✳flâneur — a map of the web's best reading

Owen Queen

owencqueen.github.io · 58 words · saved by 1 readers

Hello! I’m a computer science Ph.D. student and Knight-Hennessy Scholar at Stanford University, working on how we can build better foundation models for biomedicine. I am interested on building powerful and practical AI for scientific applications. First, my interests are in studying and enabling emergent properties in AI systems for scientific applications. This work involves training general foundation models that can perform task and data transfer across domains and under distribution shifts. Second, I want to build AI systems that can practically interact with scientists, increasing transparency and interpretability of complex reasoning systems used by AI systems to present new hypotheses. I strive to tackle the most challenging problems in biomedicine, with applications in drug discovery and personalized medicine. I grew up in Watertown, TN and earned my Bachelor’s of Science from the University of Tennessee, Knoxville (UTK) in May 2022 with majors in computer science and mathemat

ReasonOps: Operator Segmentation for LLM Reasoning Traces Daniel Lee, Owen Queen, James Zou arXiv preprint · 2026 ReasonOps is an annotation-free method that segments chain-of-thought traces into 7 recurring reasoning operators—discourse-level moves such as backtracking, inferring, and hypothesizing—that generalize across model families and benchmarks, enabling model identification, correctness estimation, and early quality assessment before a trace completes. Paper

Explore this link on the map →

related reading