TREAT Benchmark

What is TREAT?

TREAT evaluates LLMs' ability to recognize theorem identities from equivalence-preserving formula transformations, with 737 identities and 29,480 transformed…

Released

2026-07-29

Evaluates

Factuality, Language & Knowledge, Mathematics & Formal Science, cs.AI

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2608.07540

TREAT paper

• Paper https://arxiv.org/abs/2608.07540

Which sources cite TREAT?

• TREAT

Related Factuality, Language & Knowledge, Mathematics & Formal Science, cs.AI benchmarks

• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• AARRI-Bench
• ABC-Bench
• ActiveSciBench
• AdaMem-Bench
• AdvancedMathBench
• AdversaBench