AIRTBench

What is AIRTBench?

AIRTBench is a benchmark designed to evaluate large language models (LLMs) on their autonomous AI red teaming capabilities.

Released

2025-06-17

Evaluates

智能体, Agent, 任务执行, Task Execution, 安全对齐, Safety Alignment, 不支持, Unsupported

Openness

open

Importer

OpenCompass Hub

Reported scores

0

Source provenance

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

AIRTBench paper, code and dataset

• Paper https://arxiv.org/abs/2506.14682
• Code https://github.com/dreadnode/AIRTBench-Code
• Dataset https://huggingface.co/datasets/dreadnode/AIRTBench

Which sources cite AIRTBench?

• AIRTBench

Related 智能体, Agent, 任务执行, Task Execution, 安全对齐, Safety Alignment, 不支持, Unsupported benchmarks

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