FalseReject Benchmark

What is FalseReject?

Safety alignment approaches in large language models (LLMs) often lead to the over-refusal of benign queries, significantly diminishing their utility in…

Released

2025-05-12

Evaluates

推理, Reasoning, 安全, Safety, 大语言模型, LLM, 逻辑推理, 安全对齐, Safety Alignment, 不支持, Unsupported

Openness

unknown

Importer

OpenCompass Hub

Reported scores

0

Source provenance

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

FalseReject paper and dataset

• Paper https://arxiv.org/pdf/2505.08054
• Dataset https://huggingface.co/datasets/AmazonScience/FalseReject

Which sources cite FalseReject?

• FalseReject

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

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