Interpretability, Alignment, and Safety
- The Safety Tax of Cache Compression [
REPO] - Measuring Altruism, Alignment, and Social Welfare in Large Language Model Agents [
REPO] (Under Review @ NeurIPS '26) - The Spectral Geometry of Misalignment [
REPO] (Proposal) - The Depth of Deception Steering [
REPO] (Proposal) - Liar, Liar: Beyond Vocabulary Suppression [
REPO] (Proposal) - Where Reasoning Models Commit to Deception [
REPO] (Proposal)
Machine Learning
- Reinforcement Learning from Downstream Feedback [
REPO] (Proposal) - The Notion-Based Reasoning System [
REPO] (Proposal)
Mathematics
Human-Computer Interaction
Mathematics & Physics
AI/ML
- Gradient
- Open Source Machine Learning Education Platform Prototype
- Truth-grounded Digital Twins, Information Diffusion Simulations
- LLMs, Prediction Markets, and Quantitative Finance
- Lightweight C++ Neural Network
- Computer Vision & 3D Trigonometry
- Computer Vision & UAVs/F22 Raptor
- Computer Vision, Biotechnology, and ESP32/Arduino Hardware



