COFFEE-GYM: An Environment for Evaluating and Improving Natural Language Feedback on Erroneous Code
Abstract: This paper presents Coffee-Gym, a comprehensive RL environment for training models that provide feedback on code editing. Coffee-Gym includes two major components: (1) Coffee, a dataset containing humans' code edit traces for coding questions and human-written feedback for editing erroneous code; (2) CoffeeEval, a reward function that faithfully reflects the helpfulness of feedback by assessing the performance of the revised code in unit tests. With them, Coffee-Gym addresses the unavailability of high-quality datasets for training feedback models with RL, and provides more accurate rewards than the SOTA reward model (i.e., GPT-4). By applying Coffee-Gym, we elicit feedback models that outperform baselines in enhancing open-source code LLMs' code editing, making them comparable with closed-source LLMs. We make the dataset and the model checkpoint publicly available in https://huggingface.co/spaces/Coffee-Gym/Project-Coffee-Gym.
Paper Type: Long
Research Area: NLP Applications
Research Area Keywords: Code editing, Feedback, Code LLM, RL, PPO
Contribution Types: NLP engineering experiment, Publicly available software and/or pre-trained models, Data resources
Languages Studied: Python, English
Submission Number: 5476
Loading