설명 없음
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…

Similar Papers 제목 키워드 기반

MdEval: Massively Multilingual Code Debugging

2024-11-04 · Shukai Liu, Linzheng Chai, Jian Yang, Jiajun Shi 외

Code large language models (LLMs) have made significant progress in code debugging by directly generating the correct code based on the buggy code snippet. Programming benchmarks, typically consisting of buggy code snipp…

Program Repair

COAST: Enhancing the Code Debugging Ability of LLMs through Communicative Agent Based Data Synthesis

2024-08-09 · Weiqing Yang, Hanbin Wang, Zhenghao Liu, Xinze Li 외

Code debugging is a vital stage of software development, essential for ensuring the reliability and performance of Large Language Models (LLMs) in the code generation task. Human debugging typically follows a multi-stage…

Code GenerationCode Repair

Dual-Process Scaffold Reasoning for Enhancing LLM Code Debugging

2025-11-11 · Po-Chung Hsieh, Chin-Po Chen, Jeng-Lin Li, Ming-Ching Chang arxiv

Recent LLMs have demonstrated sophisticated problem-solving capabilities on various benchmarks through advanced reasoning algorithms. However, the key research question of identifying reasoning steps that balance complex…

Computational Efficiency

The Debugging Decay Index: Rethinking Debugging Strategies for Code LLMs

2025-06-23 · Muntasir Adnan, Carlos C. N. Kuhn

The effectiveness of AI debugging follows a predictable exponential decay pattern; most models lose 60-80% of their debugging capability within just 2-3 attempts, despite iterative debugging being a critical capability f…

Code Generation

CodeEditorBench: Evaluating Code Editing Capability of Large Language Models

2024-04-04 · Jiawei Guo, Ziming Li, Xueling Liu, Kaijing Ma 외

Large Language Models (LLMs) for code are rapidly evolving, with code editing emerging as a critical capability. We introduce CodeEditorBench, an evaluation framework designed to rigorously assess the performance of LLMs…

Code Generation