Eliciting Instruction-tuned Code Language Models’ Capabilities to Utilize Auxiliary Function for Code Generation

Seonghyeon Lee, Suyeon Kim, Joonwon Jang, HeeJae Chon, Dongha Lee, Hwanjo Yu


Abstract
We study the code generation behavior of instruction-tuned models built on top of code pre-trained language models when they could access an auxiliary function to implement a function. We design several ways to provide auxiliary functions to the models by adding them to the query or providing a response prefix to incorporate the ability to utilize auxiliary functions with the instruction-following capability. Our experimental results show the effectiveness of combining the base models’ auxiliary function utilization ability with the instruction following ability. In particular, the performance of adopting our approaches with the open-sourced language models surpasses that of the recent powerful language models, i.e., gpt-4o.
Anthology ID:
2024.findings-emnlp.100
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2024
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1840–1846
Language:
URL:
https://aclanthology.org/2024.findings-emnlp.100
DOI:
Bibkey:
Cite (ACL):
Seonghyeon Lee, Suyeon Kim, Joonwon Jang, HeeJae Chon, Dongha Lee, and Hwanjo Yu. 2024. Eliciting Instruction-tuned Code Language Models’ Capabilities to Utilize Auxiliary Function for Code Generation. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 1840–1846, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Eliciting Instruction-tuned Code Language Models’ Capabilities to Utilize Auxiliary Function for Code Generation (Lee et al., Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-emnlp.100.pdf