CodeUltraFeedback: An LLM-as-a-Judge Dataset for Aligning Large Language Models to Coding Preferences

TL;DR

This paper introduces CodeUltraFeedback, a dataset for evaluating and improving LLMs' alignment with coding preferences using GPT-3.5 as a judge, and demonstrates how this data can enhance model alignment and correctness.

Contribution

It presents a new dataset and methodology for aligning LLMs with coding preferences, utilizing GPT-3.5 as a judge and applying this to fine-tune CodeLlama-7B-Instruct.

Findings

GPT-3.5 and GPT-4 responses are generally preferred over open-weight LLMs.

Using CodeUltraFeedback improves the alignment of CodeLlama-7B-Instruct with coding preferences.

Aligned CodeLlama-7B-Instruct outperforms larger LLMs on the HumanEval+ benchmark.

Abstract

Evaluating the alignment of large language models (LLMs) with user-defined coding preferences is a challenging endeavour that requires a deep assessment of LLMs' outputs. Existing methods and benchmarks rely primarily on automated metrics and static analysis tools, which often fail to capture the nuances of user instructions and LLM outputs. To address this gap, we propose using the LLM-as-a-Judge methodology to evaluate the alignment of LLMs with coding preferences. Based on this approach, we present CodeUltraFeedback, a comprehensive dataset designed to facilitate the evaluation and improvement of LLM alignment. CodeUltraFeedback consists of 10,000 coding instructions, each annotated with four responses generated from a diverse pool of 14 LLMs. These responses are ranked based on five distinct coding preferences using GPT-3.5 as a judge, providing both numerical scores and detailed textual feedback. Our analysis of CodeUltraFeedback reveals that responses from GPT-3.5 and GPT-4 are generally preferred over those from open-weight LLMs, highlighting significant differences in alignment between closed and open-weight models. In turn, we explore the usage of CodeUltraFeedback as feedback data to fine-tune and align CodeLlama-7B-Instruct using supervised fine-tuning (SFT) and reinforcement learning from AI feedback (RLAIF) with direct preference optimization (DPO). The resulting aligned CodeLlama-7B-Instruct model outperforms larger LLMs in terms of alignment with coding preferences and shows improved functional correctness on the HumanEval+ benchmark compared to the original instruct model. Therefore, our contributions bridge the gap in preference tuning of LLMs for code and set the stage for further advancements in model alignment and RLAIF in automated software engineering.