Papers
arxiv:2405.04520

NaturalCodeBench: Examining Coding Performance Mismatch on HumanEval and Natural User Prompts

Published on May 7, 2024
Authors:
,
,
,
,
,

Abstract

A new code benchmark, NaturalCodeBench, evaluates large language models on a diverse set of real-world coding tasks, highlighting gaps in current benchmarks and model performance.

Large language models (LLMs) have manifested strong ability to generate codes for productive activities. However, current benchmarks for code synthesis, such as HumanEval, MBPP, and DS-1000, are predominantly oriented towards introductory tasks on algorithm and data science, insufficiently satisfying challenging requirements prevalent in real-world coding. To fill this gap, we propose NaturalCodeBench (NCB), a challenging code benchmark designed to mirror the complexity and variety of scenarios in real coding tasks. NCB comprises 402 high-quality problems in Python and Java, meticulously selected from natural user queries from online coding services, covering 6 different domains. Noting the extraordinary difficulty in creating testing cases for real-world queries, we also introduce a semi-automated pipeline to enhance the efficiency of test case construction. Comparing with manual solutions, it achieves an efficiency increase of more than 4 times. Our systematic experiments on 39 LLMs find that performance gaps on NCB between models with close HumanEval scores could still be significant, indicating a lack of focus on practical code synthesis scenarios or over-specified optimization on HumanEval. On the other hand, even the best-performing GPT-4 is still far from satisfying on NCB. The evaluation toolkit and development set are available at https://github.com/THUDM/NaturalCodeBench.

Community

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2405.04520
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 67

Browse 67 models citing this paper

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2405.04520 in a dataset README.md to link it from this page.

Spaces citing this paper 300

Browse 300 spaces citing this paper

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.