Scholar
Soufiane Hayou
Google Scholar ID: JBb5zekAAAAJ
Assistant Professor, Johns Hopkins
AI
Deep Learning
Hyperparameters
Scaling
Stochastic processes
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
1,218
H-index
15
i10-index
19
Publications
20
Co-authors
15
list available
Publications
9 items
Training-Free versus Training-Based Intent Classification in LLMs: Accuracy, Robustness, and Failure Modes
2026
Cited
0
The Myth of Expert Specialization in MoEs: Why Routing Reflects Geometry, Not Necessarily Domain Expertise
2026
Cited
0
$\mu$pscaling small models: Principled warm starts and hyperparameter transfer
2026
Cited
0
Learning Rate Scaling across LoRA Ranks and Transfer to Full Finetuning
2026
Cited
0
A Proof of Learning Rate Transfer under $μ$P
2025
Cited
0
PLoP: Precise LoRA Placement for Efficient Finetuning of Large Models
2025
Cited
0
Optimal Embedding Learning Rate in LLMs: The Effect of Vocabulary Size
2025
Cited
0
On the Stability of the Jacobian Matrix in Deep Neural Networks
2025
Cited
0
Load more
Co-authors
12 total
Arnaud Doucet
Google DeepMind
Bin YU
Professor of Statistics and EECS, UC Berkeley
Jean-Francois Ton
ByteDance Seed
Yee Whye Teh
Professor of Statistical Machine Learning, Oxford, Research Scientist, DeepMind
Chris Mingard
DPhil student, University of Oxford
Fadhel Ayed
Department of Statistics, University of Oxford
George (Yorgos) Deligiannidis
Professor of Statistics, University of Oxford
Eugenio Clerico
Universitat Pompeu Fabra (Barcelona)
