JudgeBench

What is JudgeBench?

JudgeBench is a benchmark aimed at evaluating LLM-based judges for objective correctness on challenging response pairs.

Released

2024-10-16

Evaluates

其他, Other, 大语言模型, LLM, 知识储备, Knowledge, 逻辑推理, Reasoning, 不支持, Unsupported

Openness

open

Importer

OpenCompass Hub

Reported scores

0

Source provenance

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

JudgeBench paper, code and dataset

• Paper https://arxiv.org/abs/2410.12784
• Code https://github.com/ScalerLab/JudgeBench
• Dataset https://huggingface.co/datasets/ScalerLab/JudgeBench

Which sources cite JudgeBench?

• JudgeBench

Related 其他, Other, 大语言模型, LLM, 知识储备, Knowledge, 逻辑推理, Reasoning, 不支持, Unsupported benchmarks

• A-Bench
• A-OKVQA
• AbsPyramid
• ActionAtlas
• ActiveFly-Bench
• Ada-LEval
• AdvancedMathBench
• AECBench