Paper Detail

Scaling Mathematical Proof Supervision from Structured Textbooks

Youqing Fang, Zhouqi Hua, Yinhao Tang, Yuzhe Gu, Lingkai Kong, Bin Liu, Weiming Zhang, Kai Chen, Yanan Sun, Wenwei Zhang

huggingface Score 11.0

Published 2026-10-29 · First seen 2026-11-12

General AI

Abstract

Mathematical supervision for language models is typically flattened into isolated problem-solution pairs, discarding the curricular organization and evidential structure of the original sources. We introduce TOME-Math (Textbook-Oriented Mathematical Evidence), a large-scale corpus that converts 2,118 textbooks spanning the core areas of university-level mathematics subjects into approximately 380K structured proof-supervision records while preserving chapter order, proposition-proof links, dependency references, and source provenance. Leveraging these structural signals, we explore curriculum learning across multiple mathematical subjects and characterize scaling patterns with supervision volume and source breadth. At full scale, fine-tuning on TOME-Math raises the unweighted mean score across our held-out evaluations and established public benchmarks by 16.7 percentage points over the base model, while exceeding all three public instruction-data baselines evaluated in aggregate. Our findings highlight structured textbooks as both a scalable source of mathematical supervision and an empirical framework for studying the impact of data organization on mathematical reasoning. We will release the corpus to facilitate further research on structured mathematical supervision.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@misc{fang2026scaling,
  title = {Scaling Mathematical Proof Supervision from Structured Textbooks},
  author = {Youqing Fang and Zhouqi Hua and Yinhao Tang and Yuzhe Gu and Lingkai Kong and Bin Liu and Weiming Zhang and Kai Chen and Yanan Sun and Wenwei Zhang},
  year = {2026},
  abstract = {Mathematical supervision for language models is typically flattened into isolated problem-solution pairs, discarding the curricular organization and evidential structure of the original sources. We introduce TOME-Math (Textbook-Oriented Mathematical Evidence), a large-scale corpus that converts 2,118 textbooks spanning the core areas of university-level mathematics subjects into approximately 380K structured proof-supervision records while preserving chapter order, proposition-proof links, depen},
  url = {https://huggingface.co/papers/2610.40394},
  keywords = {language models, fine-tuning, curriculum learning, scaling patterns, instruction-data baselines, mathematical reasoning, huggingface daily},
  eprint = {2610.40394},
  archiveprefix = {arXiv},
}

Metadata

{}