Summary of Large Language Models As Code Executors: An Exploratory Study, by Chenyang Lyu et al.
Large Language Models as Code Executors: An Exploratory Study
by Chenyang Lyu, Lecheng Yan, Rui Xing, Wenxi Li, Younes Samih, Tianbo Ji, Longyue Wang
First submitted to arxiv on: 9 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 The paper pioneers the exploration of Large Language Models (LLMs) as code executors, where code snippets are directly fed to the models for execution and outputs are returned. The study comprehensively examines this feasibility across various LLMs, including OpenAI’s o1, GPT-4o, GPT-3.5, DeepSeek, and Qwen-Coder. Notably, the o1 model achieved over 90% accuracy in code execution, while others demonstrated lower accuracy levels. To improve the performance of weaker models, an Iterative Instruction Prompting (IIP) technique is introduced that processes code snippets line by line. This resulted in an average improvement of 7.22% with a highest improvement of 18.96%. The study highlights the transformative potential of LLMs in coding and lays the groundwork for future advancements in automated programming. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper explores how Large Language Models (LLMs) can execute code snippets to get the output. Researchers tested different LLMs, including o1, GPT-4o, and others, to see if they could correctly run short pieces of code. They found that some models did much better than others – for example, the o1 model got over 90% of the code right! To help weaker models do better, the researchers developed a new technique called Iterative Instruction Prompting (IIP) that looks at each line of code separately. This made a big difference, with an average improvement of 7%. The study shows how LLMs could be used to make coding easier and more efficient in the future. |
Keywords
» Artificial intelligence » Gpt » Prompting
