影响力指数
论文质量、代表作、近期表现、广度与样本量置信度综合计算
30.82/100
前 23.5%
全站排名 #15,161
发表论文7 篇
平均评分
年均产出3.5 篇/年
Hrayr Harutyunyan
研究方向
large language models · long context · efficient transformers · domain generalization · knowledge distillation · semi-supervised learning · self-supervised learning · representation learning · information theory · deep learning · unsupervised learning
20
Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation
NeurIPS 2025Poster
29
Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRA
ICLR 2025Poster
三作26
A Little Help Goes a Long Way: Efficient LLM Training by Leveraging Small LMs
ICLR 2025Rejected
14
In-context learning in presence of spurious correlations
ICLR 2025Rejected
一作22
Mimetic Initialization Helps State Space Models Learn to Recall
ICLR 2025Rejected
二作