Dhruv Gupta

Pittsburgh, PA

I am a current first-year Master’s student at Carnegie Mellon University in the MIIS (Masters of Intelligent Information Systems) program within the LTI (Language Technologies Institute). I’m very passionate about building trustworthy AI, and my particular research interests currently revolve around (1) Analyzing cultural and moral biases in LLMs, (2) Improving reasoning abilities in LLMs. Some future directions that I am planning on working towards is using reinforcement learning (RL) to incorporate human feedback as to mitigate harmful behavior of language models. Always looking for interesting NLP-related projects to join!

A fun fact about me: I was born in India, but have spent 3+ years in Thailand, Kazakhstan, China and Malaysia each, where I completed my International Baccalaureate degree at the International School of Kuala Lumpur. I love learning new languages, and have spent time improving my Russian, Mandarin, French and Hindi!

selected publications

IC2S2 2024

Tackling the Challenge of Measuring Moral Appeals from Social Media Discourses: Comparing Dictionary, BERT, and Large Language Models

In International Conference on Computational Social Science 2024, 2024
IC2S2 2024

Comparing LLMs with traditional content analysis methods in measuring communication concepts

In International Conference on Computational Social Science 2024, 2024

arXiv

Evolving Domain Adaptation of Pretrained Language Models for Text Classification

Yun-Shiuan Chuang, Yi Wu, Dhruv Gupta, and 7 more authors

2023

Bib HTML

@misc{chuang2023evolving,
  title = {Evolving Domain Adaptation of Pretrained Language Models for Text Classification},
  author = {Chuang, Yun-Shiuan and Wu, Yi and Gupta, Dhruv and Uppaal, Rheeya and Kumar, Ananya and Sun, Luhang and Sreedhar, Makesh Narsimhan and Yang, Sijia and Rogers, Timothy T. and Hu, Junjie},
  year = {2023},
  eprint = {2311.09661},
  archiveprefix = {arXiv},
  primaryclass = {cs.CL},
  url = {https://arxiv.org/abs/2311.09661},
}