Rohan Khetan

Research

Identified Structural Bias in CMIP6 Sea Surface Temperature Simulations using Reanalysis Data. Presenting at the Michigan Science Fair. 2nd Place/NASA/NOAA/Statistical Society Awards

Climate models are important tools for predicting Earth's climate, but they are not always accurate. CMIP6 models simulate Earth's climate by solving simplified physics equations on a three-dimensional global grid. I hypothesize that CMIP6 climate models show larger errors at certain locations and under specific physical conditions because small-scale processes, such as cloud formation and ocean mixing, must be approximated in these models. This research investigates where and when CMIP6 climate models differ from observed data by comparing model output to ERA5 data, which combines real-world observations with physics-based modeling. The study focuses on sea surface temperature and analyzes monthly data over a selected ocean region. Model error is calculated by measuring the difference between CMIP6 predictions and ERA5 values. Machine learning techniques are then used to identify patterns in the conditions associated with larger model errors to uncover the specific factors associated with physical location and processes that cause climate models to deviate from observed data.

Publications

PoliticsBench: Benchmarking Political Values in Large Language Models with Multi-Turn Roleplay (Rohan Khetan, Ashna Khetan) ICML AI4GOOD (OpenReview), 2026. Click to view

While Large Language Models (LLMs) are increasingly used as primary sources of information, their potential for political bias may impact their objectivity. Existing benchmarks of LLM social bias primarily evaluate demographic stereotypes, and when political bias is measured, it is done so at a coarse level, overlooking the values that shape sociopolitical reasoning. We introduce PoliticsBench, a multi-stage roleplay benchmark for evaluating fine-grained value expression in LLMs. Across twenty evolving scenarios, models articulate tradeoffs, take positions, and make decisions under competing pressures. Across eight prominent LLMs, we show that scenario-based prompting elicits broader and more strongly expressed value profiles than direct political questions, with peak interaction stages increasing the number of strongly activated value dimensions by approximately 0.75 (out of 10 total dimensions), a statistically significant increase relative to baseline prompting (p < 0.05). In addition, commitment to a stance increases over the course of interaction, rising by approximately 1.4 points on a [0,5] scale from initial to decision stages. While responses become less robust to scenario paraphrasing in later interaction stages, inter-judge agreement remains relatively stable. Our results suggest that evaluating LLM political behavior requires moving beyond static prompts toward longer interactive settings that capture how values are applied in context.

Projects (click each one for more!)

Baseball Swing Similarity Analyzer

This project calculates how similar your swing is to a pro player's swing! Just upload two videos and you will get a similarity score!

I created this app using Google MediaPipe, which applies framepoints to videos. I compared the vectors between the different landmark points to yield a similarity score! See GitHub code here

Stock Sentiment Analyzer

This app analyzes news articles to figure out the sentiment for a company. This is useful for determining whether a stock will go up or down.

I used the NewsAPI and the OpenAI API to build this app.

LilBizFinder

This website is a haven for discovering and reviewing small businesses. As a business owner, you can add your own business!

I designed this website for the FBLA states tournament, but unfortunately, our school was unable to compete