RLTF: Reinforcement Learning from Unit Test Feedback

By Jiate Liu, Yiqin Zhu, Kaiwen Xiao, Qiang Fu, Wei Yang, Deheng Ye. 2023

134 citations recorded by the source corpus.

The goal of program synthesis, or code generation, is to generate executable code based on given descriptions. Recently, there has been an increasing number of studies employing reinforcement learning (RL) to improve the performance of large language models (LLMs) for code. However, current representative works either rely solely on offline frameworks, limiting the exploration of new sample spaces, or fall short in the utilization of unit test signals, not accounting for specific error locations within the code. To address these issues, we propose RLTF, i.e., Reinforcement Learning from Unit Test Feedback, a novel online RL framework with unit test feedback of multi-granularity for refining code LLMs. Our approach generates data in real-time during training and simultaneously utilizes fine-grained feedback signals to guide the model towards producing higher-quality code. Extensive experiments show that RLTF achieves state-of-the-art performance on the APPS and the MBPP benchmarks. Our code is available at: https://github.com/Zyq-scut/RLTF.

Subfields: Artificial Intelligence, Information Systems, Software, Reinforcement Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Gabriel Synnaeve
2. Yossi Adi
3. Jade Copet
4. Nicolas Usunier
5. Tal Remez
6. Hugo Touvron
7. Jianfeng Gao
8. Itai Gat

In Information Systems

1. Gabriel Synnaeve
2. Nicolas Usunier
3. Wenhan Xiong
4. Baptiste Rozière
5. Thomas Scialom
6. Jason Aaron Edward Weston
7. Aliaksei Severyn

In Software

1. Gabriel Synnaeve
2. Baptiste Rozière

In Reinforcement Learning

1. Nicolas Usunier
2. Gabriel Synnaeve
3. Wenhan Xiong
4. Thomas Scialom
5. Shimon Whiteson
6. Gregory Farquhar
7. David Churchill
8. Nantas Nardelli

Citations in the corpus

• Code Llama: Open Foundation Models for Code (2023, background): “Liu et al. (2023b) showed that the inability to recall content placed in the middle of long prompts is a common failure mode in LLMs; our retrieval task is analogous to their setup, albeit tailored to code models which are not fine-tuned to follow instructions.”
• Why Do Multi-Agent LLM Systems Fail? (2025, unknown): “(Liu et al., 2023a; Qi et al., 2024; Kirchner et al., 2024).”
• UltraFeedback: Boosting Language Models with High-quality Feedback (2023, unknown): “, 2023a) to utilize preference data more efficiently, or align LLMs with various types of feedbacks (Yao et al., 2023; Wang et al., 2023c; Liu et al., 2023).”
• A Survey on Large Language Models for Code Generation (2024, unknown): “2 Overview: LLMs for Code Generations”