ICSE 2024
Fri 12 - Sun 21 April 2024 Lisbon, Portugal

Recent advances in large language models (LLMs) significantly boost their usage in software engineering. However, training a well-performing LLM demands a substantial workforce for data collection and annotation. Moreover, training datasets may be proprietary or partially open, and the process often requires a costly GPU cluster. The intellectual property value of commercial LLMs makes them attractive targets for imitation attacks, but creating an imitation model with comparable parameters still incurs high costs. This motivates us to explore a practical and novel direction: slicing commercial black-box LLMs using medium-sized backbone models.

In this paper, we explore the feasibility of launching imitation attacks on LLMs to extract their specialized code abilities, such as “code synthesis” and “code translation.” We systematically investigate the effectiveness of launching code ability extraction attacks under different code-related tasks with various query schemes, including zero-shot, in-context, and Chain-of-Thought. We also design response checks to refine the outputs, leading to an effective imitation training process. Our results show promising outcomes, demonstrating that with a reasonable number of queries, attackers can train a medium-sized backbone model to replicate specialized code behaviors similar to the target LLMs. We summarize our findings and insights to help researchers better understand the threats posed by imitation attacks, including revealing a practical attack surface for effectively generating adversarial code examples against LLMs.

Thu 18 Apr

Displayed time zone: Lisbon change

11:00 - 12:30
Language Models and Generated Code 2Demonstrations / Research Track at Maria Helena Vieira da Silva
Chair(s): Reyhaneh Jabbarvand University of Illinois at Urbana-Champaign
11:00
15m
Talk
Exploring the Potential of ChatGPT in Automated Code Refinement: An Empirical Study
Research Track
Qi Guo Tianjin University, China, Junming Cao Fudan University, Xiaofei Xie Singapore Management University, Shangqing Liu Nanyang Technological University, Xiaohong Li Tianjin University, Bihuan Chen Fudan University, Xin Peng Fudan University
11:15
15m
Talk
Deep Learning or Classical Machine Learning? An Empirical Study on Log-Based Anomaly Detection
Research Track
BoXi Yu The Chinese University of Hong Kong, Shenzhen, Jiayi Yao The Chinese University of Hong Kong, Shenzhen, Qiuai Fu Huawei Cloud Computing Technologies CO., LTD., Zhiqing Zhong Chinese University of Hong Kong, Shenzhen, Haotian Xie The Chinese University of Hong Kong, Shenzhen, Yaoliang Wu Huawei Cloud Computing Technologies Co., Ltd., Yuchi Ma Huawei Cloud Computing Technologies CO., LTD., Pinjia He Chinese University of Hong Kong, Shenzhen
11:30
15m
Talk
TRACED: Execution-aware Pre-training for Source Code
Research Track
Yangruibo Ding Columbia University, Benjamin Steenhoek Iowa State University, Kexin Pei The University of Chicago, Gail Kaiser Columbia University, Wei Le Iowa State University, Baishakhi Ray AWS AI Labs
11:45
15m
Talk
On Extracting Specialized Code Abilities from Large Language Models: A Feasibility Study
Research Track
Li Zongjie Hong Kong University of Science and Technology, Chaozheng Wang The Chinese University of Hong Kong, Pingchuan Ma HKUST, Chaowei Liu National University of Singapore, Shuai Wang The Hong Kong University of Science and Technology, Daoyuan Wu Nanyang Technological University, Cuiyun Gao Harbin Institute of Technology, Yang Liu Nanyang Technological University
12:00
15m
Talk
When Neural Code Completion Models Size up the Situation: Attaining Cheaper and Faster Completion through Dynamic Model Inference
Research Track
Zhensu Sun Singapore Management University, Xiaoning Du Monash University, Australia, Fu Song State Key Laboratory of Computer Science and Institute of Software, Chinese Academy of Sciences., Shangwen Wang National University of Defense Technology, Li Li Beihang University
Pre-print
12:15
7m
Talk
TestSpark: IntelliJ IDEA’s Ultimate Test Generation Companion
Demonstrations
Arkadii Sapozhnikov JetBrains Research, Mitchell Olsthoorn Delft University of Technology, Annibale Panichella Delft University of Technology, Vladimir Kovalenko JetBrains Research, Pouria Derakhshanfar JetBrains Research