Tina Behnia

PhD student, University of British Columbia

Joined

May 2022

Names

Tina Behnia (Preferred)

Emails

****@ece.ubc.ca (Confirmed)

Personal Links

Google Scholar

DBLP

Career & Education History

PhD student

University of British Columbia (ubc.ca)

2023 – Present

MS student

University of British Columbia (ubc.ca)

2021 – 2023

Advisors, Relations & Conflicts

PhD Advisor

Christos Thrampoulidis

2021 – Present

Expertise

Optimization

2021 – Present

Machine learning theory

2021 – Present

Deep learning

2021 – Present

Publications

Facts in Stats: Impacts of Pretraining Diversity on Language Model Generalization
Tina Behnia, Puneesh Deora, Christos Thrampoulidis
- Submitted to ICLR 2026
- Readers: Everyone
Stats or Facts: Decomposing Generalization in Language Models with Small-Scale Models
Tina Behnia, Puneesh Deora, Christos Thrampoulidis
- MOSS@ICML2025 Oral
- Readers: Everyone
In-Context Occam’s Razor: How Transformers Prefer Simpler Hypotheses on the Fly
Puneesh Deora, Bhavya Vasudeva, Tina Behnia, Christos Thrampoulidis
- MOSS@ICML2025 Oral
- Readers: Everyone
In-Context Occam’s Razor: How Transformers Prefer Simpler Hypotheses on the Fly
Puneesh Deora, Bhavya Vasudeva, Tina Behnia, Christos Thrampoulidis
- COLM 2025
- Readers: Everyone
Can Transformers Learn Tasks of Varying Complexity In-context?
Puneesh Deora, Bhavya Vasudeva, Tina Behnia, Christos Thrampoulidis
- SCSL @ ICLR 2025
- Readers: Everyone
Implicit Geometry of Next-token Prediction: From Language Sparsity Patterns to Model Representations
Yize Zhao, Tina Behnia, Vala Vakilian, Christos Thrampoulidis
- CPAL 2025 (Recent Spotlight Track)
- Readers: Everyone
On how to avoid exacerbating spurious correlations when models are overparameterized
Tina Behnia, Ke Wang, Christos Thrampoulidis
- 2022 IEEE International Symposium on Information Theory (ISIT)
- Readers: Everyone
Implicit Geometry of Next-token Prediction: From Language Sparsity Patterns to Model Representations
Yize Zhao, Tina Behnia, Vala Vakilian, Christos Thrampoulidis
- COLM
- Readers: Everyone
Implicit Geometry of Next-token Prediction: From Language Sparsity Patterns to Model Representations
Yize Zhao, Tina Behnia, Vala Vakilian, Christos Thrampoulidis
- CoRR 2024
- Readers: Everyone
Symmetric Neural-Collapse Representations with Supervised Contrastive Loss: The Impact of ReLU and Batching
Ganesh Ramachandra Kini, Vala Vakilian, Tina Behnia, Jaidev Gill, Christos Thrampoulidis
- ICLR 2024
- Readers: Everyone

View all 22 publications

Tina Behnia

Names

Emails

Personal Links

Career & Education History

Advisors, Relations & Conflicts

Expertise

Publications

Co-Authors