UK AISI 与 EvalEval 合作开展可重复评估报告
英国人工智能安全研究所（AISI）正在使用EvalEval的基础设施公开分享评估结果，支持更加可重复和可验证的评估科学。
01
证据
- HHugging Face Blog公司披露2026-09-22
Why reproducible evaluation reporting matters What AISI is sharing Contribute to the shared mission About the EvalEval Coalition About the UK AI Security Institute Further reading The EvalEval Coalition is thrilled to share that the UK AI Security Institute (AISI) is using EvalE…
- HHugging Face Blog公司披露2026-09-22
Why reproducible evaluation reporting matters What AISI is sharing Contribute to the shared mission About the EvalEval Coalition About the UK AI Security Institute Further reading The EvalEval Coalition is thrilled to share that the UK AI Security Institute (AISI) is using EvalE…
- HHugging Face Blog公司披露2026-09-22
This builds naturally on AISI's work to make evaluation more efficient through OptStop , more statistically rigorous through HiBayES , and more standardised in areas including transcript analysis and capability elicitation. Together, AISI and EvalEval are working to diagnose gap…