Mathematics Preprints
Explore the openai/math manuscripts. Understand the question, the stated result and its assumptions before opening the PDF.
Begin with the big questions.
Original, claim-aware explanations of selected papers: the problem, what the authors state, why it matters, and where the result stops.
Three details that change a theorem.
The fine print is often the mathematics. A good reading starts with what a result actually states — and what it does not.
Every is not almost every.
“For every curve” rules out exceptions in the stated class. “For almost every real number” permits a set of exceptions of measure zero. These two phrases express fundamentally different results.
A density is not a prediction.
A claim that half of quadratic twists have rank zero is about a limiting proportion, not the rank of any chosen twist. Likewise, “for sufficiently large n” need not cover the first thousand values of n.
A bound has a direction.
Showing that five colors fail on the Euclidean plane leaves six colors unresolved. Proving an upper bound on a counting function is different from exhibiting examples that meet a lower bound.
All manuscripts
Choose any card for its research context and source abstract. Guided cards also include an independently written mathematical explanation.
No matching manuscripts
Try another term or category.