Aditya Pola

I am a research based master's student at the Indian Institute of Technology, Hyderabad. I am advised by Prof. Vineeth NB. I'm interested in Mechanistic interpretability, VLMs, VLAs, AI safety.

Email    GitHub    Google Scholar    CV

profile photo
Research

I’m currently working on reasoning in VLMs and VLAs.

Select projects

thumbnail
thumbnail
LogicCBMs: Logic-Enhanced Concept-Based Learning
Deepika SN Vemuri, Gautham Bellamkonda, Aditya Pola, Vineeth N Balasubramanian
WACV 2026
[Paper] [Code] [Project Page]

LogicCBMs introduce differentiable logic gates into concept-based models.

thumbnail
thumbnail
Where does an LLM begin computing an instruction?
Aditya Pola, Vineeth N Balasubramanian
UniReps, NeurIPS 2025
[Paper]

At what layer does the model shift from reading the prompt to computing the instructed operation?


Template adapted from this repo.