Abhinav Rao

I am a Master’s student at Carnegie Mellon University majoring in Natural Language Processing. I am currently working in Prof. Maarten Sap’s Lab on Harmlessness and Factuality in Language Models in the context of social cultures. I am particularly interested in studying biases and toxicity in Language models and hope to extend it to multiple languages and modalities (RAI).

Prior to joining Carnegie Mellon, I spent a year as a predoctoral research fellow at Microsoft Turing, where I worked on responsible AI for Large Language Models. I was part of Dr. Monojit Choudhury’s group, where we worked on exciting problems such as evaluating the ethical reasoning capabilities of Large Language Models and Jailbreaking LLMs. As part of my product work at Turing India, I was involved in building Bing Chat, where I developed multilingual classifiers for detecting harmful content!

Before this, I had interned at Microsoft Research India with Dr. Sunayana Sitaram, on project LITMUS, where I worked with the Bing Defensive team at Microsoft’s Search Technology Center of India (STCI), to develop a multilingual query expansion pipeline.

As part of my undergraduate thesis at SpeechLab, NTU, I was advised by Prof. Chng Eng-Siong on Multilingual natural language processing. [Paper] [Code]

In my free time, I like to walk around campus and Pittsburgh. I also like learning about Languages and their origins; I am involved in a project with Monojit and Quincy Amoah on the Sumerian Language! I also closely follow the demoscene (Site’s down! Alternate link). I’ve not gotten into coding anything yet, but I’d prefer to some day in the near or far future. I usually prefer demos on the Commodore or PC-DOS, and am especially awed at size-constrained ones.