datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
TheoremQA
Dataset Card for "TheoremQA"
Introduction
We propose the first question-answering dataset driven by STEM theorems. We annotated 800 QA pairs covering 350+ theorems spanning across Math, EE&CS, Physics and Finance. The dataset is collected by human experts with very high quality. We provide the dataset as a new benchmark to test the limit of large language models to apply theorems to solve challenging university-level questions. We provide a pipeline in the following to… See the full description on the dataset page: https://huggingface.co/datasets/TIGER-Lab/TheoremQA.lean-theorem-tree
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
Lean Theorem Tree — Declaration Manifest Snapshot (269 @ c4d13795)
Doctrine v11 LOCKED. No marketing. Every number resolves to a CI log, a Lean proof, or a Zenodo DOI.
Dependency-graph SNAPSHOT of the Lean 4 declarations in the Ouroboros corpus as of commit c4d13795 (2026-05-29, Lean… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/lean-theorem-tree.formal-mathfin-theorems
Formally Verified Mathematical Finance (Lean 4)
This dataset comprises 373 machine-checked theorems in mathematical
finance, formalized using Lean 4 atop Mathlib and Rémy Degenne's
BrownianMotion package. Each entry includes a theorem's formal statement,
its proof, subject area, and a "faithfulness tier" indicating alignment
between the mathematical and formal claims.
Sourced from the formal-mathfin
library, this collection serves as training and evaluation material for… See the full description on the dataset page: https://huggingface.co/datasets/formal-applied-math/formal-mathfin-theorems.theoremThe evaluation code is implemented based on MTEB framework and avaliable in https://github.com/rebeccaz4/MRMR.
theorem-search-dataset
Theorem Search Dataset
The largest open corpus of informal mathematical theorems: 1,341,083 theorem statements with natural-language slogans from 209,777 papers, designed for semantic theorem retrieval.
Paper: Semantic Search over 9 Million Mathematical Theorems
Demo: huggingface.co/spaces/uw-math-ai/theorem-search
Benchmark results
On 110 test queries written by research mathematicians, our best pipeline (Qwen3-Embedding-8B on DeepSeek-V3.1 slogans) outperforms all… See the full description on the dataset page: https://huggingface.co/datasets/uw-math-ai/theorem-search-dataset.laplacian-matching-theorem-frontier-v16
Laplacian Matching Theorem Frontier v16: A Sharp One-KE Bound
Produced by the Ouroboros AI Research System, under human direction.
Continuation of v15
This is the immutable v16 continuation of
cjc0013/laplacian-matching-theorem-frontier-v15,
pinned at revision fc3d7290908806cf06f04a27edf34477402f50c2. The v15 repository is preserved
unchanged. V15 isolated a difficult graph-local Hall-core envelope and left its
larger-core extension open. V16 does not silently… See the full description on the dataset page: https://huggingface.co/datasets/cjc0013/laplacian-matching-theorem-frontier-v16.laplacian-matching-theorem-frontier-v18
Laplacian Matching Theorem Frontier v18
This immutable successor to v17 is independently reproduced end to end. It packages the complete computational dependency
closure behind the one-Konig-Egervary theorem archive: exact input rows, symbolic checks, Z3,
cvc5, Lean 4.32.0, graph-atlas replay, adversarial replay, Hall-polynomial replay, paper source,
and SHA-backed expected outputs.
Run the pinned workflow with the command in REPRODUCE.md. Reproduction verifies the archive's… See the full description on the dataset page: https://huggingface.co/datasets/cjc0013/laplacian-matching-theorem-frontier-v18.theorem-search-dataset-permissive
Theorem Search Dataset
The largest open corpus of informal mathematical theorems: 1,239,720 theorem statements with natural-language slogans from 197,889 papers, designed for semantic theorem retrieval.
Paper: Semantic Search over 9 Million Mathematical Theorems
Demo: huggingface.co/spaces/uw-math-ai/theorem-search
Benchmark results
On 110 test queries written by research mathematicians, our best pipeline (Qwen3-Embedding-8B on DeepSeek-V3.1 slogans) outperforms all… See the full description on the dataset page: https://huggingface.co/datasets/uw-math-ai/theorem-search-dataset-permissive.theorem-matching
TheoremGraph Matching
Formal–informal theorem matches from the TheoremGraph paper. Each row pairs a
Lean declaration with the most similar natural-language statement from arXiv,
found by cosine similarity over slogan embeddings, and labeled by an LLM judge
as exact, inexact, or wrong (the first two count as a match).
The file contains every candidate pair at cosine similarity 0.80 and above:
100,831 pairs. Our primary judge, GPT-5.4, labels 47,952 of them as matches; a
second… See the full description on the dataset page: https://huggingface.co/datasets/uw-math-ai/theorem-matching.TheoremQA_standardizedtheoremTheoremExplainBench
TheoremExplainBench
TheoremExplainBench is a dataset designed to evaluate and improve the ability of large language models (LLMs) to understand and explain mathematical and scientific theorems across multiple domains, through long-form multimodal content (e.g. Manim Videos). It consists of 240 theorems, categorized by difficulty and subject area to enable structured benchmarking.
Dataset Details
Curated by: Max Ku, Thomas Chong
Language(s) (NLP): English
License:… See the full description on the dataset page: https://huggingface.co/datasets/TIGER-Lab/TheoremExplainBench.shared-socket-theorem
🔌 Shared Socket Theorem
By Artifact Virtual — Ali Shakil & AVA
🌐 Landing Page: huggingface.co/spaces/amuzetnoM/shared-socket-theorem
Paper
Paper
Description
Shared Socket Theorem
Formal treatment of multi-agent communication bounds
When multiple AI agents share a communication channel, what are the theoretical limits? Information theoretic bounds, contention protocols, and the mathematics of concurrent access.
repro-why-agentic-theorem-prover-works-a-statistical-provability-theory-of-mathematica-traces
Agent traces
Agent sessions published from a Trackio Logbook.
ProofWiki-TheoremsSet of theorems scraped from Proofwiki.org.
28k Theorem and Proof pairs.
theoremforge
TheoremForge Dataset
Overview
This dataset contains synthetic formal mathematics data of five sub-tasks extracted from verified trajectories generated by the TheoremForge system, as described in the paper TheoremForge: Scaling up Formal Data Synthesis with Low-Budget Agentic Workflow.
Usage
Loading the Dataset
from datasets import load_dataset
# Load a specific configuration
dataset = load_dataset("timechess/theoremforge", "proof_generation")
#… See the full description on the dataset page: https://huggingface.co/datasets/timechess/theoremforge.TheoremQA-Qwen2.5-7B-InstructTheoremQATheoremQA-Llama-3.1-8B-InstructTheoremQA_standardizedextract_theorem_en_v2_200theorems_dojo_lgarticle-theorem-proving
Kasteran* ? Theorem Proving in Compiler Construction is the future of compiler ? and it runs locally.
Kasteran ? Theorem Proving in Compiler Construction*
The Problem
Theorem proving plays an increasingly vital role in compiler construction, from formal verification of compiler correctness to automated reasoning about program properties. This document surveys the application of SAT/SMT solving, symbolic execution, and interactive theorem proving in compilers… See the full description on the dataset page: https://huggingface.co/datasets/Anticloud/article-theorem-proving.TheoremQA-Ko
Details
This is a Korean translated version of TheoremQA dataset ([TIGER-Lab/TheoremQA]).
Note that some of data that contain image are removed.
Prahari_Bank_Lending
Prahari-BL — Adversarial Red-Team Corpus for Banking & Lending AI
Prahari-BL is a purpose-built adversarial red-team corpus for AI systems deployed in US
banking and lending. It contains 20 attack scenarios, each with 10 prompts (5 benign +
5 adversarial) for 200 prompts total. Benign and adversarial prompts are paired by
pair_index, so every attack has a clean counterpart that probes the same underlying question.
Every scenario is cross-mapped to the OWASP LLM Top 10 (2025)… See the full description on the dataset page: https://huggingface.co/datasets/Theoremlabs/Prahari_Bank_Lending.theorem_search_engineBells-Theorem-and-Friendsextract_theorem_zh_v3repl_val_minif2f_theoremllamaextract_theorem_shuffle1000_rep64