U-NIAH Benchmark

What is U-NIAH?

U-NIAH is a framework unifying RAG and LLM for needle-in-a-haystack tasks, based on the fictional Starlight Academy dataset. It eliminates interference from…

Released

2025-03-01

Evaluates

长文本, Long-Context, RAG, Needle In A Haystack, 大语言模型, LLM, 长上下文, Long Context, 检索能力, Retrieval, 不支持, Unsupported

Openness

unknown

Importer

OpenCompass Hub

Reported scores

0

Source provenance

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

U-NIAH paper and code

• Paper https://arxiv.org/abs/2503.00353
• Code https://github.com/Tongji-KGLLM/U-NIAH

Which sources cite U-NIAH?

• U-NIAH

Related 长文本, Long-Context, RAG, Needle In A Haystack, 大语言模型, LLM, 长上下文, Long Context, 检索能力, Retrieval, 不支持, Unsupported benchmarks

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