Let be a finite Borel measure on , and let denote its singular part with respect to Lebesgue measure in the Lebesgue decomposition . Prove that there exists a dense set such that and .
Graduate coursework, qualifying-exam and olympiad-level mathematics: proofs, constructions and calculations that still challenge today's strongest models, each with a complete, rigorous solution.
What makes it worth training on.
Kept only when the problem stays hard for current frontier models.
Graduate & competition math: 1M+ problems →Every solution is checked by independent SOTA models before delivery.
98%+ sampled accuracy · 97%+ customer-validated →Newly composed problems, not copies of textbooks or contest archives.
Similarity-deduplicated against sources and public corpora →Inspect the data.
Real items from this line, shown exactly as delivered. Switch to JSON for the raw record, or download the public sample file.
See explanation
By singularity of with respect to , choose a Borel set such that and (if , take ). For each , by outer regularity of Lebesgue measure, choose a relatively open set containing with ; if , take . Also choose a relatively open dense set with , for example a union of sufficiently small open intervals around a countable dense subset of . Set . Then each is relatively open and dense in , and . Let . By the Baire category theorem, is dense in and is a set. Since for every , , so , hence equality. Finally, for every , so .
{
"id": "6a87b547d2762196b6bf3514",
"subject": "Mathematics",
"level": "Graduate",
"qtype": "Proof",
"question": "Let $\\mu$ be a finite Borel measure on $[0,1]$, and let $\\mu_s$ denote its singular part with respect to Lebesgue measure $m$ in the Lebesgue decomposition $\\mu=\\mu_{ac}+\\mu_s$. Prove that there exists a dense $G_\\delta$ set $G\\subseteq[0,1]$ such that $m(G)=0$ and $\\mu_s(G)=\\mu_s([0,1])$.",
"options": null,
"answer": "See explanation",
"solution": "By singularity of $\\mu_s$ with respect to $m$, choose a Borel set $A\\subseteq[0,1]$ such that $m(A)=0$ and $\\mu_s(A)=\\mu_s([0,1])$ (if $\\mu_s=0$, take $A=\\emptyset$). For each $n$, by outer regularity of Lebesgue measure, choose a relatively open set $U_n\\subseteq[0,1]$ containing $A$ with $m(U_n)<1/(3n)$; if $A=\\emptyset$, take $U_n=\\emptyset$. Also choose a relatively open dense set $W_n\\subseteq[0,1]$ with $m(W_n)<1/(3n)$, for example a union of sufficiently small open intervals around a countable dense subset of $[0,1]$. Set $V_n=U_n\\cup W_n$. Then each $V_n$ is relatively open and dense in $[0,1]$, and $m(V_n)\\le m(U_n)+m(W_n)<2/(3n)$. Let $G=\\bigcap_{n=1}^\\infty V_n$. By the Baire category theorem, $G$ is dense in $[0,1]$ and is a $G_\\delta$ set. Since $A\\subseteq U_n\\subseteq V_n$ for every $n$, $A\\subseteq G$, so $\\mu_s(G)\\ge \\mu_s(A)=\\mu_s([0,1])$, hence equality. Finally, $m(G)\\le m(V_n)<2/(3n)$ for every $n$, so $m(G)=0$."
}How it is sliced.
Levels
Graduate coursework and qualifying-exam level; olympiad and elite-competition level.
Disciplines
Analysis, measure theory, algebra, number theory, combinatorics, geometry, inequalities, probability, plus mathematical physics.
Question types
Proofs, calculations, constructions and short answers. Proofs ship with complete arguments.
Filtering
Retained only when verification confirms correctness and difficulty.
Record format.
| Field | Type | Description |
|---|---|---|
| id | string | Stable unique identifier. |
| subject | string | Discipline, e.g. Mathematics, Physics, Chemistry. |
| level | string | Undergraduate · Graduate · Competition. |
| qtype | string | Proof · Calculation · Short answer · Single / multiple choice · Fill-in-the-blank · True / false. |
| question | string | Problem statement in Markdown with LaTeX. |
| options | array | null | Answer options for choice questions. |
| answer | string | Final answer, verifiable where the question type allows. |
| solution | string | Full step-by-step reasoning. |
Built for.
- Advanced reasoning RLHard problems with verifiable answers for RL on long reasoning chains.
- AI for ScienceRigorous mathematics for models that must prove, not just compute.
- Frontier evaluationHeld-out sets that probe the upper limit of mathematical reasoning.
Other linesfrom the same engine.
Request a pilot slice.
Tell us the graduate math slice you need. We will send a pilot sample cut to your specification for your own evaluation.