Summary of Rlef: Grounding Code Llms in Execution Feedback with Reinforcement Learning, by Jonas Gehring et al.
RLEF: Grounding Code LLMs in Execution Feedback with Reinforcement Learning
by Jonas Gehring, Kunhao Zheng, Jade Copet, Vegard Mella, Quentin Carbonneaux, Taco Cohen, Gabriel Synnaeve
First submitted to arxiv on: 2 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents a novel approach to teaching large language models (LLMs) to utilize execution feedback in code synthesis tasks. Specifically, the authors propose an end-to-end reinforcement learning method that enables LLMs to learn from their own generations and improve iterative code development. The proposed method outperforms independent sampling on competitive programming tasks, achieving state-of-the-art results with both small and large models while reducing the number of required samples by an order of magnitude. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large language models are used as agents that solve user-specified tasks in multiple steps, keeping manual engagement to a minimum. To reliably achieve desired outcomes, these models need to ground their generations in any feedback obtained. The authors propose a method for teaching LLMs to leverage execution feedback in code synthesis, where state-of-the-art models struggle to improve code iteratively. They benchmark on competitive programming tasks and achieve new state-of-the-art results with both small and large models while reducing the number of samples required. |
Keywords
» Artificial intelligence » Reinforcement learning
