Hi, I am Anamika Lochab.
I'm currently pursuing my PhD in Computer Science at Purdue University under the guidance of Prof. Ruqi Zhang. Previously, I completed my Master's degree in Computer Science at Rutgers University, where I received an Outstanding Project Award. I earned my Bachelor's degree in Computer Science and Engineering from Vellore Institute of Technology, Bhopal, graduating as a Gold Medallist.
My research focuses on trustworthy and robust AI, particularly in aligning large language models (LLMs) through learning from human feedback. I also explore multimodal alignment and model uncertainty to improve generalization and safety across diverse model families.
Publications
VERA: Variational Inference Framework for Jailbreaking Large Language Models
A. Lochab*, L. Yan*, P. Pynadath*, X. Zhang, R. Zhang
NeurIPS 2025
Energy-Based Reward Models for Robust Language Model Alignment
A. Lochab, R. Zhang
Conference on Language Modeling (COLM), 2025
Cascade Reward Sampling for Efficient Decoding-Time Alignment
B. Li, Y. Wang, A. Lochab, A. Grama, R. Zhang
Conference on Language Modeling (COLM), 2025
Awards & Honors
Education
M.S. in Computer Science
Rutgers University | 2022 – 2024
- Received Outstanding Project Award
- Advisor: Prof. Yongfeng Zhang
B.Tech in Computer Science and Engineering
Vellore Institute of Technology, Bhopal | 2018 – 2022
- Gold Medallist – Graduated top of class
- Developed projects in NLP, ML, and assistive technologies