Tomás Dulka

According to our database, Tomás Dulka authored at least 3 papers between 2023 and 2025.

Collaborative distances:

• Dijkstra number of five.
• Erdős number of four.

2025

Feature Hedging: Correlated Features Break Narrow Sparse Autoencoders.

David Chanin, Tomás Dulka, Adrià Garriga-Alonso

CoRR, May, 2025

2024

A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders.

David Chanin, James Wilken-Smith, Tomás Dulka, Hardik Bhatnagar, Joseph Bloom

CoRR, 2024

2023

EconBERTa: Towards Robust Extraction of Named Entities in Economics.

Karim Lasri, Pedro Vitor Quinta de Castro, Mona Schirmer, Luis Eduardo San Martin, Linxi Wang, Tomás Dulka, Haaya Naushan, John Pougué-Biyong, Arianna Legovini, Samuel Fraiberger

Proceedings of the Findings of the Association for Computational Linguistics: EMNLP 2023, 2023