Product
Help
Pricing
About
Jobs
Log in
Get started
Author
Add to collection
Jesse Hoogland
Observed name:
Hoogland, Jesse
Institutions:
Commonwealth Scientific and Industrial Research Organisation,
The University of Melbourne
Observed institutions:
Commonwealth Scientific and Industrial Research Organisation,
The University of Melbourne,
University of Oxford
OpenAlex ID:
A5104240187
Open Problems in Mechanistic Interpretability
2025
·
Sharkey, Lee
, Chughtai, Bilal
, et al.
·
Transactions on machine learning research
·
380
Loss Landscape Degeneracy and Stagewise Development in Transformers
2024
·
Hoogland, Jesse
, Wang, George
, et al.
·
Transactions on machine learning research
·
64
Differentiation and Specialization of Attention Heads via the Refined Local Learning Coefficient
2024
·
Wang, George
, Hoogland, Jesse
, et al.
·
International Conference on Learning Representations
·
37
You Are What You Eat -- AI Alignment Requires Understanding How Data Shapes Structure and Generalisation
2025
·
Lehalleur, Simon Pepin
, Hoogland, Jesse
, et al.
·
arXiv (Cornell University)
·
25
PDF
Dynamics of Transient Structure in In-Context Linear Regression Transformers
2025
·
Carroll, Liam
, Hoogland, Jesse
, et al.
·
arXiv (Cornell University)
·
17
PDF
Structural Inference: Interpreting Small Language Models with Susceptibilities
2025
·
Baker, Garrett
, Wang, George
, et al.
·
International Conference on Learning Representations
·
15
The Loss Kernel: A Geometric Probe for Deep Learning Interpretability
2025
·
Adam, Maxwell
, Furman, Zach
, et al.
·
arXiv (Cornell University)
·
10
PDF
Compressibility Measures Complexity: Minimum Description Length Meets Singular Learning Theory
2025
·
Urdshals, Einar
, Lau, Edmund
, et al.
·
Transactions on machine learning research
·
9
PDF
Influence Dynamics and Stagewise Data Attribution
2025
·
Lee, Jin Hwa
, Smith, Matthew
, et al.
·
International Conference on Learning Representations
·
7
Show more works
View as search
Works count:
11
Citation count:
580
H-index:
9
I10-index:
7
Year
Topic
Type