Barbara Plank
~Barbara_Plank2
5
论文总数
2.5
年均投稿
平均评分
接收情况4/5
会议分布
COLM
2
ICLR
2
NeurIPS
1
发表论文 (5 篇)
20253 篇
-
Circuit Compositions: Exploring Modular Structures in Transformer-Based Language Models
ICLR 2025desk_rejected
4
Surgical, Cheap, and Flexible: Mitigating False Refusal in Language Models via Single Vector Ablation
ICLR 2025Poster
4
Refusal Direction is Universal Across Safety-Aligned Languages
NeurIPS 2025Poster