Darshil Doshi

Publication Activity (10 Years)

Years Active: 2021-2024
Publications (10 Years): 7

Top Topics

Experimental Results On Real World

Belief Functions

Top Venues

Publications

Darshil Doshi, Aritra Das, Tianyu He, Andrey Gromov
To Grok or not to Grok: Disentangling Generalization and Memorization on Corrupted Algorithmic Datasets. ICLR (2024)
Darshil Doshi, Tianyu He, Aritra Das, Andrey Gromov
Grokking Modular Polynomials. CoRR (2024)
Tianyu He, Darshil Doshi, Aritra Das, Andrey Gromov
Learning to grok: Emergence of in-context learning and skill composition in modular arithmetic tasks. CoRR (2024)
Darshil Doshi, Tianyu He, Andrey Gromov
Critical Initialization of Wide and Deep Neural Networks using Partial Jacobians: General Theory and Applications. NeurIPS (2023)
Darshil Doshi, Aritra Das, Tianyu He, Andrey Gromov
To grok or not to grok: Disentangling generalization and memorization on corrupted algorithmic datasets. CoRR (2023)
Tianyu He, Darshil Doshi, Andrey Gromov
AutoInit: Automatic Initialization via Jacobian Tuning. CoRR (2022)
Darshil Doshi, Tianyu He, Andrey Gromov
Critical initialization of wide and deep neural networks through partial Jacobians: general theory and applications to LayerNorm. CoRR (2021)