I'm a fourth-year undergraduate student at the Jerome Fisher M&T Program at the University of Pennsylvania, specializing in AI safety and transparency research. My work focuses on evaluating and understanding large language models, particularly in the areas of:
- Model evaluations and benchmarking
- Steering model behaviors
- AI transparency and interpretability
- Co-authored the most comprehensive empirical meta-analysis of AI safety benchmarks to date
- Work presented at the UK AI Safety Institute
- Research cited by OpenAI's Superalignment Fast Grants page
- Techniques incorporated into major open-source projects (llama.cpp, vLLM)
- Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress? (NeurIPS 2024 D&B)
- Representation Engineering: A Top-Down Approach to AI Transparency (200+ citations)
- Localizing Lying in Llama (NeurIPS 2023 SoLaR Workshop)
- High-Efficiency Scattering Probe Design for S-Polarized Near-Field Microscopy (Applied Physics Express, 2021)
- Validity of Machine Learning in the Quantitative Analysis of Complex Scanning Near-Field Optical Microscopy Signals Using Simulated Data (Physical Review Applied, 2021)
- 🌐 Personal Website
- 📚 Google Scholar
- 🎓 Semantic Scholar
- 🐦 X
I'm an experimentalist at heart with a quick, iterative, and empirically-driven approach. I like to focus on new potential research areas, rather than research that fits cleanly into pre-existing research areas.