Research

I currently focus on various aspects of model scaling (architecture, optimization, evals, etc.) across our stack; my team builds end-to-end scaling laws and studies interactions between different training paradigms to improve holistic model training.

Below, I summarize (public) aspects of my day-to-day work and previous research. My Google Scholar is linked here.

Publications¹

At OpenAI, I’ve worked on roughly every model over the last two years: o3, o4-mini, GPT-5.x, and GPT-6 models. I’ve also contributed to or coauthored some alignment papers.

Wang, Miles*, …, Jeffrey Wang, …, and Dan Mossing. “Persona Features Control Emergent Misalignment.” [Paper Link]

Prior to OpenAI, I worked on the security/privacy/fairness as well as planning/behavior/control of foundation models through the twin lenses of empirics & theory.

Li, Marvin*, Jason Wang*, Jeffrey G. Wang*, and Seth Neel. “MoPe: Model Perturbation-based Privacy Attacks on Language Models.” Empirical Methods for Natural Language Processing (2023). Also featured at the NeurIPS SoLaR Workshop. [Paper Link]

Kuchlous, Sahil*, Marvin Li*, and Jeffrey G. Wang*. “Bias Begets Bias: the Impact of Biased Embeddings on Diffusion Models.” Trustworthy Multimodal Foundation Models and Agents (TiFA) Workshop, ICML 2024. [Paper Link]

Shah, Anand*, Kehang Zhu*, Yanchen Jiang, Jeffrey G. Wang, Arif K. Dayi, John J. Horton, and David C. Parkes. “Learning from Synthetic Labs: Language Models as Auction Participants.” [Paper Link]

Wang, Jeffrey G.*, Jason Wang*, Marvin Li*, and Seth Neel. “CheckMIABench: Firm Foundations For Membership Inference Attacks on Language Models.” Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Short Paper). 2026. [Paper Link]

Bioinformatics. Even further before, I conducted bioinformatics research. I wrote an exposition of this work, building from biological and computational first principles, here; for this work, I was a national finalist of the 2021 Regeneron Science Talent Search.

Chakraborty, Abhijit*, Jeffrey G. Wang*, and Ferhat Ay. “dcHiC detects differential compartments across multiple Hi-C datasets.” Nature Communications (2022). Also featured as an oral presentation in the Regulatory and Systems Genomics Track at ISMB and as a poster at RECOMB. [Paper Link] [Open-Source Library]

A Sketch of my Research Philosophy

A unifying theme of my research is computation. I find it incredible that we can just throw more and more of the right compute toward approximating some function undergirding the mystery we wish to solve, and usually that mystery shatters.

In my opinion, the best research consists of theoretically grounded work that is empirically validated through capable engineering. In our current explosion of machine learning research, this means that I particularly value building robust systems, creating efficient implementations, and writing clean code.

In general, I take a systems approach toward epistemology. I have found that I love building knowledge by repeatedly depth-first searching into new topics, gleaning as much as I can, and building infrastructure for maintaining/revisiting this knowledge. In research, I enjoy taking approaches that synthesize techniques between several domains, and often find–in the process of DFS’ing–that hidden structure emerges between seemingly disparate areas.