Eliciting Instruction-tuned Code Language Models' Capabilities to Utilize Auxiliary Function for Code Generation
This addresses code generation efficiency for developers, though it appears incremental as it builds on existing instruction-following capabilities.
The researchers tackled the problem of improving code generation by enabling instruction-tuned models to use auxiliary functions, and found that their approaches allowed open-source models to outperform proprietary models like GPT-4o.
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 proprietary language models, i.e., gpt-4o.