About me
Hello!! I’m Shreya, a fourth year PhD student at the University of Pennsylvania, where I currently work with Penn NLP and the World Well Being Project. I am advised by Lyle Ungar and Eric Wong. My research is supported by the NSF Graduate Research Fellowship.
Previously, I attended the University of Southern California (Fight On!!) where I received a B.S. in Computer Science and Applied Mathematics. During my time at USC, I worked with Morteza Dehghani at the Morality and Language Lab (formerly the Computational Social Science Lab).
I specialize in cross-cultural NLP, focusing on quantifying and mitigating Anglocentric bias in modern LLMs. I am broadly interested in helping LLMs understand subjective constructs that vary across people and cultures, such as linguistic style (e.g. emotion, politeness), implied language, and social norms.
I am currently working on: (1) Quantifying how translation fails to preserve components of linguistic style, (2) Subjectivity in implication understanding, and (3) Benchmarking cultural awareness in LLMs.
Outside of academia, I like to cook, run, and ask strangers if I can pet their dogs. Please reach out if you want to chat about my research, the NSF GRFP, or the PhD application process!
Recent News!!
- [Oct 2024] 📝 New preprint - The FIX Benchmark: Extracting Features Interpretable to eXperts
- [June 2024] 🌎 I’ll be attending NAACL to talk about Measuring cultural variation. Come say hi!
- [March 2024] 🎉 Our paper - Building Knowledge-Guided Lexica to Model Cultural Variation was selected for an oral at NAACL!
- [Feb 2024] 📚 Excited to intern at Google Deepmind this summer, working on social understanding in LLMs!
- [Dec 2023] 🌎 I’ll be attending EMNLP in Singapore and NeurIPS in New Orleans! Come say hi and chat about Cross-cultural style comparison or Visual topic modeling
- [Nov 2023] 🎉 We won the Area Chair Award (Interpretability and Analysis of Models for NLP) at AACL!!
- [Oct 2023] 📝 Our paper - Comparing Styles across Languages - was accepted to EMNLP
- [Sep 2023] 📝 Our paper - Faithful Chain-of-Thought Reasoning - was accepted to AACL
- [Aug 2023] 📝 New preprint - Human-Centered Metrics for Dialog System Evaluation
- [Jul 2023] 🌎 I am attending the 9th International Conference on Computational Social Science (IC2S2) in Copenhagen to present our work on Measuring Regional Variation in Culture Through Embedding-Based Lexica
- [Jul 2023] 🎉 We won best paper at WASSA!! A huge shoutout to my wonderful collaborators
- [Jul 2023] 🌎 I am attending ACL in Toronto to present our work on Anglocentric bias in LMs
- [May 2023] 📝 Our paper - Multilingual Language Models are not Multicultural: A Case Study in Emotion - was accepted to the 13th Workshop on Computational Approaches to Subjectivity, Sentiment & Social Media Analysis (WASSA)
- [Mar 2023] 🎉 I am beyond honored to recieve the NSF Graduate Research Fellowship!