影响力指数

17.69/100

前 47.1%

全站排名 #30,310

发表论文4 篇

平均评分4.6

年均产出1.3 篇/年

Andis Draguns

Principal Researcher@Contramont Research·美国·OpenReview

研究方向

backdoors · AI safety · deep learning

AI Models Can Provably Hide Arbitrary Capabilities

ICLR 2026Withdrawn

Unprobeable Backdoors: Evading Runtime Detection in Transformers

ICLR 2026Withdrawn

Mitigating the Influence of Distractor Tasks in LMs with Prior-Aware Decoding

ICLR 2025Rejected

合作者 (6)

Andrew Gritsevskiy

Christian Schroeder de Witt

Sumeet Ramesh Motwani

Raymond Douglas

Tomáš Gavenčiak