AIRBench-2024

What is AIRBench-2024?

AIR-Bench 2024, the first policy-aligned AI safety benchmark, structures 8 government regulations and 16 corporate policies into four security tiers, with…

Released

2024-08-05

Evaluates

安全, Safety, 大语言模型, LLM, 安全对齐, Safety Alignment, 不支持, Unsupported

Openness

restricted

Importer

OpenCompass Hub

Reported scores

0

Source provenance

• Original evidence https://hub.opencompass.org.cn/dataset-detail/AIRBench-2024

AIRBench-2024 paper, code and dataset

• Paper https://arxiv.org/abs/2407.17436
• Code https://github.com/stanford-crfm/air-bench-2024
• Dataset https://huggingface.co/datasets/allenai/WildBench

Which sources cite AIRBench-2024?

• AIRBench-2024

Related 安全, Safety, 大语言模型, LLM, 安全对齐, Safety Alignment, 不支持, Unsupported benchmarks

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