Excited to share with you an exciting project with Jacky Yip UW-Madison Physics and Charles Arnal and François Charton Meta where we develop a transformer model to automate the generation of Calabi-Yau manifolds, extra dimensional spaces in string theory: arxiv.org/abs/2507.03732

Black-box Optimization for LLM Post-Training 💪 Strong non-vacuous generalization bounds ✔️ Privacy by design ✔️ Robustness to poisoning and data extraction ✔️ Improvement on reasoning benchmarks ✔️ AI at Meta NYU Center for Data Science (1/8)

LLMs lose diversity after RL post-training, and this hurts test-time scaling & creativity. Why does this collapse happen, and how can we fix it? Our new work introduces: 🔍 RL as Sampling (analysis) 🗺️ Outcome-based Exploration (intervention) [1/n]