James Zou
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
AI×1NLP×1
Frequent co-authors
Research Timeline
2026
ReasonOps: Operator Segmentation for LLM Reasoning Traces
ReasonOps introduces an unsupervised method to segment and analyze the common, compositional structure of LLM reasoning traces, discovering universal reasoning operators that predict model identity and problem correctness.
Highlighted terms show continued research focus across papers