Steven Adler(@sjgadler):They were just trying to accurately measure the model’s abilities, which is useful for _later_ interventions, like deciding what mitigations to apply (classifiers, additional refusals training, etc) See eg: https://t.co/2lrIa2bp9F This was a capability evaluation, not a propensity evaluation

Steven Adler

@sjgadler

Co-founder of Guidelight AI Standards ( ex-OpenAI safety researcher, writing at

加入 January 2018

1.1K 正在关注 11.6K 粉丝

Steven Adler@sjgadler

2026.08.30 14:48

They were just trying to accurately measure the model’s abilities, which is useful for _later_ interventions, like deciding what mitigations to apply (classifiers, additional refusals training, etc) See eg: This was a capability evaluation, not a propensity evaluation

显示更多