CHASE-Code Benchmark

What is CHASE-Code?

CHASE is a unified framework to synthetically generate challenging problems using LLMs without human involvement

Released

2025-02-14

Evaluates

代码, Code, 数学, Math, 大语言模型, LLM, 代码工程, 数理能力, 不支持, Unsupported

Openness

open

Importer

OpenCompass Hub

Reported scores

0

Source provenance

• Original evidence https://hub.opencompass.org.cn/dataset-detail/CHASE-Code

CHASE-Code paper, code and dataset

• Paper https://arxiv.org/pdf/2502.14678
• Code https://github.com/McGill-NLP/CHASE
• Dataset https://huggingface.co/datasets/McGill-NLP/CHASE-Code

Which sources cite CHASE-Code?

• CHASE-Code

Related 代码, Code, 数学, Math, 大语言模型, LLM, 代码工程, 数理能力, 不支持, Unsupported benchmarks

• A-Bench
• A-OKVQA
• AbsPyramid
• ActionAtlas
• Ada-LEval
• AECBench
• AFQMC
• AgentBoard