Equational Theories: 1,200 Selected Problems for Mathematics Distillation

SAIRfoundation released this dataset of 1,200 mathematical training problems in 2026, curated from the Equational Theories Project raw implication dataset. It serves as the primary training set for Stage 1 of the Mathematics Distillation Challenge: Equational Theories competition.

Use Cases

• Training models to identify implications between mathematical laws
• Distilling mathematical reasoning for the Equational Theories competition
• Benchmarking theorem provers on curated equational logic problems

Strengths

• 1,200 curated training problems
• Derived from the Equational Theories Project
• Available in JSON format for Polars and Pandas integration

Limitations

• Small sample size relative to the full 4,694-law source dataset
• Domain-specific focus restricted to equational theories
• Lack of explicit column documentation in metadata

Provenance

Source

SAIRfoundation / Equational Theories Project

Collection Method

selected from raw implication dataset

Freshness

Last updated March 2026.