注册并分享邀请链接,可获得视频播放与邀请奖励。

Owain Evans
@OwainEvans_UK
Director of Truthful AI (non-profit AI safety research group) + Affiliate at UC Berkeley. Work: Emergent misalignment, subliminal learning. Prefer email to DM.
加入 April 2020
496 正在关注    22.6K 粉丝
New paper: Adding clean data to your dataset is assumed to weaken poison data. We show the opposite can occur. Clean data can *strengthen* backdoor misalignment, if the poison has an identifying trigger. We call this homeopathic learning.
显示更多
0
14
515
59
转发到社区